(~) [NSTITUTPAsrEoR/EI.SEVlEa

Res. &iicrubiol. 1991, 142, 209-216

Paris 1991

Unique sequence requirements for the P1 plasmid replication origin T . G . Brendler, A . L . Abeles, L . D . Reaves a n d S . J . A u s t i n Laboratory o f Chromosome Biology, 4BL-Basic Research Program, NCl-Frederick Cancer Research Facility, Frederick, M D 21702 USA

SUMMARY We have carried out a detailed genetic analysis of the P1 plasmid replication origin and shown that it has four major structural requirements: the DnaA box, a aeries of five 7-base pair (bp) repeats, a GC-rich spacer and five 19-bp repeats that bind the P1 flepA protein. The origin requires the DnaA protein and its racognltlon sequence ('~hcDncA box). ,~;owever, a|though five boxes are present in two separate blocks tn the wild type, just one, peacedeither to the left or right of Ihe core origin sequences, is sufficient/or function as long as Ihe box conforms exactly to the proposed consensus. Each of the five 7-bp repeats that constitute the core of the origin is required; mutations within any of the first six base pairs can block origin function. The required bases include, but are not limited to, those constituting dam mathylation sites. Also essential is a 39-bp GC-rich sequence. We show this to be a spacer of critical length that separates the 7-bp repeats from the last required region; a series of 19-bp repeats that bind the P1 RepA initiator protein. Key-worda : Pl, Plasmid, Replication; Genetic analysis.

INTRODUCTION The PI plasmid replication origin is subject to stringent regulation that ensures that a copy number of only one per unit celt is maintained. In this respect, it resembles the host chromosome origin oriC with which it shares some structural features. Like oriC, the P1 origin requires the host DnaA protein and has multiple adenine-methylation sites which must be methylated for proper function. Unlike oriC, the Pt origin also requires the action of a Pl-encoded initiator protein, RepA. T h e m i n i m a l PI p l a s m i d replicon (the PI-R replicon) has

been defined and studied in isolation (Abeles el ak, 1984). It consists of an origin of replication followed by the gene for the essential RepA protein and a copy-control locus, incA. The origin consists of a maximum of 254 bp (Abeles et aL, 1984; fig. l). A central core region is distinguished by a group of five 7-bp direct repeats including four with nested G A T C sequences that are s u b s t r a t e s for the Escherichia coil DNA deoxyadenine mcthylase (Hattman el ak, L978; Geier and Modrich, 1979). An additional unnested G A T C sequence lies close by. To its right is a 39-bp GC rich region followed by a group of five

19-bp repeats (Abeles et aL, 1984) that are specifically bound by the essential Pt replication initiator protein, RepA (Abeles et aL, 1989). The leftmost portion of the origin contains a tandem direct repeat of the putative DnaA recognition sequence (the DnaA box) (Fuller et aL, 1984). An additional region containing three closely apposed repeats matching or resembling DnaA boxes is present outside and to the right of the minimal functional origin. It lies in the untranslated leader region of the repA gene (Hansen and Yarmolinsky, 1986; fig. IL The DnaA protein is required for P] plasmid replication, both in vivo

21o

T.G. BRENDLER ET AL.

Plori 400

500

l

I

600

I 19 bp. RepA repeats

7 bp (GATC)

t,,, ,I I /

ft I s

1 / t l

S

reDA DnaA

"

/

EcoRI Fig. I. The Pl pIasmid replication origin and associated sequences. Stippled boxes are the 7-bp imperfect repeats. All but the second box contain GATC adenine melhylation sites. Open arrows are potential DnaA boxes. Shaded arrows are 19-bp RepA binding repeats. Vertical shading distinguishes Ihe 39-bp GC-rich sequence, The upper part of the map represents the minimal origin as originally defined (Chauoraj et aL, 19851. The lower part shows associated sequences including the rightward set of DnaA boxes and the start of the repA gene. Cot*rdinates are in standard base pair units (Abeles el al., 19841.

and in vitro (Hansen and Yarmolinsky, 1986; Wickner and Chattoraj, 19871. We have carried oat detailed genetic analyses of the individual domains of the origin (Abeles el al., 1990; Brendler et aL, manuscript in preparation}. Here we present the overall picture obtained in less detail and attempt to relate it to currem models of origin function. MATERIALS AND METHODS Buffers, media, enzymes and reagents. - - Conditions for cell growth and buffers used for cl~ctrophoresis and DNA preparation were previously described

(Maniatis el aL~ 1982; Abeles and Austin, 1987; Abeles et al,, 1989, Abeles el al., 19901. Enzymes were obtained from US Biochemical Corporation (Cleveland, OH), Bcthesda Research Laboratories, Inc. (Gaithersburg, MD), New England Biolabs, Inc. (Beverly, MA), Boehringer Mannheim Biocbemicals (Indianapolis, IN) or Pharmacia LKB Biotechnology, Inc. (Piscataway, N J) and used according to the suppliers' recommendations. Details of other reagents ar.d materials were as previously described (Abeles~ 1986; Abeles and Austin, 1987; Abeles et al., 1989; Abeles et at., 1990).

Bacterial s t r a i n s . - - E , eoli strain DH5 (recAI) (Hanahan, 1983, 19851 was used for DNA cloning and plasmid preparation. Pl oriR function was tested in the strain CC708 (BR827, Abeles et el., 19901 that is polA12 (ts) that contains the plasmid pALA619 that supplies the RepA protein in trans.

DNA

preparation and

sequencing, - - Plasmid DNA was prepared as described (Manintis el aL, 19821, Preparation of DNA sequencing templates from supereoiled DNA was by the method of Chen and Seeburg (19851. DNA was sequenced by the dideoxy method using "Sequenase" modified T7 DNA polymerase (Tabor and Richardson, 19871. Deoxyoligonucleotides were synthesized on an " A p p l i e d Biosystems" DNA synthesizer (Model 380B) using the phosphite trlester method (Matteucci and Carutbers, 1981; Beaucage and Caruthers, 19811. The oligonucleotides were purified by Scphadcx G-25 chromatography and kinased according to Abeles et aL (19891. Complementary oligonucleotides were annealed under the conditions described (Wosnick et aL, 19871. Plasmids and cloning vectors. - - Plasmids p U C I 8 (YanishPerron et al., 19851, p U C I 9 (Yanish-Perron el al., 19851 and pALA619 (Abeles el al., 1990) which contains the gene for the PI rcpA protein in a pSC|01-based vector were previously described. The PI origin plasmids were derived from the PI minimal origin clone pALA618, a pUCI9 derivative (Abeles et al., 19901. The wild type origin derivative pALA630 was generated by inserting the PI oriR sequence from bp 386 to bp 502 between the Hindlll (pUt19) site and Nrul (PI bp 502) site of pALA618 using a double-stranded oligonucleotide with a HindIIl site at the PI bp

SEQUENCE REQUIREMENTS FOR P/ PLAS,'vI/D REPLICA T/ON OR/GIN 386 end. Variants with additional restriction sites were constructed similarly in order to facilitate replacement of various regions o f the o r i g i n with m u t a n t oligonucleotides. Plasmid pALA6M gains an Sspl site by an A to T mutation at bp 419 and pALA632 gains a Bglll site by changing PI hp 414 from a C to A. Plasmid pALAI204 gains a Clal site by changing bp ,166 from T to A. Derivatives with mutations of the origin core were i s o l a t e d by i n t r o d u c t i o n o f double-stranded oligonucleotides replacing the Bglll to Nrul interval of pALA632. These were designed so as to correct base 414 to wild type. Constructs varying the number and position of the leftmost DnaA boxes were made by replacing the Hindlll to Sspl interval of pALA631 with specific oligonucleotides. In this case, all the derivatives retain the silent mutation at h p 419. The construction of clones containing the rightmost group o f DnaA "Joxes and variants of it was ~reviously described (Abeles et at., 1990). Changes in the length and composition of the GC-rich spacer were achieved by replacing the Clal to Nrul interval of p A L A I 2 0 4 with the relevant sequences present on a doublestranded oligonucleotide, restoring the wild type base at bp 406. Random mntagenesis o[ P1 minimal origin with degenerate oligonucleotides. - - The region between PI bp 414 and bp 500 was targeted for random mutagenesis. This region spans four o f the five 7-bp tandem repeats, four of the five dam sites and a GC-rinh region (fig. 1). A mixed population o f oligonueleotides which contained a low level of contamination by the other three nucleotides at each position was synthesized (Oliphant et al., 1986; Hill et aL, 1987). Priming of the oligonucleotide with the universal MI3 primer, synthesis of the complementary strand, and cloning of the resultant

B a m H l / N r u l fragments were as described (Schneider and Stormo, 1989; Oliphant et aL, 1986; Hill et ul., 1987). In rive assay for P1 oriR function. - - The in rive replication assay was performed as described in Abeics et al. (Abeles et at,, 1990).

211

1o

~

i

~

I

class i class 2

7~ ol ~ a- 001

Cla~*, 3 pucls

ooo~

5

RESULTS

Analysis of the PI plssmid origin We have described a convenient system to assay the function of the PI plasmid origin in rive (Abeles et at., 1990). The minimal Pl origin or its mutant derivatives are cloned into the vector pUCI9. The plasmids are introduced by t r a n s f o r m a t i o n into a strain in which D N A polymerase [ is temperature-sensitive (CC708 polATsl2). The strain contains a compatible plasmid supplying the PI RepA protein in trans. Origin activity can be assayed in two ways. First, transformation frequency is measured at 42°C. The PI origin must be functional if transformants arise, because the pUC19 origin cannot work in the absence of DNA polymerase 1. Second, transfermams can be isolated at 30°C where the p U C I 9 origin works and the stability of the plasmid is then followed after a shift to 42°C. This gives a quantitative estimate of Pl origin function because the rate of loss is determined by the average copy number that the PI origin can sustain (Abeles er aL, 1990). Figure 2 shows the typical results obtained with the quantitative assay. Origin mutants that have a high rate of loss (class 3L resembling that of the p U C I 9 vector, are defective. Origin mutams that have low rates of loss (class 1), resembling that of the wild type PI origin construct, are functional. Mutants

lo

15

generations Fig. 2. Typical resnlts obtained in a

quantitative assay for origin function. The curves represent the proportion of cells retaining ~he plasmid in nonselective culture after a shift to 42~C. Results with the wild type PI origin plasmid pALA630 and the vector pUCI9 are shown. In the latter case, the ultimate rate of loss is as predicted for a plasmid that does not replicate at a[[ (the slope of log, [retention] ~s. number of generations ~ - l). Also showrt are results typical for class l, 2 and 3 origin mutations. The examples given are mutations A ~ T , 419; C ~ T , 427; and C-,T,457, respectively.

that give intermediate stability (class 2), are presumed to be partially defective for origin function. These mutants also give a characteristic phenotype in the transformation assay; they transform at relatively normal frequency at 42°C but form very small colonies on the selective medium. The DnaA boxes In order to test the functional significance of the two groups of DnaA boxes, w e constructed a number of deletion and rearrangement derivatives of the origin and tested their function_ The constructs are illustrated in figure 3 along with the results of t rans forma t i on assays in the

T.G, BRENDLER ET AL

212 o~

o~

500

.,£

'; ~,4~,

~ ~

Transformant5

Y" r e p A

~1 I,1~ PAt A655 DALA63 !

~

+ ~ ~

700

I

DALA657

39~__=

~

pALA645

~9=¢~-

~

pAL A654

(T[ATC/hCAC/AA; Fuller et al., 1984). A construct with a

~c~----- ~

DALA662

~ ~

pAL A646

t__

I"t+

-I-~

-I-

~

Fig, 3. Origin variants with different DnaA box configurations. All of tile constructs have the region (top line) that encompasses Ihe 7-bp repeats (stippled boxes) and 19-bp RepA-binding repeats (shaded arrows). These sequences are abbreviated as a dotted-line box in the construct maps below. Potential DnaA boxes are shown as open arrows, The results of transformation lasts in strain CC708 at 42°C are shown ( - = no transformants; = normal number relative IO wild type PI origin construct1. Aslerisk indicates that Ihe construct pALA654 has the consensus DnaA ha~ q-IATCCACA, h differs by one base from pALA645, whose box has the sequence no! .~lly preseln at this lOcUSITTATCCACTI, pALA662 also has a consensus box.

single box positioned to the left of the origin core is functional when the box has the sequence 5'TTATCCACA (pALA654), but not when it has the sequence $ ' T T A T C C A C T (pALA645), the sequence normally present in this position. From this study, it can be concluded that neither the number or position of DnaA boxes is very critical for origin function. However, the alternative arrangements of DnaA boxes do have some quantitative consequences as shown in the maintenance assays in figure 4. Pinsmid pALA658 that contains all five of the natural DnaA boxes is the most stable. Plasmids with just the two leftmost boxes (pALA631) and just one consensus box (pALA654) show progressive instability, suggesting that the number of DnaA boxes dictates the initiation frequency and, therefore, the relative copy

.1

polAts strain at 42°C. The wild type construct (pALA658) encompasses all of the previously defined origin sequences and additional bases to its right, including the second group of DnaA boxes and part of the open reading frame for the P[ RepA protein. It transforms efficiently, showing that the origin is active. As expected, transformation is dependent upon the PI RepA protein that the strain supplies in traus (data not shown), The other constructs all have the same core origin sequences but have been modified by oligonucleotide-diremed mutagenesis

to vary the number and position of the DnaA boxes. From the data, we conclude that there is considerable redundancy with respect to the boxes. The presence of boxes is essential, as a construct that lacks all of them is inactive (pALA646). However, a functional origin can be constructed with either the left or the right set of DnaA boxes. M o r e o v e r , constructs with just a single box can be active, either to the left or right of the origin core. In this case, it is essential that the single box exactly matches the published D n a A box c o n s e n s u s

~_ 01

.OOl

5 to generations

t5

Fig. 4. Maintenance assays of plasmids with different OnaA box configurations. The structures of the plasmids are shown in figure 3. Open squares: pALA658 (five bo:~es). Open circles: pALA631 Itwo boxes to the left of origin). Closed circles: pALA654 tone consensus box). Closed squares: pALA646 (no boxes).

SEQUENCE REQUIREMENTS FOR PI PLASMID REPLICA TION ORIGIN

number of the constructs when a fixed level of RepA protein is supplied. Nevertheless, it is clear that a single box couslru~,t is proficient for some degree of initiation; a finding that has theoretical implications discussed below.

Mntagenesis of the origin core The 85 bp comprising the 7.bp repeats and the adjacent GC-rich region were subjected to mutagenesls by the introduction o f random bases in a synthetic oligonueleotide (figure 5). The synthesis was designed to give an average of one base change per molecule. After synthesis of a complementary D N A strand, the double-stranded o!igonucleot ides were ligated into a pUCl9-based plasmid in such a way as to restore a complete PI plasmid origin sequence with the random mutations in place. The D N A content o f cells transformed with the ligaled products was screened

by filter hybridization using labelled s~nthetie oligonucleotide~ c'~rresponding to short :atervals of the wll(l type region q hybridization probes. Using conditions where mismatched bases partially inhibit hybridization, we were able to eliminate most of the clones that had no insert or had b~se changes in more than one interval. Fifty of the remaining clones were cultured; plasmid DNA was prepared from them and the DNA sequence of the origin region determined for each. Seventeen of the resulting clones had single base-change mutations. We also found six additional clones that had single base-deletion mutations (fig. 5). These presumably arose from errors in oligonucleotide synthesis. The single basec h a n g e m u t a n t s were supplemented by five more isolated by specific oligonueleotide mutagenesis. Each of the total of 28 mutants was tested for origin function as described above. As reproduction of the graphical

Ga i ~

AT

G

o cac'f~AT~i~,'r~Tc~'Ac~il~rr ,,o TT LT Lo qr tATCCaC,~rrATC Dfla.a,

DnaA

!!

AA A

T~

a

C

~!

G CG a

Fig. 5. Mutations of the 7-bp repeat and GC-rich regions. The DnaA boxes (open boxed regions) and %bp repeats (stippled boxes) are marked. The GATC methylafion sites are underlined. Arrows show the single base-change mutations. Upward arrows indicate class 1 mutations/little or no effect on origin function). Full