Vol 13, Issue 3, 428-442, March 2003
LETTER
Comparative Analysis of Superintegrons: Engineering Extensive Genetic Diversity in the Vibrionaceae
Dean A. Rowe-Magnus1,
Anne-Marie Guerout,
Latefa Biskri,
Philippe Bouige and
Didier Mazel2
Unité de Programmation Moléculaire et Toxicologie
Génétique–CNRS URA 1444, Département de
Microbiologie Fondamentale et Médicale, Institut Pasteur, 75724,
Paris, France
 |
ABSTRACT
|
|---|
Integrons are natural tools for bacterial evolution and innovation.
Their involvement in the capture and dissemination of
antibiotic-resistance genes among Gram-negative bacteria is well
documented. Recently, massive ancestral versions, the superintegrons
(SIs), were discovered in the genomes of diverse proteobacterial
species. SI gene cassettes with an identifiable activity encode
proteins related to simple adaptive functions, including resistance,
virulence, and metabolic activities, and their recruitment was
interpreted as providing the host with an adaptive advantage. Here, we
present extensive comparative analysis of SIs identified among the
Vibrionaceae. Each was at least 100 kb in size, reaffirming the
participation of SIs in the genome plasticity and heterogeneity of
these species. Phylogenetic and localization data supported the
sedentary nature of the functional integron platform and its
coevolution with the host genome. Conversely, comparative analysis of
the SI cassettes was indicative of both a wide range of origin for the
entrapped genes and of an active cassette assembly process in these
bacterial species. The signature attC sites of each species
displayed conserved structural characteristics indicating that symmetry
rather than sequence was important in the recognition of such a varied
collection of target recombination sequences by a single site-specific
recombinase. Our discovery of various addiction module cassettes within
each of the different SIs indicates a possible role for them in the
overall stability of large integron cassette
arrays.
[Supplemental material is available online at
www.genome.org. The sequence data from this study have been submitted
to GenBank under accession nos. listed in Table 1.]
Natural selection favors the evolution of strategies that increase
the rate of adaptation, that is, chance favors the
prepared genome (Caporale 1999 ). Although mutation generally causes
only a very small and localized change in a cell, the transfer of
genetic material involves much broader changes that may permit the
organism to carry out new functions and adapt to environmental changes
(Ochman et al. 2000). Integrons are exquisitely suited for this
purpose. Integrons are natural cloning and expression systems that
incorporate open reading frames (ORFs) and convert them to functional
genes (Rowe-Magnus and Mazel 1999, 2001). They have been expansively
identified as the constituents of transferable elements responsible for
the evolution of multidrug resistance among human, animal, and plant
pathogenic isolates during the antibiotic era. More than 70 different
antibiotic-resistance genes, covering most antimicrobials used against
Gram-negative infections, have been characterized within integrons thus
far (Rowe-Magnus et al. 2002a). The substantial impact of integrons on
bacterial evolution is underscored by the present dilemma in the
treatment of infectious disease, as the development of
multiple-antibiotic resistance can often be traced to the stockpiling
of resistance loci within integrons to create multiresistance integrons
(MRIs). MRIs harboring up to eight resistance cassettes have been
isolated from multiresistant clinical isolates (Naas et al. 2001).
The integron platform codes for an integrase (intI) that
mediates recombination between a proximal primary recombination site
(attI) and a target recombination sequence called an
attC site (or 59 base elements; 59 be). The attC site
is usually found associated with a single open reading frame in a
circularized structure termed a gene cassette (Hall and Stokes 1993;
Recchia and Hall 1995; Stokes et al. 1997; Sundstrom 1998). Insertion
of the gene cassette at the attI site, which is located
downstream of a resident promoter, Pc, internal to the
intI gene, drives expression of the encoded proteins (Levesque
et al. 1994).
Most of the attC sites of integron gene cassettes identified
to date share little homology. Their length and sequence vary
considerably (from 57–141 bp) and their sequence similarities are
primarily restricted to their boundaries, which correspond to the
inverse core site (CS; RYYYAAC) and the core site (CS; G TTRRRY,
where R is a purine, Y is a pyrimidine, and is a recombination
point; Stokes et al. 1997; Collis et al. 1998). Despite their limited
homology, attC sites do share certain structural
characteristics (see Fig. 1). Each forms an
imperfect inverted repeat, with strong complementarity being
particularly apparent over the terminal 25 bp. The CS and ICS are
perfectly complementary in the attC site of a circularized
cassette. This stretch of complementarity can extend as far as 5 nt
downstream of the CS and upstream of the ICS, giving a total of up to
12 consecutive complementary nucleotides in these regions. Two
integrase-binding sites, called LH and RH simple sites, are found
within each attC site. The LH simple site is composed of the
ICS and a second sequence that is related to the CS consensus, called
2L (Stokes et al. 1997; Collis et al. 1998). The ICS and 2L sequences
are separated by a 5-bp spacer, and it has been proposed that the
integrase binds to the LH simple site either as two monomers or a dimer
(Collis et al. 1998). The RH simple site is similarly defined, and it
includes the CS and a second sequence that is related to the ICS
consensus, called 2R. As seen for the LH simple site, the CS and 2R
sequences have a 5–6-bp spacer between them. Because it has been
demonstrated that markedly different attC sites are recognized
as substrates by the integron-integrase protein (Collis et al. 2001),
this indicated that the sequence of the attC site, outside of
the ICS and CS, is not as important as its symmetry.
Five classes of MRIs have been reported based on the divergence of
their integrase genes, which share between 39% and 58% identity. The
evolutionary divergence among the integrase genes indicated that the
activity of this system has extended far beyond the 60 yr of the
antibiotic era. This postulate was confirmed by the recent discovery of
massive ancestral versions, the superintegrons (SIs), in the genomes of
diverse proteobacterial species. The first was discovered in the
Vibrio cholerae genome (Mazel et al. 1998). Located on the
smaller of the two circular V. cholerae chromosomes
(Heidelberg et al. 2000), this SI spanned 126 kb and harbored 179
cassettes of mainly unassigned function (Rowe-Magnus et al. 1999),
dwarfing any previously described MRI. Strikingly, the attCsites associated with each captured cassette, termed VCRs
(Vibrio cholerae repeated
sequences; Barker et al. 1994), were highly homologous in length and
sequence, unlike their counterparts of MRIs. In addition, the
cassettes of the V. cholerae SI were demonstrated to be
substrates for the class 1 integrase of MRIs (Mazel et al. 1998;
Rowe-Magnus et al. 2002b). These observations led to the proposal
that each distinct attC site of MRI gene cassettes was
representative of a distinct SI. Using a systematic search, similar
SIs were identified in the genomes of diverse proteobacteria
(Rowe-Magnus et al. 2001), and they all shared the same general
characteristics as the V. cholerae SI. Other
integron-integrase genes and hundreds of gene cassettes have also
been directly isolated from a variety of soil samples (Nield et al.
2001). Although it is not known if these soil-derived integrases and
cassettes are of SI or MRI origin, it is now evident that SIs, like
MRIs, are widespread among bacterial populations.
Phylogenetic analysis based on the comparison of the 16S RNA and
SI-integrase genes revealed that the evolutionary history of the
integron systems paralleled that of the radiation, indicating that
integrons are ancient structures (Rowe-Magnus et al. 2001).
Furthermore, we observed from the limited sample of available cassettes
that their attC sites (1) appeared to be species-specific and
(2) highly conserved within each SI. We have now performed an in-depth
analysis of the SIs identified within divergent members of the
Vibrio lineage. These analyses confirmed the sedentarity of
the integron platform (intI gene and attI site). A
comparison of subsets of cassettes from the SIs of Vibrio
metschnikovii, Vibrio fischeri, and V. cholerae
indicated that the majority of the cassettes found in the SI of one
species were not found in that of another. Such cassettes may have
important roles in the exploitation of a particular niche by a
particular species; they lend credence to the hypothesis of an in vivo
cassette assembly process that is independently active within each
species. Although the signature attC sites of each species
showed limited sequence homology, they all displayed conserved
structural characteristics that support the idea that symmetry, rather
than sequence, is a key feature in the recognition of such a varied
collection of target recombination sequences by a single site-specific
recombinase. Finally, our demonstration of an active toxin/antidote
system encoded in a V. fischeri SI gene cassette that is
homologous to the control of cell
death (ccdAB) system of the F plasmid and the
presence of similar systems as gene cassettes in the V. choleraeN16961 and V. metschnikovii SIs might provide an
explanation for the apparent stability of the massive cassette arrays
in these species.
|
RESULTS
|
|---|
The Chromosomal Location of the Vibrio SI Loci
The sequence of the loci carrying the previously characterized
integrases from V. metschnikovii A267, Vibrio
parahaemolyticus 75.2, V. fischeri 103203, and
Listonella pelagia 10276.2 (Rowe-Magnus et al. 2001) were
extended downstream of the intIA genes. Novel SIs were also
identified in Vibrio natriegens 103193T, Listonella
anguillarum 63.36, and Vibrio vulnificus 75.4. Each of the
SIs identified among all the Vibrionaceae in this work were estimated
to be at least 100 kb by Southern analysis using the attC site
sequences as probes (see below). Sequence analysis showed that, as
in V. cholerae and Vibrio mimicus, the V.
metschnikovii, V. vulnificus, and L. anguillarum
intIA genes were located adjacent to the rpmL–rplT
ribosomal protein operon (Fig. 2). The
L. pelagia, V. natriegens, and V.
parahaemolyticus intIA genes were located downstream of a cluster
of conserved ORFs different from those in the V. cholerae
clade. The closest homologs of two of these, a rimJ ortholog
and ORF474, were on Chromosome 1 of V. cholerae (VC1309 and
VC1310, respectively; Fig. 2). However, a homolog of ORF588 could not
be identified on either of the two chromosomes in V. cholerae.
ORF588 did show significant homology (40% identity) to an ORF of
unknown function in Deinococcus radiodurans (accession no.
E75485). The V. fischeri integrase gene, VfiintIA,
was located next to an ORF that was most closely related (at 39%
identity) to VCA0034 of V. cholerae N16961. VCA0034
resides on chromosome 2 of V. cholerae, 250 kb away from
the V. cholerae SI-integrase gene, VchintIA.
Comparison of the intergenic sequence between these intIAgenes and the neighboring genes or ORFs did not reveal any
conserved motifs.
View larger version (38K):
[in this window]
[in a new window]
|
Figure 2. Chromosomal location of the Vibrio superintegrons.
Representation of the genetic context in which each of the SIs was
found relative to their phylogenetic distribution according to their
IntIA genes is shown. SIs in identical locations are grouped
within the same phylogenetic clade. The corresponding orthologs in
V. cholerae are marked with VC(A) followed by the number
designation for the ORF. The attI site and VXRs are also
indicated. For clarity, only the first cassettes within each SI are
shown. The integrases intI1 (Liebert et al. 1999), intI2(Sundstrom et al. 1991), intI3 (Hall et al. 1999),
intI9 SXT (Hochhut et al. 2001), and intIHS (H.
Sorum, K. Dommarsnes, K. Sandersen, L. Sundstrom, M. Gullberg, and A.
Solberg, 2001, GenBank accession no. AJ277063) are found associated
with mobile DNA elements. The integron-interases I8-2, I7-2, and I6-2
were amplified from DNA soil samples (Nield et al. 2001). The
integrases (IntIA) of V. cholerae, V. mimicus,
V. metschnikovii, V. parahaemolyticus, V.
fischeri (Vfi), L. pelagia (Lpe), Shewanella
oneidensis (Son), Shewanella putrefaciens (Spu),
Xanthomonas campestris (Xca), Xanthomonas species(Xsp), Nitrosomonas europaea (Neu), P.
alcaligenes (Pal), and Pseudomonas mendocina (Pm) have
been previously described (Rowe-Magnus et al. 2001; Vaisvila et al.
1999, 2001). The recombinases XerC and XerD are from E. coli
(Eco); (rplT and rpmL) ribosomal genes; (dark gray,
striped, or black boxes) adjacent gene(s) that are not part of the SI
structure.
|
|
Characterization of the V. metschnikovii Cassette Array
Southern hybridization of BglII- or EcoRI-digested
V. metschnikovii genomic DNA with a VMeR probe
(Vibrio metschnikovii
repeats, the signature attC site of the V.
metschnikovii cassettes) revealed a series of fragments totaling
>100 kb for the VMeR cassette array (data not shown). Four of these
nonoverlapping fragments (Table 1) were
cloned and sequenced, providing a total of 26 cassettes (Table
2). An alternative PCR strategy, using
VMeR1 and VMeR2 as primers, allowed the retrieval of 6 additional
cassettes and recloning of the c253-5 cassette (Table 2). Among the
total of 32 cassettes recovered, two were found to be repeated twice
and six times, respectively, reducing the pool to a total of 26 unique
cassettes. Interestingly, six of these unique cassettes were found to
have counterparts in the V. cholerae SI (see section below).
Twenty-three were found to carry a plausible ORF that was preceded by a
potential ribosome-binding site (RBS). The putative products of seven
cassettes did not show any significant similarity to any characterized
or hypothetical proteins found in the databases (Table 2). The
potential ORFs encoded in all but two cassettes were found in the
classical (positive) orientation for integron cassettes, that is, the
3' end of the ORF was located just upstream of or within the ICS of the
downstream VMeR. The likely start codons of most of these classically
orientated ORFs were located within the first 40 nt of the cassettes,
leaving little space for a potential promoter. The ORFs carried by
c374-2 and c374-4 were identical, and both copies were in the inverse
(negative) orientation with respect to the VMeR. Interestingly, the
start codons for the ORFs inside these negatively oriented cassettes
were located 89 and 109 bp upstream of the VMeR, which provides
sufficient space to encoded potential promoter sequences. Four
cassettes, of sizes varying from 464–690 nt, including the cassette
present in six copies, did not contain a recognizable ORF.
Characterization of the V. fischeri Cassette Array
Southern hybridization of BamHI- or
HindIII-digested V. fischeri genomic DNA with a VFR
probe (Vibrio fischeri
repeats, the V. fischeri cassette attC
sites) revealed a series of fragments equaling >100 kb for the VFR SI
cassette array (data not shown). Six fragments were cloned and
sequenced (Table 1), and two were found to overlap. These fragments
provided a total of 27 cassettes. Two cassettes were found to be
repeated twice and eight times, respectively, giving a total of 19
unique cassettes. Sixteen of these were found to carry an ORF larger
than 250 bp that was preceded by a potential RBS. As seen in V.
cholerae and V. metschnikovii, the majority of the ORFs
were found in the classical orientation, and a single cassette, c669-2,
was found to carry two ORFs in inverse orientation. This cassette
encoded a functional cytotoxic protein and its antidote (see below).
The putative products encoded in seven cassettes, totaling eight ORFs,
were found to have similarity to previously characterized proteins
(Table 3) that are related to simple
functions or previously described ORFs. Analysis of the coding
potential of the cassette repeated eight times, c667-2, revealed no
significant ORF larger than 150 bp; however, BLASTX analysis revealed a
region with significant similarity to a 40-amino-acid domain of intron
maturases. None of the V. fischeri cassettes showed any
similarity to any of the SI cassettes found in the SIs of the other
Vibrio species examined.
The V. fischeri ccdAB Cassette Encodes a Cytotoxic Protein and Its Antidote
The V. fischeri cassette c669-2 was found to carry two ORFs
related to the control of cell death systems that form the
gyrase-inhibiting family of proteins, CcdB, and their antidotes, CcdA.
Interestingly, the closest relatives of the cassette encoded Ccd
proteins were neither functionally nor structurally associated. The
closest relative of CcdBVfi was found to be the
CcdBF of the plasmid addiction system from the F plasmid
(42% amino acid identity), whereas CcdAVfi was only 19%
identical to the associated antidote, CcdAF (Miki et al.
1984). Furthermore, CcdAVfi was found to be 42% identical to
the antidote protein, CcdAO157, found on the chromosome of
Escherichia coli O157:H7 (accession no. NP_285744; Perna et
al. 2001), whereas CcdBVfi shared only 35% identity with the
toxin, CcdBO157 (accession no. NP_2857445.1). To functionally
characterize this putative ccdAB system, the 739-bp
NsiI fragment from p669 encompassing the two ORFs of V.
fischeri cassette c669-2, hereafter called ccdAB, was
cloned into pNOT218 that had been digested with PstI (Table
1). This plasmid, p1357, contained the ccdAB genes in the
opposite orientation relative to the lacZ promoter. To ensure
that the ccdAB operon would be expressed in E. coli,
the p1357 insert was recloned in pTZ19R, placing the ccdAB
genes in the same relative orientation as the inducible lac
promoter (p1400, Table 1). The toxigenic activity associated with
ccdB expression was demonstrated by subcloning the
AclI–EcoRI internal fragment of p1400, which
contained only the 3' end of ccdA and an intact ccdB,
into pSU19 (CmR) digested by AccI and EcoRI
to put ccdB under the control of Plac. The ligation
product was then used to transform DH5 and  106, a DH5 strain
containing the plasmid p1400. Whereas transformation of 106 gave
rise to thousands of CmR clones, only a single CmR
clone was obtained in transformations with DH5. Furthermore,
characterization of the single DH5 CmR clone by sequence
analysis of the corresponding plasmid (p1446) revealed a C  T
transition in the ccdB coding region. This transition resulted
in the conversion of codon 35 from an Arg (CGA) to a stop codon (TGA),
leading to early termination of ccdB translation. These
results demonstrated that expression of ccdB in the absence of
ccdA coexpression was lethal in E. coli.
Analysis and Comparison of the VXRs
We developed a program, XXR, that could detect the attC
sites of integron gene cassette arrays. We used XXR to extract the
attC sites from the cassette array of the V. cholerae
N16961 SI and recovered 176 complete VCRs. We also used the program to
extract the VMeRs (V. metschnikovii) and VFRs (V.
fischeri) from the various contigs that we built. Dendograms of the
VXRs associated with the cassettes from the different contigs and
V. cholerae N16961 were compiled using CLUSTALX and TreeView
(Fig. 3). As can be seen,
149 of the 176 VCRs showed an overall similarity of >90%. The other
27 VCRs partitioned into three more remote subclades. All but one VMeR,
VMeR 253-5, grouped together to form a distinct clade from the group
gathering the 149 VCRs. Interestingly, most of the 27 remote VCRs
appeared to be more closely related to the VMeRs as they branched among
them. The VFRs also formed a coherent clade distantly related to the
VCRs and VMeRs. The attC sites carried by the cassettes found
in multiple copies inside the V. metschnikovii and V.
fischeri SIs formed subclades within the respective VXR families
(Fig. 3), indicating that they all descended from a common ancestor.
View larger version (16K):
[in this window]
[in a new window]
|
Figure 3. Phylogenetic relationships of the VXRs. The unrooted dendrogram was
compiled following extraction of the repeat sequences from the SIs of
V. cholerae (VCRs in black ), V. metschnikovii (VMeRs
in blue), and V. fischeri (VFRs in red) using the XXR program.
The VXRs of the other species are omitted for clarity. The significance
of the red- and green-framed VXRs are discussed in the text. Filled
boxes denote VXRs that are descended from a common ancestor.
|
|
Comparison of the Related Cassettes Found in the V. metschnikovii and V. cholerae SIs
An important issue to be addressed is the degree of cassette
exchange among Vibrio species. Among the 32 V.
metschnikovii cassettes examined, 6 were found to have counterparts
in the V. cholerae SI. However, only two of these, c253-3 and
c253-5, were complete, and subsequent analysis focused on their
phylogenetic relationship to their V. cholerae counterparts.
Interestingly, these two families of cassettes showed different types
of relationships. Cassette c253-3 of V. metschnikovii was
related to two nearly identical V. cholerae cassettes, VCA0338
and VCA0415. Cassette c253-3 was 84.5% and 83% identical to VCA0338
and VCA0415, respectively. However, this overall homology was in fact
partitioned within the cassettes; that is, the ORFs of the cassettes
shared 91%–92% identity, but the VXR components shared only
62%–63% identity. In addition, the VMeR carried by cassette c253-3
was closely related to the majority of the VMeRs, whereas the VCRs
carried by the V. cholerae cassette homologs, VCA0338 and
VCA0415, branched with the majority of the VCRs (see Fig. 3).
Conversely, cassette c253-5, which is unique in our V.
metschnikovii cassette sample, is related to a V. cholerae
cassette that appears twice in the N16961 SI, cassettes VCA0414 and
VCA0425. Cassette c253-5 shares 65%–66% identity with VCA0414 and
VCA0425. Here, however, we observed that the divergences between the
ORFs and the VXRs were in the same range, 63%–64% and 73%–75%
identical, respectively. Noticeably, the repeat sequences of the
VCA0414 and VCA0425 cassettes in V. cholerae have close
homologs to unrelated V. cholerae cassettes, whereas the
repeat sequence from its V. metschnikovii counterpart was less
related to any of the VMeRs from our cassette sample (best score 70%
identity) and clearly branch out of the VMeRs (Fig. 3).
Structural Characteristics of the VXRs
We have analyzed the different families of Vibrio attC
sites for the presence of specific structural features similar to those
described for the VCRs by Manning and colleagues (Barker et al. 1994).
Like the VCRs, the VFRs and the VMeRs were also found to show imperfect
dyad symmetry with the potential to adopt stable single-strand
secondary conformations. Figure 4 shows
examples of such potential secondary structures found for different
representatives of the VXR families. We noticed that, in all cases, a
potential stem structure was formed starting 4 nt downstream of the
last "C" of the ICS sequence RYYYTAAC and hybridizing
with a sequence ending 4 nt upstream of the first "G" in the CS
sequence GTTARRY. More interestingly, we observed that the
nucleotides in positions 5 to 8 in the ICS arm of this stem, including
a protruding nucleotide (a C in 201 of the 219 VXRs analyzed), as well
as the complementary nucleotides in the CS arm of the stem were
conserved in all of the 219 VXRs analyzed (see alignment in
supplementary figure; available online at http://www.genome.org). We
also noticed that these specific characteristics were conserved in the
PAR sequences recently described for the Pseudomonas
alcaligenes SI cassettes (data not shown; Vaisvila et al. 2001). In
each case, this "protruding C" region was found to overlap with the
2L/2R complementary sequences described by Hall and collaborators;
however, we could not identify any convincing homology to the 2L and 2R
consensus sequences within these regions. As shown in Figure 4, we
observed that in addition to the protruding C mentioned above,
protruding nucleotides were always found in the ICS arm of the stem and
never in the CS arm, regardless of the primary sequence of the VXR.
The VCRs of the V. cholerae SI cassettes were found in most
cases to have a stretch of 9–11 consecutive complementary nucleotides.
The majority of the VFRs and VMeRs also displayed a complementary
stretch in the same length range, 9–11 nt. The CS consensus sequences
found for the VCRs, VMeRs, and VFRs were determined to be GTTAKGYN,
GTTAtRYK, and GTTAtRYg, respectively (the consensus is written in
capital letters if >75% of the VXRs carry the same type of nucleotide
at that position and in small letters if the incidence is
50%–75%; [K] G or T; [Y] C or T; [N] A, C, G, or T).
|
DISCUSSION
|
|---|
The recent evidence exalting the impact of lateral gene transfer
(LGT) on bacterial evolution supports the conclusion that LGT is both
an ancient and a perpetual phenomenon that has been a major force in
shaping the genomes of extant species. The relatively recent discovery
of integrons left open the question of the extent, beyond the
dissemination of antibiotic-resistance genes, to which this particular
system had an impact on bacterial evolution. The discovery of ancestral
chromosomal SIs in diverse proteobacteria established this system as a
key player in the evolution and plasticity of bacterial genomes
(Rowe-Magnus and Mazel 2001). We used several independent approaches to
analyze the SIs among remote species in the Vibrionaceae. We examined
the SIs and the regions flanking the integron integrases for (1)
clusters of orthologous genes, (2) conservation of local gene order,
(3) the distribution of orthologs between species, and (4) comparison
of phylogenetic trees constructed from 16S RNA and rplT and
intIA genes. The data presented here from our extended
analysis of known and novel SIs in members of the Vibrionaceae supports
the contention of an ancient and a perpetual role for integrons in the
engineering of extensive genetic diversity in this and other genera.
Integrons are Ancient Structures
We previously established that the evolutionary history of the
superintegron platforms, that is, the intIA genes and their
associated attI sites, paralleled that of the Vibrio
radiation according to 16S RNA analysis, indicating that integrons are
ancient structures (Rowe-Magnus et al. 2001). We have now extended our
analysis through the identification and characterization of new SI
platforms within the genomes of Vibrio species remote from
those previously characterized. Every Vibrio species examined
thus far has been found to harbor an SI. In addition, we have now
identified the SI integrase and the first cassettes carried in the SIs
of L. anguillarum and V. vulnificus, a fish pathogen
and an emerging human pathogen, respectively. V. natriegens
was also included in our study to validate the recent proposal for
reclassification of the L. pelagia species CIP 10276.2 to the
V. natriegens group (Macian et al. 2000). We observed that the
intIA genes, as well as the chromosomal genes located at the
SI boundary, were almost identical in L. pelagia CIP 10276.2
and V. natriegens 103193T, lending strong support to the
reassignment of L. pelagia strain CIP 102762 to V.
natriegens.
To provide additional support to the aforementioned congruence of the
16S rRNA and intI-based phylogenetic trees, we have
characterized the genetic context in the neighborhood of the SIs by
examining the chromosomal region flanking the intIA boundary
in each Vibrio species. As shown in Figure 2, we observed that
the location was in agreement with the grouping drawn from the
phylogeny. Indeed, the position of the SI island on the chromosome was
conserved within, but not between, the subclades, that is, the
intIA genes in the V. cholerae clade were all located
downstream of the same ribosomal operon, and all these strains belong
to the same phylogenetic clade. Likewise, the intIA genes of
the SIs in the V. parahaemolyticus clade were also located
within the same genetic context, but their chromosomal location was
different from that of the V. cholerae clade. Interestingly,
the integrase gene of the V. vulnificus SI, VvuintIA,
branched within the V. parahaemolyticus clade but was located
in the same genetic context as the intIA genes of the V.
cholerae clade. This indicated that the chromosomal rearrangement
that led to the change in SI location from the neighborhood of the
rplT gene to that of ORF474 likely occurred in the ancestor of
V. parahaemolyticus subsequent to its divergence from V.
vulnificus. In accordance with the topology of the intI
tree, the location of the V. fischeri SI was
completely unrelated to the other Vibrios, as its
intIA was found adjacent to an ORF most closely related to
VCA0034, which is located 250 kb away from the SI on Chromosome 2 of
V. cholerae N16961. The conserved chromosomal context within
the respective subclades in the intIA phylogenetic tree could
be very useful in recovering the SIs from other species within the same
subclade.
Informational genes, that is, those involved in numerous
protein–protein interactions such as ribosomal genes, are proposed to
be refractory to LGT (Jain et al. 1999). For this reason, ribosomal
genes have been used as the cornerstone for phylogenetic analysis. Our
phylogenetic analysis of the intIA genes revealed that the
V. fischeri intIA gene had the deepest branching point among
all the integron-integrases, but the V. fischeri branching
point derived from 16S rRNA analysis placed this species inside the
Vibrio radiation, between V. metschnikovii and
V. cholerae (Rowe-Magnus et al. 2001; Vaisvila et al. 2001).
Lateral transfer of the integron platform in V. fischeri would
provide a simple explanation for this inconsistency. However, although
less problematic in the prokaryotic realm, it has been shown that rRNA
trees can sometimes be misleading in inferring phylogenies because of
differences in base composition and unequal rates of evolution among
species (Philippe and Laurent 1998). Hence, phylogenies based on
alternative or multiple gene sets are being used more frequently to
support the accuracy of rRNA trees.
In a phylogenetic analysis based on the ribosomal L20 genes
(rplT), another essential chromosomal gene unlikely to be
subject to horizontal transfer, we retrieved a branching order
congruent with the intIA gene-based tree for all species
examined (Rowe-Magnus et al. 2001). This discrepancy might be
attributed to either a specific and congruent evolution of the
intIA and rplT genes compared with the 16S rRNA genes
in these species, or to an acquisition of the intIA and
rplT genes from an unknown and phylogenetically remote
bacterial source in V. fischeri. Although this question is
still open, two observations in the work described here lead us to
favor the first hypothesis. First, if we examine the integrase genes of
V. cholerae and V. fischeri we observe 39% identity
between them. The sole homolog to the ORF abutting VfiintIA is
VCA0034 of V. cholerae, at 34% identity. These values are
within the same range and are consistent with the placement of these
species in the rplT gene. This indicates that these three
genes, intIA, rplT, and the VCA0034 orthologs,
coevolved together within these species. Second, if acquired from an
exogenous source, the evolutionary congruence among intIA,
rplT, and the VCA0034 ortholog in V. fischeri implies
the simultaneous and relatively recent capture of all three genes.
Because rplT is not in the vicinity of VfiintIA and
the VCA0034 ortholog, this would necessitate the large-scale
acquisition of a genetic element bearing an essential gene, that is, a
chromosome, in a V. fischeri ancestor. Furthermore, the
replacement of a single ribosomal component, such as rplT, or
a subpopulation of components would give rise to heterogeneous
ribosomal complexes that are likely to have a drastic effect on cell
viability. Maintenance of such a chromosomal element thus seems highly
unlikely. Likewise, coevolution of the three genes following their
independent capture within a small evolutionary timeframe is also an
improbable scenario. Finally, comparison of the intergenic sequences
located between the different intIA and their adjacent
chromosomal genes did not reveal any conserved sequences or structures
to indicate that SIs are mobile. We favor the notion of a specific and
congruent evolution of the intIA and rplT genes
compared with the 16S rRNA genes in these species, and believe that the
V. fischeri SI integrase branching point is consistent with
the ancient and sedentary attributes of the system, and its coevolution
with the host genome.
The Gene Cassette Reservoir
A comparison of subsets of cassettes from the SIs of V.
metschnikovii, V. fischeri, and V. cholerae
indicated that the majority of the cassettes found in the SI of one
species were not found in that of another. None of the 19 unique
V. fischeri cassettes had a counterpart in the SIs of the
other Vibrio species examined, and only 6 out of the 32
V. metschnikovii cassettes examined were found to have
counterparts in the V. cholerae SI. Two of these, c253-3 and
c253-5, were complete and allowed further comparative study with their
V. cholerae counterparts. Interestingly, these two families of
cassettes showed different degrees of relatedness between the ORF and
attC portions of the cassettes. Our analyses indicated that
some common cassettes, such as c253-3 of V metschnikovii and
VCA0338 of V. cholerae, were independently assembled within
these bacteria or that a conversion mechanism corrected the associated
XXR sequence into the signature attC site of that species
following acquisition. However, the existence of cassettes carrying
remote attC sites within the same SI renders this second
alternative less probable. For comparison of other cassettes, such as
c253-5 of V. metschnikovii and VCA0414 of V.
cholerae, it is tempting to speculate that these specific cassettes
both descended from a common ancestral cassette that was constructed in
V. cholerae or a closely related Vibrio species and
was then subsequently acquired in V. metschnikovii via
horizontal transfer. Thus, some cassettes may have been constructed
independently in the two species, whereas others were likely
descendents of a common ancestral cassette that were acquired via
horizontal transfer. The unique cassettes may have important roles in
the exploitation of a particular niche by a particular species, and
their shear numbers lend credence to the hypothesis of an in vivo
cassette assembly process that is independently active within each
species.
Analysis of the attC sites carried by the cassettes found in
the V. cholerae N16961 SI showed that a large majority
( 85%) of the cassette-associated VCRs shared an overall similarity
of >90% (Fig. 3), as previously observed among the first 13 VCRs
identified (Barker et al. 1994). The rest of the VCRs showed at most
70% identity to the first group and were more closely related to VXRs
of other species, such as the VMeRs of V. metschnikovii (Fig.
3). Because completely unrelated ORFs were associated with almost
identical VCRs, this indicated that (1) the VCRs and the ORFs had
independent origins, and the cassettes corresponded to the addition of
an attC site to a gene through a specific assembly process;
and (2) the cassettes carrying an attC site that was poorly
related to the VCRs likely had an exogenous origin and have been
acquired as complete cassettes through lateral transfer.
Cassette-Encoded Functions
The activity of only a handful of SI cassettes has been demonstrated
experimentally. These include pathogenicity factors (van Dongen et al.
1987; Ogawa and Takeda 1993), antibiotic-resistance determinants
(Melano et al. 2002; Rowe-Magnus et al. 2002b), metabolic genes (Barker
and Manning 1997; Rowe-Magnus et al. 2001), and restriction enzymes
(Rowe-Magnus et al. 2001; Vaisvila et al. 2001; R. Vaisvila, R. Morgan,
and E. Raleigh, unpubl.). As seen for the V. cholerae and the
P. alcaligenes SI cassettes (Heidelberg et al. 2000;
Rowe-Magnus et al. 2001; Vaisvila et al. 2001), a large number of the
V. metschnikovii and V. fischeri cassette-encoded
genes have no counterparts in the databases or the sole homologs are
ORFs of unassigned function (Tables 2 and 3). We also noticed that the
putative products of a large number of cassettes are likely associated
with the cell envelope as they contain signal peptide sequences and/or
transmembrane segments (Tables 2 and 3). This might be indicative of a
role for these cassettes in membrane integrity, variation, or
adaptation to environmental changes.
Potential Secondary Structure of the attC Sites
Perhaps the most striking aspect of the activity of
integron-integrases is their ability to recognize remote DNA sequences
as targets (Mazel et al. 1998; Hall et al. 1999; Collis et al. 2001;
Rowe-Magnus et al. 2001; Drouin et al. 2002; Hansson et al. 2002).
Several common structural features among attC sites have been
previously identified from the comparison of various
antibiotic-resistance cassettes (Francia et al. 1997; Stokes et al.
1997). All attC sites possess (1) an internal imperfect dyad
symmetry, (2) a core-site (CS) consensus of sequence GTTAGSC ([S] G
or C, originally defined as GTTRRRY) and a perfectly complementary
inverse core site (ICS) of consensus GSCTAAC (Hansson et al. 1997;
Stokes et al. 1997), and (3) two sets of inversely oriented integrase
binding sites, the LH and RH simple sites. We observed that regardless
of the primary sequence of the VXR, in all cases a potential stem
structure was formed starting 4 nt downstream of the ICS and
hybridizing with a sequence ending 4 nt upstream of the CS sequence.
More interestingly, we observed that the protruding C region and the
nucleotides immediately flanking it in the stem were conserved in all
but a few exceptions of the 219 VXRs analyzed. We also noticed that
these specific characteristics were conserved in the much shorter
attC sites of other SIs that have been recently identified
(Vaisvila et al. 2001). In each case, this protruding C region was
found to overlap with the 2L/2R complementary sequences described by
Hall and collaborators; however, we could not identify any convincing
homology to the 2L and 2R consensus sequences in these regions. Despite
this, the VCRs have been experimentally shown to be substrates for the
integrase of class 1 integrons (Mazel et al. 1998; Rowe-Magnus et al.
2001, 2002b). We observed that, in addition to the protruding C,
protruding nucleotides were always found in the ICS arm of the stem and
never in the CS arm sequence of the VXR. Along with the lack of
identifiable 2L and 2R sites within the VXRs, these results support the
hypothesis that, outside of the ICS and CS, the particular structural
characteristics that result from the primary sequence in the imperfect
inverted repeat regions of attC sites are more important in
defining an attC site as a target for the integrase than the
specific sequence itself. It would thus seem that symmetry, rather than
sequence, is a key feature in the recognition of such a varied
collection of target recombination sequences by a single site-specific
recombinase. We are presently conducting experiments to test this
hypothesis.
Microevolution Versus Macroevolution in SIs
Postsegregational killing (PSK)
systems are normally found either on plasmids or within prophages.
These systems generally consist of a pair of genes organized in an
operon, with the downstream gene specifying a stable toxin and the
upstream gene specifying a specific but unstable antitoxin. Once the
operon is expressed, the cells are "addicted" to the short-lived
antidote polypeptide, because its continued de novo synthesis becomes
essential for cell survival (Engelberg-Kulka and Glaser 1999). Because
loss of the extrachromosomal elements bearing these modules selectively
kills the cured cells, these types of genetic systems have been found
to enhance plasmid segregation and phage maintenance. The chromosomal
equivalents are known as control of cell death (CCD) systems. We have
demonstrated that one of the V. fischeri cassettes encodes an
addiction module of the ccdAB family, a gyrase toxin and its
antidote. Other chromosomal- or plasmid-encoded PSK systems that
inhibit varied essential cellular functions have been characterized
(Jensen and Gerdes 1995; Couturier et al. 1998; Engelberg-Kulka and
Glaser 1999), but this is the only example, with the exception of a
ccdAB operon found on the E. coli O157:H7 chromosome
(Perna et al. 2001), of a gyrase-targeted killing system that is not
carried by a plasmid (Couturier et al. 1998). Cassettes carrying genes
with functions related to other PSK systems can also be identified in
other SIs, such as the putative parDE (VCA0359/VCA0360) and
higAB (VCA0391/VCA0392) cassette homologs of V.
cholerae, and the death on curing (doc) P1 toxin
analogs identified in the V. cholerae and V.
metschnikovii SIs. Upstream of the doc-related gene
(ORFc253-6b) in the V. metschnikovii cassette c253-6, is a
small ORF (ORFc253-6a, Table 2), which almost certainly encodes a
prevents host death (phd) antidote counterpart, even if its
product, apart from a similar small size, shows little homology to the
P1 phd gene. Interestingly, even if the annotated V.
cholerae N16961 genome did not reveal a bona fide phd-like
antidote gene associated with the doc analog, careful analysis
of the sequence located upstream of the doc analog (VCA0475)
led us to identify the same putative phd gene. However, in
V. cholerae, a deletion event occurred inside the upstream
phd–doc cassette, resulting in its fusion with an acetyl
transferase gene, masking both the gene and the real boundary with the
phd–doc cassette.
Several features, including the presence of the phd–doc,
higAB, and parDE orthologs within the V.
cholerae SI, led Heidelberg and coworkers to suggest that
Chromosome 2 of V. cholerae was originally a megaplasmid that
was captured by an ancestral Vibrio species (Heidelberg et al.
2000). The subsequent relocation of essential genes from Chromosome 1
to this megaplasmid then created a chromosome, enhancing the stable
maintenance of this smaller replicon. The similar nucleotide
composition and percentage G + C content between the two chromosomes
would support the acquisition event to have occurred
hundreds of millions of years ago. However, two features of the PSK
systems weaken the argument that Chromosome 2 was once a megaplasmid.
First, all three of these PSK systems are structured as gene cassettes
and could have been acquired at any time. Second, in addition to being
mobile, the GC composition of the aforementioned PSK gene cassettes is
quite different (40%–44%) from the rest of the chromosome (47%).
This contradicts the megaplasmid hypothesis because, if the
phd–doc, higAB, or parDE PSK gene cassettes
participated in the initial retention of the megaplasmid in an
ancestor Vibrio species, then they would be expected to have a
GC composition more in line with the genome. This indicates a more
recent origin for the PSK systems and that they were acquired by a
preexisting Chromosome 2.
An intriguing question is what is the function of these chromosomal
addiction modules? PSK/CCD systems may have varied complex roles in
bacterial behavior and evolution. Bacteria are unicellular organisms.
Hence, they would not be expected to undergo programmed cell death
(PCD). However, Glaser and colleagues recently discovered a chromosomal
PCD locus, mazEF, in E. coli. This locus has the same
organization as the PSK systems, encoding a stable toxin and an
unstable but specific antidote, and functions in a similar manner.
Although the target of the MazF toxin is not known, the operon was
shown to be subject to regulation by guanosine 3',5'-bispyrophosphate
and antibiotics that are general inhibitors of transcription or
translation (Aizenman et al. 1996; Sat et al. 2001). These compounds
act as signaling molecules to trigger PCD in E. coli. This
discovery has prompted biologists to revisit the view that bacteria may
behave in multicellular ways. Furthermore, Orejas and coworkers have
shown that the antidote of a second chromosomal ccd loci in
E. coli, chpB, can neutralize the plasmid-borne toxin
of the parD system (Santos Sierra et al. 1998). These types of
functional interactions between systems may play an important role in
bacterial evolution by permitting the acquisition of plasmids bearing
only the toxin partner of the operon or, to reduce the genetic burden
on the bacterium, the loss of plasmids carrying homologous PSK systems.
Clark et al. (2000) examined the global SI organization of 65 different
V. cholerae O serotypes by PCR and Southern hybridization.
Extensive restriction polymorphism was observed even among closely
related isolates, indicating a plasticity in the SI structures and in
their microevolution through integrase-mediated events. This is not
unexpected because the gene cassettes are independently mobile genetic
units that are supposedly subject to episodic selection and we have
demonstrated that the integron–integrase randomly excises cassettes
from SIs (Rowe-Magnus et al. 2002b). SIs may contain hundreds of
attC sites and multiple copies of gene cassettes, insertion
sequences, and pseudogenes. Because identical and likely functional
copies of the same cassettes are found in the V. cholerae SI
(such as the three copies of the glutathione-transferase, VCA0328,
VCA0341, and VCA0463; Heidelberg et al. 2000; Rowe-Magnus et al.
2002b), it is unlikely that a mechanism specifically inactivating
multicopy cassettes exists in these species to suppress homologous
recombination between them. It is also commonly proposed that
pseudogenes are rapidly eliminated from bacterial genomes by selection
to provide streamlined, compact chromosomes to minimize the genetic
burden (Lawrence et al. 2001), and insertion sequences are renowned for
their activities in restructuring genomes (Shapiro 1999). Clearly,
ample opportunities exist for large-scale deletions and rearrangements
to occur between attC sites, IS elements, or multicopy
cassettes, so the macroevolution of SIs would also be anticipated. Yet,
paradoxically, SIs can remain remarkably stable. The genetic
organization of MRIs promotes coexpression of the inserted gene
cassettes from a single promoter, PC, that is
internal to the integrase gene. Hence, selection for one resistance
determinant often coselects for the maintenance of the entire array.
However, a similar situation for SIs is difficult to imagine. We
propose that the PSK/CCD systems act to stabilize the massive arrays of
SI cassettes. A parDE paralogous gene family of seven members
has been identified in the V. cholerae genome (gvc family 331,
http://www.tigr.org/tigr-scripts/CMR2/ParalogousList.spl?db=gvc).
Interestingly, all seven of these paralogs are structured as gene
cassettes, bringing the total number for PSK-type systems in the
V. cholerae SI to nine. Kobayashi and colleagues have shown
that restriction-methylation systems (RMSs) fit all the properties of
PSK systems (Kobayashi 1998; Kobayashi et al. 1999). Indeed, once
acquired, they become essential for the survival of the bacteria
because the long half-life of the nuclease compared with the methylase
will eventually cause cell death if the RMS is lost (Kusano et al.
1995). Thus, the bacteria become dependent on the invading RMS. This
facet has been demonstrated to enhance plasmid segregation stability in
E. coli and Bacillus subtilis (Kulakauskas et al.
1995; Handa et al. 2000). We suggest that the presence of PSK systems
in the SIs of the Vibrionaceae or RMS in the SIs of
Xanthamonads (Rowe-Magnus et al. 2001) and
Pseudomonads (Vaisvila et al. 2001) may act to stabilize these
massive arrays of independently mobile genetic units by minimizing
large-scale random excisions, because a probability exists to lose or
shut off expression of the PSK or RMS cassettes. Such an event would
result in cell death. In this scenario, the maintenance of other
cassettes within these arrays would not have to arise from a direct
link to the PSK or RMS cassettes and the array could be stably
maintained in the absence of selection. Thus, the selfish nature of
PSK/CCD systems could contribute positively to the fitness of bacteria,
and this may help to explain the presence of these types of systems
within SIs. Experiments are presently underway to test this hypothesis.
|
METHODS
|
|---|
Bacterial Strains, PCR, and Sequencing
Bacterial strains were provided by the Collection de l'Institut
Pasteur (CIP). All genomic DNA was isolated by using the QIAGEN DNEasy
Kit. Genomic DNA from each species was digested with the indicated
restriction enzyme (V. metschnikovii, BamHI,
EcoRI, or EcoRV; V. fischeri,
BamHI, HindIII, or XbaI; L.
anguillarum, HindIII), and partial libraries were
constructed as previously described (Rowe-Magnus et al. 2001).
PCR-amplified genes were cloned using the TA TOPO cloning kit
(Invitrogen) and verified by sequencing of the corresponding genomic
clones performed by MWG-Biotech. Primers were obtained from Genset.
Sequence and Phylogenetic Analysis
ORFs of at least 150 nt were identified using the ORF Finder at the
National Center for Biotechnology Information
(http://www.ncbi.nlm.nih.gov/gorf/gorf.html). Each amino acid sequence
was used as a BLAST query (Altschul et al. 1997) to search the GenBank,
EMBL, SWISS-PROT, and the NCBI unfinished Microbial Genome BLAST Web
site sequence databases. Signal peptides were identified with SignalP
program v1.1 at the Center for Biological Sequence Analysis
(http://www.cbs.dtu.dk/services/SignalP; Nielsen et al. 1997).
Transmembrane proteins were predicted with TMHMM at the Center for
Biological Sequence Analysis at the Technical University of Denmark
(http://www.cbs.dtu.dk/ services/TMHMM-1.0/). Putative promoter
sequences were predicted using the BDGP promoter software program
(http://www.fruitfly.org/seq_tools/promoter.html). DNA secondary
structures were determined using MFOLD analysis
(http://bioweb.pasteur.fr/seqanal/interfaces/mfold-simple.html).
Alignments were generated with the CLUSTALX software program (version
1.8), and dendrograms were compiled by using the neighbor-joining
method (computed from 1000 independent trials) of CLUSTALX and PHYLIP
or TREEVIEW.
Developing the XXR Program
To automatically identify nucleotide sequences that are
structurally organized as integron cassette arrays in any nucleotide
sequence, we developed the program XXR. XXR is written in Perl (5.6
version) and is functional on any platform that uses this programming
language (Linux, Unix, Cygwin). In principle, using any nucleotide
sequence file in FASTA format, XXR is able to extract putative cassette
structures that fulfill the criteria established from analysis of
previously known cassettes from integrons and superintegrons. These
criteria are (1) anchoring at an AAC to define the end of the inverse
core site (ICS) and a GTT to define the start of the core site (CS);
(2) the minimum and maximum length of the complementary sequences
between the ICS and the cognate upstream CS in the cassette as well as
their identification; and (3) the maximum length of the two building
blocks of a cassette, the ORF and its associated attC site, in
order to avoid aberrant cassettes. Most parameters can be modified;
however, some, like the anchor sequences in the CS (GTT) and in the ICS
(AAC) and the minimal length of the complementarity, are fixed. The
most important point is the definition of the ICS, as this sequence is
the anchoring point for the research of the full cassette sequence. The
boundary is then searched in the neighborhood upstream and downstream
of the ICS. After analysis of a given sequence, XXR outputs two
different file types: a file of data gathering all the extracted
cassettes in FASTA format and XML (eXtensible
Markup Language) files for each sequence that
likely corresponds to a cassette. Figure 5
shows the procedure followed by XXR for the retrieval of cassette
sequences from the input sequence file. XXR can be used through our Web
site, http://www.pasteur.fr/recherche/unites/ptmg/integ.
View larger version (29K):
[in this window]
[in a new window]
|
Figure 5. Graphical representation of the search parameters used to retrieve
attC sites by the program XXR. See text for details.
|
|
Cloning Novel SI Loci and Gene Cassettes in the Vibrionaceae
Gene cassettes from the V. metschnikovii SI were
identified by screening of EcoRI and BglII partial
libraries with the primers VMeR1 (GTCCCTCTTGAAGCGTTTGTTA) and VMeR2
(GC CCCTTAAGCGGGCGTTA). The same strategy was used to screen V.
fischeri BamHI, HindIII, and XbaI
partial libraries with the primers VFR1 and VFR2 (Rowe-Magnus et al.
2001). Cloning of the integrase genes from the SI loci carried in
V. vulnificus 75.4 and V. natriegens 103193 was
performed by PCR using primers LPR1 (CGATCCCTCTTGAAGTTTGTTA) and I456/5
(TCTTGTAC(C/G)GT(A/T)CGTATATCA) or LPR1-2 (GAATCACTCTTAAACAGTTTGTTA)
and LPINT8-2 (CCTT TACCTTGCCAGACACG), respectively. LPR1 and LPR1-2
were designed from the comparison of the attC sites of the
cassettes found in the L. pelagia SI (LPRs), and correspond to
the outer end of the LPR. I456/5 was designed from comparison of known
Vibrio SI integrases and corresponds to a highly conserved
segment at the 3' end of the gene. The PCR products were cloned using
the TA TOPO cloning kit to give p1184 and p1216. Sequence analysis of
the corresponding PCR products allowed us to identify most of the
intIA genes, the cognate attI sites, and the first
cassettes of both the V. vulnificus and V. natriegens
SIs. Southern analysis of L. anguillarum genomic DNA probed
with the V. fischeri intIA gene showed that the
LanintIA gene was carried on a 2.5-kb HindIII
fragment. Screening of a L. anguillarum HindIII
partial library with the same probe allowed the isolation of p1456, a
plasmid carrying the corresponding insert. Sequence analysis revealed
that this fragment carried the intIA gene and the first three
cassettes of the SI. As the L. anguillarum intIA gene was most
closely related to the V. metschnikovii intIA gene, we tested
whether the LanintIA was located in the same chromosomal
context, that is, downstream of the rplT gene (Fig. 2). This
was achieved by PCR using RPLT (ATGCCTCGCGTAAAACGTGGTGTAC), a primer
corresponding to the highly conserved 5' region of the rplT
gene, and Lang1 (TCGGCATTGCAGCGAGCAGTTCGG), a primer internal to the
LanintIA (p2010, Table 1). The result was a 630-nt product
that was similar in size to the corresponding fragment in V.
metschnikovii. Sequencing of this fragment confirmed the
intIA location downstream of the rplT gene (Fig. 2).
Using an analogous PCR strategy with oligonucleotides Vna1
(GATTTCTTTCAGACACCGCTCTCAC) and ORF474 (GACTGAATGTCTTATTTGCCTTTGG), we
were able to show that the V. natriegens intIA gene was
located in the same genetic context as the L. pelagia and
V. parahaemolyticus IntIA genes (p2018, Table 1).
Identification of the V. vulnificus intIA chromosomal context
was achieved by using the same PCR strategy with oligonucleotides RPLT
and Vvu1 (CTG CAGAAACAGGCACTCATCAGGATG; p2019, Table 1).
Cloning the ccd Cassette of V. fischeri
A 739-bp NsiI fragment of p669 that encompassed the two
ORFs of the V. fischeri cassette c669-2 was cloned into
pNOT218 that had been digested with PstI. This plasmid, p1357,
contained the ccdAB genes in the opposite orientation relative
to the lacZ promoter. To ensure that the ccdAB operon
would be expressed in E. coli, the p1357 insert was recloned
into pTZ19R, placing the ccdAB genes in the same relative
orientation as the inducible lac promoter (p1400). The
toxigenic activity associated with ccdB expression was
demonstrated by subcloning the AclI–EcoRI internal
fragment of p1400, which contained only the 3' end of ccdA and
an intact ccdB, into pSU19 digested by AccI and
EcoRI in order to put ccdB under the control of
Plac (p1446). The ligation product was then used to transform
either DH5 or 106, a DH5 strain containing the plasmid p1400.
|
WEB SITE REFERENCES
|
|---|
http://bioweb.pasteur.fr/seqanal/interfaces/mfold-simple.html; MFOLD
analysis, Institut Pasteur.
http://www.cbs.dtu.dk/services/SignalP; SignalP program v1.1, the
Center for Biological Sequence Analysis.
http://www.cbs.dtu.dk/services/TMHMM-1.0/; transmembrane protein
prediction, the Technical University of Denmark.
http://www.fruitfly.org/seq_tools/promoter.html; BDGP promoter software
program, Berkley Drosophila Genome Project.
http://www.ncbi.nlm.nih.gov/gorf/gorf.html; ORF Finder, the National
Center for Biotechnology Information.
http://www.pasteur.fr/recherche/unites/pmtg/integ; authors' Web site.
http://www.tigr.org/tigr-scripts/CMR2/ParalogousList.spl?db=gvc;
paralogous gene families, The Institute for Genomic Research.
http://www.ncbi.nlm.nih.gov/BLAST/; Similarity search programs, the
National Center for Biotechnology Information.
|
Acknowledgements
|
|---|
We dedicate this paper to the memory of Pr. Maurice Hofnung, former
Chef de l'Unité PMTG, for his encouragement, guidance and
friendship. D.R-M. is an EMBO and a Fondation de la Recherche Medicale
(FRM) Post-doctoral fellow. This work was supported by the Institut
Pasteur, the CNRS, and the Programme de Recherche Fondamentale en
Microbiologie et Maladies Infectieuses et Parasitaires from the MENRT
and the DGA (contract no. 0134020).
The publication costs of this article were defrayed in part by payment
of page charges. This article must therefore be hereby marked
"advertisement" in accordance with 18 USC section 1734 solely to
indicate this fact.
|
Footnotes
|
|---|
1 Present address: Department of Microbiology, Sunnybrook &
Women's College Health Sciences Center, Toronto, Ontario, Canada, M4N
3N5; and the Department of Laboratory Medicine & Pathobiology, Faculty
of Medicine, University of Toronto, Toronto, Canada. 
2 Corresponding author.
E-MAIL mazel{at}pasteur.fr; FAX 33 1 45 68 87 90.
Article and publication are at
http://www.genome.org/cgi/doi/10.1101/gr.617103.
|
REFERENCES
|
|---|
Aizenman, E., Engelberg-Kulka, H., and Glaser, G. 1996. An Escherichia coli chromosomal "addiction module" regulated by guanosine [corrected] 3',5'-bispyrophosphate: A model for programmed bacterial cell death. Proc. Natl. Acad. Sci. 93: 6059-6063.[Abstract/Free Full Text]
Altschul, S.F., Madden, T.L., Schaffer, A.A., Zhang, J., Zhang, Z., Miller, W., and Lipman, D.J. 1997. Gapped BLAST and PSI-BLAST: A new generation of protein database search programs. Nucleic Acids Res. 25: 3389-3402.[Abstract/Free Full Text]
Barker, A. and Manning, P.A. 1997. VlpA of Vibrio cholerae O1: The first bacterial member of the 2-microglobulin lipocalin superfamily. Microbiology 143: 1805-1813.[Abstract]
Barker, A., Clark, C.A., and Manning, P.A. 1994. Identification of VCR, a repeated sequence associated with a locus encoding a hemagglutinin in Vibrio cholerae O1. J. Bacteriol. 176: 5450-5458.[Abstract/Free Full Text]
Caporale, L.H. 1999. Chance favors the prepared genome. Ann. NY Acad. Sci. 870: 1-21.[Abstract/Free Full Text]
Clark, C.A., Purins, L., Kaewrakon, P., Focareta, T., and Manning, P.A. 2000. The Vibrio cholerae O1 chromosomal integron. Microbiology 146: 2605-2612.[Abstract/Free Full Text]
Collis, C.M., Kim, M.J., Stokes, H.W., and Hall, R.M. 1998. Binding of the purified integron DNA integrase Intl1 to integron- and cassette-associated recombination sites. Mol. Microbiol. 29: 477-490.[CrossRef][Medline]
Collis, C.M., Recchia, G.D., Kim, M.J., Stokes, H.W., and Hall, R.M. 2001. Efficiency of recombination reactions catalyzed by class 1 integron integrase IntI1. J. Bacteriol. |
![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
Top
ABSTRACT
RESULTS
and 
), and
the asterisk (*) indicates the position of the protruding C present in
2L relative to 2R. The extents of the LH and RH sites as defined by
Stokes et al. (1997)