Mol Biol Evol 2000 Dec;17(12):1956-70
Evolution of two-component signal transduction.
free full article online
Koretke KK, Lupas AN, Warren PV, Rosenberg M, Brown JR.
SmithKline Beecham Pharmaceuticals, Collegeville, Pennsylvania 19426-0989, USA.
Two-component signal transduction (TCST) systems are the principal means for coordinating responses to environmental changes in bacteria as well as some plants, fungi, protozoa, and archaea. These systems typically consist of a receptor histidine kinase, which reacts to an extracellular signal by phosphorylating a cytoplasmic response regulator, causing a change in cellular behavior. Although several model systems, including sporulation and chemotaxis, have been extensively studied, the evolutionary relationships between specific TCST systems are not well understood, and the ancestry of the signal transduction components is unclear. Phylogenetic trees of TCST components from 14 complete and 6 partial genomes, containing 183 histidine kinases and 220 response regulators, were constructed using distance methods. The trees showed extensive congruence in the positions of 11 recognizable phylogenetic clusters. Eukaryotic sequences were found almost exclusively in one cluster, which also showed the greatest extent of domain variability in its component proteins, and archaeal sequences mainly formed species-specific clusters. Three clusters in different parts of the kinase tree contained proteins with serine-phosphorylating activity. All kinases were found to be monophyletic with respect to other members of their superfamily, such as type II topoisomerases and Hsp90. Structural analysis further revealed significant similarity to the ATP-binding domain of eukaryotic protein kinases. TCST systems are of bacterial origin and radiated into archaea and eukaryotes by lateral gene transfer. Their components show extensive coevolution, suggesting that recombination has not been a major factor in their differentiation. Although histidine kinase activity is prevalent, serine kinases have evolved multiple times independently within this family, accompanied by a loss of the cognate response regulator(s). The structural and functional similarity between TCST kinases and eukaryotic protein kinases raises the possibility of a distant evolutionary relationship.
[...]
Two-component signal transduction (TCST) pathways form the central signaling machinery in bacteria (for reviews, see Stock, Ninfa, and Stock 1989 ; Parkinson and Kofoid 1992 ; Hoch and Silhavy 1995 ). In response to a stimulus, typically extracellular, the kinase component autophosphorylates at an internal histidine (the H-box). The high-energy phosphate group is then transferred to an aspartyl residue on the response regulator component (hence, "two-component" signal transduction), which modifies cellular behavior via an effector domain (fig. 1 ). Although most systems use a linear phosphorelay from one kinase to one response regulator, some use more complicated paths, involving a branching of the signal (chemotaxis) or multiple phosphorylated components (sporulation, adaptation to anaerobic conditions).
| Code Sample |
[img]http://mbe.oupjournals.org/content/vol17/issue12/images/medium/mbev-17-12-10-f01.gif[/img]
[b]Fig. 1.[/b]—Schematic diagram illustrating phosphorelay in two-component signal transduction systems. The signal (usually coming from the extracellular environment) is transduced to the "linker" region, which is located adjacent to the last transmembrane helix of the membrane-bound protein. The linker interacts with the kinase domain to induce autophosphorylation at a histidine (shown in red) located in the His-box domain or—in [b]CheA[/b]—in the Hpt domain. The phosphate is then transferred to an aspartate residue (shown in red) in a response regulator domain. The phosphorylated response regulator elicits the appropriate response within the cell (typically via the DNA-binding activity of an effector domain). In some systems, an Hpt domain serves as a regulatory phosphate sink for the phosphorylated response regulator
|
In bacteria, TCST systems mediate adaptive responses to a broad range of environmental stimuli. These include citrate uptake and catabolism (Cit), aerobic respiration (Arc), osmoregulation (EnvZ/OmpR), stress-induced sporulation (Kin/Spo), N-acetylmuramoyl-l-alanine amidase biosynthesis (Lyt), nitrate and nitrite metabolism (Nar), nitrogen regulation (Ntr), phosphate regulation (Pho), host recognition for pathogen invasion (Vir), and chemotaxis (Che) (Stock, Ninfa, and Stock 1989 ; Parkinson and Kofoid 1992 ; Hoch and Silhavy 1995 ). TCST systems also exist in certain nonanimal eukaryotes and in some Archaea (Alex and Simon 1994 ; Loomis, Shaulsky, and Wang 1997). In plants, they mediate photosensitivity (Schneider-Poetsch et al. 1991 ; Yeh and Lagarias 1998 ) and ethylene response (Chang et al. 1993 ); in fungi, they mediate osmoregulation (Maeda, Wurgler-Murphy, and Saito 1994 ; Krems, Charizanis, and Entian 1996 ; Posas et al. 1996 ) and hyphal development (Alex, Borkovich, and Simon 1996 ; Alex et al. 1998 ); and in the slime mold, they mediate Dictyostelium discoideum osmoregulation (Schuster et al. 1996 ) and fruiting body formation (Singleton et al. 1998 ). A thorough cataloguing of protein kinases found in the genome of the nematode Caenorhabditis elegans failed to find any true orthologs to prokaryotic histidine kinases, which suggests that TCST systems do not occur in metazoans (Plowman et al. 1999 ).
[...]
An exception to the phosphorelay described here is found in the chemotaxis kinase CheA, where the H-box has lost the catalytic histidine and is used exclusively for dimerization, while an Hpt domain that may have originally served a regulatory function is now used for phosphorelay to CheY and CheB (Bilwes et al. 1999 ; Dutta, Qin, and Inouye 1999 ). Further variability in His-acceptor domains is seen in the Spo0B protein, whose H-box domain forms a dimeric four-helix bundle similar to canonical H-boxes but of opposite handedness (Varughese 1998 ).
In addition to the three phosphorelay domains described above, a fourth domain conserved broadly in TCST systems has recently been recognized (Park and Inouye 1997 ; Aravind and Ponting 1999 ). Termed the "linker" region (or HAMP domain), it is typically found at the C-terminal end of the last transmembrane segment in many histidine kinases, chemoreceptors, bacterial nucleotidyl cyclases, and phosphatases, and mutations show that it plays a critical role in signal transduction. Some proteins contain multiple copies in tandem, suggesting that it represents an autonomously folding unit. Its ability to regulate kinase activity in trans (e.g., between chemoreceptors and CheA) and the variable nature of the segments connecting it to the H-box suggest that it acts through direct interaction with the kinase domain rather than through propagation of conformational change along the polypeptide chain. Thus, four protein domains appear to be typically involved in the signal transduction pathway from the extracellular sensor domain to the cytoplasmic effector (fig. 1 ).
TCST systems represent one of the most studied and best understood areas of bacterial physiology. Recently, they have also emerged as attractive targets for anti-microbial drug development (Barrett et al. 1998 ; Lange et al. 1999 ; Throup et al. 2000). Here we report the results of a detailed phylogenetic and structural study of genomic TCST sequences, undertaken to explore the origin and evolution of TCST systems.
[...]
Several interesting observations follow from the phylogenetic clusters presented here: No cluster contained proteins from both Archaea and eukaryotes, although a specific evolutionary relationship has long been postulated between these two groups (reviewed in Brown and Doolittle 1997 ). Bacterial phylogeny similarly did not correlate well with the observed clustering. Despite the considerable sizes of some clusters, none contained representatives from each bacterial species, and no bacterium had a representative in all of the clusters. Among bacteria, no cluster predominated across all species: Pho contained nearly a third of all TCST proteins detected in E. coli, B. subtilis, and Synechocystis and two thirds of those in M. tuberculosis, but none from spirochetes. In the latter, Che which is missing from A. aoelicus and M. jannaschii, was predominant (even though both organisms encode flagellar proteins and are motile). [a little garbled, I think this means these particular guys don't have Che proteins, e.g.:
"Among the Archaea, Methanobacterium thermoautotrophicum and Archaeoglobus fulgidus contained the largest numbers of histidine kinases (16 and 14, respectively) and response regulators (10 and 11, respectively), although they still contained fewer than free-living bacteria. Pyrococcus horikoshii had only a single histidine kinase and two response regulators (corresponding to the chemotaxis proteins CheA, CheY, and CheB), while Methanococcus jannaschii, Aeropyrum pernix, and Thermoplasma acidophilum had none. "]
[...]
Origin of the Histidine Kinase Fold
Response regulators are not recognizably related to other known protein families beyond a general structural similarity to P-loop NTPases (such as Ras) (Lukat et al. 1991 ; Stock et al. 1993 ), which has hitherto not been considered sufficient to infer homology. Histidine kinases, however, are clearly related to Hsp90, MutL, and type II topoisomerases in the ATP-binding domain (Tanaka et al. 1998 ; Bilwes et al. 1999 ; Dutta, Qin, and Inouye 1999 ). The conserved structural core consists of an antiparallel, four-stranded ß-sheet flanked on one side by three -helices, which surround the ATP-binding site. In addition, an ß element, which is in an equivalent structural position, is circularly permuted in the sequence of histidine kinases relative to other proteins with this fold. The structural similarity is mirrored in a set of conserved sequence motifs, primarily associated with nucleotide binding, which strongly imply descent from a common ancestor (fig. 4A ). Phylogenetic analysis of the sequence data by distance methods indicates that all kinases in this superfamily arose from a single ancestral protokinase (fig. 4B ). Because of the low branch point of PDKs, it is unclear whether this ancestor had Ser- or His-phosphorylating activity.
Histidine Kinases and Response Regulators Have Coevolved
In this study, we analyzed the phylogenetic relationships between the TCST systems from 14 complete and 6 partial genomes. Their components represent highly evolved multigene families, and the pattern and process of proliferation of such interacting yet structurally unrelated proteins is an important problem in evolutionary biology. A priori, two competing models for the evolution of novel TCST systems appear plausible: the recruitment model and the coevolution model. The recruitment model suggests that novel TCST systems evolve through gene duplication of one component, which then co-opts components from heterologous systems to yield a new specificity. This model is supported by the structural similarity of response regulators, in which only a few residues are sufficient to determine specificity, and by the observed crosstalk of TCST systems within an organism. From the phylogenetic perspective, this mechanism would result in an incongruent clustering of cognate histidine kinases and response regulators. The coevolution model suggests that novel TCST systems evolve by global duplication of all their components and subsequent differentiation. This model is supported by the fact that many TCST systems are concurrent on the chromosome. Phylogenetically, this mechanism would produce congruent gene trees for histidine kinases and response regulators.
Our results support the coevolution model. Despite the large number of proteins considered, which limited the resolution of the analysis by lowering the ratio of aligned residues to OTUs, the trees obtained for the histidine kinase and response regulator domains showed congruent clustering (fig. 2 ). Although the precise evolutionary relationships between clusters were not supported by strong bootstrap values, the overall topology of the NJ histidine kinase tree was verified by heuristic search results for the minimal-length MP tree. No such support was obtained for the response regulator tree, which had a lower resolution, but its validity was verified by the occurrence of two superclusters (Arf/Cit/CheY/Lyt and Nar/Mth) that were also found in the histidine kinase tree. The clusters themselves were statistically much more robust, with over half receiving bootstrap support of >50% in the distance-based analysis. As required by the coevolution model, pairs of histidine kinases and response regulators that are known to interact were overwhelmingly found in cognate clusters, as were 89% of histidine kinase and response regulator pairs that are linked on the chromosome (fig. 3 ). Coevolution has previously been proposed for eukaryotic TCST proteins, as well as for two hybrid kinases of E. coli (BarA and RcsC) (Pao and Saier 1997 ). Hybrid kinases, however, also provide evidence for a recruitment mechanism at work. For example, four of the five hybrid kinases of E. coli (ArcB, TorS, RcsC, and EvgA) are known to signal through response regulators found in noncognate clusters, and the fifth one, BarA, is thought to do so as well (via OmpR, found in the Pho cluster) (Hoch and Silhavy 1995 ). Recruitment is also observed in the chemotaxis and sporulation systems. Nevertheless, coevolution appears to be the strongly predominant mechanism by which novel TCST systems arise. Similar patterns of molecular coevolution have been observed in other interacting proteins, such as neuropeptides and their receptors (Darlinson and Richter 1999 ) and chaperonin subunits (Archibald, Logsdon, and Doolittle 1999 ).
Coevolution is not limited to TCST proteins, but also extends to the domains forming them. Both domain shuffling and domain swapping are comparatively rare events, and, with few exceptions, response regulators having homologous carboxyl-terminal domains were found within one cluster. These results agree well with a previous study performed on 49 bacterial response regulators by Pao and Saier (1995) , who found that classes of response regulators, defined by homology of their carboxyl-termini, generally formed distinct phylogenetic clusters. Pao and Saier's (1995) classes 1–5 correspond—in order—to our phylogenetic clusters Ntr, Pho, Nar, CheB, and Hybrid; classes 6 and 7 were phylogenetically heterogeneous in both studies.
[...]
Histidine Kinases and Eukaryotic Protein Kinases—Homology or Analogy?
A distance-based phylogenetic analysis of proteins containing a histidine kinase–like ATP-binding domain, which include type II topoisomerases, Hsp90, and MutL, indicated that all kinase domains with this fold are monophyletic and confirmed that the PDKs form an outgroup to the histidine kinase clade (fig. 4B ). Searches for more distantly related proteins surprisingly suggested a similarity to the small lobe of protein kinases, which is also involved in ATP binding. The structurally similar region covers virtually the entire conserved core of both folds but is circularly permuted in the protein kinase small lobe (fig. 5A ). The amino- and carboxyl-termini of the histidine kinase fold are in close proximity, however (a prerequisite of circular permutation), and circular permutation events have been documented in the evolution of many protein folds (see the SCOP database at http://scop.mrc-lmb.cam.ac.uk/scop/; Lo Conte et al. 2000 ), including the histidine kinase fold (Bilwes et al. 1999 ). Although the protein kinase small lobe is part of the so-called ATP-grasp fold (jointly with the peptide-binding large lobe), a recent structural analysis by Grishin (1999) has concluded that only the large lobe is homologous among the members of this fold, with the small lobe having been recruited among structurally similar but unrelated proteins. Our analysis supports an independent evolutionary origin of the small and large lobes of eukaryotic protein kinases.
Despite the fact that the structural similarity between histidine kinases and protein kinases is remote, the signaling pathways in which the two types of kinase operate present intriguing similarities. Both form homodimers, which phosphorylate in trans and generally contain an extracellular sensory domain, which binds signaling molecules asymmetrically at the subunit interface. They have a common mode of signal transduction, as shown by the phytochromes Phy1 and Phy2 of the moss Ceratodon purpureus, which are 90% identical in the chromophore-binding domain, yet signal through protein kinase and histidine kinase domains, respectively (fig. 5C ). Functional chimeras have also been constructed between the sensory domain of the Tar chemoreceptor and both protein and histidine kinases (Moe, Bollag, and Koshland 1989 ; Utsumi et al. 1989). Finally, as discovered recently, both kinase types use adaptors with an SH3-fold for interaction with other proteins (Bilwes et al. 1999 ). Similar parallels can be drawn between response regulators and the Ras family of G-proteins (Lukat et al. 1991 ). Not only are both encoded by large multigene families, linking different sensory inputs to specific effector outputs, but they are both activated by a high-energy phoshoanhydride bond. Their striking structural similarity, particularly in the active site, has previously been interpreted as evidence for homology rather than analogy (Artymiuk et al. 1990 ). Although each of these similarities may have arisen by convergent evolution, the combination of structural and functional parallels that can be drawn throughout signaling pathways in bacteria and eukaryotes suggest that a prototypical signal transduction pathway may already have existed in the last common ancestor and that this pathway utilized protein phosphorylation. If so, yet a third group of kinases (possibly showing similarly profound structural changes) remains to be discovered in Archaea, where bacterial- and eukaryotic-type kinases are rare and clearly acquired by horizontal transfer.
|
; (ii) propagation of a deformation (kink) along the helical path; (iii) propagation of a reversal of the helical sense along the cell body; and (iv) irregular flexing and twitching, which is functionally equivalent to standard bacterial tumbling. Here, we analyse in detail only the first case (from which all the rest are derived), including switching of the helical sense.