Control of rep Gene Expression in Plasmid pGA1 from Corynebacterium glutamicum

Abstract

The cryptic multicopy plasmid pGA1 (4,826 bp) from Corynebacterium glutamicum LP-6 belongs to the fifth group of rolling-circle-replicating plasmids. A determinant, which negatively controls pGA1 replication, was localized in the leader region of the rep gene coding for the initiator of plasmid replication. This region, when cloned into the compatible vector pEC6, was found to cause decrease of segregational stability of the pGA1 derivative pKG48. A promoter and a single transcriptional start site were found in the rep leader region in orientation opposite to the rep gene. These results suggest that a small countertranscribed RNA (ctRNA) (ca. 89 nucleotides in length), which might inhibit translation of pGA1 rep gene, is formed. Analysis of predicted secondary structure of the pGA1-encoded ctRNA revealed features common with the known ctRNAs in bacteria. Inactivation of the promoter P-ctRNA caused a dramatic increase of copies of the respective plasmid, which proved a negative role of the ctRNA in control of pGA1 copy number. A region between the promoters Prep and P-ctRNA with a potential to form secondary structures on both ctRNA and rep mRNA was found to cause low activity of the rep promoter even when promoter P-ctRNA was deleted. Thus, the sequence within the rep leader region itself seems to act, in addition to the ctRNA, as a second regulatory element of a novel type, negatively influencing expression of the pGA1 rep gene.

Bacterial plasmids as extrachromosomal, self-replicating DNA molecules, encode elements controlling their replication. To define and to maintain their characteristic copy number in a given host, plasmids replicating in both theta-type and rolling-circle modes use negative regulatory circuits (34). Inhibition of plasmid replication by increasing the dosage of negatively acting elements in trans has been used to identify these negative regulators of plasmid replication (24).

The replication of rolling-circle-replicating (RCR) plasmids has been shown to be regulated by negative control of synthesis or of activity of their replication initiators (Rep proteins), which are rate limiting for replication. Small (50 to 250 nucleotides [nt]), untranslated, short-lived, and constitutively synthesized, countertranscribed RNAs (ctRNAs) complementary to the leader sequences of rep mRNAs are the main regulatory elements involved in negative control of the rep gene expression (18, 35). In plasmids of the pT181 family and in plasmid p353-2 (pC194 family), one or two ctRNAs cause premature termination (attenuation) of rep mRNA synthesis (25, 32). RCR plasmids of the pMV158 family, and also some plasmids belonging to the pC194 family, use ctRNA regulation at the translational level mediated by interference of ctRNA with Rep synthesis by sequestering the ribosome-binding site of rep mRNA (8, 20). In plasmids of the pMV158 family, still another mechanism of negative control of rep expression involves binding of a small repressor protein (CopG) to the rep promoter (7).

The cryptic multicopy plasmid pGA1 (4,826 bp) from Corynebacterium glutamicum LP-6 (39) replicates by the rolling-circle mechanism. Its minimal replicon (1.7 kb) contains the rep gene, coding for the Rep protein (22), and the double-strand origin of replication (1). The deduced amino acid sequence of the pGA1 Rep protein is highly homologous to those of the Rep proteins of other small cryptic plasmids from corynebacteria, pSR1 from C. glutamicum (2) and pYM2 from C. efficiens (GenBank accession no. AB084384). Similarity of these Rep proteins to the Rep proteins of larger plasmids from corynebacteria, namely, pNG2 from C. diphtheriae (45), pAG1 from C. glutamicum (41), pTP10 from C. striatum (40), pCG4 from C. glutamicum (GenBank accession no. AF164956), pTET3 from C. glutamicum (GenBank accession no. AJ420072), and five plasmids from C. jeikeium (pK43, pCJ84, pK64, pKW4, and pB85766; GenBank accession numbers AF364477, AY048596, AY079086, AF401314, and AF486522, respectively), was also found. Since no significant homology between these Rep proteins and the Rep proteins of other RCR plasmids belonging to the four basic families exists, these Corynebacterium plasmids were proposed to form the fifth group of RCR replicons (26).

Plasmid pGA1 contains two genes, per and aes, located outside its minimal replicon, whose products influence positively stable maintenance of the plasmid. The per gene, whose expression was found to be negatively controlled at the transcriptional level by the pGA1 gene product(s), is the main determinant involved in stable maintenance of pGA1. When present in trans, per stimulates an increase in copy number and in segregational stability of the low-copy pGA1 derivatives lacking this gene (22). The aes gene was found to code for an accessory element involved in the stable maintenance of the plasmid pGA1 (44).

To find a regulatory element negatively controlling replication of the plasmid pGA1, we analyzed in detail structure and function of the region upstream of its rep gene. This analysis proved the negative role of a small ctRNA in control of plasmid pGA1 copy number and the negative effect of a sequence in the rep leader region with potential to form secondary structures on the activity of the rep promoter.

MATERIALS AND METHODS

Bacterial strains, plasmids, and culture conditions.

Bacterial strains used in the present study were Escherichia coli DH5α (14) and C. glutamicum R127 (19). Plasmids used are listed in Table 1. E. coli strains were grown at 37°C in Luria-Bertani medium (36), C. glutamicum strains were grown at 30°C in Casitone-yeast extract (CY) medium (29). Selective conditions in the media were obtained by using antibiotics at the following concentrations: kanamycin, 20 μg/ml; streptomycin, 20 μg/ml; spectinomycin, 20 μg/ml; and chloramphenicol, 10 μg/ml (for C. glutamicum) and 20 μg/ml (for E. coli).

TABLE 1.

Plasmids used in this work

Plasmid	Relevant characteristicsa	Source or reference
pK19	Kmr; E. coli vector (2.7 kb)	33
pS19	Smr Spr; E. coli vector (3.0 kb) (pK19 derivative, Kmr determinant was exchanged for Smr Spr determinant from pCG4)b	This work
pGA1	Cryptic plasmid from C. glutamicum LP-6 (sequence: GenBank/EMBL accession no. X90817)	22, 39
pKG48	pGA1 linearized in EcoRI site (nt 1788) in pK19	22
pSG48	pGA1 linearized in EcoRI site (nt 1788) in pS19	This work
pKG32	BamHI fragment of pGA1 (nt 1607-4826) in pK19	22
pK32PctRNAmut	pKG32 with mutation in the −10 hexamer of promoter P-ctRNA	This work
pET2	Kmr; E. coli-C. glutamicum promoter-probe vector (7.5 kb, replicon of pBL1, promoterless cat gene)	42
pET2-PEXctRNA	Promoter P-ctRNA (PCR fragment of pGA1, nt 2395 to 2219) in pET2	This work
pET2-PctRNA1	Promoter P-ctRNA (PCR fragment of pGA1, nt 2395 to 2198) in pET2	This work
pET2-PctRNA2	Promoter P-ctRNA (PCR fragment of pGA1, nt 2395 to 2172) in pET2	This work
pET2-PctRNA3	Promoter P-ctRNA (PCR fragment of pGA1, nt 2395 to 2159) in pET2	This work
pET2-PctRNA4	Promoter P-ctRNA (PCR fragment of pGA1, nt 2395 to 2151) in pET2	This work
pET2-PctRNA5	Promoter P-ctRNA (PCR fragment of pGA1, nt 2395 to 2147) in pET2	This work
pET2-PctRNAmut	Promoter P-ctRNA with mutated −10 hexamer (PCR fragment of pGA1, nt 2395 to 2172) in pET2	This work
pET2-Prep1	Promoter Prep (PCR fragment of pGA1, nt 2007 to 2150) in pET2	This work
pET2-Prep2	Promoter Prep (PCR fragment of pGA1, nt 2007 to 2212) in pET2	This work
pET2-Prep3	Promoter Prep (PCR fragment of pGA1, nt 2007 to 2230) in pET2	This work
pET2-Prep4	Promoter Prep (XbaI-NheI fragment of pGA1, nt 1979 to 2275) in pET2	This work
pET2-Prep4(PctRNAmut)	Promoter Prep and promoter P-ctRNA with mutated −10 hexamer (XbaI-NheI fragment of pGA1, nt 1979 to 2275) in pET2	This work
pEC6	Cmr; derivative of pEC5c with unique BamHI site	B. J. Eikmanns
pEC6-LR	PCR fragment of pGA1 containing rep gene leader region (nt 2151 to 2395) in pEC6	This work

^aKm^r, kanamycin resistance; Sm^r, streptomycin resistance; Sp^r, spectinomycin resistance; Cm^r, chloramphericol resistance.

^bNešvera et al. (21).

^cEikmanns et al. (10).

DNA manipulations and transformation of E. coli and C. glutamicum.

Plasmid DNA from E. coli was isolated by the method of Birnboim and Doly (5). Plasmid DNA from C. glutamicum was isolated by a modified alkaline extraction procedure with lysozyme (30). DNA fragments were amplified by the PCR technique. Restriction enzymes and T4 DNA ligase were used as recommended by the manufacturers (New England Biolabs or Fermentas). The nucleotide sequences were determined by using automatic sequencer Vistra (Amersham). Transformation of E. coli DH5α was carried out by the method of Hanahan (14), and electrotransformation of C. glutamicum was done according to the method of Liebl et al. (19).

Site-directed mutagenesis.

Inactivation of the −10 hexamer of promoter P-ctRNA was performed by site-directed mutagenesis (17). Two complementary primers, PCTRNAMD (5′-AAGAAATGTCCCTGGCCGAT-3′) and PCTRNAMR (5′-ATCGGCCAGGGACATTTCTT-3′), covering positions 2236 to 2255 of pGA1 (GenBank accession no. X90817), and two external primers, pUC/M13 Reverse (Promega) and PGANRUD (5′-AGGCGCTCGCGAAATCTC-3′) covering positions 2433 to 2416 of pGA1, were used in PCRs. The resulting alteration of the −10 hexamer of the promoter P-ctRNA (TAGAGT → CAGGGA) changed only the second codon of the rep gene, leading to exchange of threonine for serine in the Rep protein. To obtain the plasmid pKG32PctRNAmut, the final PCR product (739 bp), carrying unique PstI and NruI restriction sites, was used to replace the wild-type PstI-NruI fragment in plasmid pKG32.

RNA isolation and primer extension analysis.

Total RNA from C. glutamicum R127 containing plasmid pET2-PEXctRNA was isolated according to the method of Eikmanns et al. (11). Oligonucleotide CM4 (5′-GAAATCTCGTCGAAGCGTCG-3′), covering positions +21 to +40 relative to the translational start site of the cat gene in plasmid pET2, was used as a primer. It was terminally labeled with [γ-³²P]ATP by using T4 polynucleotide kinase (Amersham). The primer extension reaction was done by the method of Peters-Wendisch et al. (31).

Determination of the incompatibility effect caused by a cloned gene.

Incompatibility effect of a cloned gene in biplasmid C. glutamicum cells was determined by monitoring maintenance of the resident plasmid during cultivation under conditions selective for the plasmid carrying the possible determinant negatively controlling replication. The overnight cultures were serially diluted (1:200) in fresh CY medium without antibiotic, and the cells were plated on CY plates without antibiotic. After growth to single colonies, replica plating on CY plates containing antibiotics selective for only the tested plasmid (kanamycin) and for both plasmids (kanamycin and chloramphenicol) was performed. The percentage of the kanamycin-resistant (Km^r) clones was determined, and the presence of plasmid DNA was checked in selected isolates.

Determination of plasmid copy number and plasmid maintenance.

The number of plasmid copies per C. glutamicum chromosome was determined in total lysates prepared by the method of Noirot-Gros and Ehrlich (23) from plasmid-harboring C. glutamicum cells grown in the presence of kanamycin. The DNA samples were run on 0.8% agarose gel and then stained with SYBR-Green I (Sigma). The electrophoreograms were scanned by Kodak Gel Scanner and analyzed by the software supplied with the scanner. Estimations of plasmid copy number per chromosome were done by using the following equation: [(signal of plasmid CCC + OC forms × 1.36)/(signal of chromosome)] × [(size of C. glutamicum chromosome/plasmid size)], where CCC is the covalently closed-circle form of the plasmid and OC is the open-circle form of the plasmid. A value of 3,309 kb (GenBank accession no. NC_003450) was used as the size of C. glutamicum chromosome.

For determination of plasmid maintenance, plasmid-harboring C. glutamicum strains grown overnight were diluted (1:200) in fresh CY medium and, after serial dilutions, the cells were plated on CY medium without antibiotics and then replica plated on CY plates containing kanamycin. The percentage of Km^r clones was determined, and the presence of plasmid DNA was checked in selected isolates.

Enzyme assay and determination of MIC.

The specific activity of chloramphenicol acetyltransferase (CAT) in C. glutamicum cell extracts was determined by the method of Shaw (38). The MIC of chloramphenicol in E. coli and C. glutamicum strains was determined on Luria-Bertani and CY plates, respectively, by the method of Ozaki et al. (27).

Calculation of theoretical efficiency of transcription termination.

Free energy (ΔG) of formation of hairpin RNA structures (computer predicted by using the DNAstar-GeneQuest program) was calculated according to the method of Freier et al. (13). The theoretical efficiency of termination was calculated by using the algorithm developed by d'Aubenton-Carafa et al. (6).

RESULTS

A determinant negatively controlling plasmid pGA1 replication is located upstream of the rep gene.

Determinants negatively controlling replication of RCR plasmids are predominantly located upstream of the rep genes (18). To find such determinant(s) on plasmid pGA1, the fragment of pGA1 DNA (nt 2151 to 2395 in Fig. 1) covering a substantial part (90 of 93 nt) of the rep gene leader region, was cloned into the multicopy vector pEC6, a derivative of pBL1 (37) compatible with pGA1. A sequence with potential to generate a secondary structure that may serve as a rho-independent transcriptional terminator in a direction opposite to the rep gene is present in the rep leader region (structure IR1′ in Fig. 1). The resulting construct pEC6-LR was introduced into C. glutamicum R127 harboring plasmid pKG48 (which contains the complete pGA1 sequence), and maintenance of pKG48 was monitored. After 25 generations of growth under conditions selective for pEC6-LR, only 58% of the C. glutamicum cells containing pEC6-LR harbored the plasmid pKG48 also. On the other hand, 100% of the C. glutamicum cells containing vector pEC6 harbored also pKG48 under the same conditions. This incompatibility effect indicates that the cloned DNA fragment carries a trans-acting determinant negatively controlling plasmid pGA1 replication.

FIG. 1.

Part of the pGA1 sequence relevant to the data presented (the coordinates used are those from the GenBank/EMBL pGA1 sequence, accession number X90817). The DNA sequence of the rep gene leader region (LR) is marked in boldface; “+1” indicates the first base of the transcript, which is shown as a boldface underlined letter; and the putative −10 and −35 hexamers of the promoters are underlined. IR, inverted repeat; SD, putative Shine-Dalgarno site whose sequence is underlined. The start codon of rep gene is underlined; the first 9 amino acids of the Rep protein are shown; and the enzyme sites are marked in italics. E1, E2, E3, E4, and E5 indicate the positions of the 3′ ends of the fragments cloned into the vector pET2 (see the resulting plasmids pET2-PctRNA1, pET2-PctRNA2, pET2-PctRNA3, pET2-PctRNA4, and pET2-PctRNA5 in Table 3).

A promoter is present in the rep leader region in orientation opposite to the rep gene.

Plasmid pGA1 fragment (nt 2198 to 2395 in Fig. 1) containing part of the rep leader region but lacking the sequence IR1′, which resembles a rho-independent terminator (Fig. 1), was cloned into the promoter-probe vector pET2 in orientation opposite to that of the rep gene and tested for promoter activity in C. glutamicum cells. This fragment carried by the resulting plasmid pET2-PctRNA1 exhibited high transcriptional activity. The level of chloramphenicol resistance (MIC) was found to be 90 μg/ml. The corresponding CAT activity in cell extract was 0.462 ± 0.010 μmol · min⁻¹ · mg of protein⁻¹. When the vector pET2 was used as a control, the MIC and CAT activity values were 5 μg/ml and < 0.001 μmol · min⁻¹ · mg of protein⁻¹, respectively. To map the 5′ end of the transcript pertinent to this newly discovered promoter, primer extension analysis of total RNA isolated from C. glutamicum R127 harboring plasmid pET2-PEXctRNA (carrying fragment nt 2219 to 2395) was performed. A single band was detected as a result of reverse transcription with oligonucleotide primer complementary to the vector pET2 (Fig. 2). The transcriptional start site, mapped as base G in Fig. 2, is located 6 bp upstream of the translational start site of the pGA1 rep gene (Fig. 1). The putative promoter hexamers (−10, TAGAGT; −35, GTGACT) (Fig. 1 and 2) show similarity to those of the C. glutamicum promoter consensus sequence (TANAAT and TTGGCA, respectively [28, 43]). On the basis of these results, we conclude that pGA1 encodes a small countertranscribed RNA (ctRNA), complementary to the leader region of the rep mRNA, which can be involved in negative control of expression of the pGA1 rep gene. The newly discovered promoter, designated P-ctRNA, was found to be constitutive, since no negative effect of presence of plasmid pSG48, containing the complete pGA1 sequence, on its activity was observed. (CAT activity in cell extract of C. glutamicum harboring pET2-PctRNA1 plus pSG48 was 0.51 ± 0.05 μmol · min⁻¹ · mg of protein⁻¹.)

FIG. 2.

Transcriptional start site respective to the promoter P-ctRNA as determined by primer extension. Letters above represent the products of the DNA sequencing (A, C, G, and T represent the dideoxynucleoside triphosphates with which the sequencing was done). “P” indicates the product of primer extension; the DNA sequence read from the gel and its complementary sequence are shown on the right side. The respective transcriptional start site is shown in boldface (✽). The −10 hexamer is also indicated.

Effect of inactivation of the promoter P-ctRNA on plasmid copy number.

To test the effect of putative ctRNA on pGA1 copy number, inactivation of promoter P-ctRNA, abolishing synthesis of ctRNA, was performed. Its −10 hexamer TAGAGT was altered by site-directed mutagenesis to the sequence CAGGGA. The fragment containing the mutated −10 hexamer was cloned into the promoter-probe vector pET2. No promoter activity of this fragment was observed in cell extracts of C. glutamicum harboring the resulting plasmid pET2-PctRNAmut, which confirmed the essential role of the −10 hexamer for transcriptional activity (the MIC of chloramphenicol was 5 μg/ml, and the CAT activity was <0.01 μmol · min⁻¹ · mg of protein⁻¹). To test the phenotypic effect of this mutation, the region carrying wild-type P-ctRNA was exchanged for that carrying its mutated −10 hexamer in the low-copy-number and segregationally unstable derivative of pGA1, plasmid pKG32. As shown in Table 2, inactivation of promoter P-ctRNA caused approximately eightfold increase in the number of copies of the plasmid pKG32PctRNAmut in comparison with that of the original plasmid pKG32. The observed segregational stability of pKG32PctRNAmut, differing from extremely unstable maintenance of pKG32, is most probably a consequence of the increased number of plasmid copies. These results proved the negative role of the ctRNA in control of pGA1 copy number in C. glutamicum cell.

TABLE 2.

Number of copies per chromosome of C. glutamicum and maintenance of the derivatives of plasmid pGA1 in C. glutamicum

Derivative of pGA1	Presence (+) or absence (−) of:		Mean copy no.a ± SD	Maintenance after 25 generations (%)b
	ctRNA	per gene
pKG48	+	+	37.1 ± 7.1	100
pKG32	+	−	6.3 ± 3.5	<0.2
pKG32ctRNAmut	−	−	49.7 ± 9.5	100

^aValues represent the means from at least four independent experiments.

^bExpressed as the percentage of Km^r clones.

Analysis of the predicted secondary structure of the pGA1-encoded ctRNA and determining its length.

Computer-predicted secondary structure of the pGA1-encoded ctRNA shows the formation of two hairpin structures (Fig. 3A). The first hairpin (S1), which would start 4 nt downstream of the transcriptional start site of the ctRNA, possesses a 10-bp GC-rich stem with a small internal loop (ΔG = −11.5 kcal/mol). The sequence of the 5′ end of the stem is complementary to the putative Shine-Dalgarno sequence of the rep gene. The sequence at the loop of the predicted secondary structure (5′-CUGAUGA-3′) exhibits the U-turn motif (12), which fits the consensus sequence of U-turns in prokaryotic antisense RNAs or in their target RNAs (5′-YUNR-3′, where Y = pyrimidine, U = uracil, N = any base with preference of purine [most often G], and R = purine) (Fig. 3B). The U-turn motif is considered to promote RNA-RNA pairing (12). The second possible structure (T) might be formed 2 bp downstream of the first secondary structure. This hairpin consists of a 14-bp GC-rich stem and a 4-nt loop. Since five U residues are present 5 nt (instead of the usual ≤3 nt) downstream of its 3′ end, we assume that this hairpin structure acts as a nontypical rho-independent transcriptional terminator (6). The predicted value of the free energy of its formation was estimated (13) to be ΔG = −20.7 kcal/mol. The calculated value of the parameter “d” (6) is 24.475, which correlates with a theoretical efficiency of termination of ca. 75%.

FIG. 3.

(A) Computer-predicted secondary structure of pGA1-encoded ctRNA. The length of the putative ctRNA is given according to the results of measuring promoter activity of promoter-containing P-ctRNA fragments, which differ in length. S1, the first hairpin, which might be formed downstream of the 5′end of pGA1-ctRNA; T, the hairpin that resembles rho-independent transcriptional terminators. The sequence of the U-turn in S1, similar to the U-turn loop consensus motif in some prokaryotic antisense RNAs or in their target RNAs, is given in bold. The sequence complementary to the putative Shine-Dalgarno sequence of pGA1 rep mRNA is marked by vertical lines. (B) Similarity of the sequence of pGA1-encoded ctRNA U-turn to the U-turn loop consensus motif in some prokaryotic antisense RNAs or their target RNAs () (according to Franch et al. [12]). Y, pyrimidine; U, uracil; N, any base, with preference for purine (most often G); R, purine; the nucleotides forming the top-stem base pairs closing the loops are underlined. chr., chromosome. (C) Computer-predicted secondary structure of the RNA encoded by the rep gene leader region of plasmid pGA1. SD, putative Shine-Dalgarno site (underlined) of rep gene. The start codon of rep gene is given in boldface letters.

To determine the length of the pGA1-coded ctRNA, P-ctRNA-containing fragments, differing in length, were cloned into the promoter-probe vector pET2 and promoter activity was measured in cell extracts of C. glutamicum cells harboring the individual plasmid constructs. Decrease of activity was obviously an evidence for the presence of a terminator. As shown in Table 3, a substantial decrease of CAT activity was observed in C. glutamicum cells harboring plasmid pET2-PctRNA3 carrying the 77-bp P-ctRNA-containing fragment ending just downstream of five U residues. This result is in agreement with the predicted termination features of the above-described hairpin structure (T). Two constructs with longer P-ctRNA-containing fragments (pET2-PctRNA4 and pET2-PctRNA5) showed still more severe decrease of CAT activity. According to the results shown in Table 3, the size of ctRNA was estimated to be about 89 nt. The 3′ ends of the tested fragments are shown in Fig. 1.

TABLE 3.

Promoter activity of P-ctRNA-containing fragments differing in length

Plasmida	Size of ctRNA (nt)	MIC in C. glutamicum (μg of Cm/ml)b	Mean CAT activityc in C. glutamicum (μmol · min−1 · mg of protein−1) ± SD	Termination (%)
pET2		5	<0.001	—
pET2-PctRNA1	38	90	0.36 ± 0.002	0
pET2-PctRNA2	64	90	0.38 ± 0.009	0
pET2-PctRNA3	77	60	0.078 ± 0.001	78
pET2-PctRNA4	85	30	0.022 ± 0.003	94
pET2-PctRNA5	89	30	0.013 ± 0.003	96

^aThe 3′ ends of the cloned fragments in plasmids pET2-PctRNA1, pET2-PctRNA2, pET2-PctRNA3, pET2-PctRNA4, and pET2-PctRNA5 are shown in Fig. 1.

^bCm, chloramphenicol.

^cValues represent the means from at least four independent experiments.

Since RNA polymerase (RNAP) from C. glutamicum is not available, we have used RNAP from E. coli (Roche) for in vitro synthesis of pGA1-coded ctRNA. The reaction was performed according to del Solar and Espinosa (8, 9). A single major RNA band was detected on the polyacrylamide gel as a product of the runoff in vitro assay, which proved recognition of the promoter P-ctRNA by the E. coli RNAP. However, position of this band on the gel (RNAs of defined size were used as size markers) corresponded to the smaller size of ctRNA (65 nt) (data not shown) in comparison with that estimated in vivo in C. glutamicum cells (89 nt). The smaller ctRNA, determined by in vitro assay, could be formed due to different recognition of ctRNA transcriptional terminator (T) by E. coli RNAP since its structure differs from that of the typical E. coli rho-independent transcriptional terminators.

Influence of the sequence of the rep leader region on promoter Prep activity.

Activity of the promoter of the pGA1 rep gene, Prep, was found to be very low when the fragment containing also the promoter P-ctRNA was used (plasmid pET2-Prep4 in Table 4). To test whether this low activity of Prep was caused by synthesis of ctRNA, a DNA fragment with a deletion of the −10 and −35 hexamers of P-ctRNA, which abolishes synthesis of ctRNA, was cloned into the promoter-probe vector pET2 (plasmid pET2-Prep3). As shown in Table 4, this deletion had no positive effect on transcription from Prep. A similarly low CAT activity (0.009 ± 0.001 μmol · min⁻¹. mg of protein⁻¹) was measured in C. glutamicum harboring the plasmid pET2-Prep4(PctRNAmut) in which the mutation abolishing function of the −10 hexamer of P-ctRNA was introduced. However, when almost the whole rep leader region (90 of 93 nt) was deleted (plasmid pET2-Prep1), a 10-fold increase of Prep activity was observed (Table 4). Thus, the sequence of the rep leader region itself seems to act as another regulatory element negatively influencing expression of the pGA1 rep gene, probably due to the formation of two distinct secondary structures (hairpins) on the pGA1 rep mRNA (Fig. 3C). Results with plasmid pET2-Prep2 (Table 4), in which the sequence causing the second hairpin was deleted in the cloned pGA1 fragment, proved that the presence of the rep leader sequence responsible for the first (longer) hairpin (designated IR1 in Fig. 1, estimated value of ΔG = −18.5 kcal/mol) is sufficient to ensure low activity of the Prep. Very similar results (high activity of Prep in pET2-Prep1 and low activity of Prep in pET2-Prep2 and in pET2-Prep3) were obtained when promoter activity was measured in C. glutamicum harboring also plasmid pSG48 or in E. coli (data not shown). No interaction of the hairpin(s) with any effector encoded by plasmid pGA1 or by C. glutamicum chromosome is thus obvious.

TABLE 4.

Promoter activity of Prep-containing fragments differing in length

Plasmid	No. of secondary structures on rep mRNA	Presence (+) or absence (−) of ctRNA	MIC in C. glutamicum (μg of Cm/ml)a	Mean CAT activityb in C. glutamicum (μmol · min−1 · mg of protein−1) ± SD
pET2		−	5	<0.001
pET2-Prep1		−	60	0.147 ± 0.007
pET2-Prep2	1	−	20	0.012 ± 0.002
pET2-Prep3	2	−	20	0.014 ± 0.003
pET2-Prep4	2	+	20	0.011 ± 0.001

^aCm, chloramphenicol.

^bValues represent the means from at least four independent experiments.

DISCUSSION

Plasmid pGA1 from C. glutamicum, classified in the fifth group of RCR plasmids (26), was found to code for a unique regulatory element, the Per protein, that positively influences plasmid copy number and maintenance (22). Although the per gene expression is negatively controlled at the transcriptional level by pGA1 gene product(s) (22), this single regulatory mechanism does not seem to be sufficient to ensure the optimum plasmid copy number in C. glutamicum cells. Indeed, we have found two other negatively acting regulatory elements, both influencing expression of the rep gene coding initiator of pGA1 replication by rolling circle mode. As described here, the function of these two elements is necessary for negative control of plasmid pGA1 replication.

The negative effect of the pGA1 rep leader region, present on a compatible plasmid, on the maintenance of a pGA1 derivative was observed. It indicates that a trans-acting regulatory element is encoded by this region. We suggest that this trans-acting regulatory element is a ctRNA since (i) a strong constitutive promoter (designated P-ctRNA) and (ii) a GC-rich sequence, 36 bp downstream of the +1 base of the transcript, with potential to generate a secondary structure that could serve as a rho-independent transcriptional terminator, were found in this region in orientation opposite to the rep gene (see Fig. 1 and 2). This notion was strongly supported by the presence of a U-turn loop in the first potential hairpin structure (S1) of the putative ctRNA with a sequence highly similar to the consensus sequence of U-turns in prokaryotic antisense RNAs or in their complementary target RNAs (Fig. 3A and B). This U-turn loop is considered to promote pairing with the complementary sequence on the rep mRNA (Fig. 3C). Efficient interaction of pGA1-encoded ctRNA with the rep mRNA can be predicted also from its other structural features typical for regulatory ctRNAs (15, 16), namely, from the size of the S1 loop (7 nt) and from presence of an internal loop in the S1 stem (see Fig. 3A). According to the approximate size of the transcript starting from P-ctRNA (ca. 89 nt), determined by measuring promoter activity of fragments differing in length, the end of pGA1-encoded ctRNA seems to correlate with start of the rep mRNA (see Table 3 and Fig. 1). The smaller ctRNA (65 nt) was detected by in vitro assay using E. coli RNAP. It was found that terminators of C. glutamicum often differ from those of E. coli both by structure and function (3). Therefore, the smaller size of the in vitro ctRNA transcript was caused very probably by transcriptional termination driven by the heterologous RNAP differing in its action from that of the native RNAP.

Strong positive influence of inactivation of promoter P-ctRNA on copy number of the low-copy-number pGA1 derivative pKG32 (Table 2) proved that the product of the gene transcribed from this promoter (i.e., ctRNA) controls negatively number of pGA1 plasmid copies. Inhibition of synthesis of the Rep protein, a positively acting initiator of pGA1 replication, by interaction of ctRNA with the complementary leader sequence of the rep mRNA is apparently responsible for the negative effect of ctRNA on pGA1 copy number. A dramatic increase in pKG32 copy number, observed when inhibitor of the Rep protein synthesis was absent, was thus caused most probably by the increase of the Rep concentration in C. glutamicum cells. These results suggest the rate-limiting role of Rep for pGA1 replication, which is in agreement with data described in other RCR plasmids (4). A simultaneously observed positive effect of inactivation of promoter P-ctRNA on segregational stability of unstable plasmid pKG32 reflects most probably correlation between plasmid copy number and maintenance, which might be common for small plasmids lacking a particular stabilizing system.

Unlike the observed positive effect of inactivation of promoter P-ctRNA on pGA1 copy number, inactivation of this promoter has no positive effect on activity of the promoter Prep (Table 4). These results indicate that the pGA1-encoded ctRNA acts negatively at the posttranscriptional level. The ctRNA could interfere with Rep synthesis by sequestering the ribosome-binding site of rep mRNA. This hypothesis is supported by the fact that the sequence of the 5′end of the stem of the first predicted ctRNA secondary structure (S1 in Fig. 3A) is complementary to the putative Shine-Dalgarno sequence of the rep gene. Such a mechanism of copy number control by ctRNA at the translational level was found in RCR plasmids of the pC194 and pMV158 families (8, 20).

The observed low activity of promoter Prep itself is another feature of negative control of the rep gene expression, which seems to affect the maintaining of an optimum number of plasmid pGA1 copies. It was found that the presence of the sequence IR1 (Fig. 1) in the rep leader region is responsible for this low activity of Prep, which is independent of synthesis of ctRNA complementary to the rep mRNA leader region. Thus, the IR1 sequence is a second regulatory element, in addition to the ctRNA, negatively influencing expression of the pGA1 rep gene. Observation that no trans-acting element encoded by the plasmid pGA1 or by C. glutamicum chromosome influenced the negative effect of the IR1 sequence on the Prep activity excluded functioning of this sequence as a DNA target (operator) for any effector. Therefore, it seems that it is the distinct secondary structure of IR1 formed within the leader region of rep mRNA, which is responsible for the negative effect on the rep gene expression. Interference of this structure with the function of RNA polymerase may be one of the possible mechanisms. A regulatory role of such structure(s) in control of number of plasmid copies has not yet been described in any bacterial plasmid.

As proved previously, maintaining of the optimum number of pGA1 copies in C. glutamicum cells also requires the functioning of the positively acting element, the Per protein (22). When possible mechanisms of its activity were tested, it was found that, regardless its positive effect on pGA1 copy number, it has no positive effect on the activity of promoter Prep, and no binding of this protein to pGA1 DNA was observed (T. Venkova-Canova, unpublished results). As shown in this study, abolishment of ctRNA synthesis also has a strong positive effect on plasmid copy number but no effect on activity of promoter Prep. Taking these two findings together, we may suggest that both copy number controlling elements (i.e., the positively acting Per protein and the negatively acting ctRNA) function at the posttranscriptional level. The positive effect of Per on the pGA1 copy number might be explained by its interaction with ctRNA, resulting in a decrease of negative effect of ctRNA on the rep gene expression. This hypothesis serves as a basis for further studies on mechanisms of Per function.