Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Open Access

Peer-reviewed

Research Article

A Role for Tn6029 in the Evolution of the Complex Antibiotic Resistance Gene Loci in Genomic Island 3 in Enteroaggregative Hemorrhagic Escherichia coli O104:H4

  • Piklu Roy Chowdhury,

    Affiliations The ithree institute, University of Technology Sydney, PO Box 123 Broadway, Sydney, New South Wales, 2007, Australia, NSW Department of Primary Industries, Elizabeth Macarthur Agricultural Institute, Private Bag 4008, Narellan, New South Wales, 2567, Australia

    ⨯
  • Ian G. Charles,

    Affiliation The ithree institute, University of Technology Sydney, PO Box 123 Broadway, Sydney, New South Wales, 2007, Australia

    ⨯
  • Steven P. Djordjevic

    Steven.Djordjevic@uts.edu.au

    Affiliation The ithree institute, University of Technology Sydney, PO Box 123 Broadway, Sydney, New South Wales, 2007, Australia

    ⨯

Correction

17 Apr 2015: The PLOS ONE Staff (2015) Correction: A Role for Tn6029 in the Evolution of the Complex Antibiotic Resistance Gene Loci in Genomic Island 3 in Enteroaggregative Hemorrhagic Escherichia coli O104:H4. PLOS ONE 10(4): e0126197. https://doi.org/10.1371/journal.pone.0126197 View correction

Abstract

In enteroaggregative hemorrhagic Escherichia coli (EAHEC) O104 the complex antibiotic resistance gene loci (CRL) found in the region of divergence 1 (RD1) within E. coli genomic island 3 (GI3) contains blaTEM-1, strAB, sul2, tet(A)A, and dfrA7 genes encoding resistance to ampicillin, streptomycin, sulfamethoxazole, tetracycline and trimethoprim respectively. The precise arrangement of antibiotic resistance genes and the role of mobile elements that drove the evolutionary events and created the CRL have not been investigated. We used a combination of bioinformatics and iterative BLASTn searches to determine the micro-evolutionary events that likely led to the formation of the CRL in GI3 using the closed genome sequences of EAHEC O104:H4 strains 2011C-3493 and 2009EL-2050 and high quality draft genomes of EAHEC E. coli O104:H4 isolates from sporadic cases not associated with the initial outbreak. Our analyses indicate that the CRL in GI3 evolved from a progenitor structure that contained an In2-derived class 1 integron in a Tn21/Tn1721 hybrid backbone. Within the hybrid backbone, a Tn6029-family transposon, identified here as Tn6029C abuts the sul1 gene in the 3´-Conserved Segment (-CS) of a class 1 integron generating a unique molecular signature that has only previously been observed in pASL01a, a small plasmid found in commensal E. coli in West Africa. From this common progenitor, independent IS26-mediated events created two novel transposons identified here as Tn6029D and Tn6222 in 2011C-3493 and 2009EL-2050 respectively. Analysis of RD1 within GI3 reveals IS26 has played a crucial role in the assembly of regions within the CRL.

Citation: Chowdhury PR, Charles IG, Djordjevic SP (2015) A Role for Tn6029 in the Evolution of the Complex Antibiotic Resistance Gene Loci in Genomic Island 3 in Enteroaggregative Hemorrhagic Escherichia coli O104:H4. PLoS ONE 10(2): e0115781. https://doi.org/10.1371/journal.pone.0115781

Academic Editor: Willem van Schaik, University Medical Center Utrecht, NETHERLANDS

Received: April 22, 2014; Accepted: December 1, 2014; Published: February 12, 2015

Copyright: © 2015 Chowdhury et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited

Data Availability: All relevant data are within the paper and its Supporting Information files.

Funding: This work was supported by a joint collaboration between the NSW Department of Primary Industries and the ithree Institute. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Competing interests: The authors have declared that no competing interests exist.

Introduction

The German outbreak of E. coli O104:H4, and a smaller number of sporadic cases of E. coli O104:H4 infections that followed a few weeks later in France, resulted in an unusually high incidence of haemolytic uremic syndrome (HUS) (850 cases) (as well as bloody diarrhoea (4320 cases) and 82 deaths (total numbers)) [1]. Isolates from these outbreaks, including 2011C-3493 isolated from a US patient with a history of travel to Germany, express Shiga toxin 2, (stx2) which is typically associated with enterohemorrhagic (EHEC) strains. The outbreak strains are however phylogenetically related to enteroaggregative (EAEC) strains. The hybrid features of O104:H4 indicated that they belong to a new lineage known as the enteroaggregative haemorrhagic E. coli (EAHEC). Strains isolated from the German outbreak and the ones from the sporadic infections in France form a separate clade (Clade 1) distinct from historical enteroaggregative O104:H4 isolates,- such as strain O104:H4 55989 recovered from Central Africa in 1995 [2–4]. EAHEC O104:H4 has also caused cases of HUS and bloody diarrhea in the Republic of Georgia. Strains isolated in 2009 from the Republic of Georgia (2009EL–2050, 2009EL–2071) are closely related to the Clade 1 strains but cluster distinctly from them and the historical isolates [2].

The 2011 outbreak strains and subsequent isolates recovered from sporadic cases of HUS in France were found to be resistant to multiple antibiotics including third generation cephalosporins (ceftiofur and ceftriaxone), ampicillin, streptomycin, trimethoprim, sulfamethoxazole, tetracycline and sulfisoxazole, severely limiting treatment options [4–7]. Unique to the 2011 outbreak strains was a large IncI1 plasmid encoding the CTX-M-15 cephalosporinase [8]. The plasmid has been sequenced [8] and extensively characterised in EAHEC O104:H4 isolates, but a detailed analysis of the antibiotic resistance genes and their arrangement in Genomic island (GI3) has not been determined.

The acquisition and loss of mobile genetic elements can play an important role in the emergence of new pathogens. The acquisition of virulence genes and antibiotic resistance are considered to be significant events in the emergence of new pathogens because of the potential to impact fitness and transmissibility [9]. Comparative whole genome based analyses of O104:H4 isolates has revealed diversity between EAHEC O104:H4 strains [2,4]. Closed genomes now exist for three EAHEC O104:H4 strains including a representative of the German outbreak strain 2011C-3493 and two strains from the republic of Georgia (2009EL-2050 and 2009EL2071). The closed genome sequences provide contextual information of the precise arrangement of resistance genes and is important for understanding the molecular events that generate complex antibiotic resistance gene loci (CRL)[3,10], [11]. Such closed genomes allowed us to determine the precise arrangement of antibiotic resistance genes within GI3 and the molecular events that created the CRL.

GI3 is a hotspot for genetic events within O104:H4 genomes [2,3] and is unique to Clade 1 O104:H4 strains [3]. It is an important region within O104:H4 genomes because it has previously been demonstrated to be mobile [12,13]. GI3 targets a widely-dispersed genomic hotspot consisting of a target 23 bp sequence and consequently it may be significant in the emergence of future MDR pathogens in the Enterobacteriaceae. Moreover, it houses a CRL containing the mercury resistance gene cluster (mer-module) and the virulence factor Ag43, implicated in biofilm formation [3]. The CRL also contains genes encoding resistance to ampicillin (blaTEM-1), streptomycin (strAB), sulfisoxazole (sul1 and sul2), tetracycline (tetA(A)) and trimethoprim (dfrA7). Notably, comparative genome studies of O104:H4 were the first to identify the blaTEM1-sul2-strA-strB gene cluster in a chromosomal location. Previously, this gene cluster had only been identified on plasmid backbones [14–18]. For example, an IncHI2 plasmid from Salmonella enterica serovar Typhimurium from Australia carries a sul2-strA-strB gene cluster together with plasmid replication genes repA and repC adjacent to a module encompassing the blaTEM-1 gene [15]. On the plasmid, these genes comprise a large part of Tn6029, a transposon flanked by direct copies of IS26 [15]. The molecular events that led to the construction of Tn6029 have been described and involve the insertion of Tn1, Tn2 or Tn3 in plasmid RSF1010 followed by the independent insertion of three copies of IS26 [15]. In 2011 a second variant of the transposon, Tn6029B [18], was described within the sequence of an IncH1 plasmid, (pHCM1 [30]) isolated from a Salmonella enterica serovar Typhi strain from Vietnam in 1993.

IS26 is recognised to be an important driver in the evolution of CRL within Gram-negative pathogens [11,19–26]. Tn6029 and related structures have been identified on multiple antibiotic resistance plasmids [14–18,27,28] suggesting they play an important role in the dissemination of the blaTEM1-sul2-strA-strB cluster within E. coli and Salmonella populations. Given that the blaTEM1-sul2-strA-strB gene cluster has now been reported to be chromosomally located on GI3 of EAHEC O104:H4 we were interested in determining whether Tn6029-family transposons have played a role in the evolution of the chromosomally located CRL within EAHEC O104:H4 genomes.

Here, we present the key defining features of the CRL within GI3 of the finished genomes of the 2011 German outbreak strain, 2011C-3493 (CP003289.1) and the 2009 Republic of Georgia strain, 2009EL-2050 (CP003297.1). Our analysis indicates that IS26-mediated events have played a major role in shaping GI3 in EAHEC O104:H4. We propose an evolutionary model for the CRL seen in EAHEC O104:H4 isolates,- based on the genetic signatures we identified from the closed genomes and iterative BLASTn searches of the microbial genome database.

Materials and Methods

Strains and sequences analysed

Regions around Genomic island 3 in the finished genomes of EAHEC strains 2011C-3493 (CP003289.1), 2009EL-2050 (CP003297.1) and 2009EL-2071 (CP003301.1) [2] were analysed in detail to decipher the arrangement of genes within the CRL. Strain 2011C-3493 is from the United States from a patient that had recently returned from Germany during the epidemic and is considered representative of German outbreak strains [2]. Strain 2009EL-2050 was recovered in 2009 from a patient with bloody diarrhoea in the Republic of Georgia and is epidemiologically unrelated to the 2011 outbreak strains [2]. In addition, we have also analysed corresponding regions on previously published plasmids—pASL01a (NC_019091.1) [27], pHCM1 (NC_003384.1) [30] and pAKU_1 (AM412236.1) [32].

Data mining

Defined sequences spanning regions within RD1 in strains 2011C-3493 and 2009EL-2050 were downloaded from GenBank and aligned using Mauve version 2.3.1 [29] as described below. The automated annotations available through NCBI genome database were manually curated using the NCBI ORF finder and iterative BLASTn and BLASTp searches. Number of nucleotides spanning the priming sites of primers L1 and JL-D2 [19] were used to determine the precise location of Tn6029-family transposons within RD1 using AmplifyX version 3.1 software.

BLASTn analysis of draft O104:H4 genomes

Fifty three draft and complete genomes of Escherichia coli O104:H4 isolates (taxid:1038927) available in the GenBank NCBI microbial genome database (on 21st July 2014) were queried with BLASTn using sequences representing different regions within the RD1 structure seen in 2009EL-2050 and 2011C-3493, and proposed to represent a progenitor in this study (see details in results section). The DNA sequence of segments 1 and 2, were derived from the genome sequence of isolate 2011C-3493 while that of fragments 3, 4 and 5 were derived from 2009EL-2050. BLAST alignments that exhibited greater than 99.95% identity over 100% of the query sequence length against E. coli O104:H4 genomes were deemed significant and analysed in detail here. BLAST hits showing sequence identity against varying lengths of the input sequence spanning Tn6029 (Fragment 4) were also considered because (1) the region is predicted to be evolving rapidly [14,16,17,27,30] and (2) the region had multiple copies of an insertion element which typically generate scaffold breaks during assembly. A visual representation of locations of the different input query start positions as it appears in S3 Table is detailed in S1 Fig.

Results

Genetic features of RD1 that distinguish O104:H4 strains 2009EL-2050 and 2011C-3493

To decipher the salient features of RD1 in 2009EL-2050 (Fig. 1A) and 2011C-3493 (Fig. 1B) we carried out an in-depth sequence analysis and annotation of the regions in both strains. Our analysis showed that the basic backbone of the CRL that defines RD1 in O104:H4 strains comprise a chimeric Tn21/Tn1721 structure with a Tn1721-associated tetA-tetR-pecM module adjacent to a mercury resistance module from Tn21. In both isolates the CRL in RD1 (Fig. 1A and 1B) harbours a class 1 integron, typical of the In2 family, with a dfrA7 cassette that encodes resistance to trimethoprim [31]. However, the class 1 integron in 2011C_3493 is inverted compared to how it appears in 2009EL-2050 (See Fig. 1A and 1B). The 3´-conserved segment (3´-CS) of the class 1 integron in both strains is partially deleted by the insertion of a composite transposon (see later) and comprises qacE∆1 and a portion of the sul1 gene (79 nucleotides at the 3´ end of the sul1 gene). The deletion in sul1 and remaining regions of the 3´-CS probably arose by insertion of a composite IS26 transposon in the 3´-CS followed by an IS26-mediated deletion event [19]. These genetic events generated variations in the sequences within RD1 in these strains and will be detailed below.

thumbnail
Download:
Fig 1. Cartoons defining unique features of the complex resistance loci seen in RD1 of EAHEC O104:H4 isolates.

A: RD1 in EAHEC isolate 2009EL-2050. B: RD1 in isolate 2011C-3493. The symbol representing different copies of IS26 within the resistance loci is highlighted as the key at the top of the figure.

https://doi.org/10.1371/journal.pone.0115781.g001

Downstream of the composite IS26 transposon within RD1 is the remnant of ∆tniA missing 428 nucleotides of the gene and the IRt inverted repeat, both of which represent key structural components of clinical class 1 integrons (Fig. 1). A mercury resistance module identical to that found in Tn21 is present beyond IRt. The insertion site of the class 1 integron in the mercury module found in EAHEC O104:H4 is typical of the In2-family. The CRL carries a second truncated copy of a tniA gene (indicated by ∆tniA ** in Fig. 1), missing 996 nucleotides from the 3´ end of the gene. ∆tniA ** (Fig. 1) and its associated IRt is found adjacent to the tetA-tetR-pecM module, an orientation identical to that seen in Tn1721-derived transposons. These data suggests that the two copies of ∆tniA have separate origins.

Notably the 3´-CS of the class 1 integron in both 2011C-3493 and 2009EL-2050 comprises 1210 nucleotides and includes the qacE∆1 gene plus a partial copy of the sul1 gene. The sul1 gene is typically found as a complete ORF in the 3´-CS of most clinical class 1 integrons. An IS26-mediated deletion resulted in the loss a region of the 3´-CS leaving only 79 nucleotides at the 3´ end of the sul1 gene. Previously we have shown that IS26-mediated deletion events create novel signatures that can be used in tracking lateral movement of CRL within bacterial populations [16,17,19,20]. In-silico PCR simulation using primers in intI1 (L1) and IS26 (JL-D2) (see methods) is expected to generate a novel 2082 bp amplicon that can be used as a genetic signature to test for the presence of the CRL in EAHEC O104:H4 (Fig. 2A). Notably, BLASTn analysis of the sequence of the 2082 bp amplicon indicated that it is identical to a homologous fragment in pASL01a (27,072 bp) [27], a small plasmid circulating within commensal E. coli in West Africa. pASL01a has a backbone of 5,168 bp that carries the plasmid replication genes and Tn21-derivative transposon, TnASL01a [27]. Transposon TnASL01a encodes resistance to ampicillin, streptomycin, sulfathiazole and trimethoprim, has a class 1 integron carrying dfrA7 (trimethoprim resistance) gene and an IS26 transposon which carries the blaTEM1-sul2-strA-strB gene cluster [27]. The composite IS26 transposon in pASL01a appears at precisely the same location as in RD1 of 2011C-3493, a remarkable observation that supports the hypothesis that these derivate Tn21 transposons share evolutionary history.

thumbnail
Download:
Fig 2. Molecular signatures created by the insertion of IS26.

A: The 3´-CS of class 1 integrons in pASL01a and pAKU_1 has been structurally modified by different IS26-mediated deletion events such that PCR with L1 and JL-D2 primers is expected to generate 2082 and 2342 bp long amplicons respectively. B: Eight base pair signature sequences created by the insertion of IS26 found flanking the inverted repeats of IS26 elements clearly suggest that the CRL in strain 2011EC-3493 is a derivative of Tn6029C while 2009–2050 is a derivative of Tn6029B described previously in pHCM1. The eight base repeat on the left end of Tn6029B and that in 2009EL-2050 are different because in pHCM1 (Tn6029B) there is a deletion of the class 1 integron whereas in 2009EL-2050 the fragment adjacent to the left hand end of the insertion site is inverted by IS26-mediated events. These signatures are also consistent with the proposition that Tn6029C evolved from Tn6029B.

https://doi.org/10.1371/journal.pone.0115781.g002

BLASTn analysis of the 2082 bp amplicon sequence also generated a partial match with plasmid pAKU_1 [32]. Alignments revealed the 2082 fragment is identical across the entire query sequence except for a gap of 260 nt. In pAKU_1, a portion of the orf5 gene (which is often found associated with 3´-CS of clinical class 1 integrons) is present, which is missing in 2011C-3493 and 2009EL-2050. Detailed sequence similarity searches confirmed that pAKU_1 retains binding sites for primers L1 and JL-D2 and the diagnostic PCR is expected to generate a 2342 bp fragment (Fig. 2) clearly distinguishing the evolutionary histories of the CRL found in pAKU_1 with pASL01a, 2009EL-2050 and 2011C-3493.

Evidence of Tn6029-derived regions within RD1

Within the 3´-CS∆ and the IRt of the class 1 integron in outbreak isolate 2011C-3493 are coupled IS26 elements (indicated by a box in Fig. 1B) which have their respective transposase genes facing away from each other, followed by repA-repC-sul2-strA-strB module and a third copy of IS26. Boundaries of the insertion element IS26 are defined by the presence of 14 nucleotide inverted repeats (TTTGCAACAGTGCC) at either end (IR left or IRl and IR right or IRr). Insertion of an IS26 creates a direct repeat of eight nucleotides at the insertion site. Eight nucleotides long sequences directly abutting IS26 can therefore be used as a signature to track the evolutionary events that constructed the CRL [19]. Within the coupled IS26 structure we i) identified a deletion of 10-nucleotides of IS26-IRl and ii) addition of an eight nucleotides CCCCATAT. As a consequence, a 12-nucleotide spacer signature, TTTGCCCCATAT, separates the two copies of IS26 in the coupled IS26 structure (Fig. 1B). Independent BLASTn analysis of the module consisting of the repA-repC-sul2-strA-strB bounded by two copies of IS26 indicated 100% identity to the module seen in pASL01a, pHCM1 and other plasmids that carry Tn6029-family transposons.

In isolate 2009EL-2050, the region flanked by direct copies of IS26 between IRi and IRt had two separate modules, one which had ∆tnpR-blaTEM1-∆CR2 genes, while the other had repA-repC-sul2-strA-strB genes (Fig. 1A). Independent BLASTn analysis of the entire IS26 transposon structure containing both the ∆tnpR-blaTEM1-∆CR2 and repA-repC-sul2-strA-strB modules showed 99.99% sequence identity to transposon Tn6029B, first described in plasmid pHCM1 (Fig. 2B). Based on our analysis of the CRL present in the two genomes the simplest likely progenitor that can give rise independently to the two CRL via separate micro-evolutionary events within a very short time is depicted in Fig. 3. Details of these micro-evolutionary events are explained below.

thumbnail
Download:
Fig 3. Structure of a proposed progenitor locus from which RD1 has evolved.

Lines under the structure represent segments, sequence of which formed the query input sequence in BLASTn analysis of the NCBI Microbial genome database.

https://doi.org/10.1371/journal.pone.0115781.g003

A comparison of the CRL found in strains 2011C-3943 and 2009EL-2050 showed: i) differences in the number, location and orientation of IS26 elements, ii) the orientation of the class 1 integron (5´-CS and deleted 3´-CS) in 2009EL-2050 is inverted compared to that of 2011C-3493 and iii) that the module containing blaTEM-1 gene is absent in the outbreak strain 2011C-3493 but present in progenitor strain 2009EL-2050 (Fig. 1B and 1A respectively).

pASL01a deserves special mention in this context as it carries an identical copy of the class 1 integron containing dfrA7 with an IS26-flanked transposon inserted at an identical location of the 3´-CS as is found in both O104:H4 genomes: 2009EL-2050 and 2011C-3493. Our analysis indicates that TnASL01a contains a novel variant of Tn6029. The Tn6029 variant is characterized by an 11 bp deletion in the ∆tnpR-blaTEM-1-ΔCR2 segment seen in Tn6029B and by the inverted orientation of the module compared to the structures seen in Tn6029 or Tn6029B. We propose the name Tn6029C for the Tn6029 variant present in pASL01a. Characteristic features of Tn6029 variants and the unique attributes that distinguish Tn6029B from Tn6029C are shown in S2B Fig.

Tn6029 and variants Tn6029B and Tn6029C each carry the repA-repC-sul2-strA-strB gene cluster acquired from RSF1010 and blaTEM-1 from Tn1, Tn2 or Tn3 [15,33]. The description of the events that led to the formation of Tn6029 has been reported [15]. Tn6029B [18] differs from Tn6029 by loss of 85 nucleotides in ΔCR2 gene. Tn6026C was likely formed by an IS26-mediated inversion of the blaTEM-1 module followed by deletion of 11 nucleotides within the Tn2-derivate (S2B Fig.). This inversion event in Tn6029C has also reversed the characteristic eight nucleotide direct repeat signature sequences that abut IS26 elements in Tn6029B (Fig. 2B). Apart from CTCGCGCC in pHCM1, all eight nucleotide signature sequences adjacent to the different copies of IS26 present in Tn6029B and Tn6029C were identical to that seen in RD1 in isolates 2009EL-2050 and 2011C-3493 respectively (Fig. 2B).

Analysis of RD1 in EAHEC O104:H4 genomes

We interrogated the microbial genome database in NCBI for the presence of RD1 in GI3. Our searches were restricted to complete and draft genomes of E coli O104:H4 isolates (taxid: 1038927) that aligned to five separate fragments spanning RD1 (See Fig. 3). The rationale for our BLASTn approach resides in our hypothesis that the CRL seen in genomes 2011C-3493 and 2009EL-2050 are the product of independent micro-evolutionary events and therefore we did not want to restrict our searches to the specific structures seen in isolates 2011C-3493 or 2009EL-2050. A summary of the entire BLASTn searches is presented in Table 1, while detailed results of all the BLAST analysis are presented in S1–S5 Tables. On the 20th of July 2014, the GenBank microbial genomic database consisted of 53 E coli O104:H4 genomes including the three completely finished genomes described by Ahmed et al., [2] and five well-scaffolded draft genomes with an N50 of 1Mbp [3]. Notably, BLASTn analysis of Fragment 2 (Fig. 3), which includes the 2082 bp signature sequence generated matches with 100% sequence identity in 35 of the 53 O104:H4 genomes. BLASTn searches using Fragment 1 (Fig. 3) which spans a portion of GI3 and its junction with the Tn21-specific transposition module and a small overlapping portion with Fragment 2 marking the beginning of integron In2 showed 100% identity in 42 of the 53 genomes. Fragment 4, which spans the Tn21 mer-module (Fig. 3), showed 100% identity in 37 genomes. Thirty-six genomes also had the Tn1721-derived tetA-tetR-pecM module and the associated ΔtniA** gene.

thumbnail
Download:
Table 1. Summary of the BLASTn analysis of the different fragments considered to contain signature sequences essential for tracing micro-evolutionary events taking place within the CRL of GI3 in E coli O104:H4 draft genomes available in GenBank.

https://doi.org/10.1371/journal.pone.0115781.t001

Table 1 depicts a summary of BLASTn searches using Fragments 1–5. Our data indicates that 22 of the 53 (~42%) O104:H4 genomes in the NCBI microbial genome database have all five fragments that span the progenitor RD1 structure depicted in Fig. 3. Genomes Ec11–4984 (AHOU01000000), Ec11–4988 (AHOY01000000) and Ec11–4986 (AHOW01000000) carry Tn6029B in addition to the other four fragments spanning the CRL in our proposed progenitor structure (see Fig. 3). While RD1 in isolate Ec11–4984 (AHOU01000000) is split into two scaffolds, isolates Ec11–4988 and Ec11–4986 have a complete RD1 in one supercontig. Collectively our data supports the presence of the progenitor structure/s depicted in Figs. 3 and 4 in EAHEC O104:H4 isolates from around the world. Isolate Ec12–0466 (AIPR01000000) has a significant portion of the Tn6029B and all other segments that constitute the CRL in RD1 in two separate contigs [3]. The contig that has Tn6029B, Tn21 and the Tn1721 modules starts at a point just past the blaTEM1 gene of the ∆tnpR-blaTEM1-∆CR2 module. Similar scenarios were noted in 14 other incomplete genome sequences (S3 Table). Given that assembly of genome sequences using next generation sequencing technologies is problematic, specifically in regions that have multiple copies of any insertion sequences, our search of specific fragments that contain distinguishing features of CRL identified from our analysis of the two completely finished genomes (2011C-3493 and 2009EL-2050) suggests that variants of the proposed progenitor/s are most likely present within GI3 of the majority of O104:H4 genomes.

thumbnail
Download:
Fig 4. Sequence of events that led to the formation of RD1 in EAHEC O104:H4 isolates 2009EL-2050 from the Republic of Georgia and 2011C-3493, the German HUS representative isolate.

A: Events that shaped RD1 in 2009EL-2050. The progenitor was modified by IS26 inserting into the tnpM gene creating a characteristic eight base duplication followed by the inversion of a fragment of DNA bounded by two IS26 elements and containing a deleted class 1 integron. These events created a new composite IS26 transposon, Tn6222. Direct repeats and reverse complementary sequences generated by IS26 mediated events are depicted with square brackets [CAATGCGC] and the class 1 integron within curved brackets {TCCAT}. B: Events that shaped RD1 in 2011C-3493. Insertion of IS26 within the IRL of an existing IS26 in the progenitor CRL followed by a deletion of blaTEM-1 gene containing module in Tn6029C generated the unique CRL depicted in the outbreak isolate 2011C-3493 and led to the formation of the novel transposon Tn6223 (bottom panel). The 12 bp sequence within Tn6223 is a unique signature created by the deletion of the proposed blaTEM-1 gene containing fragment within Tn6029C. The asterisk on the underlined sequences within Tn6223 is the portion which is identical to IRL of IS26 and the remaining eight of the 12 bases indicate the eight base region which abuts the IS26 at the other end.

https://doi.org/10.1371/journal.pone.0115781.g004

The chimeric Tn21-Tn1721 backbone present in the CRL of RD1

The chimeric backbone structure seen in isolates 2011C-3943 and 2009EL-2050 has most likely formed by a homologous recombination event between two In2 derivatives, one of which is identical to that seen in plasmid pASL01a (NC_019091.1) [27] and the other found in Salmonella enterica plasmids like pYT2 (AB605179.1) or pSal8934b (JF274992) [34] (see S2A Fig.). The fact that the two copies of ΔtniA (ΔtniA and ΔtniA** seen in Fig. 1) found in the CRL can be traced to their exact locations in both plasmids supports this contention. pASL01a or variants of it are likely progenitors of the region encompassing the In2 integron because it has the 2082 bp (Fig. 2A) molecular signature seen in 35 of 53 sequenced EAHEC O104:H4 strains. In the outbreak strain 2011C-3943 the tetA gene has been interrupted by the insertion of IS181b at a time post homologous recombination in the Tn21-Tn1721 backbones, as 10 nucleotide direct repeats of the sequence at the insertion site of IS181b and directly abutting the inverted repeats of IS181b are clearly evident. The progenitor (Fig. 3) from which RD1 in O104:H4 isolates 2011C-3943 and 2009EL-2050 evolved was likely a hybrid structure that included an In2 derivative as the insertion site of the class 1 integron on the Tn21 backbone is characteristic of In2 integrons. Our BLASTn analysis of the O104:H4 microbial genomic database using overlapping fragments 4 and 5 (see Fig. 3) which span regions of RD1 containing the Tn21-specific and the Tn1721-derived modules of CRL (Table 1) provide evidence that these hybrid structures are frequently present in O104:H4 EAHEC.

Discussion

Multiple comparative genomic studies on both finished and draft genomes of EAHEC O104:H4 [4,35] indicate that RD1 within GI3 is a hotspot for micro-evolutionary events mediated by mobile genetic elements, particularly IS26 [2]. The preponderance of IS26 in GI3 indicates that clinically-relevant antibiotic resistance genes known to be mobilised by IS26 including blaSHV-11, blaSHV-12 (β-lactam resistance) [11,36,37], qnrB19 (quinolone resistance) [14,36] and aphA1 (kanamycin/neomycin resistance) [15] can readily recombine within the CRL in O104:H4. Furthermore, homologous recombination events involving IS26 is well known to drive the transfer of genetic loci between chromosome and plasmid backbones [11,38], and has played a central role in the evolution of new transposon-like structures (This study, [15,40]). As a consequence, evolution of resistance to clinically important antibiotics in O104:H4 should be monitored.

Our analyses indicate that two separate evolutionary scenarios (Fig. 4) shaped the CRL in EAHEC O104:H4 isolates 2011C-3943 and 2009EL-2050. The formation of a chimeric backbone structure formed by a homologous recombination event between two In2 derivatives is a key event in the creation of a Tn21/Tn1721 hybrid transposon that represented a progenitor structure shown in Fig. 3. The CRL seen in 2009EL-2050 and 2011C-3493 evolved separately from Tn6029B and Tn6029C respectively. The sequence of eight nucleotides adjacent to IS26 elements in the CRL shown in Fig. 2B provides irrefutable evidence that IS26 played a key role in the evolution of the CRL seen in EAHEC strains 2011C-3493 and 2009EL-2050. The precise events that created the CRL in 2009EL-2050 is depicted in Fig. 4A and B.

In isolate 2009EL-2050, a copy of IS26 inserted into the tnpM gene of Tn21, creating an eight base sequence duplication (Fig. 2B) at the insertion site, in the proposed progenitor. This event created inverted copies of IS26 flanking the class 1 integron allowing this region to undergo an IS26-mediated inversion. Inversion events driven by inverted copies of IS26 have been previously described [15]. These events generated a novel transposon flanked by direct copies of IS26 which we have identified here as Tn6222. Tn6222 comprises two modules, one composed of a functional class 1 integron containing a dfrA7 gene cassette and another with the blaTEM-1 gene (Fig. 4A). The class 1 integron in transposon Tn6222 has the ability to assemble more resistance genes via the acquisition of resistance gene cassettes and may continue to evolve from the resistance phenotype imparted by its present configuration.

The CRL in 2011C-3493 has evolved by a separate set of events mediated by the insertion of an IS26 within the boundaries of Tn6029C. IS26 increasingly plays a key role in the evolutionary events that underpin the formation of CRL in genomic islands among a range of emerging pathogens [21,39]. We propose that a copy of IS26 inserted into the IRL of the existing IS26 that abuts the blaTEM-1 gene in Tn6029C in the proposed progenitor (Fig. 4B). A subsequent deletion event removed one copy of IS26, the blaTEM-1 gene and ΔCR2 generating a new derivative of the Tn6029-family, identified here as Tn6029D (Fig. 4B) This loss of blaTEM1 in the outbreak strain is consistent with the formation of a translocatable unit [40] formed as a consequence of the insertion of an IS26 at the IRL of an existing IS26 present in the progenitor structure. While Tn6029D is clearly related to Tn6029C by virtue of the 12-nucleotide signature sequence (TTTGCCCCATAT), the loss of blaTEM-1 is unique (Fig. 4B). Evidence that the insertion of IS26 has driven the creation of the CRL in 2011C-3493 is found in the eight base characteristic sequences shown in Fig. 2B.

The progenitor structure depicted in S2A Fig. was created by a double reciprocal cross-over between the common regions in two In2 derivatives, one of which is identical to plasmid pASL01a (NC_019091.1) [27] and the other identical to that in Salmonella enterica plasmids like pYT2 (AB605179.1) or pSal8934b (JF274992) [34]. Although a similar structure is also found in pHCM1, this plasmid is unlikely to be a progenitor of the structure seen in RD1 as it lacks a class 1 integron [15,30]. The insertion of IS26 generates eight base direct repeats [41] at the site of insertion and is known to induce deletions of variable length creating unique molecular signatures. We acknowledge that one of the limitations in our evolutionary models is that they are based on polished genome sequences of only two EAHEC isolates. To counter this, we performed extensive BLASTn analysis of 53 O104:H4 E. coli genomes in the microbial genome database. Our analyses confirm the presence of our proposed ancestral structures and identified possible variants of RD1 found in the genomes of 2011C-3493 and 2009EL-2050. In our opinion the major hindrance to rigorous assessment of evolutionary history for complex resistance loci or microbial genomes as a whole is not the quality of available genome data but a lack of statistical models for the inference of genome evolution and the availability of computational tools implementing the models. As genomes evolve their loci undergo rearrangement, insertions, deletions, segmental duplication and lateral transfer. Although efforts to combine these events into a unified statistical model of evolutionary history are ongoing [42,43], no software implementing such a unified inference model is yet available.

The 2082 bp molecular signature has been observed in pASL01a (JQ480155.1), in the CRL of EAHEC O104:H4 German outbreak strain 2011C-3493 (CP003289.1) and strains 2009EL-2050 (CP003297.1) and 2009EL-2071 (CP003301.1) from patients with bloody diarrhoea in the Republic of Georgia, and in a clinical E. coli isolate (HM999792.1) from Sydney, Australia. In strain 2009EL-2071 the CRL in RD1 has been lost [2]. RD1 is located in GI3 adjacent to the selC tRNA gene in EAHEC [3]. GI3 likely presents an evolutionary advantage to strains that carry it because it carries ag43, a gene encoding the self-associating serine protease auto-transporter of Enterobacteriaceae (SPATE) that influences biofilm formation in EAHEC O104:H4 [44] and the yeeV/yeeU toxin/antitoxin system [35]. This view is supported by a study showing that kidney damage and virulence gene expression (stx2, aggR and pgaA) correlates with the ability of O104:H4 to form biofilms in germ-free mice [45]. GI3 is potentially mobile [12] and targets the 23 bp genomic sequences that are widely dispersed in E. coli genomes, suggesting it may be a significant player in future emerging pathogens.

Supporting Information

S1 Fig. Diagrammatic guides to interpret the physical locations of the nucleotide alignment start positions (BLASTn analysis of fragment 3) along the sequence corresponding to the Tn6029-family transposon seen in isolate 2009EL-2050.

Different segments of the transposon have aligned to the different scaffolds of the draft genome sequences in the BLASTn analysis of the fragment 3 (Fig. 3) against draft E. coli O104:H4 genomes available in GenBank.

https://doi.org/10.1371/journal.pone.0115781.s001

(TIF)

S2 Fig. A: Homologous recombination events that have shaped the CRL in the theoretical progenitor structure.

The homologous recombination event has likely taken place between two multiple antibiotic resistance plasmids pASL01a and plasmids from Salmonella like pYT2. The event created a CRL encoding dfrA7 (resistance to trimethoprim), blaTEM-1 (resistance to ampicillin), strAB (resistance to streptomycin), sul2 (resistance to sulfamethoxazole), qacE∆1 (resistance to quaternary compounds), merA (resistance to mercury chloride) and tet(A)A (resistance to tetracycline). B: A cartoon depicting the structural differences and genetic signatures present in members of Tn6029-family of transposons, i.e. Tn6029, Tn6029B and Tn6029C. The top panel shows the structure of Tn6029 and the middle panel shows that Tn6029B is characterised by an 85 bp deletion in the blaTEM module (see asterix). In Tn6029C (bottom panel), the blaTEM containing module is orientated in the reverse direction compared to Tn6029 / Tn6029B and has lost an additional 11 bp from the deleted-CR2 region (indicated by the asterix). Tn6029C is the arrangement seen in pASL01a.

https://doi.org/10.1371/journal.pone.0115781.s002

(TIF)

S1 Table. Results of BLASTn analysis of E coli O104:H4 draft genomes using Fragment 1 (5730nt).

https://doi.org/10.1371/journal.pone.0115781.s003

(DOCX)

S2 Table. Results of BLASTn analysis of E coli O104:H4 draft genomes using Fragment 2 (3298nt).

https://doi.org/10.1371/journal.pone.0115781.s004

(DOCX)

S3 Table. Results of BLASTn analysis of E coli O104:H4 draft genomes using Fragment 3 (8677nt).

https://doi.org/10.1371/journal.pone.0115781.s005

(DOCX)

S4 Table. Results of BLASTn analysis of E coli O104:H4 draft genomes using Fragment4 (4670nt).

https://doi.org/10.1371/journal.pone.0115781.s006

(DOCX)

S5 Table. Results of BLASTn analysis of E coli O104:H4 draft genomes using Fragment 5 (5301nt).

https://doi.org/10.1371/journal.pone.0115781.s007

(DOCX)

Acknowledgments

We thank Aaron Darling for his comments on an early draft of this manuscript.

Author Contributions

Conceived and designed the experiments: PRC SPD. Performed the experiments: PRC. Analyzed the data: PRC. Contributed reagents/materials/analysis tools: PRC. Wrote the paper: PRC SPD IGC. Data presentation, figures, literature searching: PRC.

References

  1. 1. Beutin L, Martin A (2012) Outbreak of Shiga toxin-producing Escherichia coli (STEC) O104:H4 infection in Germany causes a paradigm shift with regard to human pathogenicity of STEC strains. J Food Prot 75: 408–418. pmid:22289607
  2. 2. Ahmed SA, Awosika J, Baldwin C, Bishop-Lilly KA, Biswas B, et al. (2012) Genomic comparison of Escherichia coli O104:H4 isolates from 2009 and 2011 reveals plasmid, and prophage heterogeneity, including shiga toxin encoding phage stx2. PLoS One 7: e48228. pmid:23133618
  3. 3. Grad YH, Godfrey P, Cerquiera GC, Mariani-Kurkdjian P, Gouali M, et al. (2013) Comparative genomics of recent Shiga toxin-producing Escherichia coli O104:H4: short-term evolution of an emerging pathogen. MBio 4: e00452–00412. pmid:23341549
  4. 4. Grad YH, Lipsitch M, Feldgarden M, Arachchi HM, Cerqueira GC, et al. (2012) Genomic epidemiology of the Escherichia coli O104:H4 outbreaks in Europe, 2011. Proc Natl Acad Sci U S A 109: 3065–3070. pmid:22315421
  5. 5. Gault G, Weill FX, Mariani-Kurkdjian P, Jourdan-da Silva N, King L, et al. (2011) Outbreak of haemolytic uraemic syndrome and bloody diarrhoea due to Escherichia coli O104:H4, south-west France, June 2011. Euro Surveill 16. pmid:22221498
  6. 6. Jourdan-da Silva N, Watrin M, Weill FX, King LA, Gouali M, et al. (2012) Outbreak of haemolytic uraemic syndrome due to Shiga toxin-producing Escherichia coli O104:H4 among French tourists returning from Turkey, September 2011. Euro Surveill 17. pmid:23241235
  7. 7. Mariani-Kurkdjian P, Bingen E, Gault G, Jourdan-Da Silva N, Weill FX (2011) Escherichia coli O104:H4 south-west France, June 2011. Lancet Infect Dis 11: 732–733. pmid:21958580
  8. 8. Kunne C, Billion A, Mshana SE, Schmiedel J, Domann E, et al. (2012) Complete sequences of plasmids from the hemolytic-uremic syndrome-associated Escherichia coli strain HUSEC41. J Bacteriol 194: 532–533. pmid:22207742
  9. 9. Riley LW (2014) Pandemic lineages of extraintestinal pathogenic Escherichia coli. Clin Microbiol Infect 20: 380–390. pmid:24766445
  10. 10. Baquero F, Tobes R (2013) Bloody coli: a gene cocktail in Escherichia coli O104:H4. MBio 4: e00066–00013. pmid:23422408
  11. 11. Hudson CM, Bent ZW, Meagher RJ, Williams KP (2014) Resistance determinants and mobile genetic elements of an NDM-1-encoding Klebsiella pneumoniae strain. PLoS One 9: e99209. pmid:24905728
  12. 12. Chouikha I, Germon P, Bree A, Gilot P, Moulin-Schouleur M, et al. (2006) A selC-associated genomic island of the extraintestinal avian pathogenic Escherichia coli strain BEN2908 is involved in carbohydrate uptake and virulence. J Bacteriol 188: 977–987. pmid:16428402
  13. 13. Blum G, Ott M, Lischewski A, Ritter A, Imrich H, et al. (1994) Excision of large DNA regions termed pathogenicity islands from tRNA-specific loci in the chromosome of an Escherichia coli wild-type pathogen. Infect Immun 62: 606–614. pmid:7507897
  14. 14. Cain AK, Hall RM (2012) Evolution of a multiple antibiotic resistance region in IncHI1 plasmids: reshaping resistance regions in situ. J Antimicrob Chemother 67: 2848–2853. pmid:22888274
  15. 15. Cain AK, Liu X, Djordjevic SP, Hall RM (2010) Transposons related to Tn1696 in IncHI2 plasmids in multiply antibiotic resistant Salmonella enterica serovar Typhimurium from Australian animals. Microb Drug Resist 16: 197–202. pmid:20701539
  16. 16. Venturini C, Beatson SA, Djordjevic SP, Walker MJ (2010) Multiple antibiotic resistance gene recruitment onto the enterohemorrhagic Escherichia coli virulence plasmid. FASEB J 24: 1160–1166. pmid:19917674
  17. 17. Venturini C, Hassan KA, Roy Chowdhury P, Paulsen IT, Walker MJ, et al. (2013) Sequences of Two Related Multiple Antibiotic Resistance Virulence Plasmids Sharing a Unique IS26-Related Molecular Signature Isolated from Different Escherichia coli Pathotypes from Different Hosts. PLoS One 8: e78862. pmid:24223859
  18. 18. Bailey JK, Pinyon JL, Anantham S, Hall RM (2011) Distribution of the blaTEM gene and blaTEM-containing transposons in commensal Escherichia coli. J Antimicrob Chemother 66: 745–751. pmid:21393132
  19. 19. Dawes FE, Kuzevski A, Bettelheim KA, Hornitzky MA, Djordjevic SP, et al. (2010) Distribution of class 1 integrons with IS26-mediated deletions in their 3'-conserved segments in Escherichia coli of human and animal origin. PLoS One 5: e12754. pmid:20856797
  20. 20. Djordjevic SP, Stokes HW, Roy Chowdhury P (2013) Mobile elements, zoonotic pathogens and commensal bacteria: conduits for the delivery of resistance genes into humans, production animals and soil microbiota. Front Microbiol 4: 86. pmid:23641238
  21. 21. Doublet B, Praud K, Weill FX, Cloeckaert A (2009) Association of IS26-composite transposons and complex In4-type integrons generates novel multidrug resistance loci in Salmonella genomic island 1. J Antimicrob Chemother 63: 282–289. pmid:19074421
  22. 22. Lee KY, Hopkins JD, Syvanen M (1990) Direct involvement of IS26 in an antibiotic resistance operon. J Bacteriol 172: 3229–3236. pmid:2160941
  23. 23. Naas T, Mikami Y, Imai T, Poirel L, Nordmann P (2001) Characterization of In53, a class 1 plasmid- and composite transposon-located integron of Escherichia coli which carries an unusual array of gene cassettes. J Bacteriol 183: 235–249. pmid:11114922
  24. 24. Post V, White PA, Hall RM (2010) Evolution of AbaR-type genomic resistance islands in multiply antibiotic-resistant Acinetobacter baumannii. J Antimicrob Chemother 65: 1162–1170. pmid:20375036
  25. 25. Roy Chowdhury P, Ingold A, Vanegas N, Martinez E, Merlino J, et al. (2011) Dissemination of multiple drug resistance genes by class 1 integrons in Klebsiella pneumoniae isolates from four countries: a comparative study. Antimicrob Agents Chemother 55: 3140–3149. pmid:21518841
  26. 26. Ruiz E, Rojo-Bezares B, Saenz Y, Olarte I, Esteban I, et al. (2010) Outbreak caused by a multi-resistant Klebsiella pneumoniae strain of new sequence type ST341 carrying new genetic environments of aac(6')-Ib-cr and qnrS1 genes in a neonatal intensive care unit in Spain. Int J Med Microbiol 300: 464–469. pmid:20547103
  27. 27. Labar AS, Millman JS, Ruebush E, Opintan JA, Bishar RA, et al. (2012) Regional dissemination of a trimethoprim-resistance gene cassette via a successful transposable element. PLoS One 7: e38142. pmid:22666464
  28. 28. Szczepanowski R, Braun S, Riedel V, Schneiker S, Krahn I, et al. (2005) The 120 592 bp IncF plasmid pRSB107 isolated from a sewage-treatment plant encodes nine different antibiotic-resistance determinants, two iron-acquisition systems and other putative virulence-associated functions. Microbiology 151: 1095–1111. pmid:15817778
  29. 29. Darling AC, Mau B, Blattner FR, Perna NT (2004) Mauve: multiple alignment of conserved genomic sequence with rearrangements. Genome Res 14: 1394–1403. pmid:15231754
  30. 30. Wain J, Diem Nga LT, Kidgell C, James K, Fortune S, et al. (2003) Molecular analysis of incHI1 antimicrobial resistance plasmids from Salmonella serovar Typhi strains associated with typhoid fever. Antimicrob Agents Chemother 47: 2732–2739. pmid:12936967
  31. 31. Partridge SR, Brown HJ, Stokes HW, Hall RM (2001) Transposons Tn1696 and Tn21 and their integrons In4 and In2 have independent origins. Antimicrob Agents Chemother 45: 1263–1270. pmid:11257044
  32. 32. Holt KE, Thomson NR, Wain J, Phan MD, Nair S, et al. (2007) Multidrug-resistant Salmonella enterica serovar paratyphi A harbors IncHI1 plasmids similar to those found in serovar typhi. J Bacteriol 189: 4257–4264. pmid:17384186
  33. 33. Yau S, Liu X, Djordjevic SP, Hall RM (2010) RSF1010-like plasmids in Australian Salmonella enterica serovar Typhimurium and origin of their sul2-strA-strB antibiotic resistance gene cluster. Microb Drug Resist 16: 249–252. pmid:20617928
  34. 34. Tamamura Y, Tanaka K, Akiba M, Kanno T, Hatama S, et al. (2013) Complete Nucleotide Sequences of Virulence-Resistance Plasmids Carried by Emerging Multidrug-Resistant Salmonella enterica Serovar Typhimurium Isolated from Cattle in Hokkaido, Japan. PLoS One 8: e77644. pmid:24155970
  35. 35. Guy L, Jernberg C, Arven Norling J, Ivarsson S, Hedenstrom I, et al. (2013) Adaptive Mutations and Replacements of Virulence Traits in the Escherichia coli O104:H4 Outbreak Population. PLoS One 8: e63027. pmid:23675451
  36. 36. Dionisi AM, Lucarelli C, Owczarek S, Luzzi I, Villa L (2009) Characterization of the plasmid-borne quinolone resistance gene qnrB19 in Salmonella enterica serovar Typhimurium. Antimicrob Agents Chemother 53: 4019–4021. pmid:19528272
  37. 37. Ford PJ, Avison MB (2004) Evolutionary mapping of the SHV beta-lactamase and evidence for two separate IS26-dependent blaSHV mobilization events from the Klebsiella pneumoniae chromosome. J Antimicrob Chemother 54: 69–75. pmid:15163647
  38. 38. Osborn AM, Bruce KD, Strike P, Ritchie DA (1997) Distribution, diversity and evolution of the bacterial mercury resistance (mer) operon. FEMS Microbiol Rev 19: 239–262. pmid:9167257
  39. 39. Siebor E, Neuwirth C (2013) Emergence of Salmonella genomic island 1 (SGI1) among Proteus mirabilis clinical isolates in Dijon, France. J Antimicrob Chemother 68: 1750–1756. pmid:23580563
  40. 40. Harmer CJ, Moran RA, Hall RM (2014) Movement of IS26-Associated Antibiotic Resistance Genes Occurs via a Translocatable Unit That Includes a Single IS26 and Preferentially Inserts Adjacent to Another IS26. MBio 5. pmid:25538196
  41. 41. Mollet B, Iida S, Arber W (1985) Gene organization and target specificity of the prokaryotic mobile genetic element IS26. Mol Gen Genet 201: 198–203. pmid:3003524
  42. 42. Boussau B, Szollosi GJ, Duret L, Gouy M, Tannier E, et al. (2013) Genome-scale coestimation of species and gene trees. Genome Res 23: 323–330. pmid:23132911
  43. 43. Sjostrand J, Tofigh A, Daubin V, Arvestad L, Sennblad B, et al. (2014) A Bayesian method for analyzing lateral gene transfer. Syst Biol 63: 409–420. pmid:24562812
  44. 44. Vejborg RM, Klemm P (2009) Cellular chain formation in Escherichia coli biofilms. Microbiology 155: 1407–1417. pmid:19383712
  45. 45. Al Safadi R, Abu-Ali GS, Sloup RE, Rudrik JT, Waters CM, et al. (2012) Correlation between in vivo biofilm formation and virulence gene expression in Escherichia coli O104:H4. PLoS One 7: e41628. pmid:22848550