| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Molecular Diagnostics and Genetics |
1 Laboratorio de Ecología Molecular, Centro de Astrobiología (INTA-CSIC), Madrid, Spain.
2 Laboratorio de Referencia de E. coli (LREC), Departamento de Microbiología y Parasitología, Facultad de Veterinaria, Universidad de Santiago de Compostela, Lugo, Spain.
aAddress correspondence to this author at: Laboratorio de Ecología Molecular, Centro de Astrobiología (INTA-CSIC), carretera de Ajalvir km 4, 28850, Torrejón de Ardoz, Madrid, Spain. Fax 34-915201074; e-mail parrogv{at}inta.es.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Methods: Type-specific oligonucleotide microarrays and a PCR scheme were designed and constructed for the detection and typing of genetic variants of the LEE genes. Oligonucleotide probes were tested for their specificity against the corresponding type strain by microarray hybridization using fluorescent DNA, either PCR-amplified (single, multiplex, long-range), chromosomal, or amplified chromosomal DNA.
Results: The PCR scheme and the oligonucleotide microarray allowed us to distinguish 16 variants (
1, 2, ß1, ß2, 
1, 2/
, 
/
, 
, 
, 
, 
, 
, µ, 
, 
, o) of the eae gene, 4 variants (1, ß1, 1, 2/) of the tir gene, 4 variants (1, ß1, ß2, 1) of the espA gene, 3 variants (1, ß1, 1) of the espB gene, and 3 variants (1, ß1, 1) of the espD gene. We found a total of 12 different combinations of tir, espA, espB, and espD genes among the 25 typed strains.
Conclusions: The PCR scheme and the oligonucleotide microarray described are effective tools to rapidly screen multiple virulence genes and their variants in E. coli strains isolated from human and animal infections. The results demonstrate the great genetic diversity among LEE genes of human and animal STEC and EPEC strains.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Differences in length, insertion sites, or nucleotide sequences have been found in the LEE region. Moreover, specific variants of intimin are related to STEC strains pathogenic for humans, whereas other intimin variants are related to human or animal EPEC strains (19)(31)(32). The N-terminal end of intimin is well conserved, whereas the C-terminal end is highly variable and is responsible for receptor binding. It has been suggested that differences in host tissue cell tropism may be attributable to different intimin types (33)(34). For example, intimin 1 from STEC O157:H7 seems to restrict colonization to Peyer’s patches of the human intestine (35). Adu-Bobie et al. (19) found that antigenic variation exists within the cell-binding domain of intimin types expressed by different clinical human EPEC and STEC isolates. Using specific oligonucleotide primers complementary to the 3'-end eae genes, they defined 4 intimin types (, ß, , and ) by type-specific PCR (TS-PCR) assays. In another molecular study, Oswald et al. (31) described a TS-PCR assay that identifies a fifth intimin variant (intimin ). They divided the intimin alleles , ß, and based on PCR–restriction fragment length polymorphism profiles into 1, 2, 1, and 2 subtypes. Tarr and Whittam (36) described 2 new types ( and ) in a work on the molecular evolution of intimin genes in human STEC and EPEC O111 clones. Recently, Zhang et al. (32) determined the sequences of 3 new intimin variant genes (, , and ) found in human STEC strains. They recommended TS-PCR protocols for detection of these types of eae genes and performed a phylogenetic analysis. The sequence of the 3' variable region of eae gene of a new intimin () has been submitted to GenBank by B. China (GenBank accession no. AF439538), and we sequenced the whole intimin gene from a human strain (GenBank accession no. AJ715409). Blanco et al. [Ref. (11) and unpublished results] identified 3 new intimin genes in human EPEC strains (intimin ß2, µ, and ) and 1 in bovine STEC strains (intimin ). Like Zhang et al. (32), we have observed that the sequences of the eae-2 gene defined by Oswald et al. (31) and the eae- gene described by Tarr and Whittam (36) are almost identical (99%). These 2 sequences should be considered as a single eae variant (2/). We have also observed that the eae- gene recently described by Zhang et al. (32) has a sequence very similar (99%) to that of the eae- gene defined by Adu-Bobie et al. (19); therefore, these 2 sequences should be considered a single eae variant (/). Furthermore, because of the high degree of sequence identity, specific primers could not be designed for distinguishing eae-2 and eae- genes or for differentiating eae- and eae- genes. In addition to the eae gene, variants in the tir (, ß, , and ), espA (, ß, and ), espB (, ß, and ), and espD (, ß, ) genes have also been described, whereas the esc and sep genes are more conserved (29)(31)(37)(38). The nomenclature for tir, espA, espB, and espD genes follows the same principle as that for eae, as described China et al. (37). Currently, only associations between the tir, espA, espB, and espD genes with 7 different types of intimin (1, ß1, 1, 2/, , , and ) have been studied in EPEC and STEC strains.
Thus, differentiation of eae, tir, and esp alleles is an important tool for STEC and EPEC typing in routine diagnostics as well as in epidemiologic and clonal studies. The fact that the 5' regions of eae genes are conserved whereas the 3' regions are heterogeneous has permitted the design of universal and allele-specific PCR primers to differentiate among 16 variants of the eae gene encoding 16 different intimin types and subtypes: 1, 2, ß1, ß2, 1, 2/, /, , , , , , µ, , , and o [Refs. (4)(19)(31)(32)(39) and this study]. Genetic diversity was also found in other LEE genes (29)(37)(38). In this work, we describe 4 variants (1, ß1, 1, and 2) of the tir gene, 4 variants (1, ß1, ß2, and 1) of the espA gene, 3 variants (1, ß1, and 1) of the espB gene, and 3 variants (1, ß1, 1) of the espD gene. We also establish a total of 12 different combinations or associations between the tir, espA, espB, and espD genes.
Although molecular methods such as PCR, multiplex PCR, and real-time PCR have been used to detect and identify pathogenic E. coli strains, an important limitation is the existence of a large number of virulence factors as well as a high allelic variation (4)(37)(40)(41). This makes PCR assessment of any single isolate for all known virulence genes and variants very laborious. Microarray technology allows assessment of hundreds of different genes and variants from a particular strain as well as comparative analysis between strains. Several studies have been reported that used DNA microarrays as a diagnostic tool (21)(40)(41)(42)(43)(44)(45)(46)(47)(48). Bekal et al. (49) presented a DNA microarray for pathotype identification, using PCR-amplified fragment from 91 E. coli virulence genes as immobilized probes. These PCR fragments work very well when the target genes are quite different in sequence, but they can produce cross-hybridization events between highly similar gene variants.
The growing allelic variance at the LEE island requires more precision during the design of a DNA microarray. PCR fragments are not appropriate as probes because of the high degree of sequence identity between alleles of the same gene. In this work, we designed and constructed an oligonucleotide microarray to detect specifically most of the known alleles of the LEE-encoded tir, eae, espA, espD, and espB genes. Each specific probe on the microarray has been checked and validated against its corresponding type strain, with fluorescently labeled PCR-amplified DNA used as target. We also present a comparison between different methods for target labeling: single PCR, multiplex PCR, long-range PCR (LR-PCR), chromosomal DNA, and multiple displacement amplification (MDA)-amplified chromosomal DNA.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
, and they are control strains (type strains) for the 16 different intimin types and subtypes. Strains were stored at room temperature in nutrient broth containing 7.5 g/L agar. E. coli K-12 MG1655, a non-LEE pathogenicity island–containing strain, was used as negative control.
|
serotyping
We used the method described by Guinée et al. (50) to determine the O and H antigens, using all available O (O1–O185) and H (H1–H56) antisera. All antisera were obtained and adsorbed with the corresponding cross-reacting antigens to remove the nonspecific agglutinins. The O antisera were produced in the Laboratorio de Referencia de E. coli, and the H antisera were obtained from the Statens Serum Institut (Copenhagen, Denmark).
pcr primers and oligonucleotide probes
PCR primers and oligonucleotide probes for the different variants of the tir, eae, espA, espD, and espB genes were designed and synthesized (after sequence alignment, by Clustal W software, from the GenBank sequences) or taken from previous works (see Tables 2 and 3 in the Data Supplement that accompanies the online version of this article at http://www.clinchem.org/content/vol52/issue2 ). For microarrays, probes were synthesized with a 5' C6 amino linker followed by a (dT)10 tail and the specific sequence (from 18 to 25 nucleotides). Probe design was done so that the mismatched positions were located in the central region of the oligonucleotides to increase the hybrid instability with sequences from other strains. All probes were selected for similar melting point temperatures so that all assays could be performed under the same experimental conditions.
pcr typing of lee genes and detection of other virulence genes
The methodology used for PCR typing of the eae, tir, espA, espD, and espB genes and PCR detection of the stx1, stx2 and bfpA virulence genes has been described elsewhere (4)(51)(52). Nucleotide sequences and predicted sizes of the amplified products for the specific eae, tir, espA, espB, espD, stx1, stx2, and bfpA oligonucleotide primers used in this study are shown in Table 1 of the online Data Supplement.
dna preparation, pcr amplification, and labeling
Genomic DNA from the strains listed in Table 1
was extracted from overnight cultures by use of a commercial DNA preparation reagent set (Gnome; Q-BIOgene). All strains were rechecked by PCR amplification with specific oligonucleotides (see Table 1 in the online Data Supplement) and sequencing of the corresponding gene fragment (not shown). Fluorescent labeling of each gene variant shown in Table 1 in the online Data Supplement was carried out by PCR as follows: 50-µL total volume containing 1 µg of each bacterial strain DNA (see Table 1 in the online Data Supplement) as template; 3 mM MgCl2; 100 µM each of dATP, dCTP, and dGTP; 50 µM dTTP; 50 µM Cy5-dUTP (Amersham Biosciences), 200 µM each of forward and reverse primers, and 1 U of Platinum Taq DNA polymerase in its reaction buffer (Invitrogen).
For the eae-1, -2, -, -, -1, -2, -, -, -, -µ, -, -o, -, and - gene variants, the forward primer was EAE-1 and the reverse primers were EAE-A, EAE-A2R, EAE-D3R, EAE-ep2R, EAE-C1, EAE-C2, EAE-et2R, LP-7, EAE-L2R, FV373-R, EAE-N2R, IH2997f-R, EAE-RB, and EAE-Z2R, respectively. For eae-ß1 and -ß2, the forward primer was EAE-F and the reverse primers were EAE-B8R and EAE-B5R, respectively. The primer pairs TIR-I1F/TIR-I1R, TIR-II1F/TIR-II1R, TIR-III1F/TIR-III1R, and TIR-IV1F/TIR-IV1R were used for tir-1, -ß1, -1, and -2, respectively. For espA-1, -ß1, -ß2, and -1, ESPA-1 was the forward and ESPA-I1R, ESPA-II1R, ESPA-III1R, and ESPA-IV1R were the respective reverse primers; for espB-1, -ß1, and -1, ESPB-F was the forward primer and ESPB-I1R, ESPB-II1R, and ESPB-III1R were the respective reverse primers; and for espD-1, -ß1, and -1, ESPD-4F was the forward and ESPD-I1R, ESPD-II1R, and ESPD-III1R were the respective reverse primers. For all cases but eae-ß1 and -ß2, the thermocycler was programmed as follows: 1 cycle of 5 min at 95 °C; 10 cycles of 20 s at 95 °C, 30 s at 60 °C, and 1.5 min at 68 °C; 30 cycles of 20 s at 95 °C, 30 s at 58 °C, and 1.5 min + 5 s/cycle at 68 °C; 1 cycle of 10 min at 68 °C; and a hold at 4 °C. For eae-ß1 and -ß2, the annealing temperature was lowered to 55 °C during the 10 first cycles and 53 °C for the remaining 30 cycles. Once checked on agarose gels, PCR products were purified from the unincorporated nucleotides by use of PCR purification reagents from Qiagen. We checked the labeling efficiency by measuring the absorbance at 260 nm (for DNA) and 635 nm (for Cy5) in a NanoDrop ND-1000TM spectrometer (NanoDrop Technologies).
multiplex pcr, lr-pcr, and labeling
For simultaneous multiplex PCR and labeling of the tir, eae, espA, espD, and espB genes from strains FV3671 and FV3676, we used the following conditions: primer pairs EAE-F/EAEet-2R for eae- (eta) in strain FV3671; EAE-F/LP7 for eae- (iota) in strain FV3676; ESPA1/ESPAII-1R for espA-ß1 in FV3671; ESPA1/ESPAI-1R for espA-1 in FV3676; and TIR-A/TIR-B, ESPB-F/ESPBI-1R, and ESPD-4F/ESPDI-1R for the tir-1, espB-1, and espD-1 genes, respectively, in both strains. The PCR reaction volume was 50 µL, with the same buffer and components used as above; the final concentrations of each primer was 200 µM. The thermocycler was set at 1 cycle of 5 min at 95 °C; 10 cycles of 20 s at 95 °C, 30 s at 50 °C, and 1 min at 68 °C; 25 cycles of 20 s at 95 °C, 30 s at 48 °C, and 1 min + 5 s/cycle at 68 °C; 1 cycle of 10 min at 68 °C; and a hold at 4 °C. PCR products, purification, and labeling efficiency were checked as above.
Amplification and labeling of the whole tir-eae-espA-espD-espB gene cluster (
10 kbp) were carried out by LR-PCR using the universal pair of primers TIR-UP (5'-TGATTAATCATGGCAAACTGACTA-3') and ESPB-DOWN (5'-CACTGCCACAAAGAAACTCCTTC-3') in a 50-µL (total volume) reaction. The reaction contained 500 µM each of dATP, dCTP, and dGTP; 250 µM dTTP; 50 µM Cy5-dUTP; 300 µM each of the primers; and 3.5 U of the Expand-Long Template DNA polymerase (Roche) in the supplied reaction 1x buffer. The thermocycler was programmed as follows: 1 cycle of 5 min at 95 °C; 15 cycles of 20 s at 95 °C, 30 s at 55 °C, and 10 min at 68 °C; 30 cycles of 20 s at 95 °C, 30 s at 55 °C, and 10 min + 20 s/cycle at 68 °C; 1 cycle of 10 min at 68 °C; and a hold at 4 °C. PCR products, purification, and labeling efficiency were checked as above.
genomic dna labeling
Genomic DNA was fluorescently labeled as follows: DNA (60 µL at 100 ng/µL in 10 mmol/L Tris-HCl, pH 8) was fragmented by ultrasonication by immersion of an Eppendorf tube containing the sample for 10 s at maximum power in the water bath on the "cup horn" of a MISONIX XL2010 sonicator. Under these conditions, we obtained fragments 1 to 4 kbp in length (the majority around 2.5 kbp), which were used for random priming labeling with Klenow DNA polymerase. The reaction mixture containing 2 µg of denatured DNA; 120 µmol/L each of dATP, dCTP, and dGTP; 60 µmol/L dTTP; 60 µmol/L Cy5-dUTP; 50 ng/µL random hexamers; and 50 U of Klenow enzyme (New England Biolabs) in its supplied 1x reaction buffer was incubated at 37 °C for 1–2 h. The reaction was stopped by addition of 5 µL of 0.5 mol/L EDTA, and the DNA was purified and checked for labeling efficiency as above.
genomic dna amplification
For genomic DNA amplification by MDA, 1 and 10 pg of bacterial DNA were subjected to amplification with Genomiphi system (Amersham Biosciences). Usually, more than 10 µg of amplified DNA was obtained. Up to 2 µg of this DNA was sonicated for 10 s and labeled as indicated above.
microarray production, hybridization assays, and analysis
Oligonucleotide probes, 60 µmol/L in spotting buffer (TeleChem International), were printed on epoxy-activated glass slides (TeleChem International) by use of a Microgrid II arrayer (BioRobotics; Genomic Solutions). Ten minutes after spotting, the slides were collected and stored at room temperature until used, as recommended by the manufacturer. Microarrays were prehybridized with 20 µL of 5x standard saline citrate containing 1 g/L sodium dodecyl sulfate, 0.1 g/L herring sperm DNA, and 10 g/L bovine serum albumin at 42 °C for 30–40 min. Slides were washed in distilled water, submerged in isopropanol, dried by centrifugation, and hybridized with 15–20 µL of fluorescently labeled DNA. Both PCR fragments and chromosomal DNA were denatured for 5 min at 95 °C and set to hybridize on the microarray under a cover slide (HybriSlip; Gracebio) in a hybridization chamber at 55 °C for 12 h. Printing and hybridization (Hybit) buffers and hybridization chambers were from TeleChem International. Slides were washed at room temperature twice for 5 min each in 2x standard saline citrate containing 1 g/L sodium dodecyl sulfate, twice for 5 min each in 0.2x standard saline citrate containing 1 g/L sodium dodecyl sulfate, and twice for 5 min each in 0.2x standard saline citrate; slides were then dried by centrifugation and scanned in a GMS 418 Array Scanner (Affymetrix). Images were analyzed and the spot signal was quantified by Genepix software (Genomics Solutions). The spot intensities are given in arbitrary units (0–65 000) as obtained from the Genepix software.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
1, 2, ß1, ß2, 1, 2/, /, , , , , , µ, , , and o), 4 tir variants (1, ß1, 1, and 2), 4 espA variants (1, ß1, ß2, and 1), 3 espB variants (1, ß1, and 1), and 3 variants (1, ß1, and 1) of the espD gene (Fig. 1
; also see Table 1 in the online Data Supplement). Because of the high degree of sequence identity, specific primers could not be designed for PCR typing between eae-2 and eae- genes and between eae- and eae- genes. Because tir, espA, espB, and espD sequences are available from only a limited number of strains, primers could be designed for the detection of only a small number of variants of these LEE genes. Listed in Table 1
are the 12 different combinations of tir, espA, espB, and espD genes that we have found among the 25 typed strains. In this study, we found that all strains with a specific eae type had the same combination of tir, espA, espB, and espD types. We found homologous gene combinations only in strains with eae-1, eae-ß1, and eae-1 variants, because all strains possessing these genes had the corresponding 1, ß1, and 1 subtypes of tir, espA, espB, and espD by PCR typing.
|
stec-epec microarray: a microarray for lee pathogenicity island diversity
We selected oligonucleotide probes for STEC diversity from multiple alignments of sequences of the LEE pathogenicity island genes (see Table 2 and Fig. 1 in the online Data Supplement). Probes for 16 variants of eae, 4 of tir, 4 of espA, 3 of espB, and 3 of espD were synthesized. All probes contained an aliphatic (C6) amine at the 5' end followed by a (dT)10 tail before the specific sequence, which ranged from 18 to 24 nucleotides. Two parallel subarrays were printed in quadruplicate spots: one containing both PCR fragments and specific oligonucleotides as probes (Fig. 2, A and C
), and one containing only the oligonucleotide probes (Fig. 2, B and D
). Although PCR fragments showed extensive cross-reactivity, oligonucleotides gave a high specific hybridization signal when tested against fluorescently labeled PCR-amplified fragments from the corresponding type strain. At least 2 oligonucleotide probes (usually 1 in the forward and 1 in the reverse direction) were tested for their specificity by microarray hybridization. Several sets of probes, such as those shown in Table 2 of the online Data Supplement, were printed in different sets of microarrays to check for their specificity with all strain types. Probes that failed to identify the strain for which they were designed were modified (extending or shortening the sequence) or were fully replaced by a different sequence from other locations on the gene (see Table 2 in the online Data Supplement). At least 1 probe per strain was selected that gave a high specific and intense signal with the corresponding fluorescent PCR-amplified fragment (see Fig. 1 in the online Data Supplement), whereas the remaining probes gave signals that were below the detection limit or were very close to background. Some cross-hybridization was still detected with certain probes, such probe IH2489a-F, which apart of the strong signal with its specific strain (eae-2) also gave some signal with strain eae-. Nevertheless, the difference between the specific vs nonspecific signal intensities was always more than 2-fold (note the saturated spots with eae-2), clearly indicating that IH2489a-F probe was specific for the eae-2 strain. Genomic DNA from a non–verotoxin-producing E. coli (strain MG1655) and a PCR-amplified fragment from 16S rRNA were used as negative control spots. No signal was detected in the probes when fluorescently labeled genomic DNA from MG1655 was used as target (not shown).
|
testing the stec-epec microarray for several genes
One of the great advantages of DNA microarrays is that they permit screening for up to thousands of different genes or probes simultaneously. For that purpose, samples to be analyzed must contain at least 1 putative target for each probe. Multiplex PCR can in part solve this issue by amplifying 2 to 6 different loci in 1 reaction, so that all PCR products can be checked on the microarray. The use of total chromosomal DNA as fluorescent target clearly overcomes the limitations of multiplex PCR and expands to the genomic level the number of loci to be analyzed. We tested the performance of our STEC microarray, using different methods for target preparation (Fig. 3
). As expected, PCR-based methods gave stronger and more specific signals for each probe, but they allowed the analysis of only a limited number of genes. For gene clusters such the LEE island, LR-PCR is an interesting alternative because all genes tested have the same number of target molecules, avoiding unequal signals resulting from differences in the amplification efficiencies, which can occur with multiplex PCR. The use of genomic DNA as target, even when it gave a lower signal and higher background, allowed us to simultaneously detect and identify all gene variants for a certain strain or sample (Fig. 3
; also see Fig. 2 in the online Data Supplement). Frequently, the amount of starting DNA is a limiting factor; for example, in environmental, clinical, or veterinary samples. In these cases, chromosomal DNA amplification techniques with enzymes such as phage 
29 DNA polymerase are extremely useful. We used this technique to amplify total DNA (initial amount, 10 pg) of chromosomal DNA from strains FV3671 (eae-) and FV3676 (eae-); we then fluorescently labeled 2 µg of the products (see Materials and Methods) and incubated the labeled products with the microarray (Fig. 3
). The result was similar to that obtained with chromosomal DNA without amplification. Use of both amplified and nonamplified chromosomal DNA enabled detection of all genes and the specific eae and tir variants, although it failed with espADB-specific probes. In those cases, we redesigned and synthesized new probes until good signals were obtained (see Fig. 2 in the online Data Supplement). Note that those spots containing chromosomal DNA from the negative control strain and 16S rRNA gene (see Fig. 3
, bottom left of the arrays), gave a positive signal only when we used fluorescently labeled chromosomal DNA (amplified or not) for hybridization.
|
determining pathotype with the stec-epec microarray in a single hybridization experiment
We also tested the specificity of the probes for all strain types, using chromosomal DNA (see Fig. 2 in the online Data Supplement). As mentioned above, some probes were redesigned or newly synthesized from a different location of the gene (see Table 2 in the online Data Supplement). All gene variants were simultaneously checked for each strain, permitting us to define a gene pathogenic pattern (pathotype) characteristic for each STEC strain. A sort of "barcode" can be obtained for each strain after hybridization, so that a unique barcode identifies a unique strain. Several oligonucleotides that were designed as universal primers for PCR amplification (EAE-F, TIR-A, TIR-B, or ESPA-1) failed in some hybridizations with chromosomal DNA (see Fig. 2 in the online Data Supplement), indicating that further optimization for microarrays is necessary. This is not a main issue because longer oligonucleotides or PCR-amplified fragments can be used as universal probes. The fragmentation of genomic DNA by ultrasonication to obtain a mean size of 2 to 4 kb before labeling was very reproducible (see Materials and Methods) while avoiding cumbersome procedures that use enzymatic digestions.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
1, 2, ß1, ß2, 1, 2/, /, , , , , , µ,, , and o), 4 variants of the tir gene (1, ß1, 1, and 2/), 4 variants of the espA gene (1, ß1, ß2, and 1), 3 variants of the espB gene (1, ß1, and 1), and 3 variants of the espD gene (1, ß1, 1). We also constructed an oligonucleotide microarray (STEC-EPEC microarray) for the detection and characterization of these genes.
China et al. (37) and Goffaux et al.(29) identified 3 subtypes of the tir (, ß, and ), espA (, ß, and ), espB (, ß, and ), and espD (, ß, and ) LEE genes. They found only 4 LEE profiles in human and bovine AEEC strains: eae-/tir-/espA-/espB-/espD-, eae-ß/tir-ß/espA-ß/espB-ß/espD-ß, eae-/tir-/espA-/espB-/espD-, and eae-/tir-/espA-/espB-/espD-. Nielsen and Andersen (38) identified 4 subtypes of tir (, ß, , and ), 3 subtypes for espD (, ß, and ), and a total of 7 LEE profiles (eae-/tir-/espD-, eae-ß/tir-ß/espD-ß, eae-/tir-/espD-, eae-/tir-/espD-NT, eae-/tir-/espD-, eae-/tir-ß/espD-ß, and eae-/tir-/espD-NT). In the present study, we identified 4 variants of tir and 3 variants each of espA, espB, and espD, as well as 12 combinations of these LEE genes. We found homologous combinations or associations [in the sense described by China et al. (37)] only in strains with eae-1, eae-ß1, and eae-1 genes because all strains possessing these genes had the corresponding 1, ß1, and 1 subtypes of tir, espA, espB, and espD. As in previous studies (29)(37)(38), we found that the strains with a specific eae variant showed the same combination of tir and esp genes. After sequence analysis of the amplification products, we also observed a homologous association with tir and esp genes in strains with eae-2/ and eae-o (data not shown). The nucleotide sequence of the amplification products of nontypeable tir, espA, espB, and espD genes confirmed that these strains possessed new variants of these LEE genes (data not shown).
The use of PCR-amplified fragments as probes in the microarray enabled detection and differentiation of genes with very different sequences, but use of these fragments produced a strong cross-reaction between different alleles of the same locus. It is necessary to focus the probe to a shorter and less complex region of the gene by use of synthetic oligonucleotides. The differences in nucleotide sequences among eae, tir, espA, espD, and espB gene variants were high enough to allow us to design specific probes for oligonucleotide microarrays. Some of them gave good sensitivity and specificity, but others had to be redesigned to achieve the desired results. For most of the gene variants, we designed 2 specific oligonucleotides, both of which gave good signals except if one of them was located at one end of the labeled target. In that case, the hybridized complex was not sufficiently stable and could not be detected. For best performance, a specific oligonucleotide must be located in the inner part the target sequence. Our STEC-EPEC microarray is the result of empirical tests for many probes to get the best combination in terms of specificity. Hybridization conditions such temperature, buffer composition, or target length may be modified to obtain higher signal intensities or better differentiation between strains.
Oligonucleotide microarrays allow the detection and differentiation of not only different virulence and non-virulence genes, but also multiple alleles of the same gene in a single hybridization event. For linked loci such as those of the LEE pathogenicity island, we found that use of LR-PCR for target amplification and labeling offered advantages compared with multiplex PCR in the sense that LR-PCR was more reproducible and gave smaller differences among different strains. Moreover, the possibility of having all genes in the same fragment ensures the same copy number for all of them. In addition, we showed that the use of chromosomal DNA as the hybridization target enables whole pathotyping, not only because many genes can be checked simultaneously, but also because their variants can be identified. Genomic DNA amplification techniques such MDA are suitable for use when the amount of the initial sample is limited, as happens for clinical or environmental samples. The proposed method for chromosomal DNA fragmentation by ultrasonication before labeling is rapid and reproducible, and it avoids the need for enzymatic digestions and procedures to concentrate DNA.
In conclusion, the combination of MDA, ultrasonic fragmentation of DNA, fluorescent labeling, and oligonucleotide microarray hybridization can simplify and improve the analysis of clinical, veterinary, or environmental samples.
|
Acknowledgments |
|---|
rskov and I. rskov (Copenhagen, Denmark), H. Schmidt (Wüzburg, Germany), H. Karch (Münster, Germany), E. Oswald (Toulouse, France), Y. Bertin (Clermont-Ferrand-Theix, France), and T.S. Whittam (Michigan) for generously providing strains. We also thank Monserrat Lamela for skillful technical assistance. This work was supported by grants from the Fondo de Investigación Sanitaria of Instituto de Salud Carlos III of Spanish Ministerio de Sanidad (grant FIS G03-025-COLIRED-O157), and from the Xunta de Galicia (grants PGIDIT02BTF26101PR and PGIDIT04RAG261014PR). P.G. is a fellow on grant FIS G03-025. M.M. and C.B. have a contract from Instituto Nacional de Técnica Aeroespacial (INTA), and V.P. has a "Ramón y Cajal" contract from the Spanish Ministerio de Educación y Ciencia. |
Footnotes |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
). J Clin Microbiol 2004;42:645-651., ß, , , and four intimin derivatives expressed by attaching and effacing microbial pathogens. J Clin Microbiol 1998;36:662-668.The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  Y. Ha, S. K. Ha, Y. H. Lee, D. Kim, S. Chae, C. H. Kim, K. K. Ahn, O. Kim, and C. Chae Prevalence of tir gene subtypes in Escherichia coli isolates from pigs Vet Rec., October 4, 2008; 163(14): 424 - 425. [Full Text] [PDF]
|
|
|
|
 |
|
 J. E. Afset, E. Anderssen, G. Bruant, J. Harel, L. Wieler, and K. Bergh Phylogenetic Backgrounds and Virulence Profiles of Atypical Enteropathogenic Escherichia coli Strains from a Case-Control Study Using Multilocus Sequence Typing and DNA Microarray Analysis J. Clin. Microbiol., July 1, 2008; 46(7): 2280 - 2290. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Ito, M. Iida, M. Yamazaki, K. Moriya, S. Moroishi, J. Yatsuyanagi, T. Kurazono, N. Hiruta, and O.-A. Ratchtrachenchai Intimin Types Determined by Heteroduplex Mobility Assay of Intimin Gene (eae)-Positive Escherichia coli Strains J. Clin. Microbiol., March 1, 2007; 45(3): 1038 - 1041. [Abstract] [Full Text] [PDF] |
