INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Streptomyces spp. are gram-positive soil bacteria that undergo a complex cycle of morphological differentiation and synthesize multiple medically and industrially useful secondary metabolites (11, 12). Plasmids isolated from Streptomyces species (reviewed in reference 17) include autonomous circular plasmids (e.g., pIJ101 [24], pSN22 [21], and pJV1 [32]) and linear replicons (40), as well as plasmids generated by site-specific excision of chromosomal DNA segments and capable of reintegrating site specifically into Streptomyces chromosomes (e.g., SLP1 and pSAM2 [reviewed in reference 28]). During their existence as extrachromosomal replicons, most integrating plasmids share with other types of Streptomyces plasmids the ability to undergo conjugal transfer and inhibit transiently the growth of plasmid recipients, yielding zones of slowed growth called pocks (3).

SLP1, which was the first reintegrating plasmid discovered in Streptomyces (4), exists normally as a 17.2-kb plasmidogenic sequence (SLP1^int) in the chromosome of Streptomyces coelicolor A3(2). Upon mating of S. coelicolor with S. lividans, a closely related strain which lacks SLP1, SLP1 can excise from the chromosome and transfer to S. lividans, where it can either enter the recipient chromosome or undergo deletion, generating autonomously replicating circular plasmids (28). The integration and excision of SLP1 occur by site-specific recombination between a chromosomal attachment site (attB) and a largely homologous site (attP) on SLP1 and are mediated by two genes, int and xis, whose products resemble, respectively, the integrase and excisase of bacteriophage  and other temperate phages (7, 8). SLP1^int includes a locus, designated imp (inhibition of plasmid maintenance), which can act both in cis and in trans by a mechanism distinct from normal plasmid incompatibility, to prevent propagation of SLP1 as an extrachromosomal element; consequently, existence of SLP1 as a plasmid requires deletion or mutation in imp (15). Two translationally coupled genes within the imp locus, impA and impC, function conjointly to autoregulate their own expression at the imp promoter (P_imp) (34). The impC gene product contains a helix-turn-helix motif and exerts negative regulation over SLP1 replication (15, 34). The ImpA protein has features in common with the GntR family of transcriptional repressors (14, 34), which includes the Kor proteins encoded by Kil-override loci affecting the transfer of certain Streptomyces plasmids (36). Together, ImpA and ImpC can also regulate a promoter located 5' to xis (6). Collectively, these findings have raised the prospect that imp may have multiple regulatory roles related to the existence of SLP1 in its extrachromosomal state and its behavior as a plasmid (34).

Here we report the cloning and characterization of an imp-regulated promoter that controls SLP1-mediated gene transfer and the identification and characterization of SLP1 genes that govern intermycelial and intramycelial dissemination of the plasmid. Our findings, which support the notion that the imp locus is a master regulator of multiple functions associated with the extrachromosomal state of SLP1, suggest that SLP1-mediated enhanced gene transfer, but not low-frequency mobilization of chromosomal genes, results from decreased imp activity in donors following the formation of mating pairs.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains, plasmids, and growth conditions. The strains and plasmids used in this study are listed in Table 1. S. lividans TK64 was used as the Streptomyces host strain. TK23 was used in genetic crosses to assay plasmid transfer and chromosome-mobilizing ability. Escherichia coli DH5 or BRL2288 was used for transformation. E. coli MB117 was used for the delivery of the Tn5 transposon. Streptomyces strains were grown in liquid culture in YEME medium (16) or in solid culture in MY (described in reference 8), R5, or MM (16) medium. Hygromycin, thiostrepton, spectinomycin, and kanamycin were used as described elsewhere (16) to select for resistance to these antibiotics (Hyg^r, Tsr^r, Spc^r, and Kan^r, respectively).

Transformation, plasmid isolation, and DNA manipulation. Transformation of Streptomyces was done as described elsewhere (16), using R5 medium; 200 ng of DNA was used in transformation mixes. Plasmids were isolated by using a plasmid DNA isolation kit from Qiagen, Inc. (Valencia, Calif.). Ligations were done as described elsewhere (31), using 5× ligase buffer from Life Technologies, Inc. (Rockville, Md.).

Construction of an SLP1 library in pXE4. pCAO106 (pACYC177 [10] containing SLP1 as a BamHI insert) was digested by BamHI. The SLP1 fragment was isolated from the gel, purified by using a gel extraction kit (QIAquick; Qiagen), and partially digested by the enzyme Sau3AI. The Sau3AI fragments generated were then cloned into the pXE4 BamHI site as described elsewhere (31).

Catechol dioxygenase activity assays. Catechol dioxygenase (XylE) activity assays were performed as described in reference 18. Briefly, qualitative assays were performed by spraying 0.1 M catechol onto plates containing 2- to 3-day-old colonies. Appearance of yellow colonies was determined visually no more than 1 h after spreading of the catechol in the initial screen and then 10 min after addition of catechol for studies of P_opimp (promoter opposite imp).

Tn5 mutagenesis. pSUM50 was used to transform strain E. coli MB117, containing the Tn5 transposon (Table 1), in 23 independent tubes. Ampicillin-resistant and Kan^r transformants were selected, and plasmid DNA was extracted from these transformants. The Tn5 insertions were then mapped by restriction analysis.

DNA sequencing and sequence analysis. To sequence the 1.85-kb SalI (8.70)-HindIII (10.20) region, pSUM50 was digested by BamHI and HindIII, and the 2.9-kb fragment separated on gels by electrophoresis was isolated by using a QIAquick gel extraction kit (Qiagen). This fragment was then digested by SalI and cloned into pUC18 digested by HindIII and SalI. To sequence the SLP1 HindIII (10.20)-BclI (14.80) region, pSUM50 was digested by HindIII, and the 6-kb fragment separated on gels by electrophoresis was isolated by using a QIAquick gel extraction kit. This fragment was then cloned into pUC18 linearized by HindIII and previously dephosphorylated by calf intestinal alkaline phosphatase (31), to prevent its religation. Sequence was performed by oligonucleotide walking. pUC forward and reverse 23-base sequencing primers were used. The other primers used were obtained from Operon Technologies, Inc. (Alameda, Calif.). Sequencing was performed with an ABI PRISM 310 Genetic Analyzer from Perkin-Elmer (Foster City, Calif.). Open reading frames (ORFs) were identified by using a modified version of the program FRAME (2), kindly provided by K. Kendall. The FRAME program is based on the biased codon usage that occurs in organisms having high G+C content. Sequence was analyzed by using BLAST (1), DNA Strider (26), and Pedro's BioMolecular Research Tools (29a).

Primer extension analysis. Total RNA, isolated from S. lividans TK64 harboring pXE4 or pSJH6 (pXE4 containing the Sau3AI SLP1 region having promoter activity) or from Stm188 (S. lividans expressing impA and impC) harboring pSJH6, was extracted as described by Hopwood et al. (16). RNA was hybridized with oligonucleotide pXE4₇₁ (5'-CGGCGGTTTCAAAGGCCAGACAGGCGGTAAG-3') (see Fig. 2) labeled with [-³²P]ATP. The products were extended as described elsewhere (27) and loaded on a 6% polyacrylamide gel. A sequencing ladder was generated by sequencing pSJH6 or pXE4 with oligonucleotide pXE4₇₁.

Nucleotide sequence accession number. The nucleotide sequence of the SLP1 transfer region has been deposited in GenBank under accession no. 538670.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Identification and characterization of an SLP1 promoter controlled by imp. Autoregulation of imp expression is mediated through a promoter, P_imp, that initiates transcription of a polycistronic operon containing the impA and impC genes (34) (Fig. 1). To identify other SLP1 promoters regulated by imp, we fused SLP1 DNA fragments generated by partial digestion with Sau3AI to the pXE4 xylE gene (Fig. 2), which encodes catechol dehydrogenase and converts colorless catechol to an intensely yellow oxidation product (hydroxymuconic semialdehyde) (18). After transformation of S. lividans, six clones producing yellow colonies were identified; these were tested for repression of yellow coloration by impA (strain Stm190 [see Materials and Methods]), impC (strain Stm189), or a cassette containing both impA and impC (strain Stm188), all integrated into the S. lividans chromosome as a single copy. As shown in Fig. 3, one of the six promoters detected in the reporter gene assay was repressed by impA or impC alone to 6 to 12% of the expression observed in an imp-minus strain and was repressed 8 to 16 times more effectively by impA and impC acting together. The pXE4 construct carrying the DNA fragment specifying imp-repressed activity was designated pSJH6 (Fig. 3).

View larger version (23K): [in this window] [in a new window]   FIG. 1.   Map of SLP1 showing the region containing Popimp and those involved in transfer and pock formation. Popimp, Pimp, the SLP1 attachment site (attP), the integrase (int) and excisase (xis) genes, as well as the genes and ORFs involved in SLP1 intermycelial (tra) and intramycelial (spdB) transfer are shown. The direction of translation is indicated by filled arrows. The numbers indicate the positions (in kilobases) of the restriction sites in the sequence of the whole plasmid.

View larger version (25K): [in this window] [in a new window]   FIG. 2.   (A) Primer extension analysis of Popimp. RNA was prepared from S. lividans TK64 harboring pXE4 (lane 1) or pSJH6 (lane 2) or from S. lividans Stm188 expressing impA and impC and harboring pSJH6 (lane 3). The asterisk represents the transcriptional start site. (B) Map of the pXE4 derivative containing the Sau3AI (8.10 to 8.66) insert (pSJH6). The primer used for primer extension analysis, pXE471, is also shown. tsr, the thiostrepton resistance gene from Streptomyces spp.; bla, a -lactamase gene from E. coli; ori, the ColE1 origin of replication; SCP2 rep/stb, the SCP2 replication and stabilization functions (16); xylE, a promoterless copy of the xylE gene (18).

View larger version (29K): [in this window] [in a new window]   FIG. 3.   Control of Popimp by imp genes. Streptomyces strains TK64, Stm188 (expressing impA and impC), Stm189 (expressing impC), and Stm190 (expressing impA), harboring or not harboring plasmid pXE4 (containing the xylE gene) or pSJH6 (whose map is shown), as indicated, were grown on MY medium containing cellophane membranes and lysed by sonication, and expression of the xylE gene of the plasmids was measured in triplicate as indicated in Materials and Methods. The numbers for the XylE activity are the rate of change in optical density at 375 nm (OD375) per minute per milligram of protein and are averages from three independent assays. "Expression" refers to the level of expression of the xylE gene in the plasmid tested compared to that in pXE4, containing a promoterless xylE gene. tsr, the thiostrepton resistance gene from Streptomyces spp.; bla, a -lactamase gene from E. coli; ori, the ColE1 origin of replication; SCP2 rep, the SCP2 replication and stabilization functions (16); xylE, a promoterless copy of the xylE gene (18).

71

71

View larger version (33K): [in this window] [in a new window]   FIG. 4.   Sequence and sequence analysis of the Sau3AI (8.10)-BclI (14.8) SLP1 region upstream of Popimp. (A) The sequence of the first 200 bp of the Sau3AI (8.10)-Sau3AI (8.66) region and of the 200 bp prior to the spdB1 gene (as determined previously [15]) are shown and are separated by dashed lines. Recognition sequences sites for the restriction enzymes shown are underlined, as are sequences which function as putative ribosome-binding sites (RBS). Translation of the different ORFs is shown above the nucleotide sequence. The arrows originating at G, 5' of impA, and at T, 5' of spdB1, indicate the Pimp and Popimp transcription start sites, respectively. Dashed arrows indicate inverted repeats. Direct repeats (DR) are underlined. The 35 and 10 hexamers are indicated by heavy lines above the sequence. (B) Complete sequence Sau3AI (8.10)-BclI (14.8). The ORFs are indicated by arrows. Putative transmembrane helices of the putative protein are indicated by TM. The restriction sites are shown. Numbers indicate nucleotide positions of the start and stop sites of the ORFs in the sequence.

Characterization of the imp-regulated DNA region proximal to P_opimp. The sequence of the 1.5-kb DNA region immediately proximal to P_opimp (SalI (8.70)-HindIII (10.2) was determined. Computer-assisted analysis using the FRAME program (2) identified two ORFs (Fig. 4 and Table 2) predicted to encode proteins of 261 and 335 amino acids (aa). The second of these ORFs showed 47% identity and 56% similarity to spdB2 gene of pJV1 (33) and consequently was designated spdB2. The ORF proximal to the promoter was given the initial designation of orf261 and later named spdB1 for reasons discussed below. Within orf261 is the Sau3AI site of fusion of SLP1 to the xylE gene of pXE4, indicating that production of catechol dehydrogenase from the reporter gene reflects expression of orf261. Additionally, the site of the orf261-xylE fusion indicates that the P_opimp-initiated transcript found by primer extension experiments to be regulated by imp corresponds to the 5' end of orf261. As the stop codon of orf261 overlaps the start codon of spdB2 (see the GenBank sequence), orf261 and spdB2 are likely to be translationally coupled. Translational coupling has also been proposed for the spdB1 and spdB2 loci of pJV1 and pSN22, which are in corresponding positions (19, 33). orf261 is preceded by a likely ribosome-binding site (AAGGA [27]). SpdB2 contains four putative transmembrane helices schematized in Fig. 4B.

View larger version (8K): [in this window] [in a new window]   FIG. 5.   Tn5 mutational analysis of the ORFs located in the Sau3AI (8.10)-BclI (14.8) SLP1 region upstream of Popimp. Arrows represent the orientation of ORFs. The small triangles and dashed lines represent locations of the Tn5 insertions. The pock sizes formed by plasmids are represented by +++, ++, +, and , for normal, small, and tiny pocks and no pocks, respectively. Plasmid transfer efficiency was calculated as described in reference 20. The ORFs are indicated by arrows. Plasmids and the Streptomyces strains harboring them are shown.

Analysis of the SLP1 transfer region. Transfer-related genes are commonly grouped together in Streptomyces plasmids (reviewed by Hopwood and Kieser [17] and in earlier reports [23, 29] cited in reference 16) have suggested that a large region extending clockwise in SLP1 from the site at which we have mapped P_opimp is implicated in transfer of this plasmid. The sequence of this entire region, which was cloned as 6-kb HindIII (10.20)-HindIII (16.20) fragment, was determined and is schematized in Fig. 4A. The G+C content is 72 mol%, which is typical for Streptomyces DNA. All of the deduced ORFs (Fig. 1) are read in the same direction as transcription initiated from P_opimp. Distal to spdB1 and spdB2 genes are four additional ORFs, which were found to affect plasmid transfer as determined by Tn5 insertion mutagenesis (Fig. 5). A 179-aa ORF distal to orf294 has features in common with the traA genes of Streptomyces plasmid pJV1 (33) (47% identity and 62% similarity) and resembles the traA gene of pSN22 (19); this SLP1 ORF has thus been designated traA. SLP1 traA also resembles a Mycobacterium tuberculosis gene encoding a protein of unknown function, designed Rv 0911 (31% identity and 40% similarity) (13). An SLP1 gene designated traB, encoding a putative protein of 682 aa, resembles traB of pJV1 (47% identity and 56% similarity), which is itself very similar to the traB gene of pSN22 (19). SLP1 TraB also presents similarities to the SpoIIIE protein of Bacillus subtilis (9, 38) (25% identity and 45% similarity). traB of SLP1 has a P-loop ATP/GTP-binding motif in approximately the same position as in pSN22 and pJV1 traB. traA and traB are preceded by likely ribosome-binding sites (37) (AGGAG and AAGG) located 11 nucleotides from their putative translational initiation codons (see the complete sequence in GenBank). Tn5 insertions in either of these genes abolished pocking and reduced the transfer frequency, identifying these loci as transfer genes (Fig. 5).

Expression of imp in recipients prevents SLP1-mediated gene transfer. To examine directly the role of imp in chromosomal and plasmid gene transfer mediated by SLP1, S. lividans TK23 harboring the integrated SLP1 derivative pCAO153 was mated with S. lividans TK64 expressing impA and impC (strain Stm188), impA alone (Stm190), or impC alone (Stm189). Transfer of plasmid and chromosomal genes was monitored by measuring transfer efficiency and chromosome-mobilizing ability as described in reference 30. As shown in Table 3, the low background frequency of chromosomal recombinants observed for TK23 (approximately 10⁷; mating pair 1) was raised 2 orders of magnitude (mating pair 3) when the conjugation-proficient SLP1 derivative pCAO153 was present in donor cells, as was shown previously (29). Expression of either impA or impC in the recipient decreased the frequency of pCAO153-mediated recombinants by approximately 50% (mating pairs 5 and 6). However, when both impA and impC were expressed in the recipient, the ability of pCAO153 to promote chromosomal transfer was totally reversed (mating pair 7). As shown in Table 4, expression of either impA or impC in the recipient also decreased the frequency of transfer of the pCAO153 plasmid by more than 2 orders of magnitude (mating pairs 3 and 4). When both impA and impC were expressed in the recipient, the efficiency of transfer was reduced by more than 3 orders of magnitude (mating pair 5). These results, which indicate that expression of imp products in the recipient prevents SLP1-mediated transfer of both chromosomal and plasmid genes, confirm the role of Imp in such transfer.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The results reported here indicate that the imp operon controls transfer functions as well as maintenance functions of the SLP1 plasmid. We identified and characterized a promoter, P_opimp, which is repressed by imp and was found to transcribe genes required for intramycelial spreading of SLP1. Two additional groups of genes distal to P_opimp were found by sequence analysis to resemble transfer genes of other Streptomyces plasmids and shown by Tn5 insertion mutagenesis to encode functions necessary for SLP1-mediated transfer. Furthermore, SLP1-mediated gene transfer was shown to be repressed by imp.

The spdB2, traA, and traB genes of SLP1 resemble similarly designated genes of Streptomyces plasmid pJV1 and Streptomyces plasmid pSN22: The TraB protein of pSN22 contains a functional nucleoside triphosphate-binding motif and is localized in the cytoplasmic membrane (25). A structural motif that has features in common with the tra gene product of Streptomyces plasmid pIJ101 (22) and the spoIIIE gene product of B. subtilis (38), which is responsible for chromosome translocation during prespore formation of B. subtilis, was identified within traB of SLP1. The pIJ101 tra gene and the traB genes of pJV1 and pSN22 specify functions lethal to the plasmid or host (19, 22, 33), and SLP1 traB may function similarly. Although the spdB1 (orf261) gene of SLP1 shows no significant structural homology to spdB1 of pJV1 or pSN22, all three genes affect pock size. Furthermore, in spdB1 and spdB2 in all three plasmids, the stop codon of the first ORF overlaps the start codon of the second. orf294 is also a gene involved in spreading. Regulation of the SLP1 transfer region presents similarity with the regulation of transfer genes of pJV1 and pSN22: the SLP1 impA function is similar to that of traR of pJV1 and pSN22, since these genes are all autoregulated and function as repressors of transfer genes. Despite the above similarities between the transfer loci of plasmids SLP1, pJV1, and pSN22, their structural organizations are not identical (Fig. 6).

View larger version (14K): [in this window] [in a new window]   FIG. 6.   Comparison of organizations of the transfer regions of SLP1 and other Streptomyces plasmids. Positions and orientation of ORFs and genes are shown for SLP1, pJV1 (33), and pSN22 (19). Arrows represent the orientation of fragments or the direction of transcription. The number of amino acids in the ORFs is indicated in parentheses below the name of each ORF.

The transfer genes of SLP1 resembles those of conjugative Streptomyces plasmids pJV1 and pSN22, suggesting that these plasmids may be derived from a common ancestor, but are different from the transfer genes of pSAM2, which like SLP1 is an integrating plasmid. Thus, SLP1 can be viewed as a genetic element composed of discrete modules: a transfer segment similar to that present in Streptomyces nonintegrating conjugative plasmids of the pJV1 and pSN22 family, and an integration/excision module containing recombinase genes similar to those of temperate bacteriophages (7, 8) and other Streptomyces integrating elements (e.g., pSAM2).

Our results indicate that the expression of imp genes in recipients sharply reduces the frequency of events associated with SLP1-mediated transfer of both plasmid and chromosomal genes. As imp has been found in our experiments to repress genes associated with transfer, this finding is most simply interpreted as being at least in part the result of such repression. However, as transfer frequency is measured by stable persistence of transferred genes in the recipient, which requires integration of transferred genes into the recipient's chromosome, events that affect chromosomal integration of genes in the recipient also have the potential to influence determinations of frequency. Because SLP1-mediated transfer of chromosomal genes, as well as of plasmid genes, is affected by expression of imp in the recipient, any affect of imp on gene integration in the recipient would necessarily have to involve both general recombination mechanisms and site-specific plasmid integration mediated by the SLP1-encoded int recombinase (7, 8).

Based on our results, we propose the following model. When SLP1 is integrated into the chromosome, the expression of imp genes represses both plasmid DNA replication (15) and expression of transfer genes, preventing SLP1-mediated transfer of chromosomal genes as well as the integrated plasmid. We suggest that donors and recipients meet independently of the expression of plasmid transfer genes and that the interaction between donor and recipient cells has a key role in activating SLP1 transfer genes in donors. According to this model, contact between donor and recipient must necessarily precede the expression of plasmid imp-regulated loci that mediate high-frequency gene transfer, instead of being the result of expression of those genes.

Potentially, such cell-cell contact or fusion between donor and recipient could dilute Imp proteins, causing derepression of the tra and spd genes controlled by imp, and consequently promoting gene transfer. Alternatively, information acquired from the recipient may alter imp expression or activity by other signal transduction mechanisms. In any case, our results imply that the initial stage of mating in Streptomyces (i.e., cell-cell contact or fusion) is not actively induced by the plasmid genes that facilitate gene transfer events. As chromosomal gene transfer can occur at low frequency during imp-mediated repression of plasmid-borne tra and spd genes, simple cell contact or fusion is sufficient for such transfer. Indeed, chromosomal gene transfer can occur at low frequency even in the absence of plasmids (reference 30 and Table 3). Our results further suggest that derepression of imp-regulated plasmid genes following interaction between donor and recipient is responsible for the efficient transfer of plasmid and chromosomal genes during mating between Streptomyces cells that contain plasmids. This two-stage model for gene transfer in Streptomyces potentially explains the distinctly different frequencies observed for plasmid-mediated (enhanced) gene transfer versus the chromosome gene transfer (i.e., chromosome-mobilizing activity) occurring in the absence of plasmids.