Excisable Cassettes: New Tools for Functional Analysis of Streptomyces Genomes

Strains and media. E. coli strain DH5α (8) was used for cloning experiments and plasmid DNA propagation. Streptomyces strains were routinely cultivated on Hickey-Tresner medium at 30°C as described previously (15). Streptomyces lividans strain TK24 and Streptomyces ambofaciens ATCC 23877 were used for transformation experiments. Transformants carrying the hyg, tsr, or aac resistance gene were selected with 150 to 200 μg hygromycin (Hm) ml⁻¹, 25 μg thiostrepton (Ts) ml⁻¹, or 50 μg geneticin (Gn) ml⁻¹, respectively, in R2YE medium (10).Transformants harboring the aac gene can also be selected with apramycin or gentamicin. Streptomyces ambofaciens ATCC 23877 produces the macrolide antibiotic spiramycin (16). Bacteria containing plasmids were routinely grown on LB medium supplemented with 100μ g ampicillin (Ap) ml⁻¹ and 20 μg Gn ml⁻¹, and bacterial cultures in liquid LB medium were supplemented with 30 μg Ap ml⁻¹ and 10μ g Gn ml⁻¹.

Plasmids, antibiotic resistance genes, and oligonucleotides.All restriction endonuclease digestions, ligation reactions, DNA modifications, and PCR amplifications were carried out according to standard protocols (19). E. coli and Streptomyces were transformed according to standard protocols (11, 19). The plasmid pGEM-T Easy (Promega) was used for cloning of PCR products, and pBluescript (Stratagene) was used for the cloning experiments. All of the plasmids used in this study are listed in Table 1. KS+DH3, a pBluescript KS (+) derivative deleted for the HindIII restriction site, was constructed to clone the antibiotic resistance genes.

The genes encoding Hm resistance and Gn resistance were provided by Ωhyg and Ωaac cassettes (2), respectively, derived from the Ω interposon as HindIII-HindIII fragments. A HindIII fragment containing only the aac gene was obtained by PCR amplification using the primers KF42 (5′AAGCTTGTACGGCCCACAGAATGATGTCAC3′) and KF43 (5′AAGCTTCGACTACCTTGGTGATCTCGCCTT3′) (HindIII sites are underlined), with an Ωaac cassette containing a plasmid as the DNA template. The resulting PCR product was inserted into pGEM-T Easy. A HindIII fragment carrying the aac gene was obtained from the resulting plasmid. Initial attL-attR cassettes were obtained after two successive PCR steps using the following primers: Cas1R (5′CATGCCGGTCGGGATATCGCGCGCTTCGTTCG3′), Cas2R (5′CATGCCGGTCGGGATATCGGCGCGCTTCGTTCG3′), Cas3R (5′CATGCCGGTCGGGATATCGCGCGCGCTTCGTTCG3′), CasR (5′AGATCTGTTAACAAGCTTCTCGAGGGATCCCTGTCAGTCATGCGGG3′), CasL (5′GGATCCCTCGAGAAGCTTGTTAACAGATCTCCCGGCTCGTCGGAC3′), Cas1/2L (5′CCCGGGGATCTGGATATCTACCTCTTCGTCCC3′), and Cas3L (5′CCCGGGGATCGTGATATCTGCCTCTTCGTCCC3′).

Spiramycin production assays.For spiramycin production, S. ambofaciens was grown in MP5 liquid medium at 27°C (14). For bioassays, the supernatant of the culture was applied to Whatman AA paper discs. The discs were laid on plates containing Micrococcus luteus, and the plates were first incubated at 4°C for 2 hours to allow antibiotic diffusion and then incubated at 37°C. The growth inhibition area was measured and compared to standards obtained using spiramycin, as described previously (14).

Construction of S. ambofaciens ATCC 23877 derivative devoid of pSAM2.Since we were using the site-specific recombination system from pSAM2 to carry out excision in S. ambofaciens, we preferred to work with an S. ambofaciens strain devoid of pSAM2 to avoid all kinds of interference due to the integrated copy of pSAM2. Therefore, experiments were undertaken to obtain an S. ambofaciens strain cured of the integrated copy of pSAM2. It has been reported that the production and regeneration of bacterial protoplasts can promote the loss of plasmids (5). Since the pSAM2 functions are well characterized, we designed a reporter system allowing positive selection for the loss of pSAM2.

The pra gene has been described as an activator of pSAM2 replication, and its inactivation leads to the disappearance of free pSAM2 (22). KorSA has been identified as a central transcriptional repressor (23) that binds to the pra gene promoter, thus repressing pra gene expression. The rationale behind this experiment is that derepression of the pra promoter should be observed in the absence of the KorSA repressor, indicating the loss of pSAM2. We used pOS527 to obtain a fragment carrying the pra promoter fused to the promoterless aph gene (22), which was inserted into the unstable replicative vector pWHM3hyg, a pWHM3 (26) derivative in which the tsr gene (conferring Ts resistance) is replaced by hyg (conferring Hm resistance). The resulting plasmid, pOSV510, was used to transform protoplasts of the S. ambofaciens ATCC 23877 strain. Selection with neomycin of clones expressing the aph reporter gene allowed S. ambofaciens strains devoid of pSAM2 to be isolated. The absence of pSAM2 was checked by Southern blotting (data not shown), and one of the obtained clones was named OSC2. The OSC2 strain was not affected in growth, sporulation, or spiramycin production.

Nucleotide sequence accession numbers.The following sequences were deposited in the EMBL database with the indicated accession numbers: att1Ωaac+, AM238621; att2Ωaac+, AM238622; att3Ωaac+, AM238623; att1Ωhyg+, AM238624; att2Ωhyg+, AM238625; att3Ωhyg+, AM238626; att1Ωaac−, AM238627; att2Ωaac−, AM238628;att3Ωaac−, AM238629; att1Ωhyg−, AM238630; att2Ωhyg−, AM238631; and att3Ωhyg−, AM238632.

Construction of attR-antibiotic resistance-attL cassettes.The pSAM2 chromosomal attB site shares a 58-bp identity segment with attP that extends from the region encoding the anticodon loop to the 3′ end of the tRNA^Pro gene. Site-specific recombination that leads to integration or excision, promoted by the pSAM2 Int and Xis proteins, has been described for E. coli (18). It has also been shown that a 26-bp att sequence (attB26) centered on the region encoding the anticodon stem-loop of the tRNA and not completely included in the identity segment retains the entire functionality of attB (18). The minimal attP site has also been defined (17). Therefore, after site-specific intermolecular recombination between attP and attB26, it is possible to obtain what may be considered the minimal attL and attR sites. This experiment aimed to clone these minimal attL and attR sites as a single fragment, with restriction sites at the junction between attL and attR allowing the insertion of antibiotic resistance markers. When required, intramolecular site-specific recombination between these attL and attR sequences will excise the entire sequence between them, including the resistance marker, reconstituting the minimal attB site. We were able to add 1 or 2 bp to the 26-bp minimal attB site to construct a set of three different cassettes, leaving sequences of n, n+ 1, and n + 2 bp after excision. Thus, depending on the length of the insertion/deletion within the coding sequence of the target gene, it is possible to choose one of the cassettes to maintain the original reading frame after excision, thus avoiding polar mutations.

Construction was undertaken starting from plasmid pOSCo26, a cointegrate resulting from site-specific recombination between pSAM2 attP and the minimal attB26 site (16). In the first step (Fig. 1a), attR and attL were amplified as individual fragments from the cointegrate molecules. The CasR and CasL primers carried a 30-nucleotide common sequence, with an inverted orientation in one of the primers. This allowed both to associate with the attR/attL molecules obtained from the first PCR and to create restriction sites that allow the antibiotic resistance genes to be inserted between attR and attL in a further step. In a second PCR step (Fig. 1a), the PCR products from the first step were used as templates for amplification, using different couples of external primers that provided EcoRV restriction sites. This allowed a set of cassettes to be generated, named CASS1, CASS2, and CASS3, containing both attR and attL and having respective sizes of 485, 486, and 487 bp (Fig. 1b). These cassettes were further inserted as EcoRV-EcoRV fragments into the plasmid KS+DH3, leading to the plasmids pOSV501, pOSV502, and pOSV503, respectively. In the next step, an antibiotic resistance cassette was inserted between the attL and attR sites in these plasmids (Fig. 1c). Both Ωhyg and Ωaac cassettes (2), carrying Hm and Gn resistance genes, respectively, were inserted as HindIII fragments into pOSV501, pOSV502, and pOSV503. Since there were three types of sequence left after excision (att1, 33 bp; att2, 34 bp; and att3, 35 bp) (Fig. 1d), two possible orientations of the marker genes (+ denotes a resistance gene transcribed in the attL-to-attR orientation, and− denotes a resistance gene transcribed in the other orientation), and two different fragments carrying the resistance genes (Ωhyg and Ωaac), this led to a set of 12 plasmids containing the cassettes att1Ωhyg+/−, att2Ωhyg+/−, att3Ωhyg+/−, att1Ωaac+/−, att2Ωaac+/−, and att3Ωaac+/− (Table 1). The derivatives of plasmid KS+DH3 containing these cassettes are designated by the letter “p” followed by the name of the cassette.

• Open in new tab
• Download powerpoint

FIG. 1.

Construction and characteristics of excisable cassettes. (a) attL and attR were amplified individually by PCRs using the CasR and CasL primer families. (b) The CasL-CasR central primers share a 30-nucleotide inverted repeat sequence that allows attL1, -2, and -3 to be associated with attR1, -2, and -3, respectively, followed by PCR amplification using external CasL and CasR primers carrying an EcoRV restriction site sequence. The second PCR step generates the cassettes, CASS1, -2, and -3 (c), carrying a multicloning site in the central part, allowing antibiotic resistance cassette cloning. (d) Representation of the 33-, 34-, and 35-bp sequences remaining after excision, with respect to the reading frame and the cassette cloning orientation. In att2 and att3, the bold characters indicate nucleotides coming from the disrupted gene. There are some constraints in the possible nucleotide that can be used to avoid a stop codon in the sequence remaining after excision.

Construction of xis-int-expressing plasmids for Streptomyces.In pSAM2, the int and xis genes are located downstream from the repSA gene that encodes the replicase, and the three genes are cotranscribed. Moreover, pra expression is required for their efficient transcription (21). Therefore, it was necessary to place these genes under the control of a heterologous promoter to obtain a simple construction expressing xis and int. A first plasmid, pOSV507, was constructed in which the xis and int genes are transcribed from the ermE* promoter and carried by the low-copy-number vector pIJ903. Another plasmid, pOSint3, expressing xis and int under the control of the E. coli trc promoter, was constructed in a previous study (18). Since the E. coli trc promoter is functional in Streptomyces, a fragment carrying trcp-xis-int isolated from pOSint3 was inserted into plasmid pWHM3 to obtain pOSV508. pWHM3 is an E. coli/Streptomyces shuttle vector that is not very stable in Streptomyces in the absence of selective pressure. In this case, the instability was an advantage as it allowed easy curing of the plasmid.

Functionality of excision in E. coli and Streptomyces.The excisable cassettes can be isolated as blunt-ended EcoRV fragments from the plasmids that carry them and easily inserted into any gene, as insertion of the cassette can be selected by the appropriate antibiotic in both E. coli and Streptomyces. Excision of the cassette can be achieved by expressing the xis and int genes and can be screened for by the loss of antibiotic resistance. The signatures left by the cassettes after excision are shown in Fig. 1d with respect to the reading frame and the cassette cloning orientation in the disrupted gene. Some constraints exist for using the cassette that leaves the att2 or att3 sequence.

We investigated the excision event by inserting the att1Ωhyg cassette, isolated as an EcoRV fragment, into pSET152 digested by EcoRV, yielding pOSV511. The vector pSET152, which can replicate in E. coli, carries a Gn resistance marker (Gn^r) and the FC31 attachment site, which allows its integration into the chromosomes of several Streptomyces species (1). pOSV511 was introduced into a strain of E. coli containing pOSint3 (18) and therefore expressing the int and xis genes. Among the transformants selected as Gn^r colonies, >90% were Hm^s, showing very efficient excision of the cassette in E. coli.

We tested the excision efficiency in Streptomyces by introducing pOSV511 into S. lividans strain TK24 by protoplast transformation. Integrative transformants were selected as Gn^r Hm^r colonies. Southern blotting with three different transformants showed that the fragment containing att1Ωhyg+ had been integrated into the S. lividans chromosome.

We introduced the plasmid pOSV507 expressing the xis and int genes from pSAM2 into S. lividans carrying integrated pOSV511 to promote the excision of the cassette (data not shown; see below). The transformants were picked and grown on plates with no Hm selection. After two rounds of sporulation under these conditions, we tested the isolated clones for their Hm resistance. Hm^s clones were readily obtained. Southern blotting with total DNA extracted from two of these Hm^s clones confirmed excision of the cassette (data not shown). The regions containing the sequence left after excision were cloned by marker rescue and sequenced; in both cases, we found the expected 33-nucleotide sequences flanked by the two EcoRV cloning sites (data not shown). All of these results clearly showed that the excision of the cassette was efficient in S. lividans as well as in E. coli and was site specific, leaving the expected attB-like sequence after excision.

Example of the use of the cassettes: insertion of a cassette into a target gene.We showed how these cassettes could be used by choosing to inactivate a gene belonging to an operon located in the spiramycin biosynthetic cluster of Streptomyces ambofaciens. Seven genes, orf1 to orf7, encoding enzymes involved in various steps of spiramycin biosynthesis form an operon (N. Oestreicher et al., unpublished). It was shown that inserting the Ωhyg cassette into orf3, the third gene of this operon, abolished spiramycin production. However, insertion of this cassette caused a polar effect on transcription of the downstream genes, orf4 to orf7. After several gene replacement steps, it was possible to obtain an in-frame deletion in orf3; this deletion still abolished spiramycin biosynthesis but did not cause a polar effect (N. Oestreicher et al., unpublished). We repeated the orf3 inactivation, using an excisable cassette to test whether there was no polar effect after excision.

A 4.5-kb EcoRI-BamHI fragment containing orf1 to orf4 of this operon was inserted into pUC19, yielding pOS49.99. In this fragment, orf3 was disrupted by inserting the att1Ωhyg cassette, isolated as an EcoRV fragment. This fragment was inserted into pOS49.99 after PmlI/Asp718I digestion, followed by filling in of the two protruding ends, giving the plasmid pOSV512. The insertion of the cassette was accompanied by deletion of 270 bp of the orf3 coding sequence. A fragment carrying orf1-2-orf3::att1Ωhyg-orf4 was inserted into pOJ260, a vector unable to replicate in Streptomyces (1), leading to pOSV513. This first step is represented schematically in Fig. 2a.

• Open in new tab
• Download powerpoint

FIG. 2.

Schematic representation of the different steps for generating unmarked mutant strains. (a) Step 1, cloning of the gene of interest, followed by its disruption by insertion of the chosen cassette. (b) Step 2, replacement of the wild-type copy of the target gene by the disrupted gene via a double recombination event. (c) Step 3, excision of the cassette at the chromosomal locus after transitional expression of Xis and Int.

Gene replacement and cassette excision in S. ambofaciens.Protoplasts of the OSC2 strain were transformed with pOSV513 DNA denatured by alkali treatment according to the method of Oh and Chater (13). Transformants were selected as Hm^r colonies and were further grown on medium containing Gn to screen for Hm^r Gn^s colonies in which gene replacement might have occurred. Total DNA was extracted from several of these clones and analyzed by Southern blotting. Two transformants in which orf3 had been replaced by orf3::att1Ωhyg were selected for further analysis. Both clones were unable to produce spiramycin (Fig. 2b and 3b).

• Open in new tab
• Download powerpoint

FIG. 3.

Representation of the different structures at the orf1-orf7 operon in different strains of S. ambofaciens. (a) S. ambofaciens OSC2, an ATCC 23877 derivative devoid of pSAM2. (b) Same strain after inactivation of orf3 by insertion of att1Ωhyg− (orf3::att1Ωhyg−). (c) Same strain after excision of the cassette (orf3::att1). (d) Construct b with orf3 carried by plasmid pOS49.52. (e) Construct c with orf3 carried by plasmid pOS49.52. Spiramycin production is indicated for all strains.

Excision was carried out using the plasmid pOSV508, in which xis and int are transcribed from the E. colitrc promoter. Excision with pOSV508 was more efficient than that with pOSV507, and pOSV508 was less stable in Streptomyces than pSOV507, allowing a rapid loss of the xis-int-expressing plasmid. We picked 35 transformants grown on solid medium without antibiotic, which were then grown for 4 days until sporulation. After being replica plated on media with Hm, with Ts, and without antibiotics, all transformants appeared to be sensitive to Hm and Ts, suggesting both efficient excision and a very rapid loss of plasmid pOSV508. Excision of the cassette was confirmed by Southern blotting and PCR on total extracted DNA. After excision of the cassette, a residual 33-bp sequence was left. Since insertion of the cassette had been accompanied by a deletion of 270 bp, the reading frame was not changed in the inactivated orf3::att1 copy.

Absence of polar effect after cassette excision.We demonstrated the absence of a polar effect after excision by studying the restoration of spiramycin production by orf3 expression in the orf3::att1Ωhyg strain (before excision) (Fig. 3d) and in the orf3::att1 strain (after excision) (Fig. 3e). A DNA fragment carrying the Streptomyces ermE* promoter upstream from the entire orf3 coding sequence was inserted into vector pIJ903, yielding pOS49.52. This plasmid did not restore spiramycin production in the orf3::att1Ωhyg strain, indicating a polar effect due to disruption of the expression of the downstream genes, whose products are necessary for spiramycin biosynthesis (Fig. 3d).

Further experiments were carried out on two independent clones from which the cassette had been excised. The two clones were transformed with pOS49.52. Transformants were selected on R2YE medium containing Ts and further allowed to sporulate on Hickey-Tresner mediumcontaining Ts. Three independent transformants were assayed for the ability to produce spiramycin in liquid medium. All produced spiramycin at the same level as the OSC2 strain (Fig. 3e), demonstrating the absence of a polar effect due to the orf3::att1 mutation.