Interaction between the replication origin and the initiator protein of the filamentous phage f1

./. Mol. Hid (1987) 197, 1577174

Interaction Between the Replication Origin and the Initiator Protein of the Filamentous Phage fl Binding Occurs in Two Steps David Greenstein and Kensuke Horiuchi The Rockefeller University, New York, NY 10021, U.S.A. (Received 12 September 1986, and in revised form 5 May 1987) The replication initiator protein of bacteriophage fl (gene II protein) binds to the phage origin and forms two complexes that are separable by polyacrylamide gel electrophoresis. Complex I is formed at low gene II protein concentrations, and shows protection from DNase I of about 25 base-pairs (from position + 2 to + 28 relative to the nicking site) at the center of the minimal origin sequence. Complex II is produced at higher concentrations of the protein, and has about 40 base-pairs (from - 7 to + 33) protected. On the basis of gel mobility, complex II appears to contain twice the amount of gene II protein as does complex I. The 40 base-pair sequence protected in complex II corresponds to the minimal origin sequence as determined by in-vivo analyses. The central 15 base-pair sequence (from + 6 to +20) of the minimal origin consists of two repeats in inverted orientation. This sequence, when cloned into a plasmid, can form complex I, but not complex II. We call this 15 base-pair element the core binding sequence for gene II protein. Methylation interference with the formation of complex I by the wild-type origin indicates that gene II protein contacts six guanine residues located in a symmetric configuration within the core binding sequence. Formation of complex II requires, in addition to the core binding sequence, the adjacent’ ten base-pair sequence on the right containing a third homologous repeat. A methylation interference experiment performed on complex II indicates that gene II protein interacts homologously with the three repeats. In complex II, gene II protein protects from DNase I digestion not only ten base-pairs on the right but also ten base-pairs on the left of the sequence that is protected in complex I. Footprint analyses of various deletion mutants indicate that the left-most ten base-pairs are protected regardless of their sequence. The site of nicking by gene II protein is located within this region. A model is presented for the binding reaction involving both protein-DNA and protein-protein int’eractions.

signals for termination and in-vitro nicking by the gene II protein are contained entirely within domain A (Dotto et al., 1984). Signals for initiation extend from four nucleotides before the nicking site (-4) to more than 100 nucleotides downstream, including most of domain A and all of domain B (Cleary & Ray, 1980, 1981; Dotto et al.. 1982a,b, 1984). Phage with lesions in domain B grow poorly and frequently acquire compensatory mutations that restore efficient replication (Dotto & Zinder, 1984a,b; Kim & Ray, 1985). These compensatory mutations are of two types: mutations which lead to an overproduction of the gene IT protein, and mutations in gene TI which lead to production of an altered protein. The compensatory mutations restore efficient replication by A+ K- origins (wildtype domain A, defective domain B), but not) hy A-B+ or A-B- origins.

1. Introduction The intergenic region of the F-specific, filamentous bacteriophages (fl, M13, and fd) is a multiregulatory element containing the origins for plus and minus strand DNA synthesis, the packaging signal, and a rho-dependent transcription terminator (for a review, see Zinder & Horiuchi, 1985; Baas, 1985). Considerable effort has been spent in characterizing the origin of plus-strand replication. Figure 1 diagrams the plus-strand origin, with its domains and signals. The plus-strand origin contains signals for both the initiation and termination of DNA synthesis (Horiuchi, 1980; Dotto & Horiuchi, 1981; Dotto et nl., 1982b). It can be divided into two domains, A (minimal origin sequence) and B (replication enhancer sequence) (Dotto et nl., 1984: Johnston & Ray, 1984). The 157

0 1987 Academic Press Limited

15x

Il. Greenstein, and K. Horiuch,i

(a)

Domain A

,T-T\

Domain El

[El

C T 'T:A'h T:A yA\ G.T, A c C& +zo\-A=T' A:T C=G -IOCRG T",;\" 4s; T=A G=C G=C A:T T=A-+JO +40 +70 +50 +60 +60 G.C T=A 5’/ I CACTCAACtCTATCTCG~~~TATTCTTTtGITTTATAA~GGATTTTGC~GAT, T T ~-GCGCAATTTAAAAACAATTTAGTCGAGTAAAAAATTGGTTAT~CGG~ +A0 +,:o +110 +A0 +90

(b)

Domain

A

Domain

B

I30 Termination Figure 1. The plus strand replication origin of bacteriophage fl. (a) Nucleotide sequence. This sequence is necessary and sufficient for efficient origin function (Cleary & Ray, 1980, 1981; Dotto et al., 1982a, 1984). Only the sequence of the plus strand is shown. Domains A and B refer to the minimal origin sequence and the replication enhancer sequence. respectively, as described in the text. The site of nicking by the gene II protein is indicated by an arrow. This site is defined as 0. Position + 1 corresponds to nucleotide 5781 on the fl sequence described by Hill & Petersen (1982). [I)] and [E] indicate 2 palindromic sequences found within domain A. These palindromic sequences are drawn as hairpins for illustrative purpose, and do not mean that such structures actually form. A base substitution at position + 50, found in the closely related phage M13, is indicated. The sequence data are from Beck & Zink (1981) Hill & Petersen (1982) and van Wezenbeek et al. (1980). (b) Signals contained within the plus strand origin. Brackets show the extent of domains A and B. The site of nicking by the gene II protein (0) is indicated by a vertical arrow. Palindromic sequences found within domain A, corresponding to hairpins D and E, are indicated by horizontal arrows. Horizontal lines indicate the sequence requirements for initiation and termination, respectively. The hatched region and the dotted region show the sequence requirements for gene II protein binding and nicking. respectively.

Gene II protein is a multifunctional protein that plays central roles in phage DNA replication. First, it introduces a single-strand break at a specific site on the plus strand of supercoiled replicative form (RFIt) DNA (Meyer et al., 1979). The 3’ hydroxyl end of the nick serves as the primer for initiation of plus strand DNA synthesis. The gene II protein also functions at a step beyond nicking: DNA molecules that have been nicked by gene II protein still require the gene II protein for their unwinding and replication (Geider et al., 1982). Upon completion of a round of synthesis, gene II protein cleaves and circularizes the displaced single strand (Harth et al., 1981). The gene II protein also has site-specific topoisomerase activity (Meyer & Geider, 1979b). f Abbreviations used: RF, replicative form; IPTG, isopropyl-j-n-thiogalactopyranoside; CIAP, calf intestinal alkaline phosphatase; bp, base-pair(s).

This paper deals with the interaction between gene II protein and the replication origin. Geider et al. (1982) demonstrated that the gene IT protein could remain associated with RF11 after nicking, thereby protecting the specifically nicked DNA molecule from B&31 nuclease digestion. Filterbinding experiments have shown that the gene II protein binds specifically to either restriction fragments or superhelical DNA containing the fl origin of replication (Horiuchi, 1986). Sequences required for binding supercoiled DNA and restriction fragments were identical and did not include the nicking site. The effects of salt concentration, temperature and time of incubation on the binding were also the same for restriction fragments and supercoiled DNA. Thus, the binding of gene II protein to the origin occurs regardless of superhelicity, while the subsequent hydrolysis reaction requires superhelicity.

Binding

of the fl Initiator

pUCB, a plasmid containing domain 1%. was constructed by subcloning the 170 baserpair (bp) BglIIBamHI fragment (from nucleotide $46 to nucleotide +217) from pD39 (Dotto et aZ., 1984) into the BamHI site of pUC19 (Norrander it al.. 1983). pDGll7. containing the wild-type domain A region and part of domain B, was constructed by inserting a RnmHI linker ($CGGGATCCCG-3’) at the unique AsuI site of R218. and cloning the resulting 152 bp EeoRTPRamHI fragment into pBR322 (Bolivar et al., 1977) between the EcoRI and RamHI sites. The defective origins A83 (a deletion from - 14 to + 3), pD30 (a 13 bp insertion at, + 8). pD29 (a deletion from the left end to +5) and A + 2!$ (a delet,ion from + 29 to the right end) were described by Dotto et nl. (1982a, 1984) and are shown in Fig. 2. ln order conveniently to end-label the defective origins for the footprinting experiments, some of them were sub-cloned. For A83 and pD30. an EcoRI site was constructed at position -57 (AsuI site) by insertion of an 8 bp EcoRI linker (5’-GGAATTCC-3’). The &oRI-HnmHI fragments (the Ra,mHI site was at the HpnII L/H border) containing each defective origin was cloned into pBR322 between the EcoRI site and the BamHI site. The deletion A+ 29 extends from position +29 (fl seyuence) into pBR322 to nucleotide 1568 (Sutcliffe. 1978). A+49 was subcloned by insertion of an EcoRI linker at the .4~uI site at position -57 as described above. The 182 bp E’coR’IXhoII fragment (XhoII site at position 1667 of pBR322) containing the A + 29 defective origin was cloned between the EcoRl and BamHI sites of pBR322. pMBS1. containing a I5 bp element corresponding to the center of domain A (positions +6 t,o +20). was constructed as follows. The t)wo 17-mers DAGRl (5’-A(:(‘TGGA(:TCTTGTTCC-3’) and DAGRB (5’.AGCTGGAA(IAAGAGTCC-3’) were synthesized on an Applied Biosystems model 38OA DNA synthesizer and purified (after de-protection) by gel filtration on a G75 column (90 cm x 1.2 cm diameter) with 10 mM-tri-ethylamine bicarbonate (pH 8.5) as

In the present study, employing DNase I protection (Galas & Schmitz, 1978), methylation interference (Siebenlist & Gilbert, 1980), and gel retardation (Garner & Revzin, 1981; Fried & (‘rothers, 1981) experiments, we localize more brecisely the binding site of the gene II protein. We show that the binding reaction occurs in two steps involving more than one gene IT protein molecule.

2. Materials and Methods (a) Bacteria,

phage and media

Our standard Escherichiu coli Hf& strain K38 (Lyons & Zinder, 1972) and a recA56 derivative K902 (Fulford & Model. 1984) were used for plasmid growth and phagefl propagation. K561 is a laqiQ (Muller-Hill et al., 1968) derivative of K38 (Davis et al., 1985). R218 (Boeke et al.. 1979) is an fl derivative with an EcoRI site inserted in domain B at the HaeTTT G/D border (for an fl restriction map, see Horiuchi et al.. 1978). Fortified broth was described by Zinder & Boeke (1982). Isopropyl-fl-D-thiogalactopyranosidr (II’TG) was from BoehringerMannheim Biochrmicals. (b) Origin

plasm,ids

and cloning

procedures

Rrst,riction enzymes. phage T4 DNA ligate, and svnthetic linkers were from New England Biolabs and (iollaborative Research and used as recommended. T4 polynucleotide kinase was from P-L Riochemicals. DNA polymerase I Klenow fragment was from Bethesda Research Laboratories. Calf intestinal alkaline phosphatase (CTAI’) was from Boehringer-Mannheim Biochemicals. l’lasmid and phage RF DNA were prepared according to the methods of Maniatis et al. (1982) or Zinder t Boeke (I 982).

Initiation fl Ori

J-1

pDG117 A83

Termination

Nicking

Binding

+

+

+

+

.d

+*

+

J-l-P4

-

-

:

+

-

-

-

+

-

-

-

+

+

+

+

I

-

-

-

N.D.

N.D.

-

+

pD29

d--f-/

pD30

d-cc

A+29

..I

pUCB pMBS1

159

Protein to the Origin

-0 l -60

0

.

.

.

.

.

l

+

’

+270

Figure 2. f’roprrt,ies of wild-type and defective origins. The wild-type and defective origins used here are listed at the left. Beside each is a diagram of its DNA sequence. The open bars are fl sequence. Thin lines are non-specific sequence. The chevron indicates the deletion in A83. The filled bar shows the 10 bp insertion in pD30. The nicking site is indicat.ed by a small arrowhead for those origins that are nicked by the gene II protein. The vertical arrows indicate convenient rest’riction sites described in Materials and Methods. The filled circles show the sites used for end-labeling as described in Materials and Methods. The properties of each origin are listed at the right. The asterisk indicates that pDG117 requires qualitative or quantitative changes in gene II protein production for efficient initiation (as described in the text). N.D.. not determined.

160

Il.

Greenstein

buffer. The purified 17-mers were annealed and cloned into the HindTTI site of pIJCI9. (c) End-labeling

of DNA fragments

To label the plus strand, 10 pg of fl RFJ were digested with AsuI, treated with 15 units of CIAP, and extracted with phenol and precipitated with ethanol twice. The 5’-end-labeling was carried out with [y-32P]ATP (Amersham) and T, polynucleotide kinase, essentially as described by Maniatis et al. (1982). After the kinase reaction, the DNA was precipitated with ethanol and digested with CZaI. The 315 bp origin fragment (containing all of domains A and B) was isolated from a 2% (w/v) agarose gel. The fragment was eluted electrophoretically, extracted with phenol and precipitated with ethanol twice. To label the minus strand, ,&&I-digested RF1 DNA was labeled at the 3’ end with [a-32P]dGTP and DNA polymerase I Klenow fragment. The endlabeled origin fragment was isolated as described above. The defective origins (A83, pD30 and A + 29) were labeled at the EcoRI site placed at, position -57. Again, labeling of the plus-strand was by kinase reaction at the EcoRT site, and labeling of the minus strand was by end-filling. The defective origin pD29 was labeled at the EcoRT site of pBR322. pDG117 was labeled at the BamHI site that was introduced at position -57. pI.JC!B and pMBS1 were labeled at the EcoRT site from the polylinker region of pUCl9. A diagram of the end-labeled fragments used in the binding experiments appears in Fig. 2. For the electrophoretic analysis of gene IT protein origin complexes, plasmids were digested with restriction enzymes and treated with CIAP. after which the origin fragments were isolated on 2?‘, agarosr gels. Following electroelution and precipitation with ethanol, the origin fragments were end-labeled with [Y-~~P]ATP and T4 polynucleotide kinase. linincorporated label was removed by a G50 spun column procedure (Maniatis et nE.. 1982). The pDG 117 origin fragment was the 190 bp EcoRIIBanI fragment, similar to the 152 bp fragment used in the footprinting experiments except that it contained an addit’ional 38 bp of pBR322 sequence (from the BamHI site to the Han1 sit,e at posibion 413 of pBR322). The 190 bp A+29 origin fragment was the EcoRT-&mHT fragment used in the footprinting experitnent,s (as described above and in Fig. 2). The 208 bp fragment, containing the 15 bp element from the cenber of domain A (DAGRl-DAGR2 oligonucleotides) was prepared by digestion of pMBS1 with BstNI. The control pBR322 fragment was the 187 hp EcoRI-EcoRV fragment. (d) Puri&ation

of gene II

protein

The gene II protein (M,=46,000) used in the experiments shown in Figs 4(h) and 5 was purified from K38 containing pD2 (Dotto et al.; 1981) by the procedure described by Meyer & Geider (1979a). and was approximately 9076 pure. The gene II protein used in the experiments shown in Figs 3. 4(a), 6, 7, 8 and 9 was purified to homogeneity by a new method involving a novel run-away expression vector and a purification procedure that relied on the ability of the gene II protein to be renatured after treatment with guanidine hydrochloride (our unpublished data). Briefly, an EcoRI restriction fragment (Fulford, 1986) containing gene II under control of the tac1 promoter (de Boer et al., 1983) was cloned into pDG117 at the EcoRI site. The resulting plasmid pDGl17IIA contains an fl A+B- origin and an TPTG-inducible gene II protein. IJpon induction,

and

K. Horiuch.i both gene II and gene X proteins are great,ly overproduced (approx. 30 mg/l culture) due to positive feedback: the gene IT protein produced acts at the origin. stimulating DNA replication and leading to amplification of the available templates for transcription. K561 bearing pDG1 17ITA was grown in 400 ml of fortified broth containing 100 ng ampicillin/ml at 37Y: to an A,,, nm of 0.3 (approx. I.5 x 10s to I.8 x 10” cells/ml). at which time TPTG was added to 2 mM. The cells were harvested 7 h later, and resuspended in 45 ml of huffer :\ (100 InM-IIIdeic* acid-NH+, (pH 6.8 at 23°C”). IO’?, (vi\-) glycerol, 1 m&r-EDT& 5 rn>r-P-mrrcaptoet~hanoi). The cells were sonicated (Heat Systems-Ultrasonics. Inc.. model W 185 F) until 909; of t,hr cells were broken. (‘ell breakage was monitored by counting in a PetroffHausser counting chamber. The sonicate was centrifuged at 100,OOOg for I h at 4’1’. The gene II protein was located in the pellet, which was washed in buffer A and resuspended in 17 ml of huffer A. Then 39 ml of 7 M-guanidine hydrochloride (Heico extreme purity) was added with stirrtng for 1 h at 4°C (I M final concentration), followed by rentrifugation at lOO,OOOg for 1 h at 4°C. The gene IT protein was located in the pellet. which was resuspended in 5 ml of buffer A containing 1 M-guanidine hydrochloride. Solid guanidine hydrochloride was added to a final concentration of 7 M with stirring for 1 h at 4°C’. followed hy crntrifugst,ion at lOO.OOOg at. 4°C‘ for 1 h. Then 2.5 ml of the supernatant fraction containing the gene II protein was loaded onto a Sephacryl S206 column (76 cm x I.5 cm diameter) equilibrated with 7 M-guanidine hydrochloride, 50 mM .Tris. Hf”1 (pH 7.5). 1 miv-Kl)TA. 50 m.n-/-mercaptoethanol. The column was run at room temperature (23Y”) and 3-ml fractions were collected. The gene IT protein was located very close to t,he void volume. and was slightly contaminated (lo/,) with gene X protein. Then 1 ml of the peak Sephacryl S200 fraction was loaded on a Sephacryl S400 column (90 cm x 1.0 cm diameter) prepared and run as described above. The peak fractions were diluted with an equal volume of t,he column buffer and renatured by dialysis against buffer B (25 tnhl-imidazole. HCl (pH 6.8 at 23°C‘). 10% glycerol. 400 mrvt-K(‘1. 1 m,n-Fl)TA. 5 mM-p-merraptoethanol) containing 1 iv-guanidine hydrochloride for 2 h at 4°C. followed by dialysis against huffer B for 24 h with 4 changes of the buffer. The gene TT protein obtained (200 pg) had a specific activity of X.5 x 105 to 5.0 x lo5 units/mg. which is comparable to the specific activity obtained hy Meyer 8r (ieider (1979a). Both m&hods of protein preparation gave the same footprint~ing and gel retardation result,s. One unit, is the amount needed for the relaxation of 6.25 pg of fl RF1 to RFD and RFTV in 30 min at 30°C in 20 nl of 20 mM-Tris. HCI (pH 8.0). 5 mM-Mg(:l,, 5 m&ldithiothreit,ol. 80 rnM-K(‘l. The amount of gene TI protein was t&mated from the intensit,y of the hand on silverstained (Iliray et ~1.. 1981) 12Oz;,Laemmli (1970) gels, and by the Bradford (1976) dye-binding protein assay. (e) DiVmu

I protection

experinwnts

DNase protection experiments (Galas & Schmitz, 1978) were performed in a 10 ~1 reaction containing the following: 60 fmol end-labeled restriction fragment, 5 to 40 ng (106 to 806 fmol) gene IT protein, 106 ng pBR322 80 mM-KCl, 70 mM-imidazole . HCI DNA, (pH 6.8), 5 mM-MgCl,, 5 mm-dithiothreitol, 4% glycerol. After addition of gene II protein, the mixture was incubated for 10 min at room temperature (23”(Z), after which 30 ng of DNase I ( Boehringer-Mannheim Biochemicals) in a

Binding of the fl Initiator volume of I /iI was added. Thirty seconds later, the digestion was stopped by addition of 90 ~1 of stop mix: 0.3 M-Ku'aOk (pH 5.1), 66 mM-NH,OAc, 6 mM-EDTA, 0. I q. sodium dodecyl sulfate. The DNA was then rxtract,ed with phenol, precipitated with ethanol and dissolved in 2 ~1 of 809; formamide containing bromphenol blue and xylene cyanol tracking dyes. The samples were rlrctrophoresed on standard 8% polyacrylamide sequencing gels (Maxam & Gilbert, 1977). The gels were dried, and autoradiography was carried out at - 70’(’ on Fllji RX50 film with an intensification screen. (f) Jlr~thylntion

in,terfrrance

experiment

The

met,hglation interference experiment was peraccording to the method of Siebenlist & Gilbert (1980). End-labeled re&riction fragments were methyIated with dimethyl sulfate (Aldrich) as described (Maxam 8r (Gilbert. 1977). The binding reaction contained .50 to 100 fmol of tsnd-labeled origin-containing fragment and IO ng (2(K) fmol) of gene II protein in 50 ~1 of binding (pH 8.0). 5 mi%-MgCl,. (20 m.w-‘l’ris . HCI bufle1 200 pg. Pentex bovine serum 5 mM-dithiothreit,ol. albumin/ml. 80 mu-KU). The binding reaction was carried out on ice for 2 min. The reaction mixture was then filtered through a Millipore filter HA (0.45 pm pore size), which had been pre-soaked in the binding buffer \vithout bovine serum albumin. The filt,er was washed 5 times with 0.5 ml of binding buffer without bovine serum albumin at room temperature. For elution of the bound DSA. the filters wcare placed in silironized glass vials, and I ml of rluting solution (10 mM-Tris. HCI soaked in (pH 74). ~~M-EII)TA, 0.2();, SDS) at 37°C for 30min with gentle agitation. The eluatr was precipitated with ethanol. dried briefly. and resuspended in 100 ~1 of I M-piperidinr. Thr csleavage react,ion was carried out at 90°C’ for 30 min. The samples were then lyophilized, resuspended in IO ,uI of water and lyophilized. The latter two steps were repeated 5 times. The samples containing were analyzed by equal amounts of radioactivity polyacrylamide ,)I1 standard 896 rlectrophorrsis sequencing gels. .iutoradiography was carried out as drs~ribrcl above. formed

(g) I’olyrccrylomide gel electrophoresis prokin-origin com,plaxes

of gene II

(ierIe II prot,ein--origin complexes were analyzed on 5% polvacrylamide gels (4.94(+; acrylamide, 0.06% bisacrylamide) as described (Fried & Crothers. 1981), except that huffel 20 mM-Tris . HCI the was (PH 89, Fl 0.1 I~sI-EDT~~. I mM-dithiothreitol. Electrophoresis was carried out at room temperature (23°C) at 6.5 Vcm-’ with recirculation of the buffer. The binding reactions t~~picallg cnontainrd 15 fmol of origin-containing restriction fragment and 0 to 100 fmol of gene II protein in 20 ~1 of buffer c,ontaining 20 mw-Tris . HCI (pH %O), 6 mM-MgCl,. 5 m&l-dithiothreitol. 200 pg bovine serum albumin/ml, X0 rnhl-KCI, and 5’!& glycerol. The reaction was c*arrird out at, room temperature for 5 min before gently loading ont,o the gel, which had been pre-run for I h. Marker lanes contained radioactive DNA molecular wright markers and tracking dyes. The gels were dried, was caarrird out at room and autoradiograph) t,ertiperat we’. (II)

b’ootprinting

of isolated

complexes

Thr binding reactions were carried out as described in the previous se&ion except that 50 fmol of end-labeled

161

Protein to the Origin

origin-containing restriction fragment (the 152 bp &oRIL BamHI fragment from pDGl17) and 5 to 40 ng (100 to 800 fmol) of gene II protein were used. Following a 10 min incubation at room temperature, 30 ng of DNase I in a volume of 1 ~1 was added for 30 s. Immediately after digestion, the samples were loaded onto the gel (Fi’l,, polyacrylamide as described above) while it was running at 2 V/cm. When all the samples were loaded the gel was run at 6.5 V/cm. Complex I, c>omplex TI and unbound DNA were visualized by autoradiography at room temperature and the corresponding gel bands were excised. The DNA was eluted electrophoretically, precipitated with ethanol and dissolved in. 4 PI of 900;, (v/v) formamidr plus tracking dyes. The samples were electrophoresed on standard 8 y0 pal>-acrylamide sequencing gels as described above. Autoradiography was carried out at, -70°C’ on Kodak XAR*-5 film with an intensification screen. (i) Methylation

interference with formation thr individual complex-w

qf

The origin-gene 11 protein complexes were isolated from a non-denaturing 596 polyacrylamide gel. and methylation interference was studied for each species. A total of 50 fmol of end-labeled, origin-cont)aining restrication fragment was first methylated with dimethyl sulfate as described by Maxam & Gilbert (1977). and then incubated with 0 to 400 fmol of gene II protein for 5 min at room temperature. Complex I. c>omplrx II and unbound DNA were separated on a 5?, polyacrylamide gel and visualized by autoradiography. The DNA fragments were excised from the gel. eluted rlect.rophoretically and recovered by precipitation with ethanol. The methylated DNA was then cleaved with I M-pip'dine at 90°C for 30 min. Following lyophilization, the samples were electrophoresed on standard Ho/0 polyacrylamide sequencing gels and autoradiography was carrkd on Kodak XAR-5 film with an out at -70°C intensification screen.

3. Results (a) Binding

of gene II yrotrin replication origin

to the

Filter-binding studies (Horiuchi, 1986) demonstrated that’ the gene II protein binds specifically to the origin of replication, requiring sequences around palindrome E (Fig. l), though the precise location of the binding site was not determined. Roth superhelical and linear DNA bound the protein. To determine the location of binding sites. we examined the binding of the gene I1 protein for each strand of the replication origin using the DNase I footprinting technique (Galas & Schmitz. 1978). The origin fragment used in this analysis was end-labeled at the unique AsuT site (position -56) on either the plus or minus strand as described in Materials and Methods. The results shown in Figure 3(a) indicat’e that the gene II protein protects 41 bases of the minus strand from DNase 1: from position -9 to position + 32 (lane a), and 46 bases of the plus strand. from position - 12 to + 34 (lane d). A site of enhanced cleavage is present on the plus strand at position +36 (thin arrow in site was Fig. 3(a)), while no such hypersensitive

162

D. Greenstein

and K. Horiuchi

-10 $1 +10 +20 +30 i-&O --'--, i 5'-ACGTTGGAGTCW\CGTTCTT*~T*GTGG*~TCTTGTTCTTGTTC1 3'-TGCAACCTCAGGTGCAAGAAATTATCAC~GAGAACAAGGTTTGACCITGTTGTGA~TGGGATAGAGCCAGATAA-5' (b)

Figure 3. Binding of the gene II protein to the origin. (a) DNase I protection experiment. Lane a, footprint to the minus strand of the origin (40 ng of gene II protein); lane b, minus strand DNase I control (no gene II protein added); lane c, Maxam-Gilbert G reaction for the minus strand; lane d, footprint to the plus strand of the origin (40 ng of gene II protein); lane e, plus strand DNase I control (no gene II protein added); lane f, MaxamGilbert G reaction for the plus strand. The 315 bp origin-containing AauI-CZaI fragment was labeled by 3’-end-filling with [a-32P]dGTP and DNA polymerase I Klenow fragment at the AsuI site for the analysis of the minus strand. The plus strand was labeled by treatment with T, polynucleotide kinase and [y-32P]ATP at the AsuI site, following treatment with CIAP. The thick arrow indicates the site of nicking by the gene II protein on the plus strand. The gene II protein nicks the origincontaining fragment at a low efficiency (0.5%). The nicking is strand-specific and occurs at the correct site (data not shown). The thin arrow indicates the site of enhanced cleavage by DBase I on the plus strand. (b) Summary of results. The region protected by the gene II protein is underlined; weak protection is indicated by broken lines. The short arrow indicates the nicking site. The longer arrow indicates a site of enhanced cleavage by DNase I. The top and bottom lines represent the plus and minus strands, respectively. found on the minus strand. The data are summarized in Figure 3(b). The protected region resides entirely within domain A. No binding to domain R was observed in the presence (Fig. 3) or absence (data not shown) of domain A. (b) Binding

of gene II protein

to defective origins

A series of defective origins had been constructed and characterized, leading to an understanding of

the overlapping signals within the origin (Dotto et al., 1982a, 1984). Figure 2 lists the properties of the defective origins used in the present study. Filterbinding studies (Horiuchi, 1986) using these delet,ion mutants have led to the following conclusions: (1) a number of the defective origins could bind the gene II protein; (2) the nicking site was dispensible for binding; (3) sequences required for binding were located around palindrome E. DNase T footprints of

Binding

of the fl Initiator

Protein

to the Origin

163

148 +48

C25

l 25 +I5 d-7 +3 -7

+48

c7 -

+48 -7

i-25

-20

-16

+32 +25 +15 +7

-7

+25-16

i-7 +15-5

+7+3-

-12 -3-

-13-

(al 3

+47 +32 A-47 c34 +12 -cl9 -47 -8

-20

-34

+47

CIS

c34

-4

.I7

CIE

-9 -13

c4

-21 -7

(b)

Fig. 4.

D. Greenstein and K. Horiuchi

164

' 3'-TGCkACCTCAGGTGCAAG~TTATCACCTCAGAG~~GGTTTGACGTTGTTGTGAGTTGGGATACAGCCAGATAA-5' 2 5'-cgGTTtttcgCCcttTgacgTtggAGTGGACTCTTGTTCT-3' 3'-gcCAAaaagcGGgaaActgcAaccTCACCTGAGAACAAGGTTTGACCTTCTTCTGAGTGAGTTGGGATAGAGC~GAT~-5' -------5'-gttTgacAGcttAtcaTCgaTAAgctTGGACTCTTGTTC~CTGG~C~CACTC~CC~ATCTCGGT~ATT-3' 3 3'-caaActgTCgaaTagtAGctAn'cgaACCTGAGAACAAGGTTTGACCTTGTTGTGAGTTGGGATAGAGCCAGAT~-5' 4

5'-cgtTctttaatagtega~ccAAgctTGGACTCTTGTTCCAAACTGGAACAACACTCAACCCTATCTCGGTCTATT-3' 3’-gcaAgaaattatcaCctGAggTTcg~C~~G~C~GGT~GAC~TG~GTGA~TGGGATAGAGC~~T~-5’

------

Figure 4. Binding of gene II protein to defective origins. (a) DBase I protection experiments with the plus strands labeled. Panels 1 to 5, footprinting to pDG117, A83, pD29, pD30 and A+29, respectively. Lanes a, Maxam-Gilbert G reaction; b, DNase I control (no gene II protein); lanes c to f, footprints with 40, 20, 10 and 5 ng of gene II protein. respectively. The lanes marked by + and 0 are the pDG117 origin fragment with and without added gene II protein, showing t.he nick at the origin. Arrows indicate the gene II protein nicking site. (b) Dru’nse I protection experiments with the minus strands labeled. The panels are labeled as in (a), except that lanes c show the footprint with 20 ng of gene II Protein. (c) Summary of footprinting results to the defective origins. Line 1. pDG117; line 2, A83; line 3, pD29; line 4. ~D30; line li, A + 29. The nucleotides that are different from the wild-type sequence are written in lower case. The arrow indieates t,he nicking site. The protected regions are underlined as for Figure 3.

gene IT protein to the defective origins are shown in Figure 4. We investigated the binding of the gene II protein to an A’B(wild-type domain A and defective domain B) origin, pDG117 (Fig. 4(a) and (b), panels 1). The gene II protein protected essentially ail of domain A in the absence of a functional domain B sequence. The extent of the protected region on both strands (data summarized in Fig. 4(c)) was identical with that observed in the presence of a functional domain B region (see Fig. 3). In addition, the A+Borigin showed footprints over the same range of protein concentrations as the wild-type origin. In the case of two deletions that extend from the left past the nicking site (Fig. 4: A83, panel 2; pD29, panel 3), gene II protein protected approximately 44 bp, a result similar to that for the wildtype origin. In both cases, gene II protein protected the non-specific sequences, corresponding to where the nicking site would have been, had it not been deleted. pD36 was constructed by insertion of a 16 bp Hind111 linker ($CCAAGCTTGG-3’) at the filled-in Hinfl site at position +8 (Dotto et al., 1984). As a consequence of the insertion, the wild-type nicking site is positioned 13 bp upstream from its normal position. pD30 RFT cannot be nicked by gene II protein in vitro (see Fig. 2). As shown in Figure 4(a) and (b), panels 4, gene II protein bound to pD30 and protected 46 bp from DNase I digestion. The pattern of protection closely resembles that of the wild-type origin (summarized in Fig. 4(c)). The inserted linker is protected, whereas the displaced nicking site is not. The result indicates that the left side sequence of domain A (up to + 11) alone cannot bind gene II protein. A83, pD29 and pD30

showed footprints over the same range of protein concentrations as the wild-type origin. The sequence requirements for binding and nicking at the right end of domain A were determined by the deletion mutants A+ 11 and A + 29 (Dotto et al., 1984). The deletion A + 11 removes all of palindrome E and fails to bind the gene II protein (Horiuchi, 1986). The right end deletion to position +29, A+ 29, can bind gene II protein, be nicked by gene II protein and terminate replication, but cannot initiate DNA synthesis. The results of the DNase I protection experiment for A+29 are shown in Figure4(a) and (b), panels 5. Gene II protein protects 31 bp on the plus strand, from position -7 to position +24. On the minus strand, gene II protein protects 33 bp, from position -9 to position +24. The results are summarized in Figure 4(c). The left boundary of the footprint is similar to that of the wild-type origin; however, the protected region is about 10 bp shorter on the right side.

(c) Essential contacts between gene I I protein and the origin Figure F, shows the results of a methylation interference experiment performed on the plus strand and the minus strand of the wild-type and various defective origins. In this experiment, origincontaining restriction fragments were first methylated with dimethyl sulfate. The complex formed with gene IT protein was isolated on a nitrocellulose filter and subsequently cleaved with piperidine for gel electrophoretic analysis. Therefore, fragments which result from cleavages at G residues that are essential for the binding should be reduced in

Binding

of the fl Initiator

intensity compared to the control. On the plus &and (see Fig. 5(a)), there are three G residues (positions + ‘7, + 8 and + 15) which, when methylated, interfere with binding. There are three G residues on thtb minus strand (see Fig. 5(b)) whose methylation interferes wit’h binding (positions + 10, + 18 and + 19). Methylation of other G residues, including one at, position + 12 of the minus strand, did not interfere with binding. All the defective origins shown in Figure 5 exhibit the same pattern of’ mct,hylation int)erference as wild-type.

(d) 7% cow binding sequence

The G residues whose methylation interfered with binding (Fig. 5) were located in the symmetric sequence 5’.TGGACTCTTGTTCCA-3’, between positions +6 and + 20 at the center of the protected region. We therefore tried to test whether this sequence was sufficient to bind the gene II protein. The 15 bp sequence from the center of domain A was synt,hesized and cloned into the Hind111 site of pUCl9 (see Materials and Methods). Figure 6(a) shows the result. of a DNase I footprinting experiment on a fragment containing the 15 bp element.

Protein to the Origin

165

Gene II protein binds to the fragment containing the 15 bp element and protects approximately 25 bp on both strands from DNase I digestion. The protected region extends from position +3 to position + 29 on the plus strand (lane d), and from position + 1 to position +24 on the minus strand (lane a). There are two sites of enhanced cleavage, one located at position - 1 on the plus strand, the other located at position -3 on t,he minus strand. The data are summarized in Figure 6(b). We call this 15 hp sequence the core-binding sequence. Although the core-binding sequence is able to bind the gene II protein, it) yields a protection pattern which is shorter on both ends than the wild-type footprinting experiment pattern. A quantitative indicat,ed that gene II protein binds the core sequence approximately fourfold less well than the wild-type origin (data not’ shown). Both halves of the core-binding sequence are necessary for binding. In foot,printing experiment’s on a fragment containing the sequence from + 12 to +32, only very weak protection by gene IT protein was found (data not. shown). In addition, the deletion mutant A+ 11. which deletes t.he right. end of the origin to position + 11. does not. bind gene IT protein (Horiuchi. 1986).

+25-+

la) Fig. 5.

166

D. Greenstein and K. Horiuchi

ob

ef

c d

+39 -

+19+I0

-

-10

-

-20

-

(cl -10

t

+1

+10

t f tt tt tt AA

+20

A

+30

+40

t tt

ACG~T~AGTCCACGTTCTn‘AATAGTGGACTCTCTTG~CC~CT~~C~CACTC~CCCTATCTC~TCTATT TGCAACCTCAGGTGCATTATCACCTGAG~C~~T~GACCTTGTTGTGAGTT~ATAGAGCCAGAT~

1

1 1 1 1 111 1

v

w

1

1

+50

tt 111

Figure 5. Methylation interference experiment. A methylation interference wa-s performed on the plus strand (a) and the minus strand (b) of wild-type (pDG117), and mutant fragments. The origins are pDG117 (lanes a and b), A83 (lanes c and d), pD29 (lanes e and f), pD30 (lanes g and h), and A+29 (lanes i and j). End-labeled DNA fragments were methylated with dimethyl sulfate, and subsequently incubated with gene II protein. The complex formed was trapped by and then eluted from a nitrocellulose membrane, as described in Materials and Methods. The recovered DNA was cleaved with piperidine and analyzed by gel electrophoresis (lanes b, d, f, h and j). Lanes a, c, e, g and i are untreated controls: the methylated fragments were cleaved with piperidine without treatment with gene II protein. The G residues at positions +7, +8 and + 15 (on the plus strand) are reduced in intensity in the samples isolated from the DNAprotein complex as compared to the controls. For pD30 (lanes g and h), the effect is not as great as for the other origins, but it is clear in the original autoradiogram. The G residues at positions + 10, + 18 and + 19 (on the minus strand) are reduced in intensity in the samples isolated from the DNA-protein complex. Some of the smaller DNA fragments have slightly reduced intensities in the treated as compared to the untreated samples. This is an artifact of gel electrophoresis, probably caused by contaminating SDS in the sample. (c) Summary of the results showing the wild-type sequence (pDG117). The 6 G residues identified by methylation interference are indicated by the large arrowheads. The small arrows indicate positions where G residues have been substituted in the various mutants without affecting the pattern of methylation interference (i.e. G residues at these sites do not cause methylation interference).

Binding

of the fl o

Initiator

Protein to the

Origin

167

bc

-I4 -8 -2 +4 A-7

.+I9 c22 ,+25

+26

(0) -10

+1 +10 +20 +30 I I I I 5'-atgacCatgaTTacgccAagcTGG*~~TGTTcc~gctTGc*tgcctgCaggtCgactctagaGgagggg-3' 3'-tactgGtactAAtgcggTtcgACCTGACAACAAGGTcgaACgTacggacGtccaGctgagatctCctcccc-5' (b) Figure 6. Binding of gene IT protein to the core-binding sequence. (a) The 15 bp element, corresponding to the center of domain A. was chemically synthesized and cloned into the Hind111 site of pUC19 (Materials and Methods). Binding to the EcoRT-PwLII fragment containing the 15 bp element was analyzed by DBase I footprinting. Lane a, footprint with 40 ng of gene II protein, with the minus strand labeled; lane b, DBase I control for the minus strand (no gene II protein): lane c, MaxamPGilbert G reaction for the minus strand; lane d, footprint with 40 ng of gene II protein, with t,he plus strand labeled; lane e, DBase I control (no gene II protein) for the plus strand; lane f. Maxam-Gilbert G reaction for the plus strand. The arrows indicate the sites of enhanced cleavage by DPu’ase I. (b) Summary of results. The protected region is underlined. Nucleotides that differ from the wild-type sequence are written in lower case. (e) A nal?ysis

qf gene

I I protein-origin

complexes

Gene TI protein-origin complexes were analyzed by polyacrylamide gel electrophoresis by the procedure described by Fried & Crothers (1981). Figure 7 (lanes a to d) shows the results obtained

with a fl origin (A+B-) fragment (M,= I.3 x 105). Gene IT protein forms two discrete complexes with the fl origin. At lower gene IT protein concentrations (lane b), what we call complex I (M, corresponding to 3.5 x IO5 using DNA molecular weight, markers) forms. At higher concentrations of gene TT

168

D. Greenstein

and

K. Horiuchi

abcdefghijklmn

Figure 7. Gel electrophoretic analysis of gene II protein-origin complexes. Radioactively labeled DNA fragments were incubated with various concentrations of gene II protein, and analyzed by polyacrylamide gel electrophoresis (Fried & Crothers, 1981). The 190 bp EcoRI-Bu?tI fragment from pDG117, containing all of domain A and part of domain B (lanes a to d) was incubated with 0 (lane a), 25 fmol (lane b), 56 fmol (lane c) or 100 fmol (lane d) of gene II protein, respectively. (This fragment was slightly contaminated with a 218 bp and a 114 bp fragment that derived from the pBR322 vector and were not affected by gene II protein.) The 190 bp A + 29 origin-containing EcoRI-BamHI fragment (lanes e to g), was incubated with 0 (lane e), 25 fmol (lane f) or 100 fmol (lane g) of gene II protein. The 208 bp B&N1 fragment from pMBS1, containing the core-binding sequence (lanes h to k), was incubated with 0 (lane h), 25 fmol (lane i), 50 fmol (lane j) or 100 fmol (lane k) of gene II protein. The 187 bp EcoRI-EcoRV fragment from pBR322 (lanes 1 to n) was incubated with 0 (lane I). 25 fmol (lane m) or 100 fmol (lane n) of gene TT protein. protein (lanes c sponding to 5.6 x obtained with the origins (data not bind to a control

and d), complex II (M, corre105) forms. Identical results were wild-type (A+B+), pD29 and A83 shown). Gene II protein did not fragment from pBR322 (Fig. 7,

lanes 1 to n).

Both complexes I and II were formed with the initiation defective origin A+29 (Fig. 7, lanes e to g). However, the formation of complex II required a higher concentration (about 4 times, compare Fig. 7 lanes b and g) of the protein than for the wild-type origin. The A+29 complexes migrate slightly faster than the wild-type (A+B-) complexes, but this may be due to the position of the binding site within

the fragments:

in the case of the wild-type

origin the binding site is in the middle of the fragment, whereas for A+29, it is closer to one end. The 15 bp core-binding sequence binds the gene II protein in one step, forming only complex I (Fig. 7, lanes h to k) and requiring about fourfold more gene II protein (compare Fig. 7 lanes b and k). ( f) Footprinting of isolated gene II protein-origin complexes

DNase I footprinting experiments were performed on the isolated gene II protein-origin complexes

I and II. An origin-containing

restriction

fragment labeled on the plus strand with gene II protein and briefly

was incubated

treated with DNase I. Separation of the complexes followed by analysis on a DNA sequencing gel permitted the detection of each respective DNA-protein interaction. A similar analysis was performed by Andrews et al. (1987) to analyze the intermediates in binding of the yeast FLP recombinase to its target site. In complex 1 (Fig. 8(a), lane b) gene TI protein protected 27 bp extending from position + 2 to position +28. The protection pattern of complex T closely resembles that obtained with the corebinding sequence (compare Figs 6 and 8). The protection pattern of complex IT (Fig. 8(a), lane c) extends 40 bp from position -7 to position + 33. Thus, the protection pattern of complex IT corresponds to complete protection of the wild-type origin sequence (compare Figs 3 and 8). The level of protection observed in complex II (Fig. 8(a), lane c) is greater than that observed in complex I (Fig. 8(a), lane b). (g) Methylation interference with formation of the individual complexes

We reasoned that the methylation experiment

performed

by filter-binding

interference (Fig. 5) may

qf the .fl

Binding

G

Initiator

a

Protein to the Origin

b

c

169

G

c49 3

-16

3 (a)

Complex

II

Figure 8. Footprint analysis of isolated gene II protein-origin complexes. (a) An end-labeled origin-containing restriction fragment from pDGl17 was incubated with gene II protein, treated with DNase I, and electrophoresed on a 5s0 polyacrylamide gel. DNA in the separated complexes was extracted from the gel and analyzed by denaturing gel electrophoresis. For detailed procedures see Materials and Methods. Lane a, DNase I control: the DNA fragment was treated with DNase I in the absence of gene II protein, and was isolated from the gel in parallel with the complexes. Lane b, footprint of complex I formed in the presence of 10 ng of gene II protein. Lane c, footprint of complex II formed in the presence of 20 ng of gene II protein. The lanes marked G are the MaxamGilbert G reactions. An arrow on the right indicates a fragment formed by site-specific nicking by the gene II protein. (b) Summary of footprinting results. The arrow indicat,es the nicking site.

170

D. Greenstein

have failed to identify essential cont,acts unique to complex II because both complexes would be retained on the filter. In order to determine the essential contacts for formation of the two complexes, we performed a methylation interference experiment. for separately isolated complexes. Tn this experiment, a methylated origincontaining restriction fragment was incubated with gene IT protein, and complex I, complex II and unbound DNA were separately isolated by polyacrylamide gel electrophoresis. Following cleavage by piperidine at methylated C residues, the products were analyzed on a DNA sequencing gel. Fragments which result’ from cleavages at

and K. Horiuchi

G residues that are essential for formation of the respective complexes should be reduced in intensity compared to the control. Nine G residues were found whose methylation inhibited the formation of complex II (Fig. 9(a), panels 1 and 2, lane b). Five of t.hese G residues are on the plus strand at positions + 7: + 8, + 15. +25 and + 26 (Fig. 9(a), panel 1, lane b), while four of these G residues are located on the minus strand at positions + 10, + 18, + 19 and +29 (Fig. 9(a), panel 2, lane b). The nine G residues required for complex TT formation are located in three clust,ers (as summarized in Fig. 9(b)): the first between positions +7 and + 10: the second between positions + 15 and + 19; and 2 abed

(a)

+25 -cl5 +7

-7 -16

(b)

-10

J(

+lc,

+2:

no

‘3:

*?

+5?

Figure 9. The essential contacts for formation of the 2 gene II protein-origin complexes were probed by methylation interference. (a) Origin-containing restriction fragments were methylated with dimethyl sulfate, incubated with gene II protein and electrophoresed on a 5% polyacrylamide gel. DNA was isolated from the separated complexes, cleaved with piperidine and electrophoresed on a standard sequencing gel. Panel 1, the DNA fragment (152 bp BurnHI-EcoRI fragment from pDG117) labeled on the plus strand; panel 2, the DNA fragment (315 bp AsuI-CZaI fragment from fl RFI) labeled on the minus strand. Lanes a, unbound DNA without addition of gene II protein; lanes b, complex II formed with 5 ng of gene II protein; lanes c, complex I formed with 5 ng of gene II protein; lanes d, unbound fragment after incubation with 20 ng of gene II protein. An arrow at the side indicates a fragment formed by site-specific nicking by the gene II protein. (b) Summary of methylation interference results. The G residues required for formation of both complexes (filled circles) and the G residue8 required only for formation of complex II (open circles) are indicated on the DNA sequence. Repeated nucleotide sequences are indicated by horizontal arrows. A vertical arrow indicates the nicking site.

Binding

of the fl Initiator

the third between positions +25 and +29. These (: residues include the six essential G residues in the tsorr-binding sequence as well as a homologous set at the right, end of domain A. There are six G residues whose methylation inhibits the forma,tion of complex 1 (Fig. 9(a), panels 1 and 2, lane c). These G residues are identical with t,he essential G residues in the corebinding sequence (summarized in Fig. 9(b)). The three G residues on the plus strand are located at’ positions +7, +8 and + 15 (Fig. 9(a), panel 1, lane (2). The t,hree Q residues on the minus strand are loc~atc~d a.1 positions + 10, + 18 and + 19 (Fig. 9(a). panel 2, lane c). Furthermore, these six G residues were preferentially detected in unbound I>?iA under conditions where most of the DNA was hound to gene II protein (Fig. 9(a), panels 1 and 2, Iant> d). This experiment indicates that complex I and t~omplrx II result from sequential DNA protein interactions. and that complex IT differs from complex J in having an extra set of homologous contacts. The DNA isolated from complex II (Fig. 9(a). panc~l 1, lane b) has an extra cleavage located on the plus strand at the gene II protein nicking site. Since this band is observed only in complex JJ, nicking most’ likely requires formation of taomplex JJ.

4. Discussion Jn this paper. we have analyzed the interaction of the gene IT protein with the fl origin of replication. The gene JJ protein binds to restriction fragments containing the origin and protects 40 base-pairs from DNase I digestion. The protected region corresponds to domain A, the minimal origin sequence. Gene II protein binds to the origin and forms two complexes that) are separable by polyacrylamide gel electrophoresis (see Fig. 7). Using DNA molecular weight markers. the increase in mass due to binding the first unit of gene II protein is equal to the increase due to binding the second unit (2 x 105 M,). Therefore, complex IT probably csontains twice as many molecules of gene II protein as does complex J. Our failure to detect two different binding patterns with the wild-type origin by DNase I footprinting may he due to interconversion or different’ial stability of the complexes during the foot,printing procedure. The ability to detect binding intermediates by polyacrylamide gel rlectrophoresis but not by direct DNase I footprinting was reported by Andrews et al. (1987) for t,hr interact’ion of the FJ,P recombinase with its target sequence. From the footprints of the isolated complexes, we c*onclude that complex J results from interaction of gene JI protein with the core binding sequence and gives a 2’7 bp protection pattern. The protection pat’tern of complex J (Fig. 8(a), lane b) is identical wit’h that obtained with a cloned copy of the core binding sequence (Fig. 6, lane d). Direct footprinting of complex IT shows the full 40 bp

Protein to the Origin

171

protection of domain A. Since the level of prot,ection observed in complex II is great,er than that observed in complex J (see Fig. 8), complex II probably results from a more stable gene JJ protein-origin interaction. The three-dimensional structure of bovine pancreatic DNase J (Suck & Oefner, 1986) suggests that the DNase I footprint pattern overestimates the size of the actual DNAprotein interaction by approximately 5 bp per end because of t’he intrinsic DNA binding surface of the DNase J molecule. Jf so, the region covered 1)~ gene II protein in complex J would be approxlmately 15 bp. the size of the core-binding sequence. The covered region in complex JJ would correspond to roughly 30 bp. or twice the length of that’ in complex J. Therefore, the foot’prints we observt> with complex I and complex II would be reasonable protection pat,terns for one and t,wo units of respect,ively. binding to a c*ontiguous protein, stretch of DNA. We probed t)he contacts between gene JJ protjrin and the origin in the two complexes by methylation interference experiments. Methylation of any ant’ of six G residues (see Fig. 9) in the core-binding sequence inhibits formation of both complex J and complex II. Methylation of one of three G residues located between positions + 25 and + 29 specifica.lly inhibits the formation of complex JJ. This is consistent with the observation that the cloned core-binding sequence can form only complex I (Fig. 7). The nine G residues are located in three symmetric clusters, suggesting that gene IT protein interacts homologously with each repeat,. Since the G residues required for formation of complex I are a subset of those required for formation of complex JI, complex I appears to be an intermediate in formation of complex JJ. Methylation interference experiments performed using filter binding only identified the six G residues in the core binding sequence as being essential contacts (see Fig. 5). The most likely explanation is that methylation of G residues located bet#ween posit’ions + 25 and + 29 still permits formation of complex 1 and retention on the nitrocellulose filter. Based on the results described above, WC propose a model for gene II protein binding as shown in Figure 10. There are four repeats of the sequence 5’.TGGAC-3’ (a and /?) or 5’-TGGAAC-3’ (y and 6) in alternat’e directions within domain A. The core binding sequence consists of repeats b and y in inverted orientation. Full protection of 40 bp requires repeats /?, y and 6. Repeat) CIis dispensable for both binding and nicking, but required for termination of replication (Dotto et ~6.. 1982h). Methylation interference experiment)s indicate that gene JJ protein makes homologous tsontacts with the repeats 8, y and 6. Figure 10 depicts the B-form DNA structure of domain A, illustrating the protected region and the close contacts (i.e. 9 sites of methylation int,erference). The points of methylation interference are located in repeats /?, y and 6 in t,he major groove. Although we do not have quantitative

172

D. Greenstein and K. Horiuchi

Y Figure 10. A model for binding of gene II protein to the replication origin. The structure of domain A is drawn as DNA. Four subunits (marked I, II, III and IV) of the gene II protein are assumed to bind the origin.

B-form Subunits subunits depicted,

III and IV are located in a symmetric configuration. The subunit II is located about 70” apart from IV. The are depicted by different shapes to indicate the rotation around the DNA helix. Subunit I is arbitrarily since we do not have any data on its configuration. The points of methylation interference in the core-binding

sequence are circled. The G residues, which specifically interfere with formation of complex II when methylated, are shown uncircled. The G residue at position + 29 is marked by an asterisk to indicate that it is deleted in mutant A + 29. The 3 repeats (/I, y and 6 as described in the text) required for binding are indica$ed. The repeat CI(position - 13 to -9) is not shown. The plus and minus strands are labeled. The nicking site is indrcatkd by a vertical arrow. The curved arrows show the 5’+3’ direction on each strand.

data to determine how many gene II protein monomers bind to the core-binding sequence, the symmetric base sequence, particularly that of guanine residues which interfere with formation of complex I upon methylation (see Fig. 9), suggests that two subunits (indicated as II and TIT in Fig. 10) bind symmetrically to it. Then, the wildtype origin would bind four subunits of gene II protein to form complex II (see Fig. 10). This notion is consistent with the symmetry between the repeats y and 6. Assuming that one gene II protein subunit binds to each of these sequence elements, the two subunits which bind to the right half of domain A (indicated as III and IV in Fig. lo), should be located in a symmetric configuration. In the deletion mutant A+29 the 6 repeat is partially destroyed and complex II is destabilized. The subunit which binds to the left-half of the corebinding sequence (indicated as II in Fig. 10) should be located about 70” apart from IV. This model predicts two different set of dimer contacts. Absence

of

symmetry

in

the

sequences

of

the

nicking site region reflects the non-symmetric nature of the nicking which occurs at a unique site on the plus strand. The configuration of the subunit bound at this region (indicated as I in Fig. 10) is unclear.

Its involvement

is inferred

from

both

the

footprint (Fig. 8) and the gel mobility (Fig. 7) of complex II. It is remarkable that 683, pD29 and pD30 all bind gene II protein to yield full protection of approximately 40 bp. Note that the sequence of the

protected region is different among the three mutants. Gene II protein must therefore recognize the sequence to the right of position +5, and thereby protect approximately ten nucleotides to the left. In the wild-type origin, the nicking site is located within this 10 bp sequence. Obviously, only a specific sequence can be nicked, since A83, pD29 and pD30 are not nicked by gene II protein. However, when gene TI protein is bound to the 15 bp core (position + 6 to + 20) it does not protect the 10 bp on the left. Evidently, repeat 6 is required for the gene II protein to bind properly and protect the 10 bp on the left. The interaction of subunit IV (see Fig. 10) with repeat 6 may influence the binding of subunit I via protein-protein interactions. An alternative to this model is that binding of subunit IV causes a conformational change in the complex which results in protection of the left most 10 bp. The initiation-defective origin A+ 29 can form complex II, but less efficiently (see Fig. 7). Thus, the sequence to the right of position + 28 affects the second binding step. Gene II protein bound to the A+29 origin, even though able to introduce the nick, fails to participate in the initiation of unwinding (our unpublished data). The requirement of sequences at the right end of domain A for the protection of the left end suggests that protein-protein interactions are involved in the binding. While gene II protein exists as a monomer in solution (Meyer & Geider, 1979a), protein-protein interactions could occur on the

Binding qf the41 Initiator 1)X.A or during binding. Protein-protein interact’ions are also implicated by the observation that the gene I1 protein binds only very weakly to repeats y and 6 in the absence of repeat j3 (data not shown). Since binding of the first unit of gene 11 protein (int,eraction with P-7) seems to aid the

binding

of the second (interaction

with

6). the

binding must be (so-operat’ive. (lo-operative binding allows full binding to occur in an all or none fashion over a small range of protein concentrations. For inst,ance. co-operative binding of the i repressor protein CT is a crucial factor in regulating the A Iysis--1ysogeny decision (for a review, see Ptashne, 19%). The intracellular level of cT is tightly regulated by both positive and negative controls. The concentration of gene II protein in the infected cell is t,ightly controlled by the phage-encoded gene \’ protein. a translational repressor (Model et

al.,

1982; Yen

binding

& Webster,

to the origin

would

1982). Co-operative cause the plus strand

initiation rate to be highly sensitive to the level of gene IT protein when It is rate-limiting for init,iation. Tn the caasr of the icosahedral single-stranded phage 4x174, the replication initiator protein, the phage-encoded gene A prot,ein, binds covalently at the 5’ end of the nick. The high stability of the complex, due to the covalent nature of the adduct, would allow DNA replication to be processive. For origins which form non-covalent’ complexes with their initiator proteins, multiple initiator protein molecules are oRen involved in the interaction. Notable examples include the OriC:-DnaA protein complex (Fuller et al.. 1984) and the Oril-0 protein

complex (Tsurimoto & Matsubara, 1981). Ry this criterion. t.he fl origin would be a member of the latter c~lass. \Vr are grateful to Norton Zinder, Peter Model, Joe Heitman, Wilder Fulford, Nick Davis, David Russell and r\my Roth for stimulating discussions and critical reading of the manuscript. We thank Beth Goldstein and Peter Model for oligonucleotide synthesis and Mike Van Dyke and Barkur Shastry for advlce on footprinting. This work was supported in part by grants from the National Scienctb Foundation and the Xational Tnstitutes of Hralth. References K. J., Heatty, L. 0. & Sadowski, P. D. (1987). *I. :Mol. Biol. 193. 345-358. Baas. P. D. (1985). Biochim. Biophys. Acta, 825, 111-139. Beck, E. & Zink, B. (1981). Gene, 16, 35-38. Booke. J. I).. Vovis, G. F. & Zinder, pu’. D. (1979). Proc.

Andrews.

*lraf. Acad. Sri.,

U.S.A.

76. 2699-2702.

Kolivar. F.. Rodriguez, R. L.. Green, P. J., Betlach, M. (1.. Heyneker, H. L., Boyer, H. W., Crosa, J. H. & Falkow, S. (1977). Gene, 2, 95-113. Bradford, M. M. (1976). Anal. Biochem. 72, 248-254. (‘leary, $I. M. & Ray, D. S. (1980). Proc. Nut. Acad. Sci., C’.S.A. 77. 4638-4642.

(.‘leary. ,J. M. RERay, D. S. (1981). J. Virol. 40, 197-203. Davis. K;. G.. Boeke. ,J. D. & Model. P. (1985). J. Mol. Riol. 181. 111-121.

to th,e

Protein

153

O&yin

de Boer. H. A.. Cornstock, Proc. .vat. Acad.

L. ,J. & Yasser. M. (1983).

1’.S.,4. 80. 21--Z. G. P. & Horiuchi. K. (1981). .J. Mol. Sci..

Dotto. Rio/. 153. 169.-176. Dotto. G. P. & Zinder. N. D. (1984a). l’roc. Saf. Acctd. St-i., f’.S.A. 81, 133&1340. Dotto. G. P. & Zinder. 1\;. D. (19846). Satlrr(~ (London). 311. 279-280. Dotto, G. P.. Enea. V. & Zinder. E. 1). (1981). Proc. ivat. Aead. Sci., I:X.A. 78, 542lL5424. Dotto. G. P.. Horiuchi. K.. Jakes, K. J. & Zinder. X. 0. (19820). .J. Mol. Biol. 162. 335.-343. Dotto. G. P., Horiurhi, K. & Zinder. X. D. (19826). Proc. ;C’nt. Acad. Sci.. I:.S.A.

79. 7122-7126.

Dotto, G. I’., Horiuchi. K. & Zinder. S. D. (1984). .J. Mol. Riol. 172. 507~521. Fried, M. & (‘rothers. I>. M. (1981). Slrcl. rlcids IMPS.9. 6505-6525. Fulford, W. (1986). Ph.D. thesis. Rockefeller Ilniversity. Fulford, W. & Model. P. (1984). J. ,Wol. Hioi. 178. 137153. Fuller, R. S.. Funnell. B. E. & Kornberg. A. (1984). C’c,lZ, 38. 88!&900. Galas. D.
Harth, G.. Baumel, 1.. Meyer, T. F. h Geider, K. (1981). Eur. J. Biochem. 119, 663-668. Hill, I). F. & Petersen. G. P. (1982). J. 1-&l. 44. 32-46. Horiuchi. K. (1980). f’roc. Nat. Acnd. Sci.. I’.S.il. 77. 522w5229.

Horiuchi. Horiuchi.

&. (1986). J. Mol. Biol. 188. 21%223. K.. Vovis. G. F. & Model, I’. (1978). In The Single-Stranded DNA Phages (Denhardt, I). T.. Dressier. D. & Ray, D. S.. eds), pp. 113-137. Cold

Spring Harbor Harbor, SY. ,Johnston, S. & Ray,

Laboratory

Press.

Cold

Spring

I). S. (1984). J. Mol. Hiol. 177. 68.5700. Kim, M. H. & Ray, D. S. (1985). J. C’irol. 53, 871-878. Laemmli. L’. K. (1970). ,lraturp (London), 227. 168@ 1685. Lyons, L. B. & Zinder.

Maniatis.

T.. Fritsch.

E. D. (1972).

F’irology,

49, 45-60.

,J. (1982). Molec’ular Cloning, A Laboratory Xanual, Cold Spring Harbor Laborat,org Press. (“old Spring Harbor.

Maxam,

E. F. & Sambrook.

NY.

A. & Gilbert,

C’.X.A. 74. 560-564. Meyer, T. F. & Geider, 1263612641.

W. (1977). f’roc. K. (1979a).

,va,t. -4cnd. Sri.,

J. Hiol.

Meyer, T. F. & Geider, K. (19796). J.

Hiol.

(‘hem.

254,

C‘h,em. 254,

1264%.-12646.

Meyer, T. F., Geider, K.. Kurz, (1. & Schaller, H. (1979). *Ta.ture (London),

278, 365-367.

Model, I’., McGill. C.. Mazur, B. & Fulford. W. D. (1982). (‘ell , 9 29 ,I 329-335. Muller-Hill, B., C’rapo. L. &I Gilbert. W. (1968). Proc. Acad. Aci., I’.S.A. 59. 1259-1264. Norrander, ,I.. Kempe. T. & Messing. ,I. (1983). Gene. 101~106. Ptashne, M. (1986). A .Genetic Switch,: Gene Control Phage 1, Cell Press, Cambridge, U.S.A. Siebenlist, U. & Gilbert, W. (1980). Proc. ‘vat. Acad. C’.A”..~. 77. 122.-126. Suck, D. & Oefner. C. (1986). Na,turs (London). 321. 625. t

Nat. 26, and ~%i.,

6Z&

174

D. Greenstein

Sutcliffe, .J. ‘2. (1978). Cold Spring Hart~or Symp. Quant. Biol. 43, 77-90. Tsurimoto, T. & Matsubara, K. (1981). Nucl. Acidn RPS. 9, 1789-1799. van Wezenbeek. P. M. G. F., Hulsebos, ,J. J. M. & Schoenmakers, ,I. G. G. (1980). Gene. 11, 129-148.

Edited

a,nd K. Horiuzhi

Wray, W., Roulikati, T.. Wray, V. I’. bt Hancock, R. (1981). Anal. Biochem. 118, 197-203. Yen. T. S. R. & Webster, R. E. (1982). Cell, 29, 337-345. Zinder. N. 11. & Roeke. ,J. 11. (1982). Gene, 19, I-10. Zinder, N. D. & Horiuchi. K. (1985). Microbial. Rev. 49, 101-106.

by M. Gottesman