Molecular investigation of tRNA genes integrity and its relation to pathogenicity islands in Shiga toxin-producing Escherichia coli ( STEC ) strains

tRNA genes are known target sites for the integration of pathogenicity islands (PAI) and other genetic elements, such as bacteriophages, into bacterial genome. In most STEC (Shiga toxin-producing Escherichia coli), the PAI called LEE (locus of enterocyte effacement) is related to bacterial virulence and is mostly associated to the tRNA genes selC and pheU. In this work, we first investigated the relationship of LEE with tRNA genes selC and pheU in 43 STEC strains. We found that 28 strains (65%) had a disrupted selC and/or pheU. Three of these strains (637/1, 650/5 and 654/3) were chosen to be submitted to a RAPD-PCR technique modified by the introduction of specific primers (corresponding to the 5’end of genes selC and pheU) into the reaction, which we called “anchored RAPD-PCR”. The PCR fragments obtained were transferred onto membranes, and those fragments which hybridized to selC and pheU probes were isolated. One of these fragments from strain 637/1 was partially sequenced. An 85-nucleotide sequence was found to be similar to the cfxA2 gene that encodes a beta-lactamase and is part of transposon Tn4555, a pathogenicity island originally integrated into the Bacteroides genome.

Pathogenicity islands (PAI) are extensive clusters of virulence genes present in pathogenic bacteria, which are horizontally transferred among bacterial species and are acquired as plasmids, transposons and bacteriophages (Carniel et al., 1996;Waldor and Mekalanos, 1996).This fact has an enormous importance in bacterial evolution, since it may transform a non-pathogenic strain into a pathogenic form in a single event.PAI are found in pathogenic strains, but rarely in non-pathogenic ones (Hacker et al., 1997).These genetic elements have been described in bacteria such as Escherichia coli (McDaniel et al., 1995), Helicobacter pylori (Censini et al., 1996), Salmonella spp (Mills et al., 1995), and Vibrio cholerae (Waldor and Mekalanos, 1996;Novais et al., 1999;Vicente et al., 1997).PAI usually integrate into tRNA loci in E.coli (Inouye et al., 1991), Pseudomonas spp (Hayashi et al., 1993) and Salmonella spp (Mills et al., 1995), and the disruption of these genes is a potential marker for the occurrence of PAI (Hacker et al., 1997).Shiga toxin-producing Escherichia coli (STEC) colonizes the gastrointestinal tract of bovines and other animals and is mostly transmitted to humans by contaminated undercooked ground beef.The major PAI in STEC, the etiological agent of hemolytic-uremic syndrome (HUS) is the locus of enterocyte effacement (LEE) that encodes a type III secretion system and E.coli-secreted proteins, required for the induction of attaching and effacing lesions in intestinal cells (Paton and Paton, 1998).
Two preferential LEE insertion sites were described in tRNA genes, selC (selenocysteine tRNA gene) and pheU (phenylalanine tRNA gene) (Sperandio et al., 1998).During the insertion process, PAI interrupt these genes, making them non-functional.The insertion makes tRNA/PAI too large to be amplified, and some authors considered negative PCR results as indicating the presence of PAI inserted into tRNA genes (Sperandio et al., 1998).
In this work, we analyzed the molecular integrity of these two tRNA genes and investigated its relation to PAI in LEE-positive and LEE-negative STEC strains using a RAPD-PCR technique modified by the inclusion of specific primers in the reaction (which we called "anchored RAPD-PCR").

Bacterial strains
STEC strains of different serotypes were previously isolated from healthy cattle from Rio de Janeiro State, Brazil, and were classified, by the detection of the eae gene, into LEE-positive or LEE-negative (Gonzalez et al., 2001) Specific PCR and anchored RAPD-PCR conditions Blanc-Potard and Groisman (1997) and Sperandio et al. (1998) described the primer pairs used to amplify the selC gene and the pheU gene, respectively.The cycling parameters for specific PCR were: 30 cycles, each cycle consisting of a denaturing step at 94 °C for 1 min, an annealing step at 50 °C for 1 min, and an extension step at 72 °C for 1 min.The components for 50 µL PCR reaction solution were: 100 ng of DNA template, 30 pmoles of each primer, 10 mM Tris-HCL (pH8.3),50 mM KCL, 3 mM MgCl 2 , 0.1 mM of each dNTP and 2 units of Taq polymerase (Invitrogen, Carlsbad, CA, USA).Negative controls were included in each experiment.For the anchored RAPD-PCR, the following conditions were used: one cycle of 5 min each at 94 °C, 32 °C, and 72 °C, respectively; then, one cycle of 1 min at 94 °C, 5 min at 32 °C, and 5 min at 72 °C, and finally 43 cycles of 1 min at 94 °C, 1 min at 32 °C, and 2 min at 72 °C each.12 pmoles of random primers (r1 or r2) and 30 pmoles of specific primers (for the 5'end selC or the 5'end pheU) were included in each reaction.The components were the same used in the specific reactions: primer r1: 5' GGGTAACGCC 3' and r2: 5'AGAG GGCACA 3'; primer cfxA1: 5' TAACATAACCTGAACC TGTC and primer cfxA2: 5' TCAGATAGCTTATACG GAAG 3'.

DNA extraction, DNA restriction, electrophoresis conditions and Southern blotting
Genomic DNA was extracted with TRIZOL reagent (Invitrogen), according to the manufacturer's instructions.DNA fragments were extracted from the gel by the use of the Gene Clean kit (Bio 101 Inc). 10 µg of DNA were digested with 10 U of EcoRI enzyme (Invitrogen).Digested DNA was submitted to electrophoresis (100 V, 2 h) in 0.8% (w/v) on agarose gel immersed in TBE buffer (90 mM Tris-borate, 2 mM EDTA, pH 8.0), and transferred onto nylon membranes, according to the Southern blotting method (Sambrok et al., 1989).

Hybridization to radioactive probes
DNA fragments including selC and pheU genes were labeled with α-32 P dCTP using the random primer labeling Kit (Amersham Biosciences).Nylon membranes were hybridized to radioactive probes at 42 °C in the presence of 6x SSC, 0.7% SDS and 50% formamide, and the filters were washed twice with 0.3x SSC and 0.1% SDS at 42 °C for 30 min.After hybridization, the filters were exposed to X-OMAT (Kodak) films for 24 h and developed.

Sequencing
The reactions were carried out according to the manufacturer's procedures included in the ABI PRISM TM Dye terminator cycle Sequencing ready reaction kit (Perkin Elmer) and run in 6% acrylamide/urea gel at 60 W, in an ABI 373 automated sequencer (Applied Biosystems, Inc.) Our results show (Table 1) that 18 (82%) out of 22 LEE-positive strains had a disrupted selC and/or pheU.Ten (47%) out of 21 LEE-negative strains had one or both genes disrupted.We also found LEE-positive strains with both genes intact, suggesting that LEE was integrated somewhere else in the genome.These findings will be investigated further.
We submitted 14 LEE-positive and negative strains to hybridization to a selC probe, after digestion with EcoRI enzyme.There is a single copy of the selC gene in the bacteria genome (Hou, 1999), and has no EcoRI sites.After hybridization, all strains which were PCR-negative for selC (lanes 11, 13, 14, and 15) showed two bands (Figure 1).The selC + strains (lanes 2, 4, 6, 7, 8, 10, and 16) showed a single band, suggesting that, in these strains, the selC gene is intact, whereas it is disrupted in the others (Figure 1, lanes 11,13,14, and 15).Strains 231/3 and 234/1 will be investigated further, once they are LEE-negative, but have a disrupted selC gene.
In order to investigate the genetic elements integrated into selC and pheU genes, in strains with one or both of these genes disrupted, we conceived a modification of the RAPD-PCR technique.Specific primers which anneal to the 5' end of genes selC and pheU were included, along with the random primers, in the RAPD-PCR technique herein called "anchored RAPD-PCR".This approach allows the amplification of portions of the disrupted tRNA genes.Amplification products (Figure 2) submitted to hybridization confirmed this assertion (Figures 3 and 4).We were able to identify these fragments following hybridization to specific pheU (Figure 3) and selC (Figure 4) probes.Three strains [650/5 (eae + ,selC -/pheU -), 637/1(eae -, selC + /pheU -), and 654/3 (eae -,selC + /pheU -)] were initially chosen to be submitted to the modified RAPD-PCR technique, once 637/1 and 654/3 are LEE-negative but disrupted at the pheU gene, suggesting that a PAI or another genetic element might be inserted at this locus.Strain 650/5 is LEE-positive but has both genes (selC and pheU) disrupted, and it is possible that, besides LEE, another PAI might be inserted at one of these loci.
Some fragments hybridized to the pheU and selC probes (Figures 3 and 4).We chose one of these fragments (a 0.75 Kb fragment from strain 637/1, Figure 3) to be isolated and sequenced first.A short preliminary 85nucleotide sequence (accession number AY191267) was compared to homologous sequences at Genbank, and the alignment showed a similarity of 89% with gene cfxA2 (data not shown).cfxA2 is a beta-lactamase gene included in transposon Tn4555 (Smith and Parker, 1998), largely distributed among bacterial species and found very often in Bacteroides genera (Tribble et al., 1999).We were also able to amplify a 250 pb fragment from strain 637/1 by using cfxA1 and cfxA2 primers which amplify part of the cfxA2 gene (data not shown).Besides Bacteroides genera, the sequence found in strain 637/1 is similar to cfxA2 sequences also found in six species of Prevotella (P.oralis, P. malaninogenica, P. intermedia, P. denticola, P. buccae, P. bivia), both genera commonly found colonizing the bovine gastrointestinal tract.Transposon Tn4555 is in fact a pathogenicity island [Morschhäuser et al. (2000) define similar DNA elements, which have genes responsible for antibiotic resistance, as PAI] that is not self-transmissible, but can be transferred between species via conjugation, in the presence of a helper element that supplies most of the necessary conjugation functions (Smith and Parker, 1998).We speculate that transposon Tn4555, which carries the cfxA2 gene, might be transposed to STEC 637/1 in the bovine intestine, once strain 637/1 was originally isolated from cattle feces.Mizan et al. (2002) showed that cattle rumen is a favorable environment for the genetic exchange of plasmids between the indigenous micro biota and resident STEC O157:H7 in the bovine host.
In summary, our results point to selC and pheU genes as preferential target sites for the integration of LEE into the STEC genome.Eighteen out of 22 LEE-positive strains had either selC or pheU or both genes disrupted, suggesting that LEE or another genetic element is inserted into one of these loci.Results of hybridization to a selC probe allow us to suggest that the selC gene may be disrupted in selC-negative strains.The use of anchored RAPD-PCR and hybridization led to the identification of fragments which include genes selC and pheU and an adjacent region.Sequencing of only 85 nucleotides was sufficient for identifying the nature of a sequence transposed to pheU locus in strain 637/1.The results obtained showed that transposon Tn4555, which is in fact a pathogenicity island, first described to belong to the genera Bacteroides, had been transposed to the STEC genome.