Functional organization and insertion specificity of IS607, a chimeric element of Helicobacter pylori

Abstract

A search by subtractive hybridization for sequences present in only certain strains of Helicobacter pylori led to the discovery of a 2-kb transposable element to be called IS607, which further PCR and hybridization tests indicated was present in about one-fifth of H. pylori strains worldwide. IS607 contained two open reading frames (ORFs) of possibly different phylogenetic origin. One ORF (orfB) exhibited protein-level homology to one of two putative transposase genes found in several other chimeric elements including IS605 (also of H. pylori) and IS1535 (of Mycobacterium tuberculosis). The second IS607 gene (orfA) was unrelated to the second gene of IS605 and might possibly be chimeric itself: it exhibited protein-level homology to merR bacterial regulatory genes in the first approximately 50 codons and homology to the second gene of IS1535 (annotated as "resolvase," apparently due to a weak short recombinase motif) in the remaining three-fourths of its length. IS607 was found to transpose in Escherichia coli, and analyses of sequences of IS607-target DNA junctions in H. pylori and E. coli indicated that it inserted either next to or between adjacent GG nucleotides, and generated either a 2-bp or a 0-bp target sequence duplication, respectively. Mutational tests showed that its transposition in E. coli required orfA but not orfB, suggesting that OrfA protein may represent a new, previously unrecognized, family of bacterial transposases.

FIG. 1

IS607 and related IS elements, including possibly chimeric origin of orfA. (A) Overview of structures. Boxes represent ORFs. These ORFs and portions of ORFs from different elements represent protein-level homologies (in range of 25% to 37% protein level identity in BLASTP [http://www.ncbi.nlm.nih.gov/blast/blast.cgi]). Of particular note are protein-level homologies between 5′ 60 amino acids of IS607 OrfA (accession no. AF189015) and entries annotated as MerR-type bacterial regulatory proteins (33% match, first 51 amino acids in the case of HI1623 [B]) and remainder of OrfA with putative resolvase of IS1535 of M. tuberculosis (C). The ORFs of MJ0014 and MJ0013 depicted here were annotated as function unknown (accession no. U67460). MJ0014 shares 37% identity with IS607 orfA overall (D), whereas MJ0013 is not related to IS607 orfB. (B) BLASTP alignment of N terminus of IS607 OrfA protein with MerR-type bacterial regulatory protein HI1623 from Haemophilus influenzae (accession no. C64133; 33% identity; best current match to this domain of IS607 OrfA). Motif searches using the pfam 5.3 program identified a match of IS607 OrfA to MerR regulatory protein family motifs in positions 9 to 44 with an E value of 3 e⁻¹⁰ (highly significant match). (C) BLASTP alignment of remainder of IS607 OrfA with corresponding region of the IS1535 resolvase (accession no. A70583; 34% identity). Also among entries with close homology to IS607 OrfA are proteins annotated as resolvase or resolvase-related from other species (e.g., an Acidianus ambivalens protein, accession no. CAB58176, 33% identity throughout; PAB2076 of Pyrococcus abyssi, accession no. B75156, 35% identity throughout). Motif searches using the pfam 5.3 program identified a match of IS607 OrfA to site-specific recombination protein at positions 66 to 76 with an E value of 0.0011, which, although better than that obtained using the IS1535 resolvase (E value of 0.22), is rather weak. (By contrast, IS607 OrfB was matched through much of its sequence to that of a canonical transposase family [positions 23 to 300; pfam E value of 7 e⁻⁵⁰].) (D) BLASTP alignment of IS607 OrfA with M. jannaschii MJ0014, an ORF annotated as function unknown in the genome sequence (best current match to IS607 OrfA) (http://www.tigr.org/tdb/CMR/arg/htmls/SplashPage.html). It is noteworthy that the product of the adjacent gene (MJ0013), although related to putative transposases in BLASTP searches, is not detectably related to IS607 OrfB protein.

FIG. 2

Diagram of plasmids containing IS607 and related elements generated for this study, as detailed in Materials and Methods. A and B refer to orfA and orfB, respectively. The left (orfA) and right (orfB) ends of these elements are represented by the filled and patterned boxes, respectively. The average frequencies of Cam^r Str^r exconjugants and yield of Amp^s isolates among Cam^r Str^r exconjugants listed were based on a number of conjugation assays, typically using several different constructs of the depicted plasmids, to avoid being misled by mutation during PCR or gene cloning, as follows: plasmid 1, 24 assays, two different constructs; plasmid 2, 6 assays, three different constructs; plasmid 3, 24 assays, five different constructs; plasmid 4, 15 assays, five different constructs; plasmid 5, 10 assays, five different constructs; plasmid 6, 8 assays, one construct; plasmids 7a and 7b, 12 assays, four different constructs (two constructs with each frameshift [7a,b]); plasmid 8, 20 assays, six different constructs.

FIG. 3

Terminal sequences of IS607 and related elements and sites of insertion. (A) Termini and sites of IS607 insertion in H. pylori strain CPY0041. IS607 ends are in lowercase letters, and flanking sequences are in capital letters. This presentation assumes 0-bp target sequence duplication (see panel E). Insertion sites are defined relative to corresponding genes in fully sequenced genomes of strains 26695 (HP numbers [27]) and J99 (jhp numbers [2]). The following insertions were identified by sequencing in H. pylori strain CPY0041: clone 2a, insertion at position 3158/9 bp in a gene designated jhp1409, a type II DNA modification enzyme (methyltransferase) in strain J99; clone 22-4, insertion at 64/5 bp in a gene designated jhp1207 in J99 and HP1287 in 26695, a putative transcription regulator; and clone 8-3, insertion at position 2509/10 bp in a gene designated jhp1044 in J99, which has 70.8% homology with the shorter HP1116 gene of 26695 (1,154 codons versus 957 codons), which are each considered hypothetical function-unknown genes. (B) Termini and sites of IS607cam insertions in the F factor. The transposition donor plasmids (Fig. 2) from which the transposed IS607cam elements derived and nucleotide positions of insertions are as follows. Three insertions were derived from plasmid 1: 2-4, position 62046/5 bp, accession no. AF074613; 3-3, position 28627/6 bp, accession no. U01159; and 3-7, position 5996/5 bp, accession no. U01159. Three insertions were derived from plasmid 3: 5-2, position 12552/1 bp, accession no. U01159; 15-1, position 11046/5 bp, accession no. AF074613; and 15-3, position 65685/6 bp, accession no. AF074613. Four insertions were derived from plasmid 4: 1-3, position 633/4 bp, accession no. AF106329; 4-1, position 5941/2 bp, accession no. U01159; 6-1, position 3771/70 bp, accession no. AF106329; and 6-3, position 57885/6 bp, accession no. AF074613. (C) Insertions of four IS607-related elements in the genome of M. tuberculosis, to illustrate a possible insertion preference for GG dinucleotide targets (taken from http://www.sanger.ac.uk/Projects/M_tuberculosis/). (D) Left and right ends of IS607 (same strand), aligned to highlight inverted repeat motifs (capital letters). (E) Alternative models for IS607 termini and target site fates: (i) 0-bp target duplication model implies that ends of IS607 are 5′-gcta… and …aacg-3′; or (ii) 2-bp target duplication model implies that IS607 is 1 bp shorter at each end. The sequence used in this illustration is that of insertion 2a in panel A.

FIG. 4

Nine-codon overlap between IS607 orfA and orfB, and strategies that generate fusions. Line 1, carboxy-terminal amino acid sequence of OrfA protein; line 2, amino-terminal amino acid sequence of OrfB protein; line 3, corresponding DNA sequence (sense strand); line 4, mutagenic primer Mut2 (−1 frameshift in first codon of overlap; antisense DNA strand); line 5, fusion protein sequence resulting from −1 frameshift in first codon of overlap; line 6, mutagenic primer Mut1 (−1 frameshift in ninth codon of overlap; antisense DNA strand); line 7, fusion protein sequence resulting from −1 frameshift in ninth codon of overlap.

FIG. 5

Structures of products of transposition, with diagram of inferred sequential action of OrfA protein on site nearest its site of synthesis, followed by action on a second appropriate site. Donor plasmid numbers (1, 2, and 4) refer to designations used in Fig. 2.