The diversity and evolution of cell cycle regulation in alpha-proteobacteria: a comparative genomic analysis
© Brilli et al; licensee BioMed Central Ltd. 2010
Received: 27 November 2009
Accepted: 28 April 2010
Published: 28 April 2010
In the bacterium Caulobacter crescentus, CtrA coordinates DNA replication, cell division, and polar morphogenesis and is considered the cell cycle master regulator. CtrA activity varies during cell cycle progression and is modulated by phosphorylation, proteolysis and transcriptional control. In a phosphorylated state, CtrA binds specific DNA sequences, regulates the expression of genes involved in cell cycle progression and silences the origin of replication. Although the circuitry regulating CtrA is known in molecular detail in Caulobacter, its conservation and functionality in the other alpha-proteobacteria are still poorly understood.
Orthologs of Caulobacter factors involved in the regulation of CtrA were systematically scanned in genomes of alpha-proteobacteria. In particular, orthologous genes of the divL-cckA-chpT-ctrA phosphorelay, the divJ-pleC-divK two-component system, the cpdR-rcdA-clpPX proteolysis system, the methyltransferase ccrM and transcriptional regulators dnaA and gcrA were identified in representative genomes of alpha-proteobacteria. CtrA, DnaA and GcrA binding sites and CcrM putative methylation sites were predicted in promoter regions of all these factors and functions controlled by CtrA in all alphas were predicted.
The regulatory cell cycle architecture was identified in all representative alpha-proteobacteria, revealing a high diversification of circuits but also a conservation of logical features. An evolutionary model was proposed where ancient alphas already possessed all modules found in Caulobacter arranged in a variety of connections. Two schemes appeared to evolve: a complex circuit in Caulobacterales and Rhizobiales and a simpler one found in Rhodobacterales.
Living cells continuously receive and process signals coming from their environment, and by integrating this information into their own internal state, are able to react with appropriate responses which coordinate each function in the cell in order to divide and produce progeny. Regulation of cell cycle progression needs to be a robust but versatile process that integrates different exogenous and endogenous signals and that guarantees fidelity and controlled progression throughout the different phases.
Bacteria have evolved different regulation systems for cell cycle coordination, probably due to different ecological and evolutionary constraints [1, 2]. Alpha-proteobacteria subdivision is a very heterogeneous group of bacteria and includes symbionts of plants (Rhizobia), pathogens for animals (Brucella, Rickettsia) and plants (Agrobacterium), photosynthetic bacteria (Rhodobacter) and also several genera metabolizing C1-compounds (Methylobacterium).
Together with this diversity of life styles and ecological niches, the alpha-proteobacteria subdivision is also one of the bacterial groups in which cell cycle regulation has been studied in more detail and one of its members, Caulobacter crescentus, has recently become a model organism in these studies [3–6]. In this organism each cell division is asymmetric—producing two functionally and morphologically different cells, the replicating "stalked" cell type and the vegetative "swarmer" type. After each initiation of DNA replication, the replication fork is kept blocked so that the Caulobacter cell cycle can follow a pattern of once-and-only-once replication per division (G1, S, and G2 phases are temporally distinguished).
Many factors are known to regulate cell cycle progression and most of them are members of the family of two-component signal transduction proteins, comprised of histidine kinases and their response regulator substrates . Among those proteins CtrA is the master regulator of the Caulobacter cell cycle, an essential response regulator whose activity as a transcription factor varies as a function of the cell cycle [7–9]. CtrA controls various functions during cell cycle progression by activating or repressing gene expression. CtrA also blocks the initiation of DNA replication through binding of the replication origin . Among genes regulated by CtrA we can find those involved in cell division (ftsZ, ftsA, ftsQ and ftsW), the protease encoding gene clpP which is essential in Caulobacter, the DNA methylase gene ccrM, flagellar biogenesis genes, stalk biogenesis regulatory genes, pili biogenesis genes such as pilA, and chemotaxis genes [10–15].
CtrA activity and stability varies during the cell cycle. Oscillation of CtrA levels, peaking at the predivisional stage before cell division, is achieved by different mechanisms: transcription, proteolysis and phosphorylation control as discussed in detail below.
DnaA and GcrA, and the DNA methyltransferase CcrM are involved in controlling ctrA transcription [11, 16]. DnaA is a key element in cell cycle regulation because it is required for the initiation of DNA replication; it also controls the transcription of about 40 genes involved in nucleotide biogenesis, cell division, and polar morphogenesis [17, 18], and it activates the transcription of the gcrA gene . 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 . As a consequence of this genetic circuit, CtrA accumulates out-of-phase with GcrA . The transcriptional loop of ctrA is closed by CcrM. CtrA activates the transcription of ccrM, which encodes for a DNA methyltransferase whose abundance is cell cycle dependent. CcrM is able to activate dnaA promoter region through methylation, closing the positive feedback composed by CtrA, DnaA and GcrA.
A second essential regulatory control on CtrA is carried out by phosphorylation. In fact, CtrA must be phosphorylated to bind DNA and its phosphorylation depends on cell cycle progression. An essential phosphorelay, composed of the hybrid histidine kinase CckA and the histidine phosphotransferase ChpT, is responsible for CtrA phosphorylation [20, 21].
DivK, which is a response regulator, plays an essential role as a positive regulator of cell cycle progression because when phosphorylated, it indirectly inactivates CtrA and thus promotes DNA replication. Two histidine kinases are known to interact with DivK: PleC and DivJ [22–25]. Bacterial histidine kinases can have alternatively both kinase and phosphatase activities and these opposite activities are modulated by conformational changes of the protein . A null Caulobacter pleC mutant produces almost symmetric cells at the division and displays abnormal polar development. The DivJ histidine kinase plays a role in controlling the length and location of the stalk and cell division. PleC and DivJ are considered the principal phosphatase and kinase, respectively, of DivK and they are in opposite locations during cell cycle progression [27, 28]. DivJ activity is also positively controlled by the TacA/SpmX pathway, which is transcriptionally activated by CtrA [10, 29].
ChpT also transfers the phosphate to a second response regulator named CpdR, which, together with RcdA, is a factor involved in CtrA proteolysis mediated by ClpP-ClpX protease [30–32]. CtrA is degraded at the stalked pole at the G1/S transition when the origin of replication needs to be cleared and also in the stalked compartment, where initiation of DNA replication occurs immediately after cell division [33, 34].
Several of these regulatory mechanisms are at least partially redundant, and it has been demonstrated that only phosphorylation of CtrA is indispensable during cell cycle progression; in fact, cell cycle regulated transcription of ctrA can be substituted by constitutive transcription  and proteolysis can also be removed.
It has recently been demonstrated that asymmetric division also takes place in Agrobacterium tumefaciens, Sinorhizobium meliloti and Brucella abortus, indicating that at least some of the features governing cell cycle progression in Caulobacter might also be present in other species. For example, in Rhodobacter capsulatus, CtrA and CckA are not essential but are required for the expression of the GTA, a system for genetic exchanges [36, 37]. CtrA in Brucella controls cellular events similar to those controlled by CtrA in Caulobacter, but through a direct effect on different targets . Moreover CtrA from Caulobacter is able to bind the B. abortus ccrM promoter in vitro  and CtrA from B. abortus has been shown to bind promoters of ccrM, pleC, rpoD, ftsE and minC but not divK, ftsZ or the origin of replication, that are known CtrA targets in Caulobacter. In Silicibacter pomeroy three known mutants affect the motility, two of which are cckA, ctrA.
ccrM is essential and cell cycle regulated in A. tumefaciens, as observed in B. abortus, where ccrM is also essential and its overexpression impairs proper intracellular replication in murine macrophages , revealing a link between cell cycle and pathogenetic activity.
We hereby undertook a comparative and integrative analysis of factors controlling cell cycle regulation, and the regulatory connections existing between them, looking for orthologous genes of factors that are involved in controlling CtrA, the master regulator, in C. crescentus. We also analyzed CtrA-regulated functions in 37 representative genomes of alphas. At the same time, binding sites of DnaA, GcrA and the presence of putative methylation sites of CcrM were also suggested through bioinformatic analysis. All this information was used to reconstruct the architecture of CtrA regulation throughout the phylogenetic tree of alpha-proteobacteria to reveal evolutionary trends and insights into this complex regulation.
Evolutionary scheme of alpha-proteobacteria
Clusterization of alpha-proteobacteria based on the orthology of cell cycle genes
By using the bidirectional best hit (BBH) approach (see Material and Methods section) on 65 available genomes of alpha-proteobacteria (Additional file 1, Table S1; legends of additional Figures and Tables are in Additional file 2) we obtained a list of genes that are orthologous to the 14 genes involved in the Caulobacter cell cycle, and the results are reported in Figure 2 (see also Additional file 3, Figure S1 and Additional file 4, Table S2).
ClpX, ClpP and DnaA are present in all alphas studied, but surprisingly all other proteins analyzed can be absent in several alphas. Transcription factors, GcrA and CtrA, the DNA methyl-transferase CcrM and the hybrid histidine kinase CckA are present in most of the alpha-proteobacteria. Other modules, such as those of the DivJ-PleC-DivK two-component system, are present only in clusters A and C of alpha-proteobacteria.
Genes with similar phylogenetic profiles (genes co-occurring in different genomes) are often functionally related  justifying the use of our profiles to investigate possible functional relationships; the dendrogram obtained describes how similar the profiles of different genes are (see Methods and upper part of Figure 2) and it confirms the functional association between divJ, pleC and divK (encoding the two component system negatively regulating CckA activity), and between cpdR and rcdA, whose products are involved in CtrA proteolysis. Weaker and possibly new functional associations concern the gene pair ccrM/gcrA and divL/chpT. In addition, we visually inspected the phylogenetic profiles of these genes between organisms, identifying seven groups (from A to G, see Figure 2). This classification, based on orthology, will be used as a reference in the following sections.
Cluster A includes Rhizobiales, Caulobacterales and several Rhodobacterales, and is composed of the largest number of sequenced genomes; this cluster is characterized by a nearly identical conservation of factors found in Caulobacter. Although similarities are evident in this cluster, a deeper analysis revealed substantial differences that will be discussed in the next sections.
Cluster B, including other Rhodobacterales, shows a substantial difference compared to cluster A; in fact, both DivJ-PleC-DivK and RcdA-CpdR systems are missing.
Magnetospirillum and Rhodospirillum, which are closely related, are the two members of cluster C. This cluster is characterized by the presence of the PleC-DivK system since DivJ is missing and by an almost complete loss of the CpdR-RcdA system although Magnetospirillum possesses an rcdA orthologous gene.
Clusters D (Rhodospirillales) and F (part of Rickettsiales), even though they are separated in the tree reported in Figure 2, share common features: i. members of the CckA-ChpT-CtrA phosphorelay are missing at different degrees (Granulibacter and Pelagibacter do not show even ctrA orthologs); ii. DivJ-PleC-DivK and RcdA-CpdR systems are missing. Despite these similarities, the two groups diverge for the presence of CcrM and GcrA in group D.
Organisms belonging to cluster E (Sphingomonadales) show conservation of the phosphorelay CckA-ChpT-CtrA and also often possess factors required for the temporally and spatially regulated proteolysis of CtrA, such as CpdR and RcdA. However cluster E is characterized by degeneration or a complete loss of the DivJ-PleC-DivK regulation system. DivK or PleC orthologs can be found in several organisms of this subgroup although their phylogeny often significantly deviate from the phylogenetic tree of housekeeping genes (Additional file 5, Figure S2).
Finally, group G (remaining Rickettsiales), composed mainly of pathogens, has few of the factors involved in Caulobacter cell cycle progression regulation. It is however interesting to find a CckA-CtrA system whereas ChpT orthologs cannot be found using the BBH approach.
CtrA regulon in alpha-proteobacteria
The regulatory circuit that controls cell cycle progression in Caulobacter is also composed of crucial transcriptional connections, such as CtrA control on divK and the CtrA-DnaA-GcrA-CcrM circuit. This transcriptional regulation level is discussed in this section and the following. In particular, results obtained for the prediction of the CtrA regulon in alpha-proteobacteria are described here.
Laub and collaborators [13, 44] were able to identify a set of genes plausibly constituting the CtrA regulon in Caulobacter by combining varying evidence: 116 genes were identified through chromatin immunoprecipitation using phosphorylated CtrA; 88 genes were identified as CtrA-dependent for normal expression levels, and 69 as cell cycle regulated in a transcriptome analysis encompassing one complete cell cycle round. The 54 genes within the overlap of all three data sets were identified as members of the CtrA cell cycle regulon, and were used here to build the position weight matrix (PWM) of CtrA binding sites. Upstream sequences of these 54 genes were retrieved and used to find enriched sequence motifs using AlignAce . The PWM obtained (Additional file 6, Table S3) corresponds to a 16-mer containing the known CtrA binding motif and was used in a sliding window approach on a non-redundant subset of the genomes used in this work. An output was obtained where genes in a given genome have a score based on the presence of CtrA sequence motifs in the region comprised of 100 nucleotides within the coding sequence to 400 nucleotides upstream of the start codon (see Methods for details).
Is the CtrA PWM, based on Caulobacter data, valid for all the alphas analyzed here? One might speculate that if a ctrA gene taken from an alpha is able to complement deletion in Caulobacter, the PWM built on Caulobacter ctrA-controlled genes would also be valid for the bacterium where the complementing ctrA comes from. It has been shown that the ctrA gene from R. prowazekii, named czcR, is able to complement the deletion of ctrA in Caulobacter, confirming that the functionality (that is, the binding site) is conserved between Rickettsia and Caulobacter. Moreover other ctrA genes from species taxonomically closer to Caulobacter, such as S. meliloti, are able to complement the ctrA deletion in Caulobacter (Biondi, unpublished data). Considering the phylogeny of alphas and positions in the tree of R. prowazekii and S. meliloti, it is reasonable to consider that the CtrA binding site might be substantially conserved across the alphas.
In Figure 3 (see also Additional file 7, Table S4) we show the p-values for the presence of CtrA binding sites upstream of analyzed genes. CtrA controls the transcription of several genes involved in regulation of cell cycle progression, including itself in most of the alphas analyzed (86%). Moreover, the number of genes controlled by CtrA is maximal in cluster A. In this cluster CtrA controls at least one gene of each of the following systems: DivJ-DivK-PleC, CpdR-RcdA-ClpPX and GcrA-DnaA-CcrM. Several genes showed evolutionary conservation of CtrA control among members of cluster A, such as DivJ, RcdA and CcrM. The second important result arising from the analysis shown in Figure 3 was that in Cluster B, where the DivJ-PleC-DivK system is missing, CtrA controls both CckA and DivL.
Each genome-wide regulon of CtrA was defined as the list of genes with a Z-score (see Methods) ≥2 (corresponding to a p-value of ca. 0.023) in an organism. In Figure 4 (see also Additional file 8, Table S5), predicted regulons of CtrA were analyzed for functional enrichment (percentage of genes in a COG category controlled by CtrA) in genes belonging to functional categories. Most enriched categories were Signal transduction mechanisms (enriched in 15 organisms), Cell wall/membrane/envelope biogenesis (enriched in 10 organisms) and Cell motility (enriched in 9 organisms), while cell cycle functions were enriched in six species, belonging to cluster A, including Caulobacter and Neorickettsia sennetsu of cluster F. These results confirmed experimental data on the functions controlled by CtrA (see Background section) in Caulobacter suggesting that (i) the analysis of the regulon is able to capture good candidates of CtrA targets and (ii) the control of CtrA over these functions is at least partially evolutionarily conserved.
In Caulobacter, transcriptional regulation of ctrA is based on a positive feedback loop that includes CtrA itself, GcrA, DnaA and CcrM.
As reported in Figure S3 (Additional file 9) (see also Additional file 7, Table S4) the presence of CcrM methylation sites upstream of cell cycle related genes was assessed (see Methods section). For this analysis we used the consensus methylation site of CcrM, i.e. GANTC . As reported elsewhere [39, 48], methylation by CcrM is conserved among alphas and ccrM genes from Caulobacter and S. meliloti which are able to cross-complement deletions, suggesting that the methylation site might be conserved. In a homogeneous background of DNA, the expected frequency of this sequence is 4/1024 nucleotides, i.e. two occurrences in the 500 bp long window that was used to define the promoter region for motif finding. The average number of occurrences found ranged from 0.4 to 1.5 methylation sites per promoter (defined as the 500 bp from the first 100 bp of a gene to 400 bp upstream of the translation start site), according to non-uniform distributions of nucleotides in these genomes. Although some genes possess more predicted methylation sites in their promoter region, it is not possible to derive a possible control of methylation in their expression.
Although GcrA is considered a DNA binding protein activating transcription, no experimental evidence has ever been proposed to demonstrate this behavior . However it is still possible that DNA sequences are associated with the presence of this factor. GcrA putative binding sites were searched for in promoter regions of genes encoding factors involved in cell cycle regulation using a strategy similar to the one followed to predict CtrA regulons (see Methods section). In C. crescentus, as reported in Figure 5 (see also Additional file 7, Table S4), the existence of a putative GcrA binding motif (Additional file 10, Figure S4) upstream of ctrA and also the presence of such motifs upstream of divJ was confirmed. Concerning the other species, patterns of occurrence did not respect phylogenetic relationships. The only gene for which most of the organisms seem to possess a GcrA binding site is clpX.
Verification of binding site prediction
CtrA binding site consensus has been previously tested experimentally in R. prowazekii, S. meliloti and B. abortus, besides Caulobacter[38, 46, 52]. Here, we compared the experimental consensus sequences with our bioinformatic PWM (Figure 6C), and our prediction coincides with experimentally identified sequences. Our PMW corresponds also with a CtrA PMW previously found .
The DnaA binding site has been studied in very diverse bacteria such as Gram-negative Escherichia coli and Gram-positive Bacillus subtilis[49, 50]. The DnaA binding site in these two species differs because of one nucleotide, suggesting that the binding site should also be very conserved in alpha proteobacteria. We compared the predicted PWM for DnaA based on C. crescentus with experimental DnaA binding sites in E. coli, B. subtilis and S. meliloti (Figure 6C) [49, 50, 54]. Our prediction, based on nucleotide sequences that bind DnaA in Caulobacter, corresponds to binding sites experimentally found in other bacteria.
This verification was possible only for DnaA and CtrA, while GcrA has been studied only in C. crescentus and experimental data are available only in this organism. It has not been clarified whether GcrA binds DNA directly or through an unknown factor X . Therefore, since the knowledge on GcrA is still preliminary and experimental work still needs to be done, we limited the experimental validation to CtrA and DnaA, for which data are available. It should be noted, however, that both DnaA and CtrA experimental verifications revealed that our method is accurate and reliable.
Reconstruction of regulatory circuits
Cluster A (Caulobacterales, Rhizobiales and several Rhodobacterales) contains the larger number of genomes analyzed here and the organization of cell cycle genes resembles that observed in Caulobacter (see bottom part of Figure 7 for details), i.e. it includes the phosphorelay CckA-ChpT-CtrA/CpdR and also the proteolysis machinery composed by the ubiquitous ClpPX protease, CpdR and RcdA, which is however absent in Xanthobacter. The DivJ-PleC-DivK system is conserved and corresponding genes are controlled by CtrA in all members of the cluster. CcrM also controls CtrA and GcrA.
Rhizobiales are different from bacteria similar to Caulobacter (B. japonicum, P. lavamentivorans and M. maris) due to the absence of the CtrA control on GcrA which is present only in the Caulobacter- like.
In cluster E, the CckA-ChpT-CtrA phosphorelay is present with the second branch also leading to phosphorylation of CpdR that, together with RcdA, are thought to be involved in controlling CtrA proteolysis. DivK is absent in this cluster although a divK-like gene has been found although it has an anomalous phylogeny (Additional file 3, Figure S1). In most members of cluster B, CtrA controls its own promoter, other genes involved in cell division and chromosome partitioning as well as ccrM.
In cluster B, the CckA-ChpT-CtrA regulon is isolated from GcrA, CcrM and DnaA while CtrA controls itself as well as two factors involved in its phosphorylation, CckA and DivL.
In this group, DivL lacks the kinase domain that is usually present only when DivK is also present (data not shown, based on SMART database). In fact, DivK is absent in cluster B together with its kinase/phosphatase. The fact that CtrA controls its kinase creates theoretically a feedback.
In cluster C, the CckA-ChpT-CtrA phosphorelay is present while CpdR is absent. Also CtrA is not connected with DnaA and CcrM and finally DnaA has binding sites on ccrM and divL. Connections between DnaA/CcrM and CtrA seem to be achieved by the PleC-DivK two-component system. No clear positive or negative transcriptional feedbacks of CtrA on other cell cycle factors are present.
Cluster D contains the two cases among alphas, Granulibacter and Pelagibacter, where a ctrA ortholog has not been found. The phosphorelay, even in organisms of cluster E that have CtrA, is degenerated; although cckA is present, no orthologous gene of chpT has been found. There is no explanation for the presence of CckA in organisms with no CtrA. Since histidine phosphotransferases are difficult to annotate [10, 20], it is possible that other phosphotransferases substitute for ChpT in those organisms containing both CtrA and CckA, as proposed here for cluster G.
The cell cycle 'engine' (controlling DNA replication, cell division, morphogenesis of polar structures) is an essential machine in every living organism, and a major goal of molecular cell biologists is to uncover how this engine works in every organism. We have performed a bioinformatic analysis of cell cycle control genes in the alpha proteobacteria, taking cues from the well-characterized control system in Caulobacter (see Background section for details).
The procedure used here is able to detect only loss of regulatory points with respect to Caulobacter. However, the knowledge of conservations and variations from the Caulobacter scheme can be extremely useful for future studies of the cell cycle network in all alpha-proteobacteria.
Conservation and variability of the CtrA regulatory system
CtrA regulation is thought to play an essential role in most alpha-proteobacteria cell cycle progression while its activity varies during each cell cycle in response to several levels of control in Caulobacter. As illustrated in Figure 7, many regulators are connected with CtrA, via epigenetic, transcriptional and post-translational mechanisms in most of the clusters. The reconstructed architecture within alpha-proteobacteria of regulatory pathways that are involved in controlling CtrA activity is however surprisingly variable, suggesting diverse evolutionary pathways in each cluster.
CtrA, even if associated with variable regulatory circuits within alphas, shows conservation of the control of certain functions (Figure 4) such as cell division, motility and signal transduction, especially in cluster A where CtrA is, in fact, the master regulator of cell cycle. The ability to detect the same functions controlled by CtrA in different organisms also suggests that the prediction capability of CtrA-regulated genes is highly reliable.
The C. crescentus regulatory scheme of CtrA in other members of cluster A shows several variations, especially in Rhizobiales. For example, the control of CtrA on the response regulator divK, observed in C. crescentus, is shifted to the gene encoding the DivK kinase (divJ) and/or the phosphatase (pleC) in most Rhizobiales. This observation may suggest that feedbacks can be conserved even when connections are rewired.
Another interesting feature revealed by this study is the control of CtrA on cckA and divL in several members of the Rhodobacterales (cluster B), which is coupled to the absence of the system regulating CckA activity (the DivK/DivJ/PleC two-component system). Thus, in cluster A members, CtrA controls the divK-divJ-pleC transcription and DivK presumably inhibits CckA activation that is indeed able to activate CtrA through a phosphorylation cascade, while in cluster B Rhodobacterales our results suggest that CtrA acts directly on cckA transcription and thus it may directly modulate the CckA quantity in the cytoplasm.
Other regulatory circuits (DivL-DivK and DnaA-GcrA-CtrA)
Several conclusions can be made at the molecular level from this bioinformatic approach on specific factors such as DivL-DivK or the DnaA-GcrA-CtrA system. DivL, that still lacks a precise function, and DivK, have been connected in previous studies since DivL was detected in a yeast-two hybrid experiment using DivK as bait . Alleles of divL were also able to complement divK defective alleles [24, 56]. A further association between DivL and DivK was found in our study whereby orthologs of DivL can be found in alpha-proteobacteria either as complete kinases (such as in Caulobacter) or without the kinase domain (data not shown); this latter form is always associated with the absence of DivK--suggesting that the kinase domain of DivL functions together with DivK. This observation is also consistent with previous studies that showed that DivL interacts with DivK using the kinase domain and has two separate functions in Caulobacter. The observation that in several alphas DivL keeps only the sensor part is also paralleled by previous studies showing that its kinase activity in Caulobacter is dispensable .
It can be concluded from this study that the DnaA-GcrA-CtrA regulatory system in Caulobacter is not conserved throughout alpha-proteobacteria. It is however true that DnaA controls gcrA in several members of cluster A (surprisingly this control in Caulobacter was not detected) and that GcrA controls ctrA in few alphas, including in Caulobacter. This observation might suggest that this important loop in the regulation of the Caulobacter cell cycle is instead dispensable in other bacteria of cluster A. This conclusion is also consistent with the result that regulation of ctrA transcription can be removed without affecting Caulobacter viability although robustness of the system might be compromised.
Evolution of alpha-proteobacteria cell cycle
Finally, the reconstruction of the evolution of the regulatory schemes found was undertaken following phylogeny proposed elsewhere [42, 57]. Rickettsiales (clusters F and G) were excluded from this model due to their massive genome reduction that has been previously shown to be a consequence of the evolution of an obligate intracellular life in eukaryotic cells .
The regulation of cell cycle progression in Caulobacter has evolved in order to respond to a lifestyle in nutrient-poor environments but other alpha-proteobacteria occupy different ecological niches, suggesting that cell cycle regulation must respond to different requirements; from an evolutionary perspective this means that features found in Caulobacter should not be completely conserved in other alpha-proteobacteria, especially those experiencing different environments. Conversely, similarities between closer organisms were expected due to common phylogenetic ancestries.
Gupta and Mok  proposed that Rhodospirillales and Novosphingomonadales (clusters C, D and E) branched earlier than the other alphas (clusters A and B), and after Rickettsiales (clusters F and G). Two of those "older" clusters (C and E) show, in fact, a different arrangement of regulatory factors although none show circularity of the regulation created by feedbacks. From this primordial situation, when alphas were experimenting different options, two situations seem to have evolved: a complex circuit (cluster A) and a minimal organization (cluster B). In both clusters two interlaced oscillators are present: besides cluster A where this organization has already been proposed , cluster B, also shows a CtrA auto-regulatory circuit and a CtrA-CckA-ChpT-CtrA oscillator. It is also interesting to point out that the minimal circuit of cluster C corresponds to a situation where CtrA is not essential.
This is the first systematic attempt to translate the information accumulated over the years on cell cycle regulation in the model Caulobacter system into a common body of knowledge regarding the whole taxonomic group of alpha-proteobacteria. Results suggest that the scheme found in Caulobacter may work in closely related bacteria such as those belonging to Caulobacterales and Rhizobiales while in other alphas this conservation is lost even though several modules are present. Finally, this analysis represents an important step for all future cell cycle studies in alpha-proteobacteria, offering many experimental scenarios designed to confirm or reject the observations made here.
To construct our reference phylogenetic tree, from the server of the RDP (ribosomal database project) , we downloaded the E. coli sequences corresponding to universal proteins in prokaryotes: FusA, IleS, LepA, LeuS, PyrG, RecA, RecG and RplB. Then we blasted these sequences against alpha-proteobacterial genomes chosen for analysis, retrieving the BBH for each protein in each genome. Each dataset was aligned using Muscle  and the obtained multi-alignments were then joined head-to-tail in a single concatamer of 5000 sites. This alignment was used with the software Mega 4  running the Neighbor-Joining algorithm, 500 bootstrap replicates and the Dayhoff model of evolution.
We used a dataset comprising all completely sequenced genomes of organisms belonging to the alpha-proteobacteria for ortholog identification (Additional file 1, Table S1). Orthologous genes were identified with the so-called bidirectional best hit (BBH) criterion: the relation of gene x in genome A and gene y in genome B is called bidirectional best hit, when x is the best hit of query y against all genes in A and vice versa. We used two different datasets of queries coming from C. crescentus and B. melitensis and we also performed a phylogenetic analysis of those proteins found using the BBH analysis; the Caulobacter dataset is derived from experimental data  while that from B. abortus is derived from both experimental  and bioinformatic analysis performed using Caulobacter genes. For multi-domain proteins we used separated domains and discarded highly variable regions from the analysis. The two sets of BBH sequences orthologous to Caulobacter and Brucella sequences were then combined to collect the maximum number of proteins. When C. crescentus and B.melitensis sequences retrieved two different proteins for the same gene, we collected both of them and used phylogenetic methods to solve the mismatch; when one of the two queries returned no BBH we used only the one available. We used a 0.0001 e-value threshold in the Blast analysis.
Regulons of CtrA, GcrA, DnaA and CcrM methylation sites characterization
A Position Weight Matrix (PWM) describing CtrA binding sites in C. crescentus was obtained by using the AlignAce program  on upstream sequences of 54 genes previously identified as being part of the CtrA regulon . We used such a matrix to scan all genomes considered in this work (see Additional File 1, Table S1 for all organisms list) with a sliding window approach and a scoring function from Schneider and collaborators (1986) : Si = (1/L) Σj [2 + log2(Fij)], where Fij is the frequency of base i at position j of the L-mer. This score, whose maximum for the best match using a CtrA position weight matrix is ≈1.22, is a measure of the information content of a potential binding site.
We retained only motifs having a score ≥30% of the maximum score attainable with the given matrix and located in the range -100 to 400 nucleotides from the start codon of a gene.
Then we applied a Z-score transformation to highlight significant occurrences and take into account the background DNA implicitly: Zi = (Si - <Si>)/σi, where Si is calculated using the above formula, and <Si> is the average score in an organism over all L-mers and σi is the corresponding standard deviation.
For GcrA and DnaA the strategy followed was the same, except that PWMs were obtained differently. For GcrA we obtained a PWM describing a motif enriched upstream of sequences controlled by GcrA as proposed by others . By using microarray and GcrA deficient strains, a list of genes whose expression changes significantly when GcrA is mutated was obtained. We used the 48 genes with maximum fold change in that list. The corresponding upstream sequences were retrieved and analyzed with AlignAce  and MDscan . In both cases a high scoring motif (16 bp, see Additional file 10, Figure S4) present in a high percentage of the input sequences was obtained, and was used for subsequent scanning of the genomes of our dataset (Figure 5 and Additional file 7, Table S4).
DnaA recognizes a sequence whose consensus in Caulobacter is [TC] [TCG] [AG]TCCACA as previously reported . In the same work, 15 DnaA binding sites were shown and were used to build a position weight matrix specific for DnaA (Additional file 10, Figure S4). However AlignAce and MDscan were not able to identify motifs for DnaA corresponding at least partially to the proposed consensus using two datasets of genes previously proposed .
Both GcrA and DnaA were then analyzed as reported in points 2 and 3 for CtrA. CcrM methylation targets were searched using a regular expression modelling the known target of this methyl-transferase (GANTC) .
We took advantage of COG  categories to evaluate the functional enrichment of the genes identified with the approach described above. Supposing that the regulon dataset of a genome is composed of N genes, we counted the number of genes for each category and then repeated the same count for 10,000 N-sized groups of genes randomly picked from the genome. We obtain a p-value describing functional enrichment in the original dataset by counting how many times a COG category is more represented in the random group and then dividing it by the total number of repetitions performed.
position weight matrix
bidirectional best hit
clusters of orthologous groups of proteins
simple modular architecture research tool
We thank Prof. Michael Laub for critical reading of the manuscript. We thank also Mary C. Forrest for editing. Emanuele G. Biondi is supported by the "Fondazione Adriano Buzzati-Traverso" and "Ente Cassa di Risparmio di Firenze".
- Fawcett P, Eichenberger P, Losick R, Youngman P: The transcriptional profile of early to middle sporulation in Bacillus subtilis. Proc Natl Acad Sci USA. 2000, 97: 8063-8068. 10.1073/pnas.140209597PubMed CentralView ArticlePubMedGoogle Scholar
- Haeusser DP, Levin PA: The great divide: coordinating cell cycle events during bacterial growth and division. Curr Opin Microbiol. 2008, 11: 94-99. 10.1016/j.mib.2008.02.008PubMed CentralView ArticlePubMedGoogle Scholar
- Bowers LM, Shapland EB, Ryan KR: Who's in charge here? Regulating cell cycle regulators. Curr Opin Microbiol. 2008, 11: 547-552. 10.1016/j.mib.2008.09.019View ArticlePubMedGoogle Scholar
- Laub MT, Shapiro L, McAdams HH: Systems biology of Caulobacter. Annu Rev Genet. 2007, 41: 429-441. 10.1146/annurev.genet.41.110306.130346View ArticlePubMedGoogle Scholar
- Purcell EB, Boutte CC, Crosson S: Two-component signaling systems and cell cycle control in Caulobacter crescentus. Adv Exp Med Biol. 2008, 631: 122-130. full_textView ArticlePubMedGoogle Scholar
- Skerker JM, Laub MT: Cell-cycle progression and the generation of asymmetry in Caulobacter crescentus. Nat Rev Microbiol. 2004, 2: 325-337. 10.1038/nrmicro864View ArticlePubMedGoogle Scholar
- Domian IJ, Quon KC, Shapiro L: Cell type-specific phosphorylation and proteolysis of a transcriptional regulator controls the G1-to-S transition in a bacterial cell cycle. Cell. 1997, 90: 415-424. 10.1016/S0092-8674(00)80502-4View ArticlePubMedGoogle Scholar
- Quon KC, Marczynski GT, Shapiro L: Cell cycle control by an essential bacterial two-component signal transduction protein. Cell. 1996, 84: 83-93. 10.1016/S0092-8674(00)80995-2View ArticlePubMedGoogle Scholar
- Reisenauer A, Quon K, Shapiro L: The CtrA response regulator mediates temporal control of gene expression during the Caulobacter cell cycle. J Bacteriol. 1999, 181: 2430-2439.PubMed CentralPubMedGoogle Scholar
- Biondi EG, Skerker JM, Arif M, Prasol MS, Perchuk BS, Laub MT: A phosphorelay system controls stalk biogenesis during cell cycle progression in Caulobacter crescentus. Mol Microbiol. 2006, 59: 386-401. 10.1111/j.1365-2958.2005.04970.xView ArticlePubMedGoogle Scholar
- Collier J, McAdams HH, Shapiro L: A DNA methylation ratchet governs progression through a bacterial cell cycle. Proc Natl Acad Sci USA. 2007, 104: 17111-17116. 10.1073/pnas.0708112104PubMed CentralView ArticlePubMedGoogle Scholar
- Jones SE, Ferguson NL, Alley MR: New members of the ctrA regulon: the major chemotaxis operon in Caulobacter is CtrA dependent. Microbiology. 2001, 147: 949-958.View ArticlePubMedGoogle Scholar
- Laub MT, Chen SL, Shapiro L, McAdams HH: Genes directly controlled by CtrA, a master regulator of the Caulobacter cell cycle. Proc Natl Acad Sci USA. 2002, 99: 4632-4637. 10.1073/pnas.062065699PubMed CentralView ArticlePubMedGoogle Scholar
- Skerker JM, Shapiro L: Identification and cell cycle control of a novel pilus system in Caulobacter crescentus. EMBO J. 2000, 19: 3223-3234. 10.1093/emboj/19.13.3223PubMed CentralView ArticlePubMedGoogle Scholar
- Wortinger M, Sackett MJ, Brun YV: CtrA mediates a DNA replication checkpoint that prevents cell division in Caulobacter crescentus. EMBO J. 2000, 19: 4503-4512. 10.1093/emboj/19.17.4503PubMed CentralView ArticlePubMedGoogle Scholar
- Collier J, Murray SR, Shapiro L: DnaA couples DNA replication and the expression of two cell cycle master regulators. EMBO J. 2006, 25: 346-356. 10.1038/sj.emboj.7600927PubMed CentralView ArticlePubMedGoogle Scholar
- Gorbatyuk B, Marczynski GT: Regulated degradation of chromosome replication proteins DnaA and CtrA in Caulobacter crescentus. Mol Microbiol. 2005, 55: 1233-1245. 10.1111/j.1365-2958.2004.04459.xView ArticlePubMedGoogle Scholar
- Hottes AK, Shapiro L, McAdams HH: DnaA coordinates replication initiation and cell cycle transcription in Caulobacter crescentus. Mol Microbiol. 2005, 58: 1340-1353. 10.1111/j.1365-2958.2005.04912.xView ArticlePubMedGoogle Scholar
- Holtzendorff J, Hung D, Brende P, Reisenauer A, Viollier PH, McAdams HH, Shapiro L: Oscillating global regulators control the genetic circuit driving a bacterial cell cycle. Science. 2004, 304: 983-987. 10.1126/science.1095191View ArticlePubMedGoogle Scholar
- Biondi EG, Reisinger SJ, Skerker JM, Arif M, Perchuk BS, Ryan KR, Laub MT: Regulation of the bacterial cell cycle by an integrated genetic circuit. Nature. 2006, 444: 899-904. 10.1038/nature05321View ArticlePubMedGoogle Scholar
- Jacobs C, Domian IJ, Maddock JR, Shapiro L: Cell cycle-dependent polar localization of an essential bacterial histidine kinase that controls DNA replication and cell division. Cell. 1999, 97: 111-120. 10.1016/S0092-8674(00)80719-9View ArticlePubMedGoogle Scholar
- Burton GJ, Hecht GB, Newton A: Roles of the histidine protein kinase pleC in Caulobacter crescentus motility and chemotaxis. J Bacteriol. 1997, 179: 5849-5853.PubMed CentralPubMedGoogle Scholar
- Ohta N, Lane T, Ninfa EG, Sommer JM, Newton A: A histidine protein kinase homologue required for regulation of bacterial cell division and differentiation. Proc Natl Acad Sci USA. 1992, 89: 10297-10301. 10.1073/pnas.89.21.10297PubMed CentralView ArticlePubMedGoogle Scholar
- Sommer JM, Newton A: Pseudoreversion analysis indicates a direct role of cell division genes in polar morphogenesis and differentiation in Caulobacter crescentus. Genetics. 1991, 129: 623-630.PubMed CentralPubMedGoogle Scholar
- Wu J, Ohta N, Newton A: An essential, multicomponent signal transduction pathway required for cell cycle regulation in Caulobacter. Proc Natl Acad Sci USA. 1998, 95: 1443-1448. 10.1073/pnas.95.4.1443PubMed CentralView ArticlePubMedGoogle Scholar
- Weiss V, Kramer G, Dunnebier T, Flotho A: Mechanism of regulation of the bifunctional histidine kinase NtrB in Escherichia coli. J Mol Microbiol Biotechnol. 2002, 4: 229-233.PubMedGoogle Scholar
- Jacobs C, Hung D, Shapiro L: Dynamic localization of a cytoplasmic signal transduction response regulator controls morphogenesis during the Caulobacter cell cycle. Proc Natl Acad Sci USA. 2001, 98: 4095-4100. 10.1073/pnas.051609998PubMed CentralView ArticlePubMedGoogle Scholar
- Wheeler RT, Shapiro L: Differential localization of two histidine kinases controlling bacterial cell differentiation. Mol Cell. 1999, 4: 683-694. 10.1016/S1097-2765(00)80379-2View ArticlePubMedGoogle Scholar
- Radhakrishnan SK, Thanbichler M, Viollier PH: The dynamic interplay between a cell fate determinant and a lysozyme homolog drives the asymmetric division cycle of Caulobacter crescentus. Genes Dev. 2008, 22: 212-225. 10.1101/gad.1601808PubMed CentralView ArticlePubMedGoogle Scholar
- Iniesta AA, McGrath PT, Reisenauer A, McAdams HH, Shapiro L: A phospho-signaling pathway controls the localization and activity of a protease complex critical for bacterial cell cycle progression. Proc Natl Acad Sci USA. 2006, 103: 10935-10940. 10.1073/pnas.0604554103PubMed CentralView ArticlePubMedGoogle Scholar
- Jenal U, Fuchs T: An essential protease involved in bacterial cell-cycle control. EMBO J. 1998, 17: 5658-5669. 10.1093/emboj/17.19.5658PubMed CentralView ArticlePubMedGoogle Scholar
- McGrath PT, Iniesta AA, Ryan KR, Shapiro L, McAdams HH: A dynamically localized protease complex and a polar specificity factor control a cell cycle master regulator. Cell. 2006, 124: 535-547. 10.1016/j.cell.2005.12.033View ArticlePubMedGoogle Scholar
- Ryan KR: Partners in crime: phosphotransfer profiling identifies a multicomponent phosphorelay. Mol Microbiol. 2006, 59: 361-363. 10.1111/j.1365-2958.2005.04980.xView ArticlePubMedGoogle Scholar
- Ryan KR, Huntwork S, Shapiro L: Recruitment of a cytoplasmic response regulator to the cell pole is linked to its cell cycle-regulated proteolysis. Proc Natl Acad Sci USA. 2004, 101: 7415-7420. 10.1073/pnas.0402153101PubMed CentralView ArticlePubMedGoogle Scholar
- Hallez R, Bellefontaine AF, Letesson JJ, De Bolle X: Morphological and functional asymmetry in alpha-proteobacteria. Trends Microbiol. 2004, 12: 361-365. 10.1016/j.tim.2004.06.002View ArticlePubMedGoogle Scholar
- Lang AS, Beatty JT: Genetic analysis of a bacterial genetic exchange element: the gene transfer agent of Rhodobacter capsulatus. Proc Natl Acad Sci USA. 2000, 97: 859-864. 10.1073/pnas.97.2.859PubMed CentralView ArticlePubMedGoogle Scholar
- Lang AS, Beatty JT: A bacterial signal transduction system controls genetic exchange and motility. J Bacteriol. 2002, 184: 913-918. 10.1128/jb.184.4.913-918.2002PubMed CentralView ArticlePubMedGoogle Scholar
- Bellefontaine AF, Pierreux CE, Mertens P, Vandenhaute J, Letesson JJ, De Bolle X: Plasticity of a transcriptional regulation network among alpha-proteobacteria is supported by the identification of CtrA targets in Brucella abortus. Mol Microbiol. 2002, 43: 945-960. 10.1046/j.1365-2958.2002.02777.xView ArticlePubMedGoogle Scholar
- Robertson GT, Reisenauer A, Wright R, Jensen RB, Jensen A, Shapiro L, Roop RM: The Brucella abortus CcrM DNA methyltransferase is essential for viability, and its overexpression attenuates intracellular replication in murine macrophages. J Bacteriol. 2000, 182: 3482-3489. 10.1128/JB.182.12.3482-3489.2000PubMed CentralView ArticlePubMedGoogle Scholar
- Miller TR, Belas R: Motility is involved in Silicibacter sp. TM1040 interaction with dinoflagellates. Environ Microbiol. 2006, 8: 1648-1659. 10.1111/j.1462-2920.2006.01071.xView ArticlePubMedGoogle Scholar
- Kahng LS, Shapiro L: The CcrM DNA methyltransferase of Agrobacterium tumefaciens is essential, and its activity is cell cycle regulated. J Bacteriol. 2001, 183: 3065-3075. 10.1128/JB.183.10.3065-3075.2001PubMed CentralView ArticlePubMedGoogle Scholar
- Gupta RS, Mok A: Phylogenomics and signature proteins for the alpha proteobacteria and its main groups. BMC Microbiol. 2007, 7: 106- 10.1186/1471-2180-7-106PubMed CentralView ArticlePubMedGoogle Scholar
- Pellegrini M, Marcotte EM, Thompson MJ, Eisenberg D, Yeates TO: Assigning protein functions by comparative genome analysis: protein phylogenetic profiles. Proc Natl Acad Sci USA. 1999, 96: 4285-4288. 10.1073/pnas.96.8.4285PubMed CentralView ArticlePubMedGoogle Scholar
- Laub MT, McAdams HH, Feldblyum T, Fraser CM, Shapiro L: Global analysis of the genetic network controlling a bacterial cell cycle. Science. 2000, 290: 2144-2148. 10.1126/science.290.5499.2144View ArticlePubMedGoogle Scholar
- Roth FP, Hughes JD, Estep PW, Church GM: Finding DNA regulatory motifs within unaligned noncoding sequences clustered by whole-genome mRNA quantitation. Nat Biotechnol. 1998, 16: 939-945. 10.1038/nbt1098-939View ArticlePubMedGoogle Scholar
- Brassinga AK, Siam R, McSween W, Winkler H, Wood D, Marczynski GT: Conserved response regulator CtrA and IHF binding sites in the alpha-proteobacteria Caulobacter crescentus and Rickettsia prowazekii chromosomal replication origins. J Bacteriol. 2002, 184: 5789-5799. 10.1128/JB.184.20.5789-5799.2002PubMed CentralView ArticlePubMedGoogle Scholar
- Stephens C, Reisenauer A, Wright R, Shapiro L: A cell cycle-regulated bacterial DNA methyltransferase is essential for viability. Proc Natl Acad Sci USA. 1996, 93: 1210-1214. 10.1073/pnas.93.3.1210PubMed CentralView ArticlePubMedGoogle Scholar
- Wright R, Stephens C, Shapiro L: The CcrM DNA methyltransferase is widespread in the alpha subdivision of proteobacteria, and its essential functions are conserved in Rhizobium meliloti and Caulobacter crescentus. J Bacteriol. 1997, 179: 5869-5877.PubMed CentralPubMedGoogle Scholar
- Hansen FG, Christensen BB, Atlung T: Sequence characteristics required for cooperative binding and efficient in vivo titration of the replication initiator protein DnaA in E. coli. J Mol Biol. 2007, 367: 942-952. 10.1016/j.jmb.2007.01.056View ArticlePubMedGoogle Scholar
- Ishikawa S, Ogura Y, Yoshimura M, Okumura H, Cho E, Kawai Y, Kurokawa K, Oshima T, Ogasawara N: Distribution of stable DnaA-binding sites on the Bacillus subtilis genome detected using a modified ChIP-chip method. DNA Res. 2007, 14: 155-168. 10.1093/dnares/dsm017PubMed CentralView ArticlePubMedGoogle Scholar
- Collier J, Shapiro L: Spatial complexity and control of a bacterial cell cycle. Curr Opin Biotechnol. 2007, 18: 333-340. 10.1016/j.copbio.2007.07.007PubMed CentralView ArticlePubMedGoogle Scholar
- Barnett MJ, Hung DY, Reisenauer A, Shapiro L, Long SR: A homolog of the CtrA cell cycle regulator is present and essential in Sinorhizobium meliloti. J Bacteriol. 2001, 183: 3204-3210. 10.1128/JB.183.10.3204-3210.2001PubMed CentralView ArticlePubMedGoogle Scholar
- McGrath PT, Lee H, Zhang L, Iniesta AA, Hottes AK, Tan MH, Hillson NJ, Hu P, Shapiro L, McAdams HH: High-throughput identification of transcription start sites, conserved promoter motifs and predicted regulons. Nat Biotechnol. 2007, 25: 584-592. 10.1038/nbt1294View ArticlePubMedGoogle Scholar
- Sibley CD, MacLellan SR, Finan T: The Sinorhizobium meliloti chromosomal origin of replication. Microbiology. 2006, 152: 443-455. 10.1099/mic.0.28455-0View ArticlePubMedGoogle Scholar
- Ohta N, Newton A: The core dimerization domains of histidine kinases contain recognition specificity for the cognate response regulator. J Bacteriol. 2003, 185: 4424-4431. 10.1128/JB.185.15.4424-4431.2003PubMed CentralView ArticlePubMedGoogle Scholar
- Reisinger SJ, Huntwork S, Viollier PH, Ryan KR: DivL performs critical cell cycle functions in Caulobacter crescentus independent of kinase activity. J Bacteriol. 2007, 189: 8308-8320. 10.1128/JB.00868-07PubMed CentralView ArticlePubMedGoogle Scholar
- Gupta RS: Protein signatures distinctive of alpha proteobacteria and its subgroups and a model for alpha-proteobacterial evolution. Crit Rev Microbiol. 2005, 31: 101-135. 10.1080/10408410590922393View ArticlePubMedGoogle Scholar
- Sallstrom B, Andersson SG: Genome reduction in the alpha-Proteobacteria. Curr Opin Microbiol. 2005, 8: 579-585. 10.1016/j.mib.2005.08.002View ArticlePubMedGoogle Scholar
- Cole JR, Wang Q, Cardenas E, Fish J, Chai B, Farris RJ, Kulam-Syed-Mohideen AS, McGarrell DM, Marsh T, Garrity GM, Tiedje JM: The Ribosomal Database Project: improved alignments and new tools for rRNA analysis. Nucleic Acids Res. 2009, 37: D141-145. 10.1093/nar/gkn879PubMed CentralView ArticlePubMedGoogle Scholar
- Edgar RC: MUSCLE: a multiple sequence alignment method with reduced time and space complexity. BMC Bioinformatics. 2004, 5: 113- 10.1186/1471-2105-5-113PubMed CentralView ArticlePubMedGoogle Scholar
- Tamura K, Dudley J, Nei M, Kumar S: MEGA4: Molecular Evolutionary Genetics Analysis (MEGA) software version 4.0. Mol Biol Evol. 2007, 24: 1596-1599. 10.1093/molbev/msm092View ArticlePubMedGoogle Scholar
- Schneider TD, Stormo GD, Gold L, Ehrenfeucht A: Information content of binding sites on nucleotide sequences. J Mol Biol. 1986, 188: 415-431. 10.1016/0022-2836(86)90165-8View ArticlePubMedGoogle Scholar
- Liu XS, Brutlag DL, Liu JS: An algorithm for finding protein-DNA binding sites with applications to chromatin-immunoprecipitation microarray experiments. Nat Biotechnol. 2002, 20: 835-839.View ArticlePubMedGoogle Scholar
- Tatusov RL, Natale DA, Garkavtsev IV, Tatusova TA, Shankavaram UT, Rao BS, Kiryutin B, Galperin MY, Fedorova ND, Koonin EV: The COG database: new developments in phylogenetic classification of proteins from complete genomes. Nucleic Acids Res. 2001, 29: 22-28. 10.1093/nar/29.1.22PubMed CentralView ArticlePubMedGoogle Scholar
- Whelan S, Lio P, Goldman N: Molecular phylogenetics: state-of-the-art methods for looking into the past. Trends Genet. 2001, 17: 262-272. 10.1016/S0168-9525(01)02272-7View ArticlePubMedGoogle Scholar