|Home | About | Journals | Submit | Contact Us | Français|
Conceived and designed the experiments: AS FVN. Performed the experiments: AS. Analyzed the data: AS TV. Contributed reagents/materials/analysis tools: FVN. Wrote the paper: AS. Provided important suggestions: AM DD PB FH. Supervised the whole project: AM DD PB FH.
CTX-M-producing Escherichia coli strains are regarded as major global pathogens.
The nucleotide sequence of three plasmids (pEC_B24: 73801-bp; pEC_L8: 118525-bp and pEC_L46: 144871-bp) from Escherichia coli isolates obtained from patients with urinary tract infections and one plasmid (pEC_Bactec: 92970-bp) from an Escherichia coli strain isolated from the joint of a horse with arthritis were determined. Plasmid pEC_Bactec belongs to the IncI1 group and carries two resistance genes: blaTEM-1 and blaCTX-M-15. It shares more than 90% homology with a previously published blaCTX-M-plasmid from E. coli of human origin. Plasmid pEC_B24 belongs to the IncFII group whereas plasmids pEC_L8 and pEC_L46 represent a fusion of two replicons of type FII and FIA. On the pEC_B24 backbone, two resistance genes, blaTEM-1 and blaCTX-M-15, were found. Six resistance genes, blaTEM-1, blaCTX-M-15, blaOXA-1, aac6'-lb-cr, tetA and catB4, were detected on the pEC_L8 backbone. The same antimicrobial drug resistance genes, with the exception of tetA, were also identified on the pEC_L46 backbone. Genome analysis of all 4 plasmids studied provides evidence of a seemingly frequent transposition event of the blaCTX-M-15-ISEcp1 element. This element seems to have a preferred insertion site at the tnpA gene of a blaTEM-carrying Tn3-like transposon, the latter itself being inserted by a transposition event. The IS26-composite transposon, which contains the blaOXA-1, aac6'-lb-cr and catB4 genes, was inserted into plasmids pEC_L8 and pEC_L46 by homologous recombination rather than a transposition event. Results obtained for pEC_L46 indicated that IS26 also plays an important role in structural rearrangements of the plasmid backbone and seems to facilitate the mobilisation of fragments from other plasmids.
Collectively, these data suggests that IS26 together with ISEcp1 could play a critical role in the evolution of diverse multiresistant plasmids found in clinical Enterobacteriaceae.
Most antibiotics used are naturally occurring chemical compounds produced by environmental fungi and bacteria, but also synthetic antibiotics (chemotherapeutic agents) are in circulation. In the course of evolution, bacteria have developed several strategies to deal with severe effects caused by these antimicrobial drugs. Bacterial plasmids play an important role in the horizontal transfer of antimicrobial drug resistance genes. The capability of plasmids to spread between bacterial cells by conjugation greatly enhances the dissemination of these resistance genes and raises a series of clinical problems –.
β-lactam antibiotics are one, if not the most important group of antimicrobial agents in human and veterinary medicine. The predominant cause of resistance to β-lactams in Gram-negative bacteria is the production of β-lactamases. Both chromosome- and plasmid-encoded β-lactamases have been described. Until now, more than 400 β-lactamases have been reported and new β-lactamases continue to emerge worldwide . Genes encoding extended-spectrum β-lactamases (ESBL), providing resistance to both penicillins and broad-spectrum cephalosporins but not carbapenems and β-lactamase inhibitors, are often located on conjugative plasmids , . Plasmid-mediated ESBLs are most commonly of the TEM-, SHV- or CTX-M-type. CTX-M enzymes have become the most prevalent family of ESBLs among Enterobacteriaceae since their first report in 1986. To date, more than 80 CTX-M enzymes have been isolated. They are divided into 5 clusters on the basis of the amino acid sequence: CTX-M-1, CTX-M-2, CTX-M-8, CTX-M-9 and CTX-M-25 , .
Several studies have reported CTX-M-producing Escherichia coli strains as major global human pathogens, primarily associated with urinary tract infections. Notably, clinical CTX-M-15-producing E. coli isolates have become more and more widespread , –. Some plasmids, isolated from bacteria of human origin and carrying blaCTX-M genes, have been studied in order to better understand their dissemination mechanisms –. Information on blaCTX-M-carrying plasmids from clinical isolates of animal origin is, however, lacking. Therefore, we determined the complete nucleotide sequence of a blaCTX-M-15-carrying plasmid originating from an E. coli isolated from a horse and compared this plasmid with a known plasmid from an E. coli isolate from human origin. This study also highlights the evolution of IncF plasmids by determining the complete nucleotide sequence of three CTX-M-15-encoding plasmids from E. coli isolates from humans, thereby enhancing our understanding of the pedigree of these plasmids.
Plasmid pEC_Bactec is a circular molecule of 92970-bp harbouring 86 open reading frames (ORFs) (Table S1). Conjugation experiments showed that it is transferable. pEC_Bactec belongs to the incompatibility group IncI1. pMLST assigned it to a new IncI1 pMLST type, sequence type (ST) 33 (repI3, pilL3, sogS3, ardA4, trbA-pndC15), which belongs to the clonal complex ST31 (http://pubmlst.org/perl/mlstdbnet/). The trbA-pndC region, one of the five selected alleles for pMLST, can vary in length due to the insertion of the finQ gene, encoding the fertility inhibitor, in the 5′ end of the pndC gene . In addition to this finQ insertion, the pEC_Bactec plasmid contains an extra insertion of two IS66 ORFs between finQ and pndC resulting in a longer trbA-pndC region (4939-bp) and a new allele variant trbA-pndC15 (Fig. 1). Similar IncI1 plasmids have been isolated from commensal Enterobacteriaceae of poultry and pigs . However, these plasmids were of a different pMLST type, ST34 (repI3, pilL3, sogS3, ardA4, trbA-pndC16), although they belong to the same ST31-complex (unpublished data). It should be noted that for these plasmids the trbA-pndC region lacked the extra insertion of two IS66 ORFs between finQ and pndC. All these findings may indicate that these plasmids, belonging to ST33 and ST34, are closely related and their prevalence is particularly to be linked to animal reservoirs.
The transfer region of R64, which is the reference plasmid for the IncI1 group, encoded the following genes: traABCD, pilI-V, rci, traE-K, nuc, sogL, traL-Y, ecx, pnd, trbABC and nikBA . In comparison with R64, the transfer region of pEC_bactec is well conserved, with the exception of traD and pilJ, which appeared to be lost, and the nuc gene apparently rearranged within the transfer region. The traD and pilJ genes may not be required for plasmid transfer since pEC_Bactec was able to transfer in vitro by conjugation. In addition, the cluster encoding the type IV pili (PilI-V) is considered as a virulence factor. The association with resistance determinants may favour the dissemination of plasmids belonging to this plasmid family .
In comparison with R64 , pEC_Bactec lacks the arsenic, tetracycline and streptomycine resistance genes, the addiction systems mck/kor and parA/parB. This deleted region is replaced by its own resistance region (see below). The absence of these addiction systems suggests that they are not important in the maintenance of this plasmid.
Plasmid pEC_bactec carries only two known resistance genes, blaTEM-1 and blaCTX-M-15 (Fig. 1). The TEM-1 enzyme is only able to hydrolyse amino- and carboxy-penicillins. Previous studies have shown that blaTEM genes are carried by three of the earliest described bacterial transposons, namely Tn1, Tn2 and Tn3. These transposons contain the transposase and resolvase genes, tnpA and tnpR, as well as a res resolution site , . The blaTEM-1 gene of pEC_Bactec was found to be located in a Tn3-like transposon possessing intact 38 bp inverted repeats (IRs) (Fig. 2). The Tn3-like transposon is flanked by 5-bp direct TTATC repeats functioning as a target site sequence. This indicates that this transposon has jumped from another plasmid or chromosome into this plasmid, a process called transposition. Interestingly, the tnpA gene of the Tn3 transposon is disrupted by ISEcp1-blaCTX-M-15 due to ISEcp1-mediated transposition (Fig. 2). ISEcp1-like elements belong to the IS1380 family of insertion sequences and have been identified in association with genes belonging to the blaCTX-M-1, blaCTX-M-2, blaCTX-M-25 and blaCTX-M-9 ESBL gene clusters –.
ISEcp1-mediated transposition was shown to create a 5-bp duplication of the target sequence TATGA (Fig. 2). The mechanism involves the left inverted repeat (IRL) of ISEcp1 and a right inverted repeat (IRR1), which resembles the IRR of ISEcp1 (Fig. 2). Despite this disruption, the Tn3 tnpA gene still encodes a protein of 929 amino acids. If this truncated tnpA gene remains functional, it would mediate a new Tn3 element encoding CTX-M-15 in addition to TEM-1. Based on these findings, we can conclude that pEC_Bactec arose by transposition of Tn3 and ISEcp1-blaCTX-M-15.
A similar disruption of the Tn3-like transposon by an ISEcp1-blaCTX-M-3 element was recently reported for the IncI1 plasmid pEK204 found in a clinical E. coli isolate of human origin . However, this ISEcp1-blaCTX-M-3 element was inserted in an inverse way, which indicates that blaCTX-M-3 and blaCTX-M-15 represent separate escape events. Blast analysis showed 92% homology between pEC_Bactec and pEK204.
To our knowledge, this is the first genomic analysis of a blaCTX-M-15-carrying plasmid found in a clinical E. coli isolate from a horse. In different members of Enterobacteriaceae of human and animal origin, IncI1 plasmids encoding CTX-M-15 have been reported earlier , , . On some IncI1 plasmids, carrying blaCTX-M-15, extra non-β-lactam resistance determinants were observed . All these findings demonstrate the high plasticity of IncI1 plasmids.
Plasmid pEC_B24 is a circular molecule of 73801-bp harbouring 76 open reading frames (ORFs) (Table S2, Fig. 3). This plasmid was transferred in vitro by conjugation and belongs to the incompatibility group IncFII.
Comparative analysis with another IncF plasmid (pC15a-1a)  revealed that both pEC_B24 and pC15a-1a plasmids carry the FII replicon (repA2, repA3, repA1 and repA4), hok/mok, psiA/psiB and parB genes and the type I partitioning locus (parM, stbB) ensuring stable plasmid inheritance. The transfer region of pEC_B24 comprises 19 tra genes (traE, traK, traB, traP, traV, traR, traC, traW, traU, traN, traF, traQ, traH, traG, traS, traT, traD, traI and traX), 7 trb genes (trbG, trbI, trbC, trbE, trbA, trbB and trbJ) and finO. In comparison with the transfer region of other IncFII plasmids , , several tra genes (traM, traJ, traY, traA and traL), the trbF and artA genes are missing. Despite this incompleteness, pEC_B24 is easily transferred by conjugation.
Another interesting finding on this plasmid is the presence of colicin B and M genes. These transmembrane toxins kill sensitive strains of E. coli and closely related species by depolarising the cytoplasmic membrane, which lead to dissipation of cellular energy . To our knowledge, colicin B and M genes have never been described on IncFII plasmids. Whether these genes have an influence on the spread of this plasmid remains unclear and needs further investigation.
Genes encoding antimicrobial resistance are blaTEM-33 and blaCTX-M-15 (Fig. 3). TEM-33, also called inhibitor resistant TEM 5 or IRT-5, has a reduced affinity for β-lactamase inhibitors. IRT enzymes have negligible activity against extended-spectrum cephalosporins and are not considered to be ESBLs . The blaTEM-33 gene of pEC_B24 is located in a Tn3-like transposon possessing intact 38-bp inverted repeats (IRs). Like in pEC_Bactec, the Tn3-like transposon was inserted after a transposition event and is flanked by 5-bp directed repeats (ATATA) (Fig. 2). Likewise, the Tn3 tnpA gene is disrupted by ISEcp1-blaCTX-M-15 due to ISEcp1-mediated transposition creating a 5-bp duplication of the target sequence TATGA (Fig. 2). The ISEcp1-blaCTX-M-15 element seems to show high affinity for this target sequence because its transposition was also described for another IncFII plasmid, pC15a-1a . BLAST analysis yielded 72, 44 and 41% homology between pEC_B24 and previously described IncFII plasmids pC15a-1a, pEK499 and pEK516, respectively. Thus, pEC_B24 represents a new variant of IncFII plasmids harbouring blaCTX-M-15.
Plasmids pEC_L8 and pEC_L46 are both circular molecules counting 118525 and 144871-bp with 133 and 162 ORFs, respectively. Both plasmids were transferred by conjugation, belong to the incompatibility group F and represent two replicons of type FII and FIA (Table S3 and S4, Fig. 3). Comparative analysis revealed 80% homology between both plasmids.
Both plasmids also comprise several addiction systems to ensure stable plasmid inheritance and post-segregation killing: (1) the post-segregation killing protein Hok and its modulator Mok, both located near the parB and psiA/psiB genes; (2) the toxin-antitoxin system pemI/pemK; (3) two copies of the vagC/vagD virulence-associated genes; (4) and the toxin-antitoxin system ccdA/ccdB, located near the FIA replicon (Fig. 3). These addiction systems have been described in other IncF plasmids , , . They may explain the success in the maintenance of these plasmids even in the absence of antibiotic selective pressure.
Both plasmids also share a region with two copies of genes encoding permeases and ATP-binding proteins of the ABC transporter family. Whether these transporters contribute to virulence and plasmid maintenance remains unknown and needs further investigation.
The pEC_L8 and pEC_L46 plasmids had a complete transfer region: 24 tra genes (traM, traJ, traY, traA, traL, traE, traK, traB, traP, traV, traR, traC, traW, traU, traN, traF, traQ, traH, traG, traS, traT, traD, traI and traX), 8 trb genes (trbG, trbI, trbC, trbE, trbA, trbB, trbJ and trbF) and finO. In comparison with other IncF plasmids , , these plasmids acquired an extra trbD and traR gene. This complete tra region makes them self-transmissible.
Plasmid pEC_L8 carries six genes encoding antimicrobial resistance: blaTEM-1, blaCTX-M-15, blaOXA-1 (oxacillin-hydrolizing capabilities, not regarded as ESBLs), aac6'-lb-cr (encodes an aminoglycoside acetyltransferase that also modifies ciprofloxacin and norfloxacin), tetA (encoding resistance to tetracyclines) and catB4 (chloramphenicol resistance). All these resistance genes, with the exception of the tetA gene, were also found on the pEC_L46 plasmid backbone. The tetA gene is located in a Tn1721-like transposon. This mobile genetic element has been described previously on plasmid backbones of the IncA/C group  which indicates that its appearance is not limited to plasmids of the IncF group. The resistance genes blaOXA-1, aac6'-lb-cr and catB4 are found in an IS26-composite transposon of which the IS26 elements lie in opposite orientations (Fig. 4). Examination of the pEC_L8 backbone revealed two IS26-composite transposon elements, which were inserted in an inverted orientation. Both transposons are separated by two oxidoreductase (yigB genes) elements (Fig. 4). Interestingly, the IS26 element of one composite transposon element, located downstream of the two yigB genes, is truncated (left composite transposon in Fig. 4). To our knowledge, this is the first observation of this type of insertional arrangement. However, it is unclear whether this duplication provides any advantage in terms of resistance to oxacillin, aminoglycosides and chloramphenicol and needs further investigation. Furthermore, the 5′end of the aac6'-lb-cr gene truncates one IS26 element, and the 3′end of the catB4 gene is truncated by the other IS26 element (Fig. 4). Extensive analysis of IS26 revealed promoter-like elements in which the putative -35 box (TTGCAA) was found in the left inverted repeat of IS26 and the putative -10 box was found 14-bp upstream of the −35 box. This may provide a mechanism for expression of the downstream resistance genes. A similar hybrid promoter involving IS26-inverted repeat sequences was previously shown to steer the expression of aminoglycoside resistance genes located in a composite transposon .
The blaTEM-1 gene of pEC_L8 is contained in a Tn3-like transposon possessing intact 38-bp inverted repeats (IRs) (Fig. 2). Inserted after a transposition event, the transposon is flanked by 5-bp directed repeats of TTATA. Likewise, the tnpA gene is disrupted by ISEcp1-blaCTX-M-15 due to ISEcp1-mediated transposition creating a 5-bp duplication of the target sequence TATGA (Fig. 2). The latter transpositional event also occurred in pEC_L46. However, the new Tn3 transposon was rearranged by IS26 elements so that the tnpR gene is now flanked by IS26. The blaTEM-1 gene is situated further downstream on the pEC_L46 backbone (Table S4, Fig. 3) and its 5′end is disrupted by an IS26 element.
Further examination of the pEC_L46 backbone revealed more structural rearrangements. Interestingly, all these rearrangements were also flanked by IS26 elements with intact left and right inverted repeats (Fig. 3). This shows the importance of IS26 in the plasticity of plasmids.
A major difference between pEC_L8 and pEC_L46 is the insertion of a 29.5-kb fragment flanked by IS26 in the pEC_L46 backbone (Table S4, Fig. 3). This fragment showed more than 90% homology with a segment of the previously described plasmid S9 from a Klebsiella strain . This finding indicates that genetic exchange between S9 and pEC_L46 may have happened and that the insertion sequence IS26 played a crucial role in this process. Insertion of this large fragment provided pEC_L46 with extra hypothetical proteins, tra genes (two copies for some) and genes encoding antirestriction proteins (Table S4).
The antirestriction proteins, ArdK and ArdR, may play a role in overcoming the host restriction barrier by self-transmissible broad-host range plasmids. They can serve as a genetic switch that controls the expression of plasmid-encoded antirestriction functions during mating . Whether the extra tra genes could affect the transfer of this plasmid by conjugation remains unclear and needs further investigation.
None of the IS26 elements, responsible for the structural rearrangements of the pEC_L46 backbone or flanking the 29.5-kb fragment, and those flanking the composite transposons of both pEC_L8 and pEC_L46, showed target site duplications. Moreover, the absence of a direct target sequence duplication at either side of these IS26-flanked regions strengthens the hypothesis of entry through homologous recombination rather than transposition.
Nevertheless, transpositional events creating direct target sequence duplications have been discovered for certain IS26-composite transposons , . In view of the wide spread of IS26 among plasmids, the acquisition of resistance genes and other genetic fragments might be realized in different ways.
Blast analysis of pEC_L8 and pEC_L46 against pEK499, another IncF plasmid encoding blaCTX-M-15, resulted in 79% and 75% homology, respectively. Plasmids of the IncF group are geographically widespread and have played a crucial role in the spread of CTX-M-15 in Enterobacteriaceae among humans .
We have determined the complete nucleotide sequence of four CTX-M-15-encoding plasmids carried by three clonally distinct clinical E. coli isolates of human origin and one clinical E. coli isolate from a horse. For the first time, the nucleotide sequence of a blaCTX-M15-carrying plasmid (IncI1) of a clinical isolate from animal origin is described. Comparative analysis of this plasmid with a blaCTX-M-3-carrying plasmid (IncI1) of a clinical isolate of human origin  revealed a high degree of homology (>90%), indicating that similar plasmids carrying different blaCTX-M genes are circulating both in the human and animal ecosystem. Notably, more plasmid genomic research is necessary to ascertain whether the human and animal environment could be experienced by bacteria in a different way.
This study also highlights the evolution of IncF plasmids, thereby enhancing our understanding of the pedigree of these plasmids carrying blaCTX-M genes. The multitude of addiction systems present on these IncF plasmids ensures their maintenance even in the absence of antibiotic selection.
This report has also provided support for the seemingly frequent transposition events of the blaCTX-M-15 gene linked to ISEcp1. This element seems to have a preferred insertion site at the tnpA gene of a blaTEM-carrying Tn3-like transposon, the latter also being inserted by a transposition event. Both transposition events prefer AT-rich target sequences, whereas the ISEcp1-blaCTX-M-15 element prefers the same target sequence for all Tn3 elements. Whether this new Tn3 element, encoding blaCTX-M-15 in addition to blaTEM, remains mobile is still unclear. The potential for transposition of two bla genes, of which one is an ESBL gene, has public health implications since extended-spectrum cephalosporins are extensively used in human and veterinary medicine. This needs further investigation.
The mobile genetic element, IS26, member of the IS6 family, is shown to play an important role in the plasticity of the investigated plasmids. It inserts by homologous recombination, as indicated by the absence of target site duplication, and causes rearrangements. IS26 seems not only to facilitate the mobilisation of chromosomal fragments  but also fragments from other plasmids. Collectively, these data suggest that IS26 and ISEcp1 play a critical role in the evolution of diverse multiresistant plasmids found in clinical Enterobacteriaceae.
E. coli B24, L8 and L46 were isolated in Belgium from human patients with urinary tract infections. B24 was collected from a urine sample from a hospitalized patient of the Ghent University Hospital. L8 and L46 were defined as community-acquired isolates from urine samples obtained from a medical centre serving only general practitioners in Leuven. All isolates belonged to sequence type ST131 and showed clonally distinct PFGE-fingerprint patterns . E. coli Bactec was isolated from the joint of a horse suffering arthritis at the Faculty of Veterinary Medicine, Ghent University. All isolates were shown to produce CTX-M-15 by PCR and sequencing . All isolates were available at the start of this study and were gathered as part of standard care.
Approval of the Ethics Review Board at the University of Ghent was not necessary, because no information about the patients was released and the identified E. coli isolates were exempt from requirements for ethical approval.
Plasmid transfer experiments were carried out in Luria Broth medium. E. coli J5, resistant to rifampicin, was used as recipient strain. Conjugation experiments were performed overnight at 37°C with a donor/recipient ratio of 0.2. Transconjugants were selected on MacConkey agar plates (Oxoid LTD, Basingstoke, Hampshire, England) supplemented with ceftiofur (8 mg/liter) and rifampicin (250 mg/liter). Plasmid DNA was isolated from the transconjugants using a plasmid midi Qiagen kit (Venlo, the Netherlands) according to the manufacturer's instructions and separated on 0.8% 1x TAE agarose gels by electrophoresis at 140V for 4 h at 4°C. The molecular size of each ESBL-carrying plasmid was estimated by using a BAC Tracker Supercoiled DNA ladder (ranging from 165 kb to 8 kb) (Epicentre Biotechnologies, Madison, Wisconsin).
Twenty µg of each plasmid DNA was extracted and purified as described above. Roche GS-FLX titanium libraries were generated, using 5 µg of each of the 4 purified plasmid DNA samples. The DNA was fragmented by nebulisation, followed by a double Solid Phase Reversible Immobilization (SPRI) bead capture size selection with Ampure beads (Agencourt Bioscience) to generate DNA fragments of 400–1,500 bp in length. Selected fragments were end-repaired and ligated to 454 sequencing adapters. Single-stranded libraries were then generated according to the Roche GS FLX Titanium General Library Preparation Method Manual (version October 2008). These single-stranded libraries were used to perform an emulsion PCR according to the Roche GS FLX titanium emPCR Method Manual (version October 2008). The 4 resulting bead libraries were sequenced on a Roche GS-FLX system following the GS FLX Titanium Sequencing Method Manual (version October 2008). A 70×75 mm picotiter plate was divided in 8 lanes using a rubber gasket. For each of the 4 bead libraries, 1 lane was loaded with 340.000 DNA library beads.
The Mimicking Intelligent Read Assembly package MIRA (version 2.9.43) was used to perform de novo genome assembly. For the pEC_Bactec sample, 50891 of the 55632 generated sequences were assembled into one relevant contig of 93799-bp. The start and the end of this contig showed significant overlap and represented the complete, circular sequence of the pEC_Bactec plasmid. The average sequencing coverage for this plasmid was 179x. For the pEC_B24 sample, 23915 of the 25319 generated sequences were assembled into one relevant contig of 74322-bp. The start and the end of the 74322-bp contig showed significant overlap and represented the complete, circular sequence of the pEC_B24 plasmid. The average sequencing coverage for this plasmid was 112x. For the pEC_L8 sample, 54048 of the 58671 generated sequences were assembled into five relevant contigs ranging from 64755 to 7430 bp. Walking reads were used to assemble the contigs and to fill-in the gaps. The average sequencing coverage for this plasmid was 109x. For the pEC_L46 sample, 31381 of the 35029 generated sequences were assembled into 4 relevant contigs ranging from 65414 to 5267 bp. Walking reads were used to assemble the contigs and to fill-in the gaps. The average sequencing coverage for this plasmid was 71x.
In order to maximize the number of quality gene annotations, an ab initio annotating approach was followed.
Theoretical open reading frames (ORFs) were first determined using the EMBOSS getorf tool (with minimum ORF length set to 90 nucleotides, and taking all alternative start codons into account). All ORFs were translated subsequently, and BLAST (more precisely the blastp program of the BLAST suite)  was performed with an e threshold of 10−15 against the Uniprot/KB universal protein database. The generalist algorithm of getorf yielded roughly a tenfold of the expected natural ORFs, reducing the risk of false negatives. In order to keep the false positive rate low, extra parameters were considered: 1) percentage alignment between query and hit ORFs; 2) percentage conservation between aligned portions of query and hit ORFs; 3) ribosome binding strength (see details further down). ORFs with a bad ranking value (arbitrary but consistent cutoff for all comparisons) for any two of these three parameters were considered as false positives and discarded.
Ribosome binding strength was estimated by applying two long established facts.
(1) On an mRNA strand, usually within 20 nucleotides before the actual start codon, the reverse complement of 5 to 7 nucleotides near the 16S rRNA 3′ end acts as an attractor and positioner for the ribosomal small subunit (which contains the 16S rRNA in addition to a set of ribosomal proteins). This sequence appears to be more than averagely conserved and is known as the Shine-Dalgarno sequence , ; (2) In Gram-negative bacteria such as organisms belonging to the former Proteobacteria division (which includes Escherichia), an AU-rich mRNA region of 16 nucleotides long and immediately preceding the Shine-Dalgarno sequence may also attract and position ribosomes to help initiate translation of the correct, biologically active gene product , . For E. coli, the Shine-Dalgarno sequence was determined to be a subsequence of AGGAGGU (which is the reverse complement of the 3′ end of the 16S rRNA), and the minimum AU-richness (equivalent to ribosome binding capacity) of the preceding region was arbitrarily set to 10/16. For each theoretical ORF a range of possible start codons was scored; the higher the similarity to the ideal Shine-Dalgarno sequence, or the AU-richer the preceding region, or the better a combination of both, the more likely the potential start codon is to be the actual start codon.
For each BLAST hit, the following annotation information was either parsed from the BLAST report or further derived from parsed data (all in an automated fashion via Perl and Bioperl scripting): 1) the full query (E. coli) plasmid-derived ORF with coordinates (contig, theoretical ORF number, frame, ORF length, start and stop position); 2) corresponding hit parameters if relevant for further comparison; 3) remarks about ORF shortcomings (e.g. missing or extra domains with regard to the reference ORF, ORF interruption upon hitting a contig end) encountered during the parsing and annotation process; 4) an estimate of the ribosome binding strength of the mRNA region preceding the start codon; 5) nucleic acid sequence of the ORF; 6) BLAST alignment data (e value, number and fraction of amino acids unaligned, fraction of identical and fraction of conserved amino acids); 7) gene name and description if available.
Sequences of the open reading frames were compared and aligned with Genbank data using BLAST and with reference plasmids (R64 (NC_005014.1; reference plasmid for the IncI1 group), pEK204 (EU935740), pEK499 (EU935739), pEK516 (EU935738) and pC15a-1a (AY458016) by two sequence alignment using Blast2seq (http://blast.ncbi.nih.gov/Blast.cgi).
The pMLST scheme for IncI1 plasmids was used for EC_Bactec as described previously . The pilL (254 bp), sogS (254 bp), ardA (343 bp), repI1 (104 bp) and tnbA-pndC (812 bp) fragments were compared with known allelic variants (http://pubmlst.org/perl/mlstdbnet/mlstdbnet.pl?file=incI1_profiles.xml&page=oneseq).
pEC_Bactec (GU371927), pEC_B24 (GU371926), pEC_L8 (GU371928), pEC_L46 (GU371929).
(0.12 MB DOC)
(0.11 MB DOC)
(0.17 MB DOC)
(0.20 MB DOC)
Competing Interests: The authors have declared that no competing interests exist.
Funding: This work was supported in part by a grant from the Federal Public Service of Health, Food Chain Safety and Environment (Grant number RT 06/3 ABRISK), by the CODA-CERVA-VAR and Ghent University. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.