Blackwell Science, LtdOxford, UKMMIMolecular Microbiology0950-382XBlackwell Publishing Ltd, 2004? 2004531918Review ArticleProphagechromosome interactionC. Canchaya, G. Fournous and H.

Brssow

Molecular Microbiology (2004) 53(1), 918

doi:10.1111/j.1365-2958.2004.04113.x

MicroReview
The impact of prophages on bacterial chromosomes

Carlos Canchaya, Ghislain Fournous and

Harald Brssow*
Nestl Research Centre, Nutrition and Health
Department/Functional Microbiology Group, CH-1000
Lausanne 26 Vers-chez-les-Blanc, Switzerland.
Summary
Prophages were automatically localized in sequenced bacterial genomes by a simple semantic
script leading to the identification of 190 prophages
in 115 investigated genomes. The distribution of
prophages with respect to presence or absence in a
given bacterial species, the location and orientation
of the prophages on the replichore was not
homogeneous. In bacterial pathogens, prophages
are particularly prominent. They frequently encoded
virulence genes and were major contributors to the
genetic individuality of the strains. However, some
commensal and free-living bacteria also showed
prominent prophage contributions to the bacterial
genomes. Lysogens containing multiple sequencerelated prophages can experience rearrangements
of the bacterial genome across prophages, leading
to prophages with new gene constellations. Transfer
RNA genes are the preferred chromosomal integration sites, and a number of prophages also carry
tRNA genes. Prophage integration into protein coding sequences can lead to either gene disruption or
new proteins. The phage repressor, immunity and
lysogenic conversion genes are frequently transcribed from the prophage. The expression of the
latter is sometimes integrated into control circuits
linking prophages, the lysogenic bacterium and its
animal host. Prophages are apparently as easily
acquired as they are lost from the bacterial chromosome. Fixation of prophage genes seems to be
restricted to those with functions that have been coopted by the bacterial host.
Accepted 27 February, 2004. *For correspondence. E-mail
harald.bruessow@rdls.nestle.com; Tel. (+41) 21 785 8676; Fax (+41)
21 785 8544. These authors contributed equally to the database
mining.

2004 Blackwell Publishing Ltd

Introduction
Microbial genomics revealed that, in some bacteria, substantial amounts of the bacterial DNA is in fact prophage
DNA (Canchaya et al., 2003; Casjens, 2003). It also
became increasingly clear that prophage DNA has played
an important role in the evolution of bacterial pathogenicity (Boyd and Brssow, 2002). These data have changed
our understanding of phagebacterium interaction from a
simple parasitehost relationship into a two-way coevolution of viral and bacterial genomes. Here, we provide
a short overview on the impact of prophage integration on
bacterial genome structure and diversification.
Methodological problems of prophage identification
Prophage identification is not an exact science, but a
labour-intensive, empirical approach that needs a lot of
insight (Casjens, 2003). To keep pace with the increasing
number of sequenced bacterial genomes, a simple script
was written in our laboratory (Fournous, 2003) that transforms the GenBank file of a bacterial genome into a FASTA
file of all its open reading frames (ORFs). A BLASTX search
is then conducted with each individual ORF, and the output of the significant database matches is searched by a
semantic program for its annotations using positive (e.g.
phage, integrase, tail, capsid, terminase, portal) and negative (e.g. macrophage, transposase, transposon, insertion) keywords. The hits are plotted for each ORF position
of the genome, and a further script transforms this hit list
into prophage genomes by performing a neighbour analysis for further phage-like genes around each individual
hit. An example of the automatic data output is shown for
Salmonella enterica serovar Typhi strain CT-18. It shows
all phage hits (ticks on the outer circle) and the prophage
identification by neighbour analysis (green bars) (Fig. 1A).
Differences between the automatic and manual annotations by Casjens (2003) shown in the second circle were
minimal and concern only remotely phage-like elements
(Sti5, Sti6; Fig. 1A, centre).
The maps of the candidate prophage elements are
shown in Fig. 1A. Sti1 and Sti4 are clearly lambda-like
prophages in their genetic organization, but not in their
sequence; the structural genes resemble a Photorhabdus

10 C. Canchaya, G. Fournous and H. Brssow

ori
Sti10
Sti9

R1

1:
4:
3:
8:
9:

Sti8

I 2
3 4T
I

Sti1
5

10:
2:
4:
5:
6:
7:

6
7
8

Sti2
1
T

Sti3

Sti4
Sti7
Sti6

Sti5
ter

ori

C
n

ter
prophage. Sti3 is a Mu-like whereas Sti8 and Sti9 are P2like prophages. The well-known gene map of these three
phage types allowed the tentative identification of morons
(extra non-phage genes inserted into the genomes of
temperate phages) (Juhala et al., 2000). Notably, the
morons encode important candidate virulence factors for

Fig. 1. Prophage distribution.

A. Salmonella enterica serovar Typhi strain
CT18 and its prophages. The two outer circles
represent the circular genome map. Prophages
identified by Casjens (2003) are indicated by
red boxes. The outermost ring displays the
computer output of our prophage detection program (Fournous, 2003). Each tick is a potential
phage hit. Prophages identified by automatic
neighbour analysis in our program are marked
by a small green box. The gene maps of the
candidate prophage elements are displayed
within the circle with their Sti number as identifier; they are not to scale. The prophage proteins are coloured according to putative
function in the top five maps representing
lambda- (1, 4), Mu- (3) and P2-like (8, 9) prophages. The colour code is as follows: red, lysogeny; violet, lysis; green, head; brown, headto-tail; blue, tail; mauve, tail fibre; black,
lysogenic conversion; orange, DNA replication;
yellow, transcriptional regulation. In the P4-like
(10) and prophage remnants (bottom five
maps), phage-related genes are coloured red,
and candidate lysogenic conversion genes are
in black. The following candidate lysogenic conversion genes were identified: 1, msgA; 2,
enterohaemolysin-1; 3, enterohaemolysin-2; 4,
type II secretion; 5, restrictionmodification; 6,
sopE; 7, protein kinase/phosphatase; 8, pertussis-like toxin subunits A and B; 9, sspH. Recombinases are identified with I (integrase), T
(transposase) and R (invertase).
B. Location of prophages from 115 investigated
genomes projected on an idealized replichore
of a bacterial genome (ori: origin, ter: terminus).
C. The ordinate gives the number n of prokaryotic genomes, the abscissa the number p of
prophages per genome. The height of the bars
gives the number of genomes showing the indicated number of prophages.

P
this typhoid fever-causing Salmonella serovar (see
Fig. 1A legend for annotations). Sti10 resembles a P4-like
prophage and also contains a potential moron (Fig. 1A).
Sti7 represents a tail gene cluster from a P2-like phage
flanked by a pin invertase. The presence of recombinases
(integrase, transposase, invertase) points to the mobile
2004 Blackwell Publishing Ltd, Molecular Microbiology, 53, 918

Prophagechromosome interaction 11
character of these DNA elements (Fig. 1A). In contrast,
Sti2 and 5 show isolated phage genes that flank pertussislike toxin genes (Sti2) and Salmonella serovar Typhimurium virulence genes (sspH and msgA) (Sti5). Sti5 and 6
were identified by a manual search (Casjens, 2003), but
not by the automatic program (Fig. 1A). On the basis of
theoretical models, prophages are predicted to decay by
deletion of phage genes resulting in prophage remnants.
In contrast, genes conferring a selective advantage such
as a moron should be retained (Lawrence et al., 2003).
Sti5 fits this prediction, but its prophage derivation must
await the discovery of related, but less decayed prophage
elements.
Distribution and orientation of prophages
One hundred and ninety prophages identified by the program in 115 sequenced prokaryotic genomes were projected on a hypothetical replichore (one half of the
chromosome) (Fig. 1B). The position of the prophages is
given as a percentage of the distance from ori (origin) to
ter (terminus of replication) regardless of which half (left
or right) of the chromosome the prophages were identified
on. The representation does not take account of the variable length of the different bacterial replichores. Overall,
the prophages were relatively evenly distributed, but the
density of the prophages was greater near ter than near
ori. Approximately half the sequenced prokaryotic
genomes do not contain either prophages or well-defined
remnants (Fig. 1C). This group contains all sequenced
Archaea and most intracellular eubacterial pathogens. It
has been argued that the latter have undergone recent
genome contractions in which all non-essential genes for
the intracellular niche have been lost. Prophages have,
however, been observed in insect endosymbionts (phage
APSE-1 and Wolbachia prophages), suggesting that intracellular bacteria can harbour phages and prophages. A
quarter of all genomes contain one or two prophages.
Approximately one-tenth of the genomes contributed the
majority of the prophage sequences (Fig. 1C). Bacteria
containing six or more prophages in their genome are
mainly pathogens, with the notable exception of Lactococcus lactis, an organism used in cheese fermentation and
under extreme pressure from bacteriophages. The comparisons of genomes from the same species suggest a
stochastic process: prophages might be present or absent
(e.g. Streptococcus agalactiae) (Fig. 2A) or different
strains within a species may contain one, two or three
prophages (e.g. Staphylococcus aureus) (Fig. 2B).
Prophages showed a preferred orientation for integration: the structural genes pointed mostly in the direction
of the majority of the surrounding bacterial genes. When
prophages changed position from the right to the left replichore, they also changed their orientation (Smoot et al.,
2004 Blackwell Publishing Ltd, Molecular Microbiology, 53, 918

2002a; Nakagawa et al., 2003; see also the circled prophage in Fig. 2B). This strong bias in phage orientation is
still unexplained, although avoidance of RNA and DNA
polymerase collision and interference with the normal
functioning of terminus functions (dif site) have been proposed (Campbell, 2002).
Prophages contribute to the genetic individuality of a
bacterial strain
The role of mobile DNA in the diversification of bacterial
genomes becomes apparent when the sequenced
genomes from two different strains of the same species
are aligned in a dot plot. For example, the alignment of
two S. agalactiae serotypes showed about a dozen small
gaps (Fig. 2A). The gaps corresponded nearly exclusively
to mobile DNA; bioinformatic analysis revealed integrative
plasmids, transposons and two prophages. Ten gaps
showed an atypical nucleotide composition suggesting
lateral gene transfer, and eight of the genomic islands
were associated with integrases/transposases (Glaser
et al., 2002). The variable genome parts defined by comparative genomics (Glaser et al., 2002), dot plot alignment
and microarray hybridization (Tettelin et al., 2002) showed
excellent concordance. The relative contribution of prophages to the strain-specific DNA varied sometimes for
strains from the same species. In some Staphylococcus
aureus strain comparisons, prophages were the major
contributors to variability (Fig. 2B) (Kuroda et al., 2001),
whereas in other comparisons, they competed with
genomic islands and transposons for this role (Baba et al.,
2002). An extreme case is presented by Streptococcus
pyogenes in which all major gaps in the alignment of
different M serotypes could be traced to prophage integration events (Smoot et al., 2002a). As all S. pyogenes
and many S. aureus prophages encode proven or suspected virulence factors, the prophage-imposed diversity
between the strains might be of clinical relevance (Beres
et al., 2002).
The role of prophages in the individuality of the
strains is not restricted to Gram-positive bacteria, but
can also be seen in Gram-negative bacteria. Two Salmonella Typhi strains could be aligned over the entire
genome when allowing for a large chromosomal inversion across two rRNA gene clusters. From 113 ORFs
specific for one or other strain, 76 were prophage genes
(Deng et al., 2002). In a Salmonella serovar Typhi and
Typhimurium comparison, two chromosomal inversions
and 12 larger alignment gaps were identified. Nine of
the gaps can be traced to prophages or prophage remnants. The gaps represent either prophage insertion/
deletion events or prophage replacements at a given
chromosomal position. Escherichia coli can serve as a
dramatic example in which half the 1 Mb DNA that dis-

12 C. Canchaya, G. Fournous and H. Brssow

A
2000000

2603V/R

SA2

Tn916

SA1

Streptooccus agalactiae NEM316

MSSA476

2500000

SCC

23

Staphylococcus aureus NCTC 8325

ori

Fig. 2. Prophages contribute to the individuality of bacterial strains.

A. Dot plot comparison of the DNA sequences from Streptococcus
agalactiae strains 2603V/R (vertical) and NEM316 (horizontal). The
red boxes identify mobile DNA elements characterized by comparative genomics (Glaser et al., 2002) and microarray analysis (Tettelin
et al., 2002). The gaps in the alignment are located by thin horizontal
and vertical lines. The numbers are bp positions. The two prophages
are annotated with SA1 and SA2. The dot plot was done with the
DOTTUP program (http://www.emboss.org/), using the direct and
reverse sequence; word size was 15, output format was POSTSCRIPT,
the program was run in the direct and reverse direction and the figures
were combined.
B. Dot plot of the DNA sequence alignment from the Staphylococcus
aureus strains MSSA476 and NCTC 8325. Prophages are annotated
as red boxes. A prophage that changed the replichore and the orientation is circled in red. The MSSA476 sequencing data were produced
by the Microbial Sequencing group at the Sanger Institute and can
be obtained from ftp://ftp.sanger.uk/pub/sa/ (NCBI accession number
NC_002953). The NCTC8325 sequencing data were produced by the
Staphylococcus aureus Genome Sequencing Project of J. Iandolo at
the University of Oklahoma Health Sciences Center (NCBI accession
number NC_002954).
C. Microarray analysis in the gut commensal Lactobacillus johnsonii.
Outer circle: eight L. johnsonii strains were hybridized against the
reference strain NCC533. Inner circle: eight different Lactobacillus
species were hybridized against the NCC533 strain (ring 1): L. gasseri ATCC19992 and DSM20234 (2, 3), L. helveticus CNRZ303 (4),
L. crispatus DSM20584 (5), L. gallinarum ATCC33199 (6), L. amylovorus DSM20531 (7), L. acidophilus ATCC4356 (8), L. reuteri
DSM20016 (9) and L. plantarum ATCC14917 (10). Red, blue and
black segments indicate regions that do or not do cross-hybridize with
NCC533 and that are not represented on the microarray respectively.
The outermost thin circle locates prophages and prophage remnants
(orange boxes), integrase (red) and transposase (green) genes. For
details, see Ventura et al. (2003a).

tinguish E. coli O157:H7 from K-12 are accounted for by

prophage DNA (Ohnishi et al., 2001). Interestingly,
prophage comparisons between strains sharing a very
recent common ancestor (E. coli O157 Sakai and
EDL933 or the two M3 S. pyogenes strains) already
showed modular exchange reactions (Makino et al.,
1999) (Fig. 3B, two tail fibre genes in P5 versus 315.2
comparison), suggesting that prophages are a highly
dynamic part of the bacterial genomes.
Even in the comparisons of some closely related sister
species such as Listeria innocua and L. monocytogenes
(Glaser et al., 2001) or Lactobacillus johnsonii and L. gasseri (Pridmore et al., 2004; T. R. Klaenhammer and F.
Desiere, unpublished), prophages account for a substantial part of the major alignment gaps.
DNADNA microarray analysis

Lactobacillus johnsonii

Microarray analysis demonstrated that, in Salmonella

serovar Typhimurium strains, genetic differences were
essentially limited to prophage content (Porwollik et al.,
2002; Chan et al., 2003). The same was true for S.
pyogenes (Smoot et al., 2002b). Vibrio cholerae strains
recovered as far back as 1910 demonstrated only 1%
difference between the strains (Dziejman et al., 2002).
The prepandemic clinical isolates and an environmen 2004 Blackwell Publishing Ltd, Molecular Microbiology, 53, 918

Prophagechromosome interaction 13
tal non-toxigenic O1 El Tor isolate lacked the CTX
prophage encoding the cholera toxin CT, identifying
this prophage as a major determinant of genetic variability. Microarray analysis showed that lateral gene
transfer has played a fundamental role in the diversification of S. aureus: up to 22% of the genome comprised variable genetic material. Prophages comprised
17% of the total amount of 250 kb mobile and variable
DNA (Fitzgerald et al., 2001). As prophages from bacterial pathogens frequently encode virulence factors,
one could suspect that the analysis of pathogenic bacteria overestimates the contribution of prophages to
the genetic individuality of a given strain. However,
similar data were obtained with the gut commensal
Lactobacillus johnsonii. Hybridization of eight molecularly distinct strains of L. johnsonii revealed that the
prophage DNA contributed approximately half the
strain-specific DNA of the reference strain (Fig. 2C)
(Ventura et al., 2003a). When lactobacilli belonging to
seven different Lactobacillus species were included in
that analysis, genes related to the prophages of the
reference strain were not detected. This observation
demonstrates that mobile DNA is not necessarily
widely distributed across species borders.

3'
4'
5'
6'

2'

It is not rare that different prophages from the same

lysogen share DNA sequence identity. As these regions
are targets for homologous recombination, it was predicted that prophages mediate rearrangements of bacterial chromosomes (Brssow and Hendrix, 2002). Several
genome alignments support this prediction. For example,
a Japanese S. pyogenes M3 strain differed from an American M3 isolate mainly by two sequential DNA inversions
(Nakagawa et al., 2003). One inversion occurred across
two prophages (Fig. 3A). Genomics analysis suggested
that the cross-over point was located in the lysis modules.
As the lysogenic conversion genes from S. pyogenes
prophages are encoded downstream of the lysis genes,
the cross-over results in a reshuffling of virulence genes
between prophages (Fig. 3B). This might allow a wider
horizontal spread of these genes as they become associated with phages with potential new host ranges and
belonging to new immunity groups. This flexibility extends
the possibilities of conversion gene permutation in
polylysogenic hosts such as S. pyogenes. A spectacular
case of apparently prophage-mediated recombinations is

M3 315
1'

Prophages as target regions for chromosomal

rearrangement

1'

B
6
5

com

1
2
3
4

ssa

M3 SSI-1

ssa

2'

com

Fig. 3. Genome rearrangements in Streptococcus pyogenes.

A. The US S. pyogenes serotype M3 strain MGAS315 (top) was aligned with the Japanese S. pyogenes serotype M3 strain SSI-1 (bottom) using
the Artemis comparison tool. MGAS315 shows the likely original constellation, SSI-1 experienced a first inversion across the duplicated com
genes and a second inversion across prophages P5 and P6. The position of the prophages is indicated by the red boxes.
B. Alignment of the prophages P5 and P6 with the corresponding prophages 315.1 and 315.2 locating the site of recombination within the lysis
cassette. The shading connects regions of high DNA sequence identity between the compared phages. The modular structure of the prophages
is colour-coded (key in Fig. 1).
2004 Blackwell Publishing Ltd, Molecular Microbiology, 53, 918

14 C. Canchaya, G. Fournous and H. Brssow

somal site to the left of Xt1 (Fig. 4B and C) and, finally,
an inversion across the prophages Xt8 and Xt2 sharing
highly related integrase genes (Fig. 4C), leading to nearly
perfectly aligned genomes (Fig. 4D). A genome inversion
between the two sequenced O157 E. coli strains can also
be traced to a recombination between prophages 993O
and 933P. However, prophages are not a privileged
recombination site: two Yersinia pestis genomes differ by
a large number of small genome inversions and contain
numerous prophage remnants, but only three inversions
are flanked at one side by a prophage sequence (Deng

presented by the Gram-negative plant pathogen Xylella

fastidiosa. The pathovars 9a5c and Temecula shared DNA
sequence identity essentially over the entire genome, but
showed a very complex alignment pattern, suggesting
three successive chromosomal inversion events (Van
Sluys et al., 2003). Genomics analysis suggested the following sequence of events (Fig. 4): inversion across
prophage elements XfP1 and XfP2 sharing extensive
DNA sequence over their entire length (Canchaya et al.,
2003) (Fig. 4A and B), followed by an illegitimate recombination between prophage remnant Xt10 and a chromo-

B
A

Xylella fastidiosa Temecula

? Xt1

XfP1

9.a.5.c

XfP2

2500000

Xt10

2500000

observed
C

D
Xt8

Xt2

reconstructed

Fig. 4. Genome rearrangements in Xylella fastidiosa.

A. Dot plot comparison of X. fastidiosa pathovar Temecula (horizontal) and pathovar 9a5c (vertical). The position of the prophages is indicated
by red boxes. The chromosome inversion between prophages XfP1 and XfP2 leads to the dot plot shown in (B). An inversion across prophage
Xt10 and an unknown chromosomal site leads to the dot plot depicted in (C). A final inversion between prophages Xt8 and Xt2 results in the
near-perfect alignment of the two genomes shown in (D).
2004 Blackwell Publishing Ltd, Molecular Microbiology, 53, 918

Prophagechromosome interaction 15
tional protein. For example, lambdoid phage 21 inserts
within the isocitrate dehydrogenase gene and introduces
an alternative 165 bp 3 end for that gene (Campbell et al.,
1992). In rare cases, the phage recombination site attP is
located within the phage integrase gene, and prophage
integration leads to an altered int gene, which apparently
stabilizes the lysogen (Magrini et al., 1999), or the bacterial attB site complements the int gene (Bruttin et al.,
1997).

et al., 2002). In other bacteria, rDNA and duplicated bacterial genes were used for genome rearrangements (com
genes in the case of Fig. 3A).
Integration sites
Many prophages from both Gram-negative and Grampositive bacteria integrate into tRNA genes in a preferred
orientation (Campbell, 1992; 2002). Mostly, but not always
(Ventura et al., 2003a), the attP site of the phage reconstitutes the tRNA upon integration. Interestingly, an
increasing number of prophages are described that carry
tRNA genes (Fig. 5D). The phage-encoded tRNA differed
from the chromosomal tRNA and were transcribed in the
lysogen (Ventura et al., 2004). Such constellations might
alleviate the consequences of prophage integration
events into tRNA genes that do not reconstitute the original tRNA gene. However, integration into tRNA genes is
far from being universal. Integration into intergenic regions
and into ORFs has also been described (Canchaya et al.,
2002). The prophage attachment site frequently faithfully
complements the coding sequence of the protein. However, cases of inactivation of the protein-encoding function
have also been described, a well-known case being the
negative lysogenic conversion of the lipase gene by an S.
aureus phage (Lee and Iandolo, 1986). In addition, integration into an ORF can also lead to an altered but func-

Transcription of prophage genomes

In the lambdoid phages, it is known that most genes in
the integrated prophages are not transcribed, and the bulk
of the prophage genes are thus passive genetic cargo to
the lysogen. However, this statement does not apply to all
lambda prophage genes: lambda genes bor and lom are
expressed from the lysogen and confer serum resistance
to the E. coli lysogen during in vivo growth (Barondess
and Beckwith, 1990). Two lambdoid prophages from E.
coli O157 encode Shiga-like toxins (Stx), the major virulence factor of this important food pathogen. Their expression is tightly controlled. Expression is achieved during
prophage induction and when iron, sensed by the Fur
transcriptional regulator, becomes growth limiting (Wagner et al., 2002). Stx induces bleeding into the gut, and
iron thus becomes available from decaying blood cells. Stx

ori

2356

124

irp6A

2162
2159 pi3
adhA

DT

539

SPIF

625

pi2

mf2spd1
speC

1734
hmuO

pi1

mf2

922

1520

D
mf4

894

1061
dtxR

1296

Fig. 5. Transcription control of prophage genes.

A. Genome map of the Corynebacterium diphtheriae strain NCTC13129 (Cerdeno-Tarraga et al., 2003). The transcriptional regulator gene dtxR
is shown together with all genes regulated by DtxR (arrows) including the DT toxin encoded by the corynephage F shown in (B). The transcribed
lysogenic conversion genes from a Streptococcus pyogenes prophage are shown in (C). They are under the control of SPIF released from
pharyngeal cells (Broudy et al., 2001). Map of Lactobacillus plantarum prophage Lp1 in (D). The modular structure of the prophages is indicated
by a colour code (key in Fig. 1). A similar modular structure is proposed for the corynephage despite the phylogenetic distance separating the
bacterial hosts. The horizontal arrows next to the prophage maps indicate the prophage transcripts. In (B) and (C), only the right prophage end
was investigated for transcription.
2004 Blackwell Publishing Ltd, Molecular Microbiology, 53, 918

16 C. Canchaya, G. Fournous and H. Brssow

has no physiological export system, but E. coli O157 has
learned to avoid this suicidal production by charging
bystander intestinal E. coli strains with its synthesis via
infection with the induced prophage (Gamage et al.,
2003). In a number of Gram-positive pathogens, important
virulence factors are encoded between the phage lysin
gene and attR, the right attachment site (Beres et al.,
2002). Only a low transcription level of these genes was
observed in broth growth, whereas contact with pharyngeal cells induced the expression of these virulence
genes in S. pyogenes prophages (Broudy et al., 2001)
(Fig. 5C). The expression of the diphtheria toxin in
Corynebacterium diphtheriae is regulated via the DtxR
transcriptional regulator that binds in an iron-dependent
way to operators of many bacterial genes (Fig. 5A),
including the diphtheria toxin gene encoded by a corynephage (Fig. 5B). DtxR and Fur are the master regulators of large iron regulons, and prophage virulence gene
expression thus comes under the control of the host bacterium. In contrast, constitutive transcription of moron-like
extra phage genes was seen in commensals and freeliving bacteria (Ventura et al., 2003b). The morons were
located in the vicinity of the left and right phage attachment sites (Fig. 5D). Interestingly, morons from pathogens
and commensals sharing the same habitat (e.g. S. pyogenes and Lactobacillus plantarum both isolated from the
oral cavity) also shared sequence-related putative
lysogenic conversion genes in their prophages (e.g. mftype DNases; Fig. 5C and D).
In S. pyogenes and S. aureus, the prophage moron
transcription is increased by prophage induction via an
increase in copy number (Broudy et al., 2002; Sumby and
Waldor, 2003). In these cases, it is not clear whether these
genes are of ecological benefit to the phage, the lysogen
or both (Broudy and Fischetti, 2003). Microarray analysis
in a number of systems revealed that prophage genes
figured prominently under the upregulated genes of the
lysogen under stress conditions. This was observed in S.
pyogenes growing in phagocytes (Voyich et al., 2003),
recovered from mice (Kazmi et al., 2001) and human
patients or experiencing growth temperature changes
(Smoot et al., 2001).
A prominent upregulation of prophage gene expression
was also seen in Gram-negative lysogens upon infection
of animals (Dozois et al., 2003) or when changed from
planktonic to biofilm growth (Pseudomonas aeruginosa;
Whiteley et al., 2001).
Outlook
There is currently a lively discussion about the relative
contribution of vertical and horizontal elements of evolution in bacteria. Some researchers argue that the tree-like
representations of Darwinian evolution should be replaced

by web-like phylogenies as a result of the rampant effect

of horizontal gene transfer (HGT) in bacteria (Doolittle,
1999). Other researchers claim that the role of HGT for
the evolution of bacterial genomes is overstated (Kurland
et al., 2003). The sequencing of multiple strains from
the same bacterial species demonstrated that HGT
accounted for the majority of the intraspecies genome
differences. Prophages are the major contributors to
genome diversification in some species; in others, prophages compete with integrative plasmids, transposons or
pathogenicity islands (which themselves show links to
prophages) for this role. Still other sequenced bacteria
lack prophages as a result of sampling bias (e.g. Streptococcus pneumoniae) or perhaps a genuine absence of
phages (e.g. Helicobacter pylori ).
Prophages played an important and, in some cases, a
decisive role in the emergence of bacterial pathogens. V.
cholerae, E. coli O157 and C. diphtheriae are examples
in which the disease-specifying toxins are encoded by
prophages. S. aureus, S. pyogenes and Salmonella are
examples of pathogens in which many disease-modifying
factors are encoded by multiple prophages and each individual prophage contributes only an incremental virulence
increase.
With respect to the impact of HGT on bacterial evolution, we believe that two processes must be distinguished.
In the field of medical microbiology and epidemiology, we
observe events that occur across time frames that rarely
exceed 100 years. Obviously, vertical modes of bacterial
genome evolution are not very efficient over these short
time periods. Therefore, the emergence of new pathogens
is likely to rely on the acquisition of lateral DNA (or loss
of DNA or a combination of both). In this context, phages
are an ideal carrier for horizontal DNA and thus a likely
motor for short-term bacterial diversification. But does this
process influence the structure of bacterial genomes over
evolutionary time periods? Over this time scale, it is not
the acquisition, but the fixation, of prophage sequences
that counts (Lawrence and Roth, 1999). In our survey, we
have seen little, if any, evidence that prophages were fixed
into the chromosome of a bacterial species. Prophages
are apparently acquired as easily as they are lost from the
chromosome. A site occupied by a prophage in one strain
might be empty in another strain or occupied by another
prophage. Within the confines of a bacterial species, a few
cases of closely related, although not identical, prophages
were reported that occupied the same chromosomal site
in different bacteria (Casjens, 2003). The Neisseria prophages may be the clearest examples of long residence
times during prophage decay. However, this case might
be an exception as it relates to Mu-like phages that very
rarely excise precisely once they are inserted.
However, fixing of entire prophages would not be
expected as it makes no evolutionary sense to bacteria.
2004 Blackwell Publishing Ltd, Molecular Microbiology, 53, 918

Prophagechromosome interaction 17
The mechanism would no doubt be characterized by fixation of particular phage genes, the function of which has
been co-opted by the host. Prophage genes without selective value to the host are therefore likely to be deleted.
Owing to the lack of surrounding prophage genes, completely fixed prophage genes are thus difficult to detect.
In the case of bacterial pathogens, a conspicuous observation is virulence genes flanked by isolated phage-like
genes. Examples are Shiga toxin genes in Shigella dysenteriae (McDonough and Butterton, 1999), sopE2 in Salmonella Typhimurium, sspH and pertussis-like toxin genes
in Salmonella Typhi (Fig. 1A). As related genes are found
as lysogenic conversion genes in well-established prophages (Stx prophages in E. coli O157, SopE prophages in
S. Typhimurium), we probably have here examples of fixed
prophage genes. As these genes are likely to extend the
ecological range of their bacterial hosts, prophages might
have a greater impact on bacterial genomes than is indicated by the presence of clear-cut prophage DNA
sequences in the sequenced bacterial genomes.

Acknowledgements
We thank Sherwood Casjens, Chris Blake, Anne Constable
and Anne Bruttin for critical reading of the manuscript, and
the Swiss National Science foundation for financial support
to Carlos Canchaya (research grant 5002-057832).

References
Baba, T., Takeuchi, F., Kuroda, M., Yuzawa, H., Aoki, K.,
Oguchi, A., et al. (2002) Genome and virulence determinants of high virulence community-acquired MRSA. Lancet
359: 18191827.
Barondess, J.J., and Beckwith, J. (1990) A bacterial virulence
determinant encoded by lysogenic coliphage lambda.
Nature 346: 871874.
Beres, S.B., Sylva, G.L., Barbian, K.D., Lei, B., Hoff, J.S.,
Mammarella, N.D., et al. (2002) Genome sequence of a
serotype M3 strain of group A Streptococcus: phageencoded toxins, the high-virulence phenotype, and clone
emergence. Proc Natl Acad Sci USA 99: 1007810083.
Boyd, E.F., and Brssow, H. (2002) Common themes among
bacteriophage-encoded virulence factors and diversity
among the bacteriophages involved. Trends Microbiol 10:
521529.
Broudy, T.B., and Fischetti, V.A. (2003) In vivo lysogenic
conversion of Tox() Streptococcus pyogenes to Tox(+)
with lysogenic streptococci or free phage. Infect Immun 71:
37823786.
Broudy, T.B., Pancholi, V., and Fischetti, V.A. (2001)
Induction of lysogenic bacteriophage and phageassociated toxin from group A streptococci during coculture with human pharyngeal cells. Infect Immun 69:
14401443.
Broudy, T.B., Pancholi, V., and Fischetti, V.A. (2002) The in
vitro interaction of Streptococcus pyogenes with human
2004 Blackwell Publishing Ltd, Molecular Microbiology, 53, 918

pharyngeal cells induces a phage-encoded extracellular

DNase. Infect Immun 70: 28052811.
Brssow, H., and Hendrix, R.W. (2002) Phage genomics:
small is beautiful. Cell 108: 1316.
Bruttin, A., Foley, S., and Brssow, H. (1997) The sitespecific integration system of the temperate Streptococcus
thermophilus bacteriophage phiSfi21. Virology 237: 148
158.
Campbell, A.M. (1992) Chromosomal insertion sites for
phages and plasmids. J Bacteriol 174: 74957499.
Campbell, A.M. (2002) Preferential orientation of natural
lambdoid prophages and bacterial chromosome organization. Theor Popul Biol 61: 503507.
Campbell, A., Schneider, S.J., and Song, B. (1992) Lambdoid phages as elements of bacterial genomes (integrase/
phage21/Escherichia coli K-12/icd gene). Genetica 86:
259267.
Canchaya, C., Desiere, F., McShan, W.M., Ferretti, J.J.,
Parkhill, J., and Brssow, H. (2002) Genome analysis of
an inducible prophage and prophage remnants integrated
in the Streptococcus pyogenes strain SF370. Virology 302:
245258.
Canchaya, C., Proux, C., Fournous, G., Bruttin, A., and Brssow, H. (2003) Prophage genomics. Microbiol Mol Biol Rev
67: 238276.
Casjens, S. (2003) Prophages and bacterial genomics: what
have we learned so far? Mol Microbiol 49: 277300.
Cerdeno-Tarraga, A.M., Efstratiou, A., Dover, L.G., Holden,
M.T., Pallen, M., Bentley, S.D., et al. (2003) The complete
genome sequence and analysis of Corynebacterium diphtheriae NCTC13129. Nucleic Acids Res 31: 65166523.
Chan, K., Baker, S., Kim, C.C., Detweiler, C.S., Dougan, G.,
and Falkow, S. (2003) Genomic comparison of Salmonella
enterica serovars and Salmonella bongori by use of an S.
enterica serovar typhimurium DNA microarray. J Bacteriol
185: 553563.
Deng, W., Burland, V., Plunkett, G., III, Boutin, A., Mayhew,
G.F., Liss, P., et al. (2002) Genome sequence of Yersinia
pestis KIM. J Bacteriol 184: 46014611.
Doolittle, W.F. (1999) Phylogenetic classification and the universal tree. Science 284: 21242129.
Dozois, C.M., Daigle, F., and Curtiss, R. (2003) Identification
of pathogen-specific and conserved genes expressed in
vivo by an avian pathogenic Escherichia coli strain. Proc
Natl Acad Sci USA 100: 247252.
Dziejman, M., Balon, E., Boyd, D., Fraser, C.M., Heidelberg,
J.F., and Mekalanos, J.J. (2002) Comparative genomic
analysis of Vibrio cholerae: genes that correlate with cholera endemic and pandemic disease. Proc Natl Acad Sci
USA 99: 15561561.
Fitzgerald, J.R., Sturdevant, D.E., Mackie, S.M., Gill, S.R.,
and Musser, J.M. (2001) Evolutionary genomics of Staphylococcus aureus: insights into the origin of methicillinresistant strains and the toxic shock syndrome epidemic.
Proc Natl Acad Sci USA 98: 88218826.
Fournous, G. (2003) Automatisation de la recherche et de
lanalyse de prophages dans les gnomes bactriens. Travail de DEA, Universit Henri Poincar, Nancy, France.
Gamage, S.D., Strasser, J.E., Chalk, C.L., and Weiss, A.A.
(2003) Nonpathogenic Escherichia coli can contribute to
the production of Shiga toxin. Infect Immun 71: 31073115.

18 C. Canchaya, G. Fournous and H. Brssow

Glaser, P., Frangeul, L., Buchrieser, C., Rusniok, C., Amend,
A., Baquero, F., et al. (2001) Comparative genomics of
Listeria species. Science 294: 849852.
Glaser, P., Rusniok, C., Buchrieser, C., Chevalier, F.,
Frangeul, L., Msadek, T., et al. (2002) Genome sequence
of Streptococcus agalactiae, a pathogen causing invasive
neonatal disease. Mol Microbiol 45: 14991513.
Juhala, R.J., Ford, M.E., Duda, R.L., Youlton, A., Hatfull, G.F.,
and Hendrix, R.W. (2000) Genomic sequences of bacteriophages HK97 and HK022: pervasive genetic mosaicism
in the lambdoid bacteriophages. J Mol Biol 299: 2751.
Kazmi, S.U., Kansal, R., Aziz, R.K., Hooshdaran, M., NorrbyTeglund, A., Low, D.E., et al. (2001) Reciprocal, temporal
expression of SpeA and SpeB by invasive M1T1 group a
streptococcal isolates in vivo. Infect Immun 69: 4988
4995.
Kurland, C.G., Canback, B., and Berg, O.G. (2003) Horizontal gene transfer: a critical view. Proc Natl Acad Sci USA
100: 96589662.
Kuroda, M., Ohta, T., Uchiyama, I., Baba, T., Yuzawa, H.,
Kobayashi, I., et al. (2001) Whole genome sequencing of
methicillin-resistant Staphylococcus aureus. Lancet 357:
12251240.
Lawrence, J.G., and Roth, J.R. (1999) Genomic flux:
genomic evolution by gene loss and acquisition. In Organization of the Prokaryotic Genome. Charlebois, R.L. (ed.).
Washington, DC: American Society for Microbiology Press,
pp. 263289.
Lawrence, J.G., Hendrix, R.W., and Casjens, S. (2003)
Where are the pseudogenes in bacterial genomes. Trends
Microbiol 9: 535540.
Lee, C.Y., and Iandolo, J.J. (1986) Lysogenic conversion of
staphylococcal lipase is caused by insertion of the bacteriophage L54a genome into the lipase structural gene. J
Bacteriol 166: 385391.
McDonough, M.A., and Butterton, J.R. (1999) Spontaneous
tandem amplification and deletion of the shiga toxin
operon in Shigella dysenteriae 1. Mol Microbiol 34:
10581069.
Magrini, V., Storms, M.L., and Youderian, P. (1999) Sitespecific recombination of temperate Myxococcus xanthus
phage Mx8: regulation of integrase activity by reversible,
covalent modification. J Bacteriol 181: 40624070.
Makino, K., Yokoyama, K., Kubota, Y., Yutsudo, C.H., Kimura,
S., Kurokawa, K., et al. (1999) Complete nucleotide
sequence of the prophage VT2-Sakai carrying the verotoxin 2 genes of the enterohemorrhagic Escherichia coli
O157:H7 derived from the Sakai outbreak. Genes Genet
Syst 74: 227239.
Nakagawa, I., Kurokawa, K., Yamashita, A., Nakata, M.,
Tomiyasu, Y., Okahashi, N., et al. (2003) Genome
sequence of an M3 strain of Streptococcus pyogenes
reveals a large-scale genomic rearrangement in invasive
strains and new insights into phage evolution. Genome
Res 13: 10421055.
Ohnishi, M., Kurokawa, K., and Hayashi, T. (2001)
Diversification of Escherichia coli genomes: are bacteriophages the major contributors? Trends Microbiol 9: 481
485.
Porwollik, S., Wong, R.M., and McClelland, M. (2002) Evolutionary genomics of Salmonella: gene acquisitions

revealed by microarray analysis. Proc Natl Acad Sci USA

99: 89568961.
Pridmore, D., Berger, B., Desiere, F., Vilanova, D., Barretto,
C., Pittet, A.C., et al. (2004) The genome sequence of the
probiotic intestinal bacterium Lactobacillus johnsonii NCC
533. Proc Natl Acad Sci USA 101: 25122517.
Smoot, L.M., Smoot, J.C., Graham, M.R., Somerville, G.A.,
Sturdevant, D.E., Migliaccio, C.A., et al. (2001) Global differential gene expression in response to growth temperature alteration in group A Streptococcus. Proc Natl Acad
Sci USA 98: 1041610421.
Smoot, J.C., Barbian, K.D., Van Gompel, J.J., Smoot, L.M.,
Chaussee, M.S., Sylva, G.L., et al. (2002a) Genome
sequence and comparative microarray analysis of serotype
M18 group A Streptococcus strains associated with acute
rheumatic fever outbreaks. Proc Natl Acad Sci USA 99:
46684673.
Smoot, L.M., McCormick, J.K., Smoot, J.C., Hoe, N.P.,
Strickland, I., Cole, R.L., et al. (2002b) Characterization of
two novel pyrogenic toxin superantigens made by an acute
rheumatic fever clone of Streptococcus pyogenes associated with multiple disease outbreaks. Infect Immun 70:
70957104.
Sumby, P., and Waldor, M.K. (2003) Transcription of the toxin
genes present within the Staphylococcal phage phiSa3ms
is intimately linked with the phages life cycle. J Bacteriol
185: 68416851.
Tettelin, H., Masignani, V., Cieslewicz, M.J., Eisen, J.A.,
Peterson, S., Wessels, M.R., et al. (2002) Complete
genome sequence and comparative genomic analysis of
an emerging human pathogen, serotype V Streptococcus
agalactiae. Proc Natl Acad Sci USA 99: 1239112396.
Van Sluys, M.A., de Oliveira, M.C., Monteiro-Vitorello, C.B.,
Miyaki, C.Y., Furlan, L.R., Camargo, L.E., et al. (2003)
Comparative analyses of the complete genome sequences
of Pierces disease and citrus variegated chlorosis strains
of Xylella fastidiosa. J Bacteriol 185: 10181026.
Ventura, M., Canchaya, C., Pridmore, D., Berger, B., and
Brssow, H. (2003a) Integration and distribution of Lactobacillus johnsonii prophages. J Bacteriol 185: 46034608.
Ventura, M., Canchaya, C., Kleerebezem, M., de Vos, W.M.,
Siezen, R.J., and Brssow, H. (2003b) The prophage
sequences of Lactobacillus plantarum strain WCFS1. Virology 316: 245255.
Ventura, M., Canchaya, C., Pridmore, R.D., and Brssow, H.
(2004) The Prophages of Lactobacillus johnsonii NCC 533:
comparative genomics and transcription analysis. Virology
320: 229242.
Voyich, J.M., Sturdevant, D.E., Braughton, K.R., Kobayashi,
S.D., Lei, B., Virtaneva, K., et al. (2003) Genome-wide
protective response used by group A Streptococcus to
evade destruction by human polymorphonuclear leukocytes. Proc Natl Acad Sci USA 100: 19962001.
Wagner, P.L., Livny, J., Neely, M.N., Acheson, D.W., Friedman, D.I., and Waldor, M.K. (2002) Bacteriophage control
of Shiga toxin 1 production and release by Escherichia coli.
Mol Microbiol 44: 957970.
Whiteley, M., Bangera, M.G., Bumgarner, R.E., Parsek, M.R.,
Teitzel, G.M., Lory, S., and Greenberg, E.P. (2001) Gene
expression in Pseudomonas aeruginosa biofilms. Nature
413: 860864.
2004 Blackwell Publishing Ltd, Molecular Microbiology, 53, 918