Paradoxes of eukaryotic DNA replication: MCM proteins and the random completion problem.

Abstract

Eukaryotic DNA replication initiates at multiple origins. In early fly and frog embryos, chromosomal replication is very rapid and initiates without sequence specificity. Despite this apparent randomness, the spacing of these numerous initiation sites must be sufficiently regular for the genome to be completely replicated on time. Studies in various eukaryotes have revealed that there is a strict temporal separation of origin "licensing" prior to S phase and origin activation during S phase. This may suggest that replicon size must be already established at the licensing stage. However, recent experiments suggest that a large excess of potential origins are assembled along chromatin during licensing. Thus, a regular replicon size may result from the selection of origins during S phase. We review single molecule analyses of origin activation and other experiments addressing this issue and their general significance for eukaryotic DNA replication.

Full text

Page 1

Paradoxes of eukaryotic DNA
replication: MCM proteins and the
random completion problem
Olivier Hyrien,1* Kathrin Marheineke,1and Arach Goldar2
Summary
EukaryoticDNAreplicationinitiatesatmultipleorigins.In
early fly and frog embryos, chromosomal replication is
very rapid and initiates without sequence specificity.
Despite this apparent randomness, the spacing of these
numerous initiation sites must be sufficiently regular for
the genome to be completely replicated on time. Studies
in various eukaryotes have revealed that there is a strict
temporalseparationoforigin‘‘licensing’’priortoSphase
and origin activation during S phase. This may suggest
that replicon size must be already established at the
licensing stage. However, recent experiments suggest
that a large excess of potential origins are assembled
along chromatin during licensing. Thus, a regular repli-
con size may result from the selection of origins during
S phase. We review single molecule analyses of origin
activation and other experiments addressing this issue
and their general significance for eukaryotic DNA repli-
cation.CopyrightBioEssays 25:116–125, 2003.
? 2003 Wiley Periodicals, Inc.
Introduction
AccurateandcompleteDNAreplicationisessentialtogenome
stability in all organisms. In eukaryotes, DNA is normally
replicated once and only once during each S phase. As repli-
cationinitiatesatmultipleoriginsalongeachchromosome,this
requires a strict control of origin density and time of activation.
Work in Xenopus and yeasts has led to significant under-
standing of the mechanisms that prevent origins from ‘‘firing’’
morethanonceinSphase.(1)Incontrast,themechanismsthat
ensure that no stretch of the entire genome is left unreplicated
beyond the normal duration of S phase are not yet clear.
Studies with replication inhibitors have shown that many
eukaryotic cells are able to delay entry into mitosis in the
presence of unreplicated DNA (the S/M checkpoint).(2)How-
ever, a mitosis-delaying mechanism cannot, by definition,
explain the timely replication of the whole genome in an
unperturbed S phase. The problem is particularly crucial in
early embryos of insects and amphibians, which have an
accelerated cell cycle but initiate replication at random
sequences(3,4)and lack an efficient S/M checkpoint.(5,6)Early
embryos treated with replication inhibitors undergo mitosis on
schedule, resulting in catastrophic chromosomal damage.
Thus, failure of a single replication origin may have disastrous
consequences. Nevertheless, unperturbed embryos develop
normallymostofthetime,showingtheexistenceofanefficient
mechanism to complete genome replication on time. We first
summarize current knowledge of eukaryotic DNA replication
origins then focus on recent studies of origin activation in
Xenopus egg extracts that begin to answer this problem.
Proteins at eukaryotic replication origins
and the ‘‘MCM paradox’’
MCM2-7 proteins are minichromosomemaintenance proteins
first identified for their role in plasmid replication or cell cycle
progression in yeast. Studies in various eukaryotes have
defined a conserved pathway of cell-cycle-regulated protein
assembly and disassembly at DNA replication origins. This
pathway (extensively reviewed elsewhere, Ref. 1) involves at
least 20 different proteins but can be summarized as two
temporally separate steps, the recruitment and the activation
of the MCM2-7 complex (Fig. 1). MCM2-7 proteins interact
with each other and all six are required both for initiation and
elongation of DNA replication. MCM2-7 proteins can form a
heterohexameraswellasothersubcomplexes,someofwhich
possess helicase activity.(7,8)It is generally believed, though
116
BioEssays 25.2
BioEssays 25:116–125, ? 2003 Wiley Periodicals, Inc.
1Ge ´ne ´tique Mole ´culaire-UMR CNRS 8541, Ecole Normale Supe ´r-
ieure, Paris.
2CEA/Saclay, Laboratoire de Biophysique de l’ADN, DBCM, Gif-
sur-Yvette, France.
Funding agencies: The O.H. lab is supported by the Association
pour la Recherche sur le Cancer, the Ligue Nationale Contre le
Cancer (Comite ´ de Paris) and the Association Franc ¸aise contre les
Myopathies.
*Correspondence to: Olivier Hyrien, Ge ´ne ´tique Mole ´culaire. UMR
CNRS 8541, Ecole Normale Supe ´rieure, 46 rue d’Ulm, 75 230 Paris
Cedex 05, France. E-mail: hyrien@wotan.fr
DOI 10.1002/bies.10208
Published online in Wiley InterScience (www.interscience.wiley.com).
Abbreviations: ARS, autonomously replicating sequence. CDC, cell
division cycle. CDK, cyclin-dependent kinase. CHO, Chinese hamster
ovary. DHFR, dihydrofolate reductase. MCM, minichromosome main-
tenance proteins. ORC, origin recognition complex. Pre-RC, prerepli-
cative complex. rDNA, DNA containing the tandemly repeated
ribosomal RNA genes. S-CDK, S-phase cyclin-dependent kinase.
Review articles

Page 2

not proven, that MCMs form the eukaryotic DNA replication
fork helicase.(9)The recruitment of the MCM2-7 complex at
origins (called replication ‘‘licensing’’) takes place during late
mitosis and the G1phase, in preparation for the next round of
chromosomeduplication.Licensingstrictlyrequiresacomplex
of six proteins first identified in budding yeast, the origin
recognition complex (ORC).(10)In yeast, ORC binds replica-
tion origins directly and stably across the cell cycle.(11)In
metazoa, the stability of chromatin association of ORC is
higher at the G1/S transition.(12)Origin licensing also requires
CDC6 and CDT1, which must both interact with ORC to load
MCM2-7 onto chromatin.(13–15)Importantly, once MCM2-7
have been loaded, ORC and probably CDC6 and CDT1
become dispensable for subsequent replication.(14,16,17)
Once loaded, MCM2-7 complexes await activation during
S phase. This process is triggered by at least two kinases,
CDC7/DBF4 and the S-CDKs, and involves the ordered
assembly of additional proteins, among which CDC45 has
emerged as a pivotal factor.(1)CDC45 origin association
triggers origin DNA unwinding and ultimately leads to the
association of DNA polymerases with the unwound DNA
(Fig. 1).(18,19)In yeast, MCM2-7 dissociate from origins either
upon replication initiation(20,21)or upon passive replication
from a neighboring origin.(22,23)Biochemical and immuno-
fluorescence studies in metazoan cells also suggest that
MCMs are progressively excluded from replicated chromatin
during S phase.(24–29)The reloading of MCMs is prevented
untilcellspassthroughmitosisbyCDKsandbygeminin.CDKs
interfere with various functions of ORC, CDC6 and MCMs in
licensing.(30)Geminin (only found in metazoa) binds and
inhibitsCDT1.(31,32)Cyclins(andthereforeCDKactivities)and
gemininareonlydestroyedinlatemitosis,ensuringthatanew
round of origin licensing can only take place after sister
chromatidsegregation.ThestricttemporalseparationofMCM
loading and activation and the release of MCMs from repli-
cated DNA ensure that no sequence is replicated more than
once in a single S phase.
Although several lines of evidence argue that the MCMs
form the replicative DNA helicase, some observations are not
easily explained by this model. First, immunofluorescence
studies in mammalian cells(26–28)and in frog egg extracts(29)
showthat,incontrasttobonafidereplicationforkproteinssuch
as RP-A and PCNA, which colocalize with newly replicated
DNA, most of the MCMs colocalize with unreplicated DNA.
One study showed a lack of co-localization of MCMs directly
with RP-A and PCNA, or with DNA synthesized during the
period preceding fixation.(26)Second, chromatin immunopre-
cipitation experiments suggest that ORC and MCMs do not
resideoncloselyadjacentsitesinmammalianchromatin,even
in cells arrested at the G1/S boundary.(33)Finally, the number
of chromatin-bound MCM complexes exceeds the number of
replication origins and ORC complexes by a factor of 10–100
in various organisms. The ‘‘MCM paradox’’(26)is that MCM
proteins are in vast excess and do not colocalize with
replication forks.
Dispersive versus site-specific initiation
While replication initiation proteins are widely conserved,
replication origins are not.(34)In S. cerevisiae, specific se-
quences that can promote autonomous plasmid replication
(autonomously replicating sequences; ARSs) define the sites
where the synthesis of new DNA strands starts both on
plasmids and within yeast chromosomes.(35)Genomic foot-
printingindicatesthatARSsarestablyboundbyORCthrough-
outthecellcycle(11)andthat,duringlatemitosisandG1phase,
MCM2-7 bind alongside ORC to form a larger prereplicative
complex (pre-RC).(36)High-resolution mapping has shown
that replication initiates precisely at the center of the pre-RC
footprint for one ARS.(37)
In contrast to yeast, attempts to isolate specific sequences
that provide autonomous replication to transfected plasmids
in animal cells have been inconclusive.(34)In one study with
human cells, any sequence appeared suitable provided it is
largeenough(>10kb),(38)andreplicationwasfoundtoinitiate
at multiple, apparently random sites on the plasmid.(39)
Physicalmappingofchromosomaloriginsinadultanimalcells
has given more complex results. At loci such as the human
lamin B2(40)and b-globin(41)genes, replication initiates at
Figure 1. Licensing and activation of replication
origins.MCM2-7areloadedduringlatemitosisand
G1phase onto replication origins by ORC, CDT1
and CDC6 (origin licensing). Pre-replication com-
plexes (Pre-RC) are activated at the G1/S transi-
tion by two kinases, CDC7/DBF4 and S-CDKs.
A key step in this transition to replication is the
recruitment of CDC45. MCM2-7 dissociate from
DNA asS phaseprogresses.ReloadingofMCM2-
7 is prevented by at least two inhibitors, geminin
and the CDKs. This inhibition persists until cells
pass through mitosis, when geminin and cyclins
are destroyed.
Review articles
BioEssays 25.2
117

Page 3

sharplydefinedsitesasfoundinyeast.Atotherlocisuchasthe
Chinese hamster DHFR(42)and rhodopsin(43)loci and human
rDNA,(44)replication can initiate at any of a large number of
sites within a broad (5–50 kb) zone. Although detailed studies
of the DHFR locus suggest that some sites within the broad
zone might be preferred, an exact quantitation is technically
difficult and remains debatable.(42)
Anextremecaseofalackofsequencespecifityforinitiation
is found during the early development of Drosophila and
Xenopus. In these organisms, the first cell cycles following
fertilization are characterized by a very brief S phase, a lack of
G1andG2phasesandalackofzygotictranscription.Chromo-
somal replication is accelerated by the use of closely spaced
origins (average interval ?10 kb). Importantly, replication
initiates with no regard to specific sequences in these early
embryos.(3,4)Circumscribed initiation zones are only detected
after the midblastula transition, when chromatin is remodelled
and zygotic transcription resumes.(45–47)The mechanism of
this developmental transition remains unexplained. However,
studies of the Chinese hamster DHFR locus suggest that
originspecificationisacquiredinthemid-G1phaseofeachcell
cycle,(48)and is an event distinct from replication licensing,
which occurs earlier, during late telophase.(49)Prior to this
‘‘origin decision point’’, early G1nuclei appear competent to
initiate at random sequences in proper experimental condi-
tions. Therefore, the developmental acquisition of a G1phase
may be relevant to the specification of replication origins after
the midblastula transition.
Replication timing programme
The time required to complete genome replication depends
not only on origin spacing but also on the temporal program of
origin activation. In most cells, origins are not synchronously
activated but fire in a reproducible order through S phase.(50)
This timing is established during the G1phase of each cell
cycle.(51)In mammalian cells, the ‘‘timing decision point’’
occurs in early G1(after replication licensing but before the
origin decision point), simultaneously with the repositioning of
sequences in the nucleus after mitosis.(52)The b-globin locus
has been shown to move to the nuclear periphery during early
G1phase coincident with theestablishment ofits mid-Sphase
replication program in CHO cells.(53)In yeast,late origins tend
to localize close to the periphery of the nucleus specifically
during G1 phase while early origins are more randomly
localized.(54)These results suggest that origins may be
modified in specific nuclear compartments during the G1
phase to determine their initiation time. The molecular nature
of these modifications is currently unknown.
Do early Drosophila and Xenopus embryos have a
replication timing program? In contrast to the earlier assump-
tion that origins fire synchronously at the onset of S phase in
these embryos,(55–57)recent studies have clearly established
thatoriginsfireasynchronouslyatleastwhenplasmidDNA(58)
or sperm nuclei(59–61)replicate in Xenopus egg extracts.
However, it is still unclear whether specific sequences
replicate at a specific time in this system. Evidence against
this is the observation that, in Xenopus egg extracts, differ-
entiated nuclear compartments are not obvious and a single
type of replication foci (intranuclear sites of DNA synthesis)
persists throughout S phase.(62)This is in contrast to adult
cells, which have a clear replication timing program, where
patches of euchromatin and heterochromatin are obvious and
different types of replication foci are active at different stages
of S phase.(63)However, it has been reported that somatic 5S
genes may replicate prior to oocyte-type 5S genes when
Xenopus sperm are replicated in Xenopus egg extracts, but
this was only seen at high sperm concentrations that resulted
in artificial extension of S phase.(64)These experimental
conditions may reproduce the slower S phase of embryos that
have passed the midblastula transition rather than the brief S
phase typical of early embryos.
The random completion problem
The regulation of origin spacing and time of activation is
particularly crucial in the early Xenopus embryo. The fertilized
Xenopus egg undergoes 12 synchronous rounds of cell
division in only 7 hours, as opposed to a somatic cell cycle
durationof?36hours.Theseacceleratedcellcyclesconsistof
a?20minuteSphaseanda?10minuteMphasewithnoG1or
G2phases. The rate of replication fork progression is 0.5 kb/
minute,(65)so that the two replication forks initiated from a
singleorigincannotreplicatemorethan20kbineachSphase.
Therefore, to replicate the entire diploid genome (6.2?
109bp), none of the required >300 000 initiation events can
be more than 20 kb from its neighbour, assuming that all
origins fire synchronously at the onset of S phase. But since
initiation events are not synchronous, they must be spaced
even more closely. The same problem applies to the early
Drosophila embryo, in which forks progress more rapidly
(2.6 kb/minute) but S phase is shorter (3–4 minutes).
The observed average spacing of replication bubbles in
eggextractsorinearlyembryosisabout8–15kbinDrosophila
or Xenopus.(3,58–61)This value overestimates origin spacing
since a single bubble may arise from the merging of two
adjacent bubbles, and some origins may have not fired at the
time that DNA is extracted for analysis. An origin spacing of
<8–15 kb may at first sight seem comfortable given the upper
limit of 20 kb for replicon size. However, if origins were
positioned randomly, there would be a geometric distribution
of interorigin distances (Fig. 3B, red curve). With a mean
spacing of 10 kb, the probability that any pair of neighboring
originsarespacedbymorethan20kbwouldbe0.2.Evenwith
ameanspacingof5kb,thisprobabilitywouldstillbe0.12.The
consequence would be that a large number of gaps of
unreplicated DNA would persist at the end of S phase. All
available data suggest that the fork rate does not increase at
Review articles
118
BioEssays 25.2

Page 4

the end of S phase. Therefore, to ensure the complete
replication of each chromosome, the spacing of replication
initiation sites has to be more regular than predicted from a
geometric distribution, despite the lack of sequence-specific
initiation. This paradox, first noticed long ago,(66)was recently
christened the ‘‘random completion problem’’.(60)
Possible solutions to the random
completion problem
There are two theoretical solutions to the random completion
problem (Fig. 2). The first one is that, despite their lack of
sequencespecificity,potentialoriginsareassembledpriortoS
phase at regular, not random, intervals.(3,60)The nature of the
length-measuringdeviceisnotobviousbutitcouldrelyonpre-
establishedchromatinfoldingoronlateralinhibitionduringthe
licensing process. By analogy, spacing patterns are known to
arisebylateralinhibition invariousdevelopmentalsystems.In
support of this ‘‘fixed spacing’’ model, it has been observed
that the binding of Xenopus ORC to sperm chromatin in egg
extract saturates at about one copy per 8–16 kb,(57,67)which
approximatelycoincideswiththerepliconsize,suggestingthat
unidentified chromatin proteins somehow constrain ORC
binding. One problem with the ‘‘fixed spacing’’ model is that
origin firing must be extremely efficient, since a single un-
replicated gap could be lethal.
The second solution is that a large excess of potential
originsisassembledpriortoS-phasebutthat,duringSphase,
the selection of those origins that fire results in a sufficiently
regulardistributionofinitiationevents.(58)Weshallrefertothis
as the ‘‘origin redundancy’’ model. Although the saturation of
sperm chromatin by relatively low concentrations of ORC
seems inconsistent with a large excess of potential origins, it
shouldbeemphasizedthat10–40foldmoreMCMsthanORC
bindtospermchromatinineggextracts,(57,68)andthatORCis
no longer required for initiation once MCMs have been
loaded.(16,17)Based on these facts, Lucas et al.(58)have
suggested that potential origins might be defined by individual
MCM complexes spread along chromatin away from ORC
rather than by ORC itself. Importantly, recent experiments by
Edwards et al.(69)support this view (see below). A key feature
of the ‘‘origin redundancy’’ model is that any stretch of un-
replicated DNAwould remain competent forinitiation through-
outSphase,despitethemechanismsthatpreventlicensingof
new origins during S phase.
Note that an excess of potential origins is not by itself
sufficienttosolvetheproblem.Ifpotentialoriginswerespaced
every 100 bp for example, but activated synchronously with a
homogeneousprobability(P¼0.01toaccountfortheaverage
10 kb spacing), a geometric distribution of inter-initiation
distances would again result, with a significant tail of >20 kb
distances. However, if initiation isnot synchronous, a different
distributionofreplication startsiteswill result.First,originsare
inactivated when they are passively replicated, reducing the
probability of closely spaced initiations. Note that origin inter-
ference may also occur by other mechanisms in advance of
replication fork passage. Second, if initiation is a stochastic
process, the evolution of this process will depend on both I(t),
thefrequencyofinitiation,andV,thespeedofreplicationforks.
The frequency of initiation, I(t) is defined as the probability of
initiation per unit length of unreplicated DNA per unit time, at
time t. In contrast to V, which is presumed to be constant, the
frequency of initiation may change over time. This may also
affect the final distribution of initiation events.
Discriminating between the ‘‘fixed spacing model’’ and the
‘‘origin redundancy model’’ requires a direct examination of
origin firing at the single molecule level and a molecular
Figure 2. Two theoretical solutions to the random completion problem in early Xenopus embryos. A: Fixed spacing model. Potential
origins are assembled before S phase at regular, not random intervals. Each origin fires with a high efficiency during S phase. B: Origin
redundancymodel.AlargeexcessofpotentialoriginsareassembledbeforeSphase.DuringSphaseonlyafractionoforiginsfireandorigin
‘‘selection’’ within unreplicated gaps ultimately ensures their timely replication.
Review articles
BioEssays 25.2
119

Page 5

definition of ‘‘potential origins’’. Some relevant work is
described below.
Electron microscopy of replicating DNA
In a remarkable pioneering study, Blumenthal et al.(55)used
electronmicroscopytoexaminetheunitsofDNAreplicationin
the early Drosophila embryo. They measured the lengths of
replication eyes (bubbles) and the distances between the
centers of consecutive eyes on segments of replicating
chromosomal DNA (see Fig. 3A). The data were classified
according to the fraction of the segment containing the eyes
that had been replicated, so as to derive a picture of origin
activation during S phase. First, the distances between the
centers of adjacent eyes were found to distribute widely
around a 7.9 kb mean, with a tendency to peak at integral
multiplesof3.4kb.Althoughthestatisticalsignificancewasnot
Figure 3. Distribution of eye-to-eye distances: observa-
tions and models. A: Eye-to-eye distances (ETED) and
replication eye lengths (EL) can be measuredon individual
fibers using electron microscopy or molecular combing
or fiber spreading of labeled DNA. B: The distribution of
ETEDs peaks around 10 kb for sperm nuclei replicated in
Xenopus egg extracts for 42 minutes and analysed
by molecular combing (n¼419, mean¼13,7 kb, mean
replication content of fibers 42%). ETEDs of all fibers were
grouped in 2 kb classes and plotted against the middle of
each size interval (squares). The red curve shows the
geometricdistributionexpectedforpotentialoriginsspaced
at random with the same mean distance (13.7 kb).
Assuming a lattice of potential binding sites spaced at
Z kb intervals, with a probability m of filling each site, the
probability that two consecutive origins are spaced by
NZ kb would be P(N)¼m(1-m)N-1. The curve shown is
assuming Z¼2 kb and m¼2/13.7. (C) Computer simula-
tionofETEDsatareplicationcontentof40%assumingthat
potential origins are abundant (one every 100 bp) and that
the frequency of initiation I(t) increases through S phase
in the way inferred from molecular combing data(75);
fitted curve in red (A. Goldar, unpublished results).
Review articles
120
BioEssays 25.2

Page 6

assessed, this result argued for periodic initiation. Second,
Blumenthal et al. found that the mean eye density rapidly
increases with replication extent to reach a plateau when 20–
30% of the segment is replicated. This plateau is maintained
until70–80%replication,thenrapidlydecreasesasreplication
is completed. Based on these data, it was suggested that
initiation events are highly synchronized and occur during
the first 30% of replication, and that the eyes grow with little
merging until 70% replication when termination becomes
predominant.
More recently, Lucas et al.(58)used electron microscopy to
examine the replication of plasmids of a broad size range in
Xenopuseggextracts.Purified DNAincubatedineggextracts
is assembled into chromatin then into synthetic nuclei and is
replicated under cell-cycle control.(70,71)Plasmid replication
initiateswithnoregardforspecificsequences,justaschromo-
somal replication in early embryos.(3,65,72)Multiple eyes were
observedpermoleculeona20kbanda42kbplasmid,butonly
a single eye per molecule on a 9 kb plasmid, consistent with
origin interference over a limited distance. When multiple
eyes were observed, they were spaced at broadly distributed
intervals with a ?10 kb mean. Initiation was not synchronous.
Small(<2kb)eyeswereobservedatallreplicationstages,and
eyes of very different sizes coexisted on single molecules.
Furthermore,themeaneyedensityincreasedwiththefraction
of the plasmid molecule that had replicated, suggesting that
neweyescontinuedtoformatahighratethroughoutSphase,
despite the dwindling length of unreplicated DNA remaining
available for initiation. The authors suggested that potential
origins are abundant and randomly distributed, but that origin
interference and the increase of initiation frequency during
S phase modulate origin firing so as to accelerate the com-
pletion of DNA replication.
These conclusions contrast with the suggestion of
Blumenthal et al.(55)that initiation is confined to the beginning
of S phase. However, their observation that the mean eye
density stays at a plateau at 30–70% replication could simply
meanthatinitiationandterminationoccuratequalfrequencies
during mid-S phase. In fact Blumenthal et al. reported a signi-
ficant fraction of small eyes on segments that are up to 90%
replicated.
Molecular combing and fiber
spreading studies
To address the possibility that plasmid DNA replication in egg
extracts may not faithfully mimic embryonic chromosome
replication, Herrick et al.(59)and Marheineke and Hyrien(61)
studied the replication of sperm nuclei in Xenopus egg
extracts. Sperm nuclei were labelled during replication by
addition of biotin-dUTP at the start of the incubation and
digoxigenin-dUTP at a varying time. After complete replica-
tion, the DNA was purified and stretched on a glass slide
by a technique called molecular combing.(73)The alter-
nating sections of early-(biotin-labelled) and late-(biotinþ
digoxigenin-labelled) replicated DNA were examined by
optical microscopy using fluorescent antibodies. A decisive
advantadge of molecular combing is that DNA molecules are
aligned in a parallel fashion and stretched to a uniform and
reproducible extent (2 kb/mm), facilitating statistical analysis
and eliminating the selection of appropriately spread fibers
inherent to other techniques.
The data largely confirmed the conclusions obtained with
plasmids. First, eye-to-eye distances were broadly distributed
arounda?10kbpeak(Fig.3B).Second,neighboringinitiation
eventswerenotsynchronous.Abroaddistributionofeyesizes
was observed at each time point and eyes of very different
sizes occured next to each other even on short (100–200 kb)
fibers. Importantly, the frequency of initiation estimated from
thedatawasfoundtoincreasewiththefractionofthesegment
that had replicated,(59,61)as previously inferred from studies
with plasmids.(58)Based on the analogy of DNA replication
to one-dimensional crystal nucleation, growth and coales-
cence, the mathematical formalism derived long ago by
Kolmogorov(74)to describe the kinetics of crystal growth in
the three-dimensional space has been applied to the combing
data to derive a refined expression for the frequency of
initiationduringSphase.(75)Thisanalysissuggestedthatthere
isamarkedincreaseofinitiationhalfwaythroughSphase.The
significance of this observation remains to be understood.
It should be noted that the distributions of eye-to-eye dis-
tances observed in Drosophila or in Xenopus differ from a
geometric distribution, showing fewer distances in the 0–5 kb
rangebutmoredistancesaround10kb(Fig.3B,comparewith
red curve). However this does not imply that potential origins
are non-randomly distributed. A computer simulation shown
on Fig. 3C illustrates that this type of distribution can result
fromasynchronousinitiationamongalargeexcessofpotential
origins.
However, a different interpretation has been suggested
by Blow et al.(60)In this study, sperm nuclei replicating in egg
extracts were labeled using a single pulse of [3H]dTTP or
BrdUTP of varying length, and the DNA was spread and
visualized by autoradiography or using a different technique
called DIRVISH.(76)Although the distributions of eye-to-
eye distances were very similar to those observed by comb-
ing, Blow et al. suggested that the clustering of distances
in the 5–15 kb range implies a non-random origin distr-
ibution. It was found that reducing the amount of ORC that
assembles on the DNA by partial ORC immunodepletion of
the extract increases the average spacing of initiation
events.(57,60)This result would be consistent with the idea
that each origin is specified by the binding of a single ORC
molecule. Computer simulations confirmed that in this case,
ORC has to be deposited in a regular pattern every 5–15 kb
in order to account for the observed distribution of eye-
to-eye distances. If there are no potential origins between
Review articles
BioEssays 25.2
121

Page 7

ORC-binding sites, a random deposition of ORC would un-
avoidably lead to a significant proportion of excessively large
replicons. However, recent experiments by Edwards et al.(69)
question the idea that potential origins coincide with ORC-
binding sites (see below).
Blowetal.(60)alsonoticedthatadjacenttrackstendtobeof
similar lengths, both in early and late replicating DNA, and
suggestedthattherearesynchronousclustersof5–10origins
firing at different times in S phase. This conclusion seemed at
odds with the observation that eyes of very different lengths
occur next to each other on combed DNA fibers.(59,61)In fact,
the correlation found by Blow et al. between the size of
adjacent eyes is significant but weak (r¼0.16, P<0.0001).
Ourcombingdatarevealasimilarcorrelationforadjacenteyes
(r¼0.2)butlessornocorrelationbetweenthefirstandthirdor
fourth eye size (unpublished). Although a few fibers do show
the appearance of clusters, most do not. The results may
depend in part on the protocol used to label and spread DNA.
Unlike molecular combing, the DIRVISH technique does not
spread DNA in a straight fashion and does not visualize
unreplicated DNA due to the use of a single labeling pulse.
Therefore the confidence that two successive eyes belong to
the same fiber declines with distance, which may bias the
selectedsamplesinfavorofmultiple,closelyspacedeyes.We
suggest that the correlation coefficients are too weak to
conclude that highly synchronous clusters are the predomi-
nant organization of DNA replication in Xenopus egg extracts.
To investigate how the time of activation of each origin is
controlled, Marheineke and Hyrien(61)used molecular comb-
ing to follow the replication of single fibers after release from a
block with aphidicolin, a DNA polymerase inhibitor. Only a
fraction of the origins was found to initiate in the presence of
aphidicolin, and the rest were found to fire asynchronously
through S phase after release. Therefore, continuing initiation
during S phase depends on the normal progression of forks
assembled at previously activated origins. This suggests that
some mechanism may limit the number of simultaneously
activeforksduringSphase.Thefactthatthemeaneye-to-eye
distancefollowsaplateaubetween35and85%replication,as
previously observed in Drosophila,(55)implies that the fre-
quency of initiation is equal to the frequency of termination
during the plateau period. In other words, once a certain
number of forks have been assembled, further initiation
seems to depend on the completion of previously active
replicons, which is prevented by aphidicolin. This regulation
may involve the recycling of some limiting component of the
replication forks, or the monitoring of total fork number by a
checkpoint. By maintaining a constant replication rate despite
random bubble mergers, this mechanism would ensure a
timely completion of DNA synthesis and explain why the
frequency of initiation increases through S phase. The
frequency of initiation (number of new initiations per unit
length of unreplicated DNA per unit time, at time t) has to
increase because, as S phase proceeds, the total number of
bubbles per nucleus is kept constant whereas the length of
unreplicated DNA decreases and an increasing fraction of
preexisting bubbles merge.
Potential origins and pre-RC assembly
A key difference between the ‘‘fixed spacing’’ and the ‘‘origin
redundancy’’ models lies in the number and molecular nature
of potential origins. In budding yeast, replication initiates im-
mediatelyadjacenttotheORCbindingsiteatonewell-studied
origin.(37)Furthermore, genome-wide studies of both initiation
sites(77)andORCandMCMproteindistributions(78)agreewith
the notion that potential origins are defined by ORC binding
sites in this organism.
In metazoa, however, neither replication start sites nor
ORC binding sites show a consensus sequence.(34)ORC and
MCMbindinghasbeeninvestigatedbyinvivoorinvitrocross-
linking in several systems. Distinct ORC binding sites have
been identified adjacentto replication start sitesin a fly(79)and
a human(80)gene. However, ORC appears to bind in a less
specific manner at the Drosophila chorion gene locus.(81)
Furthermore, MCMs are broadly distributed over the DHFR
initiation zone in Chinese hamster cells,(82)and ORC and
MCM proteins do not in general reside on closely adjacent
sites in bulk mammalian chromatin.(33)
Recent experiments by Edwards et al.(69)using a novel
chromatin-binding assay directly address this point. In this
assay, a linear DNA fragment is coupled to magnetic beads,
digestedtodifferentlengthswithrestrictionenzymesandthen
incubatedwithacytosolicXenopuseggextractthatallowspre-
RC formation. It was found that ORC binding, as well as the
ORC-dependent binding of MCM2-7, requires a minimum
fragment size of 82 bp. When DNA fragment length was
increased incrementally up to 6 kb, the amount of ORC bound
per molecule of DNA remained unchanged whereas the
MCM:ORC ratio increased from 1:1 on the 82 bp fragment
to 20:1 on the 6 kb fragment, almost as high as on sperm
chromatin. Therefore each ORC complex appears to recruit
multiple MCM complexes that spread laterally along the DNA
(Fig. 4A). Interestingly, licensing inhibition by geminin caused
ORCbindingtoincreaseinproportiontoDNAfragmentlength,
suggesting that licensing regulates ORC binding.
Edwards et al. also found that, on sperm chromatin, all
chromatin-bound MCM complexes can be phosphorylated by
CDC7/DBF4 upon addition of a concentrated nucleoplasmic
extract, suggesting that all are potential start sites. However,
theloadingoftheCDC45proteinwaslimitedtoaratioof2:1for
CDC45:ORC. This ratio was unchanged in the presence of
aphidicolin, but increased ?20 fold in the presence of actino-
mycinD,anRNAprimaseinhibitor.Theseresultssuggestthat
most chromatin-bound MCM complexes are competent to
bind CDC45 but that productive initiation at the first MCM
complex inactivates neighboring complexes up to a certain
Review articles
122
BioEssays 25.2

Page 8

distance (Fig. 4B). In summary, these data suggest that a
single ORC recruits many potential start sites that are spread
over a zone of several kb, as previously suggested.(58)
One unresolved question with this model is that, in
S. cerevisiae, which also shows an excess of bound MCM
proteins compared to ORC, there seems to be a very tight
association of replication initiation with ORC binding site.
Therefore,evenifMCMspreadingoccursinyeast,therewould
appear to be preferential activation of ORC-proximal MCM
complexes. Edwards et al.(69)point out that, in yeast, CDC7
recruitment to origins requires ORC,(83)whereas in higher
eukaryotes CDC7 recruitment is MCM-dependent but ORC-
independent.(84)This may explain why ORC-distal MCM
complexes are initiation-competent only in higher eukaryotes.
Another consideration is that potentially this model in-
creasesthetimeneededforlicensing.ItisunclearifMCMscan
be loaded away from ORC directly by some looping mechan-
ism, or if they have first to bind chromatin close to ORC and
then move away. Assuming this movement is at a similar rate
toforkmovement(0.5kb/min)andbidirectional,itwouldtakea
few minutes for MCMs to spread over a few kb around ORC
beforelicensingiscomplete.Thiswouldbeconsistentwiththe
observation that complete MCM binding is only achieved
several minutes after ORC binding to sperm chromatin in egg
extracts.(84)How the kinetics of licensing in early embryos
compares to that observed in egg extracts remains to be
determined.
It is interesting to consider the difference between a mech-
anism that limits the assembly of potential origins at regular
intervals, and one that limits the firing/binding of CDC45 to
one out of multiple MCM complexes. In the first case, origins
cannot fire at close intervals whereas, in the second case,
closely spaced bubbles may still occur if inactivation of the
neighboring MCM complexes cannot extend to those loaded
from a different ORC. The puzzling observation that only a
single initiation event occurs on a 9 kb plasmid whereas
adjacentinitiationsonlargerplasmidsarefrequentlyspacedat
intervals much smaller than 9 kb(58)would be consistent with
the second mechanism.
Conclusions
Overall,thepublisheddatastronglyargueinfavorofthe‘‘origin
redundancy’’ model (Fig. 2B). First, the observed distributions
of replication eyes along the DNA are easily explained by this
model.Second,MCMcomplexesareabundantanddispersed
and, at least in Xenopus, all appear competent for initiation as
predicted for redundant potential origins. Third, the frequency
of initiation appears to increase markedly through S phase in
order to maintain a constant fork density despite replicon
fusion. In addition, there is evidence for ‘‘lateral inhibition’’
mechanisms at both the licensing and the activation stage.
MCM loading appears to restrict ORC binding to once every
5–10 kb even on very simple DNA substrates, a potential
mechanism to regularize ORC spacing and therefore max-
imize MCM loading. In addition, productive initiation prevents
CDC45loadingoversomedistance,apotentialmechanismfor
origin interference in advance of fork progression. Several
questions remain, however. Do adjacent MCM-loading zones
coalesce or do MCM-free gaps remain in unreplicated DNA?
Are ORC-proximal and ORC-distal MCM complexes equally
likely to support initiation? How are origin interference and the
frequency of initiation controlled? In early embryos, all the
above mechanisms are likely to cooperate in order to ensure
the timely replication of the early embryonic genome. In adult
somatic cells, replication initiates either at specific sites or at
broadbutcircumscribedzones,whichmayreflectthedifferent
extent of MCM spreading at different loci.
Acknowledgments
We apologize to the many colleagues whose work was not
cited due to space limitations. We thank E. Heard, M.-N.
Prioleau,M.Chevrier-MillerandT.Germeforcriticalreadingof
Figure 4. Model of replication initiation to account for
abundant potential origins.(58,69)A: Prior to S phase, a
single ORC molecule recruits multiples MCM2-7 com-
plexes which spread several kb away from ORC. B:
Following activation by CDC7/DBF4 and S-CDKs, the
loadingofCDC45atanyoftheselargenumberofMCM2-
7complexesdefinesthestartofDNAreplication(red)and
leads to inactivation of neighbouring MCM complexes.
Adapted from Edwards et al.(69)
Review articles
BioEssays 25.2
123

Page 9

the manuscript, J.-L. Sikorav for discussions and the referees
for their constructive criticism.
References
1. Bell SP, Dutta A. DNA replication in eukaryotic cells. Annu Rev Biochem
2002;71:333–374.
2. Canman CE. Replication checkpoint: preventing mitotic catastrophe.
Curr Biol 2001;11:R121–R124.
3. Hyrien O, Me ´chali M. Chromosomal replication initiates and terminates
at random sequences but at regular intervals in the ribosomal DNA of
Xenopus early embryos. EMBO J 1993;12:4511–4520.
4. Shinomiya T, Ina S. Analysis of chromosomal replicons in early embryos
of Drosophila melanogaster by two-dimensional gel electrophoresis.
Nucleic Acids Res 1991;19:3935–3941.
5. Kimelman D, Kirschner M, Scherson T. The events of the midblastula
transition in Xenopus are regulated by changes in the cell cycle. Cell
1987;48:399–407.
6. Raff JW, Glover DM. Nuclear and cytoplasmic mitotic cycles continue in
Drosophila embryos in which DNA synthesis is inhibited with aphidicolin.
J Cell Biol 1988;107:2009–2019.
7. Ishimi Y. A DNA helicase activity is associated with an MCM4, -6, and -7
protein complex. J Biol Chem 1997;272:24508–24513.
8. Schwacha A, Bell SP. Interactions between two catalytically distinct
MCM subgroups are essential for coordinated ATP hydrolysis and DNA
replication. Mol Cell 2001;8:1093–1104.
9. Labib K, Diffley JF. Is the MCM2-7 complex the eukaryotic DNA replica-
tion fork helicase? Curr Opin Genet Dev 2001;11:64–70.
10. Bell SP, Stillman B. ATP-dependent recognition of eukaryotic origins
of DNA replication by a multiprotein complex. Nature 1992;357:128–134.
11. Diffley JF, Cocker JH. Protein-DNA interactions at a yeast replication
origin. Nature 1992;357:169–172.
12. Bell SP. The origin recognition complex: from simple origins to complex
functions. Genes Dev 2002;16:659–672.
13. Coleman TR, Carpenter PB, Dunphy WG. The Xenopus Cdc6 protein is
essential for the initiation of a single round of DNA replication in cell-free
extracts. Cell 1996;87:53–63.
14. Maiorano D, Moreau J, Mechali M. XCDT1 is required for the assembly of
pre-replicative complexes in Xenopus laevis. Nature 2000;404:622–625.
15. Nishitani H, Lygerou Z, Nishimoto T, Nurse P. The Cdt1 protein is re-
quired to license DNA for replication in fission yeast. Nature 2000;404:
625–628.
16. Hua XH, Newport J. Identification of a preinitiation step in DNA replica-
tion that is independent of origin recognition complex and cdc6, but
dependent on cdk2. J Cell Biol 1998;140:271–281.
17. Rowles A, Tada S, Blow JJ. Changes in association of the Xenopus origin
recognition complex with chromatin on licensing of replication origins.
J Cell Sci 1999;112:2011–2018.
18. Mimura S, Takisawa H. Xenopus Cdc45-dependent loading of DNA
polymerase alpha onto chromatin under the control of S-phase Cdk.
EMBO J 1998;17:5699–5707.
19. Walter J, Newport J. Initiation of eukaryotic DNA replication: origin un-
winding and sequential chromatin association of Cdc45, RPA, and DNA
polymerase alpha. Mol Cell 2000;5:617–627.
20. Aparicio OM, Weinstein DM, Bell SP. Components and dynamics of DNA
replication complexes in S. cerevisiae: redistribution of MCM proteins
and Cdc45p during S phase. Cell 1997;91:59–69.
21. Tanaka T, Knapp D, Nasmyth K. Loading of an Mcm protein onto DNA
replication origins is regulated by Cdc6p and CDKs. Cell 1997;90:649–
660.
22. Santocanale C, Sharma K, Diffley JF. Activation of dormant origins of
DNA replication in budding yeast. Genes Dev 1999;13:2360–2364.
23. Tadokoro R, Fujita M, Miura H, Shirahige K, Yoshikawa H, Tsurimoto T,
Obuse C. Scheduled conversion of replication complex architecture at
replication origins of Saccharomyces cerevisiae during the cell cycle.
J Biol Chem 2002;277:15881–15889.
24. Madine MA, Khoo CY, Mills AD, Laskey RA. MCM3 complex required for
cell cycle regulation of DNA replication in vertebrate cells. Nature 1995;
375:421–424.
25. Chong JP, Mahbubani HM, Khoo CY, Blow JJ. Purification of an MCM-
containing complex as a component of the DNA replication licensing
system. Nature 1995;375:418–421.
26. Dimitrova DS, Todorov IT, Melendy T, Gilbert DM. Mcm2, but not RPA, is
a component of the mammalian early G1-phase prereplication complex.
J Cell Biol 1999;146:709–722.
27. Todorov IT, Attaran A, Kearsey SE. BM28, a human member of the
MCM2-3-5 family, is displaced from chromatin during DNA replication.
J Cell Biol 1995;129:1433–1445.
28. Krude T, Musahl C, Laskey RA, Knippers R. Human replication proteins
hCdc21, hCdc46 and P1Mcm3 bind chromatin uniformly before S-phase
and are displaced locally during DNA replication. J Cell Sci 1996;109:
309–318.
29. Romanowski P, Madine MA, Laskey RA. XMCM7, a novel member of
the Xenopus MCM family, interacts with XMCM3 and colocalizes with
it throughout replication. Proc Natl Acad Sci USA 1996;93:10189–
10194.
30. Nguyen VQ, Co C, Li JJ. Cyclin-dependent kinases prevent DNA
re-replication through multiple mechanisms. Nature 2001;411:1068–
1073.
31. McGarry TJ, Kirschner MW. Geminin, an inhibitor of DNA replication, is
degraded during mitosis. Cell 1998;93:1043–1053.
32. Tada S, Li A, Maiorano D, Mechali M, Blow JJ. Repression of origin
assembly in metaphase depends on inhibition of RLF-B/Cdt1 by geminin.
Nat Cell Biol 2001;3:107–113.
33. Ritzi M, Baack M, Musahl C, Romanowski P, Laskey RA, Knippers R.
Human minichromosome maintenance proteins and human origin re-
cognition complex 2 protein on chromatin. J Biol Chem 1998;273:24543–
24549.
34. Gilbert DM. Making sense of eukaryotic DNA replication origins. Science
2001;294:96–100.
35. Newlon CS, Theis JF. The structure and function of yeast ARS elements.
Curr Opin Genet Dev 1993;3:752–758.
36. Labib K, Kearsey SE, Diffley JF. MCM2-7 proteins are essential com-
ponents of prereplicative complexes that accumulate cooperatively in
the nucleus during G1-phase and are required to establish, but not
maintain, the S-phase checkpoint. Mol Biol Cell 2001;12:3658–3667.
37. Bielinsky AK, Gerbi SA. Chromosomal ARS1 has a single leading strand
start site. Mol Cell 1999;3:477–486.
38. Krysan PJ, Haase SB, Calos MP. Isolation of human sequences that
replicate autonomously in human cells. Mol Cell Biol 1989;9:1026–1033.
39. Krysan PJ, Calos MP. Replication initiates at multiple locations on an
autonomously replicating plasmid in human cells. Mol Cell Biol 1991;11:
1464–1472.
40. Abdurashidova G, Deganuto M, Klima R, Riva S, Biamonti G, Giacca M,
Falaschi A. Start sites of bidirectional DNA synthesis at the human lamin
B2 origin. Science 2000;287:2023–2026.
41. Aladjem MI, Rodewald LW, Kolman JL, Wahl GM. Genetic dissection of
a mammalian replicator in the human beta-globin locus. Science 1998;
281:1005–1009.
42. Dijkwel PA, Wang S, Hamlin JL. Initiation sites are distributed at frequent
intervals in the chinese hamster dihydrofolate reductase origin of replica-
tion but are used with very different efficiencies. Mol Cell Biol 2002;22:
3053–3065.
43. Dijkwel PA, Mesner LD, Levenson VV, d’Anna J, Hamlin JL. Dispersive
initiation of replication in the Chinese hamster rhodopsin locus. Exp Cell
Res 2000;256:150–157.
44. Little RD, Platt TH, Schildkraut CL. Initiation and termination of DNA
replication in human rRNA genes. Mol Cell Biol 1993;13:6600–6613.
45. Hyrien O, Maric C, Me ´chali M. Transition in specification of embryonic
metazoan DNA replication origins. Science 1995;270:994–997.
46. Sasaki T, Sawado T, Yamaguchi M, Shinomiya T. Specification of regions
of DNA replication initiation during embryogenesis in the 65-kilobase
DNApolalpha-dE2F locus of Drosophila melanogaster. Mol Cell Biol
1999;19:547–555.
47. Maric C, Levacher B, Hyrien O. Developmental regulation of replication
fork pausing in Xenopus laevis is ribosomal RNA genes. J Mol Biol 1999;
291:775–788.
48. Wu JR, Gilbert DM. A distinct G1step required to specify the Chinese
hamster DHFR replication origin. Science 1996;271:1270–1272.
Review articles
124
BioEssays 25.2

Page 10

49. Dimitrova DS, Prokhorova TA, Blow JJ, Todorov IT, Gilbert DM. Mam-
malian nuclei become licensed for DNA replication during late telophase.
J Cell Sci 2002;115:51–59.
50. Gilbert DM. Replication timing and transcriptional control: beyond cause
and effect. Curr Op Cell Biol 2002;14:377–383.
51. Raghuraman MK, Brewer BJ, Fangman WL. Cell cycle-dependent
establishment of a late replication program. Science 1997;276:806–809.
52. Dimitrova DS, Gilbert DM. The spatial position and replication timing of
chromosomal domains are both established in early G1phase. Mol Cell
1999;4:983–993.
53. Li F, Chen J, Izumi M, Butler MC, Keezer SM, Gilbert DM. The replication
timing program of the Chinese hamster beta-globin locus is established
coincident with its repositioning near peripheral heterochromatin in early
G1phase. J Cell Biol 2001;154:283–292.
54. Heun P, Laroche T, Raghuraman MK, Gasser SM. The positioning and
dynamics of origins of replication in the budding yeast nucleus. J Cell
Biol 2001;152:385–400.
55. Blumenthal AB, Kriegstein HJ, Hogness DS. The units of DNA replication
in Drosophila melanogaster chromosomes. Cold Spring Harb Symp
Quant Biol 1974;38:205–223.
56. Blow JJ, Watson JV. Nuclei act as independent and integrated units
of replication in a Xenopus cell-free DNA replication system. EMBO
J 1987;6:1997–2002.
57. Walter J, Newport JW. Regulation of replicon size in Xenopus egg
extracts. Science 1997;275:993–995.
58. Lucas I, Chevrier-Miller M, Sogo JM, Hyrien O. Mechanisms ensuring
rapid and complete DNA replication despite random initiation in Xenopus
early embryos. J Mol Biol 2000;296:769–786.
59. Herrick J, Stanislawski P, Hyrien O, Bensimon A. Replication fork density
increases during DNA synthesis in X. laevis egg extracts. J Mol Biol
2000;300:1133–1142.
60. Blow JJ, Gillespie PJ, Francis D, Jackson DA. Replication origins in
Xenopus egg extract are 5–15 kilobases apart and are activated in
clusters that fire at different times. J Cell Biol 2001;152:15–26.
61. Marheineke K, Hyrien O. Aphidicolin triggers a block to replication origin
firing in Xenopus egg extracts. J Biol Chem 2001;276:17092–17100.
62. Mills AD, Blow JJ, White JG, Amos WB, Wilcock D, Laskey RA. Replica-
tion occurs at discrete foci spaced throughout nuclei replicating in vitro.
J Cell Sci 1989;94:471–477.
63. O’Keefe RT, Henderson SC, Spector DL. Dynamic organization of DNA
replication in mammalian cell nuclei: spatially and temporally defined
replication of chromosome-specific alpha-satellite DNA sequences.
J Cell Biol 1992;116:1095–1110.
64. Wolffe AP. Replication timing and Xenopus 5S RNA gene transcription
in vitro. Dev Biol 1993;157:224–231.
65. Mahbubani HM, Paull T, Elder JK, Blow JJ. DNA replication initiates at
multiple sites on plasmid DNA in Xenopus egg extracts. Nucleic Acids
Res 1992;20:1457–1462.
66. Laskey RA. Chromosome replication in early development of Xenopus
laevis. J Embryol Exp Morphol 1985;89 Suppl:285–296.
67. Rowles A, Chong JP, Brown L, Howell M, Evan GI, Blow JJ. Interaction
between the origin recognition complex and the replication licensing
system in Xenopus. Cell 1996;87:287–296.
68. Mahbubani HM, Chong JP, Chevalier S, Thommes P, Blow JJ. Cell cycle
regulation of the replication licensing system: involvement of a Cdk-
dependent inhibitor. J Cell Biol 1997;136:125–135.
69. Edwards MC, Tutter AV, Cvetic C, Gilbert CH, Prokhorova TA, Walter JC.
MCM2-7 complexes bind chromatin in a distributed pattern surrounding
ORC in Xenopus egg extracts. J Biol Chem 2002;277:33049–33057.
70. Blow JJ, Laskey RA. Initiation of DNA replication in nuclei and purified
DNA by a cell-free extract of Xenopus eggs. Cell 1986;47:577–587.
71. Blow JJ, Sleeman AM. Replication of purified DNA in Xenopus egg
extract is dependent on nuclear assembly. J Cell Sci 1990;95:383–391.
72. Hyrien O, Me ´chali M. Plasmid replication in Xenopus eggs and egg
extracts: a 2D gel electrophoretic analysis. Nucleic Acids Res 1992;20:
1463–1469.
73. Michalet X, et al. Dynamic molecular combing: stretching the whole
human genome for high-resolution studies. Science 1997;277:1518–
1523.
74. Kolmogorov AN. On the statistical theory of crystallization in metals. Izv
Akad Nauk SSSR, Ser Fiz 1937;1:355–359.
75. Herrick J, Jun S, Bechhoefer J, Bensimon A. Kinetic model of DNA
replication in eukaryotic organisms. J Mol Biol 2002;320:741–750.
76. Parra I, Windle B. High resolution visual mapping of stretched DNA by
fluorescent hybridization. Nat Genet 1993;5:17–21.
77. Raghuraman MK, Winzeler EA, Collingwood D, Hunt S, Wodicka L,
Conway A, Lockhart DJ, Davis RW, Brewer BJ, Fangman WL. Replication
dynamics of the yeast genome. Science 2001;294:115–121.
78. Wyrick JJ, Aparicio JG, Chen T, Barnett JD, Jennings EG, Young RA, Bell
SP, Aparicio OM. Genome-wide distribution of ORC and MCM proteins in
S. cerevisiae: high-resolution mapping of replication origins. Science
2001;294:2357–2360.
79. Bielinsky AK, Blitzblau H, Beall EL, Ezrokhi M, Smith HS, Botchan MR,
Gerbi SA. Origin recognition complex binding to a metazoan replication
origin. Curr Biol 2001;11:1427–1431.
80. Ladenburger EM, Keller C, Knippers R. Identification of a binding region
for human origin recognition complex proteins 1 and 2 that coincides
with an origin of DNA replication. Mol Cell Biol 2002;22:1036–1048.
81. Austin RJ, Orr-Weaver TL, Bell SP. Drosophila ORC specifically binds to
ACE3, an origin of DNA replication control element. Genes Dev 1999;13:
2639–2649.
82. Alexandrow MG, Ritzi M, Pemov A, Hamlin JL. A potential role for mini-
chromosome maintenance (MCM) proteins in initiation at the dihydrofo-
late reductase replication origin. J Biol Chem 2002;277:2702–2708.
83. Pasero P, Duncker BP, Schwob E, Gasser SM. A role for the Cdc7 kinase
regulatory subunit Dbf4p in the formation of initiation-competent origins
of replication. Genes Dev 1999;13:2159–2176.
84. Jares P, Blow JJ. Xenopus cdc7 function is dependent on licensing but
not on XORC, XCdc6, or CDK activity and is required for XCdc45
loading. Genes Dev 2000;14:1528–1540.
Review articles
BioEssays 25.2
125