Nucleic Acids Research, Vol. 20, No. 4 811-817

Roles of a 106-bp origin enhancer and Escherichia coliDnaA protein in replication of plasmid R6K

Frank Wu, llya Goldberg and Marcin Filutowicz*Department of Bacteriology, 1550 Linden Drive, University of Wisconsin, Madison,Wl 53706, USA

Received November 4, 1991; Revised and Accepted January 23, 1992

ABSTRACT

A dnaA 'null' strain could not support replication ofintact plasmid R6K or derivatives containingcombinations of its three replication origins (a, y, £).DnaA binds In vitro to sites In two functionally distinctsegments of the central y origin. The 277-bp coresegment is common to all three origins and containsDnaA box 2, which cannot be deleted withoutpreventing replication. Immediately to the left of thecore lies the 106-bp origin enhancer, which containsDnaA box 1. When the origin enhancer is deleted, thecore alone can still initiate replication if levels of wt Tprotein are decreased or If copy-up T mutant proteinsare provided In trans. DnaA does not effect expressionof R6K replication initiator protein *-, although severalDnaA boxes were identified in the coding segment ofthe plr gene, which encodes T. Together these datasuggest that a single DnaA box, 2, is sufficient forinitiation from the y origin and might be sufficient andrequired for the activity of the a and /3 origins as well.Implications of the DnaA protein binding to twodomains of the y origin and the role of the 106-bp originenhancer in replication are discussed.

INTRODUCTION

DnaA protein facilitates bidirectional replication from thechromosomal origin oriC in vivo (1) and in vitro (2—4). Fourhighly conserved DnaA boxes (S'-TTA^/^A0/^) arepresent in oriC of E. coli and the chromosomal origins of fiveother gram-negative bacteria (5). In the current model of theinitiation of replication at oriC, DnaA protein binds to these sites(6) and successively melts three 13-mers located in an AT-richsegment (7). It is unclear to what extent, if any, a directinteraction of DnaA protein with the AT-rich oriC segment (8)contributes to the opening of the DNA duplex. Following thisstep, the DnaB-DnaC complex is guided into the melted region(9, 10) to allow a priming event to occur.

Several plasmids analyzed to date contain DnaA boxes in theirreplication origins. Not only can some of these plasmids replicatein conditionally lethal dnaA mutants but they are also able to

suppress these mutations by integrating into chromosome (11).It was somewhat surprising, therefore, that F and PI plasmids,which are able to facilitate the integrative suppression of dnaA-temperature sensitive strains (11 -13), fail to replicate in the dnaA'null' strains (14).

Wickner etal. (15), studying the involvement of DnaA proteinin an in vitro replication system, proposed that the roles of DnaAin the replication of oriC and plasmid PI probably differ. Inaddition to the integrative suppression, two biochemical featuresappear to support this conclusion. First, the levels of DnaA whichcan activate PI replication are insufficient to activate replicationfrom oriC in vitro. Second, the ADP-bound form of DnaA isactive in PI replication, while in vitro OriC replication requiresthe ATP-bound form (7).

Plasmid R6K contains three origins of replication, termed a,P and y. Replication from each origin requires different civ-actingregions (diagrammed in figure 1), but each origin requires theR6K-encoded protein IT for replication (16 — 19). IHF protein isthe only host-encoded factor which is able to bind to the y origin(20) and is required for replication when normal levels of sprotein are made (21). In this study we present evidence that theDnaA protein also binds to the central regulatory segment ofplasmid R6K (7 origin) and is required for activity of all threeplasmid origins under all conditions examined. After completionof this work we learned that the DnaA protein is also essentialfor replication of R6K y origin plasmid in vitro (22).

MATERIALS AND METHODS

Bacterial strains and plasmids

Bacterial strains and plasmids are listed in Table 1.

TransformationsCompetent cells were transformed with supercoiled plasmid DNAusing the calcium chloride procedure (23).

DNase I footprintingThe plasmid pMF34 (24) was digested with EcoRI or Sail. TheEcoRI digests were dephosphorylated with calf intestinal alkalinephosphatase and treated with T4 polynucleotide kinase in the

* To whom correspondence should be addressed

by guest on August 14, 2016

http://nar.oxfordjournals.org/D

ownloaded from

812 Nucleic Acids Research, Vol. 20, No. 4

Table 1. Plasmids and E. coli strains

Plasmid Origin(s) Markers Source

A22, A14-pirWTA22-/*>113*, p/>116audpir20Q

pACYC177pFW8pFWIO, pFWll

pMF26pMF26pi>113*pMF32

pMF33

pMF34pMF35pMF39pMF50

pMF50 pirtOOpMFlOOpR6K

RK2

RK2pl5AR6K ypl5A

R6K 7 -0R6K 7-/3pUC, R6K 7

pUC, R6K 7

pUC, R6K 7R6K 7pUCR6K /3

R6K/3pUC, R6K 7R6K a-7-/3

Tet"

Tet»PenRKanR

PenR

CamR

PenR

Pen"Pen"

Pen"

PenR

Pen"Pen"PenR

Pen"PenR

Pen" Strep"

(26)

(21)(44)this study; see fig. 2this study; the pir gene was inserted in either orientation as a BamHI fragment frompMF39 into pACYC184 (44) digested with BamHI(24)(24)this study; the truncated R6K y origin was inserted as a HindllJ-Bgin fragment intopUC9 digested with Hindm and BamHIthis study; the R6K 7 origin was inserted as an EcoRI-BglH fragment into pUC8digested with EcoRI and BamHI(24)pMF33 with Haell fragment containing pUC origin deleted(24)this study; pMM31 (36) was partially cleaved with HindM; the fragment containing the/3 origin was ligated to the Hindm fragment of pTJS108 containing the PenR geneequivalent to pMF50, but carrying the pirQfX) mutation (51)(20)(52)

*pi>113 is a new designation of the previously described pir\, pi>104 (24) and pirX (32) mutations, all of which were revealed by sequencing to contain a Proto Ser change at the amino acid 113 (J. Hoffman and M. Rlutowicz, unpublished).

Bacterial strainAQ634AQ3563C2110p3478

Genetic markers SourcetrpA9605 his-29 argH thyA deo metD88 dnaA+ (53)AQ634 dna/1850::7>tl0 (Tel") chr pKNSOO (KanR) (53)F~ his' thy' rha polA\ (54)thyA36 deoCl \H(rmD-rmE)\ polA\ (55)

presence of T ^ P - A T P . The Sail fragments were treated with theKlenow fragment of DNA polymerase I in the presence of or^P-ATP. The EcoRI-labeled fragments were then digested with Sail;the Sail-labeled fragments, with EcoRI. The conditions forfootprinting were as described (6). DnaA purified in thelaboratory of Dr. A. Komberg was obtained from Dr. E. Boye.Sequencing reactions (A and G) were performed as described(25). EcoRI was obtained from New England Biolabs; DNaseI, from Boehringer Mannheim; all other enzymes were obtainedfrom International Biotechnologies, Inc.

•s Immunoassay and plasmid copy number determinationImmunoblotting analysis of v was carried out as previouslydescribed (26). Plasmid copy number was determined in celllysates obtained by the alkaline lysis procedure (27).

RESULTSPlasmid R6K and its derivatives cannot be transformed intodnaA 'null' mutant cells.The intact R6K plasmid is unable to transform the dnaAr.TnlOmutant strain AQ3563 (28). Because the a and /J origins are usedpredominantly in the intact R6K (29, 30), this finding could reflectthe requirement for DnaA for activity of these two origins butnot the central 7 origin. Also, in the absence of the host proteinIHF, the 7 origin is more sensitive than the other two originsto inhibitory levels of R6K initiator protein w (20, 21); thus itis possible that the origins may also differ in their requirementfor DnaA. We wished to determine if any single origin, separatedfrom the other two, would replicate in the absence of DnaA. Thus

Tabte 2. Transformation efficiency of R6K plasmids into dmA+ and dmA::Tn\0competent cells.

Transformingplasmid DNA

R6K pirWT

pMF26 pirWT

pMF26pi>113

pMF50 pirWT

pMF50 pirt.00

pFW8 p/VWT

PACYC177

R6Korigin(s)

a-y-fl

y-0

P

y

-

Penicillin GconcentrationGig/ml)

250125250125250125250125250125250125250125

Host

AQ634dnaA +

2.2X102

2.0X102

5.5X1O3

6.7X1O3

2.OX1O3

l . l x lO 3

7.2X1O2

4.3 X103

1.4X104

4.8x10*1.5X1O3

1.5X1O3

5.3X1O4

5.8X104

AQ3563dnaA::TnlQ

0000000000001.7x10*1.2X104

1.5X1O8 dnaA* and dnaA::Tn\0 competent cells were transformed with R6Kor the indicated plasmid deletion derivative (0.25 /ig of supercoiled, closed circularDNA). Numbers indicate colonies obtained per 11% of DNA per 4xlO 7 cellsplated on antibiotic-containing LB plates supplemented with 5 ^g/ml thymidine.

R6K and deletion derivatives containing one or two origins weretested for their ability to transform the isogenic AQ634 dnaA +

and AQ3563 dnaA::TnlO strains. Although R6K derivativescarrying the (3 origin (pMF50) or the /3 and y origins (pMF26)could be established in the dnaA+ strain, no transformants ofthe dnaAvJn 10 strain could be obtained, even at low (125 /xg/ml)penicillin levels (Table 2). Replication of the isolated 7 origin

by guest on August 14, 2016

http://nar.oxfordjournals.org/D

ownloaded from

Nucleic Acids Research, Vol. 20, No. 4 813

Table 3. Transformation efficiency of R6K y origin plasmid pMF35 into competent cells carrying resident x-producing plasmids.

TransformingDNA

pMF35

pUC9

Penicillin GconcentrationOig/ml)

250125250125

AQ634dnaA+

(pFWIO)

4.0X102

7.2 xlO3

4.2 xlO2

3.0X102

AQ634dnaA+

(pFWll)

6.0X102

5.3X1O2

3.5X1O2

4.2 XlO3

Host

AQ3563dnaA::TnlO(pFWIO)

002.9X1O3

1.2X1O2

AQ3563dnaA.-.TnlQ(pFWll)

001.2X102

1.2X102

1.5X1O8 competent dna/l + or dnaA/.TnlO cells carrying the resident x-producing plasmids pFWIO or pFWll were transformed with 0.25 /ig of closed circularplasmid DNA. Numbers indicate colonies obtained per pg of DNA per 4x 107 cells plated on antibiotic-containing LB plates supplemented with 5 fig/ml thymidine.

w—a—i

pFW8

pFWIO, pFWll, pMF39

H pMF26

pMF32

pMF33, pMF34, pMF35

H pMF50

pMF100

Figure 1. Replication region of E. coli plasmid R6K (not to scale). Double-headed arrows indicate cis-acting DNA segments required for replication from the a,/3 and y origins; the location of each Greek letter marks the approximate site of replication initiation for each origin. Dotted line separates the two cis-acting DNAsegments required for a origin replication (18). Sites shown here to bind DnaA (DnaA boxes) are labeled 1 and 2 (consensus sequences at - 2 4 to - 1 6 and 247to 255, respectively), a-i indicate DnaA consensus sequences that do not bind DnaA; dark hatched boxes contain one mismatch from the 9-bp consensus; lighterhatched boxes contain two. Sequences with mismatches underlined (and nt locations) are: a, TTATAAAAC (32-40); b, TTATATTAA (73-81); c, TTGTTCAAA(87-95); d, TTATCG.ACA (673-681); e, TTATTCAAA (829-837); f, TTATCAACT (976-984); y, TTATCTGAA (1301-1309); h, TTATCTAAA (1457-1465);i, GTATACAAA (1651-1659). Small squares labeled I and H represent IHF binding sites ihfl (which contains two adjacent fflF consensus sequences, 61-73and 76-88) and ihf2 (251-263) (20, 21). The seven black triangles indicate the 22-bp repeats that bind r. pir and bis indicate the structural genes encoding xand Bis proteins, respectively. Thin lines at the bottom of the figure indicate the R6K sequences included in various R6K derivative plasmids. Relevant restrictionsites are indicated: H2, HaeD (-700, 570 and 1961); E, EcoRI (-106); H, HindlD (0); S, SnaBI (243); B, Bgin (277); F2, FnuDD (325 and 1595); A, AccI(1653). The EcoRI (-106) site is artificial; the Bgin (277) site was destroyed in the process of cloning pMF32, pMF33, pMF34, and pMF35; the Sail site in theadjacent polylinker was used in digesting MF34 to obtain y origin fragments for footprint experiments (see Materials and Methods). In plasmid pMF39, EcoRIand BamHI sites in the polylinker adjoin each side of the FnuDD fragment containing the pir gene. For descriptions of the orientation of R6K fragments and constructionof plasmid pFW8, see Figure 2.

was tested using constructs that provide the R6K initiator proteinin cis (Table 2) or in trans to the origin (Table 3). Thetransformation data could indicate that neither R6K nor any ofits deletion derivatives can replicate in the absence of DnaA.

Next we determined if DnaA, like MF, was involved in controlof R6K replication. We have shown recently (21) that 7 originreplication in IHF-deficient strains is dependent on •w copy-upmutants. These mutants contain single amino acid changes thatallow an increase in plasmid copy number (24, 31, 32). Reducedlevels of wt v, known to allow an increase in copy number (26),also permit IHF-independent replication (21). Consequently weproposed that IHF restrains the negative control of replicationenforced by normal levels of wt IT protein (21). To determineif DnaA protein is similarly involved in replication control, wehave examined if R6K derivatives containing copy-up pir alleles

enable DnaA-independent replication. A derivative of 13-y originplasmid pMF26 containing the copy nip mutation pirl 13 [formerlypirl and pir\QA (24)], and a pMF50 03 origin) derivativecontaining the pir200 allele were thus transformed into AQ634and AQ3563. As shown in Table 2, none of the pir mutantproteins permitted replication of these R6K plasmidsindependently of DnaA protein. Thus the data accumulated sofar seem to indicate that a bypass of the requirement for DnaA,if it exists, cannot be achieved under conditions that allow bypassof the requirement for IHF.

DnaA binds to two sites within a minimal R6K repliconA common mechanism by which DnaA protein affects replicationinvolves binding of DnaA to an origin, with consequent effectson DNA structure and/or the binding of other proteins. There

by guest on August 14, 2016

http://nar.oxfordjournals.org/D

ownloaded from

814 Nucleic Acids Research, Vol. 20, No. 4

H..2

P«n

PFnuDU

EcoRI

1rnuDllj

Ec.RI

Figure 2. Constmction of 7 origin plasmid pFW8. pMF33 was constructed by inserting the EcoRI-BglU (—106 to 277) 7 origin fragment into pUC9 digested withEcoRI and BamHI. pMF35 was derived from plasmid pMF33 by deleting the Hae2 fragment containing the pUC origin. FnuDII indicates the junction betweenFnuDII and Hindi sites in pMF39. To make pFW8, the EcoRI fragment of pMF39 containing the pir gene was inserted into the single EcoRI site of pMF35. pFW8contains the pir gene in cis to the 383-bp 7 origin and therefore contains most, but not all, of the DNA segment required for replication from the |3 origin (17,19 and F. Wu, unpublished). The 0 origin requires the structural integrity of the DNA segment between Hindin (0) and the end of the 0 origin, which has beenalternately mapped to AccI (1652) (19 and F. Wu, unpublished) or HaeD (1961) (17). In an intact (3 origin, the pir gene lies to the right of the 7 origin; however,the pir gene in pFW8 lies to the left of the 7 origin. Therefore, we believe that pFW8 is replicating as a 7 origin and not a (3 origin.

Table 4. Transformation

TransformingDNA

pMF32pMF34

efficiency of plasmids

A22-pjVWT

03.OX1O2

carrying the truncated

A14-pirWT

1.6x10*7.6X103

and complete 7 origins (pMF32

HostA22-piV113

4.8X103

3.8X103

and pMF34, respectively).

622-pir\ 16

1.7x10*3.8X1O3

A22-/7iV20O

5.8X1O3

4.0 xlO3

Plasmids pMF32 and pMF34 (0.25 11% supercoiled, closed circular DNA) were transformed into 1.5 x 10s p3478 competent cells. Host cells harbored resident plasmidsproducing normal levels of wt T (A22-p<>WT), decreased levels of wt x (A14-pirWT), or normal levels of T mutants (A22-pirl 13, A22-p/rl 16, or A22-pi>200).Numbers indicate colonies obtained per y.g of DNA per 4 x 107 cells plated on LB plates containing 250 /ig/ml penicillin G and 5 fig/ml thymidine. As a control,no colonies were obtained when pUC9 was transformed into these cells. Also, similar numbers of colonies were obtained when transformed cells were plated ontoLB plates containing 125 jtg/ml penicillin G and 5 /ig/ml thymidine (data not shown).

were two reasons suggesting that DnaA might directly bind to7 origin DNA. First, sequence analysis revealed that the centralR6K 7 origin contains 5 putative DnaA consensus binding sites(DnaA boxes) with 1 or 2 mismatches (designated 1, a, b, c,and 2 in Figure 1). Second, the previously described 7l l l revlmutation changes the sequence of DnaA box 2 (nt +247 to +255)from £TATCAACA (two mismatches from the consensusunderlined) to QTATCAATA (three mismatches). This mutationimpairs replication of a 7 origin plasmid (33, 34). Thus we askedwhether purified DnaA protein could protect any of these putativebinding sites in DNase I footprinting experiments. DnaA wasfound to protect two areas with poorly defined boundaries butclearly identifiable DnaA boxes within each of protectedsegments. DnaA box 1 TTGTCCACA (-24 to -16) has onemismatch from consensus (underlined). The DnaA boxesdesignated a, /3 and c in Fig. 1 were not seen to be protectedin our assay, however, irregularly spaced enhancements can beclearly seen at the left-most boundary of the AT-rich domain(coordinates 0, +1) and well within it (coordinates +34, +39,+46 and +55) (Fig. 3). It is not yet known if these rightmostenhancements require the presence of the downstream DnaAboxes, or if DnaA is specifically interacting with the AT-richregion of the 7 origin as it interacts with the AT-rich region oforiC (8).

DnaA box 2 clearly resides within a second protected area(Fig. 4). The segment immediately upstream of DnaA box 2

shows periodic enhancements (coordinates +213, +221 to+222, +232 to +235, and +240 to +243). This large protectedregion punctuated by periodic enhancements is reminiscent ofthat seen when multiple DnaA monomers bound cooperativelyto oriC (6), to a region extending 40 to 50 bp to either side ofa DnaA box in pBR322 (6), and 30-bp region which lies betweenthe two promoters for the dnaA gene and surrounds a DnaA box(35).

A 106-bp segment containing DnaA box 1 is not absolutelyessential for R6K replicationAn experiment aimed to determine which of the two DnaA boxesis essential for R6K replication was to delete each of the segmentscontaining them, and ask if the resulting truncated origins remainfunctional. In order to determine if DnaA box 1 is essential forreplication, we made use of two isogenic plasmids pMF34 andpMF32 (Fig. 1). The former construct contains the entire EcoRI-BgUI 383-bp minimal 7 origin, cloned into pUC9; the latter lacksthe leftmost EcoRI-Hindin 106-bp segment of this origin andhence lacks DnaA box 1. pMF32 and pMF34 plasmids replicatevia the pUC origin if and only if DNA Polymerase I is produced.We could test the replication ability of the complete and truncated7 origins by transforming these plasmids into the polA 1 strainp3478. The recipient strain harbors one of several availableir-producing helper plasmids, because T can activate 7 originin trans and it does so even better when its intracellular

by guest on August 14, 2016

http://nar.oxfordjournals.org/D

ownloaded from

t 2 3

61

3 9 .3H_

a :

B 1 2 3 4

- 9 .

-31-1

-5»J

c /iritis Research, Vol. 20, No. 4 815

1 2 3 4

260_

265_

- 1 6 .

- 2 H - .

-H3.I

Figure 4. Footprint of the R6K 7 origin using DnaA protein and the Sail-labeledEcoRI-Sall fragment of pMF34. Lane 1, G+A sequencing; lane 2, DNasel control;lane 3, 0.14 jig DnaA; lane 4, 0.07 n% DnaA. DnaA box 2 is bracketed; numbersindicate nt location; (+) indicates an enhancement of cleavage in the presenceof DnaA; (—) indicates protection.

-6H.

Figure 3. Footprints of overlapping segments of the R6K y origin using DnaAand the EcoRJ-labeled EcoRI-Sall fragment of pMF34. Lanes Bl and Cl, G+Asequencing, lanes Al, B2 and C2, DNasel control; lanes A2, B3 and C3, 0.14Hg DnaA; lanes A3, B4 and C4, 0.07 fig DnaA. (+) indicates DNase Ienhancements in presence of DnaA; ( - ) indicates protected bases; numbers indicatent locations. DnaA box 1 is bracketed.

concentration is reduced (26). As shown in Table 4, pMF34,which contains a complete 7 origin, can be established in allrecipients producing either different concentrations of -w ormutated x variants. In contrast the plasmid carrying the truncated7 origin, pMF32, can be established only in cells producing Tcopy-up mutants or decreased levels of wt TT. Consequently, wecall the 211-bp HindlU-Bglll fragment the core origin. A similaroverall pattern of transformation frequencies with plasmidspMF32 and pMF34 was observed (data not shown) in the E. coliC background (C2110 polAl) commonly used in our previousstudies (26, 33).

Similar attempts to delete the 34-bp SnaBI-BgUI right-mostsegment of the 7 origin, which contains DnaA box 2, failed toyield a functional origin. We could not rescue an active 7 origin

replicon when this 34-bp fragment was deleted from pMFlOO(Fig. 1) in polAX hosts containing helper plasmids producingnormal or decreased levels of wt T (M. Filutowicz, unpublished).

There are three major conclusions from these data. First, DnaAbox 1 is not absolutely essential for 7 origin activity. Second,the sequence requirements for the minimal R6K replicon(7 origin) can be reduced from 383 bp to 277 bp by altering •w,either quantitatively by reducing concentration of wt protein, orqualitatively by providing copy-up -K variants. The latter pointsheds important new light on the definition of the boundaries ofR6K origins and allows us to reinterpret some puzzlingobservations published in the past (see Discussion). Finally, the7 origin requires the 34-bp fragment containing the DnaA box 2.

DnaA does not alter ir productionSeveral putative DnaA boxes within the pir-bis operon have beenidentified by computer analysis of the available R6K sequence(Figure 1). Two boxes (d and e) lying within the pir codingsegment contain one mismatch from the DnaA consensus; twoadditional sequences (f and g) each contain two mismatches. Thebis gene (17, 19, 36) also contains two putative DnaA boxes(h and i), each having one mismatch from the DnaA consensussequence. Because DnaA is able to bind to regulatory sitescontaining DnaA boxes and thereby mediate transcriptionalrepression of certain genes (35, 37-40), we asked if DnaAprotein could also influence the expression of w protein. If itcould, this second regulatory circuit, in addition to autoregulation

by guest on August 14, 2016

http://nar.oxfordjournals.org/D

ownloaded from

816 Nucleic Acids Research, Vol. 20, No. 4

AQ2563 AQ634dnaA TniO *

Ru

-chromosomal

W t Tl —•

n —

Figure 5. A, immiinoblot assays showing levels of T produoed by plasmids pFWIOand pFWll in AQ634 and AQ3563. Two bands crossreacting with x IgG'scorrespond to the 305-amino acid full-size polypeptide (wt x) and 206-aminoacid short protein variant (x*) produced from a second translational start pointwithin the pir gene (D. York, J. Hoffman, J. Gan and M. Hlutowicz, ms. inpreparation). From the somewhat higher level of x produced by pFWl 1 in AQ634,we suspect that transcription from the TetRj gene in the vector may in part overrideautoregulation of the pir promoter. B, levels of plasmids pFWIO and pFWl 1in AQ634 and AQ3563. Chromosomal and plasmid (OC, open circular; CCC,covalently closed circular) DNA bands arc indicated.

(41—43), would help in preventing the intracellular Tconcentration from climbing to levels high enough to inhibitreplication (26).

To test the possible involvement of DnaA in the regulation ofpir gene expression, the levels of T protein in dnoA+ andisogenic dnaA 'null' cells harboring pFWIO or pFWll helperplasmids were measured by immunoblotting. These plasmidscontain the pir gene, including its regulatory sequences, clonedin either orientation into the pACYC184 vector (See Table 1 forcloning strategy). Immunoblotting assays demonstrated that thelevels of x produced from these plasmids in AQ634 and AQ3563were similar (Fig. 5A). The copy number of each plasmid wasalso nearly equivalent in both strains (Fig. 5B). We also foundthat the copy number of plasmid pMF39 (24), a pUC-derivativecarrying the pir gene, was four- to five-fold lower in the absenceof DnaA than in its presence; w production was proportional toplasmid copy number (data not shown). Hence, when v synthesisis uncoupled from R6K replication, DnaA protein appears to havelittle or no effect on the expression of the pir gene. Therefore,the inability of R6K or its derivatives to replicate in the DnaA-deficient strain cannot be explained by an effect of DnaA on Tproduction.

DISCUSSION

Plasmid R6K requires DnaA for replication, because neither R6Knor deletion derivatives could replicate in a dnaAy.TnlO strain(28 and this study). Another, less likely, intrepretation of thesetransformation data is that DnaA may be required not for thereplication but for the establishment of these plasmids. DnaA doesnot alter the level of production of R6K initiator protein T.Instead, we propose that the major role of DnaA is binding toa single site, DnaA box 2, in a 277-bp fragment common to allthree R6K origins. The adjacent 106-bp origin fragment alsocontains a site, DnaA box 1, to which DnaA protein binds invitro. The central role of DnaA binding to box 2 is supportedby two lines of evidence. First, the 106-bp fragment containingDnaA box 1 is not required for replication of the R6K a origin

(18) or the /3 origin (17, 19). We show here that it is also notnecessary for replication of the y origin when certain T mutantsor decreased levels of wt x are provided in trans. Implicationsof these observations will be discussed later. Second, pointmutations in the DnaA box 2 that increase or decrease the numberof matches to the 9-bp DnaA consensus sequence respectivelyincrease or decrease the copy number of 7 origin plasmids. Thepreviously described 7l l l revl mutation changes the sequenceof the DnaA box 2 from two mismatches to three mismatches.This mutation impairs R6K replication, reducing the copy numberof plasmids replicating from the or-g and 7-/3 origins (34) or the7 origin alone (33, 34). A mutation improving the match of DnaAbox 2 to its consensus (£TATCCACA, one mismatch underlined)and decreasing the homology of the adjacent IHF binding siteihf2 to the IHF consensus replicates at a somewhat higher copynumber than isogenic plasmids carrying either the wt 7 originor a 7 origin mutated only in the ihf2 (S. Dellis, T. Schatz andM. Filutowicz, unpublished data).

The role of the 106-bp origin enhancer in 7 origin replicationIn this paper we present evidence that the 383-bp 7 origin canbe functionally separated into at least two domains: the core 7origin, which is essential for replication under all conditions sofar tested; and the adjacent 106-bp segment, which we call theorigin enhancer, which is required for replication of the 7 originwhen normal levels (4000-10,000 dimers per cell) (26) of wtx are provided in trans. However, the core 7 origin is able toreplicate under two conditions: (1) when decreased levels of wt•K are supplied or (2) when certain T mutants are provided intrans. These two conditions allow an increase in 7 origin plasmidcopy number over that seen when normal levels of wt T areprovided, consistent with previous observations indicating thatnormal levels of wt ir partially inhibit replication; this inhibitionis apparently counteracted by the enhancer.

A role for the enhancer is suggested by an earlier finding thatthe 7 origin can function when the transposon Tn5 is insertedbetween the core origin and the enhancer, or within the enhanceritself (45). We propose that the outer end of the transposon mayfunctionally substitute for the enhancer by providing a perfectDnaA box (TTATACACA) (46).

Alternatively, or in addition to this, the enhancer may behelping the 7 origin adopt an unusual structure which permitsreplication at inhibitory levels of wt v. It has been shown (47)that IHF can fold the 7 origin only when the enhancer and theseven repeats are present. Like the enhancer, IHF helps restrainthe inhibitory effect of T (21). Thus an IHF-enhancernucleoprotein structure may provide a mechanism for alleviatingTT inhibition of replication. Finally, IHF binding may allow long-range interactions between DnaA molecules bound to DnaA box1 in the enhancer and DnaA box 2 or DnaA and x protein. Theseputative DnaA-IHF-p interactions may be similar to those seenin vitro in the pSClOl origin (48).

The findings reported here also cast new light on previousaccounts of in cis incompatibility (49). 7 origin replicationrequires a cluster of repeats that act as binding sites for T, butan extra set of repeats provided in cis may inhibit replication,causing incompatibility depending on location and orientation.•K molecules can mediate coupling of two repeats clusters on thesame plasmid (50), looping out the intervening DNA segment.Like plasmids carrying the core origin, 7 origin plasmidscontaining an extra set of repeats on the enhancer side of theorigin cannot replicate when normal levels of wt w are provided

by guest on August 14, 2016

http://nar.oxfordjournals.org/D

ownloaded from

Nucleic Acids Research, Vol. 20, No. 4 817

in trans, but can when decreased levels of wt T, or IT copy-upmutants are provided in trans. Therefore, it seems possible thatds-association or 'handcuffing' of repeats (49) may inhibitreplication not by sterically hindering access of replicationproteins to the origin, as previously proposed (49); ratherhandcuffing may prevent replication by looping out, distortingor constraining the enhancer segment and AT-rich region of they origin.

Finally, the finding that the enhancer is not required for y originreplication brings into question the traditional definition of the/3 origin. While the y origin was believed to require the enhancer(16), the /? origin was believed to contain the DNA segmentextending from the core origin rightward (17, 19; see Fig. 1).Our finding that certain levels of T allow replication of the coreorigin suggests that the boundaries of the fi origin need to bereconsidered, using plasmid constructs producing quantified levelsof ir.

ACKNOWLEDGEMENTS

We thank E. Boye for DnaA protein and B. Kline for bacterialstrains. Dya Goldberg was a participant in an honors thesisprogram in the Department of Biochemistry, University ofWisconsin, Madison. This research was supported by grants toM.F. from NIH (GM40314) and NCS (IN-35-30-24).

REFERENCES

1. Tomizawa, J. and Selzer, G. (1979) Ann. Rev. Biochem 48, 999-1034.2. Fuller, R. S. and Kornberg, A. (1983) Proc. Nail. Acad. Sd. USA 80,

5817-5821.3. Kaguni, J. M., Fuller, R. S. and Komberg, A. (1982) Nature 296, 623-627.4. Fuller, R. S., Kaguni, J. M. and Kornberg, A. (1981) Proc. Natl. Acad.

Sd. USA 78, 7370-7374.5. Zyskind, J. W., Cleary, J. M., Brusilow, W. S. A., Harding, N. E. and

Smith, D. W. (1983) Proc. Natl. Acad. Sd. USA 80, 1164-1168.6. Fuller, R. S., Funnell, B. E. and Kornberg, A. (1984) Cell 38, 889-900.7. Bramhill, D. and Kornberg, A. (1988) Cell 52, 743-755.8. Yung, B. Y.-M. and Kornberg, A. (1989)7. Biol. Chem. 264, 6146-6150.9. Baker, T., Funnell and Komberg, A. (1987) 7. WoL Chem. 262, 6877-6885.

10. Bramhill, D. and Kornberg, A. (1988) Cell 54, 915-918.11. Nishimura, Y., Caro, L., Berg, C. M. and Hirota, Y. (1971)7. MoL Biol.

55, 441 -456.12. Chesney, R. H., Scott, J. R. and Vapnek, D. (1979) J. Mol. Biol. 130,

161-173.13. Chattoraj, D., Cordes, K. and Abeles, A. (1984) Proc. Natl. Acad. Sd.

USA 81, 6456-6460.14. Hansen, E. B. and Yarmolinsky, M. B. (1986) Proc. Natl. Acad. Sd. USA

83, 4423-4427.15. Wickner, S., Hoskins, J., Chattoraj, D. and McKenney, K. (1990)/. Biol.

Chem. 265, 11622-11627.16. Kolter, R. and Helinski, D. R. (1978) Plasmid 1, 571-580.17. Shafferman, A. and Helinski, D. R. (1983)/ Biol. Chem 258,4083-4090.18. Shafferman, A., Flashner, Y., Hertman, I. and Lion, M. (1987) MoL Gen.

Genet. 208, 263-270.19. Shon, M., Germino, J. and Bastia, D. (1982) 7. Bid. Chem 257,

13823-13827.20. Filutowicz, M. and Appelt, K. (1988) Nudeic Acids Res. 16, 3829-3841.21. Dellis, S. and Filutowicz, M. (1991)7. Bacteriol. 173, 1279-1286.22. MacAlIister, T. W., Kelley, W. L., Miron, A., Stenzel, T. T. and Bastia,

D. (1991) 7. Biol. Chem. 266, 16056-16062.23. Sambrook, J.. Fritsch, E. F. and Maniatis, T. (1989) Molecular Cloning:

A Laboratory Manual, Cold Spring Harbor Press, Cold Spring Harbor.24. Filutowicr, M., Uhlenhopp, E. and Helinski, D. R. (1985) 7. Mol. Biol.

187, 225-239.25. Maxam, A. M. and Gilbert, W. (1980) Meth. Enzymol. 65, 499-560.26. Filutowicz, M., McEachern, M. J. and Helinski, D. R. (1986) Proc. Natl.

Acad. Sd. USA 83, 9645-9649.27. Bimboim, H. C. and Doly, J. (1979) Nucleic Acids Res. 7, 1513-1523.

28. Kline, B. C. (1988) Can. 7. MicrobioL 34, 526-535.29. Crosa, J. H., Luttropp, L. K. and Falkow, S. (1976) 7. Bacteriol. 126,

454-466.30. Crosa, J. (1980) 7. Biol. Chem. 255, 11075-11077.31. Stalker, D. M., Filutowicz, M. and Helinski, D. R. (1983) Proc. NatL Acad

Sd. USA 80, 550T5504.32. Inuzuka, M. and Wada, Y. (1985) EMBO 7. 4, 2301-2307.33. McEachem, M. J., Filutowicz, M. and Helinski, D. R. (1985) Proc. Natl.

Acad. Sd. USA 82, 1480-1484.34. McEachem, M. J. (1987) Ph. D. thesis, University of California, San Diego.35. Braun, R. E., O'Day, K. and Wright, A. (1985) Cell 40, 159-169.36. Mukhopadhay, P., Filutowicz, M. and Helinski, D. R. (1986)7. BioL Chan.

261, 9534-9539.37. Lother, H., Kolling, R., KOcherer, C. and Schauzu, M. (1985) EMBO J.

4, 555-560.38. Adung, T., Clausen, E. S. and Hansen, F. G. (1984) Adv. Exp. Med. Biol.

179, 199-207.39. Atlung, T., Clausen, E. S. and Hansen, F. G. (1985) Mol. Gen. Genet.

200, 442-450. -40. Polaczek, P. and Wright, A. (1990) New Biologist 2, 574-582.41. Shafferman, A., Kolter, R., Stalker, D. M. and Helinski.D. R. (1982) 7.

Mol. Biol. 161, 57 r 76.42. Kelley, W. and Bastia, D. (1985) Proc. Natl. Acad ScL USA&, 2574-2578.43. Filutowicz, M., Davis, G., Greener, A. and Helinski, D. (1985) Nucleic

Adds Res. 13, 103-114.44. Chang, A. C. Y. and Cohen, S. N. (1978)7. Bacteriol. 134, 1141-1156.45. Kolter, R. and Helinski, D. R. (1982) 7. MoL Biol. 161, 45-56.46. Yin, J. C. P. and Reznikoff, W. S. (1987) 7. Bacteriol. 169, 4637-4645.47. Filutowicz, M. andlnman, R. B. (1991)7. Biol. Chem. 266, 24077-24083.48. Stenzel, T. T., MacAlIister, T. and Bastia, D. (1991) Genes and Development

5, 1453-1463.49. McEachem, M. J., Bott, M. A., Tooker, P. A. and Helinski, D. R. (1989)

Proc. Natl. Acad. Sd. USA 86, 7942-7946.50. Mukherjee, S., Erickson, H. and Bastia, D. (1988) Cell 52, 375-383.51. Greener, A., Filutowicz, M. S., McEachem, M. J. and Helinski, D. R.

(1990) Mol. Gen. Genet. 224, 24-32.52. Kontomichalou, P., Mhani, M. and Clowes, R. C. (1970)7. Bacteriol. 104,

34-44.53. Kline, B. C , Kogoma, T., Tarn, J. E. and Shields, M. S. (1986)7. BacterioL

168, 440-443.54. Kolter, R., and Helinski, D. R. (1978) 7. Mol. BioL 124, 425-441.55. DeLucia, P. and Cairns, J. (1969) Nature 224: 1164-1166.

by guest on August 14, 2016

http://nar.oxfordjournals.org/D

ownloaded from