Introduction

Transposable elements and insertion sequences (ISs) that are capable of moving from a replicon to others are important tools in the study of bacterial genetics and gene expression. The distribution of these mobile elements in the genome is also of interest in the evolution of organisms. Some transposable elements and ISs in Streptomyces strains have been isolated [2, 4, 14, 19], and the distribution of ISs in the Streptomyces genome was recently reported [10]. Among transposons, the class-II transposable element Tn4556 was identified in neomycin-producing Streptomyces fradiae in 1987 [5, 6]. Tn4556 derivatives carrying antibiotic-resistance markers were found to be useful for transposition in Streptomyces strains [16, 22]. The most preferable derivative Tn4560, which contains the resistance marker viomycin phosphotransferase, was used to target genes involved in secondary metabolite biosynthesis [9]. The 6.8-kb sequence of Tn4556 was elucidated in 1990 [18] and at least nine ORFs have been annotated from sequence data; however, some overlapped in both strands. Thus, since the original sequence may contain errors, we resequenced Tn4556 to clarify actual ORFs and examined the efficient transposition of Tn4556.

Results and discussion

Resequencing of Tn4556

A 6.64-kbp BamHI fragment of Tn4556 in pUC1232 [5], which contains the entire Tn4556 sequence, was subcloned into pUC19 and the sequence was analyzed using a next-generation sequencer (accession # LC417441). Three positions differed from the previous sequence (accession # M29297 [15]). The first position was missing one base (a −1-bp frameshift) between 719 and 720 nt from the left end of the inverted repeat (IR-L) of the original sequence (Fig. 1a). The original ORF1 (892 aa) was annotated as a transposase (TnpA), but without the N terminus region found in other TnpAs. Based on the revised sequence (Fig. 1b), the start codon of the revised ORF1 was located upstream of the original ORF1 (from 505 to 3183 nt), and the revised ORF1 was located from 200 (TTG start) to 3184 nt (TGA stop), which encodes 994-aa TnpA (a characteristic motif of PF01526: Tn3 transposase DDE domain was observed), and the deduced polypeptide matched other TnpAs in a BLAST analysis. Furthermore, the revised ORF2 was annotated from 4392 (TGA stop) to 3355 nt (ATG start; encoded in the complementary strand), which encodes a methyltransferase domain-containing protein (345 aa; WP_085572232 of Streptomyces sp. 13-12-16 shows 95% identity and 97% similarity, WP/086708490 of S. castelarensis shows 95% identity and 97% similarity, and WP_116427250 of S. spongiicola shows 95% identity and 97% similarity), in which characteristic motifs of PF13847 (methyltransferase domain: 127–236 aa), PF08241 (methyltransferase domain: 133–228 aa), and PF13649 (methyltransferase domain: 131–224 aa) were observed, and was identical to ORF5 of the original annotation (Fig. 1a, b). The second position was the insertion of one base (a + 1-bp frameshift) at 5001 nt of the original sequence (Fig. 1a) and this insertion terminated translation because the insertion of an adenine residue made the stop codon TGA. The revised ORF3 was located from 4625 (ATG start) to 5401 nt (TGA stop), which encodes putative isoprenyl diphosphate transferase (258 aa; WP_085572231 of Streptomyces sp. 13-12-16 shows 96% identity and 98% similarity, WP_037940392 of S. toyocaensis shows 96% identity and 98% similarity, and WP_116427251 of S. spongiicola shows 96% identity and 97% similarity), in which a characteristic motif of PF01255 (putative undecaprenyl diphosphate synthase: 31–257 aa) was observed. The final position was at 6194 nt as a guanine residue (Fig. 1a; the original sequence was adenine). This region was defined as a truncated ORF in the original annotation [18], whereas a gene was located from 5445 (TGA stop) to 6419 nt (ATG start; in the complementary strand) in this resequencing. The revised ORF4 (324 aa) was annotated as a resolvase (TnpR). A previous study reported that the downstream ORF (3420–3851 nt of the original sequence) of TnpA was defined as a TnpR (Fig. 1a; ref 16), whereas the ORF did not contain the conserved motif (PF00589; phage integrase family) found in the TnpRs of transposons; however, the revised ORF4 was defined as a TnpR because the motif for PF00589 was found at 587–971 aa by a Pfam search. Thus, the revised sequence revealed that Tn4556 contains four ORFs. Class-II transposons form a cointegrate intermediate with the target replicon upon transposition and the intermediate is resolved by TnpR at the internal resolution site (res site; in many cases, a palindrome-like structure is found in the res site). In the original sequence, possible res sites were defined between 3243 and 3251 nt, and between 3307 and 3299 nt; however, these two sequences were not found in the palindrome-like structure. On the other hand, since the region between 6497 and 6527 nt in the revised sequence formed a palindrome structure containing an 11-bp stem with a 1-bp mismatch and 9-base loop structure, this region flanking the right end of the inverted repeat (IR-R) may function as a res site for resolving the cointegrate intermediate during the transposition process.

Fig. 1
figure 1

Organization of ORFs and the frame plot of Tn4556 derived from the 6625-bp a original sequence and b revised sequence. The direction of transcription and relative sizes of the ORFs deduced from an analysis of the nucleotide sequences are indicated. A frame plot was calculated by a window size of 40 codons and step size of 5 codons. Dashed lines indicate average G + C % (68.4%). IR-L, IR-R, and res indicate the left end of the inverted repeat, the right end of the inverted repeat, and the internal resolution site, respectively. The genes tnpA and tnpR encode transposase and resolvase, respectively. ORF2 and ORF3 in the revised sequence b encode predicted methyltransferase and isoprenyl diphosphate transferase, respectively. Vertical arrows in a indicate different points between the original and revised sequences of Tn4556

Construction of a new delivery vector carrying a counter-selectable marker for transposition

The transposition of Tn4556 and its derivatives was efficiently performed using a delivery vector replicated in Streptomyces strains rather than non-replicative vectors, such as the Escherichia coli plasmid pUC19. After transposition, the delivery vector containing the Tn4556 derivative has to be cured from the strain. We used the temperature-sensitive replication vector pKU110, derived from the pIJ101 replicon, for transposition in previous studies [9]; however, the elimination of the vector from the strain was not fully achievable. In the present study, we constructed a new delivery vector carrying a suicide gene for the transposition of Tn4556 derivatives. Although some suicide genes were used for counter selection in Streptomyces strains, their use was limited because sensitivity depends on the strain [7, 8]. The E. coli phenylalanyl-t-RNA synthetase β-subunit (PheS) was useful as a counter-selectable marker because its A294G variant misincorporates 4-chlorophenylalanine into cellular proteins during translation, thereby causing cell death [17]. This counter-selectable system was also applied to other bacteria [1, 3, 11, 13, 21]. While wild-type Streptomyces is insensitive to 4-chloro-dl-phenylalanine (> 20 mM), we demonstrated for the first time that Streptomyces strains carrying an extra copy of the gene encoding the PheS variant are sensitive to 4-chloro-dl-phenylalanine. S. avermitilis or S. lividans harboring a plasmid containing mutant pheS (A339G; corresponding to the A294G variant of E. coli PheS) genes (designed from the amino acid sequence of S. viridochromogenes DSM 40736; accession # WP_003989071) was sensitive to ~ 2.5 mM or 10 mM 4-chloro-dl-phenylalanine, respectively, (Table 1). Since the aa position 251 (Thr) of E. coli PheS was involved in substrate recognition, this position was our target for the enhancement of sensitivity to 4-chlorophenylalanine [15]. After the introduction of an aa replacement in the second target at Thr278 of Streptomyces PheS (corresponding to Thr251 of E. coli PheS), sensitivity to 4-chloro-dl-phenylalanine was enhanced by approximately eightfold by the variant at T278S or T278A in PheSA339G (Table 1). We then constructed the new delivery vector pGM160∆aac1::oriT::pheSA339G/T278A::Tn4556-aac3(IV), and the transposition of Tn4556-aac(3)IV was efficiently performed because it was easy to isolate progeny carrying the transposon without the delivery vector after a high-temperature incubation at 37 °C and the subsequent selection of 4-chlorophenylalanine resistance.

Table 1 Susceptibility to 4-chloro-dl-phenylalanine of Streptomyces strains and their exoconjugants carrying the mutant type of the pheS gene

Improvements in transposition using perfectly matched IR-L

In bacterial transposons, IRs at both ends of class-II transposons were relatively long (38–110 bp). Tn4556 possessed 38-bp IRs with a 1-bp mismatch (Table 2). The exchange to the perfectly matched IRs of the transposon TnHad2 in the IncP-1β plasmid pUO1 from Delftia acidovorans strain B improved transposition frequency [20]. After the cytosine residue of the 5′-end of IR-L of Tn4556 was replaced to a guanine residue, the modified IR-L perfectly matched to IR-R (Table 2). Tn4556-aac(3)IV carrying perfectly matched IR-L was joined to pGM160∆aac1::oriT::pheSA339G/T278A and transposition efficiency was examined. As shown in Table 2, transposition efficiency was approximately five- to tenfold better than that of wild-type Tn4556-aac(3)IV and the transposition occurred randomly on S. avermitilis chromosome (Fig. 2).

Table 2 Transposition of wild-type Tn4556-aac(3)IV and its derivative consisting of perfectly matched IR-L in S. avermitilis
Fig. 2
figure 2

AseI-map of S. avermitilis chromosome and transposition loci of Tn4556-aac(3)IV. The dot indicates transposition locus. The 5-bp target duplication, e.g., GGGTT, TAGAG, TGCTC, GACTG, ACCAT, GGATC, GGAGC, ATGAC, AGGTA and so on, was also confirmed at the insertion site by the sequence. Abbreviations oriC and rrnA-F indicate the replication origin and ribosomal RNA (16S–23S–5S rRNA) operons

Conclusion

We corrected the Tn4556 sequence by resequencing, and the most important ORFs for the class-II transposon, TnpA and TnpR, were accurately annotated. The new delivery vector was useful for the isolation of progeny carrying transpositions because a suicide gene encoding PheSA339G/T278A as a counter-selectable marker eliminated the delivery vector by the selection of 4-chlorophenylalanine resistance after transposition. The replacement of the perfectly matched 38-bp IR-L variant enhanced transposition. Tn4556-aac(3)IV containing perfectly matched IR-L may be useful for transposon mutagenesis in Streptomyces strains.