Correct Timing of dnaA Transcription and Initiation of DNA Replication Requires trans Translation

Abstract

The trans translation pathway for protein tagging and ribosome release has been found in all bacteria and is required for proliferation and differentiation in many systems. Caulobacter crescentus mutants that lack the trans translation pathway have a defect in the cell cycle and do not initiate DNA replication at the correct time. To determine the molecular basis for this phenotype, effects on events known to be important for initiation of DNA replication were investigated. In the absence of trans translation, transcription from the dnaA promoter and an origin-proximal promoter involved in replication initiation is delayed. Characterization of the dnaA promoter revealed two cis-acting elements that have dramatic effects on dnaA gene expression. A 5′ leader sequence in dnaA mRNA represses gene expression by >15-fold but does not affect the timing of dnaA expression. The second cis-acting element, a sequence upstream of the −35 region, affects both the amount of dnaA transcription and the timing of transcription in response to trans translation. Mutations in this promoter element eliminate the transcription delay and partially suppress the DNA replication phenotype in mutants lacking trans translation activity. These results suggest that the trans translation capacity of the cell is sensed through the dnaA promoter to control the timing of DNA replication initiation.

The trans translation pathway for protein tagging and ribosome rescue is found in all bacteria and is required for normal physiology, development, or virulence in many species (14). During trans translation, a ribonucleoprotein complex consisting of tmRNA (transfer-messenger RNA) and SmpB enters substrate ribosomes that have a peptidyl-tRNA in the P site. The termini of tmRNA fold into a structure that mimics alanyl-tRNA, and tmRNA is charged with alanine. tmRNA-SmpB acts first like a tRNA, accepting the nascent polypeptide in a normal transpeptidation reaction. A specialized open reading frame within tmRNA is then decoded by the ribosome, producing a protein that has the tmRNA-encoded peptide at its C terminus and releasing the ribosome after termination at a stop codon. The tmRNA peptide is recognized by several intracellular proteases and targets the protein for rapid degradation. This trans translation reaction is important for both translational quality control and regulation of genetic circuits in many species (14, 25).

trans translation is particularly important for developmental processes in bacteria. tmRNA is required for cellular differentiation during sporulation in Bacillus subtilis (1), symbiosis in Bradyrhizobium japonicum (6), pathogenesis in Salmonella enterica (13) and Yersinia pseudotuberculosis (27), and the cell cycle of Caulobacter crescentus (17). For systems in which the trans translation phenotype has been investigated in molecular detail, developmental defects are due to misregulation of a key signaling molecule. For example, B. subtilis sporulation is disrupted in strains lacking tmRNA due to decreased expression of the SpoIVCA recombinase, an enzyme that is required for production of σ^K during the developmental sigma factor cascade (1). Likewise, pathogenesis by Y. pseudotuberculosis is impaired in the absence of trans translation because of misregulation of the VirF transcriptional regulator (27). trans translation is also clearly required for development in C. crescentus, but the molecular basis for this requirement has not previously been determined.

C. crescentus cells that lack trans translation activity because of a deletion in smpB or ssrA, the gene that encodes tmRNA, grow slower than wild-type cells due to a delay in the initiation of DNA replication (17). Newly divided C. crescentus cells in G₁ phase can be easily isolated, and these cells will pass synchronously through the cell cycle, allowing detailed studies of cell cycle-related processes (29). Unlike those of Escherichia coli, C. crescentus cells initiate DNA replication only once per cell division (22). Work with synchronized cultures has shown that initiation of DNA replication is controlled by several factors at the origin of replication. The essential response regulator CtrA binds to five sites in the origin and represses inappropriate initiation events (28). CtrA is degraded during the G₁-S-phase transition to allow replication to initiate (5). Another essential factor, DnaA, binds to the origin of replication and is assumed to act in the same manner as E. coli DnaA, unwinding the DNA and recruiting DNA polymerase to initiate replication (21). Expression of dnaA is cell cycle regulated, with maximum new protein synthesis immediately before DNA replication initiates (4, 34), and delaying production of new DnaA protein using an inducible promoter delays replication initiation, suggesting that new DnaA synthesis is required for replication initiation (11). In addition to these essential proteins, other factors also play a role in regulation of DNA replication. A strong promoter adjacent to the origin, P_s, is active during initiation and may help unwind DNA. P_s is required for replication in a plasmid model system but is dispensable in the chromosomal context (11, 20). IHF also binds within the origin and may play a role in replication control (31).

In ΔssrA and ΔsmpB cells, CtrA is degraded at the same time as in the wild type, but replication does not initiate for 30 to 45 min (17). In addition, the abundance of tmRNA and SmpB increases just before the initiation of DNA replication, and they are depleted in early S phase, consistent with a role for trans translation in replication initiation (16). In this article, the effects of trans translation on expression of dnaA and transcription from P_s were investigated to determine why the correct timing of replication initiation requires tmRNA and SmpB.

MATERIALS AND METHODS

Strains and plasmids.

Plasmids and strains are listed in Table 1. The P_dnaA-lacZ reporter was constructed by using PCR to amplify the dnaA promoter from −369 to −1 with respect to the transcriptional start site and cloning the product into pRKlac290 (9), resulting in a lacZ gene containing its own translation initiation sequences under the control of the dnaA promoter. Deletions in the leader sequence were made in a similar fashion. Mutations in the dnaA promoter from −57 to −77 and in the putative DnaA box were constructed using the QuikChange mutagenesis kit (Stratagene). Note that although some pBBR1-derived plasmids will not replicate in ΔssrA cells, RK2-derived plasmids are maintained in all strains used in this study.

TABLE 1.

Plasmids and strains used in this study

Plasmid, strain, or genotype	Description	Reference
pRKlac290	RK2 vector for making lacZ transcriptional fusions	9
Ps-lacZ	Ps promoter in pRKlac290 (pGM976)	23
PdnaA-lacZ	dnaA promoter from −369 to −1 in pRKlac290	This study
dnaA′-lacZ	dnaA gene from −369 to +163 in pRKlac290	This study
pLC17	dnaA′-lacZ with DnaA box TTATCCAAG changed to TGAGCGCAG	This study
PxylX-dnaA′-lacZ	Xylose-inducible promoter followed by dnaA gene from +1 to +140 in pRKlac290	This study
PxylX-lacZ	Xylose-inducible promoter in pRKlac290	This study
pLC43	E. coli dnaA gene from −190 to +151 in pRKlac290	This study
pLC44	E. coli dnaA gene from −190 to −1 in pRKlac290	This study
PdnaA-dnaA	PdnaA-lacZ with dnaA coding sequence replacing lacZ	This study
CB15N	Wild-type synchronizable C. crescentus	8
KCK116	CB15N ΔssrA (in-frame deletion)	17
LC202	MG1655 ΔlacX74 ΔssrA::nptII	This study

C. crescentus strains were grown in PYE or M2 medium (7) with 0.2% glucose (M2G) or 0.3% xylose at 30°C, with 5 mg/ml kanamycin, 1 mg/ml chloramphenicol, or 1 mg/ml oxytetracycline as necessary, and growth was monitored by measuring the optical density at 660 nm (OD₆₆₀). Synchronized cultures were obtained by Ludox density gradient centrifugation as described previously (8), and all synchronized cultures were grown in M2G. The timing of DNA replication initiation was determined at each time point by incubating an aliquot with 15 μg/ml rifampin (rifampicin) at 30°C for 3 h, fixing cells by addition of ethanol to 70%, staining with Syto 13 (Invitrogen), and measuring DNA content by flow cytometry as described previously (32). E. coli strains were grown in LB broth at 37°C, and growth was monitored by measuring the OD₆₆₀ (30).

Quantitative RT-PCR.

Cells were harvested from synchronized cultures, and RNA was prepared using the RNeasy mini kit (Qiagen), including two incubations with RNase-free DNase I (Qiagen) on the column according to the manufacturer's instructions. RNA samples were diluted to 200 ng/μl and tested by PCR to ensure that there was no genomic DNA contamination before use in real-time PCR (RT-PCR). Reverse transcription was performed using the High Capacity RT kit (ABI), and cDNA was added to a PCR containing TaqMan 2× universal mix and amplified with the following protocol: 50°C for 2 min, 95°C for 10 min, and 40 repeats of 95°C for 15 s and 60°C for 1 min. The primers for dnaA gene were as follows: forward primer, 5′-GAGTTCGCGCACGCTGTAG-3′ (corresponding to bases 493 to 511 of the dnaA coding sequence); and reverse primer, 5′-CGTACGGGCCGTGGAA-3′ (complementary to bases 558 to 574 of the dnaA coding sequence). The TaqMan probe for dnaA was 5′-CGGACGGTCACTTCAATCCTGTGCT-3′. The 16S rRNA gene was used as a control with the following primers: 5′-GGGTTAAGTCCCGCAACGA-3′ and reverse primer 5′-ATGATTAGAGTGCCCAGCCAAA-3′. The TaqMan probe for 16S rRNA was 5′-CGCAACCCTCGTGATTAGTTGCCATC-3′. Both TaqMan probes were synthesized by Applied Biosystems, with 6-carboxyfluorescein at the 5′ end as the reporter and 6-carboxytetramethylrhodamine at the 3′ end as the quencher. The ABI 7300 sequence detection system was used to record data, and data were analyzed by the comparative cycle threshold (C_T) method for relative quantification according to the manufacturer's instructions (Applied Biosystems). The C_T was defined as the fraction of the cycle when the fluorescent intensity reached the threshold. For each time point, the average dnaA C_T was compared to the average C_T of 16S rRNA to get the ΔC_T value, and the ΔC_T value of this time point was then compared to the ΔC_T value of time zero min to get the ΔΔC_T value. The relative amount of dnaA mRNA was calculated by using the formula 2^−ΔΔ_CT.

Pulse-labeling, pulse-chase, and immunoblotting.

Pulse-labeling experiments were performed as described previously (16). Briefly, C. crescentus strains were grown in M2G medium to an OD₆₆₀ of 0.3 to 0.4 and synchronized. At each time point, 1 ml culture was sampled and incubated with 1 μl [³⁵S]methionine (10 to 15 μCi) at 30°C for 5 min. The reaction was stopped by adding 950 μl labeled culture to 50 μl trichloroacetic acid, and labeled protein was recovered by centrifugation. Protein was resuspended in 50 μl IP-SDS buffer (10 mM Tris-HCl [pH 8], 1% SDS, 1 mM EDTA), and 750 μl RIPA buffer (50 mM Tris-HCl [pH 8.0], 150 mM NaCl, 1% NP-40, 0.5% deoxycholate, 0.1% SDS) was added. Radioactivity was quantified by scintillation counting, and equal counts for each sample were added to anti-β-galactosidase or anti-DnaA antibody in 500 μl RIPA buffer with 15 μl protein A conjugated with Sepharose beads. After incubation at 4°C overnight, samples were washed twice with 900 μl RIPA buffer, resuspended in 2× Laemmli loading buffer, and separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Dried gels were exposed to phosphor screens, scanned using a Typhoon scanner, and analyzed using Imagequant software (Molecular Dynamics).

Pulse-chase experiments were performed as described for pulse-labeling experiments, except that after incubation with [³⁵S]methionine for 5 min, cold methionine was added to a 2% final concentration at time zero. Protein was isolated and analyzed by immunoprecipitation as described above.

For Western blots, cells were harvested by centrifugation and resuspended in a volume of 2× Laemmli loading buffer normalized to maintain a constant OD₆₆₀ per sample. Electrophoresis, blotting, and development with anti-DnaA antibody were performed as described previously (30).

β-Galactosidase activity assays.

Exponentially growing C. crescentus strains harboring a lacZ reporter were sampled, and β-galactosidase activity was measured using a modified Miller assay (15). Briefly, at each time point, 1 ml culture was taken and the OD₆₆₀ was measured. Fifty μl chloroform was added to 50 μl culture and mixed with 750 μl Z buffer (composition), and o-nitrophenyl-β-d-galactopyranoside was added to 1 mg/ml. The reaction was incubated at 30°C until the color changed to yellow and stopped by addition of 500 μl Na₂CO₃, and the OD₄₂₀ was measured. The increase in β-galactosidase activity was calculated by plotting OD₄₂₀/reaction time versus OD₆₆₀. Assays with E. coli were performed using the same procedure except that the OD was measured at 600 nm.

RESULTS

Transcription of dnaA and P_s is delayed in ΔssrA cells.

To determine if trans translation is required for the correct timing of the earliest events in initiation of DNA replication, transcription from the dnaA and P_s promoters was measured throughout the cell cycle using lacZ transcriptional reporters. A lacZ gene containing its own translation initiation sequences was fused to the transcriptional start site of each promoter on a low-copy-number plasmid, and production of β-galactosidase was measured in synchronized cultures by pulse-labeling with [³⁵S]methionine followed by immunoprecipitation of β-galactosidase (Fig. 1). In these assays, the amount of radiolabeled β-galactosidase produced at each time point indicates the relative promoter activity. The initiation of DNA replication was monitored at each time point using flow cytometry assays (see Fig. S1 in the supplemental material). In wild-type cells, transcription from the dnaA promoter peaked at 30 min, 5 min before the initiation of DNA replication (Fig. 1A). This result is consistent with previously published observations (4, 34). In the ΔssrA strain, peak transcription from the dnaA promoter occurred at 60 min, again just before the initiation of DNA replication (Fig. 1A). Transcription from the P_s promoter was also delayed in the ΔssrA strain (Fig. 1B). These data are consistent with a role for trans translation in control of dnaA and P_s transcription or at an earlier step that is required for both processes. Because a delay in dnaA transcription can delay replication initiation (11), the role of trans translation in DnaA production was investigated as a possible mediator of the cell cycle delay phenotype.

FIG. 1.

tmRNA is required for correct timing of transcription from the dnaA promoter and origin-proximal promoter P_s. Cells containing the P_dnaA-lacZ reporter (A) or the P_s-lacZ reporter (B) were synchronized in wild-type (gray) or ΔssrA (black) cells. The amount of promoter activity at each time point was measured by pulse-labeling followed by immunoprecipitation of β-galactosidase. The amount of radiolabeled β-galactosidase was normalized to the highest value for each strain. Vertical lines indicate the time at which half of the cells had initiated DNA replication as determined by flow cytometry assays (see Fig. S1 in the supplemental material). Representative curves from >5 repeats are shown.

dnaA expression is delayed in absence of trans translation.

To confirm that the delay in dnaA transcription observed in ΔssrA cells results in altered cell cycle regulation of DnaA expression, the levels of dnaA mRNA, the timing of DnaA protein synthesis, and the accumulation of DnaA protein were assayed in synchronized cultures. dnaA mRNA levels were measured using quantitative RT-PCR and DNA microarrays (Fig. 2A; also data not shown). Both techniques showed that the level of dnaA mRNA in ΔssrA cells changed during the cell cycle in a pattern similar to that for wild-type cells but delayed by 30 to 45 min. Likewise, pulse-labeling assays showed that the pattern of new synthesis of DnaA protein in ΔssrA cells was delayed compared to that for the wild type (Fig. 2B). Western blotting for total DnaA protein levels showed that the peak accumulation of DnaA in the ΔssrA strain was also delayed (Fig. 2C). These results show that the delay in transcribing dnaA does result in delayed DnaA protein production.

FIG. 2.

Expression of DnaA is delayed in cells lacking ssrA. (A) Quantitative RT-PCR was used to measure the amount of dnaA mRNA in synchronized cultures of wild-type (gray) or ΔssrA (black) cells. Values were normalized to the highest level for each strain, with error bars indicating the standard deviation after assays with three synchronized cultures. Vertical lines in panels A, B, and C indicate the time at which half of the cells had initiated DNA replication as determined by flow cytometry assays. (B) The amount of newly synthesized DnaA protein was determined in synchronized cultures by labeling cells with [³⁵S]methionine at each time point and immunoprecipitating DnaA. Representative gels and quantifications from three repeats are shown. (C) The amount of total DnaA protein was determined in synchronized cultures by Western blotting. Representative gels from five repeats are shown. Arrows indicate the band corresponding to DnaA (the higher band is GroEL). Protein levels were quantified and normalized to the highest level for each strain. (D) Pulse-chase experiments to measure the half-life of DnaA protein in exponentially growing cultures show that the half life was not affected by ssrA. Representative gels from three repeats are shown.

Because the primary effect of trans translation is to target substrate proteins for proteolysis and DnaA protein is degraded during the cell cycle, the effects of tmRNA on the degradation of DnaA were measured in exponentially growing cultures. Pulse-chase assays showed that the half-life of the DnaA protein was 30 to 35 min in both wild-type and ΔssrA cells, indicating that deletion of ssrA does not have a significant effect on the overall stability of the DnaA protein (Fig. 2D). Therefore, trans translation is required for correct timing of dnaA gene expression but does not affect posttranscriptional control of DnaA levels.

dnaA expression is repressed by its 5′ leader sequence.

The dnaA gene has a 155-base leader sequence between the transcription start site and the ATG start codon (34). The amount of β-galactosidase activity produced from the P_dnaA-lacZ reporter, which does not contain the dnaA leader sequence, was significantly higher than those for previously described dnaA′-lacZ reporters in which the lacZ gene was fused after base 9 of the dnaA open reading frame (34). This observation suggested that the dnaA leader sequence is important for control of dnaA expression. To quantify the effects of the leader sequence on gene expression, β-galactosidase activity assays were performed with wild-type and ΔssrA cells (Fig. 3). Reporters without the leader sequence had >15-fold-higher β-galactosidase activity in both strains. To more precisely map which sequences are important for decreasing gene expression, a series of deletions in the leader sequence was constructed and assayed (Fig. 3). Reporters containing ≥87 bases of the leader sequence had expression levels within twofold of that of the reporter containing the full leader, suggesting that the 3′ portion of the leader sequence is not important for regulating gene expression. Conversely, when only the first 67 bases of the leader sequence were present, expression levels were similar to those when the leader sequence was completely removed, suggesting that some or all of the sequence between bases 68 and 86 is required for regulation.

FIG. 3.

The dnaA leader sequence represses gene expression. Schematic diagrams show the wild-type dnaA locus and lacZ reporter constructs used to measure gene expression. β-Galactosidase assays were used to determine the relative expression from each promoter in wild-type and ΔssrA cells. For dnaA promoter reporters, rates were normalized to the reporter with the entire leader sequence assayed in wild-type cells. For xylX promoter reporters, rates were normalized to the reporter with the dnaA 1-140 leader sequence assayed in wild-type cells. The construct with no DnaA box has four point mutations in the putative DnaA binding site.

A putative DnaA binding site in the leader region has been proposed to be involved in autoregulation of dnaA expression (34). To determine if this site affects regulation by the dnaA leader sequence, four point mutations were engineered to change the sequence from TTATCCAAG to TGAGCGCAG in the P_dnaA-lacZ reporter. β-Galactosidase production was increased by <2-fold compared to that with the unmutated P_dnaA-lacZ reporter in both wild-type and ΔssrA cells (Fig. 3), indicating that this site plays at most a minor role in regulation by the leader sequence.

To test whether the leader sequence acts only in the context of the dnaA promoter or if it is a general repressor of gene expression, the effects of the leader on the xylose-inducible promoter P_xylX (24) were tested. When bases 1 to 140 of the dnaA leader sequence were inserted before lacZ in a P_xylX-lacZ reporter, β-galactosidase expression was decreased >50-fold in both wild-type and ΔssrA cells (Fig. 3). These results indicate that the dnaA leader sequence is capable of decreasing gene expression independently of the dnaA promoter.

The similarities between cell cycle regulation patterns of transcription from P_dnaA-lacZ reporters that do not include the leader sequence (Fig. 1), results in previous studies using dnaA′-lacZ reporters that do include the leader sequence (4, 34), and expression of dnaA mRNA that includes the leader sequence (Fig. 2A) suggest that the leader sequence does not affect the timing of dnaA transcription. To confirm this hypothesis, production of β-galactosidase from the dnaA′-lacZ reporter was monitored in synchronized cultures of wild-type and ΔssrA cells. In both strains, the patterns of β-galactosidase production were the same using dnaA′-lacZ and P_dnaA-lacZ (not shown), indicating that the dnaA leader is important for the amount of dnaA expression but not for timing of dnaA transcription with respect to the cell cycle.

Because the E. coli dnaA gene also has a long untranslated leader sequence (153 nucleotides) (26), transcriptional fusions of lacZ to the start codon and the transcriptional start site of the E. coli gene were engineered and tested with E. coli MG1655 and an isogenic strain deleted for ssrA. Removal of the leader sequence increased gene expression by a factor of 5.3 ± 1.3 in MG1655 and by a factor of 6.6 ± 0.6 in ΔssrA E. coli. The dnaA leaders from C. crescentus and E. coli do not have any sequence homology or common RNA structure, but both leaders repress gene expression.

Timing of dnaA transcription is controlled through a promoter element in a trans-translation-dependent manner.

It was previously reported that a mutation of base C(−71) to T in the dnaA′-lacZ reporter decreased transcription, suggesting the presence of a regulatory element at this site (3). To determine if this promoter element is involved in tmRNA-dependent regulation of dnaA expression, the effects of mutations at −71 and surrounding bases were examined in wild-type and ΔssrA cells (Fig. 4). Consistent with previous results, in wild-type cells the C(−71)T mutation decreased lacZ expression by >80%. Likewise, mutations at positions −73 through −69, −67, and −64 through −60 decreased expression by >50%. Similar effects were observed in the ΔssrA strain. These data suggest that the GTCAANANNAATAT sequence is required for full promoter activity. One possible explanation for these results is that a transcriptional activator binds to this site to promote transcription from P_dnaA.

FIG. 4.

A dnaA promoter element is responsible for delayed transcription in the absence of ssrA. (A) Point mutations were introduced from −57 to −77 in the dnaA promoter fused to lacZ, and β-galactosidase assays were used to measure the effects on promoter activity in wild-type (gray bars) or ΔssrA (black bars) cells. The wild-type promoter sequence is shown on top, with the point mutation indicated below. The GANTC site proposed to be methylated by CcrM is boxed in gray. Data are plotted as log₂ of the activity relative to that of the wild-type promoter in wild-type cells, so negative numbers indicate a decrease in expression. The promoter motif that includes all changes of >2-fold is indicated. (B) Transcription from the C(−71)T mutant promoter in wild-type and ΔssrA cells was determined as described in the legend for Fig. 1A.

To determine if the cell cycle regulation of dnaA transcription is affected by the GTCAANANNAATAT sequence element, expression of lacZ from the P_dnaAC(−71)T promoter was assayed in synchronous cultures. In wild-type cells, peak transcription from the P_dnaAC(−71)T promoter occurred at 15 min, just before the initiation of DNA replication at 30 min (Fig. 4B). However, in the ΔssrA strain, transcription from the P_dnaAC(−71)T promoter peaked at 15 min, even though DNA replication did not initiate until 57 min. Moreover, transcription from P_dnaAC(−71)T in ΔssrA cells was indistinguishable from that from P_dnaA or P_dnaAC(−71)T in wild-type cells. Because mutation of C(−71) restores the wild-type transcription pattern in ΔssrA cells, it is likely that the GTCAANANNAATAT element is responsible for the delay in P_dnaA transcription in the absence of trans translation. It is not yet clear whether trans translation acts directly on a transcription factor that binds to this site or whether the mechanism is less direct.

Position −71 overlaps with a GANTC methylation site for CcrM, a cell-cycle-regulated DNA methyltransferase. However, this methylation site does not appear to be important for tmRNA-dependent regulation of dnaA transcription. A(−74) would be critical for methylation at this site, but mutation of A(−74) to T did not significantly affect transcription activity in wild-type or ΔssrA cells (Fig. 4A). The A(−74)T mutation also had no effect on the timing of transcription from the P_dnaA-lacZ reporter in synchronized cultures (not shown). Therefore, position −71 and methylation of the GAGTC site by CcrM are not important for regulation of dnaA transcription by trans translation.

Expression of dnaA from a mutant promoter partially suppresses ssrA phenotype.

Based on the data presented above, it is possible that the DNA replication delay in ΔssrA cells is caused by misregulation at the GTCAANANNAATAT promoter element, which disrupts the timing of dnaA transcription. This hypothesis predicts that if dnaA transcription was uncoupled from the GTCAANANNAATAT promoter element, replication would initiate at the correct time in ΔssrA cells. To test this prediction, the dnaA coding sequence was cloned under the control of the P_dnaAC(−71)T promoter. Because the C(−71)T mutation decreases transcription, the dnaA leader sequence was not included in the P_dnaAC(−71)T-dnaA construct to increase dnaA expression levels. The combined effects of the C(−71)T mutation and no leader sequence were tested using lacZ reporters, and the P_dnaAC(−71)T-lacZ reporter had ∼4-fold-higher activity than the dnaA′-lacZ reporter in exponentially growing cultures (not shown). A low-copy-number plasmid with P_dnaAC(−71)T-dnaA was introduced into wild-type and ΔssrA cells, and the growth rate and timing of DNA replication initiation were measured (Table 2). ΔssrA cells with P_dnaAC(−71)T-dnaA grew in exponential phase with a doubling time of 135 ± 9 min, significantly faster than ΔssrA cells with P_dnaAC(−71)T-lacZ (Student's t test, P = 0.05) or with no plasmid (Student's t test, P = 0.02). P_dnaAC(−71)T-dnaA did not increase the growth rate of wild-type cells, indicating that the increased growth rate in the ΔssrA strain was due to specific suppression of the ssrA phenotype and not an unrelated increase in the growth rate caused by the plasmid.

TABLE 2.

Suppression of growth phenotypes by C(−71)T mutant

Description of straina	Doubling time (min)	Time of replication initiation (min)
wt + vector	136 ± 11	26 ± 1
ΔssrA + vector	158 ± 8	55 ± 3
wt + PdnaAC(−71)T-lacZ	136 ± 6	35 ± 4
ΔssrA + PdnaAC(−71)T-lacZ	151 ± 9	62 ± 2
wt + PdnaAC(−71)T-dnaA	133 ± 8	32 ± 3
ΔssrA + PdnaAC(−71)T-dnaA	135 ± 9	44 ± 2

^awt, wild type.

Measurements of DNA replication initiation in synchronized cultures showed that P_dnaAC(−71)T-dnaA did not affect wild-type cells but caused ΔssrA cells to initiate replication earlier (Table 2). Wild-type cells containing P_dnaAC(−71)T-dnaA or P_dnaAC(−71)T-lacZ initiated replication at the same time, indicating that expression of dnaA from this promoter does not alter the timing of initiation. ΔssrA cells with P_dnaAC(−71)T-dnaA initiated replication 20 min earlier than isogenic cells with P_dnaAC(−71)T-lacZ, suggesting that earlier transcription of dnaA partially suppresses the replication delay, even in the presence of a wild-type copy of the dnaA locus. However, the possibility that suppression was due to the amount of dnaA expression from P_dnaAC(−71)T, instead of the timing of expression, could not be excluded. Providing a plasmid-borne copy of dnaA expressed from the wild-type gene (including the leader sequence) had no effect on the growth rate of ΔssrA cells, and expressing dnaA from P_dnaA with no leader sequence caused slower growth (not shown), but these promoters have fourfold-lower activity than P_dnaAC(−71)T and sixfold-higher activity than P_dnaAC(−71)T, respectively. Nevertheless, the observations that the C(−71)T mutation relieves the P_dnaA transcription delay in ΔssrA cells and expression of dnaA from the mutant promoter partially suppresses the replication delay are consistent with a model in which much of the DNA replication delay observed in the absence of trans translation is due to misregulation of dnaA expression through the GTCAANANNAATAT promoter element.

DISCUSSION

The results presented here reveal two mechanisms for regulating expression of dnaA: a promoter element that affects transcription and an untranslated leader sequence that represses expression. Previously described mechanisms for regulating DnaA include putative transcriptional repression (15), regulated proteolysis (12), and regulatory inactivation by hydrolysis of bound ATP (21). Why are so many control mechanisms required, particularly when constitutive expression of dnaA does not have severe consequences during growth in culture (11)? A likely explanation is that these mechanisms provide the opportunity for input from multiple pathways that impact the ability of the cell to successfully complete DNA replication under different growth conditions. Possible regulatory roles for the dnaA leader sequence and promoter motif are discussed below.

The first 87 nucleotides of the dnaA leader sequence decreased expression from both the dnaA promoter and the unrelated xylose-inducible promoter >15-fold, suggesting that the leader sequence acts after transcription initiation. This leader sequence might act through a transcription attenuation mechanism or by destabilizing the mRNA. Most attenuators contain transcriptional termination sequences which are controlled by changes in RNA structure (10), but these features are not evident in the dnaA leader sequence. There is no predicted hairpin structure or a run of uridines characteristic of an intrinsic transcriptional terminator. Secondary structure predictions using the software program MFold (33) did not reveal alternative structures that would be predicted to block transcription elongation or translation initiation. Nevertheless, it is possible that this sequence uses a factor-dependent transcriptional terminator that has not been characterized. Alternatively, the leader sequence could target the mRNA for degradation, for example through binding of a small RNA. Searches of the C. crescentus genome showed no other sequences similar to the dnaA leader, so this repression mechanism may be dedicated to dnaA regulation. Although the leader sequence had profound effects on the magnitude of dnaA expression, it did not significantly alter the cell cycle timing of transcription. One possible use for this leader sequence would be to allow for a large burst in dnaA expression when DnaA protein concentrations are depleted. DnaA levels decrease dramatically under some starvation conditions (12), so a high rate of dnaA expression might facilitate restarting the cell cycle when nutrients become available. A temporary inactivation of the regulator that acts at the leader sequence would provide a large increase in dnaA expression without requiring a large increase in dnaA transcription.

How does the GTCAANANNAATAT element affect transcription? Mutations in the element decreased transcription, suggesting that this element is the binding site for a transcriptional activator. However, the C(−71)T mutation eliminates the delay in promoter activity in ΔssrA cells, suggesting removal of a repressor binding site. One possibility is that two different transcriptional regulators bind to this sequence. Alternatively, a dual regulator may bind this sequence, repressing dnaA transcription in swarmer cells but switching transcription on just before the G₁-S-phase transition. There are several examples of dual regulators, including ArgP (also called IciA), a LysR-type transcription factor that binds a single promoter site to repress or activate transcription (18). Transcription is repressed when ArgP is bound to lysine but activated when it is bound to arginine. ArgP regulates dnaA transcription in E. coli (19), and C. crescentus contains several possible homologues. Identification and characterization of the proteins that bind to the GTCAANANNAATAT promoter element will clarify how this sequence regulates dnaA transcription and how the mechanism is affected by trans translation.

trans translation was previously shown to be required for correct timing of DNA replication initiation (17), and the data presented here show that trans translation is required for correct timing of transcription from the P_dnaA and P_s promoters, two of the earliest steps in DNA replication initiation. In ΔssrA cells, mutation of the GTCAANANNAATAT element affects the timing of transcription from the dnaA promoter, and the growth rate and delay of replication initiation phenotypes are partially suppressed when there is a copy of dnaA expressed from the P_dnaAC(−71)T mutant promoter. Why would dnaA expression and replication initiation be sensitive to trans translation activity? One possibility is that trans translation regulates a transcription factor that binds to the GTCAANANNAATAT sequence, and in the absence of tmRNA the factor is misregulated and slows induction of dnaA transcription. An example of misregulation in the absence of trans translation has been observed for LacI in E. coli (2). LacI is tagged by trans translation as part of an autoregulatory circuit, and excess LacI accumulates in cells deleted for ssrA. This excess LacI delays transcription of the lac operon under inducing conditions (2). A similar defect in control of a regulator that binds to the GTCAANANNAATAT sequence could delay dnaA transcription, thereby delaying initiation of DNA replication. Alternatively, there may be a regulatory checkpoint that intentionally senses trans translation levels to ensure that there is sufficient translation capacity or that there is not a large amount of aberrant translation occurring before the cell commits to S phase. In this model, trans translation could directly affect the activity of a transcription factor or trans translation activity could be sensed indirectly, for example through the presence of stalled ribosomes. These models can be tested once factors that act at the GTCAANANNAATAT element have been identified.