Region-Specific Transcriptional Activity in the Genome of Streptomyces coelicolor A3(2)

Abstract

Transcriptional analysis of microbial genomes is an important component of functional genomics. Strategies such as hybridization of labeled total RNA against ordered clone libraries or differential-display approaches have already been carried out to identify expressed genes. We describe here an additional method which applies subtractive hybridization between genome-specific DNA and total RNA followed by a PCR approach to identify expressed microbial genes. With the new strategy, the expression of genes in the terminal regions of the linear Streptomyces coelicolor A3(2) chromosome and the accessory linear plasmid SCP1 was analyzed. The results indicate that the method is useful for the identification of expressed genes in actinomycetes and other microbial systems. We demonstrate for the first time that at least 24 genes in the chromosome end regions (silent regions) of S. coelicolor are actively expressed. In addition, several expressed SCP1 genes were identified, including a gene which shows high similarity to microbial dnaN genes and which seems to play a role in SCP1 maintenance.

Several approaches have been used to investigate global transcription of microbial genomes, taking advantage of the availability of ordered clone libraries or total genome sequence data. Chuang et al. (4) hybridized labeled reverse-transcribed RNA from Escherichia coli against an ordered clone library to identify genes induced by external stimuli, such as heat shock or osmotic shock. Tao et al. (18) refined this system by generating microarrays, including the 4,290 individual E. coli open reading frames (ORFs), which were amplified by PCR and immobilized on membranes. Similar studies have been reported for the expressional analysis of ORFs from Haemophilus influenzae and Streptococcus pneumoniae on the basis of chip technology (7). Recently, Fislage et al. (8) and Gill et al. (9) demonstrated that the E. coli genome sequence information can be used to design several primers for a differential-display strategy to clone genes from E. coli which show a specific induction pattern.

Except for the method of Chuang et al. (4), all of the above-mentioned approaches rely on the availability of sequence information to accurately analyze single-gene expression and require expensive equipment, as in the case of microarray systems. We have therefore started to search for fast and inexpensive alternative strategies to analyze the transcriptional activity of microbial genomes by the isolation and enrichment of rRNA-free mRNA. Suitable protocols for such strategies have been used in the analysis of eucaryotic genes, for example, the method of Korn et al. (13) for positional cloning of genes from the human X chromosome (Xq27.3 to Xq28). In this strategy, human cDNA was hybridized with X-chromosome-specific, labeled cosmid YAC or BAC clones. cDNA molecules which formed heteroduplexes with the labeled DNA were separated with magnetic beads. The eluted unlabeled cDNA was then amplified by a PCR approach and either cloned or used as a hybridization source.

We have established a modified strategy (13) for the isolation and enrichment of bacterial mRNA and applied the method to the transcriptional analysis of the Streptomyces coelicolor termini and the 360-kb linear plasmid SCP1. We demonstrate that the method can be used successfully to isolate and enrich microbial transcripts, which can be applied as hybridization probes to identify transcriptionally active regions in a microbial genome.

MATERIALS AND METHODS

Strains and growth conditions.

S. coelicolor M145 (SCP1⁻ SCP2⁻) and M138 (SCP1⁺ SCP2⁻) were used for the isolation of fragment-specific RNA and were grown and handled according to the method of Kieser et al. (12). Streptomyces lividans TK64 was used for transformation experiments (12). The strains were grown for RNA isolation in minimal medium as described previously (17). YEME medium was used to grow cultures for the preparation of DNA and protoplasts (12). Thiostrepton-resistant clones, after transformation with recombinant plasmid DNA, were selected by plating them on R2YE agar (12) and overlaying them with thiostrepton in a final concentration of 500 μg ml⁻¹. E. coli cells with cosmids of the ordered S. coelicolor chromosome library (15) and SCP1 (16) were grown in Luria-Bertani medium with 50 μg of ampicillin ml⁻¹ (14). The recombinant plasmids used for transformation of S. coelicolor strains were isolated from E. coli GM 48 (19).

DNA preparation, PFGE, and amplification of rpoA

Cosmid DNA was isolated by the alkaline lysis technique (2). Pulsed-field gel electrophoresis (PFGE) DNA was isolated and restricted according to the method of Redenbach et al. (15). Individual PFGE fragments from S. coelicolor for subtractive hybridizations were excised from low-melting-point gels (Seakem) and extracted with phenol-chloroform as described by the supplier.

The rpoA gene of S. coelicolor M145 was amplified with the rpoA forward primer, 5′-GACCGGCCTCGAGGTCGGCTC-3′, and the rpoA reverse primer, 5′-CAGTATCCGAGCAGGCGGCCG-3′, with 30 cycles of 1 min at 94°C, 1 min at 60°C, and 3 min at 72°C in a PCT-100 thermocycler (MJ Research).

Isolation of RNA and subtractive hybridization with magnetic beads.

Cosmids of the ordered M145 SCP1 library and eluted PFGE DNA were sonified with a Sonifier 250 (Branson) three times for 30 s each time at level 3 to generate fragment sizes of 400 to 800 bp. Ten micrograms of DNA from AseI cosmids A, J, and F, corresponding to the chromosomal termini of S. coelicolor, or from SCP1 cosmids was labeled by the random-primed method with DIG-11-dUTP (Roche). Alternatively, 10 μg of PFGE-eluted SCP1 DNA was labeled instead of the ordered SCP1 cosmids. Total RNA was isolated from S. coelicolor cultures by the method of Chomczynski and Sacchi (3). A total of 5 mg of denatured RNA was mixed with 10 μg of denatured digoxigenin (DIG)-labeled target DNA [cosmids corresponding to AseI-J, -F, and -A of S. coelicolor A3(2), rpoA gene, or SCP1 DNA] and hybridized in 2 ml of 2× SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate)–2× SSPE (1× SSPE is 0.18 M NaCl, 10 mM NaH₂PO₄, and 1 mM EDTA [pH 7.7]) including 50 U of RNase inhibitor (Fermentas) for 2 h at room temperature on a rolling shaker. Single-stranded RNA and DNA were degraded with 1,000 U of S1 nuclease (Promega) for 10 min at 37°C after the hybridization and prior to the magnetic-bead selection.

Immobilization of DNA-RNA hybrids on the basis of anti-DIG magnetic particles (Roche) was done according to the protocol of the supplier. Two milligrams of magnetic particles were washed with TEN 1000 (10 mM Tris-HCl [pH 7.5], 1 mM EDTA, 100 mM NaCl). The DNA-RNA hybrids were added to the magnetic particles and incubated for 10 to 30 min at room temperature on a rolling shaker. The particles were washed three times with 0.5 ml of TEN 1000 (10 mM Tris [pH 7.5], 1 mM EDTA, 1 M NaCl). Fragment-specific mRNA was eluted with 6 M guanidinium-HCl for 5 min at 50°C. To exclude any DNA contamination due to the elution of the mRNA, a DNase treatment with RQ1-DNase (Promega) was used to reveal only fragment-specific RNA.

Reverse transcription of fragment-specific transcripts.

Fragment-specific mRNA was reverse transcribed with Expand reverse transcriptase (Roche) as described by the supplier using one of three primers: R6 (5′-NNN NNN-3′), R9 (5′-NNN NNN NNN-3′), or RSSN (5′-SSN SSN SSN-3′) (N represents G, C, T, or A; S represents G or C). Enriched mRNA (500 ng) and 50 pmol of primers were incubated for 10 min at 65°C followed by annealing of the primers at 30°C for 10 min and transcription at 42°C for an additional 45 min with 50 U of reverse transcriptase in single-strand buffer (50 mM Tris-HCl [pH 8.3], 40 mM KCl, 5 mM MgCl₂, 0.5% Tween 20).

Second-strand synthesis was performed in a volume of 100 μl with second-strand buffer (80 mM Tris-HCl [pH 7.5], 240 mM KCl, 10 mM MgCl₂, 130 mg of bovine serum albumin ml⁻¹), 0.13 U of RNase H (Roche), and 25 U of E. coli DNA polymerase I (Roche). The reaction mixture was incubated for 1 h at 12°C, 1 h at 22°C, and 10 min at 65°C. After the addition of 4 U of T4 DNA polymerase (Roche), the reaction mixture was incubated for 10 min at 37°C. The reaction was stopped with 10 μl of 0.2 M EDTA (pH 7.2) and 2 μl of 10% (wt/vol) sarcosyl.

cDNA amplification and size selection.

Phenol-chloroform-purified cDNA molecules were ligated in a 1:1 ratio with the adapter (5′-AATTCGGCAACGAATTAATCCATGGT-3′ and 3′-GCCGTTGCTTAATTAGGTACC-5′). The ligation product was diluted 1:1, 1:10, 1:50, 1:100, 1:250, 1:500, and 1:1,000 and used for PCR. For the PCR, 100 pmol of adapter-specific primers was mixed with 200 mM deoxynucleoside triphosphates, 2.5 mM MgCl₂, and 2 U of Taq polymerase (Fermentas) and amplified in 20 to 25 cycles of 1 min at. 94°C, 1 min at 55°C, and 1.5 min at 72°C. The final product was applied on a Chromo Spin column (Clontech) to exclude all fragments smaller than 150 bp and used again for a second round of subtractive hybridization and PCR as described above.

Sequencing and primer extension.

The Thermo-Sequenase fluorescently labeled primer cycle-sequencing kit with 7-deaza-dGTP (Amersham) was used for sequence reactions, which were applied to a LI-COR 4000L sequencer (MWG). Sequencing of fragments up to 500 bp was done on an ABI 373 sequencer using the Prism ready-reaction kit (Applied Biosystems). For primer extension experiments, the nonradioactive method of Altermann et al. (1) was applied using fluorescently labeled primers.

DNA labeling and filter hybridization.

Nonradioactive probes were done using a DIG-labeling kit (Roche). For radioactive labeling, 10 to 20 μCi of [α-³²P]dCTP (3,000 Ci mmol⁻¹; NEN) was used with the High-Prime labeling kit (Roche). Unincorporated nucleotides were separated with a Sephadex G-50 spin column. Hybridizations with DIG-labeled probes were performed as described by the supplier. Radioactive probes were hybridized against Northern, Southern, and colony blots by the protocol of Church and Gilbert (5). Filters were washed twice for 15 min at 65°C (0.5% [wt/vol] bovine serum albumin, 1 mM EDTA, 5% [wt/vol] sodium dodecyl sulfate, 40 mM NaHPO₄ [pH 7.2]) and once for 15 min at room temperature (1 mM EDTA, 1% [wt/vol] sodium dodecyl sulfate, 40 mM NaHPO₄ [pH 7.2]) and exposed to a Biomax MS film (Kodak) in a cassette with Biomax MS intensifying screens (Kodak) at −70°C.

Nucleotide sequence accession number.

The GenBank accession number of the 1,101-bp large gene identified in this study is AF235031.

RESULTS

Establishment of method for enrichment of region-specific transcripts.

The basic strategy of the method for enrichment of region-specific transcripts relies on the hybridization of total RNA from a microorganism with DIG-labeled DNA of specific cosmids or PFGE fragments from particular genomic regions which are devoid of ribosomal operons (Fig. 1). During hybridization, specific DNA-RNA heteroduplex molecules will be formed which can be selectively isolated by magnetic particles coated with anti-DIG Fab fragments. Prior to elution of the mRNA from the immobilized heteroduplexes on the magnetic beads, an S1 nuclease digestion is performed to degrade all single-stranded RNA and DNA molecules which were not bound to the surface of the magnetic particles. Subsequent release of the bound mRNA from the particles will be followed by a DNase I digestion, providing fragment-specific mRNA molecules that can be used for reverse transcription with random primers and subsequent second-strand synthesis by the Gubler and Hoffman approach (10). The resulting cDNA products are then ligated with specific adapters, and a PCR with adapter-specific primers can be done to amplify the generated cDNA fragments. The product of this amplification can then be used for a second cycle of hybridization and immobilization for an additional enrichment of microbial transcripts.

FIG. 1.

Procedure for the generation of fragment-specific cDNA sources. DIG-11-dUTP-labeled single-stranded DNA (solid lines with circled Ds) is hybridized with total RNA (dotted lines). Heteroduplex molecules are immobilized on magnetic beads (M) coated with anti-DIG after S1 nuclease treatment. The molecules are subsequently eluted, treated with DNase I, and taken for first-strand DNA synthesis using random primers. Double-stranded DNA synthesis is achieved according to the method of Gubler and Hoffman (10). The generated cDNA molecules (solid double lines) are ligated with an adapter (solid boxes). The ligation products are amplified with adapter-specific primers and taken for a second cycle of immobilization and PCR.

Selective enrichment of rpoA from S. coelicolor M145.

The constitutively expressed rpoA gene of S. coelicolor was used as a test system to demonstrate the efficiency of the enrichment strategy. cDNA from S. coelicolor M145 total RNA without any enrichment and cDNA after immobilization with the amplified rpoA gene as a target and following the first and second PCR steps was cloned in pBluescript (Stratagene) by TA cloning. The proportion of positive recombinants was estimated by colony blotting. Seventy-five percent of the recombinant clones were positive after one cycle of immobilization and PCR, and 90% were positive after two cycles, compared to less than 1% positive without selection. The rpoA cDNA clones contained fragments of 150 and 800 bp. This indicated that individual microbial transcripts can be enriched by this method.

Enrichment of transcripts from the terminal regions of the S. coelicolor M145 chromosome with clones of the ordered library.

To test the approach on larger genomic stretches, we used the ordered cosmids from the AseI A, F, and J regions of the S. coelicolor chromosome, corresponding to the two terminal regions, as targets to search for expressed sequences. Figure 2 shows a representative hybridization of dot blot filters containing DNA of the AseI A, F, and J cosmids with labeled M145 cDNA, which was obtained after two cycles of PCR with different primers and immobilization conditions. There is a clear enrichment in signal intensity in comparison to the control using cDNA from total RNA without any selection. In addition, it can be shown that the choice of random primers for reverse transcription has an influence on the enrichment. cDNA sources generated with oligonucleotides with a strong GC bias gave stronger signals than complete random primers, reflecting the high GC content of the actinomycetes DNA. The labeled cDNA probes from the terminal cosmids were hybridized against a SalI shotgun library of the cosmid DNA used, yielding 144 recombinant clones with inserts of 200 bp to 5 kb, which were sequenced. For small fragments (<1 kb) the complete sequence could be obtained, which allowed the unambiguous identification of a single transcriptionally active sequence. Seventy-six of the clones matched sequences in the S. coelicolor genome sequence database, and 24 small SalI fragments which could be completely sequenced were assigned to an individual gene. Expression of those candidates was confirmed by Northern analysis with total S. coelicolor M145 RNA (data not shown).

FIG. 2.

Hybridization of radioactively labeled cDNA generated under different conditions to dot blots of clones from the ordered cosmids of the AseI A, F, and J fragments of S. coelicolor M145. (A) Distribution of overlapping cosmid clones of the AseI fragments A, F, and J from S. coelicolor M145. (B) Hybridization of reverse-transcribed total RNA from S. coelicolor M145 without immobilization and amplification. Ten milligrams of total RNA was labeled by reverse transcription using the RSSN primer and hybridized against a dot blot with cosmid DNA of the AseI A, F, and J clones. (C to E) Hybridization of fragment-specific cDNA, which was generated using the R6 primer (C), the R9 primer (D), and the RSSN primer (E). Five hundred nanograms of AseI A-, F-, and J-specific mRNA was reverse transcribed, enriched by two rounds of immobilization and amplification (PCR), and radioactively labeled with 20 μCi of [α-³²P]dCTP.

Enrichment and analysis of transcripts of the linear plasmid SCP1 from S. coelicolor M138.

Eluted PFGE DNA of SCP1 or the ordered SCP1 cosmids was used to isolate SCP1-specific transcripts as described above. Interestingly, PFGE DNA-enriched SCP1 cDNA yielded stronger hybridization signals than the cDNA probes obtained with the cosmid DNA (data not shown). Hybridization of labeled SCP1 cDNA against filters with cosmid DNA digested with various enzymes indicated that specific fragments hybridized strongly and others poorly or not at all, as expected for a probe based on expressed sequences (Fig. 3).

FIG. 3.

Hybridization of radioactively labeled enriched cDNA from the linear plasmid SCP1 against a filter with SalI-digested SCP1 cosmid DNA. (A) Gel with SalI-digested clones. Lane 1, 1-kb marker (Fermentas); lane 2, C31; lane 3, C11; lane 4, C22; lane 5, C17; lane 6, C32; lane 7, C39; lane 8, C45; lane 9, C35; lane 10, C73; lane 11, C63; lane 12, C53; lane 13, 1-kb marker. (B) Southern filter with SalI-digested cosmid DNA which was hybridized with radioactively labeled cDNA from SCP1-enriched RNA. Lane 2, C31, lane 3, C11; lane 4, C22; lane 5, C17; lane 6, C32; lane 7, C39; lane 8, C45; lane 9, C35; lane 10, C73; lane 11, C63; lane 12, C53.

To analyze the expressed sequences more directly, SalI and BamHI shotgun libraries of the SCP1 cosmids 11, 17, 22, 31, 32, 35, 39, 45, 53, 63, and 73 were screened with the labeled SCP1 cDNA. Thirty-nine positive clones from the libraries with inserts of 300 bp to 6.2 kb were identified and sequenced. For 13 of these, matches to previously annotated sequences were found. Surprisingly, we detected among the sequenced SCP1 subclones a homolog to a β subunit of DNA polymerase III (dnaN). Primer extension verified the expression of the 1,101-bp large gene, which has transcriptional start sites 47, 50, and 61 bp upstream of a GTG start codon. We tried to inactivate the identified gene by a disruption experiment to evaluate whether it has any essential function in SCP1. A 300-bp internal fragment of the dnaN homolog was therefore cloned into pDH5 (11) and used to transform S. coelicolor M138. In addition, all analyzed SCP1 fragments which were identified by the transcriptional-analysis approach were cloned into pDH5 and transformed in S. coelicolor M138. All clones with SCP1 inserts, except the partial dnaN-like gene, were integrated successfully into SCP1, suggesting that the gene may play a role in the maintenance of SCP1.

DISCUSSION

Transcriptional analysis of microbial genomes is an important component of functional genomics. Chip technology provides a superb tool for the analysis of transcriptional activity and regulation but requires sequence data to generate microarrays of ORF-specific oligonucleotides or PCR-generated ORF fragments. A second approach is the analysis of global bacterial transcription with the help of differential-display methods (8, 9). Again, sufficient sequence information is required to generate efficient primer pairs which allow the amplification of transcripts. In addition, for a successful differential-display method, at least two different RNA sources (wild-type RNA and RNA of the induced system) are necessary to identify differentially expressed genes.

In contrast, the strategy for transcriptional analysis we describe here focuses on specific regions in microbial genomes, for example, large extrachromosomal elements or the ends of linear actinomycetes chromosomes. The isolation and enrichment of rRNA-free mRNA does not require sequence information, expensive equipment, or the application of polyacrylamide gel electrophoresis. The process should in theory recover all transcriptionally active regions in a selected area, which contrasts with the amplification of mRNA on a statistical basis in the differential-display method. The basic prerequisite, therefore, is a DNA target free of any ribosomal operons which can be used in a hybridization to specifically bind the complementary mRNA. This target DNA could be a clone population, as used here for the 2-Mb large terminal regions of the chromosome from S. coelicolor M145, or a gel-eluted PFGE fragment, for example, the linear plasmid SCP1.

Our data indicate that it is possible with the applied method to enrich specific bacterial transcripts that can serve as putative probes for hybridization of random shotgun libraries. It would be beneficial to clone the enriched cDNA population directly, generating bacterially expressed sequence tags. We are presently trying to optimize the procedure to establish the efficient direct cloning of the obtained bacterial cDNA.

With the help of the method described here, we are already able to demonstrate that at least 24 expressed genes are present in the dispensable terminal regions of the S. coelicolor genome under the culture conditions used. This is, to our knowledge, the first report of global transcriptional analysis of the chromosome ends of a Streptomyces species. Dary et al. (6) studied, at the protein level, the gene expression of mutants from Streptomyces ambofaciens which had lost at least 400 kb of the chromosome ends. They were able to show that at least 31 proteins were missing in the analyzed deletion mutants, although they could not prove that the genes of those proteins affected were located in the deleted area. In contrast, our data clearly indicate that even under normal culture conditions a certain number of genes of the so-called “silent arc region” are actively expressed.

We could also demonstrate that the strategy is useful to gain rapid information about the active genetic content of linear extrachromosomal DNA elements, such as SCP1. In addition to a diverse group of putative genes, we have identified a homolog to the β subunit of DNA polymerase III. It is the first gene of that class which has been found on a linear actinomycetes plasmid. The dnaN-like gene showed surprisingly low homology at the DNA sequence level to the dnaN gene of S. coelicolor, suggesting that this copy derived from a different microbial source, although our unsuccessful disruption experiments suggest that the identified gene is in some way responsible for SCP1 maintenance in the cell.