• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of pnasPNASInfo for AuthorsSubscriptionsAboutThis Article
Proc Natl Acad Sci U S A. Oct 23, 2007; 104(43): 17111–17116.
Published online Oct 17, 2007. doi:  10.1073/pnas.0708112104
PMCID: PMC2040471

A DNA methylation ratchet governs progression through a bacterial cell cycle


The Caulobacter cell cycle is driven by a cascade of transient regulators, starting with the expression of DnaA in G1 and ending with the expression of the essential CcrM DNA methyltransferase at the completion of DNA replication. The timing of DnaA accumulation was found to be regulated by the methylation state of the dnaA promoter, which in turn depends on the chromosomal position of dnaA near the origin of replication and restriction of CcrM synthesis to the end of the cell cycle. The dnaA gene is preferentially transcribed from a fully methylated promoter. DnaA initiates DNA replication and activates the transcription of the next cell-cycle regulator, GcrA. With the passage of the replication fork, the dnaA promoter becomes hemimethylated, and DnaA accumulation drops. GcrA then activates the transcription of the next cell-cycle regulator, CtrA, once the replication fork passes through the ctrA P1 promoter, generating two hemimethylated copies of ctrA. The ctrA gene is preferentially transcribed from a hemimethylated promoter. CtrA then activates the transcription of ccrM, to bring the newly replicated chromosome to the fully methylated state, promoting dnaA transcription and the start of a new cell cycle. We show that the cell-cycle timing of CcrM is critical for Caulobacter fitness. The sequential changes in the chromosomal methylation state serve to couple the progression of DNA replication to cell-cycle events regulated by the master transcriptional regulatory cascade, thus providing a ratchet mechanism for robust cell-cycle control.

Keywords: Caulobacter, DnaA, CcrM

Chromosome methylation is found in both prokaryotes and eukaryotes, where it serves to control diverse functions, including regulation of gene expression, nucleoid segregation, and host–pathogen interactions (1). Regulatory DNA methyltranferases in bacteria include the Dam protein in Escherichia coli and other γ-proteobacteria and the CcrM protein in α-proteobacteria, which methylate adenosines on the sequence GATC and GANTC, respectively. CcrM was first discovered in Caulobacter crescentus, and homologues were later found in Agrobacterium tumefaciens, Sinorhizobium meliloti, and Brucella abortus (26). Unlike Dam, CcrM is present and active for a short time at the conclusion of chromosome replication, and it is essential for cell viability (2, 3). Chromosomal loci are fully methylated at the start of DNA replication and the loci then become hemimethylated by the passage of the replication fork and are not fully methylated again until just before cell division (2, 7). Here, we show that the change in chromosomal methylation state, tied to the progression of DNA replication, acts as a timer for the expression of the DnaA master cell-cycle regulator.

Caulobacter divides asymmetrically, yielding two different progeny: a stalked cell and a swarmer cell (Fig. 1A). Before it differentiates into a stalked cell, the swarmer cell is unable to initiate chromosome replication. In the stalked cell, the initially fully methylated chromosome is replicated bidirectionally, once and only once per cell cycle, from a single origin of replication (Cori) (7, 8). Three essential transcriptional regulators, DnaA, GcrA, and CtrA, organized as a transcriptional cascade, control the expression of >200 cell cycle-regulated genes that control multiple cell-cycle events (913) (Fig. 1B). DnaA, a dual-function protein, is nearly ubiquitous in eubacteria (14, 15). It is a pivotal element in the Caulobacter cell-cycle control circuit because it is required for the initiation of DNA replication, it controls the transcription of >40 genes involved in nucleotide biogenesis, cell division, and polar morphogenesis (7, 9, 11), and it activates the transcription of gcrA (Fig. 1B) (16). GcrA controls the transcription of ctrA and genes involved in DNA metabolism and chromosome segregation, including those encoding DNA gyrase, DNA helicase, DNA primase, and DNA polymerase III (13). CtrA, another dual-function master cell-cycle regulator, accumulates out-of-phase with GcrA (10, 12, 13, 1719) (Fig. 1A). CtrA represses DNA replication initiation by binding to and silencing Cori, and it is also a transcriptional regulator of ≈95 cell cycle-regulated genes including the genes encoding CcrM and the FtsZ cell division protein. A striking feature of the operation of the DnaA/GcrA/CtrA regulatory network is that each of these three unstable regulators accumulates at distinct times in the cell cycle to activate or repress functions occurring at that time (Fig. 1A) (16, 20).

Fig. 1.
The Caulobacter cell-cycle control circuit. (A) Schematic of the cell cycle. The θ-like structure in the cell represents the replicating chromosome. Colors indicate which master regulator(s) is present in each cell type or cell compartment. ( ...

Here, we show that the cell cycle-dependent transcription of dnaA is controlled by the methylation state of its promoter and that differential DNA methylation timing enhances Caulobacter fitness. The dnaA gene lies near Cori, and the dnaA promoter is most active when in the fully methylated state. When the replication fork passes the dnaA gene at the beginning of chromosome replication, the dnaA promoter becomes hemimethylated and inactivated. The DNA methylation-based regulation of dnaA transcription modulates the temporal levels of DnaA during the cell cycle and is a mechanism for synchronizing the multiple functions of DnaA with the initiation of DNA replication. Activation of CcrM synthesis by CtrA at the end of chromosome replication results in the rapid remethylation of the chromosome, thus reactivating dnaA expression in preparation for a new cell cycle (Fig. 1B).


The Cell-Cycle Regulation of DNA Methylation Confers a Gain of Fitness in Caulobacter.

The methylation state of the GANTC sequences in Caulobacter and A. tumefaciens chromosomes varies as a function of the cell cycle, because of the cell-cycle control of ccrM transcription (2, 6). When the ccrM gene is placed under the control of a constitutive promoter, the chromosome is kept fully methylated throughout the cell cycle, resulting in defects in cell division and cell morphology in Caulobacter, A. tumefaciens, B. abortus, and S. meliloti (26). The conservation of the cell-cycle regulation of chromosomal methylation state suggests that loss of this regulation would be counterselected. We tested this hypothesis by determining whether wild-type Caulobacter has a fitness advantage over a strain in which the chromosome is maintained in the fully methylated state throughout the cell cycle. For this competitive growth experiment, we used the Caulobacter LS1 strain that expresses ccrM constitutively (from an integrated copy of ccrM under the control of the lacZ promoter), and we have shown previously that its chromosome is fully methylated throughout the cell cycle (2). We compared the fitness of the LS1 strain with the parental strain LS176, bearing a single chromosomal cell cycle-regulated copy of ccrM. We performed three competition experiments to assay fitness of the LS1 strain relative to the LS176 strain. An equal number of LS176 and LS1 cells were combined in peptone yeast extract (PYE)-rich media, and the mixed population was grown for ≈20 generations. The LS1 strain carries a selective marker (resistance to kanamycin) that we found to be neutral in the conditions tested. The relative fitness of the two strains, defined as the ratio of their growth rates during competition for the same pool of nutrients, reflected differences in reproduction and survival. The mean fitness of the LS1 strain after 20 generations was 0.887, with its final survival frequency ≈5% of the total population. Thus, the cell-cycle regulation of DNA methylation confers a significant gain of fitness to Caulobacter.

The CcrM DNA Methyltransferase Activates dnaA Transcription.

Because Caulobacter strains that are engineered to keep their chromosome fully methylated throughout the cell cycle have defects in cell division and exhibit a low fitness, we examined the cellular levels of the three cell-cycle master regulators in these strains. We observed that DnaA levels were >2-fold higher in the LS1 strain, when ccrM was under the control of the constitutive lacZ promoter than the parental LS176 strain with a normal cell cycle-regulated ccrM expression (Fig. 2A). The increase in the level of DnaA was due to increased synthesis of DnaA [see supporting information (SI) Fig. 5] and not to decreased degradation of DnaA (see SI Fig. 6). We assayed the levels of DnaA in two strains where the dnaA promoter was hemimethylated for a longer fraction of the cell cycle because of decreased availability of CcrM. There was a significant decrease in DnaA accumulation in these strains, compared with a wild-type strain (see SI Fig. 7), confirming that the presence of CcrM stimulates DnaA accumulation.

Fig. 2.
The methylation state of the dnaA promoter controls the levels of DnaA during the Caulobacter cell cycle. (A) Strain LS1, with an extra chromosomal copy of ccrM under the control of the constitutive lacZ promoter and the parental LS176 strain, with a ...

The dnaA gene is located ≈2 kb from the origin of replication (Fig. 2 B and C). It is transcribed from a single promoter (21) containing two GANTC adenine methylation sites designated Met-1 and Met-2 (Fig. 2C). It was previously shown that the Met-2 site in the dnaA promoter is methylated by CcrM and that its methylation state varies as a function of the cell cycle (2). The dnaA promoter is fully methylated in swarmer cells, it then becomes hemimethylated in stalked cells, and is not fully methylated again until the end of the cell cycle. GANTC sequences occur significantly less frequently in the Caulobacter genome than a random five-base sequence (22), and they are found predominantly in the promoter regions of cell cycle-regulated genes (A. Hottes and H.H.M., unpublished data). The observations that transcription from the dnaA promoter decreases soon after the initiation of DNA replication (21) and that DnaA accumulation was high when the chromosome is maintained in the fully methylated state (Fig. 2A) together suggest that dnaA transcription is deactivated when the dnaA promoter becomes hemimethylated upon passage of the replication fork. To determine whether the transcription of dnaA from its native dnaA promoter was increased when the chromosome was maintained in a fully methylated state throughout the cell cycle, a DNA fragment from −228 to +165 relative to the dnaA +1 transcriptional start site (21) was used to construct a transcriptional fusion to lacZ. This construct was integrated at a site next to the natural dnaA locus, also close to the Cori (at the hcrA locus, named siteO, in strain LS4380) (Fig. 2B). Previous studies have shown that methylation sites in this region of the chromosome become hemimethylated soon after the initiation of DNA replication and remain hemimethylated until the end of DNA replication in wild-type cells (7). We measured β-galactosidase activity in the LS4380 strain under conditions of constitutive transcription of ccrM under the control of a xylose-inducible promoter. The activity of the dnaA promoter in the strain that accumulated CcrM throughout the cell cycle was ≈50% higher than that observed in the wild-type strain (see SI Fig. 8A), showing that dnaA transcription is directly, or indirectly, activated by CcrM.

Differential Methylation State of the dnaA Promoter Affects the Timing of DnaA Accumulation During the Cell Cycle.

It was previously shown that GANTC sites engineered into a transposon-based methylation probe integrated at different positions near the terminus of replication on the chromosome remain fully methylated throughout most of the cell cycle (7). Thus, moving the dnaA gene to a chromosomal position near the terminus of replication would maintain the dnaA promoter in the fully methylated state throughout most of the cell cycle, without affecting the quantity of CcrM or the methylation state of other loci. We constructed strain LS4384, in which the dnaA gene was deleted at its wild-type position and placed close to the chromosome terminus of replication (named siteT) (Fig. 2B). In the LS4384 strain, DnaA levels were ≈2-fold higher than in the control strain LS3418 with dnaA at its native chromosomal position next to the Cori (and with an empty vector integrated at site T) (see SI Fig. 8B). This observed increase in DnaA levels in a mixed population is shown below to be due to the constitutive accumulation of DnaA during the cell cycle. Furthermore, this experiment eliminated the possibility that dnaA transcription was solely activated by another CcrM-regulated transcriptional regulator, independent of the methylation state of the dnaA promoter. Thus, dnaA transcription is controlled by the methylation state of its promoter, which depends on both its chromosomal position relative to Cori and the cell cycle timing of the CcrM DNA methyltransferase.

Because the CcrM DNA methyltransferase is normally present only near the end of chromosome replication, the dnaA promoter that becomes hemimethylated upon passage of the replication fork remains hemimethylated until near the end of the cell cycle (3). To determine whether the reduced transcription of dnaA just after the initiation of replication (21) is attributable to the switch from full- to hemimethylation of the dnaA promoter, we compared DnaA levels in the strain with constitutive ccrM expression with the wild-type strain over the course of a cell cycle (Fig. 2D). Swarmer cells were isolated (at time = 0 min), and samples were collected at different times during their synchronous progression through an ≈140-min cell cycle. As expected, CcrM was present throughout the cell cycle in LS1 cells, but not in wild-type cells (see SI Fig. 9). However, temporal changes in DnaA levels during the LS1 cell cycle were strongly attenuated compared with the wild-type strain (Fig. 2D). Similarly, we examined the temporal accumulation of DnaA in the LS4384 strain (with a single dnaA gene next to the terminus) and found that the temporal changes in DnaA levels were also strongly attenuated compared with wild-type (Fig. 2D). Thus, the temporal regulation of dnaA transcription is controlled by switching the methylation state of the dnaA promoter.

DnaA is essential for the initiation of bacterial DNA replication. When DnaA binds to DnaA boxes in the Cori, there is local unwinding of an adjacent AT-rich region. The unwound region is thought to provide the entry site for the DnaB/DnaC helicase complex, followed by other replisome proteins (reviewed in ref. 14). Despite the presence of DnaA at high levels throughout the cell cycle in the LS4384 strain, we found that the LS4384 strain did not overinitiate DNA replication (data not shown). There are at least three additional mechanisms that could contribute to the control of replication initiation in Caulobacter: (i) DNA replication initiation is prevented by CtrA in swarmer and predivisional cells (17). (ii) In E. coli, the activity of DnaA is cell cycle-regulated by the nucleotide (ATP/ADP) bound to it (2326). Because the Caulobacter DnaA protein also contains an ATP-binding site and an ATPase domain, we predict that its activity is also cell cycle-regulated by the nucleotide bound to it. (iii) Replication initiation may also be controlled by the methylation state of the Cori, because it contains four methylation sites (27). This would explain why the LS1 strain tends to undergo additional replication initiation events (2), whereas the LS4384 strain does not.

Methylation Sites in the dnaA Promoter Are Required for Efficient dnaA Transcription and for Normal Cell Physiology.

To further confirm that transcription from the dnaA promoter depends on its methylation state, we mutated each of the two GANTC sites in the dnaA promoter region (Fig. 2C) and constructed transcriptional fusions to lacZ in the low-copy-number plasmid pLacZ290 (Fig. 3A). We also constructed a lacZ fusion to a dnaA promoter lacking functional methylation sites. These plasmids and the control vector were introduced into wild-type cells (NA1000), and promoter activities were measured by assays of β-galactosidase activity. We observed that the β-galactosidase activity of the dnaA promoter lacking the Met-2 site [dnaAP(Met-2)] or the unmethylatable dnaA promoter [dnaAP(UM)] was nearly as low as the activity of the strain carrying the empty vector (Fig. 3A). These results show that the methylation sites in the dnaA promoter contribute to dnaA transcriptional control. To examine the physiological effects of altering the methylation state of the dnaA promoter, we constructed strains in which the wild-type dnaA gene was deleted (LS4382) or was replaced by a mutant dnaA gene lacking both promoter methylation sites (LS4383) in all cases covered by the wild-type dnaA coding sequence under the control of the inducible Pxyl promoter (Fig. 3B). We then compared the amount of DnaA that accumulated in the two mutant strains with wild-type. Four hours after shifting cells from xylose- to glucose-containing media, DnaA was no longer detected from LS4382 cells, and only very small amounts of DnaA were detected in LS4383 cells (≈20-fold less) compared with the wild-type LS4381 control (Fig. 3B). Thus, the response of the dnaA promoter to the DNA methylation state is attributable to the GANTC sequences in its promoter. When the strain with the dnaA deletion, LS4382, was shifted to glucose-containing media, cells did not complete cell division (Fig. 3B). Instead, their cell length increased progressively until they finally died, in agreement with the previously described phenotype of DnaA-depleted cells (9). The LS4383 cells shifted to glucose-containing media were longer than the wild-type control LS4381 cells, indicating a significant reduction in the efficiency of cell division, and had longer stalks than the LS4381 cells (Fig. 3B). These results show that the strain with a dnaA promoter lacking methylation sites that transcribes very low levels of dnaA, and thus synthesizes low levels of DnaA, exhibits aberrant cell-cycle progression (Fig. 2 A and B), although the mutation is not lethal.

Fig. 3.
The methylation sites in the dnaA promoter are required for high DnaA accumulation and for normal cell morphology. (A) Mutations introduced in Met-1 and/or Met-2 sites are in upper-case letters, and designation of the pLacZ290 derivatives carrying the ...

DnaA is a critical component of the genetic circuit that controls the Caulobacter cell cycle, and its expression is a pivotal element in the activation of the GcrA/CtrA/CcrM cascade (Fig. 1B) (12, 13, 16, 19). DnaA also activates the transcription of the gene encoding the MipZ spatial regulator of the FtsZ ring assembly, among other cell-cycle genes (11). In strain LS4383, in which the two methylation sites in the dnaA promoter were eliminated at the dnaA locus yielding very low levels of DnaA, the levels of four essential cell cycle regulators (GcrA/CtrA/CcrM/MipZ) in the regulatory cascade were significantly decreased (Fig. 3B), consistent with previous reports showing that DnaA directly or indirectly regulates those genes (11, 16) (Fig. 1B).


Model for the Temporal Regulation of dnaA Transcription by DNA Methylation.

We propose a model for the cell cycle regulation of dnaA transcription shown in Fig. 3C. The dnaA promoter becomes fully methylated by CcrM before cell division (2, 7). Although the dnaA promoter is activated, allowing the accumulation of DnaA, the initiation of DNA replication occurs only in the stalked cell progeny, because the Cori is silenced in the swarmer cell progeny by CtrA. Once CtrA is cleared at the swarmer-to-stalked-cell transition, the accumulated DnaA enables replication initiation. After replication initiation, the replication fork passes through the dnaA promoter, producing two hemimethylated copies of the dnaA gene. The hemimethylated dnaA promoter has low transcriptional activity, and DnaA is an unstable protein (28), so the DnaA level falls when DnaA activity needs to be eliminated to avoid a reinitiation event. Only near completion of chromosome replication, upon synthesis and accumulation of the CcrM DNA methyltransferase, does the dnaA promoter again become fully methylated. This enables the resynthesis of DnaA to initiate a new round of chromosome replication in the stalked progeny. This methylation-based mechanism of regulation therefore serves to adjust DnaA levels in response to the progression of chromosome duplication.

DNA Methylation State Is a Synchronization Mechanism for the Caulobacter Cell-Cycle Control System.

The gene encoding the inhibitor of DNA replication initiation, ctrA, is located on the chromosome further from Cori than dnaA (Fig. 4A), and the transcription from one of ctrA's two promoters is controlled by its methylation state (29) (Fig. 1B). In contrast to the regulation of the dnaA promoter, the ctrAP1 promoter is active when in the hemimethylated state but inactive when in the fully methylated state (29) (Fig. 4A). Therefore, when the replication fork passes through the ctrA gene, the production of hemimethylated copies activates ctrAP1 transcription. If the ctrA gene is moved to a chromosomal position near the terminus, CtrA accumulation is delayed during the cell cycle (29).

Fig. 4.
DNA methylation links the progression of a bacterial cell cycle to chromosome replication. (A) Diagram of the Caulobacter chromosome showing the locations and the methylation state of the origin of replication (Cori), the dnaA gene, and the ctrA gene, ...

Because DnaA and CtrA are connected by GcrA and CcrM in the cyclic control circuit (Figs. 1B and and44B), the cellular levels of DnaA correlate with the levels of CtrA. Indeed, we observed that a deficiency in CtrA leads to reduced dnaA gene expression (see SI Fig. 7B) and that a deficiency in DnaA leads to reduced levels of CtrA (Fig. 3B). This regulatory system probably serves to control the initiation of DNA replication and to reduce overinitiation events. Nevertheless, if overinitiation events should occur, the dnaA promoter and the CcrM-regulated ctrA promoter will become unmethylated by the passage of a second replication fork during the same cell cycle. As a consequence, dnaA transcription will be rapidly shut down (Fig. 3), whereas ctrA transcription will remain efficient (29) to prevent more overinitiation events. In other words, if some malfunction prevents the successful completion of the chromosome replication process, signals are sent to the master regulatory control system to delay progression to the next phase, underscoring the robustness of this cyclical circuit.

The wave of hemimethylation associated with chromosome replication progression can be viewed as a molecular clock for cell-cycle progression. This methylation-based regulation of dnaA synchronizes the initiation of DNA replication with the activation of the GcrA/CtrA/CcrM cascade that controls multiple cell-cycle events (Fig. 4B). The methylation-based regulation of ctrA then synchronizes timing of CtrA synthesis with progress of chromosome replication. In effect, the ratchet-like DnaA/GcrA/CtrA/CcrM circuit provides the forward impetus for cell cycle progression and organizes the order of activation of various genetic subsystems, whereas the tight coupling to progression of chromosome replication paces and synchronizes the system.

Materials and Methods

Bacterial Strains, Synchronization, and Growth Conditions.

Caulobacter crescentus strains were grown in PYE complex media or M2 minimal salts plus 0.2% glucose (M2G) minimal media (30) at 28°C. Plasmids and strains used are listed. (see SI Tables 1 and 2, respectively). Antibiotics used include oxytetracyclin (1 μg/ml), kanamycin (5 or 25 μg/ml), spectinomycin (25 μg/ml), streptomycin (5 μg/ml), and nalidixic acid (20 μg/ml). Plasmids were mobilized from E. coli S17-1 (31) into Caulobacter by bacterial conjugation or introduced by transformation. Bacteriophage [var phi]CR30 was used for general transduction into Caulobacter. Synchronized cell cultures were obtained by centrifugation in a Ludox density gradient, followed by isolation of swarmer cells (32). Swarmer cells were resuspended into minimal medium and allowed to proceed synchronously through the cell cycle.

Plasmids and Strains Constructions.

For plasmids and strains constructions, see SI Text.

Fitness Assays.

Estimation of the relative fitness of LS176 and LS1 strains was performed by using an adaptation of a protocol described earlier (33). The LS1 strain carries a selective marker (kanamycin resistance). The competitive strains were individually grown in PYE-rich media to saturation at 28°C. Equal volumes of each strain were diluted 1,000-fold into the same flask containing 5 ml of PYE, and a sample was immediately plated on PYE agar and PYE agar plus kanamycin, to estimate their initial densities. Every day, for 3 days, the mixed population was diluted 1,000-fold after reaching saturation. After 3 days, a sample was plated on PYE agar and PYE agar plus kanamycin to obtain the initial density of each competitor. For each competitor, we calculated its growth rate as m = ln(1,0003·Nt/N0)/3, where N0 and Nt are initial and final densities, respectively. The relative fitness of two genotypes is defined as the ratio of their growth rates during competition for the same pool of nutrients. We also performed a control competition assay with another strain (LS4318) carrying the same selective marker (kanamycin resistance) as LS1 and its isogenic wild-type strain (LS101). We found that the LS3418 strain was as competitive as the LS101 strain, showing that the selective marker we used is neutral in the conditions tested.

Promoter Activity Assays.

For promoter activity assays, see SI Text.

Immunoblot Analysis.

DnaA and CcrM proteins were resolved on 10% SDS/PAGE (34). GcrA, CtrA, and MipZ proteins were resolved on 15% SDS/PAGE. Gels were electrotransferred to a PVDF membrane (Millipore, Bedford, MA). Immunodetection was performed with polyclonal antibodies. Anti-DnaA, anti-CtrA, anti-MipZ, and donkey anti-rabbit-conjugated to horseradish peroxidase (Jackson ImmunoResearch, West Grove, PA) sera were diluted 1:10,000. Anti-CcrM serum was diluted 1:5,000. Anti-GcrA serum was diluted 1:2,000. A chemiluminescent reagent (PerkinElmer, Wellesley, MA) and Kodak (Rochester, NY) Bio-Max MR films were used. Images were processed with Photoshop (Adobe, Mountain View, CA), and relative band intensities were determined by using ImageQuant software (Molecular Dynamics, Sunnyvale, CA).


Cells were fixed in 2.5% formaldehyde and 30 mM sodium phosphate buffer (pH7.5). Cells were then placed on a 1% agarose pad, and Nomarski differential interference contrast (DIC) images were taken with ×100 DIC objectives on a E800 microscope (Nikon, East Rutherford, NJ) with 5 MHz Micromax 5600 cooled CCD camera controlled through Metamorph (Universal Imaging, Downingtown, PA). Images were prepared by using Adobe Photoshop and Metamorph version 4.5.

Supplementary Material

Supporting Information:


We thank Craig Stephens for the pPxyl::ccrM plasmid. We thank members of the L.S. and the H.H.M. laboratories for helpful discussions and comments. J.C. was the recipient of a Stanford Dean's Fellowship. This work was supported by National Institutes of Health Grants GM32506 and GM051426 (to L.S.) and 5R24GM73011-2 (to H.H.M. and L.S.).


The authors declare no conflict of interest.

This article contains supporting information online at www.pnas.org/cgi/content/full/0708112104/DC1.


1. Wion D, Casadesus J. Nat Rev Microbiol. 2006;4:183–192. [PMC free article] [PubMed]
2. Zweiger G, Marczynski G, Shapiro L. J Mol Biol. 1994;235:472–485. [PubMed]
3. Stephens C, Reisenauer A, Wright R, Shapiro L. Proc Natl Acad Sci USA. 1996;93:1210–1214. [PMC free article] [PubMed]
4. Wright R, Stephens C, Shapiro L. J Bacteriol. 1997;179:5869–5877. [PMC free article] [PubMed]
5. Robertson GT, Reisenauer A, Wright R, Jensen RB, Jensen A, Shapiro L, Roop RM., II J Bacteriol. 2000;182:3482–3489. [PMC free article] [PubMed]
6. Kahng LS, Shapiro L. J Bacteriol. 2001;183:3065–3075. [PMC free article] [PubMed]
7. Marczynski GT. J Bacteriol. 1999;181:1984–1993. [PMC free article] [PubMed]
8. Dingwall A, Shapiro L. Proc Natl Acad Sci USA. 1989;86:119–123. [PMC free article] [PubMed]
9. Gorbatyuk B, Marczynski GT. Mol Microbiol. 2001;40:485–497. [PubMed]
10. Laub MT, Chen SL, Shapiro L, McAdams HH. Proc Natl Acad Sci USA. 2002;99:4632–4637. [PMC free article] [PubMed]
11. Hottes AK, Shapiro L, McAdams HH. Mol Microbiol. 2005;58:1340–1353. [PubMed]
12. Quon KC, Marczynski GT, Shapiro L. Cell. 1996;84:83–93. [PubMed]
13. Holtzendorff J, Hung D, Brende P, Reisenauer A, Viollier PH, McAdams HH, Shapiro L. Science. 2004;304:983–987. [PubMed]
14. Messer W, Weigel C. Mol Microbiol. 1997;24:1–6. [PubMed]
15. Gil R, Silva FJ, Zientz E, Delmotte F, Gonzalez-Candelas F, Latorre A, Rausell C, Kamerbeek J, Gadau J, Holldobler B, et al. Proc Natl Acad Sci USA. 2003;100:9388–9393. [PMC free article] [PubMed]
16. Collier J, Murray SR, Shapiro L. EMBO J. 2006;25:346–356. [PMC free article] [PubMed]
17. Quon KC, Yang B, Domian IJ, Shapiro L, Marczynski GT. Proc Natl Acad Sci USA. 1998;95:120–125. [PMC free article] [PubMed]
18. Kelly AJ, Sackett MJ, Din N, Quardokus E, Brun YV. Genes Dev. 1998;12:880–893. [PMC free article] [PubMed]
19. Reisenauer A, Quon K, Shapiro L. J Bacteriol. 1999;181:2430–2439. [PMC free article] [PubMed]
20. Wright R, Stephens C, Zweiger G, Shapiro L, Alley MR. Genes Dev. 1996;10:1532–1542. [PubMed]
21. Zweiger G, Shapiro L. J Bacteriol. 1994;176:401–408. [PMC free article] [PubMed]
22. Chen SL, Shapiro L. J Bacteriol. 2003;185:4997–5002. [PMC free article] [PubMed]
23. Mott ML, Berger JM. Nat Rev Microbiol. 2007;5:343–354. [PubMed]
24. Kurokawa K, Nishida S, Emoto A, Sekimizu K, Katayama T. EMBO J. 1999;18:6642–6652. [PMC free article] [PubMed]
25. Sekimizu K, Bramhill D, Kornberg A. Cell. 1987;50:259–265. [PubMed]
26. Speck C, Weigel C, Messer W. EMBO J. 1999;18:6169–6176. [PMC free article] [PubMed]
27. Marczynski GT, Shapiro L. J Mol Biol. 1992;226:959–977. [PubMed]
28. Gorbatyuk B, Marczynski GT. Mol Microbiol. 2005;55:1233–1245. [PubMed]
29. Reisenauer A, Shapiro L. EMBO J. 2002;21:4969–4977. [PMC free article] [PubMed]
30. Ely B. Methods Enzymol. 1991;204:372–384. [PubMed]
31. Simon R, Priefer U, Puhler A. BioTechnology. 1983;1:784–790.
32. Evinger M, Agabian N. J Bacteriol. 1977;132:294–301. [PMC free article] [PubMed]
33. Elena SF, Lenski RE. Nature. 1997;390:395–398. [PubMed]
34. Sambrook J, Fritsch EF, Maniatis T. Molecular Cloning: A Laboratory Manual. Cold Spring Harbor, NY: Cold Spring Harbor Lab Press; 1989.

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...