INFECTION AND IMMUNITY, Nov. 2002, p. 6365–6372
Copyright © 2002, American Society for Microbiology. All Rights Reserved.
Vol. 70, No. 11
Genetic Structure and Distribution of Four Pathogenicity Islands
(PAI I536to PAI IV536) of Uropathogenic
Escherichia coli Strain 536
Ulrich Dobrindt,1Gabriele Blum-Oehler,1Gabor Nagy,1,2Gyo ¨rgy Schneider,1Andre ´ Johann,3
Gerhard Gottschalk,3and Jo ¨rg Hacker1*
Institut fu ¨r Molekulare Infektionsbiologie, Universita ¨t Wu ¨rzburg, D-97070 Wu ¨rzburg,1and Institut fu ¨r Mikrobiologie und
Genetik, Labor fu ¨r Genomanalysen, Universita ¨t Go ¨ttingen, 37077 Go ¨ttingen,3Germany, and Department of
Medical Microbiology and Immunology, University Medical School, Pe ´cs, Hungary2
Received 29 May 2002/Returned for modification 8 July 2002/Accepted 29 July 2002
For the uropathogenic Escherichia coli strain 536 (O6:K15:H31), the DNA sequences of three pathogenicity
islands (PAIs) (PAI I536to PAI III536) and their flanking regions (about 270 kb) were determined to further
characterize the virulence potential of this strain. PAI I536to PAI III536exhibit features typical of PAIs, such
as (i) association with tRNA-encoding genes; (ii) G?C content differing from that of the host genome; (iii)
flanking repeat structures; (iv) a mosaic-like structure comprising a multitude of functional, truncated, and
nonfunctional putative open reading frames (ORFs) with known or unknown functions; and (v) the presence
of many fragments of mobile genetic elements. PAI I536to PAI III536range between 68 and 102 kb in size.
Although these islands contain several ORFs and known virulence determinants described for PAIs of other
extraintestinal pathogenic E. coli (ExPEC) isolates, they also consist of as-yet-unidentified ORFs encoding
putative virulence factors. The genetic structure of PAI IV536, which represents the core element of the so-called
high-pathogenicity island encoding a siderophore system initially identified in pathogenic yersiniae, was
further characterized by sample sequencing. For the first time, multiple PAI sequences (PAI I536to PAI IV536)
in uropathogenic E. coli were studied and their presence in several wild-type E. coli isolates was extensively
investigated. The results obtained suggest that these PAIs or at least large fragments thereof are detectable in
other pathogenic E. coli isolates. These results support our view that the acquisition of large DNA regions, such
as PAIs, by horizontal gene transfer is an important factor for the evolution of bacterial pathogens.
Pathogenicity islands (PAIs), as a distinct type of genetic
element, were described for the first time for uropathogenic
Escherichia coli strain 536 (O6:K15:H31) (2, 17), which is one
of the model organisms of extraintestinal pathogenic E. coli
(ExPEC) used for studies on ExPEC pathogenesis and the
evolution of bacterial pathogens. The PAI type of genetic el-
ements is characterized by a large size (?10 kb), the presence
of virulence-associated genes, frequent association with tRNA-
encoding genes or other att sites for temperate bacteriophages,
and a G?C content different from that of the rest of the
chromosome. These elements are frequently flanked by repeat
structures and carry many fragments of other mobile and ac-
cessory genetic elements, such as bacteriophages, plasmids,
and insertion sequence (IS) elements. Some PAIs are unstable
regions and can spontaneously disappear from the chromo-
some. Therefore, PAIs are considered to have evolved from
mobile genetic elements by horizontal gene transfer. It can also
be assumed that these DNA regions, since their acquisition,
underwent and will continue to undergo further evolutionary
changes, resulting in an ongoing evolution of bacterial patho-
In addition to the PAIs of E. coli strain 536, several PAIs
have been characterized for other ExPEC strains, and for many
of them at least partial DNA sequence information is available.
Two PAIs have been identified for each of the uropathogenic
isolates J96 and CFT073 (13, 30, 36) as well as for the sepsis
isolate AL862 (23). One island has been described for the
meningitis K1 isolate EV36 (5). Although some of these is-
lands resemble each other, as they carry identical genes, they
are markedly diverse with respect to size, genetic content and
organization, chromosomal insertion site, and stability.
One aim of this study involved the characterization and
sequence determination of already identified PAIs of uro-
pathogenic E. coli strain 536 to obtain a detailed picture of
these genetic elements. Additionally, to improve knowledge of
the evolution and distribution of PAIs, the presence of PAI
I536- to PAI IV536-specific sequences in different E. coli strains
MATERIALS AND METHODS
Bacterial strains and culture conditions. A collection of 132 E. coli strains
were used in this study and include the Institut fu ¨r Molekulare Infektionsbiologie
collection, which has already been used for the investigation of the flanking
regions of determinants encoding S-family adhesins (8). Additional strains have
been isolated during a long-term study of women with chronic urinary tract
infections: 2A1, 5A1, 5A1U, 2A2, 16A2U, 16A3, 8B1, 16B1, 19B1, 8B2, 3B5,
1E1, 1E5, 8F4, 3N1, 3N2, and 3N5. All strains were grown in Luria-Bertani
DNA techniques. Isolation of plasmid DNA and recombinant DNA was per-
formed as previously described (33).
PCR. A description of the primers used in this study is available as supple-
mentary material (http://www.uni-wuerzburg.de/infektionsbiologie). Screening
for the presence of PAI I536- to PAI IV536-specific sequences was performed by
* Corresponding author. Mailing address: Institut fu ¨r Molekulare
Infektionsbiologie, Ro ¨ntgenring 11, D-97070 Wu ¨rzburg, Germany.
Phone: 49 (0)931 312575. Fax: 49 (0)931 312578. E-mail: j.hacker
PCR with primers specific for the individual region to be amplified. Chromo-
somal DNAs of E. coli strains 536 and MG1655 were used as positive and
negative controls, respectively. DNA primers were purchased from MWG Bio-
tech (Ebersberg, Germany). The Taq DNA polymerase used for the detection of
genes in different E. coli strains was purchased from Qiagen (Hilden, Germany).
Grouping into the main phylogenetic lineages of the ECOR strain collection was
done by a triplex PCR described previously (6). The virulence assessment of the
extraintestinal E. coli strains included a multiplex PCR specific for a set of typical
virulence-associated genes of extraintestinal E. coli (20).
DNA sequencing and sequence analysis. Overlapping cosmid clones covering
the entire regions of PAI I536to PAI III536and their flanking regions of E. coli
strain 536 were sequenced as follows. Small insert libraries (2 to 2.5 kb) were
generated by mechanical shearing of cosmid DNA (29). After end repair with T4
polymerase, the fragments were ligated into the prepared pTZ19R vector. Iso-
lated plasmids were sequenced from both ends by using dye-terminator chem-
istry and analyzed with ABI-377 automated DNA sequencers (Applied Biosys-
tems, Weiterstadt, Germany). After assembly, the remaining gaps were closed by
primer walking on the plasmid clones. The Phrap software implemented in the
STADEN software package was used for assembly and editing of the sequence
data (35). The genetic structure of PAI IV536was confirmed by sample sequenc-
ing of the left and right borders and of several open reading frames (ORFs)
located on PAI IV536. The resulting sequences were compared with those of the
already published high-pathogenicity islands (HPIs) of Yersinia enterocolitica and
Yersinia pseudotuberculosis (3, 4).
Homology searches were performed with the BLASTN, BLASTX, PSI-
BLAST, and PHI-BLAST programs from the National Center for Biotechnology
Information (1). Sequence annotation was performed by using Artemis software
(32). Codon usage tables for PAI I536to PAI IV536were determined with
Artemis and compared with codon usage tables for E. coli K-12, for different
gram-negative and gram-positive bacterial pathogens, and for several bacterio-
phages of E. coli strain K-12, E. coli O157:H7 strain EDL933, Salmonella spp.,
and Shigella spp. (http://www.kazusa.or.jp/codon/cgi-bin).
Southern hybridization. Several PCR results from the screening for PAI I536
to PAI IV536sequences in different E. coli isolates were confirmed by Southern
hybridization of EcoRI-digested chromosomal DNA from the investigated E. coli
isolates. After agarose gel electrophoresis, the EcoRI-digested E. coli genomic
DNA was transferred to Biodyne B nylon membranes (PALL, Ro?dorf, Ger-
many). The probes used for hybridization were obtained by PCR with the primer
pairs used for PCR-based screening and the chromosomal DNA of E. coli strain
MG1655 as a template. Hybridization and detection were carried out by using an
enhanced chemiluminescence labeling and signal detection system (Amersham
Pharmacia Biotech, Freiburg, Germany) according to the manufacturer’s recom-
Nucleotide sequence accession numbers. The complete nucleotide sequences
of PAI I536to PAI III536were submitted to the EMBL database under accession
numbers AJ488511 (PAI I536), AJ494981 (PAI II536), and X16664 (PAI III536).
RESULTS AND DISCUSSION
Genetic features of PAIs of E. coli strain 536. Three PAIs of
E. coli strain 536 (PAI I536to PAI III536) comprising about 270
kb were subcloned and sequenced. The genetic organization of
PAI IV536, which represents the core element of the HPI
initially described for Yersinia, was confirmed by PCR and by
sample sequencing. The general characteristics of PAI I536to
PAI IV536of uropathogenic E. coli strain 536 (e.g., chromo-
somal insertion site, association with tRNA-encoding genes,
size, and encoded virulence factors) are compiled in Fig. 1 and
Table1. The genetic organization of PAI I536to PAI IV536is
depicted in Fig. 2. A detailed list of the known and putative
ORFs identified on these islands is available as supplementary
Generally, PAI I536to PAI III536are mosaic-like structures
consisting of many DNA fragments which show the highest
homology on the nucleotide level to chromosomal regions of
other pathogenic E. coli strains (e.g., O157:H7 strains EDL933
and Sakai and uropathogenic O6 strain CFT073) and of Shi-
gella flexneri (she and SHI-2 PAIs). Many fragments of PAI I536
to PAI III536are also highly homologous to regions of different
virulence plasmids of E. coli (pColV, pB171, pO157, and pA-
PEC-1), Shigella spp. (pWR100 and pWR501), and Yersinia
spp. (pMT1 and pYVe227). Other fragments represent as-yet-
unidentified DNA sequences without homology on the DNA
level. These different regions are interspersed by each other
and consist of already known functional and nonfunctional or
truncated ORFs, of as-yet-unidentified putative ORFs with
homology on the amino acid level to already known proteins,
or of putative ORFs with unknown functions. A marked frac-
tion of ORFs located on these PAIs is derived from mobile
accessory genetic elements, such as bacteriophages, plasmids,
and IS elements. Interestingly, PAI IV536, which represents a
broad-host-range PAI present in many different enterobacte-
ria, comprises only functional ORFs. The ORFs detected on
PAI I536to PAI IV536can be functionally grouped as shown in
In the following text, ORFs are designated in a numerical
way according to their localization and order on a PAI (as
indicated by the index), starting downstream of the tRNA-
encoding gene, which serves as the chromosomal insertion site
for the respective PAI. Consequently, the nonfunctional bac-
teriophage P4-like integrase-encoding ORF on PAI I536which
is located downstream of selC is designated ORF 1I-536.
Characteristics of PAI I536. PAI I536is associated with the
tRNA-encoding gene selC. This island is 76,843 bp in size, is
flanked by 16-bp direct repeats, and has a G?C content of
46%. The nonfunctional bacteriophage integrase gene (ORF
1I-536) immediately downstream of selC exhibits high homology
to other intP4-like genes described for PAIs as well as to that
of phage ?R73 located at this tRNA-encoding gene. In addi-
tion to already known virulence-associated genes located on
PAI I536, such as the alpha-hemolysin-encoding gene cluster,
the annotation of PAI I536sequences revealed two as-yet-
unidentified putative adhesin determinants with no homology
on the nucleotide level. The deduced amino acid sequences
together with their genetic organization indicated that putative
FIG. 1. Comprehensive map of PAI I536to PAI IV536of uropatho-
genic E. coli strain 536. The map is based on the chromosome of E. coli
strain MG1655. PAIs are indicated according to their chromosomal
insertion sites next to tRNA-encoding genes.
6366 DOBRINDT ET AL.INFECT. IMMUN.
ORF 15I-536to ORF 18I-536and ORF 37I-536to ORF 42I-536
represent gene clusters coding for F17- and CS12-like ad-
hesins, respectively. Interestingly, the gene product encoded by
putative ORF 18I-536shows homology to the F17a fimbrial
subunit as well as to the uroepithelial cell adherence protein
UcaA of Proteus mirabilis (EMBL accession no. CAA54703),
indicating that fimbriae containing this subunit may indeed be
involved in urinary tract infections. It will be important to
determine whether these putative determinants are expressed
and how the encoded fimbrial adhesins, whose homologues
have so far been described only for enterotoxigenic E. coli,
contribute to the virulence of extraintestinal E. coli. Another
adhesin-like protein may be encoded by putative ORF 47I-536,
which is preceded by two putative ORFs (ORF 45I-536and
ORF 46I-536) which may encode an ATP-binding cassette
transporter. These ORFs have in common the facts that they
also show no homology on the DNA level but that the deduced
amino acid sequences are homologous to proteins (NMB0586,
NMB0587, and NMB0588) encoded by three adjacent genes in
Neisseria meningitidis strain MC58. Other interesting ORFs
which have not been described so far are putative ORF 2I-536
and ORF 3I-536. These overlapping ORFs are similar in size,
and their deduced amino acid sequences (314 and 310 amino
acids, respectively) are 43 and 39% identical to that of modi-
fication methylase NgoFVII of Neisseria gonorrhoeae, which
consists of 374 amino acids. However, both ORFs show only
78% identity on the DNA level over the first 89 nucleotides,
thus excluding the possibility of gene duplication. As the meth-
ylation status of chromosomal DNA is also regulated in re-
sponse to different stimuli and affects gene expression, it will be
interesting to investigate whether these modification methy-
lase-encoding ORFs (in case they are expressed) also influence
the gene expression of E. coli strain 536. The importance of the
dam-encoded methylase for general gene expression and ex-
pression of virulence-associated genes in Salmonella and E. coli
was previously reported (18, 19).
Characteristics of PAI II536. PAI II536is associated with the
tRNA-encoding gene leuX. This island is 102,200 bp in size, is
flanked by 18-bp direct repeats, and has a G?C content of
46%. The functional bacteriophage integrase gene (ORF 1II-
536) immediately downstream of leuX exhibits the highest ho-
mology to the intB gene of bacteriophage P4 located at this
tRNA-encoding gene in E. coli K-12. Already known virulence
determinants located on PAI II536are the prf determinant
(ORF 6II-536to ORF 17II-536), which codes for the P-related
fimbrial adhesin, and another alpha-hemolysin-encoding gene
cluster (ORF 22II-536to ORF 25II-536). Other putative viru-
lence-associated genes present on this PAI are ORF 4II-536,
which codes for the Hek adhesin described for E. coli (EMBL
accession no. AY040859), and two putative ORFs (ORF
40II-536and ORF 41II-536) without homology on the DNA
level. The encoded gene product of ORF 40II-536shows ho-
mology to filamentous hemagglutinin-like adhesins of Borde-
tella pertussis, Pseudomonas aeruginosa, and Yersinia pestis, and
ORF 41II-536shows homology to a conserved ORF which is lo-
cated upstream of the hemagglutinin-encoding one and which
is required for secretion of the adhesins. Another interesting
ed gene product shows homology to a fragment of modification
methylase HgiDII of Herpetosiphon aurantiacus (EMBL acces-
sion no. P25265). The right-hand direct repeat structure is not
immediately followed by E. coli K-12-specific sequences rep-
resenting the conserved E. coli chromosomal backbone but by
another 4-kb DNA region which is not present in E. coli K-12
strain MG1655. In this region, only one putative ORF coding
for a hypothetical protein of E. coli O157:H7 strain Z5892 is
located. This 4-kb DNA stretch is then followed by the E. coli
K-12 chromosomal backbone starting with yjhS.
Characteristics of PAI III536. PAI III536is associated with
the tRNA-encoding gene thrW. This island is 68,124 bp in size
and has a G?C content of 47%. The functional bacteriophage
integrase gene (ORF 1III-536) immediately downstream of thrW
exhibits the highest homology to the int gene of bacteriophage
SfX, which recognizes this tRNA-encoding gene as a chromo-
somal insertion site. This PAI is flanked by 47-bp direct re-
peats, as a truncated thrW gene located at the right-hand end
of PAI III536can, together with the functional copy of thrW at
the left-hand end, serve as a direct repeat structure. However,
approximately 7 kb of DNA is located between the truncated
thrW gene at the right-hand end of PAI III536and the transi-
tion into the E. coli K-12-like chromosomal backbone. This
DNA stretch also represents foreign DNA, absent in E. coli
strain MG1655, which contains an ORF with homology to the
hemoglobin protease-encoding genes of pColV-K30 and
pAPEC-1 (EMBL accession no. AJ223631 and AF218073, re-
spectively). The Tsh hemoglobin protease has been shown to
be involved in the virulence of, e.g., avian-pathogenic E. coli
(10). As previously described (8), the right junction site of this
composite structure includes sequences with homology to frag-
ments of integrase genes of different S. flexneri and E. coli
bacteriophages, followed by DNA sequences with homology to
fragments of the insB and insA genes of the Iso IS1 element.
These results demonstrate that the chromosomal region be-
tween thrW and yagU represents a composite structure of PAI
TABLE 1. Main features of PAI I536to PAI IV536of E. coli strain 536
Known or putative virulence factor(s)
Alpha-hemolysin, F17-like fimbriae,aCS12-like fimbriaea
Hek adhesin, P-related fimbriae, alpha-hemolysin, hemagglutinin-like adhesina
S fimbriae, iro siderophore system, HmuR-like heme receptor,aSap adhesin, Tsh-like
Yersiniabactin siderophore system
aPutative ORF or operon shows no homology on the DNA level but does with respect to the deduced amino acid sequence and genetic structure.
bIntegrase-encoding gene is nonfunctional due to internal stop codons or a mutated start codon.
cTsh-like hemoglobin protease-encoding gene is located in the 7-kb region outside of PAI III536, which is absent from the E. coli K-12 strain MG1655 genome.
VOL. 70, 2002PAIs OF UROPATHOGENIC E. COLI STRAIN 5366367
FIG. 2. Comparison of the genetic organization of PAI I536to PAI IV536. Known or putative ORFs are grouped according to the following characteristics: blue, functional, known ORFs;
green, truncated ORFs with a start codon and a stop codon; grey, as-yet-unidentified ORFs without homology on the DNA level. Nonfunctional ORFs (e.g., due to internal stop codons
or frameshifts) are indicated by hatched symbols. ORF numbers are indicated below the corresponding ORF symbols. Functional or truncated tRNA-encoding genes are marked in red.
Direct repeat (DR) structures flanking PAIs are indicated. Thick black lines below the PAIs represent regions of PAI I536PAI IV536which were detected by PCR. Several PAI-specific
PCRs were grouped into PAI regions.
6368DOBRINDT ET AL.INFECT. IMMUN.
III536and other horizontally acquired sequences that could be
considered to be flanked by sequences derived from Shigella
bacteriophages which integrate into highly homologous attP
sites next to the tRNA-encoding gene thrW. The occurrence of
truncated tRNA-encoding genes within PAIs or DNA regions
presumably acquired by horizontal gene transfer has also been
observed for PAI II of CFT073 (30) and indicates the impor-
tance of these hot spots for chromosomal integration of foreign
In addition to virulence determinants carried on PAI III536,
such as the S fimbrial adhesin-encoding gene cluster sfa (ORF
17III-536to ORF 25III-536) and the siderophore system-encod-
ing gene cluster iro (ORF 27III-536toORF 31III-536), sequence
analysis of the entire region of PAI III536revealed the pres-
ence of other known genes coding for virulence factors of E.
coli; an example is ORF 52III-536, with homology to sap, which
codes for an autotransporter-adhesin and which is located on
the she PAI of S. flexneri 2a (EMBL accession no. AF200692).
The as-yet-unidentified putative ORF 36III-536 encodes a
HmuR-like heme receptor which has been described for Y.
pestis (EMBL accession no. Q56989). Only short fragments of
putative ORF 47III-536and ORF 48III-536, coding for lysine de-
carboxylase (CadA) and cadaverine/lysine antiporter (CadB)
homologues, respectively, show homology on the DNA level to
the corresponding cadA and cadB genes of E. coli or Salmo-
nella enterica serovar Typhimurium, indicating either their ac-
quisition by horizontal gene transfer or the occurrence of mul-
tiple DNA rearrangement processes resulting in these cadA
and cadB variants. Interestingly, ORFs with high homology to
these genes are also located on PAI II of CFT073 (30).
Whereas CadA and CadB have been proposed to be involved
in acid tolerance mechanisms of Salmonella (12), the encoding
genes are part of the so-called “black hole” in Shigella and
enteroinvasive E. coli, which describes a large chromosomal
deletion in these bacteria (25). The loss of this chromosomal
region promotes the virulence of Shigella and enteroinvasive
E. coli, as the product of lysine decarboxylase activity, cadav-
erine, blocks the action of Shigella enterotoxins (26).
Characteristics of PAI IV536. PAI IV536is associated with
the tRNA-encoding gene asnT. This island represents the core
element of the so-called HPI of pathogenic Yersinia species. It
has been completely sequenced for several Yersinia strains (3,
4, 11). The left and right junction sites of PAI IV536have also
been sequenced before (34). We therefore verified the genetic
organization of this PAI in E. coli strain 536 by sample se-
quencing of PCR products obtained with primers described
elsewhere (22). The obtained sequences showed between 97
and 100% identity on the nucleotide level to other HPI-specific
sequences of Y. pseudotuberculosis. Sequence determination of
the left and right junction sites of PAI IV536revealed only
minor differences in the right-hand end of PAI IV536compared
to previous results. According to our sequence analysis, the 5?
end of PAI IV536shows 98% identity to the sequence pub-
lished earlier but comprises a 1,002-bp intergenic region be-
tween fyuA and ?b1978 instead of 1,226 bp, as described by
Schubert and colleagues (34). According to these results, we
conclude that by analogy to the HPI of Y. pseudotuberculosis,
PAI IV536is about 30.2 kb in size and has a G?C content of
57%. Flanking repeat structures are absent in PAI IV536, which
contains the gene cluster required for biosynthesis of the sid-
erophore system yersiniabactin.
Genes of mobile genetic elements present on PAI I536to PAI
IV536. Besides the characteristic bacteriophage integrase-en-
coding genes located immediately downstream of the tRNA-
encoding gene serving as a chromosomal insertion site for
these PAIs, a considerable fraction of ORFs on PAI I536to
PAI III536are remnants of various IS elements and trans-
posons of different origins. The majority represent functional
or nonfunctional transposase-encoding genes. Intact IS ele-
ments or transposons, however, have not been identified on
these PAIs. Twenty-six fragments of IS elements have been
detected on PAI I536to PAI III536. They belong to different
types of IS elements, including IS1, IS2, IS3, IS4, IS10, IS50R,
IS100, IS110, IS629, IS630, IS679, IS911, IS1328, and IS1353.
Many of the corresponding families of IS elements not only are
restricted to enterobacteria but also are present among other
bacteria (24). These IS elements are frequently located on
virulence plasmids. Interestingly, at least one copy of each of
the ORFs encoding the two subunits of the IS100 transposase
is located on PAI I536to PAI III536. Furthermore, several
ORFs (copies of L0007 to L0010) of prophage CP-933 of E.
coli O157:H7 strain EDL933 are always detectable on PAI I536
to PAI III536. Genes of prophage CP-933 are also present on
PAIs of ExPEC strains CFT073 and AL862 (13, 23, 30). In
addition, the putative ORF 63I-536exhibits the highest homol-
ogy to a fragment of an intron-associated reverse transcriptase/
maturase-encoding ORF. The gene products of the adjacent
putative ORF 52II-536and ORF 53II-536, which share no se-
quence homology, are homologous to the quaternary ammo-
nium compound resistance protein QacE of Brucella melitensis
(EMBL accession no. AE009544) and E. coli (EMBL accession
no. AF205943), respectively. In E. coli, qacE is known to be
part of intron In53, a class 1 plasmid- and composite transpo-
son-located integron (27). These data indicate that these PAIs
in their present state result from repeated recombination
events, including integration of different mobile genetic and
other accessory genetic elements. PAI IV536contains no rem-
nants of IS elements.
Codon usage of ORFs located on PAI I536to PAI IV536. A
comparison of codon usage tables for every PAI demonstrated
that several codons occur with markedly different frequencies
in these islands and in the chromosomal E. coli K-12-specific
TABLE 2. Functional categories of known and putative
ORFs located on PAI I536to PAI IV536
Known or putative function
No. of ORFs that are:
Related to mobile genetic elements
Other virulence-associated function
Hypothetical or unknownb
aDue to internal stop codons or frameshifts.
bPutative ORFs or their encoded products show homology on the DNA or
VOL. 70, 2002PAIs OF UROPATHOGENIC E. COLI STRAIN 5366369
backbone. Details are available as supplementary material
(http://www.uni-wuerzburg.de/infektionsbiologie). The codon
usage of PAI I536to PAI III536differs from that of PAI IV536.
The usage of codons such as ATA, ACA, AGA, and AGG is
increased by at least a factor of 2 in these PAIs, whereas that
of codon GCG is more than twofold lower than that in E. coli
strain MG1655. This tendency is visible in several bacterial
species, such as S. flexneri, Shigella sonnei, Salmonella enterica,
Y. pseudotuberculosis, Y. enterocolitica, Staphylococcus aureus,
and Bacillus subtilis, as well as in many bacteriophages of
enterobacteria, such as 933W, ES18, P4, SfV, SfVI, and 7887.
In PAI IV536, codons AGG and CCC are used at least twice as
much as in E. coli strain MG1655, whereas the usage of codons
ACT, AAA, and GGT is more than twofold lower in PAI
IV536. These differences are indicative of the (more or less
recent) acquisition of these PAIs by horizontal gene transfer.
Analysis of the presence of PAI I536- to PAI IV536-specific
sequences in different E. coli isolates. Based on the DNA
sequences of PAI I536to PAI IV536, we selected DNA primer
combinations in order to detect specific sequences of these
islands in a set of 132 different E. coli strains including many
ExPEC and intestinal pathogenic E. coli isolates as well as
several nonpathogenic E. coli strains (Fig. 2). Detailed results
of this extensive screening for multiple PAIs, which complete
recently published results on the presence of PAI III536homo-
logues in E. coli isolates (8), are available as supplementary
material (http://www.uni-wuerzburg.de/infektionsbiologie). As
shown in Table 3, PAI IV536-specific sequences showed a
higher percentage of occurrence among the different E. coli
strains tested in this study than sequences specific for PAI I536
to PAI III536. PAI III536-specific sequences were more fre-
quently detectable than sequences specific for PAI II536. PAI
I536-specific DNA regions were detected at the lowest fre-
quency. PAI I536- to PAI III536-specific sequences were usually
not detectable in intestinal pathogenic E. coli isolates, indicat-
ing that PAI I536to PAI III536represent a part of the ExPEC-
specific gene pool. However, with a few exceptions due to local
sequence differences, 17 to 28% of the intestinal pathogenic
isolates were positive for all PAI IV536-specific PCRs. Between
67 and 92% of the ExPEC isolates were positive for all PAI
IV536-specific PCRs, suggesting that they harbor the entire
HPI core element. In contrast, the other three PAIs were not
always completely detectable in some of the strains. In these
instances, only certain PAI I536- to PAI III536-specific PCRs
gave positive PCR results for the different strains used in this
As already speculated, strains harboring determinants cod-
ing for members of the S family of fimbrial adhesins were
shown to possess a common module of different forms of a
PAI, depending on the different S adhesin types (8). The PCR
screening based on the sequence of the complete region of PAI
III536confirmed this finding, as only some of the urinary tract
infection isolates were positive for all PAI III536-specific PCRs.
However, the chromosomal insertion site for their PAI III536
equivalent is so far unknown.
For some strains tested, all PAI I536-, PAI II536-, or PAI
III536-specific sequences (with the above-mentioned exception
for PAI III536) were detectable. Positive results in at least 13
specific PCRs for PAI I536to PAI III536strongly suggested that
these strains contain the complete islands, as described for E.
coli strain 536. Strains in which the complete PAIs of uro-
pathogenic E. coli strain 536 have been detected are compiled
in Table4. As shown by a previously described triplex PCR (6),
all strains belong to major phylogenetic group B2, and the
majority carry the same repertoire of virulence-associated
genes as E. coli strain 536. The approximate genome sizes of
these strains were determined by pulsed-field gel electrophore-
sis after CeuI digestion. The restriction patterns obtained in-
dicated that, with the exception of those of strains RZ454 and
RZ532 (4.6 Mb), the genome sizes are similar, ranging be-
tween 4.9 and 5.1 Mb. These results indicate that one or even
multiple PAIs which have been described for uropathogenic E.
coli strain 536 are also detectable in other ExPEC isolates. For
the first time, we provide evidence by an extensive analysis of
specific sequences of multiple PAIs in 132 different E. coli
strains that identical PAIs, other than the HPI core element,
most likely are present in different E. coli strains. This evidence
indicates that these large chromosomal regions have been
transferred by horizontal gene transfer.
PAIs and their importance for the evolution of ExPEC. The
results of this study increase knowledge on the variability of the
genetic structures of PAIs in ExPEC. PAI IV536represents a
broad-host-range PAI also present in many other E. coli patho-
types and different enterobacteria and differs with respect to
G?C content and genetic organization from PAI I536to PAI
III536, which have similar G?C contents and structural fea-
tures and which have so far been detected only in ExPEC.
However, the latter islands share several ORFs which are also
present on many PAIs of other ExPEC isolates, such as bac-
teriophage CP-933-related sequences. This fact may be indic-
ative of continously ongoing DNA rearrangements among PAI
sequences resulting in evolution in ExPEC strains of different
PAI variants deriving from an ancestral PAI which also con-
sisted of phage CP-933. Although many PAIs of ExPEC strains
TABLE 3. Detection of PAI I536- to PAI IV536-specific sequences in different E. coli isolates
E. coli strains (n)a
% Positive specific PCRs for the indicated region of PAI:
Human and animal MNEC
and SEPEC (28)
3.2 20.9 3.2 4.8 3.2 43.5 24.2 24.2 22.6 66.1 22.6 12.9 1.6 9.7 59.7 64.5 11.3 40.3 11.3 72.6 82.3 85.5 85.5 85.5
014.3 0 0032257.1 21.4 35.7 2510.7 07.17.1 46.47.1 28.67.1 64.3 85.7 89.3 89.3 89.3
7.1 14.3 3.6 7.1 7.1 21.4 17.9
78.6 78.6 78.6
27.8 27.8 16.7 27.8011000
aUTI, urinary tract infection; MNEC, meningitis-causing E. coli; SEPEC, sepsis-causing E. coli; ECOR, E. coli reference collection.
6370 DOBRINDT ET AL.INFECT. IMMUN.
superficially resemble each other with respect to the presence
and/or genetic linkage of certain virulence determinants, there
is nevertheless considerable variability in PAI composition and
structural organization (8, 13, 21, 23, 30, 31, 36). The majority
of homologous DNA sequences present on PAIs, such as frag-
ments of IS elements and other accessory genetic elements,
allow recombination and consequently PAI rearrangements
and deletion and/or acquisition of foreign DNA that has been
acquired by horizontal gene transfer. This feature facilitates
rapid and continuous DNA rearrangements and remodeling of
PAIs. In addition, some PAIs of ExPEC represent unstable
genetic elements which can be deleted from the chromosome
(2, 9, 28). It is assumed that such deleted PAIs, or at least large
parts thereof, could be transferred to suitable recipients and
thus contribute to the ongoing evolution of pathogenic bacteria
(7, 15, 28).
In this study, we provide evidence that large regions of PAI
I536to PAI III536can be detected in different ExPEC isolates,
thus arguing for en bloc acquisition of these DNA regions by
horizontal gene transfer. In summary, the study of the genetic
organization of PAIs of uropathogenic E. coli strain 536 re-
vealed the existence of several putative virulence determinants
so far unknown in E. coli. A comparison of PAI I536to PAI
III536with other PAIs of ExPEC supports a model in which
PAI mobilization and horizontal gene transfer as well as con-
tinous reorganization of PAIs by recombination, for example,
mediated by multiple fragments of accessory DNA elements
present on one or multiple PAIs, contribute to the ongoing
evolution of ExPEC variants.
The work of the group in Wu ¨rzburg, Germany, was supported by the
SFB479, the Fonds der Chemischen Industrie, the Bayerische Fors-
chungsstiftung, and the FAZIT-Stiftung. The Labor fu ¨r Genomana-
lysen received support from the Forschungsmittel des Landes Nied-
We thank G. Balling and B. Middendorf (Wu ¨rzburg, Germany) for
sequence determination and analysis of the left and right borders of
PAI IV536as well as M. Schultheiss and B. Plaschke (Wu ¨rzburg, Ger-
many) for excellent technical assistance.
1. Altschul, S. F., T. L. Madden, A. A. Schaffer, J. Zhang, Z. Zhang, W. Miller,
and D. J. Lipman. 1997. Gapped BLAST and PSI-BLAST: a new generation
of protein database search programs. Nucleic Acids Res. 25:3389–3402.
2. Blum, G., M. Ott, A. Lischewski, A. Ritter, H. Imrich, H. Tscha ¨pe, and J.
Hacker. 1994. Excision of large DNA regions termed pathogenicity islands
from tRNA-specific loci in the chromosome of an Escherichia coli wild-type
pathogen. Infect. Immun. 62:606–614.
3. Buchrieser, C., C. Rusniok, L. Frangeul, E. Couve, A. Billault, F. Kunst, E.
Carniel, and P. Glaser. 1999. The 102-kilobase pgm locus of Yersinia pestis:
sequence analysis and comparison of selected regions among different Yer-
sinia pestis and Yersinia pseudotuberculosis strains. Infect. Immun. 67:4851–
4. Carniel, E., I. Guilvout, and M. Prentice. 1996. Characterization of a large
chromosomal “high-pathogenicity island” in biotype 1B Yersinia enteroco-
litica. J. Bacteriol. 178:6743–6751.
5. Cieslewicz, M., and E. Vimr. 1997. Reduced polysialic acid capsule expres-
sion in Escherichia coli K1 mutants with chromosomal defects in kpsF. Mol.
6. Clermont, O., S. Bonacorsi, and E. Bingen. 2000. Rapid and simple deter-
mination of the Escherichia coli phylogenetic group. Appl. Environ. Micro-
7. Dobrindt, U., and J. Hacker. 2001. Whole genome plasticity in pathogenic
bacteria. Curr. Opin. Microbiol. 4:550–557.
8. Dobrindt, U., G. Blum-Oehler, T. Hartsch, G. Gottschalk, E. Z. Ron, R.
Fu ¨nfstu ¨ck, and J. Hacker. 2001. S-fimbria-encoding determinant sfaI is lo-
cated on pathogenicity island III536of uropathogenic Escherichia coli strain
536. Infect. Immun. 69:4248–4256.
9. Dobrindt, U., U. Hentschel, J. B. Kaper, and J. Hacker. 2002. Genome
plasticity in pathogenic and nonpathogenic enterobacteria. Curr. Top. Mi-
crobiol. Immunol. 264:157–175.
10. Dozois, C. M., M. Dho-Moulin, A. Bree, J. M. Fairbrother, C. Desautels, and
R. Curtiss III. 2000. Relationship between the Tsh autotransporter and
pathogenicity of avian Escherichia coli and localization and analysis of the
Tsh genetic region. Infect. Immun. 68:4145–4154.
11. Fetherston, J. D., P. Schuetze, and R. D. Perry. 1992. Loss of the pigmen-
tation phenotype in Yersinia pestis is due to the spontaneous deletions of 102
kb of chromosomal DNA, which is flanked by a repetitive element. Mol.
12. Foster, J. W., Y. K. Park, I. S. Bang, K. Karem, H. Betts, H. K. Hall, and E.
Shaw. 1994. Regulatory circuits involved with pH-regulated gene expression
in Salmonella typhimurium. Microbiology 140:341–352.
13. Guyer, D. M., J.-S. Kao, and H. L. T. Mobley. 1998. Genomic analysis of a
pathogenicity island in uropathogenic Escherichia coli strain CFT073: distri-
bution of homologous sequences among isolates from patients with pyelo-
nephritis, cystitis, and catheter-associated bacteriuria and from fecal sam-
ples. Infect. Immun. 66:4411–4417.
14. Hacker, J., and J. B. Kaper. 1999. The concept of pathogenicity islands, p.
1–11. In J. B. Kaper and J. Hacker (ed.), Pathogenicity islands and other
mobile virulence elements. ASM Press, Washington, D.C.
15. Hacker, J., and J. B. Kaper. 2000. Pathogenicity islands and the evolution of
microbes. Annu. Rev. Microbiol. 54:641–679.
16. Hacker, J., G. Blum-Oehler, B. Janke, G. Nagy, and W. Goebel. 1999.
Pathogenicity islands of extraintestinal Escherichia coli, p. 59–76. In J. B.
TABLE 4. Characterization of different ExPEC isolates presumably containing complete PAIs of E. coli strain 536 or large fragments thereof
I536 II536 III536
ECOR 52 ND
ECOR 60 ND
malX papAH papC papEF papGIII fimH fyuA sfalfocDE sfaS hlyA kpsMTK5
malX papAH papC papEF papGII/III fimH fyuA sfalfoc focG cnf1 iutA hlyA kpsMTK5
malX papAH papC papEF papGII/III fimH fyuA sfalfoc focG cnfl iutA hlyA rfc kpsMTK5
malX papAH papC papEF papGII/III fimH fyuA sfalfoc sfaS cnf1 hlyA kpsMTK5
malX papAH papC papEF papGII/III fimH fyuA sfalfoc sfaS hlyA kpsMTK5
malX papAH papC papEF papGII/III fimH fyuA sfalfoc sfaS cnf1 hlyA kpsMTK5
malX papAH papC papEF papGII/III fimH fyuA sfalfoc sfaS cnf1 iutA hlyA kpsMTK5
malX papAH papC papEF papGII/III fimH fyuA sfalfoc focG cnf1 cdtB traT hlyA kpsMTK5
malX papAH papC papEF papGII/III fimH fyuA sfalfoc sfaS cnf1 traT hlyA kpsMTK5
malX papAH papC papEF papGII/III fimH fyuA sfalfoc sfaS cnf1 hlyA kpsMTK5
malX papAH papC papEF papGII/III fimH fyuA sfalfoc sfaS cnf1 hlyA kpsMTK5
aND, not determined.
bUTI, urinary tract infection.
cAs determined by pulsed-field gel electrophoresis after CeuI digestion.
dAs determined by multiplex PCR; designation of PCR products is according to McCormick et al. (26).
eChromosomal insertion site of PAI III536homologues is unknown. ?, present; ?, not present.
VOL. 70, 2002 PAIs OF UROPATHOGENIC E. COLI STRAIN 536 6371
Kaper and J. Hacker (ed.), Pathogenicity islands, plasmids, and other mobile
elements. ASM Press, Washington, D.C.
17. Hacker, J., S. Knapp, and W. Goebel. 1983. Spontaneous deletions and
flanking regions of the chromosomally inherited hemolysin determinant of
an Escherichia coli O6 strain. J. Bacteriol. 154:1145–1152.
18. Hale, W. B., M. W. van der Woude, B. A. Braaten, and D. A. Low. 1998.
Regulation of uropathogenic Escherichia coli adhesin expression by DNA
methylation. Mol. Genet. Metab. 65:191–196.
19. Heithoff, D. M., R. L. Sinsheimer, D. A. Low, and M. J. Mahan. 1999. An
essential role for DNA adenine methylation in bacterial virulence. Science
20. Johnson, J. R., and A. L. Stell. 2000. Extended virulence genotypes of
Escherichia coli strains from patients with urosepsis in relation to phylogeny
and host compromise. J. Infect. Dis. 181:261–272.
21. Johnson, J. R., T. T. O’Bryan, M. Kuskowski, and J. N. Maslow. 2001.
Ongoing horizontal and vertical transmission of virulence genes and papA
alleles among Escherichia coli blood isolates from patients with diverse-
source bacteremia. Infect. Immun. 69:5363–5374.
22. Karch, H., S. Schubert, D. Zhang, W. Zhang, H. Schmidt, T. O ¨lschla ¨ger, and
J. Hacker. 1999. A genomic island, termed high-pathogenicity island, is
present in certain non-O157 Shiga toxin-producing Escherichia coli clonal
lineages. Infect. Immun. 67:5994–6001.
23. Lalioui, L., and C. Le Bougue ´nec. 2001. afa-8 gene cluster is carried by a
pathogenicity island inserted into the tRNAPheof human and bovine patho-
genic Escherichia coli isolates. Infect. Immun. 69:937–948.
24. Mahillon, J., and M. Chandler. 1998. Insertion sequences. Microbiol. Mol.
Biol. Rev. 62:725–774.
25. Maurelli, A. T., R. E. Fernandez, C. A. Bloch, C. K. Rode, and A. Fasano.
1998. Black holes and bacterial pathogenicity: a large genomic deletion that
enhances the virulence of Shigella spp. and enteroinvasive Escherichia coli.
Proc. Natl. Acad. Sci. USA 95:3943–3949.
26. McCormick, B. A., M. I. Fernandez, A. M. Siber, and A. T. Maurelli. 1999.
Inhibition of Shigella flexneri-induced transepithelial migration of polymor-
phonuclear leucocytes by cadaverine. Cell. Microbiol. 1:143–155.
27. Naas, T., Y. Mikami, T. Imai, L. Poirel, and P. Nordmann. 2001. Charac-
terization of In53, a class 1 plasmid- and composite transposon-located
integron of Escherichia coli which carries an unusual array of gene cassettes.
J. Bacteriol. 183:235–249.
28. Ochman, H., J. G. Lawrence, and E. A. Groisman. 2000. Lateral gene
transfer and the nature of bacterial innovation. Nature 405:299–304.
29. Oefner, P., S. P. Hunicke-Smith, L. Chiang, F. Dietrich, J. Mulligan, and
R. W. Davis. 1996. Efficient random subcloning of DNA sheared in a recir-
culating point-sink flow system. Nucleic Acids Res. 24:3879–3886.
30. Rasko, D. A., J. A. Phillips, X. Li, and H. L. T. Mobley. 2001. Identification
of DNA sequences from a second pathogenicity island of uropathogenic
Escherichia coli CFT073: probes specific for uropathogenic populations. 184:
31. Redford, P., and R. A. Welch. 2002. Extraintestinal Escherichia coli as a
model system for the study of pathogenicity islands. Curr. Top. Microbiol.
32. Rutherford, K., J. Parkhill, J. Crook, T. Horsnell, P. Rice, M.-A. Rajan-
dream, and B. Barrell. 2000. Artemis: sequence visualisation and annotation.
33. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular cloning: a
laboratory manual, 2nd ed. Cold Spring Harbor Laboratory Press, Cold
Spring Harbor, N.Y.
34. Schubert, S., A. Rakin, D. Fischer, J. Sorsa, and J. Heesemann. 1999.
Characterization of the integration site of Yersinia high-pathogenicity island
in Escherichia coli. FEMS Microbiol. Lett. 179:409–414.
35. Staden, R., K. F. Beal, and J. K. Bonfield. 2000. The Staden package 1998.
Methods Mol. Biol. 132:115–130.
36. Swenson, D. L., N. O. Bukanov, D. E. Berg, and R. A. Welch. 1996. Two
pathogenicity islands in uropathogenic Escherichia coli J96: cosmid cloning
and sample sequencing. Infect. Immun. 64:3736–3743.
Editor: J. T. Barbieri
6372 DOBRINDT ET AL.INFECT. IMMUN.