Bacteriophage replication modules

Christoph Weigel & Harald Seitz

Max-Planck-Institut für molekulare Genetik, Berlin, Germany

Correspondence: Christoph Weigel, Abstract

Technical University Berlin, Faculty III, Inst. for
Biotechnology, Sekr. MA 5-11, Straße des 17.
Bacteriophages (prokaryotic viruses) are favourite model systems to study DNA
Juni 136, D-10623 Berlin, Germany. replication in prokaryotes, and provide examples for every theoretically possible
Tel.: 149 30 84131614; replication mechanism. In addition, the elucidation of the intricate interplay of
e-mail: weigel@molgen. mpg.de phage-encoded replication factors with ‘host’ factors has always advanced the
understanding of DNA replication in general. Here we review bacteriophage
Received 22 March 2005; revised 6 September replication based on the long-standing observation that in most known phage
2005; accepted 8 November 2005. genomes the replication genes are arranged as modules. This allows us to discuss
First published online 21 March 2006.
established model systems – f1/fd, fX174, P2, P4, l, SPP1, N15, f29, T7 and T4 –
along with those numerous phages that have been sequenced but not studied
doi:10.1111/j.1574-6976.2006.00015.x
experimentally. The review of bacteriophage replication mechanisms and modules
Editor: Ramon Diaz Orejas
is accompanied by a compendium of replication origins and replication/recombi-
nation proteins (available as supplementary material online).
Keywords
DNA polymerase; helicase; helicase loader;
initiator; primase; replication origin; replication
fork.

Contents
Introduction fP4a-type helicase-primase encoding replication
Replication mechanisms modules
Initiation by nicking: ‘rolling circle’-type DNA replication Phages-encoding DNA polymerases
Initiation by melting: theta (y)-type DNA replication The phage T4-type replication module
Initiation at the ends of linear DNA: protein-primed DNA replication The phage T7-type replication module
Initiation of DNA replication by transcription The phage D29-type replication module
Recombination-dependent DNA replication The replication modules of the phages K, Bxz1 and T5
Replication restart The phage f29-type replication module
Bacteriophage replication modules Replication modules of phages replicating by RCR
Phages encoding initiator proteins Phage replicons lacking replication protein genes
‘Initiator-solo’ replication modules Evolutionary considerations
‘Initiator-helicase loader’ replication modules The different types of phage-encoded helicases
‘Initiator-helicase’ replication modules Phage-encoded homologues of the E.coli DanaB helicase
‘Initiator-helicase loader-helicase’ replication modules Chromosomally encoded homologues of phage helicase loaders
Conclusions for ‘Phages encoding initiator proteins’ Perspectives
Replication module exchange among bacteriophages Acknowledgements, Supplementary material, References

Introduction
Chromosomes, plasmids and bacteriophages (bacterial propagation are a replication origin – the replicator – and
viruses) represent three of the four genetic elements that an initiator, in most cases a protein (Jacob et al., 1963).
are, permanently or transiently, present in a prokaryotic cell. Transposable elements, the fourth type, are covalently linked
They are entities – replicons in the terminology of the to one of the other genetic elements and therefore not
‘replicon model’ – whose key regulatory elements for considered as replicons.

FEMS Microbiol Rev 30 (2006) 321–381

c2006 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 322 C. Weigel & H. Seitz

In spite of this clear-cut definition of a replicon, a proteins from host proteins recruited for replication. This
satisfactory definition of ‘bacteriophages’ remains proble- approach leads to the elucidation of its replication mechan-
matic. Mu, to give just one example, is a typical temperate ism. Alternatively, one can determine the replication module
phage of its host Escherichia coli for all but one stage of its by subjecting a phage genome to a thorough homology search
‘life-cycle’: its genome propagates as transposon (Nakai in the available databases. A replication mechanism cannot be
et al., 2001; Morgan et al., 2002). Taxonomy-oriented reliably predicted by this approach unless the replication
biology tends to classify phages, plasmids and transposons genes of the phage genome under study are similar to those
as ‘mobile genetic elements’ to account for such overlaps of one of the established model systems.
(Toussaint & Merlin, 2002) and resorts to chromosomal In addition to discussing the replication modules of
genes for the definition of bacterial species. However, established model systems we will explore whether the
bacterial chromosomes contain a variety of intact, defective, concept of ‘replication modules’ can lead to a better under-
and degraded prophages, i.e. phages in the integrated state standing of the replication of those numerous phages that
and their remnants. Hence the taxonomy-oriented approach have been sequenced but not studied experimentally. We will
cannot cut the Gordian knot: prophage genes account for a thus evaluate which of the well-studied phages are valid
notable portion of the genes in bacterial chromosomes model systems and which should be regarded as unique
(Casjens, 2003; Canchaya et al., 2003) and are largely cases.
responsible for phenotypic variations among the strains of Taking into account that most readers prefer the printed
bacterial species, including such important traits as patho- version of a paper for studying a topic from a broader
genicity (Banks et al., 2002; Bruessow et al., 2004). Replica- viewpoint and the online version for selective searches, we
tion research has traditionally focused on chromosomes, have decided to present the different aspects of phage
plasmids and phages as experimental systems without caring replication in two parts:
much about their exact classification, and we follow this (1) Bacteriophage replication mechanisms and replication
route here. modules are discussed in this part of the review.
Bacterial chromosomes are fully autonomous genetic ele- (2) A compendium of phage replication origins and phage
ments, i.e. they carry all genes whose products are required replication/recombination proteins is presented in the sup-
for their replication. Also bacteriophage and plasmid repli- plementary material available online.
cons multiply as intact entities but their autonomy is limited Note: all parts of the ‘compendium of origins and pro-
due to their partial or complete dependence on factors teins’ referred to in the following are marked with ‘COM’ to
encoded by chromosomal genes for reproduction. The eluci- encourage and facilitate navigation between the two parts.
dation of the intricate interplay of phage- or plasmid-encoded In addition, all numbers of Sections, Tables and Figures of
replication factors with chromosomally encoded ‘host’ factors the compendium are ‘tagged’ with the prefix ‘C’.
has always advanced the understanding of both systems.
Chromosome replication has been studied for decades in E.
coli and Bacillus subtilis – more recently also in Streptomyces
Replication mechanisms
lividans (reviewed in Marians, 1996; Messer & Weigel, 1996; The structure of double-stranded B-DNA ‘immediately
Kogoma, 1997; Moriya et al., 1999; Messer, 2002; Messer & suggests a possible copying mechanism for the genetic
Zakrewska-Czerwinska, 2002). A comprehensive review of material’ (Watson & Crick, 1953), and three possible
plasmid replication has been presented by del Solar and molecular mechanisms for the initiation of this copying
colleagues (del Solar et al., 1998). Reviews of the replication process, DNA replication: (1) ‘nicking’, i.e. the breakage of
of phages with DNA genomes focus on favourite model the covalent phosphodiester bond between two neighbour-
systems: l (Campbell, 1994; Taylor & Wegrzyn, 1995), T4 ing bases on one strand; (2) ‘melting’, i.e. the localised
(Mosig, 1998), f29 (Meijer et al., 2001), f1/fd (Horiuchi, disruption of the hydrogen bonds that tether together the
1997), SPP1 (Alonso et al., 2005) and T7 (Richardson, 1983). two complementary DNA strands; and (3) melting of the
Several recent research papers and reviews cover the field of terminal hydrogen bonds of linear double-stranded DNA
RNA phages, which will not be discussed here (Bollback & (dsDNA) molecules. All three initiation mechanisms gen-
Huelsenbeck, 2001; Chetverin, 2004; Makeyev & Grimes, 2004; erate single-stranded regions as templates for the synthesis
Mindich, 2004; Poranen & Tuma, 2004). of complementary daughter strands, resulting in what is
We will discuss bacteriophage replication based on a long- known as semi-conservative DNA replication since the hall-
standing observation: genes encoding replication functions mark experiments of Meselson & Stahl (1958).
tend to be located close to each other in many phage These three possible initiation mechanisms have been
genomes, resulting in what has been termed a ‘replication studied in detail for circular and linear dsDNA phage
module’. The replication module of a phage can be deter- replicons. The replication of linear dsDNA molecules seems
mined experimentally by dissecting cognate (phage-encoded) straightforward, irrespective of the initiation mechanism.

c 2006 Federation of European Microbiological Societies FEMS Microbiol Rev 30 (2006) 321–381
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 Bacteriophage replication modules 323

However, all known (replicative) DNA polymerases synthe- replication mechanism operating in a phage replicon under
sise DNA exclusively in 5 0!3 0 direction and require a study.
primer, mostly a short oligo-ribonucleotide. Thus, the 5 0 -
ends of both strands cannot be replicated, resulting in a loss Initiation by nicking: ‘rolling circle’-type DNA
of genetic information during successive rounds of replica- replication
tion. Research on phage replication has revealed four
Replicons that propagate by the ‘rolling circle’ mode of DNA
different molecular mechanisms to overcome this problem:
replication (RCR) include bacterial phages and plasmids
(1) Bacilus subtilis phage f29 uses a specialised protein as
with circular dsDNA genomes (Khan, 1997, 2000). In all
‘portable primer’ that remains covalently attached to both
cases that have been studied experimentally, a DNA-bound
genome ends after completion of replication.
initiator protein nicks one strand of the dsDNA molecules.
(2) Escherichia coli phage T7 uses direct terminal repeats
The 5 0 -end of the disrupted DNA strand becomes covalently
that are regenerated by processing of genome concatemers
linked to a specific tyrosine residue of the initiator, while the
during the packaging of monomeric genomes into phage
free 3 0 -OH end is elongated by a replisome. After one round
heads.
of replication, the strand-transfer reaction is reversed,
(3) Escherichia coli phage N15 employs a specialised en-
liberating a single-stranded from a double-stranded mole-
zyme, protelomerase, to (re)generate the covalently closed
cule in a reaction that does not require recombination. DNA
ends of the linear double-stranded prophage genome after
replication is completed by the synthesis of the complemen-
replication of a circular intermediate.
tary strand of the single-stranded molecule. For the phages
(4) Many phage genomes that enter the cell in a linear form
with single-stranded circular DNA genomes, e.g. fX174,
are converted to a circular form prior to replication or,
M13 and fd, this ‘completion-step’ is the conversion of the
alternatively, integration into the host chromosome as
viral or (1)-strand to the double-stranded ‘replicative form’
prophage.
(RF).
The genetic information is faithfully conserved during
It was shown by Horiuchi and co-workers for the
replication of covalently closed circular dsDNA molecules
filamentous E. coli phage fd that the nicking reaction is
but the helical nature of DNA creates a topological problem:
preceded by a localised DNA unwinding around the nick-
the progeny molecules are intertwined and require a recom-
site (Higashitani et al., 1994) (COM section C2.1.1.). This
bination step for resolution. A comparable problem arises
points to the reason why RCR has so far only been found for
from cutting a Moebius ribbon with 2n twists along the
circular dsDNA replicons: protein-induced DNA unwinding
middle. Escherichia coli phage l replication is initiated by
is apparently only possible with negatively supercoiled, i.e.
the ‘melting’ mechanism early after infection, and proceeds
undertwisted, DNA. Small linear dsDNA molecules cannot
by simultaneous synthesis of both daughter strands, thus
be undertwisted because both strands rotate freely around
creating catenated progeny molecules that are resolved by
each other, with one exception: linear dsDNA with cova-
host topoisomerases. Later during infection, the circular
lently closed ends as in the fN15 prophage (see next
progeny molecules are converted, probably by recombi-
section). The situation is more complex with large linear
nation proteins, to structures that allow the continuous
DNA molecules, e.g. bacterial chromosomes, where an
synthesis of (linear) concatemeric phage DNA. These con-
intricate interplay of topoisomerases, gyrases and a number
catemers are finally processed by the phage packaging
of nucleoid-associated proteins (among others: HU, H-NS)
apparatus to yield monomeric linear genomes. A complete
creates transient ‘topological domains’ of undertwisted
understanding of the replication mechanisms of circular
DNA that are anchored to cell structures (Worcel & Burgi,
replicons includes knowledge of cognate recombination
1972; Postow et al., 2004).
processes, therefore. Escherichia coli phage P2 avoids the
topological problem by replicating each parent DNA strand Phage fd: Replication in the ‘rolling circle’ mode was for
of its circular(ised) genome separately, involving a single- some time considered specific for small plasmid and phage
stranded replication intermediate not catenated with the replicons. As we know today, large conjugative plasmids use
dsDNA molecule. RCR coupled to a specific secretion system for DNA transfer
The genomes of phages P2 (33.6 kb), T7 (39.9 kb), N15 to recipient cells (Llosa et al., 2002) and also a number of
(46.4 kb) and l (48.5 kb) are fairly similar in size but their phages with mid-sized genomes (30 kb) replicate via RCR
replication follows different routes, as outlined briefly (see below). We first discuss the successive steps of RCR of
above. Only the experimental and/or computational search the filamentous E. coli phage fd as an example (Fig. 1),
for replication origin structures and genes encoding replica- which largely resembles RCR of the isometric phage fX174:
tion proteins, i.e. the elucidation of the ‘replication module’, Step 1. The single-stranded circular (1)-strand DNA of ffd
together with a comparison with the known mechanisms that enters the host cell is covered by host single-strand
discussed below can lead to a prediction of the likely binding protein (SSB) except for the single-strand origin

FEMS Microbiol Rev 30 (2006) 321–381

c 2006 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 324 C. Weigel & H. Seitz

Fig. 1. The mechanism of ‘rolling circle’ repli-

cation: phage fd. Successive steps are indicated
by open arrows/arrowheads plus numbering
(see text for details). Complementary DNA
strands are shown as parallel lines; twisted lines
indicate (negative) supercoiling of the DNA
molecule. Dark blue/green: parental DNA
strands; light blue/green: daughter DNA
strands; yellow: RNA primers. Red dots indicate
free 3 0 -hydroxyls used for DNA synthesis.
Phage-encoded proteins are shown as coloured
circles, host factors as coloured triangles. The
colour code used for individual proteins
corresponds to the coloured protein names in
the text.

(sso). The sso is recognised by the host (s70)RNA polymer- Step 3. The dso of ffd is located close to the sso, and consists
ase although its structure does not resemble a typical of two structural elements: the so-called nick-site and,
promoter (Kaguni & Kornberg, 1982) (COM section adjacent, the binding sites for gpII. gpII bound to its binding
C2.1.). RNA polymerase synthesises a short untranslated sites on a supercoiled substrate induces a conformation of
transcript (20 nt; Higashitani et al., 1993) that is elongated the nucleoprotein complex that results in the localised
for (  )-strand synthesis by host DNA Pol III holoenzyme. unwinding of 7 bp encompassing the nick-site as a prerequi-
At this step, fX174 requires assembly of the restart primo- site for the nicking reaction (COM section C2.1.1C3.1.1.).
some (formerly PriA primosome, see ‘Replication restart’ Steps 4 and 5. The nicking reaction is performed by an
section) at the sso, which directs primer synthesis by DnaG appropriately positioned gpII protomer within the oligo-
primase and (  )-strand synthesis by DNA Pol III holoen- meric complex II. Nicking occurs simultaneously with the
zyme. A third variation exists for complementary- (transient) covalent linkage of the (1)-strand 5 0 -end to a
strand synthesis of phage G4: DnaG primase alone was specific tyrosine residue of gpII. Only for clarity, this
found to be required for efficient primer synthesis in vitro reaction is shown as two separate steps in Fig. 1.
(Bouché et al., 1978; Stayton & Kornberg, 1983; Stayton Step 6. The unwound region serves as entry site for the host
et al., 1983; Hiasa et al., 1989). In all three systems, primer Rep helicase, which dimerises upon DNA-binding and un-
removal is performed by PolA and gap sealing by DNA winds the duplex in 3 0 ! 5 0 direction; it is not known
ligase. whether gpII attracts Rep by direct physical interaction
Step 2. The conversion of the viral (1)-strand DNA into the (Hours & Denhardt, 1979; Takahashi et al., 1979; Meyer &
covalently closed circular RF is completed upon introduc- Geider, 1982; Chao & Lohman, 1991).
tion of negative supercoils by the host gyrase. The phage Step 7. Strand-displacement synthesis starting from the free
initiator gpII binds to the double-strand origin (dso) in the 3 0 -OH end is performed by host DNA Pol III holoenzyme
linear or (relaxed) circular forms (‘complex I’) but origin (Meyer & Geider, 1982). Replication intermediates at this
unwinding and nicking by ‘complex II’ requires the nega- stage appear in the electron microscope as dsDNA circles
tively supercoiled form (COM section C2.1.1.). with attached single-stranded loops, thus the term ‘rolling

c 2006 Federation of European Microbiological Societies FEMS Microbiol Rev 30 (2006) 321–381
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 Bacteriophage replication modules 325

Fig. 2. The mechanism of ‘rolling circle’ repli-

cation (RCR): phage P2. Successive steps are
indicated by open arrows/arrowheads plus
numbering (see text for details). Complemen-
tary DNA strands are shown as parallel lines;
twisted lines indicate (negative) supercoiling of
the DNA molecule. Dark blue/green: parental
DNA strands; light blue/green: daughter DNA
strands; yellow: RNA primers. Red dots indicate
free 3 0 -hydroxyls used for DNA synthesis.
Phage-encoded proteins are shown as coloured
circles, host factors as coloured triangles. The
colour code used for individual proteins corre-
sponds to the coloured protein names in the
text.

circle’ DNA replication (Gilbert & Dressler, 1968). Depend- to the tyrosine residue of gpII is transferred back to the free
ing on the conditions used for sample preparation, the 3 0 -OH end.
single-stranded part can appear as a single-stranded tail Step 9a. The single-stranded replication intermediate
rather than as circle, giving the molecules a shape resem- of step 8 represents the (1)-strand, i.e. the phage genome,
bling a Greek sigma (s) (Allison et al., 1977). RCR and s- which associates with gpV SSB already during liberation
type DNA replication (sDR) are frequently used synony- from the double-stranded replication intermediate. The
mously in the literature (see also Kornberg & Baker, 1992). rapid association of the (1)-strand DNA with gpV SSB
However, the lengths of the ‘tails’ derived from RCR never and, subsequently, with other coat proteins (gpVII, gpIX)
exceed the contour lengths of the circular parent molecules. (Feng et al., 1997) prevents strand-switching or coupled
In this review, we reserve the term sDR for a replication leading- and lagging-strand DNA synthesis by the host
mechanism that also produces ‘tailed’ molecules but in- replisome.
volves a recombination step for initiation; an example is the Step 9b. The RF is restored from the double-stranded
switch from yDR to sDR at later stages during l replication replication intermediate of step 8 by the action of host
(see below). ‘Tails’ derived from sDR are always (partially) DNA ligase and gyrase, allowing the resumption of the
double-stranded owing to coupled leading- and lagging- replication cycle with step 2. It has not been firmly proven
strand DNA synthesis and are, as genome concatemers, that the gpII initiator remains bound to the dso throughout
usually much longer than the circular parent molecules to all steps of the replication cycle, as tentatively shown in Fig.
which they are attached. 1. Horiuchi and co-workers have shown that a small stretch
Step 8. When the replisome reaches the initial nick-site, the upstream of the nick-site is important for the termination
displaced single strand is liberated from the double-stranded reaction (step 8), suggesting that gpII may remain bound to
circle by a reversal of step 5: the 5 0 -end transiently bound the dso throughout DNA synthesis (Dotto et al., 1984).

FEMS Microbiol Rev 30 (2006) 321–381

c2006 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 326 C. Weigel & H. Seitz

Similar to other initiation systems, however, one or more in addition to host replisomal proteins (COM section 3.2.)
binding sites for gpII could be occupied throughout the (Liu et al., 1993). It is not known which segment of the fP2
replication cycle without allowing the formation of complex genome serves as ‘single-strand origin’, and also the mole-
II. This would localise the displaced single strand, via its cular mechanism of primosome recruitment is presently not
linkage to gpII, close to the position required for the reversal known.
of the strand-transfer reaction (step 5). Step 9. When the replisome reaches the initial nick-site, the
The molecular mechanism responsible for the balanced completely displaced single strand is liberated from the
synthesis of ffd RF and (1)-strand is not known precisely double-stranded circle by a reversal of step 5: the 5 0 -end
but gpX may be involved. During in vitro replication of transiently bound to Y454 of A is transferred back to the free
fX174, the C protein is involved in DNA packaging and, by 3 0 -OH end; the second ‘active tyrosine’ Y450 is apparently
binding to the initiation complex formed by A protein, instrumental in binding of the free 5 0 -end (Odegrip &
promotes multiple rounds of (1)-strand synthesis while Haggård-Liungquist, 2001). It is not known how A protein
preventing the accumulation of RF (Aoyama & Hayashi, is kept ‘in place’ to perform the strand-transfer reaction.
1986). After gap sealing by the host DNA ligase and adjustment of
negative superhelicity by gyrase, the closed circular dsDNA
Phage P2: As a second example for RCR, we discuss the may undergo a new round of replication starting with step 2
replication of E. coli phage P2 (Fig. 2): or serve as substrate for DNA packaging into phage heads. It
Step 1. fP2 DNA enters the host cell as linear dsDNA with is not known precisely which molecular mechanism triggers
19 bp complementary 5 0 -overhangs (cos). Following intra- the choice between ongoing replication and packaging, but
molecular circularisation, the gaps are sealed by the host it may be the availability of packaging proteins.
DNA ligase. Step 10. After completion of ‘lagging-strand’ synthesis by the
Step 2. The conversion of the circular fP2 DNA into the host replisome, the resulting double-stranded progeny mo-
replication-proficient form is completed upon introduction lecule is processed by PolA, DNA ligase and gyrase. The
of negative supercoils by the host gyrase. Binding of the closed circular dsDNA may undergo a new round of
phage initiator A to dsDNA in vitro has not yet been replication starting with step 2 or serve as substrate for
demonstrated, and although it is clear that no other fP2 DNA packaging into phage heads.
protein is involved, additional host factor(s) that could Step 11. The fP2 terminase consists of the P and M
support unwinding have not yet been identified (Liu & subunits. M was proposed to contribute the endonuclease
Haggård-Liungquist, 1994). activity required for the linearisation of the circular replica-
Step 3. The initiator protein A binds to the partially single- tion intermediates at the cos-sites during packaging of the
stranded nick-site, located within the fP2 ori (COM section phage DNA (Linderoth et al., 1991).
3.1.).
Steps 4 and 5. The nicking reaction is performed by an Although not all steps are yet known in necessary detail,
appropriately positioned A monomer (COM section fP2 replication demonstrates that (1) RCR is not confined
C3.1.1.). Nicking occurs simultaneously with the (transient) to replicons with small genomes (o 10 kb), and (2) RCR is
covalent linkage of the (1)-strand 5 0 -end to a specific easily adopted for the replication of ss and dsDNA genomes.
tyrosine residue of A, most likely Y454 (Odegrip & A highly specific feature of phage replication in the ‘rolling
Haggård-Liungquist, 2001). Only for clarity, this reaction is circle’ mode is the involvement of the Rep helicase during
shown as two separate steps in Fig. 2. strand-displacement synthesis. Also plasmid propagation by
Step 6. The unwound region serves as entry site for the host RCR depends on Rep helicase – or PcrA, its homologue in
Rep helicase and, subsequently, host replisomal proteins. Gram-positive bacteria (Petit et al., 1998). Rep and PcrA
Step 7. Strand-displacement synthesis starting from the free belong to the superfamily I helicases and are involved in
3 0 -OH end is performed by host DNA Pol III holoenzyme. recombination processes rather than in chromosome repli-
Step 8. The displaced single strand serves as lagging-strand cation of their hosts (COM section C3.3.) (Petit & Ehrlich,
template already during ongoing strand-displacement 2002). In fact, the inability of plasmids or phages to replicate
synthesis on the double-stranded phage DNA. Haggård- in a rep/pcrA mutant host may be taken as an indication that
Liungquist and co-workers were able to show that single- these replicons propagate via RCR.
strand DNA (ssDNA) replication intermediates of fP2
minichromosomes are converted to dsDNA solely by host
proteins, i.e. DnaB helicase and its helicase loader DnaC, Initiation by melting: theta (y)-type DNA
DnaG primase and DNA Pol III holoenzyme. By contrast, P2 replication
phage ssDNA intermediates require the fP2 B protein as Replicons that propagate by the theta (y)-mode of DNA
helicase loader, and probably other phage-encoded proteins replication (yDR) include bacterial chromosomes, plasmids

c 2006 Federation of European Microbiological Societies FEMS Microbiol Rev 30 (2006) 321–381
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 Bacteriophage replication modules 327

and phages. In the known cases, a specialised protein (the functions during yDR: the initiator RepC is responsible for
initiator) binds to its recognition site(s) adjacent to an AT- oriV unwinding, and RepA is the cognate helicase whose
rich region within a replication origin. The nucleoprotein action is followed by the RepB’ primase. Following these
complex formed by the initiator protein and the origin DNA initiation steps, the host DNA Pol III holoenzyme synthe-
results in ‘melting’, i.e. partial unwinding of the AT-rich sises the progeny molecules, probably by strand displace-
region. The unwound region serves as entry site for the ment (Rawlings & Tietze, 2001).
primosomal proteins (helicase loader/helicase1primase) A number of y-replicating plasmids use dual initiators, and
and, subsequently, the replisomal proteins (DNA polymer- their replication origins contain DnaA binding sites in addi-
ase1accessory proteins). The initiator performs the func- tion to binding sites for the cognate initiator. In these systems,
tion of a primosomal protein by recruiting the helicase to DnaA is either used in support for the unwinding step
the unwound region. Because the primosomal proteins, (pSC101 (Datta et al., 1999), F plasmid (Kawasaki et al.,
helicase1primase in particular, promote the assembly of 1996), fP1 prophage plasmid (Park & Chattoraj, 2001), R6K
the replisome, the initiator protein is also often called Lu et al., 1998), for the recruitment of the replicative helicase
‘replisome organiser’. We use the term ‘yDR’ for initiation DnaB, or for both functions (RK2/RP4, Konieczny et al.,
by a specialised initiator protein, and use the term ‘tDR’ for 1997; reviewed in Messer, 2002). The intricate host-depen-
initiation by transcription, which also involves ‘DNA melt- dent interplay of the TrfA initiator, DnaA and DnaB for
ing’ for primosome and replisome assembly on the (locally) replication of RK2 is discussed in more detail in the ‘Evolu-
single-stranded template (see ‘Initiation of DNA replication tionary considerations’ section. A particularly intriguing
by transcription’ section). finding was the observation that a mutation in the repA
The chromosome of E. coli replicates in the y-mode initiator gene of the Pseudomonas sp. plasmid pPS10 resulted
(Cairns, 1963). It contains a unique replication origin, oriC in a protein that extended the host range of the plasmid,
(Meijer et al., 1979; Sugimoto et al., 1979), and DnaA is the allowing its replication in E. coli through interaction of RepA
initiator (Kohiyama et al., 1966; Chakraborty et al., 1982; with DnaA (Giraldo & Fernandez-Tresguerres, 2004).
Fuller et al., 1984). DnaA is responsible for origin ‘melting’ Contrary to the expectation of Campbell & Botstein
(Fuller et al., 1984; Roth & Messer, 1995; Krause et al., 1997) (1983), phages that encode DnaA homologues have not yet
and directs the replicative helicase DnaB to the unwound been found, and also, contrary to the mentioned plasmid
region (Marszalek & Kaguni, 1994; Weigel & Seitz, 2002). systems, phage replicons that propagate via yDR and engage
Subsequently, DnaB recruits the DnaG primase, and both the DnaA protein of their host for replication are not
proteins together promote the assembly of the DNA Pol III known. An exception is the regulation of pR-mediated
holoenzyme for bidirectional coupled leading- and lagging- ‘transcriptional activation’ of l replication by DnaA, but in
strand DNA synthesis (Fuller et al., 1981; Kaguni et al., 1982; this case DnaA acts as transcription factor rather than as
Kaguni & Kornberg, 1984). Present research efforts concen- primosomal protein (Glinkowska et al., 2003).
trate on a better understanding of the regulation of DnaA
activity in the cell cycle (Speck & Messer, 2001; Katayama, Phage l: Escherichia coli phage l was the first phage
2001; Su’etsugu et al., 2004) and of the molecular details of replicon for which replication in the y-mode could be
the multiple protein–protein (Zechner et al., 1992; Weigel demonstrated in all details (reviewed in Taylor & Wegrzyn,
et al., 1999; Chang & Marians, 2000; Seitz et al., 2000) and 1995) (Fig. 3):
protein–DNA interactions (Fujikawa et al., 2003; McGarry Step 1. l DNA enters the host cell as linear dsDNA with
et al., 2004). Most sequenced bacterial chromosomes con- 12 bp complementary 5 0 -overhangs (cos). Following intra-
tain detectable oriC structures (Mackiewicz et al., 2004) and molecular circularisation, the gaps are sealed by host DNA
encode dnaA gene(s), and we may assume that DnaA/oriC- ligase.
dependent yDR is the ‘normal’ route for chromosome Step 2. The conversion of the linear l DNA into the
replication in bacteria (Messer, 2002). Initiation of replication-proficient form is completed upon introduction
chromosome replication has not been studied in detail in of negative supercoils by the host gyrase. The phage initiator
the (very) few bacterial species that either lack a dnaA O binds to the replication origin (oril) in the linear or
homologue, e.g. Wigglesworthia glossinidia, or where disrup- (relaxed) circular forms but origin unwinding requires the
tion of the dnaA gene does not produce a phenotype negatively supercoiled form (Schnos et al., 1988).
(Richter et al., 1998). Step 3. oril is located in the middle of the O gene (COM
Studies of plasmid and phage replication revealed that the section C2.2.). O protein bound to its binding sites on a
‘ABC-pathway’ of E. coli is just one possibility for yDR. For supercoiled substrate in vitro induces a conformation of the
example, the broad host-range plasmid RSF1010 (IncQ) nucleoprotein complex that results in origin unwinding.
encodes a set of replication proteins that are entirely Origin unwinding in vivo, however, requires ‘transcriptional
unrelated to the E. coli proteins but perform analogous activation’, i.e. transcription driven by the pR promoter

FEMS Microbiol Rev 30 (2006) 321–381

c2006 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 328 C. Weigel & H. Seitz

Fig. 3. The mechanism of replication in the

theta (y)-mode for circular DNA: phage l.
Successive steps are indicated by open arrows/
arrowheads plus numbering (see text for de-
tails). Complementary DNA strands are shown
as parallel lines; twisted lines indicate (negative)
supercoiling of the DNA molecule. Dark blue/
green, parental DNA strands; light blue/green,
daughter DNA strands; yellow, RNA primers.
Red dots indicate free 3 0 -hydroxyls used for
DNA synthesis. Phage-encoded proteins are
shown as coloured circles, host factors as co-
loured triangles. The colour code used for
individual proteins corresponds to the coloured
protein names in the text.

upstream of the O gene and extending to a region downstream Step 8. oril-bound O protein recruits a second P  DnaB
of oril (COM sections C2.2. 1 C3.1.2.) (Hase et al., 1989). complex to the replication bubble, but now with the opposite
Step 4. l P forms a tight 3:6 complex with the host replicative orientation. As before, DnaB helicase is released from the
helicase DnaB. The P  DnaB complex is recruited to the O  P  DnaB complex by the action of chaperones, and DnaB
unwound region by direct interaction of P with oril-bound recruits DnaG primase for priming of DNA synthesis.
O. The initial binding of P  DnaB to the unwound region Step 9. Two replication forks are engaged in coupled
involves the cryptic ssDNA-binding property of P. DnaB is leading- and lagging-strand synthesis, resulting in bidirec-
liberated from the P  DnaB complex by the action of the host tional replication away from the origin. Replication inter-
chaperones DnaJ and DnaK (COM section C3.2.). mediates at this step appear as structures resembling the
Step 5. DnaB helicase action widens the single-stranded Greek letter theta (y) with (mostly) double-stranded loops
region. DnaB recruits the host DnaG primase for priming of in the electron microscope, thus the name. It has been
leading-strand synthesis. shown that O monomers are removed from oril by the
Step 6. The DnaB  DnaG primosome recruits host DNA Pol action of ClpX/ClpP protease (Zylicz et al., 1998), but some
III holoenzyme for leading-strand synthesis. Strand-switch- may remain bound to binding sites in oril throughout the
ing of DnaG results in priming for lagging-strand synthesis. replication cycle (for clarity, binding of O to sites in oril
Step 7. Unidirectional coupled leading- and lagging-strand is not shown for this step in Fig. 3).
synthesis by the host replisome. Early after infection, l Step 10. After completion of DNA synthesis the RNA primers
replication proceeds with step 8. At later stages, prior to the are removed by the 5 0 ! 3 0 -exonuclease activity of PolA, the
switch to sDR (see below), unidirectional replication be- gaps simultaneously filled by the DNA polymerase activity of
comes prevalent, probably by a combination of (1) cessation PolA, the gaps sealed by DNA ligase and negative superhelicity
of transcriptional activation from the pR promoter, and (2) introduced by gyrase. Dimer resolution is performed by host
depletion of host DnaB. topoisomerase IV (Espeli & Marians, 2004).

c 2006 Federation of European Microbiological Societies FEMS Microbiol Rev 30 (2006) 321–381
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 Bacteriophage replication modules 329

Fig. 4. Changing the replication mode: switch

from y type to s-type DNA replication. Dark colour,
parental strands; light colour, daughter strands. Dark
blue, leading-strand template; dark green, lagging-
strand template; light green, leading strand; light
blue, lagging strand with Okazaki fragments indi-
cated by dashed line. RNA primers synthesised by
DnaG primase are shown in yellow. For clarity, single-
strand binding protein, helicase, primase, polymer-
ase and single-strand annealing protein (SAP) are not
shown. Instead, filled red arrows indicate site(s) of
enzymatic action(s). Red dots indicate free 3 0 -hydro-
xyls used for DNA synthesis. Subsequent steps are
indicated by open arrows.

Step 11. Both progeny molecules may resume yDR starting molecules. Takahashi found s-like structures, i.e. head-to-
from step 2 or switch to the s-mode of replication (see tail genome concatemers protruding from a circular parent
below). The packaging of l DNA into phage (pro)heads molecule, by electron microscopy of l replication inter-
requires genome concatemers for processing, at the cos-sites, mediates (Takahashi, 1975). Interestingly, s-structures but
by the small subunit of l terminase; the closed-circular no replication intermediates typical for yDR have been
monomeric progeny molecules from step 11 are not appro- observed in fSPP1-infected B. subtilis cells, although initia-
priate substrates (Collins & Hohn, 1978; Feiss et al., 1985; tion of fSPP1 replication occurs by the unidirectional
Smith & Feiss, 1993; Sippy & Feiss, 2004). y-mode (Missich et al., 1997). In addition, mutations in
The O (initiator) and P (helicase loader) proteins are the the fSPP1 genes 34.1 (exonuclease) and gene 35 (SAP) have
only phage-encoded proteins required for l replication in been shown to result in a replication arrest phenotype
the y-mode; all other replication factors and accessory (Burger & Trautner, 1978; Weise et al., 1994). For propaga-
factors, gyrase, DNA ligase, PolA and Topo IV, are recruited tion of fSPP1, the switch from yDR to sDR seems therefore
from the host. The general outline of the ‘replication to depend on cognate recombination functions.
scheme’ for l is virtually identical for the large family Recombination steps that could lead to a switch from
‘lambdoid’ phages (discussed in detail in ‘Bacteriophage yDR to sDR are shown schematically for a circular model
replication modules’ section). Note that here we use the replicon in Fig. 4. The switch is initiated by an interruption
fuzzy term ‘lambdoid’ exclusively to characterise phages of replication fork progression. Theoretically, it does not
with replication modules showing similarity to the l repli- matter whether a progressing fork encounters a strand
cation module. break, or whether a fork is halted by a ‘road-block’, i.e. a
nucleoprotein complex. Next, a scission is set in the fork
Phage SPP1: Recombination is a prerequisite for the region by an endonuclease, thus creating a double-strand
propagation of fSPP1, and could play a role for l and other break (DSB). As shown in Fig. 4, the lagging strand is
phages that replicate via yDR (see above). Hints come from partially degraded by the action of a 5 0 ! 3 0 exonuclease.
the following observations: (1) the linear dsDNA genomes The exposed 3 0 -OH end of the lagging-strand template is
of lambdoid phage circularise upon entry into the host cell, covered by a single-strand annealing protein (SAP) and
(2) linear head-to-tail genome concatemers are required for annealed to the leading-strand template, thus displacing
the packaging of replicated dsDNA into phage capsids parts of the leading strand. Depending on the size of the
(Taylor & Wegrzyn, 1995), and (3) the lambdoid phages single-stranded gap in the leading-strand template, the
replicate via the y-mode, which leads to circular progeny annealing reaction can be performed either by host RecA as

FEMS Microbiol Rev 30 (2006) 321–381

c 2006 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 330 C. Weigel & H. Seitz

strand invasion, or by the phage-encoded SAP as strand of unidirectional replication, the 5 0 -end of the leading
annealing (Stahl et al., 1997). The resumption of replication strand is displaced by the arriving replication fork (Dodson
results in the complete displacement of the first leading et al., 1986). However, also a DSB in one circular genome
strand. The successive displacement of the linearised lag- could be trimmed to obtain the linear substrate required for
ging-strand template from the circular leading-strand tem- invasion of another circular genome in trans mediated by
plate circle gives the resulting structure, the typical s shape host recombination proteins. Although l Rap may be
observed in the electron microscope. Ongoing coupled instrumental for creating a DSB, the major source of
leading- and lagging-strand DNA synthesis produces the progeny molecules with a DSB are probably those that were
genome concatemers required for DNA packaging into cut at the l cos site but failed to be packaged into virion
virion capsids. capsids because they were not part of concatemers (Stahl
Only parts of this model are supported by experimental et al., 1985).
evidence. Proteins with 5 0 ! 3 0 exonuclease activity and In comparison with fSPP1, the dependence on the
with single-strand annealing activity are encoded by l, cognate recombination proteins seems to be less strict for l
fSPP1 and by the Rac prophage in the E. coli K12 genome, replication. As a consequence, the presence of recombina-
and have been studied to detail (COM section C3.6.2.). tion genes in the genomes of other lambdoid phages can be
Experimental evidence for a phage-encoded endonuclease taken as an indication but not as proof of their (essential)
responsible for the initial scission in one of the parental role during replication. In each case, only experiments can
strands of the circular replicon is elusive (Taylor & Wegrzyn, help to find decisive answers.
1995). More recent results indicate, however, that replica- A noteworthy difference among the ‘lambdoid’ phages
tion fork arrest may lead to fork regression and thus the exists with respect to the form in which the (linear) phage
formation of a Holliday junction by annealing of the newly genomes are delivered into the host cell and packaged to
synthesised strands (Seigneur et al., 1998; McGlynn & Lloyd, phage heads, following replication. All l monomers end
2000). Accordingly, (pro)phage-encoded Holliday junction with identical 12 bp 5 0 -overhangs (cos-sites) generated by l
resolvases like RusA or l Rap could be the long-sought terminase, and are thus packaged from identical sites along a
candidate nucleases responsible for the initial scission concatemer. By contrast, the ‘headful packaging’ mechanism
(COM section C3.6.2.). of fSPP1 driven by the hetero-oligomeric G1P  G2P (1 : 10)
In Fig. 4, we show the cut in the parental lagging-strand terminase generates, using pac-site(s), a heterogeneous
template because a cut in the leading-strand template would population of terminally redundant and partially circularly
create an unusually orientated structure: leading-strand permuted DNA molecules with 2 bp overhangs (Chai et al.,
synthesis would use the linear ‘tail’ template, and the 1992). The individual steps of replication of both phages are
synthesis of Okazaki fragments would be directed by the virtually identical and not influenced by this ‘logistic’
circular template. To our knowledge, such structures have difference, however.
never been found in experiments. It is completely unknown
whether and how the scission in one of the parental strands Phage N15: As mentioned above, the ‘melting’ step during
is specifically directed to the parental lagging-strand tem- yDR and RCR requires: (1) interaction among origin-bound
plate, in order to create the known s-structure. We show initiator protomers and (2) negative superhelicity of the
the displacement of the leading strand by the lagging- origin DNA. All established in vitro assays for the ‘melting’
strand template after exonucleolytic resection of the lagging step use, besides the purified initiator, closed-circular DNA
strand. Whether this kind of ‘self-invasion’ in cis occurs carrying the replication origin under study (COM section
during phage replication in vivo is uncertain, but it demon- C3.1.). For practical reasons, closed-circular DNA is purified
strates that a second molecule providing the invading strand from cells: it is negatively supercoiled, and the degree of
is not necessary for a switch from yDR to sDR – at least in superhelicity is easily controlled. Hence it has become a
theory. general notion that RCR and yDR are initiation mechanisms
Although l encodes a pair of exonulease/SAP recombina- for circular DNA replicons. The replication of the linear
tion proteins (Reda/Redb) like fSPP1, their contribution to prophage of E. coli phage N15 demonstrates that it is
the switch from yDR to sDR remains controversial. Bidir- negative superhelicity and not circularity that is important
ectional replication is altered to unidirectional prior to the for initiation in the y-mode (reviewed in Ravin, 2003;
switch, probably as response to (1) a cessation of DnaA- Fig. 5):
dependent transcriptional activation (Baranska et al., 2001), Steps 1 and 2. Like l DNA, fN15 DNA enters the host cell as
and (2) a decrease in available DnaB helicase. Zylicz et al. linear dsDNA with 5 0 -overhangs (cos). Following intramo-
(1998) showed that a decrease in availability of the host lecular circularisation, the gaps are sealed by host DNA
ClpX/ClpP protease promotes unidirectional replication. ligase. Host gyrase action provides the negative superhelicity
Echols and co-workers proposed that, following one round required for origin unwinding.

c 2006 Federation of European Microbiological Societies FEMS Microbiol Rev 30 (2006) 321–381
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 Bacteriophage replication modules 331

Fig. 5. The mechanism of replication in the

theta (y) mode for linear DNA: phage N15.
Successive steps are indicated by open arrows/
arrowheads plus numbering (see text for de-
tails). Complementary DNA strands are shown
as parallel lines; twisted lines indicate (negative)
supercoiling of the DNA molecule. Dark blue/
green, parental DNA strands; light blue/green,
daughter DNA strands; yellow, RNA primers.
Red dots indicate free 3 0 -hydroxyls used for
DNA synthesis. Phage-encoded proteins are
shown as coloured circles, host factors as co-
loured triangles. The colour code used for
individual proteins corresponds to the coloured
protein names in the text. The recogniciton
sites for the TelN protelomerase are indicated as
telLR, telL, telLL, telR, telRR and telRL (see text
for details).

Step 3. Initiation is performed by the multifunctional RepA at this step, a Y-structure with telL hairpin ends is formed
protein at the replication origin residing in the 3 0 part of the (step 8a). Otherwise, replication ends with the formation of
repA gene. a circular head-to-head dimer.
Step 4. Origin unwinding and priming is performed by Step 7. The head-to-head dimer is processed by PolA, DNA
RepA, in analogy to the action of a protein during initiation ligase and gyrase for completion of the DNA synthesis, just
of replication of fP4 (COM section C3.1.2.) (Mardanov as in the case of l replication (Fig. 3; step 10).
et al., 2004). In addition, the homologous RepA initiator of Step 8a. Successive cleaving-joining by TelN at telLL and
fPY54 has very recently been shown to possess primase and telRR generates two linear progeny molecules with hairpin
helicase activity (Ziegelin et al., 2005). The phage-encoded ends. This form of the fN15 prophage is fully competent to
protelomerase TelN recognises the palindromic telLR se- enter new rounds of replication beginning with step 4.
quence as a dimer and, by a cleaving-joining reaction, Step 8b. Co-ordinated cleaving-joining by TelN at telLL and
generates the telL and telR ‘hairpin’ ends (Deneke et al., telRR generates two circular progeny molecules with telRL
2000, 2002). The covalently closed ends preserve the nega- and telRL sites, respectively. This form of fN15 follows the
tive superhelicity in the linearised molecule. If TelN activity ‘l-scheme’ for yDR and the subsequent switch to sDR,
is suppressed at this step, fN15 replication follows the ‘l- which leads to the formation of genome concatemers that
scheme’ for yDR and the subsequent switch to sDR, which are processed by the fN15 packaging apparatus.
leads to the formation of genome concatemers that are
processed by the fN15 packaging apparatus (Ravin, 2003). Efficient propagation of linear minichromosomes carrying
Step 5. Replication of fN15 proceeds bidirectionally and the repA gene together with the telN gene and a telLR site has
with coupled leading- and lagging-strand synthesis (Ravin been demonstrated (Ravin et al., 2001). It is presently not
et al., 2003). known, however, by which mechanism the activity of TelN is
Step 6. Replication encompassing the telL site (re)generates suppressed to allow the linear prophage to enter yDR with
a palindromic telLL site that is substrate for cleaving-joining circular molecules to provide – after the switch to sDR – the
by TelN. If TelN already acts on the replication intermediate genome concatemers for DNA packaging during the lytic cycle.

FEMS Microbiol Rev 30 (2006) 321–381

c 2006 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 332 C. Weigel & H. Seitz

Fig. 6. The mechanism of protein-primed DNA

replication (ppDR): f29. Successive steps are
indicated by open arrows/arrowheads plus
numbering (see text for details). Complemen-
tary DNA strands are shown as parallel lines.
Dark blue/green, parental DNA strands; light
blue/green, daughter DNA strands. Red dots
indicate free 3 0 -hydroxyls used for DNA synth-
esis. The colour code used for individual pro-
teins (coloured circles) corresponds to the
coloured protein names in the text.

Initiation at the ends of linear DNA: protein- Examples for linear, dsDNA replicons with proteins
primed DNA replication covalently attached to both ends include pro- and eukar-
yotic viruses, e.g. Adenoviruses, and eukaryotic plasmids,
e.g. the Kalilo plasmid(s) of Neurospora. The linear chromo-
All known (replicative) DNA polymerases require a ‘primer’
somes of Streptomyces species and of several linear Strepto-
– a free 3 0 -hydroxyl group provided by the strand comple-
myces plasmids contain terminal proteins; these replicons
mentary to the template strand – because they cannot start
initiate DNA synthesis at internal origins, and the terminal
de novo DNA synthesis. DNA polymerases can elongate
proteins are part of a special mechanism allowing the full
either the 3 0 -OH end of nicked dsDNA as in RCR (see
replication of the partially single-stranded telomers (Bao &
‘Initiation by nicking: ‘rolling circle’-type DNA replication’
Cohen, 2003). Note that the linear chromosome and several
section) or the 3 0 -OH end of a short RNA primer synthe-
linear plasmids of Borrelia burgdorferi do not contain
sised by a specialised RNA polymerase, a primase, as in yDR
terminal proteins but possess hairpin ends like E. coli phage
(see ‘Initiation by melting: theta (y)-type DNA replication’
fN15 (see above), and also a cognate protelomerase has
section). In the forthcoming section, we will discuss ‘prim-
been identified (Deneke et al., 2004; Huang et al., 2004).
ing’ by transcripts that remain bound to their templates, and
in the subsequent section, ‘priming’ by the 3 0 -end of ssDNA Phage 29: The initiation of replication of B. subtilis phage
invading a duplex. In addition, the CCA-3 0 stem of (un- f29 occurs exclusively at the phage ends, and depends on
charged) tRNA can serve as ‘primer’, as was found for the terminally attached protein. f29 is therefore the ‘model
retrovirus replication and has been proposed for the replica- of choice’ for studying replication by ‘melting’ of the
tion of Cauliflower Mosaic Virus (CaMV) (Pfeiffer & Hohn, terminal hydrogen bonds of linear dsDNA molecules (me-
1983). Finally, a sterically favourably positioned hydroxyl- chanism 3; see above). More importantly, ppDR has been
group of Ser, Thr, or Tyr residues in a protein can serve as studied in great detail for f29 (reviewed in (Meijer et al.,
‘primer’, which led to the terminus ‘protein-primed DNA 2001), and the individual steps are discussed in the follow-
replication’ (ppDR). In all known cases, this ‘portable ing (Fig. 6):
primer’ protein remains attached to the 5 0 -end of the newly Step 1. The linear, double-stranded phage DNA enters the
synthesised DNA strand, and is called ‘terminal protein’ host cell with p3 TP covalently bound to the 5 0 -terminal
(TP) by convention. bases. To discriminate between TP bound to genomes that

c 2006 Federation of European Microbiological Societies FEMS Microbiol Rev 30 (2006) 321–381
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 Bacteriophage replication modules 333

have already been replicated and free, newly synthesised TP, ‘f29 replication factories’, i.e. stationary protein complexes
the former is called ‘parental’ TP. through which the DNA is threaded during replication.
Step 2. The f29 p6 double-strand binding protein has been
shown to form oligomers (Abril et al., 1997, 1999), and to
bind preferentially to sites with intrinsic DNA curvature; Initiation of DNA replication by transcription
two such sites are located at a distance of 46–62 bp from the The mechanism of initiation of DNA replication by tran-
left end of the phage genome, and 68–125 bp from the right scription (tDR) has been studied in four experimental
end, respectively (Serrano et al., 1989). Nucleoprotein com- systems representative for all three types of prokaryotic
plexes formed by p6 heavily distort the bound DNA, and replicons: constitutively stable DNA replication (cSDR) of
this distortion has been suggested to be responsible for helix the E. coli chromosome, replication of ColE1-type plasmids,
destabilisation at the phage genome ends (Serrano et al., ‘early’ replication of E. coli phage T4 and replication of E.
1993). coli phage T7. The basic features of tDR are remarkably
Step 3. The penultimate T residue of f29 DNA serves as similar in all four systems. Briefly, RNA polymerase binds to
template for dAMP incorporation to TP catalysed by p2 a promoter on dsDNA and synthesises a short (untrans-
DNA polymerase (Méndez et al., 1992) (sequence shown in lated) transcript that remains attached to its template. The
Fig. 6). f29 p2 DNA polymerase binds newly synthesised TP transcript provides the 3 0 -OH end used by DNA polymerase
in solution and performs the deoxyadenylation reaction in for displacement synthesis of a leading strand. Formally,
vitro without a DNA template, albeit with higher selectivity RNA polymerase performs the triple function of: (1) an
for dATP and more efficiently in the presence of parental p3 initiator (DNA melting), (2) a helicase (DNA unwinding)
TP bound to f29 DNA (Blanco et al., 1992). Binding of the and (3) a primase (providing the 3 0 -OH end of an RNA
p2  p3 complex to f29 ends is activated, in addition, by the primer). Strand-displacement synthesis switches to unidir-
correctly positioned p6 nucleoprotein complex (Freire et al., ectional, coupled leading- and lagging-strand synthesis
1996). upon assembly of a primosome (primase/helicase) on the
Step 4. p2 catalyses the deoxyadenylation of Ser232 in p3 TP displaced strand (R-loop). Lagging-strand synthesis on
(Hermoso et al., 1985). Following this reaction, the the displaced strand further widens the loop allowing
p2  p3  dAMP complex slides back 1 nt, and p2 starts DNA primosome assembly on the opposite strand. The net
synthesis at position 12 (Méndez et al., 1992). result of these reactions is bidirectional replication of the
Step 5. The ‘elongation phase’ of f29 replication starts with template.
the dissociation of p3 TP and p2 DNA polymerase after In ColE1-type plasmids the RNAII transcript assumes a
template-directed synthesis of 6–9 nt (Méndez et al., 1997). complex tertiary structure upon trimming by RNase H, and
Covering of the ssDNA by p5 SSB protects the displaced its elongation is performed by PolA. Accordingly, ColE1-
strand from nuclease digestion. In addition, p5 prevents type plasmids (e.g. pBR322) cannot be propagated in polA
hairpin formation that might slow down DNA synthesis by mutants of E. coli. A primosome-assembly site (PAS) that
p2, prevents template-switching of p2 and supports p2 becomes single-stranded during PolA-driven displacement
processivity by helix destabilisation (Martı́n et al., 1989; synthesis serves as entry site for the restart primosome
Soengas et al., 1995; Esteban et al., 1997). (PriA, PriB, PriC, DnaT) (see ‘Replication restart’ section).
Steps 6 and 7. During the elongation phase, p2 displaces the Subsequently, the primosome recruits the replicative heli-
p6 nucleoprotein complex. DNA synthesis is initiated at case, DnaB, and the primase, DnaG, and bidirectional
both ends of the f29 genome, and the two parental strands replication is performed by DNA Pol III holoenzyme
become separated when the replication forks pass each (reviewed in del Solar et al., 1998). To be operative in E. coli,
other. Both p2 DNA polymerases continue with DNA the mechanism of cSDR requires inactivation of the rnhA
synthesis until the end of the single-stranded template. Each gene encoding RNase H, resulting in a longer half-life of
progeny molecule contains p3 TP bound to its 5 0 -ends. various transcripts (reviewed in Kogoma, 1997). cSDR is not
possible in priA or polA mutant strains, indicating that both
Salas and colleagues point to the intriguing observation proteins perform essential functions, similar to their func-
that two additional phage-encoded proteins participate in tions for ColE1 replication. cSDR can sustain chromosome
f29 replication in vivo: p1 and p16.7 (reviewed in Bravo replication in E. coli dnaA or oriC-deletion mutants, show-
et al., 2005). The f29 replisome could be targeted to a ing that cSDR bypasses the ‘normal’ pathway of initiation of
membrane-associated p1 multimeric structure by interac- chromosome replication (Messer, 2002).
tion between p1 and primed TP. The integral membrane
protein p16.7 is thought to recruit the f29 DNA replisome Phage T7: The replication pathways of phage T4 will be
through interaction with both the parental TP and the addressed in the following section. Here we discuss tDR of
ssDNA. Both proteins can thus be envisaged as parts of fT7, which can be divided into the following steps (Fig. 7):

FEMS Microbiol Rev 30 (2006) 321–381

c 2006 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 334 C. Weigel & H. Seitz

Fig. 7. The mechanisms of replication initiated

by transcription (tDR): phage T7. Successive
steps are indicated by open arrows/arrowheads
plus numbering (see text for details). Comple-
mentary DNA strands are shown as parallel
lines. Dark blue/green, parental DNA strands;
light blue/green, daughter DNA strands; yellow,
RNA primers. Red dots indicate free 3 0 -hydro-
xyls used for DNA synthesis. Phage-encoded
proteins are shown as coloured circles, host
factors as coloured triangles. The colour code
used for individual proteins corresponds to the
coloured protein names in the text. The 160 bp-
long direct terminal repeats in fT7 DNA are
indicated by white arrows.

Step 1. Entry of the linear dsDNA of T7 into the host cell Step 4. Although it would be reasonable to assume
starts with the ‘left end’, and host RNA polymerase tran- that gene 2.5 SSB could cover the replication ‘bubble’
scribes the ‘early’ genes, including gene 1 encoding T7 RNA also during earlier steps, Fuller & Richardson (1985a)
polymerase. Gene 1 RNA polymerase is responsible for found a measurable positive influence of gene 2.5
transcription of the ‘primary origin’ (Saito et al., 1980; protein only for the priming of bidirectional DNA synthesis
Fuller & Richardson, 1985b). Deletion of the primary origin in vitro.
results in initiation of replication from other T7 RNA Step 5. Bidirectional coupled leading- and lagging-strand
polymerase promoters in the genome (Tamanoi et al., 1980; synthesis results in Y-shaped replication intermediates ob-
Wever et al., 1980). served in the electron microscope at some time after
Step 2. The transcript is elongated through displacement initiation (Dressler et al., 1972).
synthesis by gene 5 DNA polymerase; priming of lagging- Step 6. Enzymes involved in processing of the progeny
strand synthesis is performed by gene 4 primase-helicase molecules include gene 6 protein with RNase H activity,
upon binding to a gene 4 recognition site (5 0 -GGGTC) that gene 5 and gene 1.3 DNA ligase.
becomes single stranded during displacement synthesis Step 7. The replication of the linear fT7 DNA is inherently
(Fuller & Richardson, 1985b). incomplete. Owing to the presence of 160 bp long direct
Step 3. Coupled leading- and lagging-strand synthesis terminal repeats, replication intermediates with 3 0 -over-
widens the replication ‘bubble’. Such bubble structures were hangs can hybridise to each other, forming head-to-tail
instrumental in defining the replication origin by early concatemers. Covalent linkage of the concatemers is
electron microscopic studies (Dressler et al., 1972). achieved by the action of gene 1.3 DNA ligase. Concatemers

c 2006 Federation of European Microbiological Societies FEMS Microbiol Rev 30 (2006) 321–381
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 Bacteriophage replication modules 335

are subject to secondary initiation events, resulting in the DNA to a homologous strand in dsDNA in such a way that
cyclical growth of concatemer lengths. the free 3 0 -hydroxyl end of the ‘invading’ strand can serve as
Step 8. Secondary initiation events are instrumental for the primer for DNA polymerase. The annealing of the two
overall growth of the concatemeric phage DNA, but in some complementary strands during this recombination process
cases collapsing or aborted forks result in Holliday struc- is promoted by proteins with strand-annealing property
tures. In addition, partially synthesised strands resulting (SAPs), helicase(s) and SSBs. Homologous recombination
from displacement synthesis create branched structures. resulting in displacement loops (D-loops) may proceed with
Because branched DNA molecules are not appropriate further annealing – including also the complementary
substrates for packaging into phage heads, gene 3 Holliday- strand of the invading 3 0 end – and branch migration. The
junction resolvase, gene 6 5 0 ! 3 0 exonuclease activity and recombination intermediates are finally resolved by struc-
gene 1.3 DNA ligase are required for proper trimming of the ture-specific endonucleases, e.g. Holliday junction resol-
concatemeric DNA. vases. Depending on the pathway, the resulting structures
Step 9. fT7 DNA packaging into phage heads starts with the are ‘splice’ or ‘patch’ variants of ‘join-break’ recombination.
‘right end’ (Son et al., 1993). During packaging, the con- If the D-loop created by homologous recombination serves
catemeric DNA is cut by a site-specific nuclease in order to for primosome and replisome assembly, we may talk of
liberate genome monomers with single-stranded 5 0 -over- ‘join-copy’ recombination (Mosig, 1994), or RDR. The
hangs. The nuclease activity is associated with capsid formation of a D-loop as the first step to initiate DNA
proteins, and the recognition sites are known, but the synthesis classifies RDR as a ‘melting’ mechanism, by formal
responsible phage-encoded protein has not yet been unequi- criteria.
vocally identified (White & Richardson, 1987; Chung & RDR is not suitable for complete de novo replication of a
Hinkle, 1990). replicon, unless it contains tandemly repeated sequences.
Step 10. Processing of phage concatemers occurs in a way However, if (partially) duplicated replicons are present in a
that allows the restoration of the terminal repeats by a fill-in cell, RDR can serve as an efficient bypass mechanism for
reaction. This reaction requires, in addition to gene 5 DNA replication initiation, e.g. when the ‘normal’ initiation path-
polymerase, gene 6 exonuclease to prevent strand-displace- way is disabled. RDR of the E. coli chromosome was first
ment synthesis (White & Richardson, 1987; Serwer et al., detected by Lark and Kogoma (Kogoma & Lark, 1975), and
1990). Successful packaging also requires gene 2 protein as studied in great detail by Kogoma (1997) as ‘induced stable
inhibitor of host RNA polymerase (LeClerc & Richardson, DNA replication’ (iSDR). iSDR can sustain chromosome
1979). replication for several hours in the absence of protein
synthesis upon induction of the SOS-response. Because the
When the replication schemes for fT7 (Fig. 7) and l (Fig. induction of the SOS-response inhibits cell division, iSDR is,
3) are compared, it is apparent that they are virtually unlike cSDR, not a replication bypass mechanism allowing
identical for the steps following priming of lagging-strand cell proliferation. iSDR does not require DnaA, but depends
synthesis up to the end of the first round of DNA synthesis. crucially on intact recombination functions (RecA, RecBC)
The similarities and differences among the proteins respon- and PriA to form restart primosomes (see ‘Replication
sible for performing the successive enzymatic steps are restart’ section). The importance of RDR for chromosome
discussed in detail in COM section C3. tDR and yDR are replication in E. coli under normal growth conditions is still
replication mechanisms that rely on duplex melting: either a matter of debate; that RDR serves to rescue broken
by RNA polymerase or on an initiator. We have discussed chromosomes and stalled replication forks is, however,
above that yDR requires a negatively supercoiled substrate. generally accepted (Kuzminov, 1999; Cox et al., 2000;
By contrast, tDR can be initiated by RNA polymerase on a Maisnier-Patin et al., 2001) (see ‘Replication restart’
relaxed linear substrate. However, this difference may section).
not be as significant as it appears at first sight: RNA
polymerases are known to modulate the local superhelicity Escherichia coli phage Mu depends entirely on host
of their templates during transcription, and this (local) enzymes for the replication of its genome. The phage-
superhelicity has been proposed to be important for R-loop encoded, oligomeric MuA transposase complex transfers
stability (Liu & Wang, 1987; Rahmouni & Wells, 1992; fMu ends to (nonhomologous) target DNA. The MuA
Drolet et al., 1994). ‘transpososome’ creates a fork at each end, and remains
tightly bound to both forks. The host ClpX chaperone is
required for a ‘loosening’ of the DNA interaction(s) of the
Recombination-dependent DNA replication
transpososome. An as yet unidentified host factor further
The basic reaction in recombination-dependent DNA repli- displaces the transpososome and promotes the assembly of a
cation (RDR) is the annealing of a single-stranded stretch of restart primosome, which subsequently recruits the DnaBC

FEMS Microbiol Rev 30 (2006) 321–381

c2006 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 336 C. Weigel & H. Seitz

Fig. 8. The mechanism of recombination-de-

pendent DNA replication (RDR): phage T4.
Successive steps are indicated by open arrows/
arrowheads plus numbering (see text for de-
tails). Complementary DNA strands are shown
as parallel lines. Dark blue/green, parental DNA
strands; light blue/green, daughter DNA
strands; yellow, RNA primers. Red dots indicate
free 3 0 -hydroxyls used for DNA synthesis. The
colour code used for individual proteins (shown
as coloured circles) corresponds to the coloured
protein names in the text. The terminally re-
dundant sequences of the fT4 genome are
indicated by small blocks within the DNA
strands. For steps 7 and 8 the replisome formed
during tDR is shown as a grey silhouette to
avoid the diagram becoming overly compli-
cated.

helicase complex (see ‘Replication restart’ section). ‘Repli- present in Fig. 8 a scheme for fT4 replication, which
cative transposition’ of fMu thus ends up in the ‘normal’ combines the initial phase of tDR with the subsequent RDR
pathway for RDR (reviewed in Nakai et al., 2001). phase. This scheme should be regarded as ‘minimal’ in the
sense that the essential steps are included but not the
Phage T4: Among phage replicons that (unlike Mu) amazing number of known bypass mechanisms, which have
encode cognate replication proteins, RDR is best under- always made fT4 replication a topic suitable for mono-
stood for E. coli phage T4. The interdependence of recombi- graphs rather than for reviewing articles. As above, we
nation and replication was already the subject of a review by discuss individual steps (Fig. 8):
Broker & Doermann (1975), at that time mostly based on Step 1. ‘Early’ replication of fT4 is initiated by transcription
results of genetic and electron microscopic studies. Later it from one of several origins, oriA, F, G and E. These origins
was established that fT4 replication proceeds in two stages: are promoters that are specifically recognised by host RNA
the initial, rifampicin-sensitive stage that depends on (host) polymerase after replacement of the s70 subunit by the
RNA polymerase, and the second, ‘burst’ stage that is phage-encoded AsiA s-factor. In addition, promoter-recog-
suppressed in recombination mutants (Luder & Mosig, nition by the modified RNA polymerase requires phage-
1982). The genome of fT4 has a size of 168.8 kb (Miller encoded transcriptional activators: MotA in the case of oriA,
et al., 2003). Linear fT4 DNA entering the host cell has a oriF and oriG, and DbpC in the case of oriE (Mosig et al.,
size of 173 kb and is circularly permuted, i.e. it contains 1995). The transcripts synthesised by RNA polymerase
3–5 kb terminal redundancy. As Mosig pointed out, the remain attached to their template strands, thus forming an
terminal redundancy is ‘sufficiently large to allow homo- R-loop structure. A structure downstream of the promoter
logous recombination between the terminal regions of a has been shown to posses properties of a ‘DNA-unwinding
single chromosome, allowing successful infection of a host element’ (DUE) and might be required for the stability of
cell by a single T4 particle’ (Mosig et al., 1995, p. 86). We the RNA  DNA heteroduplex (Carles-Kinch & Kreuzer,

c 2006 Federation of European Microbiological Societies FEMS Microbiol Rev 30 (2006) 321–381
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 Bacteriophage replication modules 337

1997). In addition, transcript stability may require forma- H-type structures indicative for (double) branch migration
tion of complex secondary structures including partial could only be found under conditions when replication was
hybridisation to the displaced DNA strand (Mosig et al., repressed (Broker, 1973). Therefore, the switch from tDR to
1995). RDR during fT4 replication seems to be highly efficient
Step 2. Nossal et al. (2001) were able to show that the fT4 under normal conditions. Alberts and Formosa simulated
replication proteins perform DNA synthesis in vitro on an the initial step of RDR in vitro: replication could be obtained
artificial substrate mimicking the R-loop structure at ori- with the appropriate DNA substrates and purified proteins
F(uvsY) using the 3 0 -OH end of the RNA transcript as DNA polymerase ‘holoenzyme’ (gps 43, 44, 45 and 62), gp32
primer. Binding of gp43 DNA polymerase requires the SSB, Dda helicase and UvsX SAP (Formosa & Alberts, 1986).
gp45 sliding clamp for processive DNA synthesis. The Step 8. We have followed fT4 replication up to the point
homo-trimeric gp45 sliding clamp is assembled around the where strand invasion by the unreplicated ‘loose’ 3 0 -OH end
ssDNA downstream of the free 3 0 -OH end of the RNA  DNA into the terminal redundancy of the same molecules solves
heteroduplex by the gp44/gp46 clamp loader complex (to the problem to replicate fully the linear phage DNA. Gp43
avoid ‘molecular crowding’, gps 45, 44 and 46 are not shown DNA polymerase elongates the annealed 3 0 -OH end, and
in Fig. 8; for details see COM section C3.5.2.). It appears replisome formation occurs in the D-loop as above in step 2.
that efficient DNA synthesis in vitro is synchronised by the Thus, the switch from tDR to RDR is completed. The
versatile gp59 helicase loader: (1) gp59 removes gp32 SSB complex topological structure (Fig. 8, boxed area) created by
from the displaced DNA strand in the R-loop upon interac- the recombination step can be resolved through endonucleo-
tion (Ishmael et al., 2001), (2) it loads the gp41 helicase to lytic cleavage by gp49 endonuclease VII. Alternatively, addi-
the displaced strand (Venkatesan et al., 1982; Barry & tional priming of DNA synthesis may occur in the opposite
Alberts, 1994b) and (3) it can slow down polymerase activity direction because the gp59 helicase loader shows preferential
of gp43 until the helicase is completely loaded (Nossal et al., binding to branched fork structure (three-way or four-way
2001). Barry & Alberts (1994a) identified an alternative junctions) (Jones et al., 2000). During the ‘burst’ phase of
pathway for gp41 loading in vitro: in the absence of gp59, fT4 replication, steps 6–8 are repeated until the exhaustion
the Dda helicase can remove the RNA polymerase ahead of of the dNTP pools, but secondary origin-dependent initia-
the synthesizing gp43 DNA polymerase, which in turn tions for tDR occur rarely (Mosig et al., 1995). UvsW helicase,
allows recruitment of gp41 directly by gp43. expressed later during infection, may participate in suppres-
Step 3. The gp41 helicase recruits the gp61 primase for sion of origin-dependent initiations by removing the RNA
synthesis of the first lagging-strand primer (Burke et al., 1985). from the R-loop (Dudas & Kreuzer, 2001).
Steps 4 and 5. Elongation of the lagging-strand primer is
performed by a second gp43 DNA polymerase recruited to Packaging of fT4 DNA into phage heads requires genome
the forming replisome. The complete fT4 replisome is now concatemers. Therefore, the ‘network’ of interwoven recom-
composed of gp41 helicase, gp61 primase for cyclical prim- bination structures created by RDR has to be ‘trimmed’, i.e.
ing of lagging-strand synthesis (Okazaki fragments), two Holliday junctions resolved, branches created by dismissed
gp45 sliding clamps plus their gp44/gp62 clamp loaders, and replication forks eliminated and all gaps sealed. The phage-
two gp43 DNA polymerases for coupled leading- and encoded proteins mentioned above can perform all the
lagging-strand synthesis (Salinas & Benkovic, 2000; Kadyrov required functions, and make fT4 replication independent
& Drake, 2001). The first lagging strand is converted to the of host functions up to this last step, DNA packaging.
leading strand by a second replisome assembled for DNA
synthesis in the opposite direction (step 5). We wish to emphasise again that this scheme for fT4
Step 6. Replication of the linear fT4 DNA is inherently replication presents a ‘minimal version’ and only includes
incomplete. The 3 0 -end of the lagging-strand template is the recombination ‘pathway II’ believed to play the major
covered by gp32 SSB and UvsX SAP. UvsX requires UvsY as role for fT4 replication under normal growth conditions
accessory protein, but the function of UvsY is not known (Mosig, 1998). The multiple replication and recombination
exactly. Recent results suggest the UvsY supports UvsX pathways encoded by fT4 are probably the result of
loading by weakening the association of gp32 with ssDNA consecutive adaptations of the phage to a great variety of
(Bleuit et al., 2004). growth conditions, preserving its (almost complete) inde-
Step 7. Analyses by electron microscopy revealed that the pendence of host functions.
unreplicated ends of newly replicated fT4 DNA molecules
preferentially ‘invade’ the terminally redundant region at the
Replication restart
other end of the same molecule, or in other chromosomes in
the case of multiple infections (small coloured blocks at the Replication research has always been greatly influenced
chromosome ends in Fig. 8) (Dannenberg & Mosig, 1983). by the ‘replicon model’ (Jacob et al., 1963) and the

FEMS Microbiol Rev 30 (2006) 321–381

c2006 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 338 C. Weigel & H. Seitz

physiological studies of Maaloe and his collaborators, sum- PriA primosome in chromosome replication remained
marised in their statetement: ‘We are therefore led to believe elusive until Kogoma discovered that E. coli priA(null)
that the overall production of DNA, RNA, and protein is strains are defective in recombination and DSB repair and
regulated by mechanisms that control the frequencies with proposed that PriA is responsible for replisome assembly at
which the synthesis of individual nucleotide and amino acid recombination intermediates, leading to RDR (Kogoma
chains are initiated’ (Maaloe & Kjeldgaard, 1966, p. 163). et al., 1993; Masai et al., 1994).
Translation, transcription and replication have indeed been Escherichia coli PriA has detectable homologues in the
shown to be regulated primarily at the initiation step (see genomes of most bacteria across all phyla and is genetically
the previous subsections). However, replication research has and biochemically well characterised (reviewed in Sandler &
for many years neglected the fact that any premature Marians, 2000). PriA is an SF2-type helicase (COM section
abortion of DNA synthesis is not just a waste of energy but C3.3.) that can unwind DNA in 5 0 ! 3 0 and 3 0 ! 5 0
a challenge to replicon integrity, so severe that the existence direction, but the helicase activity is dispensable for its
of a salvage pathway(s) could have been anticipated. For function as primosomal protein (Zavitz & Marians, 1992).
about 10 years, the elucidation of molecular pathway(s) PriA has a marked preference for binding to branched DNA
promoting restart of DNA synthesis at stalled or dismissed structures in vitro, and binding to D-loops occurs down-
replication forks has developed into a new field bringing stream of the invading strand, albeit without strand pre-
together recombination and replication research (Cox et al., ference in the absence of SSB (Jones & Nakai, 1999; Liu &
2000; Lusetti & Cox, 2002). Replication restart is mostly Marians, 1999; Cadman & McGlynn, 2004). The results of
studied for chromosome replication, but two bacteriophage recent genetic and biochemical experiments suggest that two
replication systems have greatly influenced the present (partially overlapping) pathways exist for restart primosome
models: (1) the conversion of (1)-strand DNA of fX174 assembly: the PriA-dependent pathway involving PriA, PriB
into the replicative form by E. coli enzymes (see ‘Initiation and DnaT, and a second, PriA-independent pathway
by nicking: ‘rolling circle’-type DNA replication’ section) mediated by PriC in conjunction with the E. coli Rep
and (2) the recombination-dependent DNA replication of helicase (Sandler et al., 1999; Sandler & Marians, 2000;
fT4 (see ‘Recombination-dependent DNA replication’ sec- Heller & Marians, 2005). Earlier, Seufert and Messer de-
tion). scribed yet another pathway for replication restart: if a
The conversion of (1)-strand DNA of fX174 into the replisome encounters a block shortly downstream from
replicative form in vitro depends on the E. coli proteins N oriC, it re-initiates at PAS sites 2 kb away. The interpreta-
(PriB), N 0 (PriA or Y), N00 (PriC), I (DnaT), DnaB, DnaC, tion was that the helicase remains attached and unwinds
DnaG and DNA polymerase III holoenzyme; the conversion until a PAS site is exposed as a single-strand and used for
is completed upon removal of the RNA primer by PolA and (PriA-dependent) replisome re-assembly (Seufert & Messer,
gap sealing by DNA ligase (Schekman et al., 1975). PriA, 1986). Homologues of PriB and PriC are present in the
PriB, PriC and DnaT are required to load the replicative sequenced genomes of various Gram-negative bacteria, but
helicase as DnaB6C6 hetero-hexamer to the SSB-coated could not be detected by BLAST searches in the genomes
template. DnaC dissociates from the complex after helicase of Gram-positive bacteria (in addition, we could not
loading, and the remaining proteins are collectively called detect phage-encoded homologues of PriA, PriB or PriC).
the preprimosome. The recruitment of the DnaG primase by Homologues of DnaT, however, are only present in
DnaB converts the preprimosome into the primosome the genomes of those species that also encode homologues
(Tougu & Marians, 1996; Chang & Marians, 2000). DnaB of DnaC, and partial homologues are present in several
and DnaG recruit DNA polymerase III through multiple phage genomes that code for a DnaC-type helicase loader.
protein–protein interactions, thus forming the replisome We will discuss in the ‘Evolutionary considerations’ our
(Zechner et al., 1992; reviewed in Kornberg & Baker, 1992; hypothesis that the dnaTC gene pair was acquired by E. coli
Marians, 1996). Results from in vitro studies led to the from a replication module of an ancient lambdoid phage.
suggestion that the primosomal proteins PriA, PriB, PriC From the above, we anticipate that further variants of the
and DnaT remain in physical contact with the replisome protein composition of PriA primosomes will be revealed
during DNA synthesis (Ng & Marians, 1996). During E. coli upon analyses of replication systems in Gram-positive
chromosome replication from oriC, primosome formation bacteria.
requires DnaA, DnaB6C6 and DnaG (Messer & Weigel, 1996; During the elongation phase, replisomes may encounter
Hiasa & Marians, 1999). The terms ‘DnaA primosome’ (or two types of nonprogrammed stops that result in replication
‘ABC primosome’; Masai et al., 1990) and ‘PriA primosome’ fork stalling or collapse, and disassembly of the replisome:
(or ‘fX primosome’) reflect the differences of both primo- (1) chemically modified bases in one of the template
somes with respect to protein composition. Despite its strands, or (2) nicks in one of the two template strands or
established function for fX174 replication, a role for the DSBs. In the first case, replication fork stalling may lead to

c 2006 Federation of European Microbiological Societies FEMS Microbiol Rev 30 (2006) 321–381
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 Bacteriophage replication modules 339

fork regression and annealing of the newly synthesised Marians and colleagues have proposed to use the term
strands, i.e. formation of a Holliday junction or ‘chicken- ‘co-ordinated processing of damaged replication forks’
foot’ structure. Fork regression regenerates templates for (CPR) to account for the observation that recombination-
repair of the damage by either the nucleotide excision or dependent replication restart of chromosome replication in
base excision repair pathways. Repair can result in reversal of E. coli is an (essentially) error-free process, in contrast to
the fork regression and the fork structure itself may serve as error-prone DNA repair-synthesis during SOS induction
substrate for primosome assembly. Alternatively, supported (Sandler & Marians, 2000). To avoid ‘abbreviation overload’,
by recent experimental results, the resolution of the Holliday we prefer to use the term RDR, at least as long as no
junction by RuvC or RusA could trigger replication restart in fundamental mechanistic differences between ‘recombina-
a reaction similar to that found for restart triggered by DSB tion-dependent DNA replication’ (RDR) in fT4 and
(Seigneur et al., 1998; McGlynn & Lloyd, 2000). In the ‘recombination-dependent restart’ (RDR) of E. coli chromo-
second case, a nick in either of the template strands will lead some replication are revealed.
to fork collapse and results in a DSB (Michel et al., 1997).
Resection of the linearised arm by the 5 0 ! 3 0 exonuclease
activity of RecBCD leaves a 3 0 -OH tail that is covered by
Bacteriophage replication modules
RecA. Subsequently, RecA-mediated strand invasion in the The term ‘replication module’ is often used in recent papers
‘intact’ arm creates a D-loop (Kuzminov, 1999) that can be dealing with the architecture of bacteriophage genomes to
used as substrate for replication restart (Liu et al., 1999). account for the recurrent observation that replication genes
Because in all these cases replication restart could be shown co-localise in a distinct segment of phage genomes. In some
to depend on the PriA primosome, Sandler & Marians cases, the detection of similarities of one or more predicted
(2000) proposed its re-naming as ‘(replication) restart ORFs to particularly well-conserved proteins (e.g. helicases,
primosome’. By genetic analysis of an E. coli gyr mutant DNA polymerases) were thought sufficient to pinpoint the
strain, Ehrlich and co-workers showed that a requirement ‘replication module’ of a particular phage genome. We do
for the restart primosome also exists under conditions not reject this somewhat sloppy use of the term ‘module’
where replication restart did not involve recombination because it results in positive ‘hits’ in most cases. However,
(Grompone et al., 2003). It seems possible, therefore, that only a more precise definition of the replication module can
the ‘backup’ properties of the restart primosome are also prevent the misleading impression that the replication of a
required to face less severe impairments of replication fork given phage is understood by pinpointing its ‘replication
progression. Estimates vary but it is likely that under most module’ the sloppy way.
growth conditions a replisome starting from oriC has a Following accepted practice in molecular biology, a
15–50% chance of being inactivated before reaching a Ter definition of bacteriophage replication modules should rely
site (Maisnier-Patin et al., 2001). This again emphasises the largely on the results of genetic and biochemical studies. A
importance of the restart primosome. straightforward approach would start with phage DNA
During fT4 replication, the origin-dependent initial fragments ligated to a selectable marker, searching for
phase (tDR) is followed by the ‘burst’ phase that (almost) autonomous replicating plasmids after transformation of
entirely occurs in the RDR-mode (see previous section). It an appropriate host. Comparable strategies led to the
was shown by Kreuzer and co-workers that in vitro not only detection of ldv plasmids (Matsubara & Kaiser, 1968), of
the invading 3 0 -ends of fT4 can efficiently trigger RDR but the E. coli prophage Rac replication module (Dı́az &
also artificially introduced DSBs (George et al., 2001). There Pritchard, 1978), of the fadh replication module (Alter-
is thus a convincing similarity between RDR in fT4 and mann et al., 1999) and of the replication module of fc2-type
recombination-dependent replication restart of chromo- phages (Rakonjac et al., 2003). However, this ‘functional
some replication: with respect to the mechanism, but also approach’ is unsatisfactory at present, mainly for three
with respect to the enzymatic functions involved (see Table 1 reasons. One trivial reason is the lack of functional studies
in Cox, 2001). However, both systems differ with respect to for the vast majority of known phage replicons. Another
(1) the timely order of primosome and replisome assembly trivial reason is the implicit assumption that replication
and (2) the properties (of some) of the primosomal pro- genes occur tightly packed in a single cluster, which is the
teins. In E. coli, the assembly of the restart primosome is a case in most but not all known phage groups. The third
prerequisite for replisome assembly. By contrast, the fT4 reason becomes apparent when one looks more closely at the
gp59 helicase loader promotes the loading of the gp41 long record of research on the ldv plasmids, which were
helicase to the D-loop and slows down simultaneously discovered in 1968 (!) by Matsubara & Kaiser (1968). The
ongoing DNA synthesis by gp43, probably for efficient initially studied plasmids contained the replication origin
‘coupling’ of gps 41 and 43 in the replisome (Barry & oril located within O, and the O (initiator) and P (helicase
Alberts, 1994a). loader) genes transcribed from the pR promoter together

FEMS Microbiol Rev 30 (2006) 321–381

c 2006 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 340 C. Weigel & H. Seitz

with the cII and cro genes in an ill-defined context. It was replication machinery. Therefore, we must include
shown in numerous subsequent studies that the cII and the all phage-encoded replication functions in order to obtain
cro regulatory loops are not essential for l plasmid replica- a useful definition of ‘bacteriophage replication modules’.
tion; transcription from the po promoter seems important As we will show in the following, the close linkage of
but not the transcript, oop RNA (for details, see Taylor & replication genes in most phage genomes justifies this
Wegrzyn, 1995). Finally it was shown that pR can be replaced expansion of the ‘replicon model’, and even suggests possible
by a different (inducible) promoter (Herman-Antosiewicz functions for experimentally uncharacterised proteins in
et al., 2001), which relieves l plasmid replication from the some cases.
intricate control by host DnaA (Glinkowska et al., 2003). For a precise definition of phage replication modules, the
These results led to the functional definition of the l-type emphasis on the initiation step in the ‘replicon model’
replication module being composed of the O (oril) and P appears as a weak point. The replication of many phage
genes (Wrobel & Wegrzyn, 2002). l plasmids could thus genomes requires recombination steps that are, in most
serve as excellent model systems for the initiation – and cases, performed by cognate recombination proteins in
initiation control – of bidirectional l replication in the y order to provide the relinearised form that is the substrate
mode. In addition, unidirectional replication of l plasmids, for packaging into phage capsids (COM section C3.6.2.).
which precedes the switch from yDR to sDR during l phage As we will show in the following, there is a striking co-
replication, could be shown (Baranska et al., 2002). How- localisation of replication and recombination genes in many
ever, the switch from yDR to sDR – characteristic for l phage genomes. Therefore, we include known and putative
phage replication – was never observed with l plasmids. It is recombination genes in our definition of phage replication
not clear at present whether this is due to the lack of the modules.
required recombination functions Reda/Redb (Exo/Bet), The discussion in this section will focus on four
and RapA (NinG) in l plasmids, or due to the lacking Gam major types of replication modules: (1) modules containing
function (inhibitor of host RecBCD). This demonstrates initiator genes, (2) modules containing DNA polymerase
that the straightforward ‘functional approach’ to define genes, (3) modules containing fP4a-type helicase-primase
replication modules can eventually fail to reveal auxiliary genes and (4) the replication modules of filamentous
components. phages. Although this formal division seems somewhat
Historically, the first useful definition of a prokaryotic eclectic, it reflects the present knowledge – but not phage
replication module was given in the ‘replicon model’ by systematics, nota bene. Where possible, the definition
Jacob, Brenner and Cuzin: ‘The replicon is assumed to be a of the individual types of replication modules is based on
circular structure carrying two specific genetic determi- experimental results. We will include, in addition, the
nants. A structural gene determines the synthesis of a results of similarity searches discussed in COM section C3.
diffusible active element, the initiator. The initiator acts on Furthermore, the definitions will be based on the gene
a replicator, allowing the beginning of the replication which arrangements of fully functional phages as represented
proceeds along the circular structure’ (Jacob et al., 1963, p. in the completely sequenced phage genomes. We include
331). A particularly startling aspect of the ‘replicon model’ in the discussion several prophage genomes but because
was the hypothesis that the initiation of replication is their replication/recombination genes might have under-
positively regulated, which is indeed the case for all known gone rearrangements and/or inactivation in the prophage
bacteriophage replicons (Nordström, 2003). However, this state they cannot serve as a basis for the definition.
clear-cut definition can only be applied to phage replicons We do not discuss in depth the important point of the
with several important modifications. (1) The replicon may transcription, and its regulation, of the bacteriophage repli-
be circular or linear DNA. Many linear phage genomes cation/recombination genes because experimental results
recircularise prior to replication, but others initiate are too scarce and predictions doubtful. We expect, never-
replication on the linear substrate. (2) The replicator (in theless, that a formal classification of phage replication
modern terms: replication origin) is a unique structure modules will help to improve the assignment of putative
in most phage replicons, but multiple origins are (pro)phage gene functions in future genomic sequencing
known for those phages where replication is initiated at projects.
D- or R-loops. (3) Many phages encode cognate initiators.
However, phage replicons using R-loops for replisome
Phages encoding initiator proteins
assembly do not encode a cognate initiator in the strict
sense. We have discussed in COM section C3.1.2. the phage-
With the notable exception of the fT4-type phages, encoded initiator proteins for yDR, with l O and fSPP1
bacteriophages are semiautonomous replicons and have G38P as the best understood examples. Both initiator genes
evolved various strategies to recruit components of the host contain the phage replication origin, a common feature of

c 2006 Federation of European Microbiological Societies FEMS Microbiol Rev 30 (2006) 321–381
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 Bacteriophage replication modules 341

Fig. 9. ‘Initiator-solo’ replication modules of Gram(1)-specific phages: part A. Genes encoding replication and recombination functions are shown in
their genomic context. fP335 (IL-type module) was included for better comparison. The alignment is shown with homologues of fA118 gp49 (initiator)
at fixed positions. Blocks with solid colours indicate (putative) gene functions detected by BLAST comparison: int, integrase; repr, l cI-type phage
repressor; recE, recT, rusA, ruvC, erf, putative recombinases; ini, initiator; loader, DnaC-type helicase loader; meth, methylase; ssb, single-strand DNA
binding protein; dut, dUTPase. Apparently truncated genes are shown in square brackets. ORFs with significant similarity ( 4 30% identical residues) are
indicated by striped colouring. Homologues of fPVL orf63 and fBK5-T orf63 (pink colour marked with an asterisk) were found in fE125 (gp70) (see Fig.
12) and in f3626 (orf50) (see Fig. 15). Dark and light grey colouring indicates ORFs lacking homologues in any other phage genome. Dark and light grey
colouring with black outline indicates ORFs with homologues in (completely sequenced) phages other than compared here. The ORF size is indicated by
block height:  100 residues = 1 U,  200 residues = 2 U,  300 residues = 3 U, etc. The relative positions of the ORFs in the phage genomes are
indicated by distances in kilobases (not to scale). Except where indicated by ‘ 4 ’ the direction of transcription is from up to down. The sequences were
taken from the genome entries for fA118 (40.8 kb) [NC_003216], L. innocua pf (3.0 Mb) [NC_003212], C. tetani pf (2.8 Mb) [NC_004557], L. lactis
pf Lp1 (3.3 Mb) [NC_004567], L. gasseri pf (10.7 kb, partial sequence) [NZ_AAAO02000006], L. monocytogenes pf (2.94 Mb) [NC_003210], fLC3
(32 kb) [NC_005822], f31.1 (9.9 kb, partial sequence) [AF208055], ful36.1 (8.1 kb, partial sequence) [AF212846], fTP901-1 (37.7 kb) [NC_002747],
fPVL (41.4 kb) [NC_002321], fBK5-T (40 kb) [NC_002796], fTuc2009 (38.3 kb) [NC_002703] and fP335 (36.6 kb) [NC_004746].

this type of initiator gene (COM section C2.2.). In their ‘Initiator-solo’ replication modules
respective genomes, both initiator genes are directly fol-
lowed by other replication genes: the l P gene encoding the A number of Gram(1)-specific phages possess a (putative)
helicase loader and the fSPP1 genes 39 and 40 encoding the initiator gene – containing the phage replication origin – in
helicase loader and the helicase, respectively (COM sections addition to recombination genes, but lack detectable heli-
C3.2. 1 C3.3.). The initiation of replication of both phages case loader or helicase genes. These phage genomes share a
differs with respect to the entry point of the host replication common architecture: genes encoding integrases and phage
machinery: the origin-bound l O recruits the host replica- repressors are found upstream of the initiator gene, and
tive helicase as l P3  DnaB6 complex, while the origin- transcribed in the opposite direction. Genes encoding
bound fSPP1 G40P helicase recruits the host primase after exonuclease/SAP gene pairs (COM section C3.6.2.) are
dissociation of the unstable G38P  G39P  G40P ATP inter- located between the initiator and repressor gene, and the
mediate complex (COM section C3.2.). We therefore discuss direction of their transcription is the same as for the
‘initiator-helicase loader’ (IL-type) and ‘initiator-helicase repressor gene. Genes encoding Holliday junction resolvases
loader-helicase’ (ILH-type) replication modules separately. are, when present, invariably found downstream of the
A third type, the ‘initiator-helicase’ (IH-type) replication initiator gene. Genes encoding SSBs are present in several
module, is represented by Salmonella sp. phage P22: the phage genomes, but at varying positions. Among the genes
initiator gene 18, containing the fP22 replication origin, is that are also invariably found downstream of the initiator
directly followed by the helicase gene 12. We will start the gene are the genes encoding (putative) dUTPases, most
discussion with yet a fourth type, the ‘initiator-solo’ (I-solo) frequently found in the genomes of Lactococcus/Lactobacil-
type of replication modules. lus phages (Fig. 9).

FEMS Microbiol Rev 30 (2006) 321–381

c 2006 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 342 C. Weigel & H. Seitz

Fig. 10. Secondary structure prediction for the fA118 gp49 initiator, and the Bacillus subtilis DnaD and DnaB helicase loaders. Secondary structure
predictions for fA118 gp49 [NP_463514], B. subtilis DnaB [NP_390777] and B. subtilis DnaD [NP_390116] were obtained by the Jpred method (Cuff
et al., 1998). Colour code: red, a-helical region; green, b-strand; black line, unstructured. Regions with significant similarity are indicated by grey blocks.
‘2D similarity’ indicates a region showing a comparable secondary structure prediction but lacking protein sequence similarity. ‘% ident.’ indicates
regions with significant protein sequence similarity in addition to a similar secondary structure prediction.

The initiator, recombination and ssb genes are embedded their hosts, but with DnaBBsu only for fA118 gp49 (Table 1,
in a highly variable context of mostly small ORFs with lanes 1–3; Fig. 10).
unknown function. A number of these small ORFs are found Several putative initiators encoded by prophages of
at corresponding or at different positions in other phage bacillales genomes also show this particular arrangement of
genomes. But there are also ORFs lacking known homo- their C-terminal domains (Table 1, lanes 4–6). Interestingly,
logues. We can exclude the possibility that these small ORFs the replication initiators of various Staphylococcus sp. plas-
encode essential replication functions because none of them mids, which are structurally not related to phage initiators,
is found conserved in the entire set of phages compared in also contain at their C-termini a stretch of 50 residues
Fig. 9. The I-solo type of replication modules can thus be showing significant similarity with the DnaB helicase loa-
defined as: an initiator gene containing the phage replication ders of their hosts (Table 1, lanes 7 and 8) and with the C-
origin, exonuclease/SAP genes either of the recE/recT- or the terminal 50 res. of the phage initiators, respectively (not
erf-type, and resolvase genes of the rusA-type. Because shown). ‘MultAlin’ analysis of the protein sequences from
recombination genes are, like ssb genes, not found in all Table 1 did not produce a reasonable consensus sequence
phage genomes we have to define them as accessory func- (Corpet, 1988). ‘JPred’ and ‘PHD’ secondary structure
tions. prediction analysis suggests, however, that the C-terminal
Whether it is appropriate to include fTuc2009 in the I- half of the 50-residue-long ‘DnaB-tail’ assumes a-helical
solo type group cannot be answered satisfactory at present: conformation in most proteins preceded by an unstructured
orf17 (241 res.) downstream of the orf18 initiator occupies loop with a conserved tryptophan residue (Fig. 10). The
the position where the helicase loader gene is found in unstructured loop containing the conserved tryptophan is
fP335 (Fig. 9). Homologous genes of fTuc2009 orf17 are also detectable at the extreme C-terminus of most known
found at corresponding positions in phages fbIL285, ful36, DnaDBsu orthologues, but which all lack the predicted a-
ful36.2 fQ33, fQ30 and L. lactis pf pi2. BLAST searches helix. The relatively small size of this ‘DnaB-tail’
suggest a distant relationship of fTuc2009 orf17 to DnaI suggests that it is of functional rather than of structural
proteins of Firmicutes but the characteristic Walker A-type importance.
NTP-binding motif is not detectable. It seems possible that Phages fTP901-1, fBK5-T, f31.1 and ful36.1 of Lacto-
fTuc2009 orf17 represents a yet unknown type of helicase bacillus sp. and related prophage genomes encode putative
loader. initiators with a detectable ‘DnaB-tail’ (Table 1, lanes
The initiation of replication in these phages has not been 13–16). The initiators of fBK5-T, f31.1, and ful36.1 show
studied experimentally. There is good reason, however, to an overall similarity of 30% among each other, with the
hypothesise that the mechanism resembles the intricate similarity rising to 90% (identical residues) within the C-
mechanism for helicase loading in B. subtilis, which involves terminal 50 residues. In contrast to the fTP901-1 REP gene
the concerted action of the DnaB, DnaD and DnaI helicase and the pfLp1 gene20 the putative initiators of phages
loaders to recruit the replicative helicase DnaC (see COM fBK5-T, f31.1 and ful36.1 and the Lactobacillus gasseri pf
section C3.2. for details). The (putative) initiator gp49 of gene lgas0588 lack detectable similarity with the cognate
Listeria sp. phage fA118, and the almost identical (putative) DnaD. The putative initiators encoded by gene20 of the
initiators of Staphylococcus aureus phages fPVL (orf46) and Lactobacillus plantarum prophages Lp1 and Lp2 lack detect-
fN315 (sa1791) contain in their C-terminal domain a able similarity in their N-termini but have virtually identical
region of similarity with the DnaD helicase loader of their C-termini; only the former is therefore included in Table 1.
hosts and with DnaDBsu [pfam04271] (Fig. 10), directly For pfLp1 we found no similarity with DnaDBsu but with
followed by – and partially overlapping with – a stretch of DnaD of Bacillus halodurans instead (Table 1, lane 15).
50 residues that shows significant similarity (Z40% ident. Orthologues of DnaBBsu are only detectable among
residues) with the C-termini of the DnaB helicase loaders of species of the bacillales and lactobacillales subgroups of the

c 2006 Federation of European Microbiological Societies FEMS Microbiol Rev 30 (2006) 321–381
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 Bacteriophage replication modules 343

Table 1. Similarity of the initiators of Gram(1)-specific phages to the DnaD and DnaB helicase loaders of their hosts
Host DnaB (%) DnaBBsu (%)
Host DnaD DnaDBsu
f /pf/p/chr. Accession no. Gene Residue (entire length) (entire length) n  100 n  50 n  100 n  50
1. f A118 (Listeria monocytogenes) NP_463514 gp49 310 29% 172–275 25% 177–257 34 43 30 
2. f PVL (Staphylococcus aureus) NP_058485 orf46 297 24% 87–273 23% 152–265 21 42  
3. f N315 (S. aureus) NP_835538 sa1791 297 24% 87–273  21 42  
4. Listeria innocua pf NP_471742 lin2412 310 27% 152–273 25% 177–257 34 43 29 
5. L. innocua pf NP_469432 lin0086 303 22% 37–266 28% 171–249 30 41 33 
6. L. monocytogenes pf NP_465841 lmo2317 324 27% 161–287 25% 191–271 30 43 32 
7. pIP1629 (Streptococcus AAD02381 Rep1 285   40 41  
epidermidis)
8. pN315 (S. aureus) NP_395563 sap027 286    45  
9. S. aureus NP_374796 DnaB 466 24% 268–394  nd nd 29 
10. L. innocua NP_470932 DnaB 458   nd nd 43 35
11. L. monocytogenes NP_465086 DnaB 458   nd nd 37 44
12. Bacillus subtilis NP_390777 DnaB 472 nd  nd nd nd nd
13. f TP901-1 (Lactococcus lactis) NP_112676 REP 272 27% 142–229 29% 145–245  48  
14. f BK5-T (L. lactis) NP_116541 orf49 269    48  
15. Lactobacillus plantarum pf Lp1 NP_784408 gene 20 310 29% 184–257 31%w 194–266 26 32 21 
16. Lactobacillus gasseri pf ZP_00046421 lgas0588 307   30 43  

Values are percentage identical residues; a dash indicates no significant homology detectable by BLAST (bl2seq); nd, not done (self-comparison). BLAST
(bl2seq; Tatusova & Madden, 1999 similarity searches were performed for ‘host DnaD’ and ‘DnaDBsu’ [NP_390116] with the complete sequence as
query; ‘host DnaD’ were: L. monocytogenes DnaD [NP_465419], S. aureus DnaD [NP_374567], L. innocua DnaD [NP_471343], S. epidermidis DnaD
[NP_764696], L. lactis DnaD [NP_267226], L. plantarum DnaD [NP_785314] and L. gasseri DnaD [ZP_00045943]. For the columns showing the BLAST
results with host DnaD and DnaDBsu as queries, the percentage identity value is given together with the position of the matching region in the subject
sequence. BLAST (Bl2seq) similarity searches were performed for ‘host DnaB’ and DnaBBsu sequentially (1) for the C-terminal 100 residues (n–100) and
(2) for the C-terminal 50 residues (n–50); ‘host DnaB’ were: L. lactis DnaB [NP_266907], S. epidermidis DnaB [NP_764914], L. gasseri DnaB
[ZP_00046732] and L. plantarum DnaB [NP_785118].
Similarity detected in a stretch of 25 residues by genome BLAST with the C-terminal 50 residues of REP and orf49.
w
Similarity with B. halodurans DnaD [NP_242563].

firmicutes (COM section C3.2.). This observation corre- 4 70% identical residues Bacillus/Listeria, 60% identical
sponds to the finding that initiators containing a ‘DnaB-tail’ residues Bacillus/Staphylococcus).
could only be detected in the genomes of phages that infect We deduce from the above that the initiators in the ‘I-
species from these two phylogenetic groups. Several of the solo’ type replication modules of Gram(1)-specific phages
(putative) phage initiators analysed here contain in addition contain in their C-terminal domains subdomains that inter-
to the ‘DnaB-tail’ a region of similarity with DnaDBsu act with the DnaC replicative helicases of their hosts, and/or
preceding the ‘DnaB-tail’ in their C-terminal domains. The with the third helicase loader, DnaI, in addition. A more
direct comparison of the DnaB and DnaD proteins of S. precise hypothesis would require a more detailed knowledge
aureus reveals a region of 100 residues of limited similarity about helicase loading in firmicutes. It is safe, however, to
(Table 1, lane 9). This region corresponds to the region of assume that the initiators are sufficient to direct the host
similarity with DnaD found in the initiators of phages replication machinery to the phage replication origin at the
fA118, fPVL, fN315 and fTP901-1. Given the related step of helicase loading.
function of the DnaD and DnaB proteins, this finding may With a length of only 137 and 88 residues, the (putative)
point to a common evolutionary origin of both proteins and initiators of phages fNIH1.1 and fMM1, respectively, are
might help to unravel the origin of the phage initiators unusually short (Fig. 11). The fNIH1.1 orf08 initiator
containing a ‘DnaB-tail’. By contrast, the DnaD and DnaB protein shows similarity (27% identical residues) to the C-
proteins of B. subtilis, L. gasseri, Listeria sp. (Table 1, lanes terminus of the fTP901-1 REP initiator. orf07 and orf08 are
10–12), and Lactococcus/Lactobacillus sp. seem unrelated. separated by an untranslated stretch of 473 bp (NCBI entry
We note, in addition, the rather low conservation of the NC_003157; position 5719–6192). A 111-residue-long ATG-
DnaB proteins (C-termini) among the closely related less ORF could be readily identified within this stretch that
genera Bacillus, Listeria and Staphylococcus. By contrast, shows 42% identity with the N-terminus of the fTP901-1
other replication proteins are highly conserved, e.g. DnaA: REP initiator, and which may represent the ‘missing’ orf08

FEMS Microbiol Rev 30 (2006) 321–381

c 2006 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 344 C. Weigel & H. Seitz

gene orf6 is too short to accommodate a gene encoding a

putative initiator C-terminus, which was also not found
elsewhere in the fMM1 genome, assuming the possibility of
gene splitting. López and co-workers could identify a
putative replication origin downstream of orf5, but a phage
replication mechanism involving a (putative) initiator de-
void of an oligomerisation/interaction domain has so far
not been studied experimentally (Obregon et al., 2003). The
architecture of the fMM1 genome resembles most closely
that of fNIH1.1: in both genomes several genes usually
found upstream of the initiator gene are located down-
stream instead. It seems possible that the (putative) initiator
gene fMM1 orf5 suffered a deletion during this rearrange-
ment (see Fig. 11).
Gram(  )-specific phages with I-solo type replication
modules include fSfV, fST64B and phage e15 of entero-
bacteria, and the Burkholderia mallei (Betaproteobacteria)
phage fE125. All four phages show some gross similarity in
genome architecture among each other (Fig. 12) and in
Fig. 11. ‘Initiator-solo’ replication modules of Gram(1)-specific phages: comparison to the Gram(1)-specific phages (Fig. 9). fSfV
part B. Genes encoding replication and recombination functions are and fST64B are closely related with long stretches of
shown in their genomic context. f31.1, fTP901-1 and Lactobacillus significant DNA sequence similarity along their entire
gasseri pf (I-solo type module), as well as fSM1 (IL-type module) were
genomes. Not surprisingly, the (putative) initiator genes
included for better comparison. The alignment is shown with homo-
logues of fTP901-1 REP (initiator) at fixed positions. Blocks with solid
fSfV orf39 and fST64B sb42 are homologous (87% iden-
colours indicate (putative) gene functions detected by BLAST compar- tical residues), and the arrangement of the flanking genes is
ison: int, integrase; repr, l cI-type phage repressor; erf, rusA, putative well conserved. By contrast, the (putative) initiators e15 p42
recombinases; ini, initiator; ssb, single-strand DNA binding protein; dut, and fE125 gp60 share no protein sequence similarity, only
dUTPase. In fNIH1.1, an asterisk marks the ATG-less ORF representing the latter shows some weak similarity to the N-terminal
the putative initiator N-terminus (see text for details) upstream of the DNA-binding domains of fSfV orf39 and fST64B sb42. All
orf08 initiator (C-terminus). ORFs with significant similarity ( 4 30%
four initiator genes contain the (putative) phage replication
identical residues) are indicated by striped colouring. Dark and light grey
colouring indicates ORFs lacking homologues in any other phage
origins (COM section C2.2.). The apparent differences in
genome. Dark and light grey colouring with black outline indicates ORFs the regions flanking the initiator genes of the four phages
with homologues in (completely sequenced) phages other than com- allows for a straightforward description of the replication
pared here. The ORF size is indicated by block height:  100 re- module: an initiator gene containing the replication origin,
sidues = 1 U,  200 residues = 2 U,  300 residues = 3 U, etc. The and – as accessory functions – genes encoding Holliday
relative positions of the ORFs in the phage genomes are indicated by junction resolvases of the RusA- (fSfV, fST64B) or RuvC-
distances in kilobases (not to scale). Except where indicated by ‘ 4 ’ the
type (e15). As found for the Gram(1)-specific phages with
direction of transcription is from up to down. The sequences were taken
from the genome entries for fTP901-1 (37.7 kb) [NC_002747], f31.1
I-solo-type replication modules, the resolvase genes are
(9.9 kb; partial sequence) [AF208055], L. gasseri pf (10.7 kb; partial located downstream of the initiator genes. A pair of recE/
sequence) [NZ_AAAO02000006], fNIH1.1 (41.8 kb) [NC_003157], recT-type recombination genes is only encoded by e15, and
fMM1 (40.2 kb) [NC_003050] and fSM1 (34.7 kb) [NC_004996]. located between the integrase and the phage repressor genes.
This localisation seems to be conserved in Gram(  )-
specific phage genomes. The function of the ParB-like
N-terminus (marked with an asterisk in Fig. 11). Although protein gp58 of fE125 remains to be studied. Despite the
we have to await a revision of the fNIH1.1 DNA sequence, different arrangement of the recombination genes, the
we tentatively assume that fNIH1.1 does not posses an replication module of these four Gram(  )-specific phages
‘unusual’ I-solo type replication module. The situation is is identical to that of the ‘I-solo’-type replication module of
more complex for fMM1. The (putative) initiator encoded the Gram(1)-specific phages.
by orf5 shows significant similarity to the N-terminal DNA- The initiation of replication of these four phages has not
binding domains of several related phages, including the been studied, but we can assume that the phage initiators
111-residue-long (putative) ORF upstream of fNIH1.1 recruit the host replicative helicase directly, i.e. without
orf08 (37% identical residues), but lacks a C-terminal involving a specific helicase loader, in order to gain access
oligomerisation domain. The distance to the downstream to the host replication machinery. It should be kept in mind

c 2006 Federation of European Microbiological Societies FEMS Microbiol Rev 30 (2006) 321–381
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 Bacteriophage replication modules 345

not include the 20 C-terminal residues of DnaT, which

may be important for the interaction with DnaC (see
‘Evolutionary considerations’ section). It is possible, there-
fore, that the ‘DnaT-like’ region in e15 p42 includes a site for
interaction with DnaB. Such a ‘DnaT-like’ region could not
be detected in the other three initiators, though. It thus
remains unresolved how they attract the host helicase. A
hint may be the observation that the fSfVorf40 protein (162
residues) shares a region of significant similarity with the E.
coli primosomal protein PriA (residues 34–75; 38% identical
residues). This region is missing in the truncated orf40
homologue p31 of phage e15. The molecular architecture
of E. coli PriA is not well understood, and therefore an easily
testable hypothesis cannot be derived from this observation.
In the E. coli K12 genome, gene yfdN encodes a homologue
of fSfV orf40 (95% identical residues). The preceding yfdO
gene encodes a protein with significant similarity to the C-
Fig. 12. ‘Initiator-solo’ replication modules of Gram(  )-specific phages.
terminus of the fSfV orf39 initiator. YfdO and YfdN are
Genes encoding replication and recombination functions are shown in
their genomic context. The alignment is shown with homologues of fSfV thus the remnants of the replication module of the highly re-
orf39 (initiator) at fixed positions. Blocks with solid colours indicate arranged and truncated KpLE1 prophage (Fig. 12).
(putative) gene functions detected by BLAST comparison: int, integrase; The small orf59 (86 residues) upstream of the orf60
repr, l cI-type phage repressor; recE, recT, rusA, parB, putative recombi- initiator in the fE125 genome shows (BLAST) similarity to
nases; ini, initiator. Apparently truncated genes are shown in square the N-terminus of the fVorf39 initiator, but no similarity to
brackets. ORFs with significant similarity ( 4 30% identical residues) are
orf60. Although orf59 encodes a (putative) DNA-binding
indicated by striped colouring. Homologues of fE125 gp70 (pink colour
domain, we believe that the protein is a recombination relic
marked with an asterisk) were found in fPVL (orf63), fBK5-T (orf63) (see
Fig. 9) and f3626 (orf50) (see Fig. 15). Dark and light grey colouring – a partial duplication – rather than a functional important
indicates ORFs lacking homologues in any other phage genome. Dark and protein. The (putative) initiator encoded by the recently
light grey colouring with black outline indicates ORFs with homologues in sequenced Burkholderia pseudomallei phage f1026b is vir-
(completely sequenced) phages other than compared here. The ORF size is tually identical with fE125 gp60 (97% identical residues),
indicated by block heigth:  100 residues = 1 U,  200 re- and the initiator gene is also preceded by the partial initiator
sidues = 2 U,  300 residues = 3 U, etc. The relative positions of the ORFs
duplication found in fE125. Both phage genomes are
in the phage genomes are indicated by distances in kilobases (not to scale).
therefore very closely related, and f1026b is not discussed
Except where indicated by ‘ 4 ’ the direction of transcription is from up to
down. The sequences were taken from the genome entries for Escherichia separately here.
coli K12 pf KpLE1 (4.6 Mb) [NC_000913], fSfV (37.1 kb) [NC_003444],
fST64B (40.1 kb) [NC_004313], f e15 (39.7 kb) [NC_004775] and fE125
(53.4 kb) [NC_003309].
‘Initiator-helicase loader’ replication modules
In addition to the O initiator and the P helicase loader as
that the mechanism of helicase loading is fundamentally essential factors, replication of l phage may require the
different in E. coli and B. subtilis: the E. coli DnaB helicase is action of the Reda, Redb and RapA recombination proteins
recruited to the site of loading as a stable hetero-hexameric (see ‘Initiation by melting: theta (y)-type DNA replication’
complex (DnaB6C6). By contrast, the B. subtilis DnaC heli- section). Genes encoding all these functions are found at
case hexamer is assembled at the site of loading by the corresponding positions in the genomes of f933W and
concerted action of the DnaB, DnaD and DnaI helicase f4795, and define the components of the IL-type replication
loaders (see above). Despite their identical I-solo-type module. There is a considerable degree of similarity in
replication modules the initiator proteins of Gram(1)- and arrangement and type of ORFs downstream of the initiator
Gram(  )-specific phages have (had) to adapt to these and helicase loader genes in l, f933W and f4795, but this
specific requirements of the host proteins. An interesting similarity is not found in the region upstream of the phage
feature of the phage e15 p42 initiator is its significant C- repressor gene (Fig. 13). Thus, the three phages are clearly
terminal similarity to the E. coli primosomal protein DnaT distinct though related. fH-19B encodes an Erf-type instead
(residues 157–217; 40% identical residues). DnaT directs the of a Redb-type SAP at the corresponding position in its
DnaB6C6 double-hexamer to the restart primosome (DNA- (partially sequenced) genome, which adds further support
bound PriA, PriB and PriC) during replication restart (see to the notion that these proteins are functionally equivalent
‘Replication restart’ section). The region of similarity does (COM section C3.6.2.). We found homologues of the l ren

FEMS Microbiol Rev 30 (2006) 321–381

c2006 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 346 C. Weigel & H. Seitz

Fig. 13. ‘Initiator-helicase loader’ replication modules of Gram(  )-specific phages. Genes encoding replication and recombination functions are
shown in their genomic context. fVT2-Sa (IH-type module) was included for better comparison. The alignment is shown with homologues/functional
analogues of l O (initiator) at fixed positions. Blocks with solid colours indicate (putative) gene functions detected by BLAST comparison: int, integrase;
repr, l cI-type phage repressor; exo, reda-type exonuclease; bet, redb-type single-strand annealing protein (SAP); erf, fP22 erf-type SAP; recE, recT,
RecE-, RecT-type recombinases; ini O, initiator; loader, DnaC-type helicase loader; P, l P-type helicase loader; hel, F4-type helicase; rapA, RapA-type
resolvase; ruvC, RuvC-type resolvase; ssb, single-strand DNA binding protein. For the Rac prophage, the gene names of the Escherichia coli genome are
indicated. ORFs with significant similarity ( 4 30% identical residues) are indicated by striped colouring. Dark and light grey colouring indicates ORFs
lacking homologues in any other phage genome. Dark and light grey colouring with black outline indicates ORFs with homologues in (completely
sequenced) phages other than compared here. The ORF size is indicated by block height:  100 residues = 1 U,  200 residues = 2 U,  300
residues = 3 U, etc. The relative positions of the ORFs in the phage genomes are indicated by distances in kilobases (not to scale). Except where indicated
by ‘ 4 ’ or by an arrow spanning several ORFs, the direction of transcription is from up to down. The sequences were taken from the genome entries for
fVT2-Sa (60.9 kb) [NC_000902], l (48.5 kb) [NC_001416], f933W (61.7 kb) [NC_000924]; f4795 (57.9 kb) [NC_004813], fH-19B (18.4 kb; partial
sequence) [AF034975], f21 (4.9 kb; partial sequence) [B21237660], f80 (6 kb; partial sequence) [BP80ER], fBcep22 (63.9 kb) [NC_005262];
Salmonella enterica (typhimurium) LT2 pf Gifsy-2 (4.95 Mb) [NC_003197]; E. coli K12 pf Rac (4.6 Mb) [U00096]. (a) l P-type helicase loader; (b)
DnaC-type helicase loader.

gene in all phages encoding a P-type helicase loader, and We found only few examples for IL-type replica-
invariably downstream of the P gene homologue, but not tion modules with a DnaC-type helicase loader among
outside this phage group. ren provides resistance to Rex Gram(  )-specific (pro)phages (Fig. 13b). The Burkholderia
exclusion, and has therefore not to be considered to have a sp. (Betaproteobacteria) phage fBcep22 has a gross archi-
replication function (Campbell, 1994). tecture resembling that of l with three major differences: (1)
The architectures of l and fVT2-Sa are almost identical it lacks a detectable phage repressor gene, (2) the exonu-
(Fig. 13). In fact, there is a stretch of 5 kb of almost clease/SAP gene pair is of the recE/recT-type and (3) a ruvC-
complete DNA identity in both phage genomes, encompass- type resolvase gene is located upstream of the (putative)
ing the redab genes. However, both phages posses different initiator-helicase loader gene pair. The direction of tran-
replication modules, an IH-type module in the case of scription of the fBcep22 recET gene pair corresponds to
fVT2-Sa and the IL-type in the case of l. Although both that of l redab, and therefore fBcep22 is probably no
replication modules are characterised by a different mechan- exception from the ‘rule’ that recombination genes are
ism and point of attraction of the host replication machin- located between the integrase and repressor genes in Gram-
ery, this observation gives strong support to the notion that negative-specific phages. The small size (77 residues) of the
both modules are functionally equivalent and may be ex- ruvC-type resolvase fBcep22 gp15 suggests that the protein
changed between phages by recombination, a topic dis- is inactive and represents a recombination relic.
cussed to more detail in the ‘Replication module exchange The architectures of the Salmonella sp. prophage Gifsy-2
among bacteriophages’ section. and the E. coli K12 Rac prophage closely resemble that of

c 2006 Federation of European Microbiological Societies FEMS Microbiol Rev 30 (2006) 321–381
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 Bacteriophage replication modules 347

fBcep22, and also that of l. Given the highly similar for the I-solo-type replication modules of Gram(1)-specific
architecture of Gifsy-2 and Rac, they may represent deriva- phages: integrase and phage repressor genes are located
tives of a common ancestral phage. Both prophages encode upstream of the initiator gene, and transcribed in the
fully functional replication genes, and are interesting exam- opposite direction. Exonuclease/SAP recombination genes
ples for phage replicons where a phage-encoded helicase of the recE/recT- or erf-type are in all cases located down-
loader competes with a host orthologue, DnaC, for the host stream of the phage repressor genes, and either upstream
replicative helicase. The extremely long recE genes are (fSLT, fPV83, fr1t, fP335, fLL-H, fmv4) or, less fre-
uniquely found in these two prophages, with the exonu- quently, downstream (fEJ-1, f7201) of the initiator/heli-
clease activity residing in the C-terminal 300 residues case loader gene pairs. Resolvase genes, mostly rusA-type,
(COM section C3.6.2.). Although RecE and RecT are are invariably found downstream of the initiator/helicase
functionally analogous to the l Reda and Redb proteins, loader gene pairs (Fig. 14). The position of the ssb genes is,
we note that the gene order is usually reversed in the recET as observed before, highly variable. An ssb gene could not be
gene pairs (Fig. 13). detected in the fP335 genome, but we note that orf10 (306
As mentioned in COM section C3.2., the protein encoded residues) upstream of the initiator orf11 has a highly acidic
by the pf Rac ydaV gene (b1360) shows significant homol- SSB-like C-terminus. Although the IL-type replication
ogy to E. coli DnaC (50% identity) (Wrobel & Wegrzyn, modules of the Gram(1)-specific phages is identical to that
2002; Casjens, 2003). Wróbel & Wegrzyn (2002) found, in of the Gram(  )-specific phages with respect to gene
addition, that the protein encoded by the pf Rac ydaU gene composition, there is definitively a higher degree of similar-
upstream of ydaV shows significant homology (40% iden- ity within each group with respect to the positioning of the
tity) to the C-terminal 80 residues of E. coli DnaT (179 exonuclease/SAP recombination genes. Within the group of
residues). In the genomes of E. coli K12 and the other Gram(1)-specific phages with IL-type modules, a compar-
completely sequenced E. coli strains CFT073, O157:H7 and ison of the smaller ORFs surrounding the replication/recom-
O157:H7 EDL933, dnaT is located directly upstream of bination genes allows a further differentiation at the level of
dnaC in a small operon (Masai & Arai, 1988). dnaTC gene host species subgroups: one can clearly distinguish between
pairs are also present in the chromosomes of Shigella flexneri Staphylococcus, Streptococcus and Lactobacillus/Lactococcus
2a (DnaT: 100% identical residues/DnaC: 100% identical phages. In Fig. 14 we have chosen red-striped colouring for
residues), Salmonella enterica (typhimurium) LT2 (81%/ ORFs exclusively present in Staphylococcus phages (a), green-
93% identical residues), Klebsiella pneumoniae (74%/93% striped colours for ORFs specific for Streptococcus phages (b),
identical residues), Buchnera aphidicola APS (42%/65% and violet-striped colours for ORFs specific for Lactobacillus/
identical residues) and B. aphidicola Sg (36%/65% identical Lactococcus phages (c). ORFs with homologues in more than
residues). As mentioned in COM section C3.2, the dnaTC one subgroup are indicated by blue-striped colours, and
gene pair is not present in the chromosomes of sequenced these ORFs are in many cases located distantly to the
Yersinia sp. strains (Thomson et al., 2002), and also not in initiator/helicase loader gene pairs. This indicates that re-
those of other species outside the enterobacterial branch of combination events that lead to viable progeny occur fre-
the Gammaproteobacteria. We discuss in the ‘Evolutionary quently and preferentially among phages with the same host
considerations’ the evolutionary relationship of the E. coli range. Although this remark seems somewhat trivial, it
dnaTC and the pfRac ydaUV initiator-helicase loader gene points to a possible reason why the gross architecture of
pair. many phages with different replication modules is so strik-
The Gram(1)-specific phages with IL-type replication ingly similar: the replication modules are apparently func-
modules encode initiators of the l O-type but exclusively tionally equivalent.
helicase loaders of the DnaCEco-type (COM section C3.2.).
Also in this group, the phage replication origins reside
‘Initiator-helicase’ replication modules
within the initiator genes (COM section C2.2.). fSM1
provides the rare exception from the ‘rule’ that the (puta- Salmonella phage P22 encodes two essential replication
tive) initiator gene is located directly upstream of the heli- proteins: gene 18, the l O-type initiator, and gene 12, the
case loader gene: both genes are separated by a small DnaB-type replicative helicase. In addition, the fP22 erf
intervening ORF (52 residues) of unknown function (Fig. gene encodes an essential recombination function (Botstein
14). The conserved linkage of initiator/helicase loader genes & Matz, 1970). Wickner (1984b) showed that purified fP22
allows us to assign the function of a helicase loader, e.g. to gene 12 protein can replace E. coli DnaB helicase in the
f7201 orf5 protein, which gives equally significant hits with fX174 DNA in vitro DNA synthesis assay, and bypasses the
the structurally similar IstB-like transposase small subunits requirement for DnaC. This suggests that during initiation
in BLAST searches. The initiator/helicase loader gene pairs of fP22 replication the gene 18 initiator recruits the gene 12
are embedded in a comparable context as already described helicase to the unwound origin by direct interaction.

FEMS Microbiol Rev 30 (2006) 321–381

c2006 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 348 C. Weigel & H. Seitz

Fig. 14. ‘Initiator-helicase loader’ replication modules of Gram(1)-specific phages. Genes encoding replication and recombination functions are shown
in their genomic context. The alignment is shown with homologues/functional analogues of fSLT orf256 (initiator) at fixed positions. Blocks with solid
colours indicate (putative) gene functions detected by BLAST comparison: int, integrase; repr, l cI-type phage repressor; erf, recE, recT, single-strand
annealing proteins; ini, initiator; loader, DnaC-type helicase loader; res, RusA- and RuvC-type resolvases; ssb, single-strand DNA binding protein; dut,
dUTPase. Apparently truncated genes are shown in square brackets. ORFs with significant similarity ( 4 30% identical residues) are indicated by striped
colouring. Red-striped colouring indicates ORFs conserved within the group of Staphylococcus phages (a), green-striped colours indicate ORFs
conserved within the group of Streptococcus phages (b) and violet-striped colours indicate ORFs conserved within the group of Lactobacillus/
Lactococcus phages (c). ORFs with homologues in more than one subgroup are indicated by blue-striped colours. Dark and light grey colouring indicates
ORFs lacking homologues in any other phage genome. Dark and light grey colouring with black outline indicates ORFs with homologues in (completely
sequenced) phages other than compared here. Diagonal stripes indicate chromosomal genes flanking prophage genes. The ORF size is indicated by
block heigth:  100 residues = 1 U,  200 residues = 2 U,  300 residues = 3 U, etc. The relative positions of the ORFs in the phage genomes are
indicated by distances in kilobases (not to scale). Except where indicated by ‘ 4 ’ the direction of transcription is from up to down. The sequences were
taken from the genome entries for fSLT (42.9 kb) [NC_002661], fPV83 (45.6 kb) [NC_002486], f77 (41.7 kb) [NC_005356], fETA (43 kb)
[NC_003288], f13 (42.7 kb) [NC_004617], fEJ-1 (42.9 kb) [NC_005294], fSM1 (34.7 kb) [NC_004996], f7201 (35.5 kb) [NC_002185], fr1t
(33.4 kb) [NC_004302], fAT3 (39.2 kb) [NC_005893], fP335 (36.6 kb) [NC_004746], fbIL286 (41.8 kb) [NC_002667], fLL-H (7 kb, partial sequence)
[LLHSSB] and fmv4 (8.7 kb, partial sequence) [AF182207].

Subsequently, origin-bound gene 12 protein attracts the host The replication genes are surrounded by a dazzling
DnaG primase. fP22 can thus serve as a model for the IH- diversity of small ORFs of mostly unknown function. Many
type replication module that is found in a number of of these small ORFs are well conserved among several closely
Gram(  )- and Gram(1)-specific (pro)phages. related phages – fP22 and its ‘cousins’ serve as example here
For the initiator proteins of this group, a specific signa- – but their positions vary considerably. Other small ORFs
ture for initiators of IH-type modules could not be detected lack known homologues, and there are in fact no two phage
(COM section C3.1.2.). The helicases in the IH-type mod- genomes where the initiator and helicase genes are em-
ules are invariably DnaBEco-type helicases (F4 superfamily; bedded in an identical context. Remarkably, however, the
see COM section C3.3.). All phage genomes compared in replication genes are always found directly adjacent to each
Fig. 15 show a striking similarity in their gross architecture, other, the initiator gene upstream of the helicase gene. This
whether their replication module is located in the left (e.g. probably reflects a selective advantage of coupled genes
fP22, f11) or right half (e.g. fHK97, f3626). The archi- encoding interacting proteins in a genetic context that is
tecture closely resembles that of l (included for comparison subject to rearrangements by recombination.
in Fig. 15). In all cases, the integrase and phage repressor This conserved arrangement of the (putative) initiator
genes are located upstream of the initiator and helicase and helicase genes in mind, it should not be forgotten that
genes, and transcribed in the opposite direction. there is considerable variation among the replication genes

c 2006 Federation of European Microbiological Societies FEMS Microbiol Rev 30 (2006) 321–381
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 Bacteriophage replication modules 349

Fig. 15. ‘Initiator-helicase’ replication modules. Genes encoding replication and recombination functions are shown in their genomic context. l (IL-type
module) was included for better comparison. The alignment is shown with homologues/functional analogues of l O (initiator) at fixed positions. Blocks
with solid colours indicate (putative) gene functions detected by BLAST comparison: int, integrase; repr, l cI-type phage repressor; exo, reda-type
exonuclease; bet, redb-type single-strand annealing protein (SAP); erf, fP22 erf-type SAP; ini, initiator; loader, l P-type helicase loader; hel, F4-type
helicase; res, RusA- and RapA-type resolvases; ssb, single-strand DNA binding protein; dut, dUTPase. ORFs with significant similarity ( 4 30% identical
residues) are indicated by striped colouring. Dark and light grey colouring indicates ORFs lacking homologues in any other phage genome. Dark and
light grey colouring with black outline indicates ORFs with homologues in (completely sequenced) phages other than compared here. Diagonal stripes
indicate chromosomal genes flanking prophage genes. Homologues of f3626 orf50 (pink colour marked with an asterisk) were found in fPVL (orf63),
fBK5-T (orf63) (see Fig. 9) and fE125 (gp70) (see Fig. 12). The ORF size is indicated by block heigth:  100 residues = 1 U,  200 residues = 2 U,
 300 residues = 3 U, etc. The relative positions of the ORFs in the phage genomes are indicated by distances in kilobases (not to scale). Except where
indicated by ‘ 4 ’ the direction of transcription is from up to down. The sequences were taken from the genome entries for l (48.5 kb) [NC_001416], P.
putida pf (6.18 Mb) [NC_002947], fD3 (56.4 kb) [NC_002484], fP22 (41.7 kb) [NC_002371], fST104 (41.4 kb) [NC_005841], fVT2-Sa
(60.9 kb) [NC_000902], fHK620 (38.3 kb) [NC_002730], fHK97 (39.8 kb) [NC_002167], fHK022 (40.8 kb) [NC_002166], fSf6 (39 kb)
[NC_005344], fST64lT (40.7 kb) [NC_004348], f11 (43.6 kb) [NC_004615], f3626 (33.5 kb) [NC_003524], B. cereus pf (5.4 Mb) [NC_004722], B.
anthracis pf (5.2 Mb) [NC_003997]. (a) Gram(  )-specific (pro)phages; (b) Gram(1)-specific (pro)phages. In the lower part, grey blocks indicate
different degrees of similarity of the initiators and helicases as discussed in the text (see also Table 2); filled plus open boxes indicate the two
different dIR-type iterons and filled plus open circles the two different rep-type iterons in the (putative) phage replication origins, respectively, as
discussed in the text.

themselves, including variations of the (putative) replication termini, and the replication origins of both subgroups are
origins. The almost identical helicases of fP22, fHK97 and different (Table 2, Fig. 15).
fST104 are only distantly related to the helicases of the Exonuclease/SAP recombination genes in addition to the
other phages from the ‘P221cousins’ group, which also initiator/helicase gene pair are only found in the genomes of
form a subgroup of almost identical proteins (Table 2). the Gram(  )-specific phages, and should be included in
fP22 on one side and fHK971fST104 on the other side the definition of the IH-type module. The exonuclease/SAP
have initiators that are distinct enough from each other to gene pairs are invariably located between the integrase and
result in different replication origins (Table 2; see also Fig. phage repressor genes. fVT2-Sa encodes homologues of the
15). In addition, the initiators (and replication origins) of l reda/redb gene pair, while the other phages encode only an
phages fSf61fST64T are clearly different from the initia- Erf-like SAP. As an exception, fD3 encodes an exonuclease
tors of fHK620, fHK022 and fVT2-Sa. In this case, the together with an Erf-like SAP. The remarkable conservation
similarity of the initiators is largely confined to the C- of the position of these exonuclease/SAP gene pairs suggests

FEMS Microbiol Rev 30 (2006) 321–381

c 2006 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 350 C. Weigel & H. Seitz

Table 2. IH-type replication modules of Gram(  )-specific phages: similarities of initiators and helicases.

The values given are percentage identical residues obtained by BLAST bl2seq; Tatusova & Madden (1999) similarity searches. A dash indicates no
similarity detected. Boxed values indicate almost complete DNA sequence identity. Stippled boxes indicate significant DNA sequence similarity,
particularly in the 3 0 -part of the gene.

that the two ORFs at the corresponding position in the 1996). The related Rep1 genes of both plasmids contain the
fHK620 genome also encode recombination functions. replication origins, and the Rep1 proteins show the same
The genes encoding (putative) Holliday junction resol- domain structure as l O-type initiators, unusual for plasmid
vases are invariably located downstream of the initiator and initiators. The Rep2 gene downstream of Rep1 encodes a
helicase genes in the genomes of the Gram(  )-specific (putative) F4-type helicase, with Mycobacteriophage CJW1
phages, and at roughly corresponding positions. The resol- gp82 as closest homologue (29% identical residues, full
vases are of two types: RapA-like (e.g. fP22, fVT2-Sa, length).
fD3) and RusA-like proteins (e.g. fST64T, fHK97). This
suggests that both resolvase types are functionally equiva-
‘Initiator-helicase loader-helicase’ replication
lent. The rapA-like gene in the fSf6 genome is interrupted
modules
by insertion element IS911, and does probably not allow the
synthesis of a functional protein. This observation suggests The replication of B. subtilis phage SPP1 has been studied in
that a Holliday junction resolvase is not an essential protein detail (see above). The essential replication functions in-
for phage propagation, probably because backup functions clude the G38P initiator, the replication origin oriL located
exist in the host cell. within gene 38, the unique helicase loader G39P and the
fSf6, f11 and f3626 encode SSBs but there is no G40P helicase (DnaBEco-type). The replication of fSPP1
preferential position of the ssb gene, as observed also for DNA requires, in addition, as essential recombination func-
the other module types. A ssb gene is not an essential tions the G34.1P exonuclease and the G35P SAP proteins,
component of the IH-type replication module but should and the origin-like structure oriR (Ayora et al., 2002). Gene
be considered as an accessory replication gene. 45 may be a truncated version of a rusA-like gene, and a
The ‘IH-type’ modules of prophages in the genomes of Holliday junction resolvase has not yet been determined as
B. cereus and B. anthracis are highly homologous, located at essential for fSPP1 replication. In addition, the G36P SSB is
corresponding positions in the genomes of both species, and not essential under laboratory conditions. fSPP1 is most
flanked by orthologous host genes (Fig. 15). This suggests closely related to Listeria sp. fA118 (see Fig. 16). The
that a prophage was already present in the genome of the apparent differences are due to the fact that the lytic phage
ancestor of both species. However, the direct neighbour- SPP1 lacks a detectable repressor gene, and for the integrase
hood of both initiator and helicase genes shows that gene, only a putative remnant can be found at an unusual
prophage degradation proceeded differently after diver- position: gene 37 downstream of gene 36 (SSB). fSPP1 is
gence. the prototype Gram(1)-specific phage with an ILH-type
Relatively few phages are known from Streptomyces sp., replication module.
and the best studied case, phage C31, belongs to the group of A unique variant of the IHL-type replication module is
phages that encode DNA polymerases (see ‘Phages encoding present in pf315.1 in the Streptococcus pyogenes MGAS315
DNA-polymerases’ section). Interestingly, the Streptomyces genome: the helicase gene (locus tag SpyM3_0690) is located
plasmids pSLA2 and pSCL have replication modules, which directly upstream of the initiator-helicase loader gene pair.
suggest that they were originally derived from a bacterio- Because the gene context of this prophage closely resembles
phage(s) with IH-type replication module (Chang et al., the organisation of other phages, recombinatorial

c 2006 Federation of European Microbiological Societies FEMS Microbiol Rev 30 (2006) 321–381
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 Bacteriophage replication modules 351

shorter at the N-terminus than other helicase loaders of this

type, and its functioning uncertain.
Escherichia coli phage P27 is the prototype Gram(  )-
specific phage with a ILH-type replication module, and
encodes a helicase loader of the DnaCEco-type. The overall
architectures of fP27 and fSfV (I-solo type) are very
similar, and both phages share (limited) regions with almost
perfect DNA sequence homology (Fig. 16; compare grey
shaded regions). When the architecture of fP27 is com-
pared with that of fHK97 and fHK620, and with the
architecture of fSfV in addition, the emergent picture
suggests that all three phages are closely related despite their
different replication modules (Fig. 16). Because phages
fP27 and fSfV lack detectable DNA sequence similarity
directly downstream of the L19 helicase and the orf39
initiator, respectively, it is impossible to address the intri-
guing question of whether the ILH-type module of fP27
came about by acquisition of a helicase-loader gene (moron;
Hendrix et al., 2000) by an IH-type phage, or whether the I-
solo module of fSfV came about by loss of the
helicase loader/helicase gene pair (lesson) from an ILH-type
phage.
Two other (putative) ILH-type modules of prophages
Fig. 16. ‘Initiator-helicase loader-helicase’ replication modules. Genes
could be identified in the genomes of B. bronchiseptica RB50
encoding replication and recombination functions are shown in their
(Betaproteobacteria) and P. luminescens (Gammaproteobac-
genomic context. fA118 and fSfV (I-solo type module), as well as
fHK97 and fHK620 (IH-type module) were included for better compar- teria), respectively (Fig. 16). Neither of these prophages is
ison. The alignment is shown with homologues/functional analogues of closely related to fP27. The putative DnaCEco-type helicase
fP27 L17 (initiator) at fixed positions. Blocks with solid colours indicate loader (locus tag plu3472; 194 residues) of the P. luminescens
(putative) gene functions detected by BLAST comparison: int, integrase; prophage is considerably shorter at the N-terminus than
repr, l cI-type phage repressor; recE, recT, RecE-, RecT-type recombi- other helicase loaders of this type, and its functioning
nases; ini, initiator; loader, DnaC-type (dark green) or fSPP1 G39P (pale
uncertain. The putative helicase loader gene (locus tag
olive) helicase loader; hel, F4-type helicase; rusA, RusA-type resolvase;
bb2208; 123 residues) separating the initiator and helicase
ssb, single-strand DNA binding protein; terS, terL, terminase subunits.
ORFs with significant similarity ( 4 30% identical residues) are indicated genes in the B. bronchiseptica prophage lacks homologues in
by striped colouring. Dark and light grey colouring indicates ORFs lacking the databases. It may represent another novel type of
homologues in any other phage genome. Dark and light grey colouring helicase loader, because it is similar in size to the fSPP1
with black outline indicates ORFs with homologues in (completely G39P helicase loader, and too large to be considered simply a
sequenced) phages other than compared here. The ORF size is indicated recombination remnant.
by block height:  100 residues = 1 U,  200 residues = 2 U,  300
residues = 3 U, etc. The relative positions of the ORFs in the phage
genomes are indicated by distances in kilobases (not to scale). Except Conclusions for ‘Phages encoding initiator
where indicated by ‘ 4 ’ or by an arrow spanning several ORFs, the proteins’
direction of transcription is from up to down. The sequences were taken
from the genome entries for fHK97 (39.8 kb) [NC_002167], fHK620 We have described in the preceding paragraphs the replica-
(38.3 kb) [NC_002730], fSfV (37.1 kb) [NC_003444], fA118 (40.8 kb) tion modules of 40 bacteriophages with fully sequenced
[NC_003216], fSPP1 (44 kb) [NC_004166], Streptococcus pyogenes pf genomes and, in addition, 21 modules of several partially
(1.9 Mb) [NC_004070], fP27 (42.6 kb) [NC_003356], B. bronchiseptica
sequenced phage genomes and of various prophages. At the
RB50 pf (5.3 Mb) [NC_002927] and Photorhabdus luminescens pf
end of 2004, 220 bacteriophage genone sequences were
(5.68 Mb) [NC_005126]. (a) Gram(1)-specific (pro)phages; (b)
Gram(  )-specific (pro)phages. available in the databases, and the initiator-encoding phages
of the ‘lamdoid’ type clearly made up a considerable
percentage. With respect to statistical significance of the
sample, however, their number is still too low to allow for a
rearrangements due to the prophage state seem unlikely fair judgement as to whether one of the four module types
(Fig. 16). The putative DnaCEco-type helicase loader (locus has a particular selective advantage. In addition, there
tag SpyM3_0692; 167 residues) of pf315.1 is considerably certainly exists a strong bias in phage sampling, e.g. the

FEMS Microbiol Rev 30 (2006) 321–381

c2006 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 352 C. Weigel & H. Seitz

fermentation industries have always had a strong interest in (6) Genes encoding Holliday junction resolvases are located
understanding dairy phages, while interest in pathogenicity (almost) invariably downstream of the initiator gene. The
determinants strongly favoured the study of phages of encoded proteins may be of the RusA-, RapA- or RuvC-type.
Staphylococci and Streptococci. It is perhaps only circum- These proteins provide accessory functions because several
stantial that ILH-type replication modules are the least phage genomes lack a detectable resolvase gene. Given the
frequently found among Gram(1)- as well as Gram(  )- already known diversity of phage-encoded resolvases, differ-
specific phages. However, the number of 60 examples for ent novel types may be discovered in the future.
I-solo, IL- IH- and ILH-type replication modules is suffi- (7) ssb genes with similarity to E. coli ssb are found too
cient to derive a unified model for a virtual ‘lambdoid’ frequently in ‘lambdoid’ phage genomes to be neglected,
replication module: but not frequently enough to assume a more than
(1) The (putative) initiator genes always contain the accessory function for phage replication.
(putative) phage replication origin at a position that This definition of the virtual ‘lambdoid’ replication
corresponds to the linker region separating the N-term- module would have been impossible to derive without a
inal DNA-binding domain from the C-terminal oligo- thorough analysis of the occurrence and the positioning of
merisation/interaction domain in the initiator protein. the high number of mostly small ORFs that separate the
(2) In temperate phages, the integrase and phage repressor genes described above. Moreover, the positional conserva-
genes are invariably located upstream and transcribed tion of several of these small ORFs in phage pairs with
divergently from the initiator gene. Integrase and/or phage different replication modules (e.g. l/fVT2-Sa) was instru-
repressor genes may not be present (e.g. fSPP1) or inactive mental in developing the model that the four different
(e.g. fSf6) in the genomes of lytic phages, but the gene order module types are functionally equivalent although they
is comparable with that of the temperate phages. mediate the link of phage replication to the host replication
(3) Helicase loader genes are (almost) invariably located machinery at different stages of replisome formation.
directly downstream of the initiator genes within the same Phages of the fadh-type (see the subsequent section)
transcription unit. The genes may encode helicase loaders of encode entirely different replication proteins, including
the l P-type, the DnaCEco-type, the fSPP1 G39P-type or yet fP4a-type helicase-primases, but the gross architectures of
unknown types (e.g. fTuc2009). these phages closely resemble that of l, in particular with
(4) Helicase genes are (almost) invariably located di- respect to the localisation and orientation of the replication
rectly downstream of the initiator gene (IH-type module) genes to integrase and phage repressor genes. It seems
or the helicase loader gene (ILH-type module) within the possible that the definition of the virtual ‘lambdoid’ replica-
same transcription unit. Invariably, these genes encode tion module requires an even wider approach than given
F4 helicases of the DnaBEco-type. here.
(5) Many phages encode exonuclease/SAP recombination A unique organisation of replication genes is present in
genes, mostly in closely linked gene pairs. Although their E. coli phage P1. This phage is a rare example for a naturally
absolute requirement has been shown in some cases (e.g. occurring ‘joint replicon’: plasmid replication of the P1
fSPP1), these proteins are considered to have accessory prophage is driven by the R replicon, while replication
functions because a number of phage genomes lack detect- in the lytic cycle is driven by the L replicon (COM sec-
able exonuclease/SAP genes. Despite similarity on the tion C3.1.2.; Lobocka et al., 2004). Both replicons are
protein level, one can distinguish between the Reda-type separated physically, use different initiators and require a
and RecE-type exonucleases on the one hand, and the different subset of host replication proteins. With the RepL
Redb, RecT and Erf SAPs on the other hand. Gram(1)- initiator and the Ban helicase, the L replicon of fP1 would
and Gram(1)-specific phages can be distinguished by the formally fit to the ‘IH-type’ but we propose to exclude
preferential localisation of the exonuclease/SAP genes: it from this definition because these two replication genes
they are (almost) invariably located downstream of the are not part of the same transcription unit. In addition,
repressor gene and upstream of, and transcribed conver- the gross architecture of fP1 is hardly comparable with
gently to, the initiator gene in Gram(1)-specific phages, that of l.
but are not necessarily part of the same transcription unit An important lesson of our comparison of the different
(e.g. fSPP1). In Gram(  )-specific phages the exonu- types of ‘lambdoid’ replication modules may come from the
clease/SAP genes are located between the phage repressor observation that within a highly mosaic population an
and integrase genes and transcribed convergently with underlying pattern of similarity between individual mem-
them. The positional conservation of the exonuclease/ bers only becomes apparent when methods are applied that
SAP genes in either phage group further supports the allow the (quasi-)simultaneous visualisation of the entire
speculation that yet unknown recombination proteins data set. Conventional one-by-one comparison does not
may be identified in several phage genomes (e.g. fHK620). suffice.

c 2006 Federation of European Microbiological Societies FEMS Microbiol Rev 30 (2006) 321–381
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 Bacteriophage replication modules 353

Replication module exchange among genome size of 40 kb, a highly similar overall genome
bacteriophages architecture and share significant sequence similarity with
other S. aureus phages, namely fPVL, fSLT, fPV83 and
The phenomenon of genetic mosaicism was first detected fETA (Fig. 17). The ‘replication module’ of f11, f12, and
among close relatives of E. coli phage l, and has been studied f13 was localised between the ‘lysogeny module’ and the
most extensively for members of this group (Highton et al., ‘transcriptional regulation module’ in one half of the phage
1990). Among many instructive studies, the genetic analyses genomes (Iandolo et al., 2002). Within the ‘replication
of lreverse and laltSF show that infecting phages can pick modules’ Iandolo and coworkers found helicase, polymerase
up genes from cryptic prophages in the host genomes by and SSB encoding genes but the precise ends of the replica-
recombination, thus resulting in a replacement of their tion module could not be determined.
recombination/replication modules by a different type (Kai- The replication modules of the three phages are clearly
ser & Murray, 1979; Friedman et al., 1981). We were distinct, implying different replication mechanisms. Phages
particularly interested to study to what extent mosaicism f11 and f13 contain (putative) replication origins for yDR
also affects replication functions in phages that are not within their initiator genes (see ‘Phages encoding initiator
closely related to l (by DNA sequence). proteins’ section). Both phages probably replicate via yDR,
Iandolo et al. (2002) have studied the genome organisa- therefore. Concerning the replication module and its flank-
tion of the three S. aureus phages f11, f12 and f13 in ing regions, f12 is apparently a chimera of phages most
detail. The three temperate phages have a comparable closely related to fSLT on one side, and Bordetella sp. phage

Fig. 17. Replication module exchange: part A. Genes encoding replication and recombination functions are shown in their genomic context. For easier
comparison the alignment is shown with homologues of f12 p12 (DNA polymerase) and f11 gene 15 (initiator) at fixed positions. None of the proteins
has been analysed biochemically and their putative function assigned by BLAST similarity to known proteins. Blocks with solid colours indicate gene
functions: ini, initiator; hel loader, DnaC-type helicase loader; ssb, single-strand DNA binding protein; erf, fP22 Erf-type protein; roi, Roi-type phage
antirepressor; dut, dUTPase; int, phage integrase; repr, l cI-type phage repressor; DNA pol, DNA polymerase (Pol I-type); SF2 hel, superfamily 2 helicase;
F4 hel, F4 family helicase (DnaBEco-type); P4a hel, fP4a-type primase-helicase; recE, recT, rusA, ruvC, putative recombinases. ORFs with significant
similarity ( 4 30% identical residues) are indicated by striped colouring. Proteins with similarity to ORFs encoded by f12 are shown in red/yellow striped
colours; proteins lacking homologues in the f12 genome are shown in blue/green striped colours (see text for details). Dark and light grey colouring
indicates ORFs lacking homologues in any of the other phage genomes compared here. Light grey shaded blocks point to regions containing
homologous ORFs in phages f12 and fSLT, and in f12 and fBPP-1. The ORF size is indicated by block height:  100 residues = 1 U,  200
residues = 2 U,  300 residues = 3 U, etc. The relative positions of the ORFs in the phage genomes are indicated by distances in kilobases (not to scale).
Except where indicated by ‘ 4 ’ the direction of transcription is from up to down. The sequences were taken from the genome entries for fAPSE-1
(36.5 kb) [NC_000935], fBcepNazgul (58.1 kb) [NC_005091], fBPP-1 (42.5 kb) [NC_005357], f12 (45 kb) [NC_004616], fSLT (42.9 kb) [NC_002661],
f11 (43.6 kb) [NC_004615], fETA (43.1 kb) [NC_003288], f13 (42.7 kb) [NC_004617], fPVL (41.4 kb) [NC_002321], fEJ-1 (42.9 kb) [NC_005294]
and fPV83 (45.6 kb) [NC_002486].

FEMS Microbiol Rev 30 (2006) 321–381

c 2006 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 354 C. Weigel & H. Seitz

BPP-1 on the other (Fig. 17; compare shaded areas). The phage genomes both (putative) replication genes are em-
lack of a functional analogue of the (putative) primase- bedded in a highly similar gene context.
helicase protein encoded by fBPP-1 leads us to speculate On the nucleotide level, the comparison of the region
that f12 replicates by a mode similar to that known from encompassing genes 13–18 of phages f11 and f13, respec-
ColE1-type plasmids: a primary f12 transcript synthesised tively, suggests that genes 15 and 16 were replaced by a
by the host RNA polymerase is elongated by f12 DNA ‘cut1paste’-type recombination event in a common ances-
polymerase in a unidirectional strand-displacement reaction tor of both phages (Fig. 18a). However, when the region
until the host replication machinery is recruited to replicate encompassing ORFs 18–22 of fPV83 is also considered, the
fully the phage replicon, probably at a PAS site (del Solar emerging picture is considerably more complex: although
et al., 1998). We discuss in COM section C3.3. the lack of there is significant sequence homology between the 5 0 halves
experimental evidence for the participation of SF2-type of f11 gene 15 and fPV83 orf20, the homology between
helicases in replication. As with fBPP-1, fBcepNazgul and f13 gene 15 and fPV83 orf20 starts within the 3 0 -ends and
fAPSE-1, f12 also codes for a SF2-type helicase (p26). remains uninterrupted until the end of f13 gene 16 (Fig.
Intriguingly, a SF2-type helicase gene is also present in fSLT, 18b, c). This sequence patchwork results in partial protein
but in an apparently truncated form (116 residues; orf116a). sequence homologies: whereas the N-terminal 104 residues
Comparison of the fBPP-1 gene arrangement with that of of fPVL orf20 protein are identical (one mismatch) with
f12 suggests that independent recombination events led to f11 gene 15 protein, the C-terminal 14 residues are identical
the acquisition of the DNA polymerase and SF2 helicase with f13 gene 15 protein. The DNA and protein sequence
genes by the parent(s) of f12. However, fBPP-1 was similarity among the three proteins in the ‘middle’ part is
certainly not the direct source of these genes, as judged from low. Note that a DNA sequence patchwork can also be
the moderate degree of similarity of the proteins (32% observed for the upstream and downstream neighbouring
identical residues for DNA polymerase, 40% identical resi- genes. Therefore, we have to assume several successive
dues for SF2 helicase). recombination events among related phages to explain this
Yet another instructive example for replication module mosaicism, which makes it virtually impossible to trace the
exchange is provided by a comparison of phages f11, f13 exact descent of the individual genes.
and fPV83. We have shown in COM section C3.1.2. that the We derive confidence from the above observations that
proteins encoded by gene 15 of f11 and f13, respectively, the concept of phage replication modules has a molecular
and the orf20 protein of fPV83 are bona fide initiators for basis and is not just a useful theoretical tool for the
yDR, and contain the (putative) replication origins of these classification of the various replication module types of
phages. In all three phage genomes, a second (putative) phages. However, recombination does not necessarily occur
replication gene is located directly downstream of a initiator exactly between genes, but also at seemingly random points
gene: genes encoding a DnaCEco-type helicase loader in f13 within genes. Subsequent selection of functional recombi-
and fPV83 (IL-type replication module), and an F4-type nants then leads to the impression that entire modules are
helicase in f11 (IH-type module) (Fig. 17). In all three replaced (Hendrix et al., 2000). If this were true some useful

Fig. 18. Replication module exchange: part B. Dot-plot matrix analysis (nucleotide sequences) of the replication module plus adjacent regions of
phages f11, f13 and fPV83. f11, positions 8375–12 858, 4483 bp, genes 13–18 [NC_004615]; f13, positions 8133–12 214, 4081 bp, genes 13–18
[NC_004617]; fPV83, positions 8367–11 867, 3500 bp, ORFs 18–22 [NC_002486]. The dot-plots were obtained using ‘method 2’ (K_tuple value = 8) of
the dot matrix subprogram of the DNAMANs software (version 4.0; Lynnon Inc.). (a), f11  f13; (b), f11  fPV83; (c), f13  fPV83. The coloured
blocks indicate the individual proteins encoded by the regions of the phage genomes subjected to dot-plot analysis (exactly to scale). Colouring
corresponds to that in Fig. 17; orange, initiator; green, DnaC-type helicase loader; dark blue, F4-type helicase.

c 2006 Federation of European Microbiological Societies FEMS Microbiol Rev 30 (2006) 321–381
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 Bacteriophage replication modules 355

information could be derived from a careful elucidation of knowledge, however, it has not been rigorously examined
such mosaics. We have discussed in COM section C3.1.2. whether all phages can propagate efficiently in their staphy-
that many initiators seem to be composed of a N-terminal lococcal hosts in the absence of (helper) prophages provid-
DNA-binding domain and a C-terminal oligomerisation ing such additional replication functions. Although
domain, which can be interchanged to a certain extent. With expression of most prophage genes is repressed in lysogens
their homologous N-termini but only weakly homologous by the cognate repressor, this situation may change upon
C-termini the f11 gene 15 and fPV83 orf20 proteins belong infection by a hetero-immune phage (J. Iandolo, pers.
to this category. More importantly, however, the conserva- commun.).
tion of the extreme C-termini of fPV83 orf20 and f13 gene
15 proteins suggests that this part of the proteins is
responsible for the interaction with the DnaC-type helicase /P4a -type helicase-primase encoding
loader encoded by the downstream gene. This is reminiscent replication modules
of the situation observed for the extreme C-terminus of The fadh replication module was defined experimentally by
YdaU in the pfRac YdaUV IL-type replication module (see Henrich and co-workers, who found autonomous replica-
also COM Fig. C9). tion for plasmids carrying a particular 7 kb fadh fragment
In addition to the IL-type replication modules (fPV83, together with a selectable marker (Altermann et al., 1999).
fETA, f13, fSLT) and the IH-type (f11), also the ‘I-solo’ This fragment contained in addition to the putative replica-
type is found in this group of highly related phages: in tion origin (downstream of orf771) the presumed replica-
fPVL. Recombination and SSB genes are found in some but tion genes orf223 (ntp), orf455 (SF2-helicase), orf175 (SSB),
not all of the I/IL/IH-phages (Fig. 17). Formally, this orf771 (fP4a-type helicase) and four small ORFs encoding
suggests that these are accessory rather than essential func- polypeptides with unknown functions (indicated by an
tions for phage replication (see previous section). To our extended bracket in Fig. 19). Attempts to reduce the size of

Fig. 19. fP4a-type helicase-primase encoding replication modules. Genes encoding replication functions are shown in their genomic context. For
easier comparison the alignment is shown with homologues of fadh orf771 (fP4a-type helicase) at a fixed position. Blocks with solid colours indicate
(assigned) gene functions. ORFs with significant similarity ( 4 30% identical residues) are indicated by identical striped colouring. Dark and light grey
colouring indicates ORFs lacking homologues in any other phage genome. Dark and light grey colouring with black outline indicates ORFs with
homologues in (completely sequenced) phages other than compared here. The ORF size is indicated by block height:  100 residues = 1 U,  200
residues = 2 U,  300 residues = 3 U, etc. The relative positions of the ORFs in the phage genomes are indicated by distances in kilobases (not to scale).
Except where indicated by ‘ 4 ’ the direction of transcription is from up to down. The sequences were taken from the genome entries for fbIL310
(14.9 kb) [NC_002669], fSfi11 (39.8 kb) [NC_002214], fO1205 (43.1 kb) [NC_004303], fDT1 (34.8 kb) [NC_002072], fadh (43.8 kb) [NC_000896], f
-105 (39.3 kb) [NC_004167], fPSA/f2389 (37.7 kb) [NC_003291], fA2 (43.4 kb) [NC_004112], fAT3 (39.2 kb) [NC_005893], fBK5-T (40 kb)
[NC_002796], fTP901 1 (37.7 kb) [NC_002747], fr1t (33.4 kb) [NC_004302] and fg1e (42.3 kb) [NC_004305]. The partially known sequence of f31
(10.8 kb) was taken from entry AJ292531.

FEMS Microbiol Rev 30 (2006) 321–381

c 2006 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 356 C. Weigel & H. Seitz

this fragment were unsuccessful, probably because smaller

or partly overlapping fragments lacked appropriate tran-
scription signals, thus pointing to the limitations of the
experimental approach (Altermann et al., 1999). By a phage
replication interference approach, Moscoso & Suarez (2000)
located the (putative) fA2 replication origin at a position
corresponding to that of fadh, van Sinderen and co-work-
ers the replication origin of fO1205 (Stanley et al., 2000),
Klaenhammer and co-workers the replication origin of f31
(Madsen et al., 2001), and Brüssow and co-workers the
replication origins of fSfi19 and fSfi21 (relatives of fSfi11
Fig. 20. Putative primase domains of fP4a-type helicases. The sources
shown in Fig. 19) (Foley et al., 1998).
are indicated at the left side; gene names (locus tags) are shown under-
A comparison of the fadh replication module with the neath together with the sizes of the encoded proteins (not shown to size);
corresponding regions of phages f105, fPSA and fA2 the direction of transcription is indicated by arrowheads. The ‘%’ values
suggests that the small ORFs between fadh orf223 and (Bl2seq) shown give the percentage of identical residues for the com-
orf771 are not essential and should not be considered as pared regions of two proteins (Tatusova & Madden, 1999). Colour code:
parts of the fadh-type replication module (Fig. 19). The dark blue, fP4a-type helicase domain plus C-terminal DNA-binding
domain; pink/red/violet, different types of (putative) primase domains.
comparison of the flanking regions upstream of fadh
orf223 and downstream of orf771 with the corresponding
regions in the other three phages corroborates the experi- represent yet uncharacterised primase domains. Here we
mentally derived definition of the fadh replication module, extend this hypothesis, based on the results of BLAST
underlining the usefulness of combined experimental and searches, which identified two prophage-encoded N-term-
genomic analyses. inal domains of fP4a-type helicases with similarity to
Despite the lack of detectable DNA sequence similarity, the orf35 (Fig. 20). This makes it likely that also fDT1 orf35
highly similar gene order and the homology of the proteins encodes a yet uncharacterised primase. Phage-encoded
encoded by fA2 and fAT3 suggests that both phages have a DnaG-type primases occur either fused to a known helicase
common evolutionary origin (Fig. 19). It appears that the IL- domain as in fP4, fC31 and fT7, or as separate polypep-
type replication module of fAT3 and the orf35 protein of tides as, for example, in fKMV or fVpV262 (see below: the
fA2 are functionally equivalent for replication of the two fT7-type replication module). It appears that the putative
phages. Further examples of phages with partial similarities primases follow the same scheme: they occur fused to a
of their replication genes and flanking regions to fA2 genes fP4a-type helicase domain as in f105 or in the S. coelicolor
are fBK5-T, fTP901-1 and, to a lesser extent, fr1t (Fig. 19). prophage gene SCO5612, or as single polypeptides as in
In all these phages, we could detect a (putative) replication fDT1 (Fig. 20). Fusions of a DnaG-type primase to a
origin in the intergenic region downstream of the gene helicase are found for fP4a-type and F4-type helicases –
encoding the fP4a-type helicase (COM section C2.2.). but so far not for SF2-type helicases – and therefore seem to
Most probably, orf771 protein performs the same func- be a common theme among phage-encoded replicative
tion for fadh replication as the fP4 a-protein for fP4 helicases (see below). But despite these suggestive observa-
replication, i.e. a combined initiator-helicase function, but tions, the primase function of the fadh orf39 N-terminus
experimental evidence for this hypothesis is not available. and the fDT1 orf35 proteins have to be confirmed experi-
The C-termini of orf771 and fP4 a-protein are homolo- mentally.
gous, but the function of the extended N-terminal domain BLAST searches readily identify fadh orf223 protein as a
of orf771 is not known (COM sections C3.3. 1 C3.4.). The member of the AAA-family of NTPases but do not allow
set of (putative) replication proteins of fDT1 corresponds function to be predicted. Homologues of the fadh orf223
to that of fadh except that the fDT1 orf36 protein (504 protein are found in closely related replicon modules (e.g.
residues) lacks the extended N-terminus of its homologue, fA2), more distantly related replication modules (e.g.
fadh orf771 (771 residues). In addition, fDT1 orf35 up- fDT1), but also in unrelated modules (e.g. fBK5-T).
stream of orf36 (pale pink label in Fig. 19) encodes a protein However, an orf223 homologue is not encoded by f105,
that lacks a homologue in fadh. However, homologues of which suggests an accessory rather than essential function of
fDT1 orf35 are invariably found in those phage genomes orf223 protein for the fadh-type replication module.
that encode the shorter variant of the fP4a-type helicase, fadh orf455 encodes an SF2-type helicase with an as yet
i.e. fbIL310, fSfi11, f31 and fO1205. In COM Section unknown function for phage replication. The SSB encoded
C3.4., we speculate that the extended N-termini of fadh by fadh orf175 contains the characteristic acidic C-termi-
orf771 and its homologues in fA2, f105 and fPSA may nus of the fA2 orf34-group of SSBs. These ‘Group 3’ SSBs

c 2006 Federation of European Microbiological Societies FEMS Microbiol Rev 30 (2006) 321–381
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 Bacteriophage replication modules 357

are related to the chromosomally encoded SSBs of bacilli, known sequences of Aeromonas sp. phages f25 and f65
but probably not entirely interchangeable for functioning (Fig. 21). Because early and late replication of fT4 involves
during phage replication (COM section C3.6.1.). Homo- several and different replication origins, it is not possible to
logues of the fadh orf455 SF2-type helicase and orf175 SSB include origins in the module concept for this phage group
are part of the fadh-type replication module, but neither (Miller et al., 2003). With one exception, all genes encoding
protein is encoded by fbIL310 (Fig. 19). fbIL310 (15 kb) is the replication and recombination proteins discussed in the
probably a satellite phage that depends on a helper phage for ‘Recombination-dependent DNA replication’ section could
its propagation, as with E. coli phage P4 (Chopin et al., be localised in a comparable context: f44RR2.8t does not
2001). It is likely, therefore, that fbIL310 contains only the encode a homologue of the fT4 UvsX SAP, and a functional
basic components of an fadh-type replication module: analogue has yet to be identified.
replication origin1initiator function of the orf24 fP4a- There is a striking conservation of gene order, and
type helicase, and the (putative) orf23 primase. direction of transcription, in all phage genomes despite
To conclude this subject, we wish to point out that in highly varying numbers of ORFs separating the genes
some cases the ‘module concept’ does not promote a better encoding essential replication/recombination functions. In
understanding of replication gene assortments than mere some cases these intervening ORFs may be simply predic-
BLAST searches: tion artefacts and not actually expressed. For many of these
(1) fg1e encodes a homologue of the fadh orf223 NTP- ORFs, however, homologous sequences are found encoded
binding protein (Hel), an unusually large (putative) SSB (224 by at least one of the other phages at a different genome
residues; Rorf224) with similarity to f31 SSB, and a DnaC- position. This suggests that many of these ORFs are func-
type helicase loader (Ntp) similar to fr1t orf12 protein (Fig. tional although certainly not essential for phage propaga-
19) (Kodaira et al., 1997). No putative replication origin tion. It seems possible that the recurring recombination
structure could be detected in this segment of the fg1e events that are essential for replication of fT4-type phages
genome (not shown), and also a putative initiator gene could created these extensive mosaics, but without upsetting the
not be identified. The comparison of the regions upstream overall gene order. An example of this recombinatorial chaos
and downstream of these replication genes with the corre- is the fT4 alt gene, which is surrounded by one or more
sponding regions of other phages suggests that this is in fact copies of complete and/or partial duplications in fT4,
the fg1e replication module, but a possible molecular fRB69 and f44RR2.8t (see Fig. 21; orange label down-
mechanism for fg1e replication cannot be deduced from stream of gp30). The observation that in the fKVP40
this highly atypical assortment of replication genes. genome entire gene blocks have been apparently transposed
(2) The Streptomyces sp. phages C31 and BT1 encode fP4a- would be in line with this hypothesis.
type helicases, and the N-termini of these proteins are bona We suggest that the fT4-type replication module is
fide primase domains of the DnaG-type (COM section composed of two gene clusters with additional ‘orphan’
C3.4.), in contrast to fadh orf771. No other genes are genes (see Fig. 21): Cluster 1 contains the genes encoding (in
present in the fC31 and fBT1 genomes with similarity to order of trancription) the gp47 and gp46 recombination
the genes of the fadh-type replication module. Both phages proteins, the gp45 clamp, the gp44 and gp62 clamp loaders,
encode Pol I-type DNA polymerases (see Fig. 23). Even a the gp43 DNA polymerase, the UvsX SAP, the gp41 helicase
distant similarity of the replication modules of these two and the gp61 primase. The size of this cluster ranges from
phages with the fadh-type replication module is hardly 15 kb (fAeh1) to 18.5 kb (fT4), depending on the num-
detectable, despite the fP4a-type helicase common to both. ber and sizes of intervening ORFs. Cluster 2 contains the
(3) The mycobacteriophage Barnyard encodes a fP4a-type genes encoding the 5 0 ! 3 0 exonuclease Rnh, the gp59
helicase, and an SF2-type helicase. There is nevertheless no helicase loader and the gp32 SSB. The size of this cluster is
detectable similarity to the fadh-type replication module: about 3.8 kb. The genes encoding DNA ligase (gp30), UvsW
fBarnyard encodes a Pol III-type DNA polymerase, and all (SF2 helicase) and endonuclease VII (gp49) are located at
replication genes are found at rather large distances from corresponding positions in all genomes, except for fKVP40,
each other. but not in a larger context of other replication/recombina-
tion genes. This is also observed for Dda (SF1 helicase),
which has a conserved position in the genomes of four of the
Phages-encoding DNA polymerases
six fully sequenced genomes. The conservation of the gene
order in both gene clusters supports the identification of
The phage T4-type replication module
gene function by BLAST comparisons also for those proteins
In order to identify a possible fT4-type replication module, that have not been characterised biochemically.
we aligned the genomes of phages fT4, fRB69, fRB49, Krisch and co-workers proposed a division of the fT4-
f44RR2.8t, fAeh1 and fKVP40, also including the partially type phage group into three subgroups based on a sequence

FEMS Microbiol Rev 30 (2006) 321–381

c2006 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 358 C. Weigel & H. Seitz

Fig. 21. The fT4-type replication module. Genes

encoding replication and (most) recombination
functions are shown in their genomic context. For
the definition of ‘cluster 1’ and ‘cluster 2’ see text.
For easier comparison the alignment is shown
with homologues of gp43 (DNA polymerase) at a
fixed position. ORFs with significant similarity
( 4 30% identical residues) are indicated by iden-
tical colouring. Dark and light grey colouring
indicates ORFs lacking homologues in any of the
other phage genomes compared here. White
blocks in the fRM378 genome indicate ORFs with
similarity to ORFs of (at least) one of the other
phages, but not shown in this alignment. The ORF
size is indicated by block height:  100 re-
sidues = 1 U,  200 residues = 2 U,  300 re-
sidues = 3 U, etc. The relative positions of the ORFs
in the phage genomes are indicated by distances
in kilobases (not to scale). Except where indicated
by ‘ 4 ’ the direction of transcription is from up to
down. The sequences were taken from the gen-
ome entries for fT4 (169 kb) [NC_000866],
fRB69 (167.5 kb) [NC_004928], fRB49 (164 kb)
[NC_005066], f44RR2.8t (173.6 kb)
[NC_005135], fAeh1 (233.2 kb) [NC_005260]
and fKVP40 (244.8 kb) [NC_005083]. The par-
tially known sequences for f25 (7.5 kb) and f65
(25.2 kb) were taken from entries AY497556 and
AY303350, respectively.

comparison of the major head and tail proteins: T-even genome. A homologue of the fT4 UvsX gene was not detected,
phages (fT4, fRB69), pseudo T-even phages (f44RR2.8t, but p018 encodes a fP22 Erf-like SAP (COM section C3.6.2).
fRB49) and schizo T-even phages (f65, fKVP40, fAeh1) Genes encoding a helicase loader, a SSB and clamp1clamp
(Tétart et al., 2001). The BLAST similarities of the replica- loader proteins with similarity to the fT4-type proteins were
tion proteins discussed in COM section C3. support this not detected. Despite the presence of several ORFs with
subgrouping. However, the conserved direction of transcrip- similarity to predicted or known proteins of fT4-type phages,
tion and the conserved gene order in both clusters of neither the ORFs flanking the replication/recombination genes
replication/recombination genes of all eight phages, except nor the gene order in fRM378 suggest a relationship to T-even
for fKVP40, justify the proposal of a ‘common’ replication phages. Clearly, several more phages related to fRM378 would
module for the fT4-group of phages. have to be isolated and sequenced before conclusions concern-
The fT4-type gene replication gene clusters could not be ing the relationship of this phage to the enterobacterial T-even
identified unambiguously in the fRM378 genome (Fig. 21). In phages can be made.
addition, this phage and its host Rhodothermus marinus
(Bacteroidetes/Chlorobium group) are poorly characterised
microbiologically. The p092 DNA polymerase of fRM378 The phage T7-type replication module
belongs to the Pol II-type DNA polymerases but lacks the
canonical 30 ! 50 exonuclease domain, which is encoded by a T7 is the prototype of the ‘T-odd’ group of E. coli phages,
separate gene (locus tag p024). Like the corresponding fT4 following the traditional nomenclature. This group is ‘odd’
proteins, the fRM378 p019 helicase is a member of the F4- in several aspects: phages fT1 and fT5 have a genome
family, and the p101 primase a member of the DnaG-type architecture that deviates significantly from that of fT7, and
family. In addition, a 50 ! 3 0 exonuclease gene (p012) and a encode only very few ORFs with limited similarity to fT7
SF2-type helicase gene (p104) are present in the fRM378 proteins – they certainly do not belong to this group. In

c 2006 Federation of European Microbiological Societies FEMS Microbiol Rev 30 (2006) 321–381
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 Bacteriophage replication modules 359

addition, half of the fXp10 genome codes for structural and Figure 22 shows a set of 15 phage genomes, which
host lysis proteins that are related to l in an arrangement resemble fT7 with respect to types and arrangement of
typical for lactococcal phages (Yuzenkova et al., 2003). their replication genes. The alignment shows the consider-
However, the replication proteins of fT7 are among able variation in number and size of small intervening ORFs
the best understood examples of bacteriophage replication with (mostly) unknown functions even for closely related
proteins, and we will refer to the ‘fT7-group’ in the phage pairs with identical gene order of their replication
following for phages with a comparable set of replication genes (e.g. fT7/fA1122, fT3/fYeO3-12, fK1-5/fSP6).
genes, irrespective of their classification by systematic Also, there is no conservation of the order of the replication
criteria. genes, although BLAST analysis suggested their (near)
Replication of fT7 in vitro requires the assembly of the homology (COM section C3.). To ‘distil’ the fT7 replication
phage-encoded proteins SSB (gene 2.5), primase-helicase module from this complex picture we have to discuss the
(gene 4A and 4B) and DNA polymerase (gene 5) at a replication proteins individually.
preformed R- or D-loop, in addition to host thioredoxin as The DNA polymerases encoded by all phages belong to the
processivity factor. In vivo, host RNA polymerase transcribes T7 gene 5 subfamily of Pol I-type DNA polymerases, except
the fT7 RNA polymerase gene (gene 1), and fT7 RNA those of fXP10, fKMV and fVpV262. The latter are Pol I-
polymerase subsequently transcribes the fT7 genome from a type DNA polymerases but they lack the subfamily-specific
(known) set of highly specific promoters. R-loops formed by residues within the signature motifs. The fPaP3 DNA
these transcripts serve as assembly sites for the fT7 replisome. polymerase may be composed of two separate polypeptides:
In addition to the cognate RNA polymerase, fT7 codes for a p39 contains the 3 0 ! 5 0 exonuclease and DNA polymerase
cognate DNA ligase (gene 1.3), a cognate 5 0 ! 3 0 exonuclease signature motifs; the function of p32 protein, which is
(gene 6) and a Holliday-junction resolvase (gene 3). similar to the fT7 gene 5 N-terminus, remains to be studied.

Fig. 22. The fT7-type replication module. Genes encoding replication and recombination functions are shown in their genomic context. For easier
comparison the alignment is shown with homologues of fT7 gene 4A (DNA primase-helicase) at a fixed position. ORFs with significant similarity
( 4 30% identical residues) are indicated by identical colouring. Dark and light grey colouring indicates ORFs lacking homologues in any of the other
phage genomes compared here. The ORF size is indicated by block height:  100 residues = 1 U,  200 residues = 2 U,  300 residues = 3 U, etc. The
relative positions of the ORFs in the phage genomes are indicated by distances in kilobases (not to scale). Except where indicated by ‘ 4 ’ the direction of
transcription is from up to down. The sequences were taken from the genome entries for fT7 (39.9 kb) [NC_001604], fA1122 (37.5 kb) [NC_004777],
fT3 (38.2 kb) [NC_003298], fYeO3-12 (39.6 kb) [NC_001271], fgh-1 (37.4 kb) [NC_004665], fSP6 (43.8 kb) [NC_004831], fK1-5 (44.4 kb)
[AY370674], fP60 (47.9 kb) [NC_003390], fSIO1 (39.9 kb) [NC_002519], fFelix01 (86.2 kb) [NC_005282], fXp10 (44.4 kb) [NC_004902], fPaP3
(45.5 kb) [NC_004466], fKMV (42.5 kb) [NC_005045], fVpV262 (46.0 kb) [NC_003907] and pf P. putida strain KT2440 (positions
2 586 633–2 597 674) [NC_002947].

FEMS Microbiol Rev 30 (2006) 321–381

c 2006 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 360 C. Weigel & H. Seitz

Thirteen of the 15 phages encode a fT7 gene 4A-type speculate that this protein serves as processivity factor for the
primase-helicase. In the fXP10 and fKMV genomes, the DNA polymerase, in analogy to E. coli thioredoxin for fT7
primase genes are located directly upstream of the helicase gene 5 DNA polymerase. It is not known why fKMV and
genes – a split helicase gene in the case of fXP10 (see COM fXp10 code for a second 3 0 ! 50 exonuclease, because their
Table C16). The primase gene of fVpV262 is located DNA polymerases (gp19 and p39, respectively) already contain
upstream of the helicase gene but separated by a small the typical 30 ! 50 exonuclease domains within their Pol I-
intervening ORF. The latter three examples demonstrate type DNA polymerases. The fKMV orf24 and fXp10 p35
that although a fused primase-helicase is prevalent among 3 0 ! 50 exonucleases are similar to each other (44% identical
the fT7-type replication modules, different gene-arrange- residues) but have no detectable similarity to the 3 0 ! 50
ments are possible. exonuclease domains of their Pol I-type DNA polymerases.
Proteins with similarity to the fT7 gene 2.5 SSB are From the above, we conclude that the ‘basic’ fT7-type
encoded by six of the 15 phages only. Given its essential role replication module is composed of five genes: (1) a gene
for fT7 replication this is somewhat surprising. It is not encoding a Pol I-type DNA polymerase lacking the 5 0 ! 3 0
known whether the phages that lack a cognate SSB gene exonuclease domain of E. coli PolA, (2) a gene encoding a
recruit host SSB for their replication or whether they encode 5 0 ! 3 0 exonuclease, (3) a gene encoding a DnaGEco-type
yet unknown SSB proteins. By simple eye-screening, how- primase, (4) a gene encoding a DnaBEco-type helicase (F4
ever, we could not detect ORFs with the characteristic of family) and (5) a gene encoding an RNA polymerase. These
SSBs, e.g. a highly acidic C-terminus. We propose that a five genes are arranged in a gene cluster 15 kb in length in
cognate SSB is an accessory rather than essential component one half of the phage genome in most cases, arranged in the
of the fT7-type replication module. same direction of transcription, but with highly varying
Homologues of the 5 0 ! 3 0 exonuclease encoded by fT7 numbers of intervening ORFs. But, in contrast to the fT4-
gene 6 are found in all phage genomes, except fSIO1. We type replication module, no prediction can be made about
therefore consider this gene as part of the fT7-type replica- the gene order because all permutations are observed, the
tion module. only conserved feature being the gene order: primase-heli-
Ten of the 15 phages encode their own RNA polymerase case. In principle, this basic set of replication proteins would
(fT7 gene 1). Because we were unable to identify in the be able to initiate and drive replication of the phage genome
remaining phages genes encoding replication proteins in- by the mechanism known from fT7. Accessory functions
dicative of an alternative replication mode, we speculate that would, when present, increase replication specificity, i.e.
these phages use unknown mechanisms to redirect the host render the phage replicon less dependent on the host
RNA polymerase to their own promoters, e.g. a specific s transcription, recombination and replication machinery.
factor. Until this question is answered experimentally we
tend to consider the RNA polymerase gene as part of the
The phage D29-type replication module
fT7-type replication module.
Cognate DNA ligases (fT7 gene 1.3), endonucleases The temperate mycobacteriophages fD29, fL5, fBxz2,
(fT7 gene 3) or proteins similar to fT4 endonuclease VII fBxb1, fRosebush and fPG1, as well as the Streptomyces
are encoded only by a subset of the phages compared here. sp. phages fBT1 and fC31 encode a DNA polymerase with
These genes are therefore best classified as accessory func- significant similarity to fD29 gp44 (COM section C3.5.).
tions for the basic fT7-type replication module. For none of the phages has the replication mechanism been
In two instances, we could identify small proteins with studied in detail. David et al. (1992) reported that cloning of
significant similarity to proteins encoded by other phages of a 2.6-kb PstI fragment from fD29 into a selectable plasmid
the set compared here: (1) the fFelix01 p181 protein (266 resulted in efficient and stable transformation of Mycobac-
residues) is considerably shorter than fT7 gene 1 RNA terium smegmatis. They concluded that this fragment carries
polymerase (883 residues) and therefore probably nonfunc- the fD29 replication origin. Because this fragment carries
tional, and (2) the fVpV262 p21 protein (85 residues) is the intact fD29 gp33 (putative) integrase gene together
probably an N-terminally truncated SSB and nonfunctional. with the att (attachment) site, the observed plasmid stability
From our type of analysis it is impossible to decide whether could have been also due to efficient integration into the M.
these apparently truncated genes represent unsuccessful at- smegmatis chromosome. In order to identify a possible
tempts at gene aquisition (‘moron’) or remnants of a pre- fD29-type replication module, and replication mechanism,
viously complete gene (‘lesson’) (Hendrix et al., 2000). we aligned the genomes of the phages for the region flanking
Cyanophage P60 encodes a small protein (87 residues; locus the DNA polymerase genes (Fig. 23).
tag P60_19) with significant similarity to C3-type thioredox- We exclude phages fBT1 and fC31 from the further
ins. As this gene is located between the genes encoding the discussion of a ‘fD29-type replication module’ for two
primase-helicase and DNA polymerase, it is attractive to reasons. The functional equivalent of the primase and

c 2006 Federation of European Microbiological Societies FEMS Microbiol Rev 30 (2006) 321–381
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 Bacteriophage replication modules 361

with other phages encoding fP4a-type primase–helicase

(see above).
For fPG1, but not for fRosebush, we could detect two
smaller ORFs with homologues in fD29. In addition, these
two phages lack detectable endonuclease VII and exonu-
clease genes, and they encode a SF2-type helicase, in contrast
to fD29. It is not known whether the very large F4-type
helicases of these phages contain an N-terminal primase
domain (COM section C3.4.). Although it may seem some-
what arbitrary, we exclude these two phages from the
discussion of the ‘fD29-type replication module’.
The remaining four phages are closely related, even on the
DNA sequence level (Bruessow & Desiere, 2001) (see Fig.
23). Other phage replicons with a split primase gene are not
known, and all four encode a particularly small version of an
F4-type helicase (COM sections C3.3. 1 C3.4.). When the
‘fD29-type replication module’ is examined for the genes
flanking the replication/recombination genes, it appears that
the fBxb1 genome contains the ‘minimal version’: the
region from the gp41 DNA polymerase gene to the gp62
exonuclease gene spans 10.2 kb, in contrast to 15.5 kb in the
Fig. 23. The fD29-type replication module. Genes encoding replication fD29 genome. In all four phage genomes, the gene order of
and recombination functions are shown in their genomic context. For
the replication/recombination genes and the direction of
easier comparison the alignment is shown with homologues of fD29
their transcription are conserved.
gp57 (DNA primase) at a fixed position. None of the proteins has been
analysed biochemically and their putative function assigned by BLAST When the set of replication/recombination genes of fD29
similarity to known proteins. Blocks with solid colours indicate gene is compared with the fT7-type replication module, a
functions: DNA pol, DNA polymerase (Pol I-type); pri N-term, primase, N- striking similarity becomes apparent. Both contain a Pol I-
terminus; pri C-term, primase, C-terminus; endo VII, endonuclease VII; type DNA polymerase, a DnaGEco-type primase and an F4-
exo, RecB-type exonuclease; F4 hel, F4 family helicase (DnaBEco-type); family helicase. In addition, the ‘fD29-type replication
P4a hel, fP4a-type primase-helicase; SF2 hel, superfamily 2 helicase;
module’ contains an endonuclease VII with similarity to
red, ribonucleotide reductase. ORFs with significant similarity ( 4 30%
the fXp10 p36 protein, and a (putative) ribonucleotide
identical residues) are indicated by striped colouring. Proteins with
similarity to ORFs encoded by fD29 are shown in blue/yellow striped reductase with similarity to the fSIO1 p21 protein – the
colours; proteins with similarity to ORFs encoded by fPG1 are shown in latter not being a replication protein in the strict sense.
red/violet striped colours; proteins with similarity to ORFs encoded by However, fD29 lacks a cognate RNA polymerase and SSB,
fC31 are shown in green striped colours. Dark and light grey colouring and the (putative) 5 0 ! 3 0 exonuclease is clearly nonhomo-
indicates ORFs lacking homologues in any of the other phage genomes logous, being more related to RecB-type exonuleases [COG
compared here. The ORF size is indicated by block height:  100
2887]. There is thus no reason to differentiate between a
residues = 1 U,  200 residues = 2 U,  300 residues = 3 U, etc. The
‘fD29-type replication module’ and a ‘fT7-type replication
relative positions of the ORFs in the phage genomes are indicated by
distances in kilobases (not to scale). Except where indicated by ‘ 4 ’ the module’, and we propose to include the former in a more
direction of transcription is from up to down. The sequences were taken relaxed definition of the latter. Clearly, the replication
from the genome entries for fD29 (49.1 kb) [NC_001900], fL5 (52.3 kb) mechanism of fD29 has to be analysed experimentally to
[NC_001335], fBxz2 (50.9 kb) [NC_004682], fBxb1 (50.5 kb) justify this classification.
[NC_002656], fBT1 (41.8 kb) [NC_004664], fC31 (41.2 kb)
[NC_001978] and fRosebush (67.5 kb) [NC_004684].
The replication modules of the phages K, Bxz1
and T5
helicase of fD29 are the fP4a-type primase-helicase pro- Escherichia coli phage T5, mycobacteriophage Bxz1 and
teins of these two phages. This suggests that fBT1 and fC31 Staphylococcus sp. phage K are among the relatively few
replicate by initiator-dependent yDR, but this remains to be (known) phages with large genomes ( 4 100 kb) that are,
confirmed. In addition, fBT1 and fC31 do not encode a with respect to the types of replication proteins they encode,
(detectable) endonuclease VII and exonuclease, and also no not closely related to the T-even phages. Like the fT4-type
ORFs with similarity to any of the smaller ORFs of fD29 phages, however, all three encode DNA polymerases, pri-
were found. Therefore, both phages are discussed together mases, and F4- as well as SF2-helicases. The observation by

FEMS Microbiol Rev 30 (2006) 321–381

c 2006 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 362 C. Weigel & H. Seitz

electron microscopy of multiple origins and branched con-

catemeric structures of fT5 replication intermediates
suggested a replication mechanism similar to fT4 (Bour-
guignon et al., 1976). The replication mechanisms of fK
and fBxz1 are not known.
In the fT5, fK and fBxz1 genomes the replication genes
are arranged in a 15 kb (fT5, fBxz1) to 25 kb (fK)
segment together with (putative) recombination genes and
with ORFs of unknown function (Fig. 24). fT5 and fK
encode Pol I-type DNA polymerases (the fK sequence
became available after completion of COM section C3.5.).
Unlike most other Pol I-type polymerases the fT5 DNA
polymerase does not require additional factors for proces-
sivity. The processivity factor requirements of the fK orf86/
88/90 DNA polymerase are unknown. fBxz1 encodes a Pol
III-type DNA polymerase. Genes encoding the proofreading
activity (3 0 ! 5 0 exonuclease) and the 5 0 ! 3 0 exonuclease
have not yet been identified in the fBxz1 genome. Given the
high degree of conservation of these proteins it seems likely
that fBxz1 recruits the respective host proteins for replica-
tion. In addition, genes encoding DNA polymerase acces-
sory proteins could not be identified in the fBxz1 genome
(Pedulla et al., 2003). It is presently not known whether the Fig. 24. The fK, fT5 and fBxz1 replication modules. Genes encoding
three polymerases perform strand-displacement synthesis or replication and recombination functions are shown in their genomic
assemble into dimeric replisomes for coupled leading- and context. Few of the proteins have been analysed biochemically; in most
lagging-strand synthesis. All three phages encode F4-type cases, their putative function could be assigned by BLAST similarity to
(replicative) helicases and DnaG-type primases. fBxz1 known proteins. Gene names/locus tags are shown below the (assigned)
gp192 located upstream of the gp193 helicase gene encodes functions. Dark and light grey colouring indicates ORFs lacking homo-
logues in any of the other phage genomes compared here. The ORF size
a DnaC-type helicase loader, an arrangement also found in
is indicated by block height:  100 residues = 1 U,  200 re-
fP27. In addition to the F4-type helicases, all three phages sidues = 2 U,  300 residues = 3 U, etc. The relative positions of the ORFs
encode SF2-type helicases. The second SF2-type helicase of in the phage genomes are indicated by distances in kilobases (not to
fT5 (locus tag T5.108) shows significant similarity to the scale). Except where indicated by ‘ 4 ’ the direction of transcription is
UL9 helicase involved in replication initiation of Herpes from up to down. The sequences were taken from the genome entries
simplex virus 1 (HSV1) but its role for fT5 replication is for fK (127.4 kb) [NC_005880], fT5 (121.8 kb) [NC_005859] and fBxz1
unknown. As noted in COM section C3.3., the presence of (156.1 kb) [NC_004687].

SF2-type helicase genes in so many phage replicons – in

many cases located within the replication modules – makes past. Although D5 binds to ssDNA, it binds with higher
it necessary to understand the role of this helicase for the affinity to dsDNA, which it covers stoichiometrically. D5 is
replication process better. fK orf24 and fT5 rnh encode the required for the regulation of late transcription in fT5, but
cognate 5 0 ! 3 0 exonucleases for primer removal. Both a role for replication has also been proposed. The DNA
phages encode DNA ligases, although the assignment of fK binding properties of D5 make a role for this protein as
orf21 as DNA ligase is questioned by the observation that replication initiator unlikely. D5 actually inhibits fT5 DNA
this protein is similar to the fRB69 RnlB RNA ligase. A polymerase in vitro possibly through direct interaction
putative ligase gene was not detected in the fBxz1 genome. (Fujimura & Roop, 1983). Interestingly, none of the three
fK orf70 encodes a protein for which the assignment phages encodes a (detectable) SSB. However, there may exist
‘putative Rep protein’ was chosen, but experimental results more than the known SSB types (COM section C3.6.1.).
confirming this role are not available (O’Flaherty et al., The replication modules of fBxz1, fK and fT5 provide
2004). Rep has a (predicted) helix–turn–helix (HTH) motif further examples for the co-localisation of (putative) re-
in its N-terminus but shows no similarity to known phage combination genes and replication genes in phage replicons.
or plasmid initiators. We could detect a pronounced AT- In most cases a direct participation of cognate recombina-
peak approximately in the middle of the orf70 gene but tion proteins in phage replication has not been demon-
iterons could not be identified. fT5 encodes the DNA- strated. Their exact enzymatic function remains uncertain
binding protein D5 that has attracted some attention in the and their assignment is based mostly on BLAST similarities.

c 2006 Federation of European Microbiological Societies FEMS Microbiol Rev 30 (2006) 321–381
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 Bacteriophage replication modules 363

However, the clustering of recombination and replication

genes suggests that mechanisms related to RDR are operat-
ing during replication of most phages. Although the simi-
larity of fK orf93 protein to RecA proteins is weak, fBxz1
gp201 is one of the rare examples of a phage-encoded RecA
homologue (COM section C3.6.2.). Whether fK orf94
protein is a (putative) s-factor (O’Flaherty et al., 2004) or
a member of the RecJ-family of recombination proteins as
also suggested by BLAST searches remains to be seen. fBxz1
gp205 shows all signature motifs of a RusA-type Holliday
junction resolvase, and also the fK orf78 and fT5 D14
proteins give BLAST hits with Holliday junction resolvases.
fK and fT5 encode proteins with significant similarity to
the SbcDC-type recombinases. fT5 encodes, in addition,
the D15 exonuclease with high affinity for fork structures.
From the phylogenetic viewpoint fK, fT5 and fBxz1
are, at best, only distantly related to each other, and also only
distantly related to the T-even phages despite the compar- Fig. 25. The replication module of the f29-type phages. Genes encod-
able genome sizes. In addition, their hosts belong to ing replication functions are shown in their genomic context. ORFs
different branches of the bacterial kingdom. However, all encoding proteins with known replication functions are shown in solid
three phages possess highly similar replication modules with colour. Striped colours indicate similar proteins in the phages. Dark and
respect to the set of replication/recombination genes they light grey colouring indicates ORFs lacking homologues in any other
phage genome. Dark and light grey colouring with black outline
encode. If only the replication module is considered, a
indicates ORFs with homologues in (completely sequenced) phages other
hypothetical common ancestor of all DNA polymerase- than compared here. The ORF size is indicated by block height:  100
encoding phage replicons could be envisaged that carried residues = 1 U,  200 residues = 2 U,  300 residues = 3 U, etc. The
genes for a primase, a replicative helicase and a DNA relative positions of the ORFs in the phage genomes are indicated by
polymerase. The highly differentiated types of extant repli- distances in kilobases (not to scale). Except where indicated by ‘ 4 ’ the
cons would then reflect gene replacement, e.g. in the case of direction of transcription is from up to down. The sequences were taken
the different DNA polymerase types, and acquisition of from: f29 (19.4 kb) [NC_001423], fB103 (18.6 kb) [NC_004165], fGA-
1 (21.1 kb) [NC_002649], fCp-1 (19.3 kb) [NC_001825], fBam35C
additional genes during evolution, e.g. genes encoding
(14.9 kb) [NC_005258], fPR772 (14.9 kb) [AY441783], fPRD1 (14.9 kb)
processivity factors, RNA polymerase genes and ssb genes. [NC_001421], f44HJD (16.8 kb) [NC_004678], f68 (18.2 kb)
Alternatively, we could assume that the fortuitous co- [NC_004679], fC1 (16.7 kb) [NC_004814] and fP1 (11.7 kb)
localisation of a DNA polymerase gene of any of the three [NC_002515] (not to be confused with Escherichia coli phage P1).
known polymerase types and an F4- or fP4a-type primase-
helicase on a DNA string creates the potential for its
ing these replication factors constitute the replication mod-
autonomous replication. According to the latter model,
ule of f29 together with the ends of its linear genome. We
DNA polymerase-encoding phage replicons could have
could detect 10 proteins with similarity to the f29 DNA
evolved independently several times during evolution. Also
polymerase that are encoded by phages with linear dsDNA
in this model, the different types of extant replicons would
genomes, ranging in size from 12 to 21 kb (COM section
reflect the acquisition of additional genes and their eventual
C3.5.1.). It is, however, difficult to trace the other replication
exchange by recombination. Whether the over-simplified
proteins of f29 in the entire set of 11 phages (Fig. 25). The
version of a tree-like phylogeny of phage replicons is more
TP gene can be detected in the majority of the sequences,
inspiring for future research than the concept of a web-like
but homologues of the f29 DPB gene could only be detected
phylogeny remains to be seen.
in fB103 and fGA-1. Therefore, we cannot discuss the
possible variations of the ppDR mechanism driving f29
The phage f29-type replication module replication in detail.
As discussed in detail in the ‘Initiation at the ends of linear
DNA: protein–primed DNA replication’, ppDR of the B. Replication modules of phages replicating by
subtilis phage f29 requires the cognate DNA polymerase RCR
(gene 2 protein), the TP (gene 3 protein), the cognate SSB
(gene 5A) and the DNA-binding protein (DBP; gene 6 The replication modules of the phages that replicate by RCR
protein) (reviewed in Meijer et al., 2001). The genes encod- are composed of: (1) an initiator gene, (2) a double-strand

FEMS Microbiol Rev 30 (2006) 321–381

c 2006 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 364 C. Weigel & H. Seitz

origin (dso) and (3) a single-strand origin (sso) (see ‘Initia- Omega remains enigmatic, and the ‘incomplete’ sets of
tion by nicking: ‘rolling circle’-type DNA replication’ sec- replication genes in the phages Barnyard and Rosebush do
tion). not provide a clue at present (Pedulla et al., 2003). Probably
We can distinguish four groups of initiator proteins, the most intriguing ‘white spot’ is the Pseudomonas aerugi-
which, despite poor overall similarity, have the conserved nosa phage KZ with the largest phage genome known so far
‘active tyrosine’ motif 3 in common and perform identical (280 kb) (Mesyanzhinov et al., 2002).
functions for the replication of their cognate replicons. The The identification of the replication origins of lactococcal
initiators of fX174, ffd, fP2 and CTXf have been studied phages belonging to the groups of fbIL67-, f923- and fc2-
in detail (COM section C3.1.1.). type phages underlines the importance of established ex-
The known ssos contain a nick-site for the initiation of perimental strategies for gaining insight into the replication
replication, and a region to which the initiator binds (COM mechanism of any phage under study (COM section C2.3)
section C2.1.). We can distinguish three different types of (Rakonjac et al., 2003). These phages do not encode cognate
initiator-binding sites: (1) an array of repeats (ffd), (2) a replication proteins. Their replication depends instead on
region of 25 bp with pronounced dyad symmetry and the the synthesis of an untranslated transcript that ‘initiates’
potential to form a stemloop (CTXf) and (3) a stretch of replication via tDR by host factors. Thus, also the replication
30 bp lacking any detectable sequence or structural motifs of phages can occur by a mechanism that is known for a long
(fP2, fX174). Although a particular type of sso is recog- time from the ColE1-type plasmids. Further search for
nised by a particular type of initiator, the structural basis for phage replication genes and mechanisms will have to take
this interaction is presently not well enough understood to into account that the failure to identify replication genes by
derive rules from it for reliable predictions. There are four protein homology/similarity to known examples calls for
known localisations for the sso with respect to the initiator experiments to determine the host factors required for
gene: (1) in the 5 0 -part of the initiator gene (fX174), (2) in replication of a phage under study.
the 3 0 -part of the initiator gene (fP2), (3) in the intergenic
region 5 0 -upstream of the initiator gene (ffd) and (4) in an Evolutionary considerations
intergenic region elsewhere in the phage genome (CTXf).
There are three known structures in the single-stranded Bacteriophages present a wider spectrum of replication
form of these phages that can serve as sso: (1) a secondary mechanisms than bacterial plasmids or chromosomes and
structure ‘mimicking’ a promoter (ffd), (2) a secondary an impressive variety of different types of enzymes that
structure with the quality of a primosome-assembly site perform particular steps during replication. When, in addi-
(fX174) or (3) a primase binding-site (fG4). The co- tion, the variability of related phage replication modules is
localisation of sso and dso in the filamentous E. coli phages considered, it becomes immediately clear that the evolution
fd, f1 and M13 is exceptional: in all other known systems, sso of phage replicons cannot be discussed in depth in the
and dso are not linked. For the phages with mid-sized context of this review. Therefore, we confine ourselves to a
genomes that replicate by RCR, the fP2 B-type helicase discussion of two particularly interesting types of replication
loader should be considered as part of the replication proteins: helicases and helicase loaders.
module. However, the exact mechanism of complemen- We first discuss possibilities to identify the replicative
tary-strand synthesis of fP2 is presently not known, and helicase in phage genomes that encode more than one
other cognate proteins might be involved in addition to B helicase (see ‘The different types of phage-encoded helicases’
(Liu et al., 1993). This topic requires further research. section). We then discuss the evolutionary origin of phage-
encoded homologues of the E. coli DnaB helicase proposed
by Moreira (2000) (see ‘Phage-encoded homologues of the
Phage replicons lacking replication protein E. coli DnaB helicase’ section). Lastly, we present a hypoth-
genes esis on the evolutionary origin of bacterial helicase loaders:
Comparing the number of 220 completely sequenced the DnaC-type helicase loader of several Gammaproteobac-
phage genomes with the number of phages discussed in this teria, and the DnaBI and DnaD helicase loaders of the
review, we realise that for approximately 30 – excluding the bacillales (see ‘Chromosomally encoded homologues of
phages with RNA genomes – no putative replication protein phage helicase loaders’ section).
genes could be identified by comparison with known
examples. Among the phages lacking known replication The different types of phage-encoded helicases
protein genes are the ‘relatives’ of E. coli phage Mu and a
number of small phages with ssDNA genomes that may also Our survey of 87 completely sequenced (pro)phage genomes
replicate via transposition, e.g. Spiroplasma f1-R8A2B. The (dsDNA) that encode one or more helicases revealed the
replication of the mycobacteriophages Che9c, Corndog and presence F4-type helicases in 50 of them, either as the only

c 2006 Federation of European Microbiological Societies FEMS Microbiol Rev 30 (2006) 321–381
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 Bacteriophage replication modules 365

helicase (39) or together with a SF2-type helicase (11).

fP4a-type helicases were found less frequently (20), and in
the majority of the cases in combination with a SF2-type
helicase (13). We could not find an example for a phage
genome that encodes a F4-type helicase together with a
fP4a-type (primase-)helicase. Altogether, 40 phage gen-
omes were found to encode SF2-type helicases. SF1-type
helicases were only found in the group of fT4-like phages
(6). This confusing scenario of different types of helicases
encoded by bacteriophages provokes the questions: (1)
whether an underlying pattern exist for the occurrence of
particular types of helicase gene(s) within the specific set of
replication genes in a given phage replicon, and (2) whether
such a pattern allows us to pinpoint the replicative helicase,
i.e. the helicase associated with the replication fork during
the elongation step of replication.
As the basis for an answer to these two questions, we
collected from the fully sequenced (dsDNA-)phage genomes
that encode helicases the data for initiators (COM section
C3.1.), helicase loaders (COM section C3.2.), primases and
DNA polymerases (COM section C3.5.). We present in Fig. 26. Phage-encoded helicases in the context of other replication
Fig. 26 a scheme that includes examples for all detected proteins. Helicase genes detected in the (completely sequenced) gen-
variants of replication gene assortments, omitting only the omes of the listed phages are shown by their gene/gene product/locus
SF1-type Dda helicases of phages from the fT4 group. For tag names. The three helicase/helicase domain types are shown by
completeness, we added the data for l and fA118, neither of different colouring (see colour code in Fig.). Initiators, helicase loaders,
primases and DNA polymerases detected in the genomes of the listed
which encodes helicases.
phages are included to allow for a comparison of the phage-encoded
There are two possible approaches to answer our initial
replication functions. Gene products are not shown to size, and the gene
questions: (1) a strict approach demanding that every order is not reflected in the figure. Bars indicate the failure to detect the
(putative) replicative helicase is experimentally analysed for respective genes by signature motif and BLAST searches of known
this property prior to a decisive statement, and (2) a more helicases and polymerases, respectively. For initiators, helicase loaders
relaxed approach that allows us to classify and hypothesise and primases, a blank indicates the failure to detect one of the respective
on the basis of reliable experimental results obtained only genes by comparison with one of the known types. The corresponding
Escherichia coli proteins are shown on top to allow for an easy
for a subset of the systems to be compared. The first
comparison. The role of the E. coli YejH protein is not known; it has been
approach is inherently less error-prone and therefore more
included here solely for completeness. Phage genomes NCBI accession
attractive. We cannot neglect the second approach, however, numbers are: fA118 [NC_003216], l [NC_001416], fP22
because it is better suited to promote a deeper understand- [NC_002371], fSPP1 [NC_004166], fMAV1 [NC_001942], fBxz1
ing of the fundamental biological process of replication by [NC_004687], fKMV [NC_005045], fT7 [NC_001604], fD29
allowing the prediction of interesting model systems for [NC_001900], fPG1 [NC_005259], fK [AY176327], fT4 [NC_000866],
experimental studies. This becomes particularly important fRM378 [NC_004735], fP4 [NC_001609], fC31 [NC_001978], fBPP-1
[NC_005357], fBarnyard [NC_004689], fSfi21 [NC_000872], fBcep1
when one considers that only a small percentage of the many
[NC_005263], fT1 [NC_005833] and fN15 [AF064539]. The genome
phage replicons known to date will ever be analysed by
sizes of the phages are shown in the rightmost column.
genetic or biochemical experiments.
The ‘strict approach’ gains support when the helicases of
eukaryotic viruses are also taken into consideration – in a sequence would suggest (Bird et al., 1998; Korolev et al.,
brief survey. The helicase subunit (UL5) of the trimeric 1998). Several viruses employ for their replication super-
UL5-UL8-UL52 primase-helicase complex of HSV1 is a family 3 helicases (SF3), for which no members were
member of the helicase superfamily 1 (SF1). The UL9 detected in phage genomes. Examples include the Rep40
protein of HSV1, which is responsible for origin recognition helicase of adeno-associated virus 2 (AAV2), the helicase
and unwinding together with the UL29 (ICP8) SSB, belongs domain of simian virus 40 (SV40) T antigen (James et al.,
to the helicase superfamily 2 (SF2) (Marintcheva & Weller, 2003) and the E1 origin-binding protein of human papillo-
2001a, b). However, the SF1- and SF2-type helicases have mavirus (HPV) (Masterson et al., 1998). Several virus-
probably a higher degree of similarity in their three-dimen- encoded SF2-type helicases are involved in the replication
sional structure than the comparison of their primary of viral RNA genomes, e.g. hepatitis C virus (HCV) NS3

FEMS Microbiol Rev 30 (2006) 321–381

c 2006 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 366 C. Weigel & H. Seitz

helicase (Lam et al., 2003). We could not detect (by BLAST The ‘staged initiation’ model postulates that the origin-
searches) proteins with significant similarity to the F4-type bound replicative helicase DnaB (preprimosome) recruits
helicases of prokaryotes and bacteriophages in the genomes the primase DnaG for the synthesis of the leading-strand
of eukaryotes and their viruses, except for some candidate primer (primosome). A physical interaction of DnaB and
proteins in genomes of chloroplasts and mitochondria. Also DnaG could be shown in vitro for the E. coli system (Tougu
for the fP4 a-type (primase-) helicase type, reasonably & Marians, 1996), and also that the binding of fSPP1 G40P
similar proteins were not detected in virus genomes, except helicase to ssDNA is stabilised by the addition of host DnaG
for a protein of unknown function of the Ectocarpus (Ayora et al., 1998). Replisome assembly then occurs at the
siliculosus (marine brown alga) EsV-1 virus. The application DNA  DnaB  DnaG complex, involving yet another set of
of more refined bioinformatic methods than the crude multiple protein interactions with Pol III holoenzyme sub-
BLAST approach would be necessary to ascertain a relation- units. The tight interaction of replicative helicase and
ship of several (putative) pox virus helicases with the fP4 a primase is particularly important for the repeated priming
helicase domain. Apparently, the already complex pattern of of Okazaki fragments during co-ordinated leading- and
different helicases encoded by phage replicons is only a part lagging-strand synthesis by the replisome. The finding of
of the puzzle, and one can hardly repress the notion that primase domains fused to helicase domains, e.g. fT7 gene 4
virtually all types of helicases can be adapted to the specific protein and fP4a, emphasises the importance of the inter-
requirements of a particular step in a nucleic acid metabolic action of the replicative helicase with the primase. There is
pathway, e.g. to the role as replicative helicase. presently only one example of a SF2-type helicase
The relaxed approach would start from the background of fused to a primase domain: the fN15 RepA protein and its
the classical ‘staged initiation’ model for chromosome homologues in fKO2, fPY54 and fVP882. The modular
replication in E. coli that positions the loading step of the architecture of the fN15 RepA protein is strikingly similar
replicative helicase DnaB at the interface of open complex to that of fP4 a, as is its role for fN15 replication, but the
formation and replisome formation (Kaguni & Kornberg, primase and helicase functions have yet to be established
1984; Kornberg & Baker, 1992). Escherichia coli DnaB (F4- (Ravin et al., 2003). Also, it is not known precisely to
type) is the prototype replicative helicase of prokaryotes what extent fN15 replication is independent of host
with orthologues in all sequenced bacterial genomes; its functions.
participation in recombination processes (e.g. branch mi- A tendency becomes apparent if the working hypothesis is
gration) has been suggested only recently (Kaplan & O’Don- accepted that the monofunctional SF2-type helicases do not
nell, 2002). It is therefore reasonable to assume that all function as the replicative helicases of bacteriophages:
phage-encoded DnaB homologues function as replicative replicons that code for a helicase and, in addition, a DNA
helicases for their cognate replicons; this has been shown polymerase also encode a primase/primase domain, but
experimentally for fT4 gp61, fT7 gene4A and fSPP1 G40P. phage replicons devoid of a DNA polymerase gene are also
The a protein is the replicative helicase for fP4 replication lacking a primase gene. This observation holds for phage
(Briani et al., 2001). The orf382 and orf504 proteins of replicons encoding F4-type helicases and fP4a-type heli-
fSfi21 and fSfi11, respectively, lack the N-terminal primase cases, likewise. Probably the assortment of replication genes
domain of their homologue, fP4a. Presumably, the fSfi21 simply reflects the degree of dependence of a given phage on
orf382-type proteins function as combined ‘initiator-heli- host replication proteins:
case’ (COM section C3.1.2.). There is presently no phage (1) Phage replicons lacking a cognate helicase gain access to
genome known that encodes an F4-type helicase together the host replication machinery by attracting the host repli-
with a fP4a-type (primase-) helicase, and we therefore cative helicase through interaction with their initiator
hypothesise that phages encode for one particular type of (fA118) or helicase loader (l). The host replicative helicase
replicative helicase only. The replication protein sets en- then attracts the host primase and DNA polymerase.
coded by the phages fKMV and fK differ only with respect (2) Phage replicons encoding a helicase gain access to the
to the presence/absence of a SF2-type helicase (see Fig. 26). host replication machinery by attracting the host primase
Together, these observations suggest that the phage-encoded through interaction with their helicase, after loading of the
SF2-type helicases are unlikely to function as replicative latter to the phage replication origin by interaction with the
helicases, despite their importance for phage replication, e.g. initiator (fP22) or the helicase loader (fSPP1). The phage
for a switch to RDR. Although the sequence similarity helicase  host primase complex then recruits the host DNA
between the phage-encoded UvsW-type helicases and the polymerase.
bacterial PriA homologues is very low, it may be informative (3) Phage replicons that encode a replicative helicase,
that PriA, the ‘initiator’ of the E. coli restart primosome, is primase and DNA polymerase are independent from the
also a SF2-type helicase (Marians, 1996) (see ‘Replication host replication machinery for elongation (fT4, fT7,
restart’ section). fC31).

c 2006 Federation of European Microbiological Societies FEMS Microbiol Rev 30 (2006) 321–381
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 Bacteriophage replication modules 367

With bacteriophages, there are always exceptions from Phage-encoded homologues of the E. coli
rules or patterns (quoting the complaint of Casjens DnaB helicase
2003): fP4 and fN15 encode a primase as domain of the
a or RepA proteins, respectively, but no DNA poly- Although a review does not usually allow the authors to
merase. Therefore, we cannot claim that a ‘helicase pattern’ present their own data as part of the discussion, we would
exists, and prefer to talk of a tendency, rather. This tendency like to point out that the phylogenetic tree for DnaB
is weakened, however, by the fact that for three phages homologues presented by Moreira based on a maximum-
shown in Fig. 26 the primase domains can only be likelihood analysis seems highly questionable: a fT7/fP22/
called ‘putative’, at best. These phages encode F4-type fHK022 group splits from the E. coli branch very early, and
(fPG1) or fP4a-type (fBPP-1, fBarnyard) helicases then differentiates into the fT7/fHK022 and fP22 sub-
with extended N-termini that lack similarity to known branches; the branching points of E. coli/fP1 and fT7/
protein sequences, except to other closely related phage fHK022 are at roughly equal distance from the common
proteins. In analogy to the fT7 gene 4 and fP4a primase- origin (see Fig. 1 in Moreira, 2000). Our analysis of
helicase fusion proteins, we speculate that these extended signature motif conservation among F4-type helicases sug-
N-termini represent yet unknown primase domains. Ex- gests, in contrast, that the P and gene 12 proteins of phages
perimental evidence in support of this speculation is fHK022 and fP22, respectively, belong to the DnaB sub-
not available. It should be noted that, for example, the family whereas the gene 4A helicase of fT7 is the prototype
RepB 0 primase of plasmid RSF1010 also lacks any of a distinct subfamily (compare COM Tables C11 and C14).
sequence similarity with DnaGEco-type primases, showing We performed a BLAST sequence comparison of individual
that much has still to be learned about the existing primase domains of DnaB homologues analysed by Moreira based
families. on the known domain structure of E. coli DnaB (Biswas &
The ‘relaxed approach’ cannot provide a conclusive Biswas, 1999). In this comparison, we included the site
model at present, but it allows us to suggest experiments: within the N-terminus of the b domain, which is responsible
promising candidates for a test of whether SF2-type heli- for the primary interaction with E. coli DnaA and several
cases can act as replicative helicases for their cognate plasmid initiator proteins (Datta et al., 1999; Seitz et al.,
replicons are fN15 RepA, fBcep1 gp66 and fT1 gene 22 2000). We found the highest degree of conservation for the
(see Fig. 26). Recent results indicate that the SF2-type RepA C-terminal DNA-binding g domain (signature motifs 2–4),
helicases are indeed the replicative helicases of fN15 and the followed by slightly lower conservation of the nucleotide-
related phage fPY54 (Mardanov et al., 2004; Ziegelin et al., binding b domain (signature motifs 1 and 1a) (Table 3).
2005). fBcep1 has not yet been studied genetically, and the Similarity in the N-termini was in general 20% lower than
available experimental data for fT1 replication are some- in the two C-terminal domains. We were not surprised to
what contradictory: the phage can be propagated in E. coli find the lowest degree of conservation for the primary site of
dnaB(ts) and dnaC(ts) hosts (Bourque & Christensen, 1980) DnaA–DnaB interaction (residues 154–210). The still sig-
but fT1 mutants with an inactivated gene 22 could be nificant degree of similarity between ban and DnaB in this
isolated (cited in Roberts et al., 2004). In addition, fT1 region certainly contributes to the successful substitution of
replication requires a functional DnaG primase of its host, DnaB by ban described by Lemonnier et al. (2003). But all
although gene 24 encodes a DnaG homologue (Bourque & other phage-encoded helicases (compared in Table 3) en-
Christensen, 1980). code, at best, distantly related initiator proteins, which in
Finally, there is no apparent correlation between the type turn would require a specific adaptation of the helicase,
of helicase(s) and the type of DNA polymerase encoded assuming that the same site is responsible for interaction.
by phage replicons. However, for a thorough discussion All compared helicases showed the lowest degree of
we would have to consider also the processivity factors. similarity to fT7 gene 4A helicase, and similarity was only
For example, it has been shown for E. coli that all three detectable in the C-terminal DNA-binding domain.
DNA polymerases can interact with the DNA polymerase Although it is difficult to determine the evolutionary rate
III b-sliding clamp, and form a replication fork together of phage-encoded genes exactly, this suggests that the
with the DnaB helicase through multiple molecular interac- divergence of fT7 gene 4A from the other phage-encoded
tions (Lopez de Saro & O’Donnell, 2001). At present, the helicases occurred before the divergence of the latter from E.
processivity factors have only been characterised for the coli DnaB and B. subtilis DnaC. Clearly, the P and gene 12
fT4 and fT7 replicons, while the fT5 DNA polymerase proteins of fHK022 and fP22, respectively, are more closely
possesses high intrinsic processivity. Virtually nothing is related to DnaBEco than to fT7 gene 4A protein, in support
known about the possible interaction of fP4a-type helicases of the subfamily grouping by signature motifs (see also
with DNA polymerases. Therefore, we have to leave this Ilyina et al., 1992). We can confirm, however, the result of
point open. Moreira that Ban is evolutionary closer to DnaB of its host

FEMS Microbiol Rev 30 (2006) 321–381

c 2006 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 368 C. Weigel & H. Seitz

Table 3. Cross-comparison (BLAST) of family 4 helicases by their domains

Escherichia coli dnaB f P1 ban Bacillus subtilis dnaC f SPP1 gene 40 f HK022 P f P22 gene 12 f P27 L19

1272179
1802276
2772463
1252173

1742273

2742458
1382193
1942305
3062478
1332181

1822247

2482442
2852454

1372189

1902281

2822454
3032471

1362191
1922284
1542210

2112302

12126
12124
12137
12132
12136
12135
12153
Replicon Gene

Escherichia coli dnaB    

f P1 ban 73 66 85 82    
Bacillus subtilis dnaC 39 38 48 55 36 35 51 55    
f SPP1 gene 40 28 – 31 49 29 – 34 52 30 – 48 63    
f HK022 P 50 – 37 45 44 – 37 42 37 – 31 39 25 – – 38    
f P22 gene 12 23 – 36 49 – – 36 47 – – 38 41 – – 38 40 30 – 29 38    
f P27 L19 – – 38 44 – – 36 43 – – 37 40 – – 39 37 25 – 36 32 28 – 51 49    
f T7 gene 4A – – – 23 – – – 28 – – – 22 – – – 26 – – – – – – – – – – – 25

Domains of the helicases were compared by the bl2seq program using the blastportion with default settings [gap open: 11; gap extension: 1;
gap  dropoff: 50; expect: 10.0; word size: 3; no (low complexity) filter] Tatusova & Madden (1999). The compared domains were adjusted to match the
known domain structure of Escherichia coli DnaB: residues 1–153 = N-terminal a domain; residues 154–210 = interaction site with DnaA Seitz et al.
(2000) and with several plasmid initiators Datta et al. (1999), residues 211–471 = b domain containing the signature motifs 1 plus 1a of the F4-type
helicases; C-terminal g domain containing the signature motifs 2, 3 and 4 of the F4-type helicases (Hall & Matson, 1999; Biswas & Biswas, 1999). Values
shown are percentage identical residues; a dash indicates no detectable similarity ( o 20% identical residues).

E. coli than fSPP1 G40P is to the DnaC replicative helicase by the E. coli dnaTC gene pair and the initiator plus helicase
of its host B. subtilis. loader proteins encoded by the ydaUV replication module of
Despite disputable branching points, the (theoretical) the E. coli K12 Rac prophage, respectively (Wrobel &
evidence that DnaB homologues were acquired from their Wegrzyn, 2002). Because of the elaborate functional inter-
enterobacterial hosts independently by phage replicons play of DnaT and DnaC with other replication proteins the
related to the extant phages fP1 on one side, and fSf6, dnaT and dnaC genes are considered housekeeping genes of
fP22, and fP27 on the other side, is compelling. It also E. coli (see ‘Replication restart’ section). Several lines of
seems reasonable to conclude that the ancestor of fSPP1 evidence lead us to speculate, however, that a progenitor of
acquired gene 40 from a Gram-positive host. The root of the E. coli acquired the dnaTC gene pair approximately 108 years
‘DnaB tree’, however, requires revision, including not only ago from a bacteriophage replicon, i.e. by horizontal gene
the fT7 gene 4A-type but also the fD29 gp65-type and fT4 transfer (HGT):
gp41-type helicase subfamilies. To improve the value of such (1) Initiator plus helicase loader gene pairs are common in
phylogenetic studies, we suggest including available data on replication modules of Gram(  )-specific phages replicat-
protein domain architecture, conserved structural and func- ing by yDR (see ‘‘Initiator-helicase loader’ replication mod-
tional motifs, and experimentally defined interaction sites, ules’ section). The propagation of these phage replicons
rather than relying exclusively on the raw protein sequences. depends strictly on the recruitment of the replicative heli-
Interestingly, fHK022 P protein, fP22 gene 12 protein case of their hosts. Conversely, the chromosomal replicons
and fP27 L19 protein seem to be at roughly equal evolu- of Gram-negative bacteria have to compete with invading
tionary distance from each other (Table 3). BLAST analysis is phage replicons for the replicative helicase. The acquisition
therefore not sensitive enough to detect subtle differences in of a phage-encoded helicase loader by the host chromoso-
the primary sequence of the helicases that would allow us to mal replicon could potentially improve its fitness for com-
predict the protein-specific mechanism for helicase loading: petition with an invading phage replicon. Although a newly
fP27 encodes a DnaCEco-type helicase loader (COM section acquired helicase loader would be already shaped for opti-
C3.2.), but not fHK620. In addition, fP22 does not encode mal functioning with the host helicase through a long-
a cognate helicase loader, and Wickner (1984a, b) was able to lasting coevolution of phage and host, it would have to
show that loading of gene 12 helicase to ssDNA in vitro does replace the established molecular mechanisms for helicase
not require DnaCEco. loading in order to confer any selective advantage to the host
replicon. Therefore, the acquisition of the helicase loader
gene, dnaC, can only be one aspect of why the progenitor of
Chromosomally encoded homologues of phage
E. coli kept the dnaTC gene pair.
helicase loaders
(2) DnaT shows significant similarity with the C-terminus
We have discussed in COM section C3.2. the significant of YdaU (see ‘Bacteriophage replication modules’ section)
similarities between the two primosomal proteins encoded but no characteristics of a phage initiator protein in its

c 2006 Federation of European Microbiological Societies FEMS Microbiol Rev 30 (2006) 321–381
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 Bacteriophage replication modules 369

Fig. 27. Similarity network for the Escherichia coli

dnaTC genes. The bacterial and (pro)phage genes
are shown roughly to size with the source indi-
cated on the left side; the direction of transcription
of the genes is indicated by arrowheads; gene
names (locus tags) are shown underneath to-
gether with the sizes of the encoded proteins. The
‘%’ values shown indicate the percentage of
identical residues for the compared regions of two
proteins.

N-terminus (Fig. 27; see also COM sections C2.2. 1 C3.1.2.). or P. luminescens (Thomson et al., 2002; Duchaud et al.,
To account for this observation, we speculate that a for- 2003). Interestingly, the former four species, but not Yersinia
tuitous recombination event in the E. coli progenitor created sp., belong to one of three clusters of enteric bacteria that
a translational fusion between a chromosomal gene and the could be grouped by their specific combinations of aromatic
3 0 -half of a prophage-encoded initiator gene, dnaT. The amino acid synthesis enzymes, and are therefore believed to
chromosomal recipient gene had the potential to interact represent very closely related species (Ahmad et al., 1990).
with one of the PriABC primosomal proteins, possibly PriA The presence of a dnaTC gene pair in the genomes of very
(Sandler, 2000). Interestingly, a gene was found in the few closely related species makes it likely that their common
genomes of two enteric bacteria (not in E. coli) encoding a ancestor had acquired it. It is less likely, by contrast, that an
small protein that shares in its N-terminus 30% identical ancestral dnaTC gene pair was lost from the chromosomal
residues with the DnaT N-terminal half (Hayes, 1998). The replicons of the other gammaproteobacteria during their
presumed recombination event resulted in dnaT in its evolution.
present form, relieved the transcription of the newly created (4) According to phylogenetic studies, S. enterica and E. coli
gene fusion from the control of the prophage repressor, and diverged as distinct species 1.2–1.6  108 years ago (Och-
eliminated – by deletion of the dnaT 5 0 -half – the replica- man & Wilson, 1987). The dnaTC gene pair is embedded in
tion origin of the prophage, activity of which would other- a larger well-conserved context in the chromosomes of both
wise be detrimental to the host. The novel fusion protein species. This makes it likely that the genome of their
DnaT could recruit the replicative helicase DnaB to PriA- common ancestor already contained the dnaTC gene pair.
primosomes at stalled replication forks by interaction with The alternative, that one species has inherited the dnaTC
its cognate helicase loader, DnaC. We assume that this gene pair from the other by HGT at some later time, seems
statistically highly improbable recombination event became less likely.
genetically fixed because it resulted in a selective advantage (5) Given the great timely distance, it is not surprising that,
for the chromosomal replicon, adding considerably to the except for the dnaTC gene pair, all other parts of the
improvement of competition fitness by the acquisition of a presumed ancient prophage have been eliminated from the
helicase loader alone. Note that these extensive speculations E. coli genome in its present version. In addition, differences
are in principle amenable to experimental analysis. in the G1C content or the codon usage between the
(3) Orthologues of the E. coli dnaTC gene pair are found presumed prophage-encoded dnaTC genes and the chromo-
exclusively in the genomes of enterobacteriales, and within some have been reduced. Accordingly, the dnaTC gene pair
this subfamily of the gammaproteobacteria so far only in the escaped detection in analyses aimed to identify recent
genomes of all E. coli, S. flexneri, S. enterica, K. pneumoniae acquisitions of the E. coli genome, and which were successful
and B. aphidicola strains, but not in others, e.g. Yersinia sp. in the cases of Rac, other prophages or the lac operon

FEMS Microbiol Rev 30 (2006) 321–381

c 2006 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 370 C. Weigel & H. Seitz

(Lawrence & Ochman, 1998; Canchaya et al., 2003; Casjens, containing initiator gene (e.g. ‘AT-peak’, iterons) although
2003). Several lysogenic phages of E. coli integrate into its both proteins bind to DNA (Marsin et al., 2001).
chromosome preferentially at or close to tRNA genes, and Only the C-termini of the DnaB and DnaD proteins of
many phage-derived genes are found at such sites (Lawrence S. aureus contain a stretch of reasonable similarity, but not
& Ochman, 1998). Intriguingly, the dnaTC gene pair is those of Listeria or Bacillus species. In addition, the con-
located at a distance of 5 kb from the leuVPQ genes servation of the dnaBI and dnaD genes in the bacilli
encoding tRNAs. genomes is considerably lower than that of several other
We were encouraged to speculate about the evolutionary replication proteins, e.g. DnaA, DnaCBsu. It is therefore
origin of the E. coli dnaTC gene pair by the observation that difficult to decide whether the dnaD gene was acquired
also the B. subtilis dnaBI gene pair has probably a similar independently from the dnaBI gene pair or whether it is a
origin, together with the dnaD gene. The details of the duplicated dnaB gene. Although the replication modules of
molecular mechanisms for helicase loading are strikingly bacilli phages and prophages frequently encode helicase
different in both species, but the DnaT and DnaC proteins of loaders with similarity to DnaI, in all these cases the
E. coli and the DnaB, DnaD and DnaI proteins of B. subtilis similarity of the respective phage initiator with either DnaB
perform analogous functions as primosomal proteins or DnaD was low or not detectable. Owing to this complex
(COM section C3.2.). As discussed in ‘Phages encoding pattern of mutual relatedness – reminiscent of mosaicism –
initiator proteins’, the initiator genes of several y-replicating no simple line of descent can be drawn (Fig. 28).
(pro)phages of bacilli show significant sequence similarities All sequenced genomes of the bacillales subfamily of the
in their C-terminal half with the C-termini of the DnaB bacilli (i.e. Bacillus sp., Listeria sp., Staphylococcus sp.,
protein (e.g. fBK5-T), the DnaD protein (e.g. fSM1) or Oceanobacillus iheyensis) contain in a fairly well-conserved
both proteins (e.g. fA118) of their specific hosts. Also, genomic context a dnaBI gene pair, and – unlinked to it – a
(putative) helicase loaders, which show significant similarity dnaD gene directly upstream of the nth gene encoding
with B. subtilis DnaI, are frequently found in initiator plus endonuclease III. In addition, all sequenced genomes of the
helicase loader replication modules of Bacilli phages. We second bacilli subfamily, the lactobacillales, contain a dnaBI
speculate that the dnaBI gene pair of B. subtilis was acquired gene pair in a genomic context, which is reasonably similar
from a prophage because it resembles the initiator plus to that in the bacilli genomes. The situation is different for
helicase loader gene pairs found in phage replicons. We the dnaD genes of lactobacillales: (1) in the first group
cautiously extend this hypothesis to the dnaD gene, which (Enterococcus faecium, E. faecalis V583, Streptococcus epider-
resembles a phage initiator gene. As we found for dnaT in midis, S. pyogenes MGAS8232, L. lactis) the dnaD genes
the dnaTC gene pair of enteric bacteria, the dnaB and dnaD are located next to the nth genes, as in the genomes of
genes of B. subtilis lack the main characteristics of an origin- bacillales; (2) in the second group (S. pneumoniae TIGR4,

Fig. 28. Similarity network for the Bacillus subtilis

dnaBI and dnaD genes. The bacterial and (pro)ph-
age genes are shown roughly to size with the
source indicated at the left side; the direction of
transcription of the genes is indicated by arrow-
heads; gene names (locus tags) are shown under-
neath together with the sizes of the encoded
proteins. The ‘%’ values shown indicate the per-
centage of identical residues for the compared
regions of two proteins.

c 2006 Federation of European Microbiological Societies FEMS Microbiol Rev 30 (2006) 321–381
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 Bacteriophage replication modules 371

S. pneumoniae R6, Streptococcus mutans, S. agalactiae) the proteins of the other two major information processing
dnaD genes are not linked to the nth genes; (3) in the third systems – translation and transcription – in all three
group (L. plantarum, L. gasseri, Leuconostoc mesenteroides, domains (Edgell & Doolittle, 1997; Leipe et al., 1999; Tye,
Oenococcus oeni) the dnaD genes are present but nth 2000; Woese, 2000, 2002). Forterre’s (1999) suggested solu-
orthologues missing. However, these three groups do not tion to this puzzle was that in the prokarya many replication
correspond to the intrabranch clustering of the lactobacil- genes were replaced by nonorthologous plasmid- or virus
lales species based on the conservation of their 16S rRNA (phage)-derived genes shortly after the divergence of the
(Ludwig et al., 1985). three domains of life from LUCA, the last universal common
We could not detect homologues of the dnaBI gene pair ancestor (Penny & Poole, 1999). This hypothesis was ques-
in the sequenced genomes of species of the clostridia branch tioned by Moreira (2000), who found that several plasmid-
of the firmicutes. Genes with significant similarity to dnaD or phage-encoded replication proteins resemble the chro-
were rare, and in all cases coupled to a gene related to phage mosomal counterparts of their specific hosts more than
helicase loaders (DnaIBsu-type). These gene pairs may rather other plasmid- or phage-encoded orthologues, even when
be prophage replication modules (COM section C3.1.2.). In taking higher mutational rates of virus genes into account
the sequenced genomes of species of the mollicutes branch (Drake et al., 1998). Prima vista, our hypothesis seems to be
of the firmicutes we could not detect dnaD homologues. All in line with the proposal of Forterre, who explicitly men-
sequenced genomes contain a homologue of the B. subtilis tions DnaCEco and DnaIBsu. There are three points of
dnaB gene, but only in Mycoplasma penetrans and Ureaplas- reserve, however: (1) the helicase loader genes were intro-
ma urealyticum linked to a dnaI homologue. Interestingly, duced into the chromosomal replicons of enterobacteriales
the similarity to DnaBBsu is confined to the N-termini of and firmicutes independently, and in both cases apparently
these proteins, in contrast to the C-terminal similarity of at a later stage of branch differentiation; (2) the helicase
DnaBBsu to phage initiators (see Fig. 28). We were unable to loader genes were introduced together with initiator genes
locate a dnaI homologue in the highly rearranged and size- into the chromosomal replicons – the latter, however, did
reduced genomes of Mycoplasma pneumoniae, M. gallisepti- not replace the cognate initiator of the recipient replicons,
cum and M. genitalium. The M. penetrans gene pair DnaA, but were converted into primosomal proteins in-
MYPE2020 and MYPE2030, encoding proteins similar to stead; (3) it is presently unclear whether the newly intro-
the B. subtilis DnaB and DnaI proteins, respectively, is duced helicase loaders replaced the genes driving the
located closely upstream of the MYPE2050 gene, encoding primordial mechanism for helicase loading to ssDNA, or
the replicative helicase. This arrangement is reminiscent of simply added more specificity. All three types of prokaryotic
the ILH-type replication module found in fSPP1 and few replicons provide us with examples suggesting that it is likely
other phages. However, this arrangement is not found in U. that helicase loaders added specificity to the existing mole-
urealyticum, and may therefore represent a fortuitous gene- cular mechanisms rather than replacing the gene(s) respon-
coupling event. sible for primordial pathways for helicase loading.
We infer from these observations that the common Among the various types of phage replication modules
ancestor of the bacilli branch of the firmicutes acquired the discussed in the previous section, the initiator plus
dnaBI gene pair and the dnaD gene from different pro- helicase modules are particularly informative. They are
phages. How the three phage-derived genes were subse- found in Gram(1)-specific phages (f3626, f11) and
quently reshaped to create the intricate three-protein Gram(  )-specific phages (fD3, fST64T, fP22). It has
helicase loading mechanism of B. subtilis remains ‘in the not been shown experimentally, but it is reasonable to
cloud of unknowing’ at present (Donachie, 2001). Our assume that the replication of these phage replicons depends
hypothesis that chromosomal replicons of bacteria acquired on the specific interaction of their initiators with their
genes encoding initiator and helicase loader proteins from cognate helicases.
bacteriophages twice during the evolution of this domain of Helinski, Konieczny and coworkers elucidated the intri-
life lacks an experimental basis, and the evidence obtained cate mechanism evolved by the broad host range plasmid
by mere comparisons of protein sequences and gene con- RK2 (IncP-group) to ensure its propagation in different
texts is circumstantial, although consistent. Admittedly, the hosts (reviewed in Konieczny, 2003). The RK2 initiator gene
most speculative part of our hypothesis is the presumed trfA allows for the synthesis of a longer TrfA-44 protein, and,
conversion of phage initiator proteins into host primosomal using an internal secondary start site, a shorter TrfA-33
proteins. protein. For the in vitro formation of a prepriming complex
It has frequently been claimed that there is an apparent with E. coli replication proteins, both TrfA proteins are fully
lack of orthologous replication proteins in the prokarya active. DnaA, DnaB and DnaC are strictly required, and
domain when compared with archaea and eukarya, in DnaA was shown to recruit the helicase from the DnaB6C6
marked contrast to the considerable orthology found for double-hexamer to the replication origin oriV of RK2 (Jiang

FEMS Microbiol Rev 30 (2006) 321–381

c2006 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 372 C. Weigel & H. Seitz

et al., 2003). Using P. putida DnaB (60% identical residues ple of an approach concentrating on one essential replica-
with DnaBEco) in the assay, TrfA-44 is more active than TrfA- tion protein instead of using poorly defined protein families
33 in prepriming complex formation. The cognate DnaA is was recently presented by Giraldo, who was able to trace the
required when using TrfA-33; DnaAPpu and DnaBPpu were relationship of prokaryotic, archaeal and eukaryal replica-
shown to interact (Caspi et al., 2001). With P. aeruginosa tion initiator proteins very close to LUCA (Giraldo, 2003).
DnaB (85% identical residues with DnaBPpu; 61% identical
residues with DnaBEco), however, only the longer TrfA-44 is
active and recruits the DnaBPae hexamer. Notably, an acces-
Perspectives
sory helicase loader protein is not required for the recruit- The ongoing race/rage of genomic sequencing will provide
ment of both Pseudomonas helicases to RK2 oriV. In addition, bacteriophage research with many promising novel objects,
helicase loading during in vitro initiation of chromosome but only their exploration by genetic, biochemical and
replication in Pseudomonas sp. does not require an accessory structural studies can transform the present data overflow
protein(s) (Y. Jiang, D. Helinski and A. Toukdarian, pers. into something coming closer to knowledge (Brenner,
commun.). Scherzinger et al. (1991, 1997) showed that the in 2000). In particular, knowing more about phage and plas-
vitro replication initiation of another broad host range mid replication will contribute to a better understanding of
plasmid, RSF1010 (IncQ-group), depends on three plasmid- the spread of virulence genes among human pathogens, a
encoded genes: repC (initiator), repA (helicase) and repB 0 task of obvious importance (for recent publications see, e.g.,
(primase). Strand opening at oriV of RSF1010 requires the Davis & Waldor, 2003; Ferretti et al., 2004; Munson et al.,
RepC protein, and the hexameric RepA helicase is recruited to 2004; Nair et al., 2004; Nelson et al., 2004; Summer et al.,
oriV without an accessory helicase loader. The initiation of 2004).
RSF1010 replication is independent from the host replicative The long-standing observation of mosaicism among
helicase and the replication linked to the host machinery at a phage genomes also extends to their replication/recombina-
later stage of replisome formation. tion genes. Our study of bacteriophage replication modules
Both sequenced Yersinia pestis genomes lack a dnaTC suggests that probably every theoretically possible combina-
gene pair (see above), and BLAST searches failed to produce tion exists with respect to replication protein assortment,
matches with other known helicase loaders (COM section and also with respect to the various types of proteins that
C3.2.). Except for DnaT and DnaC, the entire set of replica- carry out a particular enzymatic function. This is to some
tion and recombination proteins of Y. pestis is highly similar degree in conflict with one of the fundamental dogmas of
to that in E. coli (Z70% identical residues) with the evolution theory: that selective pressure drives evolution in
exception of the PriC primosomal protein (36% identical favour of the best-adapted genome (replicon) and results in
residues). Because also the replication origins of the E. coli the ‘survival of the fittest’, as Darwin’s contemporary Spen-
and Y. pestis chromosomes are virtually identical, it is likely cer phrased it. However, if we tentatively understand ‘the
that replication initiation and restart follow the same routes. fittest’ as plural instead of singular we could start to define a
Whether the missing DnaC – and also the missing DnaT, we ‘fitness threshold’, and study its variation over time depend-
would add – is responsible for the long generation times of ing on the ever-changing environmental conditions. A
Y. pestis as suggested by Thomson et al. (2002) has not been thorough discussion of this topic is far beyond the scope of
thoroughly examined. Y. pestis is an evolutionary young this review. But we speculate that the recombination func-
species ( o 2  104 years), and conclusions drawn by com- tions encoded by many phages are, in addition to their role
paring sequence data should therefore be regarded with for phage replication, responsible for the creation and the
caution (Achtman et al., 1999). However, the genomes of Y. maintenance of mosaicism: they could promote the repeated
pseudotuberculosis IP 32953 (ancestor of Y. pestis), and of re-creation of different replication modules with a compar-
Yersinia enterolytica (almost completely sequenced) lack a able selective fitness. Temperate phage replicons could be of
dnaTC gene pair. At present, we may assume that helicase particular importance for any experimental approach to
loading during replication initiation and restart does not define a ‘fitness threshold’ because they represent a unique
require accessory protein(s) in Yersinia. class of replicons: a decrease in their fitness is not necessarily
In conclusion, we propose to exclude proteins other than accompanied by their extinction as long as ‘backup’ proph-
those known to be directly involved in the essential steps of age copies reside in host genomes. In general, evolutionary
replication from interdomain comparisons aimed to eluci- considerations by microbiologists are not readily accepted
date the evolutionary origin of this basic information by evolutionary biologists, including the authoritative Mayr
processing system of all extant cells. Helicase loaders do not (1998). However, we wish to remind the reader that the
belong to this group of essential proteins of prokaryotic ‘phage fluctuation test’ presented by Luria & Delbrück
replicons because initiation reactions have been described (1943) is still the most convincing experimental proof for
(others are plausible) which dispense with them. An exam- yet another fundamental dogma of evolution theory: the

c 2006 Federation of European Microbiological Societies FEMS Microbiol Rev 30 (2006) 321–381
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 Bacteriophage replication modules 373

random occurrence of mutations. After all, phage biology review on phage SPP1 available to us prior to publication.
may still be a promising route to study experimentally the This work was supported by the European Union (MolTool,
principles of the evolution of life. grant LSHG-CT-2004-503155) and the Max-Planck-Society.
The enzymology of chromosome replication in E. coli was
established in the 1980s by the Kornberg lab among others,
and readily accepted as the comprehensive and valid gram- Supplementary material
mar, syntax and vocabulary – literally spoken – of one of the
For all figures, high resolution versions are available from
central cellular pathways of nucleic acid transactions. Ge-
the authors on request.
netic and biochemical studies of the propagation of the two
The following supplementary material is available for
other types of prokaryotic replicons, plasmids and phages,
this article online.
provided evidence for a common grammar: initiation,
Appendix S1. A compendium of phage replication origins
priming, elongation and termination of DNA synthesis.
and phage replication/recombination proteins (including 37
However, various syntactic variants, i.e. molecular mechan-
Tables and 22 Figures).
isms, were revealed by studying initiation, priming, elonga-
The material is available as part of the online article
tion and termination of DNA replication in individual
from http://www.blackwell-synergy.com.
plasmids and phages. Research on plasmid replication con-
tributed broad knowledge to the different mechanisms for
initiation and, in particular, copy number control and
References
segregation (del Solar et al., 1998; Giraldo, 2003). Studies of
phage replication revealed further mechanisms for initia- Abril AM, Salas M, Andreu JM, Hermoso JM & Rivas G (1997)
tion, priming and replication fork restart involving recom- Phage phi29 protein p6 is in a monomer-dimer equilibrium
bination processes. The analysis of phage-encoded that shifts to higher association states at the millimolar
replication proteins teaches us that it is the vocabulary that concentrations found in vivo. Biochemistry 36: 11901–11908.
varies most: a plethora of different types of initiators, Abril AM, Marco S, Carrascosa JL, Salas M & Hermoso JM (1999)
Oligomeric structures of the phage phi29 histone-like protein
helicase loaders, helicases, primases, DNA polymerases and
p6. J Mol Biol 292: 581–588.
polymerase accessory proteins fulfil their enzymatic roles
Achtman M, Zurth K, Morelli G, Torrea G, Guiyoule A & Carniel
within a strikingly similar grammatical and fairly similar
E (1999) Yersinia pestis, the cause of plague, is a recently
syntactic framework. Therefore, the E. coli way of chromo-
emerged clone of Yersinia pseudotuberculosis. Proc Natl Acad
some replication can be considered a local dialect rather Sci USA 96: 14043–14048.
than the lingua franca. Ahmad S, Weisburg WG & Jensen RA (1990) Evolution of
If we abandon thinking about E. coli as ‘the model aromatic amino acid biosynthesis and application to the fine-
prokaryote’ we can approach a more abstract model for tuned phylogenetic positioning of enteric bacteria. J Bacteriol
replication by defining enzymatic steps, rather than de- 172: 1051–1061.
manding that particular types of enzymes carry out these Allison DP, Ganesan AT, Olson AC, Snyder CM & Mitra S (1977)
steps (Benkovic et al., 2001). This ‘universal textbook of Electron microscopic studies of bacteriophage M13 DNA
replication’ would have to include the still missing descrip- replication. J Virol 24: 673–684.
tions of: (1) the DNA-unwinding by initiators in thermo- Alonso JC, Tavares P, Lurz R & Trautner TA (2005) Bacteriophage
dynamically solid terms, and (2) the involvement of cell SPP1. The Bacteriophages. 2nd edn (Calendar R, ed), pp.
membrane (components) in replication. Therefore, we are 331–349. Oxford University Press, Oxford, UK.
not yet in a position to close this chapter of molecular Altermann E, Klein JR & Henrich B (1999) Primary structure and
biology. features of the genome of the Lactobacillus gasseri temperate
bacteriophage (phi)adh. Gene 236: 333–346.
Aoyama A & Hayashi M (1986) Synthesis of bacteriophage phi
Acknowledgements X174 in vitro: mechanism of switch from DNA replication to
DNA packaging. Cell 47: 99–106.
We thank Walter Messer for generous support over all these
Ayora S, Langer U & Alonso JC (1998) Bacillus subtilis DnaG
years – and for sharing our curiosity. In addition, we thank primase stabilises the bacteriophage SPP1 G40P helicase-
Ramón Dı́az Orejas for continuous encouragement and ssDNA complex. FEBS Lett 439: 59–62.
inspiring discussions. We gratefully acknowledge helpful Ayora S, Missich R, Mesa P, Lurz R, Yang S, Egelman EH &
comments and communication of unpublished work by Alonso JC (2002) Homologous-pairing activity of the Bacillus
Erich Lanka, Juan C. Alonso, Sandra Weller, Don Helinski, subtilis bacteriophage SPP1 replication protein G35P. J Biol
John Iandolo, Grzes Wegrzyn, Kenneth Marians, Elisabeth Chem 277: 35969–35979.
Haggård and Jasna Rakonjac. Special thanks are due to J. C. Banks DJ, Beres SB & Musser JM (2002) The fundamental
Alonso, P. Tavares, R. Lurz and T. Trautner for making their contribution of phages to GAS evolution, genome

FEMS Microbiol Rev 30 (2006) 321–381

c2006 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 374 C. Weigel & H. Seitz

diversification and strain emergence. Trends Microbiol 10: Briani F, Dehò G, Forti F & Ghisotti D (2001) The plasmid status
515–521. of satellite bacteriophage P4. Plasmid 45: 1–17.
Bao K & Cohen SN (2003) Recruitment of terminal protein to the Broker TR (1973) An electron microscopic analysis of pathways
ends of Streptomyces linear plasmids and chromosomes by a for bacteriophage T4 DNA recombination. J Mol Biol 81: 1–16.
novel telomere-binding protein essential for linear DNA Broker TR & Doermann AH (1975) Molecular and genetic
replication. Genes Dev 17: 774–785. recombination of bacteriophage T4. Annu Rev Genet 9:
Baranska S, Gabig M, Wegrzyn A, et al. (2001) Regulation of the 213–244.
switch from early to late bacteriophage lambda DNA Bruessow H & Desiere F (2001) Comparative phage genomics
replication. Microbiology 147: 535–547. and the evolution of Siphoviridae: insights from dairy phages.
Baranska S, Konopa G & Wegrzyn G (2002) Directionality of Mol Microbiol 39: 213–222.
lambda plasmid DNA replication carried out by the heritable Bruessow H, Canchaya C & Hardt WD (2004) Phages and the
replication complex. Nucleic Acids Res 30: 1176–1181. evolution of bacterial pathogens: from genomic
Barry J & Alberts B (1994a) A role for two DNA helicases in the rearrangements to lysogenic conversion. Microbiol Mol Biol
replication of T4 bacteriophage DNA. J Biol Chem 269: Rev 68: 560–602.
33063–33068. Burger KJ & Trautner TA (1978) Specific labelling of replicating
Barry J & Alberts B (1994b) Purification and characterization of SPP1 DNA: analysis of viral DNA synthesis and identification
bacteriophage T4 gene 59 protein. A DNA helicase assembly of phage DNA-genes. Mol Gen Genet 166: 277–285.
protein involved in DNA replication. J Biol Chem 269: Burke RL, Munn M, Barry J & Alberts BM (1985) Purification
33049–33062. and properties of the bacteriophage T4 gene 61 RNA priming
Benkovic SJ, Valentine AM & Salinas F (2001) Replisome-
protein. J Biol Chem 260: 1711–1722.
mediated DNA replication. Annu Rev Biochem 70: 181–208. Cadman CJ & McGlynn P (2004) PriA helicase and SSB interact
Bird LE, Subramanya HS & Wigley DB (1998) Helicases: a
physically and functionally. Nucleic Acids Res 32: 6378–6387.
unifying structural theme? Curr Opin Struct Biol 8: 14–18. Cairns J (1963) The bacterial chromosome and its manner of
Biswas EE & Biswas SB (1999) Mechanism of DnaB helicase of
replication as seen by autoradiography. J Mol Biol 6: 208–213.
Escherichia coli: structural domains involved in ATP
Campbell A (1994) Comparative molecular biology of lambdoid
hydrolysis, DNA binding, and oligomerization. Biochemistry
phages. Annu Rev Microbiol 48: 193–222.
38: 10919–10928. Campbell A & Botstein D (1983) Evolution of the lambdoid
Blanco L, Bernad A, Esteban JA & Salas M (1992) DNA-
phages. Lambda II (Hendrix RW, Roberts JW, Stahl FW &
independent deoxynucleotidylation of the phi 29 terminal
Weisberg RA, eds), pp. 365–380. Cold Spring Harbor
protein by the phi 29 DNA polymerase. J Biol Chem 267:
Laboratory Press, Cold Spring Harbor, NY.
1225–1230.
Canchaya C, Proux C, Fournous G, Bruttin A & Bruessow H
Bleuit JS, Ma Y, Munro J & Morrical SW (2004) Mutations in a
(2003) Prophage genomics. Microbiol Mol Biol Rev 67:
conserved motif inhibit single-stranded DNA binding and
238–276.
recombination mediator activities of bacteriophage T4 UvsY
Carles-Kinch K & Kreuzer KN (1997) RNA–DNA hybrid
protein. J Biol Chem 279: 6077–6086.
formation at a bacteriophage T4 replication origin. J Mol Biol
Bollback JP & Huelsenbeck JP (2001) Phylogeny, genome
266: 915–926.
evolution, and host specificity of single-stranded RNA
Casjens S (2003) Prophages and bacterial genomics: what have we
bacteriophage (family Leviviridae). J Mol Evol 52: 117–128.
learned so far? Mol Microbiol 49: 277–300.
Botstein D & Matz MJ (1970) A recombination function essential
Caspi R, Pacek M, Consiglieri G, Helinski DR, Toukdarian A &
to the growth of bacteriophage P22. J Mol Biol 54: 417–440.
Bouché JP, Rowen L & Kornberg A (1978) The RNA primer Konieczny I (2001) A broad host range replicon with different
synthesized by primase to initiate phage G4 DNA replication. J requirements for replication initiation in three bacterial
Biol Chem 253: 765–769. species. EMBO J 20: 3262–3271.
Bourguignon GJ, Sweeney TK & Delius H (1976) Multiple origins Chai S, Bravo A, Lüder G, Nedlin A, Trautner TA & Alonso JC
and circular structures in replicating T5 bacteriophage DNA. J (1992) Molecular analysis of the Bacillus subtilis bacteriophage
Virol 18: 245–259. SPP1 region encompassing genes 1 to 6. The products of gene
Bourque LW & Christensen JR (1980) The synthesis of coliphage 1 and gene 2 are required for pac cleavage. J Mol Biol 224:
T1 DNA: requirement for host dna genes involved in 87–102.
elongation. Virology 102: 310–316. Chakraborty T, Yoshinaga K, Lother H & Messer W (1982)
Bravo A, Serrano-Heras G & Salas M (2005) Purification of the E. coli dnaA gene product. EMBO J 1:
Compartmentalization of prokaryotic DNA replication. FEMS 1545–1549.
Microbiol Rev 29: 25–47. Chang P & Marians KJ (2000) Identification of a region of
Brenner S (2000) Biochemistry strikes back. Trends Biochem Sci Escherichia coli DnaB required for functional interaction with
25: 584. DnaG at the replication fork. J Biol Chem 275: 26187–26195.

c 2006 Federation of European Microbiological Societies FEMS Microbiol Rev 30 (2006) 321–381
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 Bacteriophage replication modules 375

Chang PC, Kim ES & Cohen SN (1996) Streptomyces linear Dodson M, Echols H, Wickner S, et al. (1986) Specialized
plasmids that contain a phage-like, centrally located, nucleoprotein structures at the origin of replication of
replication origin. Mol Microbiol 22: 789–800. bacteriophage lambda: localized unwinding of duplex DNA by
Chao KL & Lohman TM (1991) DNA-induced dimerization of a six-protein reaction. Proc Natl Acad Sci USA 83: 7638–7642.
the Escherichia coli Rep helicase. J Mol Biol 221: 1165–1181. Donachie WD (2001) Co-ordinate regulation of the Escherichia
Chetverin AB (2004) Replicable and recombinogenic RNAs. FEBS coli cell cycle or The cloud of unknowing. Mol Microbiol 40:
Lett 567: 35–41. 779–785.
Chopin A, Bolotin A, Sorokin A, Ehrlich SD & Chopin M (2001) Dotto GP, Horiuchi K & Zinder ND (1984) The functional origin
Analysis of six prophages in Lactococcus lactis IL1403: different of bacteriophage f1 DNA replication. Its signals and domains.
genetic structure of temperate and virulent phage populations. J Mol Biol 172: 507–521.
Nucleic Acids Res 29: 644–651. Drake JW, Charlesworth B, Charlesworth D & Crow JF (1998)
Chung YB & Hinkle DC (1990) Bacteriophage T7 DNA Rates of spontaneous mutation. Genetics 148: 1667–1686.
packaging. I. Plasmids containing a T7 replication origin and Dressler D, Wolfson J & Magazin M (1972) Initiation and
the T7 concatemer junction are packaged into transducing reinitiation of DNA synthesis during replication of
particles during phage infection. J Mol Biol 216: 911–926. bacteriophage T7. Proc Natl Acad Sci USA 69: 998–1002.
Collins J & Hohn B (1978) Cosmids: a type of plasmid gene- Drolet M, Bi X & Liu LF (1994) Hypernegative supercoiling of the
cloning vector that is packageable in vitro in bacteriophage DNA template during transcription elongation in vitro. J Biol
lambda heads. Proc Natl Acad Sci USA 75: 4242–4246. Chem 269: 2068–2074.
Corpet F (1988) Multiple sequence alignment with hierarchical Duchaud E, Rusniok C, Frangeul L, et al. (2003) The genome
clustering. Nucleic Acids Res 16: 10881–10890. sequence of the entomopathogenic bacterium Photorhabdus
Cox MM (2001) Historical overview: searching for replication luminescens. Nat Biotechnol 21: 1307–1313.
help in all of the rec places. Proc Natl Acad Sci USA 98: Dudas KC & Kreuzer KN (2001) UvsW protein regulates
8173–8180. bacteriophage T4 origin-dependent replication by unwinding
Cox MM, Goodman MF, Kreuzer KN, Sherratt DJ, Sandler SJ & R-loops. Mol Cell Biol 21: 2706–2715.
Marians KJ (2000) The importance of repairing stalled Edgell DR & Doolittle WF (1997) Archaea and the origin(s) of
replication forks. Nature 404: 37–41. DNA replication proteins. Cell 89: 995–998.
Cuff JA, Clamp ME, Siddiqui AS, Finlay M & Barton GJ (1998) Espeli O & Marians KJ (2004) Untangling intracellular DNA
JPred: a consensus secondary structure prediction server. topology. Mol Microbiol 52: 925–931.
Bioinformatics 14: 892–893. Esteban JA, Blanco L, Villar L & Salas M (1997) In vitro evolution
Dannenberg R & Mosig G (1983) Early intermediates in of terminal protein-containing genomes. Proc Natl Acad Sci
bacteriophage T4 DNA replication and recombination. J Virol USA 94: 2921–2926.
45: 813–831. Feiss M, Sippy J & Miller G (1985) Processive action of terminase
Datta HJ, Khatri GS & Bastia D (1999) Mechanism of recruitment during sequential packaging of bacteriophage lambda
of DnaB helicase to the replication origin of the plasmid chromosomes. J Mol Biol 186: 759–771.
pSC101. Proc Natl Acad Sci USA 96: 73–78. Feng JN, Russel M & Model P (1997) A permeabilized cell system
Davis BM & Waldor MK (2003) Filamentous phages linked to that assembles filamentous bacteriophage. Proc Natl Acad Sci
virulence of Vibrio cholerae. Curr Opin Microbiol 6: 35–42. USA 94: 4068–4073.
David M, Lubinsky-Mink S, Ben Zvi A, Ulitzur S, Kuhn J & Suissa Ferretti JJ, Ajdic D & McShan WM (2004) Comparative genomics
M (1992) A stable Escherichia coli–Mycobacterium smegmatis of streptococcal species. Indian J Med Res 119(Suppl): 1–6.
plasmid shuttle vector containing the mycobacteriophage D29 Foley S, Lucchini S, Zwahlen MC & Bruessow H (1998) A short
origin. Plasmid 28: 267–271. noncoding viral DNA element showing characteristics of a
Deneke J, Ziegelin G, Lurz R & Lanka E (2000) The protelomerase replication origin confers bacteriophage resistance to
of temperate Escherichia coli phage N15 has cleaving-joining Streptococcus thermophilus. Virology 250: 377–387.
activity. Proc Natl Acad Sci USA 97: 7721–7726. Formosa T & Alberts BM (1986) DNA synthesis dependent on
Deneke J, Ziegelin G, Lurz R & Lanka E (2002) Phage N15 genetic recombination: characterization of a reaction catalyzed
telomere resolution. Target requirements for recognition and by purified bacteriophage T4 proteins. Cell 47: 793–806.
processing by the protelomerase. J Biol Chem 277: Forterre P (1999) Displacement of cellular proteins by functional
10410–10419. analogues from plasmids or viruses could explain puzzling
Deneke J, Burgin AB, Wilson SL & Chaconas G (2004) Catalytic phylogenies of many DNA informational proteins. Mol
residues of the telomere resolvase ResT: a pattern similar to, Microbiol 33: 457–465.
but distinct from, tyrosine recombinases and type IB Freire R, Serrano M, Salas M & Hermoso JM (1996) Activation of
topoisomerases. J Biol Chem 279: 53699–53706. replication origins in phi29-related phages requires the
Dı́az R & Pritchard RH (1978) Cloning of replication origins recognition of initiation proteins to specific nucleoprotein
from the E. coli K12 chromosome. Nature 275: 561–564. complexes. J Biol Chem 271: 31000–31007.

FEMS Microbiol Rev 30 (2006) 321–381

c2006 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 376 C. Weigel & H. Seitz

Friedman D, Tomich P, Parsons C, Olson E, Deans R & Flamm E Heller RC & Marians KJ (2005) The disposition of nascent
(1981) lambda altSF: a phage variant that acquired the ability strands at stalled replication forks dictates the pathway of
to substitute specific sets of genes at high frequency. Proc Natl replisome loading during restart. Mol Cell 17: 733–743.
Acad Sci USA 78: 410–414. Hendrix RW, Lawrence JG, Hatfull GF & Casjens S (2000) The
Fujikawa N, Kurumizaka H, Nureki O, Terada T, Shirouzu M, origins and ongoing evolution of viruses. Trends Microbiol 8:
Katayama T & Yokoyama S (2003) Structural basis of 504–508.
replication origin recognition by the DnaA protein. Nucleic Herman-Antosiewicz A, Obuchowski M & Wegrzyn G (2001) A
Acids Res 31: 2077–2086. plasmid cloning vector with precisely regulatable copy number
Fujimura RK & Roop BC (1983) Interaction of a DNA-binding in Escherichia coli. Mol Biotechnol 17: 193–199.
protein, the product of gene D5 of bacteriophage T5, with Hermoso JM, Mendez E, Soriano F & Salas M (1985) Location of
double-stranded DNA: effects on T5 DNA polymerase the serine residue involved in the linkage between the terminal
functions in vitro. J Virol 46: 778–787. protein and the DNA of phage phi 29. Nucleic Acids Res 13:
Fuller CW & Richardson CC (1985a) Initiation of DNA 7715–7728.
replication at the primary origin of bacteriophage T7 by Hiasa H & Marians KJ (1999) Initiation of bidirectional
purified proteins. Initiation of bidirectional synthesis. J Biol replication at the chromosomal origin is directed by the
Chem 260: 3197–3206. interaction between helicase and primase. J Biol Chem 274:
Fuller CW & Richardson CC (1985b) Initiation of DNA 27244–27248.
replication at the primary origin of bacteriophage T7 by Hiasa H, Sakai H, Tanaka K, Honda Y, Komano T & Godson GN
purified proteins. Site and direction of initial DNA synthesis. J (1989) Mutational analysis of the primer RNA template region
Biol Chem 260: 3185–3196. in the replication origin (oric) of bacteriophage G4: priming
Fuller RS, Kaguni JM & Kornberg A (1981) Enzymatic replication signal recognition by Escherichia coli primase. Gene 84: 9–16.
of the origin of the E. coli chromosome. Proc Natl Acad Sci Higashitani N, Higashitani A & Horiuchi K (1993) Nucleotide
USA 78: 7370–7374. sequence of the primer RNA for DNA replication of
Fuller RS, Funnell BE & Kornberg A (1984) The dnaA protein filamentous bacteriophages. J Virol 67: 2175–2181.
complex with the E. coli chromosomal origin (oriC) and other Higashitani A, Greenstein D, Hirokawa H, Asano S & Horiuchi K
sites. Cell 38: 889–900. (1994) Multiple DNA conformational changes induced by an
George JW, Stohr BA, Tomso DJ & Kreuzer KN (2001) The tight initiator protein precede the nicking reaction in a rolling circle
linkage between DNA replication and double-strand break replication origin. J Mol Biol 237: 388–400.
repair in bacteriophage T4. Proc Natl Acad Sci USA 98: Highton PJ, Chang Y & Myers RJ (1990) Evidence for the
8290–8297. exchange of segments between genomes during the evolution
Gilbert W & Dressler D (1968) DNA replication: the rolling circle of lambdoid bacteriophages. Mol Microbiol 4: 1329–1340.
model. Cold Spring Harbor Symp Quant Biol 33: 473–484. Horiuchi K (1997) Initiation mechanisms in replication of
Giraldo R (2003) Common domains in the initiators of DNA filamentous phage DNA. Genes Cells 2: 425–432.
replication in Bacteria, Archaea and Eukarya: combined Hours C & Denhardt DT (1979) ‘Nick translation’ in Escherichia
structural, functional and phylogenetic perspectives. FEMS coli rep strains deficient in DNA polymerase I activities. Mol
Microbiol Rev 26: 533–554. Gen Genet 172: 73–80.
Giraldo R & Fernandez-Tresguerres ME (2004) Twenty years of Huang WM, Robertson M, Aron J & Casjens S (2004) Telomere
the pPS10 replicon: insights on the molecular mechanism for exchange between linear replicons of Borrelia burgdorferi. J
the activation of DNA replication in iteron-containing Bacteriol 186: 4134–4141.
bacterial plasmids. Plasmid 52: 69–83. Iandolo JJ, Worrell V, Groicher KH, et al. (2002) Comparative
Glinkowska M, Majka J, Messer W & Wegrzyn G (2003) The analysis of the genomes of the temperate bacteriophages phi
mechanism of regulation of bacteriophage lambda pR 11, phi 12 and phi 13 of Staphylococcus aureus 8325. Gene 289:
promoter activity by Escherichia coli DnaA protein. J Biol 109–118.
Chem 278: 22250–22256. Ilyina TV, Gorbalenya AE & Koonin EV (1992) Organization and
Grompone G, Ehrlich SD & Michel B (2003) Replication restart evolution of bacterial and bacteriophage primase–helicase
in gyrB Escherichia coli mutants. Mol Microbiol 48: 845–854. systems. J Mol Evol 34: 351–357.
Hall MC & Matson SW (1999) Helicase motifs: the engine that Ishmael FT, Alley SC & Benkovic SJ (2001) Identification and
powers DNA unwinding. Mol Microbiol 34: 867–877. mapping of protein–protein interactions between gp32 and
Hase T, Nakai M & Masamune Y (1989) Transcription of a region gp59 by cross-linking. J Biol Chem 276: 25236–25242.
downstream from lambda ori is required for replication of Jacob F, Brenner S & Cuzin F (1963) On the regulation of DNA
plasmids derived from coliphage lambda. Mol Gen Genet 216: replication in bacteria. Cold Spring Harbor Symp Quant Biol
120–125. 28: 329–348.
Hayes F (1998) A family of stability determinants in pathogenic James JA, Escalante CR, Yoon-Robarts M, Edwards TA, Linden
bacteria. J Bacteriol 180: 6415–6418. RM & Aggarwal AK (2003) Crystal structure of the SF3

c 2006 Federation of European Microbiological Societies FEMS Microbiol Rev 30 (2006) 321–381
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 Bacteriophage replication modules 377

helicase from adeno-associated virus type 2. Structure 11: viability of the Escherichia coli cell that lacks RNase H and
1025–1035. exonuclease V activities. Biochimie 75: 89–99.
Jiang Y, Pacek M, Helinski DR, Konieczny I & Toukdarian A Kohiyama M, Cousin D, Ryter A & Jacob F (1966) Mutants
(2003) A multifunctional plasmid-encoded replication thermosensibles d’E. coli K-12. I. Isolement et characterisation
initiation protein both recruits and positions an active helicase rapide. Ann Inst Pasteur 110: 465–486.
at the replication origin. Proc Natl Acad Sci USA 100: Konieczny I (2003) Strategies for helicase recruitment and
8692–8697. loading in bacteria. EMBO Rep 4: 37–41.
Jones JM & Nakai H (1999) Duplex opening by primosome Konieczny I, Doran KS, Helinski DR & Blasina A (1997) Role of
protein PriA for replisome assembly on a recombination TrfA and DnaA proteins in origin opening during initiation of
intermediate. J Mol Biol 289: 503–516. DNA replication of the broad host range plasmid RK2. J Biol
Jones CE, Mueser TC & Nossal NG (2000) Interaction of the
Chem 272: 20173–20178.
bacteriophage T4 gene 59 helicase loading protein and gene 41 Kornberg A & Baker TA (1992) DNA Replication. W.H. Freeman
helicase with each other and with fork, flap, and cruciform
and Company, New York, NY.
DNA. J Biol Chem 275: 27145–27154.
Korolev S, Yao N, Lohman TM, Weber PC & Waksman G (1998)
Kadyrov FA & Drake JW (2001) Conditional coupling of leading-
Comparisons between the structures of HCV and Rep helicases
strand and lagging-strand DNA synthesis at bacteriophage T4
reveal structural similarities between SF1 and SF2 super-
replication forks. J Biol Chem 276: 29559–29566.
Kaguni JM & Kornberg A (1982) The rho subunit of RNA families of helicases. Protein Sci 7: 605–610.
Krause M, Lurz R, Rückert B & Messer W (1997) Complexes at
polymerase holoenzyme confers specificity in priming M13
viral DNA replication. J Biol Chem 257: 5437–5443. the replication origin of Bacillus subtilis with homologous and
Kaguni JM & Kornberg A (1984) Replication initiated at the heterologous DnaA protein. J Mol Biol 274: 365–380.
origin (oriC) of the E. coli chromosome reconstituted with Kuzminov A (1999) Recombinational repair of DNA damage in
purified enzymes. Cell 38: 183–190. Escherichia coli and bacteriophage lambda. Microbiol Mol Biol
Kaguni JM, Fuller RS & Kornberg A (1982) Enzymatic replication Rev 63: 751–813.
of E. coli chromosomal origin is bidirectional. Nature 296: Lam AMI, Keeney D & Frick DN (2003) Two novel conserved
623–627. motifs in the hepatitis C virus NS3 protein critical for helicase
Kaiser K & Murray NE (1979) Physical characterisation of the action. J Biol Chem 278: 44514–44524.
‘Rac prophage’ in E. coli K12. Mol Gen Genet 175: 159–174. Lawrence JG & Ochman H (1998) Molecular archaeology of the
Kaplan DL & O’Donnell M (2002) DnaB drives DNA branch Escherichia coli genome. Proc Natl Acad Sci USA 95:
migration and dislodges proteins while encircling two DNA 9413–9417.
strands. Mol Cell 10: 647–657. LeClerc JE & Richardson CC (1979) Gene 2 protein of
Katayama T (2001) Feedback controls restrain the initiation of bacteriophage T7: purification and requirement for packaging
Escherichia coli chromosomal replication. Mol Microbiol 41: of T7 DNA in vitro. Proc Natl Acad Sci USA 76: 4852–4856.
9–17. Leipe DD, Aravind L & Koonin EV (1999) Did DNA replication
Kawasaki Y, Matsunaga F, Kano Y, Yura T & Wada C (1996) The evolve twice independently? Nucleic Acids Res 27: 3389–3401.
localized melting of mini-F by the combined action of the Lemonnier M, Ziegelin G, Reick T, Gomez AM, Diaz Orejas R &
mini-F initiator protein (RepE) and HU and DnaA of Lanka E (2003) Bacteriophage P1 Ban protein is a hexameric
Escherichia coli. Mol Gen Genet 253: 42–49. DNA helicase that interacts with and substitutes for
Khan SA (1997) Rolling-circle replication of bacterial plasmids.
Escherichia coli DnaB. Nucleic Acids Res 31: 3918–3928.
Microbiol Mol Biol Rev 61: 442–455.
Linderoth NA, Ziermann R, Haggård-Liungquist E, Christie GE
Khan SA (2000) Plasmid rolling-circle replication: recent
& Calendar R (1991) Nucleotide sequence of the DNA
developments. Mol Microbiol 37: 477–484.
packaging and capsid synthesis genes of bacteriophage P2.
Kodaira KI, Oki M, Kakikawa M, Watanabe N, Hirakawa M,
Yamada K & Taketo A (1997) Genome structure of the Nucleic Acids Res 19: 7207–7214.
Liu Y & Haggård-Liungquist E (1994) Studies of bacteriophage
Lactobacillus temperate phage phi g1e: the whole genome
sequence and the putative promoter/repressor system. Gene P2 DNA replication: localization of the cleavage site of the A
187: 45–53. protein. Nucleic Acids Res 22: 5204–5210.
Kogoma T (1997) Stable DNA replication: interplay between Liu J & Marians KJ (1999) PriA-directed assembly of a
DNA replication, homologous recombination, and primosome on D loop DNA. J Biol Chem 274: 25033–25041.
transcription. Microbiol Mol Biol Rev 61: 212–238. Liu LF & Wang JC (1987) Supercoiling of the DNA template
Kogoma T & Lark KG (1975) Characterization of the replication during transcription. Proc Natl Acad Sci USA 84: 7024–7027.
of Escherichia coli DNA in the absence of protein synthesis: Liu Y, Saha S & Haggård-Liungquist E (1993) Studies of
stable DNA replication. J Mol Biol 94: 243–256. bacteriophage P2 DNA replication. The DNA sequence of the
Kogoma T, Hong X, Cadwell GW, Barbard KG & Asai T (1993) cis-acting gene A and ori region and construction of a P2
Requirement of homologous recombination functions for mini-chromosome. J Mol Biol 231: 361–374.

FEMS Microbiol Rev 30 (2006) 321–381

c 2006 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 378 C. Weigel & H. Seitz

Liu J, Xu L, Sandler SJ & Marians KJ (1999) Replication fork Marintcheva B & Weller SK (2001b) Residues within the
assembly at recombination intermediates is required for conserved helicase motifs of UL9, the origin-binding protein
bacterial growth. Proc Natl Acad Sci USA 96: 3552–3555. of herpes simplex virus-1, are essential for helicase activity but
Llosa M, Gomis-Ruth FX, Coll M & de la Cruz Fd F (2002) not for dimerization or origin binding activity. J Biol Chem
Bacterial conjugation: a two-step mechanism for DNA 276: 6605–6615.
transport. Mol Microbiol 45: 1–8. Marsin S, McGovern S, Ehrlich SD, Bruand C & Polard P (2001)
Lobocka MB, Rose DJ, Plunkett G III, Rusin M, Samojedny A, Early steps of Bacillus subtilis primosome assembly. J Biol
Lehnherr H, Yarmolinsky MB & Blattner FR (2004) Genome of Chem 276: 45818–45825.
bacteriophage P1. J Bacteriol 186: 7032–7068. Marszalek J & Kaguni JM (1994) DnaA protein directs the
Lopez de Saro FJ & O’Donnell M (2001) Interaction of the beta binding of DnaB protein in initiation of DNA replication in
sliding clamp with MutS, ligase, and DNA polymerase I. Proc Escherichia coli. J Biol Chem 269: 4883–4890.
Natl Acad Sci USA 98: 8376–8380. Martı́n G, Lázaro JM, Méndez E & Salas M (1989)
Lu YB, Datta HJ & Bastia D (1998) Mechanistic studies of Characterization of the phage phi 29 protein p5 as a single-
initiator–initiator interaction and replication initiation. stranded DNA binding protein. Function in phi 29 DNA-
EMBO J 17: 5192–5200. protein p3 replication. Nucleic Acids Res 17: 3663–3672.
Luder A & Mosig G (1982) Two alternative mechanisms for Masai H & Arai K (1988) Operon structure of dnaT and dnaC
initiation of DNA replication forks in bacteriophage T4: genes essential for normal and stable DNA replication of
priming by RNA polymerase and by recombination. Proc Natl Escherichia coli chromosome. J Biol Chem 263: 15083–15093.
Acad Sci USA 79: 1101–1105. Masai H, Nomura N & Arai K (1990) The ABC-primosome. A
Ludwig W, Seewaldt E, Kilpper-Balz R, Schleifer KH, Magrum L, novel priming system employing dnaA, dnaB, dnaC, and
Woese CR, Fox GE & Stackebrandt E (1985) The phylogenetic primase on a hairpin containing a dnaA box sequence. J Biol
position of Streptococcus and Enterococcus. J Gen Microbiol Chem 265: 15134–15144.
131(Part 3): 543–551. Masai H, Asai T, Kubota Y, Arai K & Kogoma T (1994) Escherichia
Luria SE & Delbrück M (1943) Mutations of bacteria from virus coli PriA protein is essential for inducible and constitutive
sensitivity to virus resistance. Genetics 28: 491–511. stable DNA replication. EMBO J 13: 5338–5345.
Lusetti SL & Cox MM (2002) The bacterial RecA protein and the Masterson PJ, Stanley MA, Lewis AP & Romanos MA (1998) A C-
recombinational DNA repair of stalled replication forks. Annu terminal helicase domain of the human papillomavirus E1
Rev Biochem 71: 71–100. protein binds E2 and the DNA polymerase alpha-primase p68
Maaloe O & Kjeldgaard NO (1966) Control of Macromolecular subunit. J Virol 72: 7407–7419.
Synthesis. Benjamin, New York. Matsubara K & Kaiser AD (1968) Lambda dv: an autonomously
Mackiewicz P, Zakrzewska-Czerwinska J, Zawilak A, Dudek MR replicating DNA fragment. Cold Spring Harbor Symp Quant
& Cebrat S (2004) Where does bacterial replication start? Rules Biol 33: 769–775.
for predicting the oriC region. Nucleic Acids Res 32: Mayr E (1998) Two empires or three? Proc Natl Acad Sci USA 95:
3781–3791. 9720–9723.
Madsen SM, Mills D, Djordjevic G, Israelsen H & Klaenhammer McGarry KC, Ryan VT, Grimwade JE & Leonard AC (2004) Two
TR (2001) Analysis of the genetic switch and replication region discriminatory binding sites in the Escherichia coli replication
of a P335-type bacteriophage with an obligate lytic lifestyle on origin are required for DNA strand opening by initiator DnaA-
Lactococcus lactis. Appl Environ Microbiol 67: 1128–1139. ATP. Proc Natl Acad Sci USA 101: 2811–2816.
Maisnier-Patin S, Nordström K & Dasgupta S (2001) Replication McGlynn P & Lloyd RG (2000) Modulation of RNA polymerase
arrests during a single round of replication of the Escherichia by (p)ppGpp reveals a RecG-dependent mechanism for
coli chromosome in the absence of DnaC activity. Mol replication fork progression. Cell 101: 35–45.
Microbiol 42: 1371–1382. Meijer M, Beck E, Hansen FG, Bergmans HE, Messer W, von
Makeyev EV & Grimes JM (2004) RNA-dependent RNA Meyenburg K & Schaller H (1979) Nucleotide sequence of the
polymerases of dsRNA bacteriophages. Virus Res 101: 45–55. origin of replication of the E. coli K-12 chromosome. Proc Natl
Mardanov AV, Strakhova TS & Ravin NV (2004) RepA protein of Acad Sci USA 76: 580–584.
the bacteriophage N15 exhibits activity of DNA helicase. Dokl Meijer WJ, Horcajadas JA & Salas M (2001) Phi29 family of
Biochem Biophys 397: 217–219. phages. Microbiol Mol Biol Rev 65: 261–287.
Marians KJ (1996) Replication fork propagation. Escherichia coli Méndez J, Blanco L, Esteban JA, Bernad A & Salas M (1992)
and Salmonella, Cellular and Molecular Biology (Neidhardt FC, Initiation of phi 29 DNA replication occurs at the second 3 0
Curtiss III R, Ingraham J, Lin ECC, Low KB, Magasanik B, nucleotide of the linear template: a sliding-back mechanism
Reznikoff WS, Riley M, Schaechter M & Umbarger HE, eds), for protein-primed DNA replication. Proc Natl Acad Sci USA
pp. 749–763. ASM, Washington, DC. 89: 9579–9583.
Marintcheva B & Weller SK (2001a) A tale of two HSV1 helicases: Méndez J, Blanco L & Salas M (1997) Protein-primed DNA
roles of phage and animal virus helicases in DNA replication replication: a transition between two modes of priming by a
and recombination. Prog Nucleic Acid Res Mol Biol 70: 77–118. unique DNA polymerase. EMBO J 16: 2519–2527.

c 2006 Federation of European Microbiological Societies FEMS Microbiol Rev 30 (2006) 321–381
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 Bacteriophage replication modules 379

Meselson M & Stahl F (1958) The replication of DNA in E. coli. Mosig G, Colowick N, Gruidl ME, Chang A & Harvey AJ (1995)
Proc Natl Acad Sci USA 44: 671–682. Multiple initiation mechanisms adapt phage T4 DNA
Messer W (2002) The bacterial replication initiator DnaA. DnaA replication to physiological changes during T4’s development.
and oriC, the bacterial mode to initiate DNA replication. FEMS Microbiol Rev 17: 83–98.
FEMS Microbiol Rev 26: 355–374. Munson RS Jr, Harrison A, Gillaspy A, et al. (2004) Partial
Messer W & Weigel C (1996) Initation of chromosome analysis of the genomes of two nontypeable Haemophilus
replication. Escherichia coli and Salmonella, Cellular and influenzae otitis media isolates. Infect Immun 72: 3002–3010.
Molecular Biology (Neidhardt FC, Curtiss III R, Ingraham J, Nair S, Alokam S, Kothapalli S, et al. (2004) Salmonella enterica
Lin ECC, Low KB, Magasanik B, Reznikoff WS, Riley M, serovar typhi strains from which SPI7, a 134-kilobase island
Schaechter M & Umbarger HE, eds), pp. 1579–1601. ASM with genes for Vi exopolysaccharide and other functions, has
Press, Washington, DC. been deleted. J Bacteriol 186: 3214–3223.
Messer W & Zakrewska-Czerwinska J (2002) Streptomyces and Nakai H, Doseeva V & Jones JM (2001) Handoff from
Escherichia coli, model organisms for the analysis of the recombinase to replisome: insights from transposition. Proc
initiation of bacterial chromosome replication. Arch Immunol Natl Acad Sci USA 98: 8247–8254.
Ther Exp (Warsz) 50: 393–398. Nelson KE, Fouts DE, Mongodin EF, et al. (2004) Whole genome
Mesyanzhinov VV, Robben J, Grymonprez B, Kostyuchenko VA, comparisons of serotype 4b and 1/2a strains of the food-borne
Bourkaltseva MV, Sykilinda NN, Krylov VN & Volckaert G pathogen Listeria monocytogenes reveal new insights into the
(2002) The genome of bacteriophage phiKZ of Pseudomonas core genome components of this species. Nucleic Acids Res 32:
aeruginosa. J Mol Biol 317: 1–19. 2386–2395.
Meyer TF & Geider K (1982) Enzymatic synthesis of Ng JY & Marians KJ (1996) The ordered assembly of the
bacteriophage fd viral DNA. Nature 296: 828–832.
phiX174-type primosome .2. Preservation of primosome
Michel B, Ehrlich SD & Uzest M (1997) DNA double-strand
composition from assembly through replication. J Biol Chem
breaks caused by replication arrest. EMBO J 16: 430–438.
271: 15649–15655.
Miller ES, Kutter E, Mosig G, Arisaka F, Kunisawa T & Ruger W
Nordström K (2003) The replicon theory 40 years. An EMBO
(2003) Bacteriophage T4 genome. Microbiol Mol Biol Rev 67:
workshop held in Villefranche sur Mer, France, January 18–23,
86–156.
2003. Plasmid 49: 269–280.
Mindich L (2004) Packaging, replication and recombination of
Nossal NG, Dudas KC & Kreuzer KN (2001) Bacteriophage T4
the segmented genome of bacteriophage Phi6 and its relatives.
proteins replicate plasmids with a preformed R loop at the T4
Virus Res 101: 83–92.
ori (uvsY) replication origin in vitro. Mol Cell 7: 31–41.
Missich R, Weise F, Chai S, Lurz R, Pedré X & Alonso JC (1997)
O’Flaherty S, Coffey A, Edwards R, Meaney W, Fitzgerald GF &
The replisome organizer (G38P) of Bacillus subtilis
Ross RP (2004) Genome of staphylococcal phage K: a new
bacteriophage SPP1 forms specialized nucleoprotein
complexes with two discrete distant regions of the SPP1 lineage of myoviridae infecting Gram-positive bacteria with a
genome. J Mol Biol 270: 50–64. low G1C content. J Bacteriol 186: 2862–2871.
Moreira D (2000) Multiple independent horizontal transfers of Obregon V, Garcia JL, Garcia E, Lopez R & Garcia P (2003)
informational genes from bacteria to plasmids and phages: Genome organization and molecular analysis of the temperate
implications for the origin of bacterial replication machinery. bacteriophage MM1 of Streptococcus pneumoniae. J Bacteriol
Mol Microbiol 35: 1–5. 185: 2362–2368.
Morgan GJ, Hatfull GF, Casjens S & Hendrix RW (2002) Ochman H & Wilson AC (1987) Evolution in bacteria: evidence
Bacteriophage Mu genome sequence: analysis and comparison for a universal substitution rate in cellular genomes. J Mol Evol
with Mu-like prophages in Haemophilus, Neisseria and 26: 74–86.
Deinococcus. J Mol Biol 317: 337–359. Odegrip R & Haggård-Liungquist E (2001) The two active-site
Moriya S, Imai Y, Hassan AK & Ogasawara N (1999) Regulation tyrosine residues of the a protein play non-equivalent roles
of initiation of Bacillus subtilis chromosome replication. during initiation of rolling circle replication of bacteriophage
Plasmid 41: 17–29. p2. J Mol Biol 308: 147–163.
Moscoso M & Suarez JE (2000) Characterization of the DNA Park K & Chattoraj DK (2001) DnaA boxes in the P1 plasmid
replication module of bacteriophage A2 and use of its origin of origin: the effect of their position on the directionality of
replication as a defense against infection during milk replication and plasmid copy number. J Mol Biol 310: 69–81.
fermentation by Lactobacillus casei. Virology 273: 101–111. Pedulla ML, Ford ME, Houtz JM, et al. (2003) Origins of highly
Mosig G (1994) Homologous recombination. Molecular Biology mosaic mycobacteriophage genomes. Cell 113: 171–182.
of Bacteriophage T4 (Karam J, Drake JW, Kreuzer KN, Mosig G Penny D & Poole A (1999) The nature of the last universal
& Hall DH, eds), pp. 54–82. ASM, Washington, DC. common ancestor. Curr Opin Genet Dev 9: 672–677.
Mosig G (1998) Recombination and recombination-dependent Petit MA & Ehrlich SD (2002) Essential bacterial helicases that
DNA replication in bacteriophage T4. Annu Rev Genet 32: counteract the toxicity of recombination proteins. EMBO J 21:
379–413. 3137–3147.

FEMS Microbiol Rev 30 (2006) 321–381

c 2006 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 380 C. Weigel & H. Seitz

Petit MA, Dervyn E, Rose M, Entian KD, McGovern S, Ehrlich SD and priC double mutants in Escherichia coli K-12. Mol
& Bruand C (1998) PcrA is an essential DNA helicase of Microbiol 34: 91–101.
Bacillus subtilis fulfilling functions both in repair and rolling- Schekman R, Weiner JH, Weiner A & Kornberg A (1975) Ten
circle replication. Mol Microbiol 29: 261–273. proteins required for conversion of phiX174 single-stranded
Pfeiffer P & Hohn T (1983) Involvement of reverse transcription DNA to duplex form in vitro. Resolution and reconstitution. J
in the replication of cauliflower mosaic virus: a detailed model Biol Chem 250: 5859–5865.
and test of some aspects. Cell 33: 781–789. Scherzinger E, Haring V, Lurz R & Otto S (1991) Plasmid
Poranen MM & Tuma R (2004) Self-assembly of double-stranded RSF1010 DNA replication in vitro promoted by purified
RNA bacteriophages. Virus Res 101: 93–100. RSF1010 RepA, RepB and RepC proteins. Nucleic Acids Res 19:
Postow L, Hardy CD, Arsuaga J & Cozzarelli NR (2004) 1203–1211.
Topological domain structure of the Escherichia coli Scherzinger E, Ziegelin G, Bárcena M, Carazo JM, Lurz R & Lanka
chromosome. Genes Dev 18: 1766–1779. E (1997) The RepA protein of plasmid RSF1010 is a replicative
Rahmouni AR & Wells RD (1992) Direct evidence for the effect of DNA helicase. J Biol Chem 272: 30228–30236.
transcription on local DNA supercoiling in vivo. J Mol Biol Schnos M, Zahn K, Inman RB & Blattner FR (1988) Initiation
223: 131–144. protein induced helix destabilization at the lambda origin: a
Rakonjac J, Ward LJH, Schiemann AH, Gardner PP, Lubbers MW prepriming step in DNA replication. Cell 52: 385–395.
& O’Toole PW (2003) Sequence diversity and functional Seigneur M, Bidnenko V, Ehrlich SD & Michel B (1998) RuvAB
conservation of the origin of replication in lactococcal prolate acts at arrested replication forks. Cell 95: 419–430.
phages. Appl Environ Microbiol 69: 5104–5114. Seitz H, Weigel C & Messer W (2000) The interaction domains of
Ravin NV (2003) Mechanisms of replication and telomere the DnaA and DnaB replication proteins of Escherichia coli.
resolution of the linear plasmid prophage N15. FEMS Mol Microbiol 37: 1270–1279.
Microbiol Lett 221: 1–6. Serrano M, Gutiérrez J, Prieto I, Hermoso JM & Salas M (1989)
Ravin NV, Strakhova TS & Kuprianov VV (2001) The Signals at the bacteriophage phi 29 DNA replication origins
protelomerase of the phage-plasmid N15 is responsible for its required for protein p6 binding and activity. EMBO J 8:
maintenance in linear form. J Mol Biol 312: 899–906. 1879–1885.
Ravin NV, Kuprianov VV, Gilcrease EB & Casjens SR (2003) Serrano M, Gutiérrez C, Salas M & Hermoso JM (1993)
Bidirectional replication from an internal ori site of the linear Superhelical path of the DNA in the nucleoprotein complex
N15 plasmid prophage. Nucleic Acids Res 31: 6552–6560. that activates the initiation of phage phi 29 DNA replication. J
Rawlings DE & Tietze E (2001) Comparative biology of IncQ and Mol Biol 230: 248–259.
IncQ-like plasmids. Microbiol Mol Biol Rev 65: 481–496. Serwer P, Watson RH & Son M (1990) Role of gene 6 exonuclease
Richardson CC (1983) Bacteriophage T7: minimal requirements in the replication and packaging of bacteriophage T7 DNA. J
for the replication of a duplex DNA molecule. Cell 33: Mol Biol 215: 287–299.
315–317. Seufert W & Messer W (1986) Initiation of Escherichia coli
Richter S, Hagemann M & Messer W (1998) Transcriptional minichromosome replication at oriC and at protein n’
analysis and mutation of a dnaA-like gene in Synechocystis sp. recognition sites. Two modes for initiating DNA synthesis in
PCC6803. J Bacteriol 180: 4946–4949. vitro. EMBO J 5: 3401–3406.
Roberts MD, Martin NL & Kropinski AM (2004) The genome Sippy J & Feiss M (2004) Initial cos cleavage of bacteriophage
and proteome of coliphage T1. Virology 318: 245–266. lambda concatemers requires proheads and gpFI in vivo. Mol
Roth A & Messer W (1995) The DNA binding domain of the Microbiol 52: 501–513.
initiator protein DnaA. EMBO J 14: 2106–2111. Smith MP & Feiss M (1993) Sites and gene products involved in
Saito H, Tabor S, Tamanoi F & Richardson CC (1980) Nucleotide lambdoid phage DNA packaging. J Bacteriol 175: 2393–2399.
sequence of the primary origin of bacteriophage T7 DNA Soengas MS, Gutiérrez C & Salas M (1995) Helix-destabilizing
replication: relationship to adjacent genes and regulatory activity of phi 29 single-stranded DNA binding protein: effect
elements. Proc Natl Acad Sci USA 77: 3917–3921. on the elongation rate during strand displacement DNA
Salinas F & Benkovic SJ (2000) Characterization of bacteriophage replication. J Mol Biol 253: 517–529.
T4-coordinated leading- and lagging-strand synthesis on a del Solar G, Giraldo R, Ruiz-Echevarria MJ, Espinosa M & Dı́az
minicircle substrate. Proc Natl Acad Sci USA 97: 7196–7201. Orejas R (1998) Replication and control of circular bacterial
Sandler SJ (2000) Multiple genetic pathways for restarting DNA plasmids. Microbiol Mol Biol Rev 62: 434–464.
replication forks in Escherichia coli K-12. Genetics 155: Son M, Watson RH & Serwer P (1993) The direction and rate of
487–497. bacteriophage T7 DNA packaging in vitro. Virology 196:
Sandler SJ & Marians KJ (2000) Role of PriA in replication fork 282–289.
reactivation in Escherichia coli. J Bacteriol 182: 9–13. Speck C & Messer W (2001) Mechanism of origin unwinding:
Sandler SJ, Marians KJ, Zavitz KH, Coutu J, Parent MA & Clark sequential binding of DnaA initiator protein to double-
AJ (1999) dnaC mutations suppress defects in DNA stranded and single-stranded DNA in the AT-rich region of the
replication- and recombination-associated functions in priB replication origin. EMBO J 20: 1469–1476.

c 2006 Federation of European Microbiological Societies FEMS Microbiol Rev 30 (2006) 321–381
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018
 Bacteriophage replication modules 381

Stahl FW, Kobayashi I & Stahl MM (1985) In phage lambda, cos is Venkatesan M, Silver LL & Nossal NG (1982) Bacteriophage T4
a recombinator in the red pathway. J Mol Biol 181: 199–209. gene 41 protein, required for the synthesis of RNA primers, is
Stahl MM, Thomason L, Poteete AR, Tarkowski T, Kuzminov A & also a DNA helicase. J Biol Chem 257: 12426–12434.
Stahl FW (1997) Annealing vs. invasion in phage lambda Watson JD & Crick F (1953) Molecular structure of nucleic acids:
recombination. Genetics 147: 961–977. a structure for deoxyribose nucleic acid. Nature 171: 737.
Stanley E, Walsh L, van der Zwet A, Fitzgerald GF & van Sinderen Weigel C & Seitz H (2002) Strand-specific loading of DnaB
D (2000) Identification of four loci isolated from two helicase by DnaA to a substrate mimicking unwound oriC. Mol
Streptococcus thermophilus phage genomes responsible for Microbiol 46: 1149–1156.
mediating bacteriophage resistance. FEMS Microbiol Lett 182: Weigel C, Schmidt A, Seitz H, Tuengler D, Welzeck M & Messer
271–277. W (1999) The N-terminus promotes oligomerisation of the
Stayton MM & Kornberg A (1983) Complexes of Escherichia coli Escherichia coli initiator protein DnaA. Mol Microbiol 34:
primase with the replication origin of G4 phage DNA. J Biol 53–66.
Chem 258: 13205–13212. Weise F, Chai S, Lüder G & Alonso JC (1994) Nucleotide sequence
Stayton MM, Bertsch L, Biswas S, et al. (1983) Enzymatic and complementation studies of the gene 35 region of the
recognition of DNA replication origins. Cold Spring Harb Bacillus subtilis bacteriophage SPP1. Virology 202: 1046–1049.
Symp Quant Biol 47(II): 693–700. Wever GH, Fischer H & Hinkle DC (1980) Bacteriophage T7
Su’etsugu M, Shimuta TR, Ishida T, Kawakami H & Katayama T DNA replication in vitro. Electron micrographic analysis of T7
(2004) Protein associations in DnaA-ATP hydrolysis mediated DNA synthesized with purified proteins. J Biol Chem 255:
by the replicase clamp-Hda complex. J Biol Chem 280: 6528–6. 7965–7972.
Sugimoto K, Oka A, Sugisaki H, Takanami M, Nishimura A, White JH & Richardson CC (1987) Processing of concatemers of
Yasuda Y & Hirota Y (1979) Nucleotide sequence of bacteriophage T7 DNA in vitro. J Biol Chem 262: 8851–8860.
Wickner S (1984a) DNA-dependent ATPase activity associated
Escherichia coli replication origin. Proc Natl Acad Sci USA 76:
with phage P22 gene 12 protein. J Biol Chem 259:
575–579.
14038–14043.
Summer EJ, Gonzalez CF, Carlisle T, Mebane LM, Cass AM, Savva
Wickner S (1984b) Oligonucleotide synthesis by Escherichia coli
CG, LiPuma JJ & Young R (2004) Burkholderia cenocepacia
dnaG primase in conjunction with phage P22 gene 12 protein.
phage BcepMu and a family of Mu-like phages encoding
J Biol Chem 259: 14044–14047.
potential pathogenesis factors. J Mol Biol 340: 49–65.
Woese CR (2000) Interpreting the universal phylogenetic tree.
Takahashi S (1975) The starting point and direction of rolling-
Proc Natl Acad Sci USA 97: 8392–8396.
circle replicative intermediates of coliphage lambda DNA. Mol
Woese CR (2002) On the evolution of cells. Proc Natl Acad Sci
Gen Genet 142: 137–153.
USA 99: 8742–8747.
Takahashi S, Hours C, Chu A & Denhardt DT (1979) The rep
Worcel A & Burgi E (1972) On the structure of the folded
mutation. VI. Purification and properties of the Escherichia
chromosome of Escherichia coli. J Mol Biol 71: 127–147.
coli rep protein, DNA helicase III. Can J Biochem 57: 855–866. Wrobel B & Wegrzyn G (2002) Evolution of lambdoid replication
Tamanoi F, Saito H & Richardson CC (1980) Physical mapping of modules. Virus Genes 24: 163–171.
primary and secondary origins of bacteriophage T7 DNA Yuzenkova J, Nechaev S, Berlin J, Rogulja D, Kuznedelov K,
replication. Proc Natl Acad Sci USA 77: 2656–2660. Inman R, Mushegian A & Severinov K (2003) Genome of
Tatusova TA & Madden TL (1999) BLAST 2 sequences, a new tool Xanthomonas oryzae bacteriophage Xp10: an odd T-odd
for comparing protein and nucleotide sequences. FEMS phage. J Mol Biol 330: 735–748.
Microbiol Lett 174: 247–250. Zavitz KH & Marians KJ (1992) ATPase-deficient mutants of the
Taylor K & Wegrzyn G (1995) Replication of coliphage lambda Escherichia coli DNA replication protein PriA are capable of
DNA. FEMS Microbiol Rev 17: 109–119. catalyzing the assembly of active primosomes. J Biol Chem 267:
Tétart F, Desplats C, Kutateladze M, Monod C, Ackermann HW 6933–6940.
& Krisch HM (2001) Phylogeny of the major head and tail Zechner EL, Wu CA & Marians KJ (1992) Coordinated leading-
genes of the wide-ranging T4-type bacteriophages. J Bacteriol and lagging-strand synthesis at the Escherichia coli DNA
183: 358–366. replication fork. III. A polymerase-primase interaction
Thomson N, Sebaihia M, Cerdeno-Tarraga A & Parkhill J (2002) governs primer size. J Biol Chem 267: 4054–4063.
Sibling rivalry. Trends Microbiol 10: 396–397. Ziegelin G, Tegtmeyer N, Lurz R, Hertwig S, Hammerl J, Appel B
Tougu K & Marians KJ (1996) The extreme C terminus of & Lanka E (2005) The repA gene of the linear Yersinia
primase is required for interaction with DnaB at the enterocolitica prophage PY54 functions as a circular minimal
replication fork. J Biol Chem 271: 21391–21397. replicon in Escherichia coli. J Bacteriol 187: 3445–3454.
Toussaint A & Merlin C (2002) Mobile elements as a combination Zylicz M, Liberek K, Wawrzynow A & Georgopoulos C (1998)
of functional modules. Plasmid 47: 26–35. Formation of the preprimosome protects lambda O from RNA
Tye BK (2000) Insights into DNA replication from the third transcription-dependent proteolysis by ClpP/ClpX. Proc Natl
domain of life. Proc Natl Acad Sci USA 97: 2399–2401. Acad Sci USA 95: 15259–15263.

FEMS Microbiol Rev 30 (2006) 321–381

c 2006 Federation of European Microbiological Societies
Published by Blackwell Publishing Ltd. All rights reserved

Downloaded from https://academic.oup.com/femsre/article-abstract/30/3/321/546048

by guest
on 03 April 2018