Genetic background of Escherichia coli and extended-spectrum [beta]-lactamase type.

To assess the implication of the genetic background of Escherichia coli strains in the emergence of extended-spectrum [beta]-lactamases (ESBL), 55 TEM-, 52 CTX-M-, and 22 SHV-type ESBL-producing clinical isolates involved in various extraintestinal infections or colonization were studied in terms of phylogenetic group, virulence factor (VF) content (pap, sfa/foc, hly, and aer genes), and fluoroquinolone resistance. A factorial analysis of correspondence showed that SHV type, and to a lesser extent TEM type, were preferentially observed in B2 phylogenetic group strains that exhibited numerous VFs but were fluoroquinolone-susceptible, whereas the newly emerged CTX-M type was associated with the D phytogenetic group strains that lacked VF but were fluoroquinolone-resistant. Thus, the emergence of ESBL-producing E. coli seems to be the result of complex interactions between the type of ESBL, genetic background of the strain, and selective pressures in ecologic niches.

**********

Extended-spectrum [beta]-1actamases (ESBL) that mediate resistance to oxyimino-cephalosporins, such as cefotaxime, aztreonam, and ceftazidime, are now observed worldwide in all species of Enterobacteriaceae (1). Traditionally, ESBLs are derived by point mutation from the common TEM and SHV-1 [beta]-lactamases. However, recently, new families of ESBLs have been described (2). The CTX-M-type ESBLs have become particularly widespread and are mainly found in strains of Salmonella and Escherichia coli (3,4). These enzymes probably evolved from chromosomal [beta]-lactamases of Kluyvera spp. by gene transposition from mobile elements and mutation (5,6). ESBLs are usually described as acquired [beta]-lactamases that are encoded mainly by genes located on plasmids. Some ESBL-encoding genes are located within transposons or integrons, which facilitates transfer between organisms. ESBL-producing organisms are responsible for nosocomial infections, and many hospitals have experienced outbreaks (1,2,7). The lower digestive tract of colonized patients has been recognized as the major source of ESBL-producing organism (2,8). These organisms pose a therapeutic challenge, since they are frequently resistant to other kinds of antimicrobial drugs, including aminoglycosides, quinolones, and cotrimoxazole (2).

E. coli in humans is a commensal inhabitant of the gastrointestinal tract. It can also cause various intestinal and extraintestinal diseases (9). Strains causing infections harbor numerous virulence factors encoded on plasmids, bacteriophages, or the bacterial chromosome within pathogenicity islands (9). Several studies have shown that pathogenic E. coli strains may be derived from commensal strains by acquiring chromosomal or extrachromosomal virulence operons (10,11). Phylogenetic analyses have shown that E. coli strains fall into 4 main phylogenetic groups (A, B1, B2, and D) (12,13). Although virulence determinants are considered to be mobile, a link between strain phylogeny and virulence has been reported. Virulent extraintestinal strains belong mainly to group B2 and, to a lesser extent, to group D, whereas most commensal strains belong to groups A and B1. Strains of phylogenetic groups B2 and D often carry virulence determinants that are lacking in group A and B1 strains (10,14-17). In addition, a trade-off between resistance and virulence has been observed. Prevalence of antimicrobial resistance was shown to be greater in non-B2 phylogenetic group strains (18). In urinary tract infections, fluoroquinolone-resistant E. coli represented predominantly low-virulence phylogenetic groups A and B1 (19). These resistant strains were also associated with a decrease in the presence or the expression of some virulence factors and a decreased invasive capacity (20,21).

The intrinsic virulence potential of ESBL-producing E. coli is unknown. They may represent traditional virulence clones of extraintestinal pathogenic E. coli (ExPEC) or low-virulence opportunists whose ability to cause disease is largely limited to compromised hosts, in which antimicrobial resistance might provide relevant selective advantage. To assess the relationships between the genetic background of the strains and the presence of an ESBL, we analyzed a collection of ESBL-producing E. coli clinical isolates involved in various extraintestinal infections or in colonization in terms of phylogenetic grouping, virulence determinant content, and fluoroquinolone resistance.

Material and Methods

Bacterial Strains

We collected 157 E. coli isolates from clinical samples on the basis of their positive double-disk synergy test from 1997 to 2002 in different areas in France: Paris area (4 hospitals), Brest, and Amiens. From these isolates 129 strains were analyzed on the basis of 3 criteria: 1) the strains produced an ESBL, 2) the strains were epidemiologically unrelated, and 3) the strains were unambiguously classified as responsible for infection or colonization. ESBLs were characterized by isoelectric focusing with ceftriaxone and penicillin as substrates (7), specific polymerase chain reaction (PCR) amplification, and direct sequencing of PCR products. The oligonucleotide primer sets specific for the [beta]-1actamase gene (bla) amplification and sequencing were taken from the literature ([bla.sub.TEM] and [bla.sub.SHV]) (22) or designed in this study ([bla.sub.CTX-M]) (Table 1). As the family of CTX-M ESBLs belongs to 4 clusters on the basis of their protein sequences, the CTX-M-1 cluster (CTX-M-1, CTX-M-3, CTX-M-10, CTX-M-12, CTX-M-15), the CTX-M-2 cluster (CTX-M-2, CTX-M-4 to CTX-M-7, Toho-1), the CTXM-9 cluster (CTX-M-9, CTX-M-14, CTX-M-16, CTX-M-18, CTX-M-19, Toho-2), and the CTX-M-8 cluster, specific primers for each cluster of the CTX-M family were designed. PCR products of [bla.sub.TEM] were subjected to direct sequencing to identify TEM-ESBLs, only when isolates produced a single [beta]-lactamase indicated by isoelectric focusing. For isolates carrying a second [beta]-1actamase of pI 5.4 or 5.6 shown by penicillin only (putative TEM-1 or TEM-2 [beta]-1actamase), sequences were obtained after plasmid transfer into E. coli K-12 J53-2 [rif.sup.r] (23). PCR product sequences were then compared to reported ESBL sequences and assigned to specific types or clusters. To identify any epidemiologic relationship between the strains, they were compared by using enterobacterial repetitive intergenic consensus (ERIC)-PCR with ERIC1 and ERIC2 as primers (24,25). When strains had identical electrophoretic profiles with both ERIC1 and ERIC2 primers, they were considered identical, and only 1 isolate per electrophoretic profile type was selected for further analysis. Among the collection of 129 strains selected for the study, 86 strains were involved in infections (urinary tract infection [UTI]: 64, bacteriemia: 7, pus production from miscellaneous infections: 15), and 43 strains were isolated from colonization (rectal samples: 39, gastric aspirate: 1, abdominal drainage: 1, vaginal sample: 1, tracheal aspirate: 1) (Table 2). The collection included 55 strains that produced a TEM-type ESBL, 22 strains produced a SHV-type ESBL, and 52 strains produced a CTX-M type ESBL (Table 2).

Susceptibility Testing, Phylogenetic Grouping, and Virulence Factors

Susceptibility to ciprofloxacin was tested by the disk diffusion technique according to the guidelines of the Antibiogram Committee of the French Society for Microbiology (www.sfm.asso.fr) with MIC criteria of [less than or equal to] 1 mg/L (diameter [greater than or equal to] 22 mm) used to define susceptibility. Phylogenetic grouping of the E. coli isolates was determined by a PCR-based method developed by Clermont et al. (26) that uses a combination of 3 DNA markers (chuA, yjaA, and an anonymous DNA fragment, TspE4.C2). Strains were assigned to phylogenetic groups on the basis of presence or absence of the 3 DNA fragments: chuA-, TspE4.C2-, group A; chuA-, yjaA-, Tspe4.C2+, group B1; chuA+, yjaA+, group B2; chuA+, yjaA-, group D. Because 2 possible profiles can be obtained for the groups A, B2, and D, each was subdivided as follows: chuA-, yjaA-, Tspe4.C2-, group A subgroup [A.sub.0]; chuA-, yjaA+, Tspe4.C2-, group A subgroup [A.sub.1]; chuA+, yjaA+, Tspe4.C2-, group B2 subgroup [B2.sub.2]; chuA+, yjaA+, Tspe4.C2+, group B2 subgroup [B2.sub.3]; chuA+, yjaA-, Tspe4.C2-, group D subgroup [D.sub.1]; chuA+, yjaA-, Tspe4.C2+, group D subgroup [D.sub.2]. Virulence genes (pap, sfa/foc, hly, aer) were detected from DNA by PCR as described previously (15,27). These genes code for 2 adhesins (pyelonephritis-associated pill system and S fimbrial adhesin), 1 toxin ([alpha]-hemolysin), and 1 iron captation system. These genes are good representatives of the intrinsic extraintestinal virulence of the strains (28).

Statistical Analysis

Data were summarized in 2 two-way tables, and each table had 129 rows, one for each E. coli strain. The first table had 16 columns corresponding to the variables, origin of the strains, phylogenetic group or subgroup, type of ESBL, and virulence factors. The second table had 12 columns corresponding to the variables, phylogenetic groups, type of ESBL, and resistance to ciprofloxacin. For each column, each strain was coded as a binary code: present = 2, absent = 1. A factorial analysis of correspondence (FAC) (29) was conducted from this table with SPAD.N software (Cisia, Saint Mande, France). To confirm the significance of the correlation observed with FAC, [chi square] tests were carried out.

Results

Characterization of ESBL Strains

Among the 129 E. coli strains analyzed, phylogenetic group B2, which is the source of most ExPEC clones, was represented by 36.4 % of the strains (8.5% were subgroup [B2.sub.2] and 27.9% were subgroup [B2.sub.3]). Phylogenetic group D, which is also a source of ExPEC but to a lesser extent, was represented by 25.5% of the strains (17% were subgroup [D.sub.1] and 8.5% were subgroup [D.sub.2]). Of the remaining strains, phylogenetic groups A and B1 were represented by 27.9% (9.3% were subgroup [A.sub.0] and 18.6% were subgroup [A.sub.1]) and 10% of the strains, respectively. The virulence determinants most represented in the collection were aer and pap, with 53 (41%) and 38 (29.5%) strains carrying these genes, respectively. Less prevalent were sfa/foc and hly determinants, with only 18 (14%) and 19 (15%) positive strains, respectively. Fluoroquinolone resistance was present in 34.8% of the strains.

ESBL-producing strains were found in all E. coli phylogenetic groups. Of the strains, 60% and 24% harbored at least 1 or 2 extraintestinal virulence determinants, respectively. Coresistance to fluoroquinolones was frequent.

Multidimensional Analysis

To assess relationships between phylogenetic groups, VFs, type of ESBL produced, and origin of the strains (infection or colonization), a FAC was constructed with the 129 E. coli strains as individuals and the 16 characteristics as qualitative variables. Projections of the variables on the plane Fl/F2 (Figure A), which accounted for 34.5% of the total variance, showed a correlation between the type of ESBL produced and several phylogenetic group/subgroups of E. coli. Thus, SHV type and subgroup [B2.sub.3] are projected on the positive values of F1 and negative values of F2, whereas TEM type and subgroup [B2.sub.2] are projected on the positive values of F1 and F2. CTX-M type and subgroup [D.sub.2] are projected on the negative values of F1 and F2. Correlation between SHV type and subgroup [B2.sub.3] was confirmed by [chi square] tests (p < 0.001) and the CTX-M type and the subgroup [D.sub.2] (p < 0.001) (Table 3).

As previously reported, sfa/foc and hly VFs were exclusively found in strains of the subgroups [B2.sub.2] and [B2.sub.3] (10,16). Pairwise comparisons between individual subgroups showed that subgroups [B2.sub.2] and [B2.sub.3] each had mean VF scores (1.45 and 1.8, respectively) significantly higher than either phylogenetic groups and subgroups [A.sub.0], [A.sub.1], B1, [D.sub.1], or [D.sub.2] (p < 0.02 for all comparisons), but they were not significantly different from one another. Likewise phylogenetic groups and subgroups [A.sub.0], [A.sub.1], B1, [D.sub.1], and [D.sub.2] were not significantly different from one another with respect to mean VF scores (mean scores 0.5, 0.66, 0.54, 0.5, and 0.63, respectively). When the type of ESBL produced was considered, the frequency of VFs was higher in SHV-producing strains (mean score = 1.8) than in TEM-producing strains (mean score = 0.96). The lowest frequency was found in the CTX-M-producing strains (mean score = 0.6). FAC stressed these 2 observations, as it showed that the pap, sfa/foc, and hly VFs were projected on the positive values of the first axis with the subgroup [B2.sub.3] and the SHV type. The correlation between SHV type and the presence of the 3 VFs was also confirmed by [chi square] tests (pap, p < 0.01; sfa/foc, p < 0.001; hly, p < 0.001). Aerobactin was found in all the phylogenetic groups and subgroups, and no correlation was observed with the FAC (Table 4).

Projection of the colonization and infection variables on the plane showed that they were clearly distinguished by the first factor and that there was a correlation with some phylogenetic groups (Figure A). The colonization characteristic was projected on the positive values of F1 with phylogenetic subgroups [A.sub.0] and [D.sub.1]. The association was close to significance ([A.sub.0], p = 0.05; [D.sub.1], p = 0.06): strains of subgroups [A.sub.0] and [D.sub.1] were isolated more frequently from colonization (relative risk [RR] of 3.15 and 2.34, respectively) (Table 3). If we consider the clones usually to be the major source of ExPEC, strains of the subgroup [B2.sub.2] were equally distributed among the strains responsible for infection or colonization (8.1% versus 9.3%), but strains of subgroup [B2.sub.3] were more numerous among the strains responsible for infection than for colonization (32.5% versus 18.6%); the correlation was close to significance (p = 0.09, RR = 2.11) (Table 3). TEM type was also projected on the positive values of F2 with the colonization characteristic, and the [chi square] test confirmed the correlation (p = 0.03).

The mean VF score of the strains responsible for infection was significantly higher (p = 0.03) than the mean VF score of the strains responsible for colonization (1.1 and 0.76, respectively). However, when each VF was considered, only the frequency of aerobactin was significantly higher among the strains responsible for infection (p = 0.03) than the strains responsible for colonization.

To assess the relationships between phylogenetic groups and subgroups, ESBL type, and resistance to fluoroquinolones, a second FAC was performed, taking into account only these variables (Figure B). Projection of the variables on the plane F1/F2, which accounted for 34% of the total variance, showed a correlation between resistance to ciprofloxacin and type of ESBL produced. Thus, the ciprofloxacin-resistant characteristic was projected on the negative values of the first factor with CTX-M-type, and the ciprofloxacin-susceptible characteristic was projected on the positive values of the first factor with TEM and SHV types. Significant differences were observed between the rate of resistance to fluoroquinolones among the CTXM-(51.9%) and among the SHV- and TEM-producing strains (13.6% and 27.7%, respectively): CTX-M type / SHV type, p = 0.002 and CTX-M type / TEM type, p = 0.009. FAC stressed also the correlation between the subgroup [D.sub.1] and the resistance to ciprofloxacin, which were projected together on the negative values of the first factor and on the positive values of the second factor. The correlation was confirmed by the [chi square] test (p = 0.03). Strains of phylogenetic subgroup [D.sub.1] had the highest resistance rate (54%), and strains of subgroups [B2.sub.2], [B2.sub.3], and [A.sub.0] had the lowest resistance rates (18%, 25%, and 25%, respectively). Group/subgroups B1, [D.sub.2], and [A.sub.1] had ciprofloxacin resistance rates of 30.7%, 36%, and 45%, respectively. No significant difference was seen in the frequencies of ciprofloxacin resistance among strains from infection or colonization (38.3% versus 27.9%). The mean VF score of the ciprofloxacin-susceptible strains was significantly higher (p < 0.001) than the one of the ciprofloxacin-resistant strains (1.2 and 0.6, respectively) (Table 4). We found hly and sfa/foc exclusively in ciprofloxacin-susceptible strains, and the frequency of pap was significantly higher among ciprofloxacin-susceptible strains (p = 0.04) than among ciprofloxacin-resistant strains. No difference was observed in the frequency of aerobactin between the 2 groups (Table 4). Although the frequency of CTX-M type was higher among UTI strains than among non-UTI strains (Table 2), FAC analysis and [chi square] tests did not show any significant association between UTI strains, phylogenetic group or subgroup, individual VFs, and ciprofloxacin resistance (data not shown), which could explain some of the previously observed correlations.

Therefore, strains harboring ESBL of SHV and TEM types belonged preferentially to the B2 phylogenetic group. They possessed extraintestinal VFs, but ESBL TEM-type strains were more likely to be isolated from cases of colonization; they were also susceptible to fluoroquinolones. On the other hand, strains harboring ESBL of CTX-M type were associated with [D.sub.2] phylogenetic subgroup, had few VFs, but were resistant to fluoroquinolones.

Discussion

This study was designed to assess the role of the genetic background of strains of E. coli in the emergence of ESBL. Strains were sampled from hospitals in several distant areas, which allowed us to build up a collection of strains producing variants of the most prevalent ESBL types. Thus 3 groups of ESBL-types were collected, TEM-, SHV-, and CTX-M-type, having enough strains in each group to be compared. Spread of clones of ESBL-producing organisms can occur from cross-contamination among patients (2,7,23). Therefore, to avoid redundant strains, we used ERIC-PCR as a typing method, and strains with similar profiles were eliminated.

Several studies suggested that extraintestinal pathogenic E. coli strains are mostly derived from the B2 phylogenetic group and to a lesser extend from the D group (15,16,30-34). It had been estimated in collections dating from before the emergence of ESBL, or in collections not selected for ESBL production, that group B2 strains account for approximately two thirds of all extraintestinal E. coil infections, including UTI, bacteremia, meningitis, and other miscellaneous infections. When all ESBL-producing E. coli strains were considered, whatever their types were, group B2 represented only 39.4% of the strains responsible for infection in our study. Thus, production of ESBL among E. coli clinical strains isolated from infection was associated with shifts in phylogenetic distribution toward non-B2 phylogenetic groups, in particular groups D and A. The distribution of group B2 among strains isolated from infection or from colonization was not very different even if it was pointed out that subgroup [B2.sub.3] strains had a tendency to be isolated more frequently in clinical infections. Johnson et al., in 1991 (18), observed that E. coli strains belonging to phylogenetic groups other than group B2 have a greater prevalence of antimicrobial resistance, such as to ampicillin, tetracycline, chloramphenicol, streptomycin, and sulfonamide; express significantly fewer virulence factors; and invade more commonly compromised hosts. ESBL-producing organisms, which are resistant to [beta]-lactams, except carbapenems and cephamycins, and are often resistant to other antimicrobial drugs, are responsible for nosocomial infections, mostly in immunocompromised patients. ESBL-producing organisms also frequently colonize the lower digestive tract, and therefore are a major source for ESBL propagation (8). This finding may explain why two thirds of the strains in our study were not traditional virulence clones of ExPEC but clones whose ability to cause infection is limited to compromised hosts, in whom antibiotic resistance might provide selective advantage.

ESBLs are acquired [beta]-lactamases that are encoded mainly by genes located on plasmids (2). As such, they are a recent evolutionary development. Even if the genetic element that carries resistance is a mobile element, the multidimentional analysis showed a preferential association between the genetic background and the type of ESBL produced by the strains. Thus, an association was seen between SHV type and subgroup [B2.sub.3], between TEM type and subgroup [B2.sub.2], and CTX-M type and subgroup [D.sub.2]. Even more, the pap, sfa/foc, and hly VFs were associated with the genotype SHV type/subgroup [B2.sub.3], defining a potentially high-virulence group of ESBL-producing E. coli strains. In contrast, the genotype CTX-M type/subgroup [D.sub.2], characterized by a low VF score, defined a potentially low-virulence group of ESBL-producing E. coli strains. The type of ESBL produced by E. coli could be a predictive factor for intrinsic virulence potential.

Organisms that produce ESBL are frequently resistant to other antimicrobial agents, such as aminoglycosides, tetracycline, and trimethoprim-sulfamethoxazole, as many of these additional resistance genes are encoded on the ESBL-associated plasmid. Fluoroquinolone resistance, which is also frequently associated with ESBL production, is usually chromosomally encoded, unlike the other coresistances. However, plasmid-mediated quinolone resistance has been discovered recently (35). Prevalence of fluoroquinolone resistance among ESBL-producing strains varies according to geographic regions (36), from 13.7% in Canada to 65.5% in the western Pacific. In our study, 34.8% of strains were resistant, which is close to the prevalence (34.2%) reported in Europe (36). Correlation with phylogenetic background and VF profiles showed highly fluoroquinolone-resistant strains of subgroup [D.sub.1], with the lowest VF score and association with colonization. In contrast, strains of phylogenetic group B2, which had the highest VF score, were among the strains with the lowest fluoroquinolone-resistance rates. These data agree with the work of Johnson et al. (19,37) and show a clear trade-off between resistance to fluoroquinolones and virulence. In addition, our study highlights an association between these fluoroquinolone-resistant strains and CTX-M-producing strains, which are devoid of VFs. However, the search for the gene responsible for plasmid-mediated quinolone resistance, qnr, by PCR was negative in our collection of ESBL-producing strains (O. Zamfir, E. Denamur, C. Branger, unpub, data). Thus, the observed association is not due to a genetic link between resistance to expanded-spectrum [beta]-1actams and quinolones on a mobile element, as was recently reported (38).

During the last 2 decades, most of the ESBL found in E. coli and, in general, in gram-negative bacilli, has been of TEM or SHV lineage. Recently TEM and SHV types have been replaced by CTX-M-type ESBL, whose emergence and proliferation are particularly noteworthy (39). The current spread may be explained in part by the ability of some insertion sequence elements to mobilize and promote the expression of [beta]-1actamase (40). However, the high rate of fluoroquinolone resistance and the low virulence of the strains canting CTX-M ESBL could provide them selective advantage to spread, especially under strong environmental antimicrobial pressure with fluoroquinolones.

In summary, mobile elements encoding ESBL are not randomly distributed among the genetic diversity of the E. coil species. The arrival, expression, and maintenance of such elements seem to be the result of complex interactions between the type of ESBL, the phylogenetic background, the intrinsic virulence of the strains, and the presence of associated fluoroquinolone resistance. Such complexity reflects very likely the diversity of ecologic niches with different selective pressures.

Table 1. Sequence of primers used to detect bla genes *

                      Primer                          Reference or
PCR target             name      Primer sequence      accession no.

[bla.sub.TEM]           A       ATGAGTATTCAATTCCG         (22)
                        B      CTGACAGTTACCAATGCTTA
[bla.sub.SVH]           P4     GGTTATGCGTTATATTCGCC       (22)
                        P5      TTAGCGTTGCCAGTGCTC
[bla.sub.CTX-M]        MenA    AAGACTGGGTGTGGCATTGA      X92506
  (CTX-M-1 cluster)    MenB    AGGCTGGGTGAAGTAAGTGA
[bla.sub.CTX-M]        M2A     CTGGAAGCCCTGGAGAAAAG      X92507
  (CTX-M-2 cluster)    M2B     TACCTCGCTCCATTTATTGC
[bla.sub.CTX-M]        ToA     GCTTTATGCGCAGACGAGTG     AF174129
  (CTX-M-9 cluster)    ToB     GCCAGATCACCGCAATATCA
[bla.sub.CTX-M]         A8      GCCTGTATTTCGCTGTTG      AF189721
  (CTX-M-8 cluster)     B8     TGTCATTCGTCGTACCATAA

* PCR, polymerase chain reaction.

Table 2. Distribution of ESBL types according to
strain origin *

                     No. strains isolated from

ESBL type                 Other
(no. strains)     UTI   infections   Colonization

TEM (55)
TEM-24            11        9             11
TEM-52             3        0              7
TEM-21             5        1              2
TEM-3              1        0              3
TEM-10             0        0              1
TEM-20             1        0              0
SHV (22)
SHV2               3        1              2
SHV4               1        1              1
SHV5               2        0              1
SHV12              5        4              1
CTX-M (52)
CTX-M-1 cluster   20        2              7
CTX-M-2 cluster    3        0              1
CTX-M-9 cluster    9        4              6

* ESBL, extended-spectrum [beta]-lactamase;
UTI, urinary tract infection.

Table 3. Distribution of extended-spectrum [beta]-lactamase
types among Escherichia coli strains isolated from infection
or colonization, according to phylogenetic group

                     No. strains     No. (%) of isolates producing
                       in the
Phylogenetic group    group (%)       TEM         SHV        CTX-M

Infection
  [A.sub.0]            5 (5.8)      2 (6.4)     1 (5.8)     2 (5.2)
  [A.sub.1]           18 (20.9)     11 (35)        0       7 (18.4)
  B1                  9 (10.4)      2 (6.4)    3 (17.6)    4 (10.5)
  [B2.sub.2]           7 (8.1)      5 (16)      1 (5.8)     1 (2.6)
  [B2.sub.3]          28 (32.5)    8 (25.8)    10 (58.8)   10 (26.3)
  [D.sub.1]           11 (12.7)     3 (9.6)    2 (11.7)    6 (15.7)
  [D.sub.2]            8 (9.3)         0           0        8 (21)
  All groups          86 (100)     31 (100)    17 (100)    38 (100)

Colonization
  [A.sub.0]           7 (16.2)     5 (20.8)        0       2 (14.2)
  [A.sub.1]           6 (13.9)      2 (8.3)        0        4 (28)
  B1                   4 (9.3)     3 (12.5)        0        1 (7.1)
  [B2.sub.2]           4 (9.3)      2 (8.3)     1 (20)      1 (7.1)
  [B2.sub.3]          8 (18.6)     5 (20.8)     3 (60)         0
  [D.sub.1]           11 (25.5)     6 (25)      1 (20)     4 (28.4)
  [D.sub.2]            3 (6.9)      1 (4.1)        0       2 (14.1)
  All groups          43 (100)     24 (100)     5 (100)    14 (100)

All strains
  [A.sub.0]           12 (9.3)     7 (12.7)     1 (4.5)     4 (7.6)
  [A.sub.1]           24 (18.6)    13 (23.6)       0       11 (21.1)
  B1                   13 (10)       5 (9)     3 (13.6)     5 (9.6)
  [B2.sub.2]          11 (8.5)     7 (12.7)      2 (9)      2 (3.8)
  [B2.sub.3]          36 (27.9)    13 (23.6)    13 (59)    10 (19.2)
  [D.sub.1]            22 (17)     9 (16.3)    3 (13.6)    10 (19.2)
  [D.sub.2]           11 (8.5)      1 (1.8)        0       10 (19.2)
  All groups          129 (100)    55 (100)    22 (100)    52 (100)

Table 4. Frequency of virulence factors among
ciprofloxacin-susceptible and ciprofloxacin-resistant
Escherichia coli strains involved in infection or colonization,
according to extended-spectrum [beta]-lactamase (ESBL) type

                                     No. (%) strains carrying

ESBL type (no. strains)        pap       sfa/foc     h/y       aer

Ciprofloxacin resistance
  TEM (15)                    4 (20)        0         0      8 (53)
  SHV (3)                       0           0         0      1 (33)
  CTX-M (27)                  2 (7)         0         0      12 (44)
  All types (45)              6 (13)        0         0      21 (46)

Ciprofloxacin sensitivity
  TEM (40)                   14 (35)     9 (22)    9 (22)    9 (22)
  SHV (19)                   13 (68)     8 (42)    9 (47)    10 (52)
  CTX-M (25)                  5 (20)      1 (4)     1 (4)    13 (52)
  All types (84)             32 (55)     18 (21)   19 (22)   32 (38)

All strains
  TEM (55)                   18 (33)     9 (16)    9 (16)    17 (31)
  SHV (22)                   13 (59)     8 (36)     9 (4)    11 (50)
  CTX-M (52)                  7 (13)      1 (2)     1 (2)    25 (48)
  All types (129)           38 (29.5)    18 (14)   19 (15)   53 (41)

                            Virulence
                              factor
ESBL type (no. strains)     mean score

Ciprofloxacin resistance
  TEM (15)                     0.8
  SHV (3)                      0.33
  CTX-M (27)                   0.51
  All types (45)               0.6

Ciprofloxacin sensitivity
  TEM (40)                      1
  SHV (19)                     2.1
  CTX-M (25)                   0.8
  All types (84)               1.2

All strains
  TEM (55)                     0.96
  SHV (22)                     1.8
  CTX-M (52)                   0.6
  All types (129)               1

Acknowledgments

We thank Paulette Berry for technical assistance, Jean Philippe Emond for providing some of the strains, Matthieu Eveillard for advice about statistical methods, and Emmanuelle Cambau for help in the detection of the qnr gene.

This work was partially supported by grant from "La Fondation pour la Recherche Medicale" and the INSERM.

References

(1.) Bradford PA. Extended-spectrum [beta]-1actamases in the 21st century: characterization, epidemiology, and detection of this important resistance threat. Clin Microbiol Rev. 2001;14:933-51.

(2.) Gniadkowski M. Evolution and epidemiology of extended-spectrum [beta]-lactamases (ESBLs) and ESBL-producing microorganisms. Clin Microbiol Infect. 2001;7:597-608.

(3.) Paterson DL. Extended-spectrum [beta]-lactamases: the European experience. Curr Opin Infect Dis. 2001 : 14:697-701.

(4.) Saladin M, Cao VT, Lambert T, Donay JL, Herrmann JL, Ould Hocine Z, et al. Diversity of CTX-M [beta]-1actamases and their promoter regions from Enterobacteriaceae isolated in three Parisian hospitals. FEMS Microbiol Lett. 2002;209:161-8.

(5.) Humeniuk C, Arlet G, Gautier V, Grimont P, Labia R, Philippon A. [beta]-lactamases of Kluyvera ascorbata, probable progenitors of some plasmid-encoded CTX-M types. Antimicrob Agents Chemother. 2002;46:3045-9.

(6.) Poirel L, Kampfer P, Nordmann P. Chromosome-encoded Ambler class A [beta]-1actamase of Kluyvera georgiana, a probable progenitor of a subgroup of CTX-M extended-spectrum [beta]-1actamases. Antimicrob Agents Chemother. 2002:46:4038-40.

(7.) Branger C, Bruneau B, Lesimple AL, Bouvet PJ, Berry P, Sevali Garcia J, et al. Epidemiological typing of extended-spectrum [beta]-1actamase-producing Klebsiella pneumoniae isolates responsible for five outbreaks in a university hospital. J Hosp Infect. 1997;36:23-36.

(8.) Lucet JC, Regnier B. Enterobacteria producing extended spectrum [beta]lactamases. Pathol Biol. 1998;46:235-43.

(9.) Donnenberg MS. Escherichia coli: virulence mechanisms of a versatile pathogen. San Diego, California: Academic Press; 2002.

(10.) Johnson JR, Delavari P, Kuskowski M, Stell AL. Phylogenetic distribution of extraintestinal virulence-associated traits in Escherichia coli. J Infect Dis. 2001;183:78-88.

(11.) Finlay BB. Falkow S. Common themes in microbial pathogenicity revisited. Microbiol Mol Biol Rev. 1997:61:136-69.

(12.) Desjardins P, Picard B. Kaltenbock B, Elion J, Denamur E. Sex in Escherichia coli does not disrupt the clonal structure of the population: evidence from random amplified polymorphic DNA and restriction-fragment-length polymorphism. J Mol Evol. 1995;41:440-8.

(13.) Herzer PJ, Inouye S, Inouye M, Whittam TS. Phylogenetic distribution of branched RNA-linked multicopy single-stranded DNA among natural isolates of Escherichia coli. J Bacteriol. 1990;172:6175-81.

(14.) Duriez P. Clermont O, Bonacorsi S, Bingen E, Chaventre A, Elion J, et al. Commensal Escherichia coli isolates are phylogenetically distributed among geographically distinct human populations. Microbiology. 2001;147:1671-6.

(15.) Picard B, Garcia JS, Gouriou S, Duriez P, Brahimi N, Bingen E, et al. The link between phylogeny and virulence in Escherichia coli extraintestinal infection. Infect Immun. 1999;67:546-53.

(16.) Bingen E, Picard B, Brahimi N, Mathy S, Desjardins P, Elion J, et al. Phylogenetic analysis of Escherichia coli strains causing neonatal meningitis suggests horizontal gene transfer from a predominant pool of highly virulent B2 group strains. J Infect Dis. 1998;177:642-50.

(17). Boyd EF, Hartl DL. Chromosomal regions specific to pathogenic isolates of Escherichia coli have a phylogenetically clustered distribution. J Bacteriol. 1998:180:1159-65.

(18.) Johnson JR, Goullet P, Picard B, Moseley SL, Roberts PL, Stamm WE. Association of carboxylesterase B electrophoretic pattern with presence and expression of urovirulence factor determinants and antimicrobial resistance among strains of Escherichia coli that cause urosepsis. Infect Immun. 1991;59:2311-5.

(19.) Johnson JR, van der Schee C, Kuskowski MA, Goessens W, van Belkum A. Phylogenetic background and virulence profiles of fluoroquinolone-resistant clinical Escherichia coli isolates from the Netherlands. J Infect Dis. 2002;186:1852-6.

(20.) Vila J. Simon K, Ruiz J, Horcajada JR Velasco M, Barranco M, et al. Are quinolone-resistant uropathogenic Escherichia coli less virulent? J Infect Dis. 2002:186:10394-2.

(21.) Velasco M, Horcajada JP, Mensa J, Moreno-Martinez A, Vila J. Martinez JA. et al. Decreased invasive capacity of quinolone-resistant Escherichia coli in patients with urinary tract infections. Clin Infect Dis. 2001:33:1682-6.

(22.) Rashced JK. Jay C, Metchock B, Berkowitz F, Weigel L, Crellin J, et al. Evolution of extended-spectrum [beta]-lactam resistance (SHV-8) in a strain of Escherichia coli during multiple episodes of bacteremia. Antimicrob Agents Chemother. 1997;41:647-53.

(23.) Galdbart JO, Lemann F, Ainouz D, Feron P, Lambert Zechovsky N, Branger C. TEM-24 extended-spectrum [beta]-1actamase-producing Enterobacter aerogenes: long-term clonal dissemination in French hospitals. Clin Microbiol Infect. 2000:6:316-23.

(24.) Struelens MJ, Carlier E, Maes N, Serruys E, Quint WG, van Belkum A. Nosocomial colonization and infection with multiresistant Acinetobacter baumannii: outbreak delineation using DNA macrorestriction analysis and PCR-fingerprinting. J Hosp Infect. 1993;25:15-32.

(25.) Versalovic J, Koeuth T, Lupski JR. Distribution of repetitive DNA sequences in eubacteria and application to fingerprinting of bacterial genomes. Nucleic Acids Res. 1991 ; 19:6823-31.

(26.) Clermont O, Bonacorsi S, Bingen E. Rapid and simple determination of the Escherichia coli phylogenetic group. Appl Environ Microbiol. 2000;66:4555-8.

(27.) Picard B, Duriez P, Gouriou S, Matic I, Denamur E, Taddei F. Mutator natural Escherichia coli isolates have an unusual virulence phenotype. Infect Immun. 2001;69:9-14.

(28.) Johnson JR, Kuskowski M, Denamur E, Elion J, Picard B. Clonal origin, virulence factors, and virulence. Infect Immun. 2000;68:424-5.

(29.) Greenacre M. Correspondence analysis in medical research. Stat Methods Med Res. 1992;1:97-117.

(30.) Johnson JR, Stell AL. Extended virulence genotypes of Escherichia coli strains from patients with urosepsis in relation to phylogeny and host compromise. J infect Dis. 2000;181:261-72.

(31.) Johnson JR, O'Bryan TT, Kuskowski M, Maslow JN. Ongoing horizontal and vertical transmission of virulence genes and papA alleles among Escherichia coli blood isolates from patients with diverse-source bacteremia. Infect Immun. 2001;69:5363-74.

(32.) Bingen-Bidois M, Clermont O, Bonacorsi S, Terki M, Brahimi N, Loukil C, et al. Phylogenetic analysis and prevalence of urosepsis strains of Escherichia coli bearing pathogenicity island-like domains. Infect Immun. 2002;70:3216-26.

(33.) Bonacorsi S, Clermont O, Houdouin V, Cordevant C, Brahimi N, Marecat A, et al. Molecular analysis and experimental virulence of French and North American Escherichia coli neonatal meningitis isolates: identification of a new virulent clone. J Infect Dis. 2003;187:1895-906.

(34.) Zhang L, Foxman B, Marrs C. Both urinary and rectal Escherichia coli isolates are dominated by strains of phylogenetic group B2. J Clin Microbiol. 2002;40:3951-5.

(35.) Martinez-Martinez L, Pascual A, Jacoby GA. Quinolone resistance from a transferable plasmid. Lancet. 1998;351:797-9.

(36.) Winokur PL, Canton R, Casellas JM, Legakis N. Variations in the prevalence of strains expressing an extended-spectrum [beta]-lactamase phenotype and characterization of isolates from Europe, the Americas, and the Western Pacific region. Clin Infect Dis. 2001;32(Suppl 2):S94-103.

(37.) Johnson JR, Kuskowski MA, Owens K, Gajewski A, Winokur PL. Phylogenetic origin and virulence genotype in relation to resistance to fluoroquinolones and/or extended-spectrum cephalosporins and cephamycins among Escherichia coli isolates from animals and humans. J Infect Dis. 2003;188:759-68.

(38.) Wang M, Sahm DF, Jacoby GA, Hooper DC. Emerging plasmid-mediated quinolone resistance associated with the qnr gene in Klebsiella pneumoniae clinical isolates in the United States. Antimicrob Agents Chemother. 2004;48:1295-9.

(39.) Bou G, Cartelle M, Tomas M, Canle D, Molina F, Moure R, et al. Identification and broad dissemination of the CTX-M-14 [beta]-1actamase in different Escherichia coli strains in the northwest area of Spain. J Clin Microbiol. 2002;40:4030-6.

(40.) Poirel L, Decousser JW, Nordmann P. Insertion sequence ISEcp1B is involved in expression and mobilization of a bla (CTX-M) [beta]-lactamase gene. Antimicrob Agents Chemother. 2003;47:2938-45.

Catherine Branger, * [dagger] Oana Zamfir, ([dagger]) Sabine Geoffroy, * Genevieve Laurans, ([double dagger]) Guillaume Arlet, ([section]) Hoang Vu Thien, ([paragraph]) Stephanie Gouriou, (#) Bertrand Picard, (#) and Erick Denamur ([dagger])

* Hopital Louis Mourier, AP-HP, Colombes, France; ([dagger]) INSERM E0339, Paris, France; ([double dagger]) Hopital Nord, Amiens, France; ([section]) Hopital Tenon, Paris, France; ([paragraph]) Hopital Trousseau, Paris, France; and (#) Hopital Morvan, Brest, France

Address for correspondence: Catherine Branger, Laboratoire de Microbiologie, Hopital Louis Mourier, 178 Rue des Renouillers, Colombes 92701 Cedex, France; fax: [phone]; email: [e-mail]

Dr. Branger is a medical microbiologist at the Hopital Louis Mourier, a teaching hospital in Paris, France. Her main research interests are nosocomial infections and molecular epidemiology of emerging pathogens.