Transposon Tn7 Is Widespread in Diverse Bacteria and Forms Genomic Islands^,

Abstract

We find that relatives of the bacterial transposon Tn7 are widespread in disparate environments and phylogenetically diverse species. These elements form functionally diverse genomic islands at the specific site of Tn7 insertion adjacent to glmS. This work presents the first example of genomic island formation by a DDE type transposon.

Tn7 is a 14-kb bacterial transposon that was originally discovered in Escherichia coli. This transposon is tightly regulated and activates transposition only when specific targets are found in the cell. Tn7 can use two transposition pathways that recognize different types of target sites with distinct but overlapping sets of transposon-encoded proteins, TnsA, TnsB, TnsC, TnsD, and TnsE (TnsABCDE) (Fig. 1) (reviewed in reference 6). The TnsABC proteins constitute the core transposition machinery that interacts with one of the two target site-selecting proteins, TnsD or TnsE, to carry out transposition. TnsB is a member of the retroviral integrase superfamily containing the characteristic “DDE” motif, composed of aspartic acid and glutamic acid residues that form the active site of the transposase (23). TnsD is a sequence-specific DNA binding protein that recognizes a sequence in the glmS gene that encodes the C terminus of glucosamine-6-phosphate synthase. TnsD-mediated transposition is directed at a high frequency into the attTn7 site, in the transcriptional terminator for the glmS gene, and has no detectable negative effect on the host. TnsE does not recognize any particular DNA sequence but preferentially directs transposition into mobile plasmids by recognizing an aspect of transfer-associated DNA replication (17, 18, 27). Presumably, the TnsE-mediated pathway would facilitate the dissemination of Tn7 to new hosts, while the TnsD-mediated pathway would provide a “safe haven” once in a new bacterial host. As described below, we find multiple elements in diverse bacteria that are clearly Tn7 relatives. They contain homologs of all five Tn7 transposition proteins and are almost always inserted into the attTn7 site of the host organism. These elements appear to facilitate the accumulation of mobile DNA within the attTn7 site in the chromosome. The accumulated DNA observed within the attTn7 sites appears to fit the definition of genomic islands, i.e., large segments (>30 kb) of a bacterial genome that have been acquired by horizontal transfer and confer fitness-enhancing qualities (10).

FIG. 1.

attTn7 locus in E. coli and Shewanella spp. (a) Representation of Tn7 in the attTn7 site of E. coli (6). (b) The attTn7 locus in Shewanella oneidensis MR-1 (AE014299) (9, 12) and in six other Shewanella species: Shewanella baltica OS195 (AATK00000000), Shewanella sp. strain MR-4 (CP000446), Shewanella sp. strain W3-19-1 (AALN01000053), Shewanella sp. strain MR-7 (AALI01000045), Shewanella sp. strain SAR-1 (AACY01051759), and Shewanella sp. strain ANA-3 (AALH01000050). (c) A single Tn7-like insertion in Shewanella loihica (PV-4) (AAL201000023). (d) A large Tn7-like insertion in S. baltica OS155 (AAIO01000013, AAIO01000122). (e) Two Tn7-like insertions in Shewanella putrefaciens CN-32 (AALB01000024, AALB01000006). (f) A possible degenerate Tn7-like insertion in S. denitrificans OS-217 (CP000302). Arrows indicate ORFs. Open arrows indicate canonical Tn7 genes, crosshatched arrows indicate noncanonical Tn7 genes found between the left and right ends of the elements, and filled arrows indicate genes found outside the ends of the elements. Ovals indicate the left (L) and right (R) ends of the elements. Dashed lines indicate additional open reading frames that are not shown. Dotted lines indicate the presumed course of genomic island evolution. The open reading frames were arbitrarily named orfY and orfZ simply to indicate homologous genes in the Shewanella spp.

Since the TnsE pathway is used to preferentially transfer Tn7 to mobile plasmids, we expected Tn7-like transposons to be present in a wide variety of hosts. Indeed, other reports have described clinically isolated β- and ɛ-proteobacteria containing Tn7 (8, 20). Laboratory experiments have shown that Tn7 is capable of transposition in many different hosts, but few naturally occurring examples have been isolated (7). We also suspected that Tn7 relatives might be present in organisms from diverse habitats, given that the Tn7-like transposon Tn5468 had been found in the attTn7 site of the acidophilic organism Acidithiobacillus ferrooxidans ATCC 33020 (16). To address these hypotheses, we used the BLAST and PSI-BLAST algorithms to query the GenBank and Comprehensive Microbial Resource databases, using the amino acid sequence of TnsE (2-4, 19). Our rationale was that TnsE is essential for moving between bacteria, and it is the gene most distal from glmS that is necessary for transposition. We found TnsE homologs in bacterial hosts from the γ-proteobacteria, δ-proteobacteria, and low-G+C, gram-positive firmicutes (Table 1; see also the supplemental material). The bacterial hosts were from both terrestrial environments and marine environments ranging from surface waters at various latitudes to deep-sea hydrothermal vents.

TABLE 1.

Tn7 and related elements identified in this study^a

Organism	Lineage	Site(s) of insertion	Size(s)	Notable function(s) encoded by the element	Source or GenBank accession no.	Reference(s)
Escherichia colib	γ-Proteobacteria	attTn7, plasmidc	14 kb	Antibiotic resistance	AP002527	11
Shigella sonnei Ss046b	γ-Proteobacteria	attTn7	14 kb	Antibiotic resistance	CP000038	28
Raoultella terrigenab	γ-Proteobacteria	unknown	NA (>13 kb)	Antibiotic resistance	DQ268533
Enterobacter cloacaeb	γ-Proteobacteria	unknown	NA (>13 kb)	Antibiotic resistance	DQ275532
Pseudomonas aerugenosab	γ-Proteobacteria	plasmid, unknownc	NA (>13 kb)	Antibiotic resistance	AM261760, DQ176869
Shewanella putrefaciens CN-32	γ-Proteobacteria	attTn7, attTn7c	21 Kb, NA (>130 kb)	Heavy metal resistance	AALB01000024, AALB01000006b
Shewanella baltica OS155	γ-Proteobacteria	attTn7	NA (>80 kb)	Type III secretion system	AAIO01000013, AAIO01000122
Shewanella loihica PV-4	γ-Proteobacteria	attTn7	21 kb	Transposases, putative hydrolase	AALS01000023
Idiomarina loihiensis L2TR	γ-Proteobacteria	attTn7	16 kb	DNA-methyltransferase	AE017340	13
Acidithiobacillus ferrooxidans ATCC 23270	γ-Proteobacteria	attTn7, terminusc	39 kb, 31 kb	Restriction modification, oxido-reductase	CMRd	16, 19
Hahella chejuensis KCTC 2396	γ-Proteobacteria	attTn7	46 kb	DNA repair-associated proteins	CP000155	14
Pelobacter carbinolicus DSM 2380	δ-Proteobacteria	attTn7	25 kb	Restriction modification system	CP000142
Bacillus cereus ATCC 10987	Firmicutes	attTn7, attTn7c	27 kb, 11 kb	Bacteriocin transporter, DNA helicase	AE017194	21
Staphylococcus sp. strain 639-2	Firmicutes	Plasmid	14 kb	Antibiotic resistance	DQ390456	26
Environmental	Unknown	Unknown	NA	Unknown	AACY01193247, AACY01164117, AACY01058851, AACY01058852, AACY01177854, AACY01101269, AAFX01066149	25

^aNA, not available (the sequence from GenBank does not extend to the ends of the element to definitively determine the size of the element; a minimum size is shown in parentheses).

^bAs far as they are sequenced, elements are identical to Tn7 as originally described for E. coli.

^cMore than one element is present in a given host.

^dCMR, Comprehensive Microbial Resource database.

To confirm that the TnsE homologs were indeed contained within Tn7-like elements, we used the Artemis Comparison Tool (release 4) to observe the synteny of tnsABCDE genes and host genes flanking the attTn7 locus (1, 5). We found that the order and orientation of transposon genes are highly conserved, although there are some examples in which other open reading frames (ORFs) have been inserted between the tns genes (Fig. 1; see also the supplemental material). None of the tns genes had nonsense mutations that might prevent these elements from remaining active. In every case, the tnsABCD genes could be found 5′ of tnsE, where sufficient DNA sequence information was present for analysis. Most of the Tn7-like elements were found in the site that was equivalent to the attTn7 site as found in E. coli. Unexpectedly, there were multiple examples where two nonidentical Tn7-like transposons had been inserted in tandem into the attTn7 locus. These tandem insertions are very likely the result of independent transposition events. While Tn7 is typically discouraged from inserting more than one element into the attTn7 locus by target site immunity, it is likely that differences in the transposon ends and TnsBC proteins from the nonidentical Tn7-like elements diminished the robustness of this process (6, 24). We found three previously undescribed Tn7 relatives that were in genomic locations that are likely to have been targeted by TnsE. These were found in plasmids, and one was found in a putative terminus region (Table 1). In E. coli, the region of the chromosome where DNA replication terminates has been shown to be an alternate TnsE target (17).

To further characterize the genetic diversity found within the Tn7-like transposons, we used Artemis (release 7) and the Artemis Comparison Tool to search the DNA sequences for the 5-bp target site duplications and the TnsB binding sites associated with the Tn7 end sequences (1, 5, 15, 22). Once the ends of each element had been identified, we classified the genes between the left and right transposon ends by their annotated functions (Fig. 1; see also the supplemental material). None of the elements contained the same complement of genes, aside from tnsABCDE. The annotated functions of the genes from each of the elements were diverse; however, DNA restriction and modification systems were common.

Because multiple nonidentical Tn7 relatives were found in Shewanella species, we also analyzed the contents of the putative attTn7 loci in the available Shewanella genomic sequences by searching for glmS and menB homologs, genes that define the attTn7 site found in this genus (Fig. 1). Out of six Shewanella species that had genes inserted into the attTn7 site, three contained Tn7-like elements, one contained repeats that may have been a degenerate Tn7 end (Fig. 1), and two contained no identifiable Tn7 components (Shewanella frigidmarina NCIMB 400 [CP000447] and Shewanella amazonensis SB2B [AAIN01000058]). The genetic contents of the attTn7 site in multiple species match the description of genomic islands. We found multiple examples of genes that are typically associated with genomic or pathogenicity islands, such as type III secretion systems, bacteriophage-related genes, non-Tn7-like transposases, heavy metal detoxification genes, and DNA restriction and modification systems. Genomic islands commonly form proximal to tRNA genes, presumably through the integration of bacteriophage or integrative conjugal elements (formerly called conjugative transposons) (10). Bacteriophage and integrative conjugal elements use conservative site-specific recombinases that require DNA sequence homology between target and donor molecules. We propose that Tn7 is able to initiate genomic island formation by a transposition process that requires no DNA sequence homology.

From the examples of Tn7 relatives shown here, especially those found within the Shewanella species, we can suggest a sequence of events that would lead from one insertion to the localization of genomic islands in the attachment site and that may lack almost all recognizable features of Tn7 (Fig. 1). One Tn7 relative may operate as a “founder element” that is able to locate and safely transpose into the bacterial chromosome, bringing with it other mobile elements and possibly attachment sites for bacteriophages, integrons, or other transposons. After multiple Tn7 transposition events accumulate in the attachment site, recombination between these elements may enhance evolution by reassortment (e.g., Fig. 1c to e). If either the left or the right end of Tn7 is lost, the compromised element would presumably be subject to reductive evolution and only the most highly selected components would remain (e.g., Shewanella denitrificans OS-217 in Fig. 1f). The formation of genomic islands in the attTn7 site is likely a collaborative process that draws on the ability of the TnsD pathway to target this highly conserved “safe site” within the chromosome and the ability of the TnsE pathway to move the element into the mobile DNA pool.