THEJOURNAL OF BIOLOGICAL CHEMISTRY Vol. 262, No. 5, Issue of February 15,pp.

2304-2309 1987
0 1987 by The American Society of Biological Chemists, Inc. Printed in ~ . . s . A .

Formation of Rolling-circle Molecules during 4x174 Complementary

Strand DNA Replication*
(Received for publication, June 18, 1986)

Minsen MokS and Kenneth J. Mariansg

From the Graduate Program in Molecular Biology,Memorial Sloan-Kettering Cancer Center, New York, New York 10021

The primosome is a mobile multiprotein priming ap- also possibly involvedin DNA repair(12). On the other hand,
paratus that requires seven Escherichia coli proteins the phage T 7 gene 4 protein (13, 14) and phage T4 gene 411
for assembly (the products of the dnaB, dnaC and dnaG 61 proteins (15-18) are unique enzymes possessing both hel-
genes; replication factor Y (protein n‘); and proteins i, icase activity (19, 20) and primase activity(15, 21-23). These
n, and n”). While the primosome is analagous to the
phage T7 gene 4 protein and phage T4 gene 41/61 proteins actat thereplication fork while bound to the lagging-
proteins in its DNA G-catalyzed priming function, its strand template functioning both to unwind the parental DNA
ability to act similarly also as a DNA helicase has and synthesize primers that are used to initiate synthesis of
remained equivocal. The role of the primosome in un- Okazaki fragments. The equivalent E. coli replication pro-
winding duplex DNA strands was investigated in the tein(s) is the primosome (24), a multi-enzyme mobile priming
coliphage 4x174 SS(c)+ replicative form DNA repli- apparatus discovered because of its requirement for the initi-
cation reaction in vitro, which requires the E. coli ation of 6x174 SS(c) + R F DNA replication (25, 26). Al-
single-stranded DNA binding protein, the primosomal though the DNA G protein-catalyzed priming activityof the
proteins, and the DNA polymerase I11 holoenzyme. primosome has long been documented (25, 26), its function
Multigenome-length, linear, double-stranded DNA as a helicase has remainedequivocal. Recently, LeBowitz and
molecules were generated in thisreaction, presumably McMacken (27) have demonstrated that the DNAB protein,
via a rolling circle-type mechanism. Synthesis of these one of the components of the primosome, has an intrinsic
products required the presence of a helicase-catalyzed
strand-displacement activity to permit multiple cycles helicase activity.
of continuous complementary (-) strand synthesis. The Studies from this laboratory have demonstrated that low
participation of the primosome in this helicase activity concentrations of topoisomerase I are required to generate
was supported bydemonstrating that other SS(c) DNA template specificity in the reconstituted pBR322 DNA repli-
templates (G4 and a-3),which lack primosome assem- cation system(28). One of the major synthetic products under
bly sites, failed to support significant linear multimer these conditionswas multigenome-length, linear,duplex DNA
production and that replication of 4x174 with the molecules presumably formed by a rolling circle-type replica-
general priming system (the DNA B and DNA G pro- tion mechanism. The appearance of these large molecules
teins and DNA polymerase I11 holoenzyme) resulted in required the primosomal proteins.’ Since rolling-circle DNA
a 13-fold lower rate of linear multimer synthesis. replication requires a helicase-type activity, it was inferred
~ ~~~ that the primosome, or some sub-assembly or component of
the primosome, was the agent responsible (28).
The replicationof duplex DNA molecules requires that the The4x174SS(c) + R F DNA replication system was
parental strands be unwound at the replication fork. Enzy- chosen for preliminary evaluation of the ability of the pri-
matic activities, termed helicases, that denature duplex DNA mosome to act as a helicase because it had simpler protein
in an ATP-dependent fashion have been identified in Esche- requirements than thepBR322 DNA replication system and
richia coli and in phage-infectedE. coli. did not generatea complex spectrum of synthetic products. It
E. coli helicase I (1)is encoded by the F factor (2) and is was expected that helicase activity would be manifested dif-
assumed to be involved in transfer of DNA strands during ferently in the 4x174 DNA replication system because of the
conjugation. Helicase I1 (3) is the uurD gene product (4-7) use of single-stranded DNA templates as opposed to double-
and is probablyinvolved inDNA repair. The function of stranded templates: this activity would permit DNA synthesis
helicase I11 (8) remains unclear. Another E. coli helicase, the
product of the rep gene (9), isrequired for the RF’ + RF and to progress beyond the formI1 structure by a strand-displace-
R F + SS(c) stagesof 4x174 DNA replication (10, 11)and is ment mechanism to create rolling-circle forms. It should be
noted that the structures and generation of replicating rolling-
* These studies were supported by National Institutes of Health circle DNA molecules considered in the present study are not
Grants GM34557 and GM34558. The costs of publication of this related to those generated in the RF + SS(c) stageof 4x174
article were defrayed in part by the payment of page charges. This replication that requires the viral gene A protein and the E.
article must therefore be hereby marked “advertisement” in accord- coli rep protein (10, 11). In this report, the generation of
ance with 18 U.S.C. Section 1734 solelyto indicate this fact. replicating rolling-circle DNA molecules in the 4x174 SS(c)
$ Robert Wood Johnson Jr. Charitable Trust Fellow in the Center’s
+ RF DNA replication system is demonstrated and evidence
Medical Scientist Training Program.
Recipient of an American Cancer Society Faculty Research is provided that their formation is primosome dependent.
Award and an Irma T. Hirschl, Monique Weill-Caulier Career Sci-
entist Award. MATERIALSANDMETHODS
The abbreviations used are: RF, replicative-form DNA; SSB, E.
coli single-stranded DNA binding protein; pol 111, DNA polymerase DNA and Enzymes-4x174 (+) SS(c) DNA was prepared by an
I11 holoenzyme; SS(c), single-stranded circular DNA form 11, relaxed, established procedure (29). Viral SS(c) DNAof flR208 (30), an fl
circular, duplex DNA; form 111, full-length, linear, duplex DNA;
Hepes, 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid. J. Minden and K. Marians, unpublished data.

2304

This is an Open Access article under the CC BY license.

 DNA Iielicase Activity A s s ~ i ~ t with
e d the E. coli P ~ i r n ~ s o ~ e
recombinant phage (a gift of Dr. P. Model, Rockefeller University),
was prepared as previously described (31). ThePstI restriction
endonuclease was from New England BioLabs. Bovine pancreatic
deoxyribonuclease I was from Sigma. O l i ~ n u c l ~ t i dwere
e s prepared
using an Applied Biosystems 380A DNA synthesizer. PstI (5'-
A A C T C T G C A G G T T G - 3 ' ) andPstIII(5"AGCGATAAA-
ACTCTGCAGGTTGGATACGCC-3') represented (-1 strand 6x174
DNA sequences at the PstI endonuclease cleavage site. PstII (5'-
CAACCTGCAGAG~-3')represented (+) strand sequences at the
PstI site.
Replication Proteins-E. coli replication proteins were purified as
previously described (28); the specific activities in units/mgare
indicated in parentheses: the E. coli single-stranded DNA binding
protein (SSB) (485), the DNA B protein ( 7 , 0 ~ )the, DNA C protein
(12,800), the DNA G protein (19,200), the DNA N protein (13,500),
DNA polymerase III* (29,700), protein i (2,390), protein n (8,900),
protein n" (lO,OOO), and factor Y (114,000).
&X174 SS(c) -+RF DNA R e p ~ ~ a t w ~ - ~ p l i c a t ireaction
on mix-
tures (25 p l ) containing 50 mM Hepes-KOH buffer (pH 8.0 a t 20 "C),
10 mM magnesium acetate, 10 mM dithiothreitol, 0.2 mg/ml bovine
serum albumin, 2 mM ATP, 100 p~ each of CTP, GTP, and UTP, 40
p~ each of dATP, dCTP, dGTP, and dTTP, 225 pmol (as nucleotide)
of &X174SS(c) DNA, 0.75 fig of SSB, 0.55 unit of the DNA B protein,
0.9 unit of the DNA C protein, 0.6 unit of the DNA G protein, 0.7
unit of protein i, 0.45 unit of protein n, 0.75 unit of protein n", 0.55
unit of factor Y, 1.8 units of DNA polymerase III*, and 0.5 unit of
the DNA N protein were incubated at 30 "C. Replication proteins
were added last as a combined protein mix to reaction mixtures
containing #X174 SS(c) DNA and SSB. Depending on the type of
subsequent analysis, reaction products were labeled with either [3HJ
dTTP (200-250 cpm/pmol) or [cx-~'P]~CTP (500-5000 cpm/pmol)
purchased from Amersham Corp. Reactions containing 13H]dTTP
were stopped by the addition of 0.1ml of 0.2M sodium pyrophosphate,
0.1 ml of 1 mg/mf heat-denatured salmon sperm DNA (as carrier),
and 4 ml of 5% trichloroacetic acid. After 10 min on ice, acid-insoluble
material was collected on glass fibers filters (Enzo), which were then
washed with 1%trichloroacetic acid and 95% ethanol anddried under
a heat lamp. The radioactivity retained was then determined. Reac-
tions containing [cv~'P)~CTP were stopped by heating a t 65 "C for 5
min followed by the addition (5 p1/25 pl reaction mixture) of a dye
mixture containing 1 mg/ml xylene cylanol, 1 mg/ml bromphenol
blue, 20 mM EDTA, 2% (w/v) sarkosyl, and 50% glycerol.
Gel ~ ~ c t r o pAndysis~ r e ~of ~Reaction P ~ ~ t s - L a ~ l syn-
ed
thetic DNA products were typically resolved by electrophoresis
through 0.8% agarose (SeaKem ME) gels at 1.8-5.5 V/cm at room
temperature in TAE buffer (50 mM Tris-HC1 (pH 7.9 at 20 "C), 40
mM sodium acetate, and 1 mM EDTA). Gels were then dried under
vacuum on a gel-drying apparatus andautoradiographed with Kodak
XAR-5 fi1m. Two-dimensional gel electrophoretic analyses were per-
formed on a horizontal gel apparatus with 0.8% agarose gels. After
electrophoresis in the first dimension under native conditions (2.7 V/
cm for 9 h) using TAE buffer, gels were soaked in four changes of 8
gel-volumes each of 50 mM NaOH, 2 mM EDTA over 1 h at room
temperature. Electrophoresis in the second dimension (2.6 V/cm for
8.5 h) was performed in the same apparatus after repositioning the
reequilibrated gels by a 90 angle. The running buffer under dena-
turing conditions was 30 mM NaOH, 2 mM EDTA. Southern blot
analysis (32) of replication reaction products resolved by a two-
dimensional gel electrophoresis was performed by transfer with 20 X
ssc (1 X ssc = 0.15 M NaCI, 15 mM sodium citrate) to a 0.45 pm
nitrocellulose membrane (Millipore) after acid depurination, alkali
denaturation, and neutralization. The baked membrane was prehy-
bridized for 2 h a t room temperature in 50 mM sodium phosphate
buffer (pH 7.0 at 20 "C), 6 X SSC, 0.25% SDS, 2.5 X Denhardt's
solution (1x Denhardt's solution = 0.2 mg/ml bovine serum albumin,
0.2 mg/ml polyvinylpyrrolidone, 0.2 mg/ml Ficoll) and 50 pg/ml
denatured salmon sperm DNA. Hybridizations were carried out in
the same solution containing 32P-5'-end-labeled single-stranded oli-
gonucleotide probes instead of salmon sperm DNA. Autoradiography
followed washing for 30 min at 4 "C with 6 X SSC. Oligonucleotides
were phosphorylated with [y-32P]ATP(Amersham Corp.) using T4
polynucleotide kinase (Pharmacia P-L Biochemicals) and isolated by
gel filtration of a 50-pl sampfe through a 6 X 25-cm Sephadex G-50
column.
2305
RESULTS
Structural Characterization of Linear Multimer Products-
The use of the 4x174SS(c) "+ RF DNA replication reaction
to evaluate the ability of the primosome to act as a helicase
was based on several observations. The predicted manifesta-
tion of a helicase activity present in such a reaction was the
generation of rolling circle-type molecules. ~ o m p l e m e n ~ ~
(-) strand DNA synthesis on the inputviral (+) strand SS(c)
templates could continue beyond the form I1 structures for
multiple cycles if a strand-displacement activity were pro-
vided. The structure of theseproducts would be relaxed,
double-stranded, unit-length circles possessing multigenome-
length, linear tails. In fact, Shlomai et al. (33), in a previous
study of the 4x174 SS(c) -4RF DNA replication reaction
with purified proteins, described a subpopulation of synthetic
products that appeared by electron microscopic analysis to be
duplex, unit-length, relaxed circles with duplex, linear tailso f
various lengths.
The kinetics of [3H]dTMP incorporation into acid-insolu-
ble material during 4x174 SS(c) "+ RF DNA synthesis was
monitored (Fig. 1).Within 5 min, synthesis of DNA products
reached an amount equivalent to that of the input template;
however, the reaction continued linearly for another 40 min.
After 90 min, the accumulation of synthetic product was 7-
fold greater than the amount of input template, suggesting
that thereaction products were not solely form I1 molecules.
Two major species of labeled products were identified by
native agarose gel electrophoretic analysis (Fig. 2). Form I1
molecules appeared early during the incubation but did not
increase in amount significantly thereafter. The second spe-
cies of products migrated very slowly and did not begin to
accumulate to an appreciable amount until after 10 min of
incubation.
In order to obtain an initial appraisal of the structure of
the larger molecules, a complete PstI restriction endonuclease
digestion of total reaction products was performed. The
4x174 genome possesses a unique PstI cleavage site. Both
species of DNA products were converted to form I11 molecules
by this treatment, suggesting a repetitive, duplex nature in
the larger products (data not shown). This could be further
demonstrated by performing apartial PstI endonuclease
digestion of the synthetic products (Fig. 3 A ) . Based on their
400 g
I4 I
5 15 x) 45 60 75 90 0
-
Time (min)
FIG. 1. Kinetics of dTMP incorporation into acid-insoluble
material in the4x174 SS(ce) RF DNA replication reaction.
A 200-p1 (8 X) standard reaction mixture containing 13H]dTTPwas
incubated at 30 "C. Ateach time point, a portion of the reaction was
stopped as described under "Materials and Methods" to determine
the amount of dTMP incorporation per 25 pl. The synthesis of
products equivalent to the amount of input DNA corresponded to
225 pmol of nucleotide or 56.25 pmol of dTMP incorporation/25 pl
of reaction mixture.
2306 DNA Helicase
Activity
Associated
-
1 2 3 4 5 6 7
. .
.- u - lineor
multimer
A u cl.3)w w r-r <;;;;
FIG. 2. Kinetics of linear multimer formation in the 6x174
SS(c) 4 RF reaction. A 50-pl (2 X) standard reaction mixture
T P incubated a t 30 "C. At each time point,
containing [ ( u - ~ P I ~ C was
6 pl was removedand thatportion of the reaction stopped as described
under "Materials and Methods." 3ZP-labeledsynthetic DNA products
were separated by agarose gel electrophoresis at 2.3 V/cm for 10 h.
Lane I, 2.5 min; lane 2, 5 min; lane 3, 10 min; lane 4, 20 min; lane 5,
30 min; lane 6,45 min; lane 7,60 min.
A. 0.
1 2 3 1 2 345
1--
kbp
'a I -:Xner -linear
multimer
23 -
9.4-
6.6 -
4.4-
n m <;;;; ::I -form I1
FIG.3. Partial Pet1 restriction endonuclease and deoxyri-
bonuclease I digestions of reaction products. A, a standard
reaction mixture containing [w3'P]dCTP was incubated a t 30 "C for
40 min and stopped by heating a t 65 "C for 5 min. "P-labeled products
were partially digested with the PstIrestriction endonuclease (1unit
per 6 pl) on ice after increasing the concentration of NaCl in the
reaction mixture to 100 mM. The digestion products were separated
by electrophoresis through a horizontal 0.3% agarose gel in TAE
buffer at 1.7V/cm for 17 h. Lane I, undigested reaction products;
lane 2, products digested for 60 s; lane 3, products digested for 120 s.
R, a standard reaction mixture was incubated a t 30 "C for 40 min and
stopped by heating a t 65°C for 5 min."'P-labeled products were
partially digested with deoxyribonuclease I on ice for 5 min after the
addition of CaCI2to a final concentration of 0.15 mM to the reaction
mixture. The digestion products were separated by electrophoresis
through a horizontal 0.3% agarose gel in TAE buffer a t 2.1 V/cm for
15 h. Lane I, undigested reaction products; lane 2, products partially
digested with PstI restriction endonuclease; lane 3, products digested
with deoxyribonuclease I (1 ng/6 pl); lane 4, products digested with
deoxyribonuclease I (3 ng/6 pl); lane 5,products digested with deox-
yribonuclease I (10 ng/6 pl).
relative electrophoretic mobilities, the molecular weights of
the molecules generated by this treatment appeared todiffer
by integral multiples of the unit 4x174genome size. At least
15 discretely resolved species could be discerned.
Two types of structures could conceivably account for the
pattern generated by partial digestion with the PstI endonu-
clease: a long, linear molecule containing tandemly repeated
4x174 genomic sequences or a catenated arrayof unit-length
4x174 circles. If the latter possibility were the case, limited
digestion by pancreatic deoxyribonuclease I should result in
a discrete pattern identical to that generated by partial diges-
tion with the PstI endonuclease. Agarose gel electrophoresis
of syntheticproductstreatedwith low concentrations of
DNase I resulted in a smear rather thana pattern containing
discrete species, eliminating the possibility that the large
DNA product was a catenated structure (Fig. 3 B ) . Based on
with the E. coli
Primosome
the results of the partial digestion by the PstI endonuclease,
the size of the large DNA was estimated a t 100 kilobases.
Multigenome-length, linearproducts of this type were an
indication of a rolling-circle mechanism occurring during the
replication reaction.
The conversion of these linear multimers toform I11 prod-
ucts by complete digestion with the PstIendonuclease, which
does not cleave single-stranded 4x174 DNA (34), implied
that theywere double-stranded in structure. In addition, these
large molecules were sensitive todigestion by exonuclease I11
but not by nuclease S1 (data not shown).
Additional structural information was gathered from anal-
ysis of two-dimensional agarose gel electrophoretic separation
of the reaction products. This method was chosen to allow
examination of the larger products separatefrom the form I1
molecules. First dimensional electrophoreses were under na-
tiveconditions, while seconddimensionalelectrophoreses
were under alkaline denaturingconditions. The large rolling-
circle molecules could be resolved into two subpopulations
(Fig. 4A). One comprised DNA that migratedvery slowly in
both native and alkaline conditions, probably representing
tails generated directly by continued elongation of the com-
plementary (-) strand during its displacement. The other
subpopulation containedmolecules between 1 and 5 kilobases
in length that were presumed to be (+) strand Okazaki frag-
ments complementary to the(-) strand tails. The absenceof
discrete species within this subpopulation was consistent with
initiation events thatwere random withrespect to the 4x174
genomic map.
These assignments were confirmed by Southern blot anal-
yses. The "P-labeled single-stranded oligonucleotide probe
PstII, representing (+) strand 4x174 sequences, hybridized
to thelarger subpopulation but not to the Okazaki fragments
A, 2nd Alkaline-
Native
- -linear mullimer
B. (-) 4x174 Products c. (+) 4x174 Products
~~
-linear -linear mullimer
multimer
,form I1 ,form II
'form I11 -form III
FIG. 4. Two-dimensional agarose gel electrophoretic anal-
ysis of reaction products. A, "'P-labeled synthetic DNA products
of a standard reaction incubated for 45 min were resolved on a two-
dimensional gel as described under "Materials and Methods." Elec-
trophoresis in the first dimension was under native conditions and in
the second dimension under denaturing conditions. The synthetic
products that migrated through these gels as diagonals underneath
the form I1 and form 111 molecules were not analyzed further, since
they did not appear to be immediately related to thelinear multimers.
B,unlabeled products of a standard reaction incubated for 90 min
were similarly resolved and transferred to a nitrocellulose membrane
as described under "Materials and Methods." The 32P-5'-end-labeled
PstII single-stranded oligonucleotidewas usedto probe DNA products
having (-) strand 6x174sequences. C, after removing the hybridized
probe, the Southern blot was reprobed with the 32P-5'-end-labeled
PstIII single-stranded oligonucleotide to identify DNA species pos-
sessing (+) strand 6x174 sequences.
 DNA Helicase Activity Associated with the E. coli Primosome 2307
(Fig. 4B).TheOkazakifragmentsubpopulation couldbe A.
detected using the "P-labeled PstIII probe representing (-) 123456789
""-
strand 6x174 sequences (Fig. 4C). Some of the material in
the larger DNAsubpopulation also hybridized with thisprobe,
implying that generationof the strand complementary to the
initial single-stranded tail in therolling-circle did not always
-form I1
involve synthesis of short Okazaki fragments. In addition, the
original (+) strandcirculartemplates associatedwith the
rolling-circle molecules could be detectedwiththePstIII
probe.
Together, these observations demonstrated that a rolling-
B.
circle product was being generated in the 6x174 SS(c) + RF
1 2 3 4 5 6 7 8 9lO1112
DNA replication reaction. These molecules possessed multi- ""

genome-length, double-stranded tails. L Y m .IO0 - lhnear

A Helicase Activity Associated with the Primosome-Since rnUlIlmer

rolling-circle DNA replication requires strand displacement,

which the DNA polymerase I11 holoenzyme cannot accom-
plish (35, 36), the observations detailed above suggested the
presence of a helicase activity associated with the primosome.
To eliminatethefortuitouscontamination of one of the
preparations of primosomal proteins with a helicase activity,
the productsof DNA replication by the entire complementof
4x174 SS(c) + RF replication proteins using G4 and a-3
phage SS(c) DNAs as templates were examined. These DNAs
lack a primosome assembly site and require only theDNA G
primase and SSB for initiation of DNA replication. Use of
the G4 and a-3 SS(c) DNAs as templates failed to support
significant rolling-circle product formation(Fig. 5), indicating
that therewas no helicase contamination of any consequence
in the protein preparations in use andprimosomethat assem-
bly was required to generate thehelicase activity.
The observations reported here and the demonstration of
the intrinsic helicase activity of the DNA B protein (27) 25
suggested that theDNA B protein was the agentresponsible A "
for the primosome-associated helicase activity.Thus theabil- 30 60 90 120 150 180
ity of the general priming system (37,38),which requires only Time (min)
the DNA B and DNA Gproteins to prime any single-stranded FIG. 6. A comparison of the kinetics of6x174 product for-
DNA in the absence of SSB, to support the formation of mation with the general priming system and with the specific
rolling-circle structures was investigated (Fig. 6A). Somelarge priming system. One hundred microliter (4 X ) standard reaction
DNA molecules could be detected. However, a comparison of mixtures containing only the DNA B (2.2 units) and DNA G (2.4
linear multimer formation during general priming-dependent units) proteins plus pol 111 (7.2 units) in the absence of SSB ( A ) or
(Fig. 6A) and primosome-dependent (Fig. 6B) DNA replica- containing SSB and all theprimosomal proteins pluspol 111 ( B )were
incubated a t 30 'C. Samples were taken a t various times after the
tion indicated that, even though both systems yielded the start of incubation for product analysis using agarose gel electropho-
same conversion of input SS(c) DNA templates to form I1 resis (2.3 V/cm for 13 h). A, lane 1, 5 min; lane 2, 15 min; lane 3, 30
molecules, the latter system produced a rate of rolling-circle min; lane 4, 45 min; lane 5, 60 min; lane 6, 90 min; lane 7, 120 min;
DNA synthesis 13-fold greater than the former system (Fig. lane 8, 150 rnin; lane 9,180 min. B, lane 1,20 s; lane 2,40 s; lane 3 , l
min; lane 4, 2 min; lane 5,5 min; lane 6, 15 min; lane 7,30 min; lane
1 2 3 4 5 8, 45 min; lane 9, 60 min; lane 10, 90 min; lane 21, 120 min. C,
following autoradiography, the bands in the dried gels corresponding
to the form I1 and linear rnultimer DNA species were excised to
quantitate radioactivity. Form I1 molecules (0-0) and linear mul-
timers (W".) synthesized in the general priming system; form I1
molecules (0-0) and linear multimers (0-0) synthesized in the
specific priming system.

612). Thus,thesedataindicatethat whereas the DNA B

FIG. 5. Synthetic DNA products formed using other SS(c)
DNA templates that lack primosome assembly sites. Standard
reaction mixtures containing [a-"PJdCTP and 225 pmol of SSB-
coated 6x174, G4, or a-3 SS(c) DNA templates were incubated at
30 "C. 32P-labeled synthetic DNA products were resolved by agarose
gel electrophoresis performed at 1.8 V/cm for 12.5 h. Lane 1, 6x174
products formed after a 40-min incubation;lanes 2 and3, G4 synthetic
products formed after 5 min and 40 min, respectively; and lanes 4
and 5, a-3 syntheticproductsformedafter 5 minand 40 min,
respectively. Prolonged incubation for up to 120 min did not result
in any significant linear multimer formation from G4 or a-3 SS(c)
DNA templates (data not shown).
protein is sufficient to catalyze rolling-circle DNA replication,
formation of the complete primosome generates a more effi-
cient helicase, probably through somemodification of the
DNA B helicase activity.
DISCUSSION
These studiesdescribe the identification of a helicase activ-
ity associated with the E. coli primosome asrevealed through
the characterizationof a rolling-circle DNA replication reac-
tion thatoccurred during 6x174 SS(c)+ RF DNA synthesis.
Highmolecularweight, double-stranded,linear,synthetic
2308 DNA Helicase Activity Associated with the E. coli Primosome
DNAproducts of approximatelytwenty4x174 genome
lengths could be detected during 4x174 SS(c) + RF DNA
synthesis, consistent with the presence of a rolling-circle DNA
replication mechanism. These linear multimers were com-
posed of two subpopulations of DNA molecules. One consisted
of extremely long molecules representing (-) strand DNA
chains formeddirectly by strand-displacement synthesis. The
other represented shorter,(+) strand Okazaki-type fragments.
Similar rolling-circle structures have been observed in pre-
vious studies of 4x174replication where the primosomal
proteins were present (33, 39, 40). It is possible that the
extremely high molecular weight products observed in other
DNA replication systemsin vitrohave a similar structure. In
fact, under certain conditions, the major synthetic product in
the pBR322 replicationsystemstudiedinthislaboratory
appears to be generated from rolling-circle structures (41).
A specific sequence (a primosome assembly site) on 4x174
SS(c) DNA is required efficient
for primosome assembly (42).
Substitution of the 6x174 DNA in the reaction mixture with
either G4 or a-3 DNA, two other similar SS(c) DNA templates
that lack primosome assembly sites, did not result in appre-
ciable linear multimer formation. By employing the general
priming system,which involves only the DNA B and DNAG FIG. 7. Model of rolling-circle DNA synthesis in the #X174
proteins acting on single-stranded DNA templates not coated SS(c) 4 RF replication reaction. i, following assembly, the pri-
with SSB (37, 38), instead of the specific priming system (i.e. mosome (0) migrates on the (+) strand 6x174 SS(c) template in an
primosome-dependent), the production of linear multimers anti-elongation direction to synthesize primers for complementary
(-) strand DNA synthesis (b) catalyzed by pol I11 (0).ii, opposite
was greatly reduced, although still evident, under conditions directionalitiesof movement on the template will result in a confron-
that were otherwise identical. tation between the primosome and pol I11 beforecomplementary
LeBowitz and McMacken (27) have recently demonstrated strand DNA synthesis is completed.Complementary strand DNA
that the DNA B protein iscapable of acting asa helicase. The synthesis can be completed if some mechanism allows them to pass
DNA B protein is a component of the primosome and may each other (iii), orif the migration of the primosome is neglible
well be the protein primarily responsible for denaturing du- strandduring the time it takes pol I11 to polymerize the complementary
completely (iu). u, conversion of this form I1 DNA-protein
plex DNA strands in the activity described in this report. complex to one capable of supporting rolling-circle DNA synthesis
Previous experimental evidence has been presented that the wouldinvolve transfer of the DNA B protein/primosome to the
DNA B protein functions more processively in the specific complementary strand. ui,rolling-circle DNAsynthesis proceeds with
primingsystemthaninthe general primingsystem(43). the DNA B protein/primosome situated at the junction to displace
Conceivably, this could explain why involvement of the pri- the (-) strand ahead of pol I11 (and to synthesize primers for the
formation of Okazaki fragments).
mosome results in much greaterrolling-circle DNA synthesis
than when the general primingsystemis involved,even activity (data notshown). In fact, it appeared that if the DNA
though both systems contain the DNA B protein. The DNA B protein or factor Y was the proteinresponsible for providing
B protein by itself may behave too distributivelyt o function the helicase activity, the nucleoside triphosphate hydrolysis
optimally as a helicase. In support of this, in 4x174 SS(c)+ requirementsunderconditions where the polymerase and
R F reactions primed with synthetic oligonucleotides and con- helicase activities were coupled could be different from those
taining only polI11 and a concentration of the DNAB protein where helicase activity was assayed independently. Kornberg
optimal for specific priming, only form I1 products were made et al. (46) demonstrated this to be case the for therep protein
(data not shown). The natureof the conversion of the DNA ATPase.There were distinct differences in K , for ATP,
B protein from its intrinsically distributive mode to the ap- substrate specificity, and sensitivity to ATP analogs when the
parently more functional processive mode remains to be de- rep ATPase effector activity of single-strandedDNA was
termined. However, itis likely thatthis conversion is a compared to thatof a replicating fork.
fundamental necessity of DNA replication in E. coli. All of It now appears that the E. coli primosome may be function-
the replication systems currently under study usingpurified ally analagous to the T7 gene 4 protein (13,14) and the phage
proteins seem to employ a mechanism to transfer the DNA B T 4 gene 41/61 proteins (15-18) in being capable of synthesiz-
protein to the DNAso that it can act ina processive fashion ing primers (15, 21-23) and acting as a helicase (19, 20). A
(28, 44, 45). multiprotein complex containing these functions canreadily
Two primosomal proteins, factor Y and the DNA B protein, be envisaged at the replicationfork to denature the parental
have nucleoside triphosphatase activity. Thus, either could duplex strands and to generate primers for lagging-strand
provide the helicase activity described here. Arai et al. (43) DNA synthesis. The concept that the primosome moves in a
proposed that the ATPase activity of protein n' (factor Y) direction opposite that of pol I11 (24) is consistent with the
moved the primosome along the DNA. Experiments were required structure-function relationships at the replication
performed in which the effects of various nonhydrolyzable fork. Inthegeneration of 9x174 rolling-circle molecules
analogs of ATP, GTP, and dATP on the production of mul- described here, the topology still needs further clarification
tigenome-length DNA molecules by isolated protein. DNA (Fig. 7).At the start of the replication reaction following
complexes known to contain at least factor Y, the DNA B assembly, the primosome must reside on the viral (+) strand
protein, and the DNApolymerase I11 holoenzyme were stud- to synthesize primers during the conversion of SS(c) mole-
ied; however, they proved unable to differentiate clearly be- cules to form I1 molecules. Subsequent formation of rolling
tween a DNA B- and a factor Y-driven primosomal helicase circles requires pol I11 to continue elongation of the nascent
 DNA Helicase Activity Associated with
the
(-) strand. If the primosome were to remain on the (+) strand,
acting as a helicase, it would have to migrate in the same
direction as pol 111, defying the currentconcept of primosome
movement. One way to resolve this difficulty would be for the
primosome to relocate to the nascent (-) strand at thejunc-
tion. In this position, it would be able to stay at thejunction,
moving in an anti-elongation direction, acting as a helicase
and forming primers for Okazaki fragment synthesis. Relo-
cation of the primosome, or itsactive component, would most
likely be a direct intrastrand transfer. Whether andhow this
occurs remains to be demonstrated.
The dataobtained from these studiesprovide evidence that
there is a helicase activity associated with the E. coli primo-
some. Manifestation of this helicase activity in rolling-circle
formation may be an indicator of the unwinding functions
operative in other reconstituted DNA replication systems as
well.
Acknowledgments-We thank Dr. Jerard Hurwitz for his critical
reading of this manuscript, Dr. Jonathan Minden for numerous
stimulating discussions during the course of this work, and David
Valentin for his technical assistance and excellent artwork.
REFERENCES
1. Abdel-Monem, M., and Hoffman-Berling, H. (1976) Eur. J.
Bwchem. 65,431-440
2. Abdel-Monem, M., Taucher-Scholz, G., and Klinkert, M.-Q.
(1983) Proc. Natl. Acad. Sci. U. S. A. 80, 4659-4663
3. Abdel-Monem, M., Chanal, M.-C., and Hoffman-Berling, H.
(1977) Eur. J. Biochem. 7 9 , 33-38
4. Oeda, K., Horiuchi, T., and Sekiguchi, M. (1982) Nature 2 9 8 ,
98-100
5. Maples, V. F., and Kushner, S. R. (1982) Proc. Natl. Acad. Sci.
U. S. A. 79,5616-5620
6. Hickson, I. D., Arthur, H. M., Bramhill, D., and Emmerson, P.
T. (1983) Mol. Gen. Genet. 1 9 0 , 265-270
7. Taucher-Scholz, G., and Hoffman-Berling, H. (1983) Eur. J.
Biochem. 137,573-580
8. Yarranton, G. T., Das, R. H., and Gefter, M. L. (1979) J. Biol.
Chem. 254,11997-12001
9. Scott, J. F., Eisenberg, S., Bertsch, L. L., and Kornberg, A. (1977)
Proc. Natl. Acad. Sci. U. S. A. 7 4 , 193-197
10. Eisenberg, S., Scott, J. F., and Kornberg, A. (1976) Proc. Natl.
Acad. Sci. U. S. A. 73, 1594-1597
11. Sumida-Yasumoto, C., Yudelevich, A., and Hurwitz, J. (1976)
Proc. Natl. Acad. U. S. A. 7 3 , 1887-1891
12. Taucher-Scholz, G., Abdel-Monem, M., and Hoffman-Berling, H.
(1983) in UCLA Symposia of Molecular and Cellular Biology
New Series (Cozzarelli, N., ed) pp. 65-76, Alan R. Liss, New
York
13. Hinkle, D. C., and Richardson, C. C. (1975) J. Biol. Chem. 2 5 0 ,
5523-5529
14. Scherzinger, E., and Seiffert, D. (1975) Mol. Gen. Genet. 1 4 1 ,
213-232
15. Nossal, N. G. (1979) J. Biol. Chem. 254,6026-6031
E. coli Primosome 2309
16. Morris, C. F., Moran, L. A., and Alberts, B.M. (1979) J. Biol.
Chem. 254,6797-6802
17. Alberts, B., Barry, J., Bedinger, P., Burke, R., Hibner, U., Liu,
C.-C., and Sheridan, R. (1980) in ZCN-UCLA Symposiaon
Molecular and Cellular Biology (Alberts, B., ed) pp. 449-473,
Academic Press, Orlando, FL
18. Silver, L. L., and Nossal, N. G. (1982) J . Bwl. Chem. 257,11696-
11705
19. Matson, S. W., Tabor, S., and Richardson, C.C. (1983) J. Biol.
Chem. 258, 14017-14024
20. Venkatesan, M., Silver, L. L., and Nossal, N.G. (1982) J. Bwl.
Chem. 2 5 7 , 12426-12434
21. Scherzinger, E., Lanka, E., Morelli, G., Seiffert, D., and Yuki, A.
(1977) Eur. J. Bwchem. 72.543-558
22. Kolodner, R., Masamune, Y.,-LeClerc, J. E., and Richardson, C.
C. (1978) J. Biol. Chem. 253. 566-573
23. Liu, C.-C., Burke, R.L., Hibner, U., Barry, J., and Alberts, B.
(1979) Cold Spring Harbor Ssmp. Quant. Biol. 43.469-487
24. Arai, K.-I., and-Korkberg, A. (i981) Proc. Natl. Acad. Sci. U. S.
A. 78, 69-73
25. Wickner, S., and Hunvitz, J. (1975) in ICN-UCLA Symposia on
Molecular and Cellular Biology (Goulian, M., and Hanawalt, P.,
eds) pp. 227-238, Academic Press, Orlando, FL
26. Weiner, J. H., McMacken, R., and Kornberg, A. (1976) Proc.
Natl. Acad. Sci. U. S. A. 7 3 , 752-756
27. LeBowitz, J. H., and McMacken, R. (1986) J. Biol. Chem. 2 6 1 ,
4738-4748
28. Minden, J. S., and Marians, K. J. (1985) J. Biol.Chem. 260,
9316-9325
29. Francke, B., and Ray, D. S. (1971) Virology 4 4 , 168-187
30. Boeke, J. D., Vovis, G. F., and Zinder, N.D. (1979) Proc. Natl.
Acad. Sci. U. S. A. 76,2699-2702
31. Model, P., and Zinder, N. D. (1974) J . Mol. Biol. 8 3 , 231-251
32. Southern, E. (1975) J. Mol. Biol. 9 8 , 503-517
33. Shlomai, J., Polder, L., Arai, K., and Kornberg, A. (1981) J. Biol.
Chem. 256,5233-5238
34. Blakesley, R.W., Dodgson, J. B., Nes, I. F., and Wells, R.D.
(1977) J. Bwl. Chem. 2 5 2 , 7300-7306
35. O’Donnell, M. E., and Kornberg, A. (1985) J. Bwl. Chem. 2 6 0 ,
12875-12883
36. O’Donnell, M. E., and Kornberg, A. (1985) J. Biol. Chern. 2 6 0 ,
12884-12889
37. Arai, K., and Kornberg, A. (1979) Proc. Natl. Acad. Sci. U. S. A.
76,4308-4312
38. Arai, K., and Kornberg, A. (1981) J. Biol. Chem. 256,5267-5272
39. Arai, K.-I., and Kornberg, A. (1981) J . Biol.Chem. 256, 5253-
5259
40. Arai, K.-I., Arai, N., Shlomai, J., and Kornberg, A. (1980) Proc.
Natl. Acad. Sci. U. S. A. 7 7 , 3322-3326
41. Minden, J. S., and Marians, K. J. (1986) J. Biol.Chem. 2 6 1 ,
11906-11917
42. Shlomai, J., and Kornberg, A. (1980) Proc. Natl. Acad. Sci. U. S.
A. 7 7 , 799-803
43. Arai, K.-I., Low, R. L., and Kornberg, A. (1981) Proc. Natl. Acad.
Sci. U. S. A. 7 8 , 707-711
44. Dodson, M., Roberts, J., McMacken, R., and Echols, H. (1985)
Proc. Natl. Acad. Sci. U. S. A. 8 2 , 4678-4682
45. Funnell, B. A., Baker, T. A,, and Kornberg, A. (1986) J. Biol.
Chem. 261,5616-5624
46. Kornberg, A., Scott, J. F., andBertsch, L. L. (1978) J. Bwl. Chem.
253,3298-3304