|Home | About | Journals | Submit | Contact Us | Français|
Urinary tract infections are one of the most frequent bacterial diseases in humans, and Escherichia coli is most often the relevant pathogen. A specific pathotype of E. coli, known as uropathogenic E. coli (UPEC), often causes serious and difficult-to-treat infections of the urinary tract. We propose a new single-tube screening tool that uses an (N)6(CGG)4 primer to generate fingerprint profiles that allow rapid discrimination and epidemiology of this group of bacteria. We found 71 different CGG-PCR profiles among 127 E. coli strains, while enterobacterial repetitive intergenic consensus (ERIC)-PCR of the same group yielded only 28 profiles. Additionally, the (CGG)4-based PCR test turned out to be very effective for clustering UPEC strains exhibiting multiple virulence genes and usually belonging to the B2 phylogenetic group, and it separated these strains from E. coli strains lacking most of the UPEC-specific virulence factors. Since the reproducibility of the CGG-PCR screen is higher than that of ERIC-PCR, our test should be a valuable means of increasing the discriminatory power of current UPEC typing schemes.
Gram-negative rods are the major etiological agents in urinary tract infections (UTIs) in humans, and Escherichia coli comprises most of these agents (20, 30, 32, 34, 38, 42). In some cases, UTI treatment is difficult because of persistent recurrences. Furthermore, UTIs are often asymptomatic at the beginning of the infection process. Particular phenotypic features of uropathogenic E. coli (UPEC) strains facilitate their persistence in urinary tracts and differentiate them from the other pathogenic and commensal E. coli strains (7, 29, 31). UPEC-specific virulence factors (VFs), which are mostly adhesins (P and S fimbriae), toxins (cytotoxic necrotizing factor type 1, α-hemolysin), bacteriocin (uropathogenic-specific protein), and siderophores (aerobactin and yersiniabactin), are important for colonization of the urinary tract (7, 8, 27). Also, type 1 fimbriae and afimbrial adhesin I are beneficial in this type of infection. Additionally, phylogenetic analyses have revealed that UPEC strains differ substantially from other E. coli strains (2, 10, 43). Pathogenic E. coli strains, including UPEC strains, belong mainly to groups B2 and D (2, 5, 14).
In the case of E. coli, 16S rRNA gene sequence analysis, phylogenetic studies, and VF profiles are valuable for detailed genetic identification (4, 5, 11, 35). PCR-based methods are very efficient, inexpensive, and rapid (44). Previously, two distinct prokaryotic repetitive elements were used for gram-negative enterobacterial strain discrimination: repetitive extragenic palindromic (REP) elements and enterobacterial repetitive intergenic consensus (ERIC) sequences (16, 37, 40). Because the ERIC-PCR band patterns were less complex than the REP-PCR band patterns, differences within the analyzed species were easier to distinguish with ERIC-PCR.
The goals of this work were to develop a novel genetic test (termed CGG-PCR) for the differentiation and epidemiological investigation of UPEC strains and to compare it to ERIC-PCR. In the comparison, the following factors were taken into consideration: cluster cutoff values, clustering capability with regard to virulence profiles, phylogenetic groups and quinolone susceptibility, reproducibility of band patterns, number of different profiles, and discriminatory indices. Our assay is based on the presence in bacterial genomes of microsatellites-trinucleotide repeat sequences. Since the different trinucleotide repeat sequence elements vary in copy number and distribution in bacterial genomes, they have the potential to serve as valuable markers for phylogenetic and epidemiological studies. A (CGG)5 hybridization probe has been successfully used in conjunction with restriction fragment length polymorphism to type Mycobacterium tuberculosis (28). However, such hybridization techniques require the isolation of large amounts of genomic DNA and are time-consuming and expensive. Also, (GTG)5-PCR was tested for its ability to track the origins of E. coli, Lactobacillus spp., and Enterococcus spp. isolated from various sources (12, 24, 25, 39). We propose an improved PCR methodology that employs an N6(CGG)4 primer with a high annealing temperature. Trinucleotide repeats are present on both DNA strands, enabling us to design a single PCR primer harboring the CGG motif that yields characteristic electrophoretic CGG-PCR band patterns. Considering the high reproducibility and specificity of the CGG-PCR profiles, this test has potential both as an alternative to and as an additional screening tool for the rapid and efficient genotyping of E. coli strains.
A collection of 127 clinical E. coli strains was assembled between June 2005 and September 2006 from the urine of patients (75% women) in various wards of Military Teaching Hospital No. 2, Medical University of Lodz, Poland. The urine samples were characterized by the presence of >104 CFU/ml of bacteria. E. coli strains were identified on the basis of their differential growth on CPS3 medium (Biomerieux). For further studies, all strains were grown at 37°C on LB plates. Each isolate was screened for susceptibility (by the disk diffusion method) to amoxicillin (amoxicilline), amoxicillin-clavulanic acid, piperacillin, cephalothin (cefalotin), cefoxitin, cefotaxime, ceftazidime, imipenem, tobramycin, amikacin, gentamicin, netilmicin, nalidixic acid, norfloxacin, ciprofloxacin, nitrofurantoin, co-trimoxazole, and fosfomycin.
Oligonucleotide primers were synthesized at the Institute of Biochemistry and Biophysics PAS and used for PCR amplification of the appropriate genes/regions. Sequences, locations, and predicted sizes of amplification products for the specific oligonucleotide primers used in this study are shown in Table Table11.
DNA for amplification was released from whole cells by boiling. Single colonies were harvested from the LB plates, suspended in 200 μl of sterile water, incubated at 100°C for 5 min, and centrifuged. The supernatant was used in all PCRs as described below.
A triplex-PCR technique was used to detect phylogroups of E. coli. Amplification of bacterial DNA fragments was performed in a volume of 50 μl containing 5 μl of 10× PCR buffer, 20 pmol of each primer (TspE4C1, TspE4C2, YjA1, YjA2, ChuA1, and ChuA2 [Table [Table1]),1]), 1.5 mM MgCl2, 0.2 mM of each deoxynucleoside triphosphate (dATP, dGTP, dCTP, and dTTP), 3 μl of chromosomal DNA solution, and 1 unit of Taq polymerase (Invitrogen). The DNA concentration range was about 20 to 50 ng of genomic DNA per microliter. Cycling conditions were as follows: denaturation at 95°C for 5 min, followed by amplification for 30 cycles of 95°C for 1 min, 55°C for 1 min, and 72°C for 2 min, followed by a final extension at 72°C for 7 min. The PCRs were performed in a T3000 Biometra thermal cycler. PCR products were separated on 1.5% agarose gels in 1× Tris-acetate-EDTA (TAE) buffer. In the case of strains belonging to phylogroup A, the amplification of the E. coli uidA housekeeping locus was performed according to a procedure described in detail elsewhere (http://www.shigatox.net/cgi-bin/mlst7/index).
A multiplex PCR technique was used to detect fragments of genes encoding selected E. coli VFs. The components of the reaction mixture were the same as mentioned above, except for (i) 5 pmol each of primers pap1, pap2, sfa1, sfa2, cnf1a, cnf2a, usp1mod, usp2mod, hly1mod, and hly2mod (Table (Table1),1), (ii) 2 mM MgCl2, and (iii) 1 μl dimethylsulfoxide. Cycling conditions were as follows: denaturation at 95°C for 3 min, followed by amplification for 35 cycles of 95°C for 1 min, 60°C for 1.5 min, and 72°C for 3 min and a final extension step at 72°C for 8 min. The PCRs were performed in a T3000 Biometra thermal cycler. PCR products were separated on 1.5% agarose gels in 1× TAE buffer.
An amplification reaction was performed in a volume of 50 μl containing 5 μl of 10× PCR buffer, 100 pmol of the (N)6(CGG)4 primer, 1.5 mM MgCl2, 0.2 mM of each deoxynucleoside triphosphate (dATP, dGTP, dCTP, and dTTP), 3 μl dimethylsulfoxide, 3 μl of chromosomal DNA solution, and 1 unit of Taq polymerase (Invitrogen). The DNA concentration range was about 20 to 50 ng of genomic DNA per microliter. Cycling conditions were as follows: denaturation at 95°C for 3 min, followed by amplification for 35 cycles of 95°C for 1 min and 72°C for 3 min, followed by a final extension step at 72°C for 8 min. The PCRs were performed in a T3000 Biometra thermal cycler. Gels were optimized according to recommendations provided with the BioNumerics software. CGG-PCR products (ranging from 0.1 kbp to 2.5 kbp) were electrophoresed (8 μl of 50 μl) on 2% agarose gels (15 × 15 cm, 4 mm thick) in 1× TAE buffer (2.5 V/cm) until the dye (bromophenol blue) migrated 6 cm from the beginning of the gel.
The components of the reaction mixture were the same as in the N6(CGG)4 primer-based PCR protocol, except for the addition of 10 pmol of each ERIC primer (Table (Table1).1). The DNA concentration range was about 20 to 50 ng of genomic DNA per microliter. Cycling conditions were as follows: denaturation at 95°C for 3 min, followed by amplification for 35 cycles of 95°C for 0.5 min, 51.2°C for 1 min and 72°C for 2 min, followed by a final extension step at 72°C for 5 min. The PCRs were performed in a T3000 Biometra thermal cycler. Gels were optimized as described above.
All gels were stained with ethidium bromide (1 μg/ml). Amplified DNA fragments of specific sizes were visualized with a UV transilluminator and photographed.
Agarose gels were normalized with a 100-bp DNA size marker (Perfect 100 bp DNA ladder; EurX) containing 13 fragments that range in size from 100 to 1,000 bp in 100-bp increments plus additional fragments at 1.5 kbp, 2 kbp, and 2.5 kbp. Clustering analyses were performed with BioNumerics software (Applied Maths), and dendrograms were calculated with the unweighted-pair group method using average linkages Pearson coefficient. The cluster cutoff values for CGG-based PCR and ERIC-PCR were set automatically by BioNumerics software at 54.79% and 39.85% similarity, respectively. The discriminatory index described by Hunter and Gaston (17) was used as a numerical index for the discriminatory power of both typing methods. The reproducibility of both CGG-PCR and ERIC-PCR was assessed by comparison of nine separate band pattern-based densitometric curves characteristic of four selected strains. The clustering coefficient based on densitometric curve characteristics was chosen to avoid problems with appearance and determination of minor bands.
We applied a commonly used technique (triplex-PCR) (5) to determine the distribution of phylogenetic groups among our collection of E. coli strains. A correlation between phylogroups and virulence in extraintestinal pathogenic E. coli has recently been demonstrated, and drug resistance is also associated with this phenomenon. Recent data have proven that quinolone-resistant extraintestinal pathogenic E. coli strains are less able to cause upper urinary tract infections and have fewer VFs than quinolone-susceptible E. coli strains (15, 26, 36). In our analysis, 62 strains were classified within group B2, 19 were within group D, 17 were within group B1, and 29 belonged to group A. In the Clermont method (5), group A strains may or may not produce PCR products (13). Additionally, the majority of these strains were VF negative and no E. coli-specific amplicon was produced. Therefore, the amplification of the uidA housekeeping locus in all group A strains confirmed that all these strains were in fact E. coli. We also assessed the distribution of six VFs, including those specific for UPEC strains. Multiplex-PCR let us identify fragments of genes encoding S fimbriae (sfaD/sfaE), P fimbriae (papC), α-hemolysin (hlyA), cytotoxic necrotizing factor 1 (cnf1), and uropathogenic-specific protein (usp), as well as the fimG/fimH region encoding subunits of type 1 fimbriae commonly present in other E. coli pathotypes (7, 21, 22, 27). We identified two major groups of E. coli in our collection of strains. Of the 127 strains analyzed, 60 possessed only type 1 fimbriae. This gene, however, was present in almost all of the investigated strains, except for five isolates that did not encode any of the VFs mentioned above. This group of strains may represent a population of either avirulent or asymptomatic bacteriuria (ABU)-causing E. coli, which lack a number of VFs (45). Zdziarski et al. (45) have defined these ABU strains as commensal-like E. coli, which may establish bacteriuria without provoking the host response. Additionally, these strains were frequently resistant to quinolones and also other antibiotics (data not shown) that may hinder their eradication from the urinary tract. In contrast, the second group encoded many VFs and belonged mostly to phylogenetic group B2. Among these, 30 strains possessed four or five of the studied UPEC-related genes (excluding fimG/fimH) and belonged to the phylogroup B2. These results implicate these strains as pathogenic or members of the ABU group of E. coli marked by the presence of special but not functional virulence determinants (2, 5, 14, 45). The resulting virulence profiles and phylogroups of the E. coli strains in our collection are presented in Table Table22 .
Next, we decided to fingerprint our collection of clinical E. coli strains using the N6(CGG)4-based PCR and ERIC-PCR techniques. The CGG-based primer used for our PCR method demonstrated a high annealing temperature (72°C), which facilitated DNA template denaturation and restricted nonspecific primer binding, which was likely to be essential for obtaining reproducible PCR profiles. First, a thorough investigation of the reproducibility of the fingerprints generated by both methodologies was carried out. In each test, nine separate PCR analyses were performed on four selected UPEC strains. Amplifications were repeated with the same strain on different days. The reproducibilities of the CGG-PCR band patterns of the strains UPEC 68, UPEC 7, UPEC 6, and UPEC 10 were 94.3%, 91.0%, 96.8%, and 92.6%, respectively. The reproducibilities of the ERIC-PCR band patterns of the same strains were 86.4%, 65.5%, 78.2%, and 78.4%, respectively (Fig. (Fig.1).1). Based on this, we estimated the overall reproducibility value for ERIC-PCR to be 77.1% ± 8.6%, while this value for CGG-PCR was 93.7% ± 2.5%. Therefore, the similarity values of 68.5% (77.1% − 8.6% in the case of ERIC-PCR) and 91.2% (93.7% − 2.5% in the case of CGG-PCR) were set as thresholds below which the evaluated fingerprint profiles were declared to be different. The low reproducibility of ERIC-PCR has been reported previously (18, 19). Significantly improved ERIC-PCR reproducibility (approximately 80%) was achieved with an elevated annealing temperature (65°C), possibly due to more-specific primer binding (19). The other REP-PCR methodologies yielded even lower levels of reproducibility under these conditions. Hence, in the case of our analyses, the relatively modest inconsistencies between the repeated PCRs proved that CGG-PCR is a valid means of determining the genetic similarity of UPEC isolates.
The collection of clinical E. coli strains was PCR genotyped with the N6(CGG)4 primer. The similarity of the fingerprint profiles was analyzed, and five clusters of E. coli were distinguished (Fig. (Fig.2).2). Interestingly, clusters I, II, and III differed substantially from clusters IV and V both phylogenetically and in their VF profiles. Type I fimbriae were not taken into consideration in our analyses, as they are ubiquitous in E. coli. Strains within the first three clusters contained no or very few VFs. A full 82.6% of these strains contained no VFs, and 17.4% of them contained up to three factors. Also, the majority of these strains belonged to the phylogenetic groups A and B1 (63.7%), and most were simultaneously resistant to three quinolones (nalidixic acid, norfloxacin, and ciprofloxacin). In contrast, 63.8% of the strains from clusters IV and V encoded at least three VFs. The strains from clusters IV and V were mostly in phylogroups B2 and D (89.7% and 6.9%, respectively). There were no strains from group B1, and only 3.4% belonged to group A. Drug resistance analyses showed that strains from these two clusters were generally sensitive to quinolone antibiotics, which was in accordance with their tendency toward high virulence and quinolone susceptibility (9, 15, 26, 36). In sum, using the (CGG)4-containing primer in the PCRs, we were able to distinguish two major subgroups of UPEC strains (Fig. (Fig.3).3). These included UPEC group I, consisting of strains that lack most of the examined VFs and belong to different phylogenetic groups. The other subgroup was UPEC group II, which consists of strains that harbor many selected VFs. Group II strains belong primarily to phylogroup B2. Furthermore, our genetic approach allowed deep discrimination of E. coli strains isolated from urine samples of patients. Taking into account the reproducibility of N6(CGG)4-based PCR (91.2%; Fig. Fig.2),2), our test was able to identify 71 different profiles: 37 in the group I UPEC strains lacking most of the UPEC-specific VFs (clusters I, II, and III) and 34 in the group II UPEC strains exhibiting multiple virulence genes (clusters IV and V).
The ERIC-PCR technique was previously described as a valuable tool for the genotyping of enteric bacteria (40). The discriminative power of the REP-PCR group, including ERIC-PCR, has been determined to be sufficient for small-scale epidemiological studies involving E. coli strains. It possesses a higher discriminatory ability for many microorganisms than do other quick-typing techniques, leading to its increased frequency of use (1, 6, 33, 41). However, in the literature, the conclusion that ERIC-PCR possesses high discriminatory power is independent of any examination of reproducibility, which significantly impacts the determination of strain differences (19, 23). Taking into consideration the quite low reproducibility of ERIC-PCR when used on our collection of UPEC strains (68.5%; Fig. Fig.4),4), it was possible to distinguish only 28 different profiles, and a discriminatory index of 0.92 was achieved. In contrast, CGG-PCR provided (i) a higher discrimination index of 0.985, (ii) a greater number of unique fingerprints, and (iii) substantially higher reproducibility (Table (Table3).3). The ERIC-PCR-based dendrogram analysis yielded five clusters, while three strains were not clustered. The calculated cluster cutoff values were also different between the two fingerprint analyses (54.79% with CGG-PCR versus 39.85% with ERIC-PCR), consistent with the higher discriminatory power of our test. Focusing on ERIC-PCR, several strains that belong to phylogenetic group B2 and that are characterized by high pathogenicity potential were clearly separated as a subgroup inside the extensive cluster I. Strains of that kind also comprised a cluster, cluster II (except one strain from phylogroup A lacking examined VFs). The other clusters were composed of a mix of phylogenetically different strains with a wide variety of VFs. Interestingly, modified ERIC-PCR (65°C) has been shown to be able to distinguish the B2 phylogenetic group from other E. coli strains (19). Nevertheless, it is important to consider the occurrence of B2 E. coli strains lacking examined VFs. ERIC-PCR band profiles may be sufficient for epidemiological investigations. However, unlike CGG-PCR, ERIC-PCR did not yield noticeable associations among the band patterns, the phylogenetic groups, and the pathogenicity profiles of the strains.
In sum, PCR analyses using the N6(CGG)4 primer proved to be more reproducible than ERIC-PCR, likely because of the higher annealing temperature of the N6(CGG)4 oligonucleotide primer. This feature can be a critical factor in PCR-based strain genotyping. Although both methods were capable of genotyping E. coli strains, reproducibility analyses gave a more accurate picture of their relative abilities to discriminate UPEC strains. In our opinion, CGG-PCR can be successfully used as a powerful alternative screening tool for genotyping E. coli strains that cause urinary tract infections and possibly also for epidemiological investigations. Furthermore, N6(CGG)4-based PCR might have practical applications in microbiology, as it could be used to detect associations between CGG-PCR fingerprints and specific bacterial properties such as phylogroups, virulence profiles, and quinolone susceptibility.
This work was partially supported by the Ministry of Science and Higher Education grant number N404 097 32/3354. Publication is supported by the European Social Fund and Budget of State implemented under the Integrated Regional Operational Programme, GRRI-D project.
Published ahead of print on 21 October 2009.