It is more than 50 years since protein histidine phosphorylation was first discovered in 1962 by Boyer and co-workers; however, histidine kinases are still much less well recognized than the serine/threonine and tyrosine kinases. The best-known histidine kinases are the two-component signalling kinases that occur in bacteria, fungi and plants. The mechanisms and functions of these kinases, their cognate response regulators and associated phosphorelay proteins are becoming increasingly well understood. When genomes of higher eukaryotes began to be sequenced, it did not appear that they contained two-component histidine kinase system homologues, apart from a couple of related mitochondrial enzymes that were later shown not to function as histidine kinases. However, as a result of the burgeoning sequencing of genomes from a wide variety of eukaryotic organisms, it is clear that there are proteins that correspond to components of the two-component histidine kinase systems in higher eukaryotes and that operational two-component kinase systems are likely to occur in these organisms. There is unequivocal direct evidence that protein histidine phosphorylation does occur in mammals. So far, only nucleoside diphosphate kinases have been shown to be involved in protein histidine phosphorylation, but their mechanisms of action are not well understood. It is clear that other, yet to be identified, histidine kinases also exist in mammals and that protein histidine phosphorylation may play important roles in higher eukaryotes.
Protein histidine phosphorylation was discovered to occur in protein extracts from [32P]Pi-labelled bovine mitochondria by Boyer et al. in 1962 . In spite of the fact that this discovery was made more than 50 years ago, until relatively recently, the biological roles of protein histidine phosphorylation and the associated histidine kinases remained obscure. The biological signalling role of protein histidine phosphorylation in bacteria, fungi and plants is now widely recognized and understood; however, this is not the case in animals. The slow progress of research in the field of histidine kinases largely stems from the technical difficulties associated with the analysis of protein histidine phosphorylation that have been reviewed extensively elsewhere [2–4]. One of these difficulties is that phosphohistidine is labile, especially under acidic conditions, whereas the phosphohydroxyamino acids are not. Unlike the phosphohydroxyamino acids that contain a phosphoester (P–O) bond, phosphohistidine contains a phosphoramidate (P–N) bond. It is likely that the lability of phosphohistidine stems from the propensity of the non-phosphorylated imidazole nitrogen to become protonated, thus making the histidine a good leaving group . The P–N bond, however, is also more thermodynamically unstable than the P–O bond with ΔG° values of hydrolysis of −12 to −14 kcal·mol−1 (1 kcal=4.184 kJ) for phosphohistidine  compared with −6.5 to −9.5 for phosphohydroxyamino acids . It is this thermodynamic propensity of phosphohistidine to transfer its phosphoryl group to other molecules, including amino acids in proteins, that underpins the mechanism of action of a large class of histidine kinases called two-component histidine kinases.
Two-component histidine kinases
Two-component histidine kinases are generally receptor-like proteins that span the bacterial cell membrane and have a sensor domain on the extracellular side of the membrane (Figure 1A). A well-characterized exception to this generalized structure of the two-component histidine kinase is the histidine kinase involved in bacterial chemotaxis, CheA (Figure 1A). CheA does not have a sensor domain of its own and is entirely cytoplasmic. Instead, it interacts with the cytoplasmic domains of a number of chemoreceptors. In the general two-component histidine kinases, a stimulus is detected by the sensor domain, which then induces conformational changes in the intracellular part of the molecule; this may induce dimerization of the histidine kinase molecules or they may be pre-dimerized. From a functional perspective, the most important result of the conformational changes is the activation of the histidine kinase domain (CA), which then phosphorylates a highly conserved histidine residue in the dimerization (Dhp) domain. This may occur either by a cis mechanism where the CA domain phosphorylates the histidine residue in its own DHp domain, or by a trans mechanism in which the CA domain of one molecule phosphorylates the histidine in its dimer partner's Dhp domain (Figure 1A). Again, CheA is different and phosphorylates a histidine residue in a histidine phosphotransfer (Hpt) domain that is analogous to the phosphotransfer domains and proteins in the more complex phosphorelay systems (Figure 1B).
Two-component histidine kinases and more complex phosphorelay systems
The signalling process then continues with the transfer of the phosphoryl group from the phosphohistidine of the histidine kinase to an aspartate residue in the receiver domain of a response regulator protein. This induces a conformational change in the effector domain, allowing DNA binding and activation, if the response regulator protein is a transcription factor, of which approximately 63% are . In other response regulators, their own enzymic activity is activated, e.g. CheB which has methyltransferase activity , or they bind to other enzymes to modulate their activity, e.g. Ssk1p that binds to and activates MAPKKK (mitogen-activated protein kinase kinase kinase) in yeast [8,9].
In the case of the phosphorelay systems, the initial phosphoryl transfer from the phosphohistidine of the kinase is to an aspartate residue on a domain that is analogous to the receiver domain of the response regulator that is part of the kinase molecule (e.g. Sln1p, ArcB) or is on a separate protein (e.g. Spo0F) (Figure 1B). There follows another phosphotransfer, this time from the phosphoaspartate to a histidine residue on an Hpt domain that is again part of the histidine kinase molecule (as with the ArcB system) or on a separate protein (e.g. Spo0B, Ypd1p). The final phosphotransfer then occurs to the aspartate residue on the response regulator. It is thought that these phosphorelay systems provide more opportunities for more sophisticated regulation and a higher discrimination than the simple two-component systems . In eukaryotes, it is the phosphorelay systems that predominate and where these systems directly control transcription, such as in Arabidopsis thaliana, the response regulators are constitutively found in the nucleus [11,12]. In these systems, the Hpt proteins shuttle between the sensor histidine kinases that contain receiver domains in the plasma membrane and the nucleus, and it is thought that the higher stability of phosphohistidine compared with phosphoaspartate may be a factor in employing an Hpt protein as a shuttle rather than the response regulator or a receiver domain-containing protein.
The phosphotransfer reaction between the phosphohistidine of the Dhp domain of the histidine kinase or that of a Hpt domain to the aspartate residue of a receiver domain appears to be catalysed by residues in the receiver domain and not the Dhp or Hpt domains [13,14]. Indeed, Lukat et al.  showed that response regulators could catalyse their own phosphorylation using small-molecule phosphoryl donors such as phosphoramidate and acetyl phosphate. Interestingly, some of the histidine kinases appear to be able to act as phosphatases that catalyse the dephosphorylation of phosphoaspartate in the receiver domain [16,17], although the biological relevance of this activity under normal cellular conditions has been questioned . This reaction is not the simple reverse of the phosphorylation reaction and the conserved histidine residue of the kinase appears to act as a general base catalyst with respect to nucleophilic attack on the phosphoryl group by water .
Higher eukaryotes: possible two-component histidine kinases
There is some evidence of the presence of two-component histidine kinase-like proteins in higher eukaryotes on the basis of sequence comparisons. In mammals, BCKDHK (branched-chain α-ketoacid dehydrogenase kinase) and PDHK (pyruvate dehydrogenase kinase) were both reported to contain prototypical two-component histidine kinase motifs [20,21]. These two enzymes also possess the Bergerat ATP-binding fold that is also characteristic of two-component histidine kinases . However, the homology of the sequence around the conserved histidine residue is not strong compared with that in the Dhp domains of two-component histidine kinases (Figure 2A). When the structures of both BCKDHK  and PDHK  were determined, it became apparent that this histidine residue is buried in the hydrophobic core of one of the domains of these enzymes and is not accessible for phosphorylation. It also became clear that this conserved histidine residue does not participate in ATP cleavage, and the mechanism of action of these two enzymes appears more closely related to that of the ATPases, DNA gyrase B and Hsp90 (heat-shock protein 90), which also have Bergerat ATP-binding folds .
Sequence-based evidence of the occurrence of two-component histidine kinases elements in higher eukaryotes
A cursory search of sequence databases, however, reveals a number of hypothetical proteins in a wide range of higher eukaryotes that contain two-component histidine kinase motifs where there is strong sequence homology around the conserved histidine residue compared with that of the Dhp and Hpt domains of a number of two-component histidine kinase systems (Figure 2A). Some of these proteins are also predicted to have Bergerat fold ATP-binding domains, whereas some were also predicted to have response regulator receiver domains. In the case of the hypothetical bee protein (GenBank® accession number XP_003494847.1), tentatively identified as a prolyl-tRNA synthetase, a predicted response regulator effector domain was also present in the form of a winged-helix DNA-binding domain. The human protein, p143, was identified in a lymphoma cDNA library and consists solely of the putative Hpt domain.
The likelihood that two-component histidine kinase-like systems may be present in higher eukaryotes from humans, macaque, frog, various insects and Hydra is borne out by the presence of a number of hypothetical proteins that have response regulator sequence motifs (Figure 2B). The fact that some of these hypothetical proteins contain both a predicted phosphotransfer domain and a DNA-binding effector domain suggests that these are transcription factors whose activity is modulated by phosphorylation of the conserved aspartate residue in the phosphotransfer domain by a histidine kinase. The hypothetical bee protein is interesting as it appears to contain all of the domains of a two-component histidine kinase system, except for a clearly defined sensor domain. At the N-terminus, where the sensor domain of two-component systems are usually found, are predicted domains associated with prolyl-tRNA synthetase, raising the possibility that these may regulate the protein's putative histidine kinase activity.
Higher eukaryotes: known histidine kinase activities
There have been a number of reviews that have focused on protein histidine phosphorylation in animals and particularly mammalian histidine kinases [25–27]. Much work has been performed on a histidine kinase(s) that phosphorylate histone H4 on its two histidine residues and whose activity is correlated with regeneration in rat liver following partial hepatectomy [28–31] and in rat liver precursor cells [30,31]. In addition, its activity has also been correlated with fetal growth and hepatocellular carcinoma in human liver [30,31]. A similar kinase has also been defected in porcine thymus . However, this enzyme(s) has never been identified or fully characterized.
The only mammalian histidine kinases that have been identified and characterized are two isoforms of NDPK (nucleoside diphosphate kinase): NDPK A (Nm23H1)  and NDPK B (Nm23H2) [34,35]. It has been reported that NDPKs were capable of phosphorylating ATP citrate lyase , succinate thiokinase  and CheA and EnvZ  on a histidine residue. However, in all cases, these proteins are capable of autophosphorylating that residue and some doubt has been cast as to whether NDPK directly phosphorylates them [39,40]. NDPK A has also been reported to phosphorylate annexin 1 on a histidine residue [41,42]. It is clear that NDPK B promotes the histidine phosphorylation of the potassium channel KCa3.1, which results in the activation of the channel [35,43]. In addition, NADPK B has been shown to phosphorylate His266 of the β-subunit (Gβ) of trimeric G-proteins in a process that leads to the non-receptor-mediated activation of Gs through the transfer of the phosphoryl group to GDP to form GTP which binds to and activates the α-subunit of Gs [34,44].
In part, the mechanism of action of NDPK as a histidine kinase seems clear. In its role in maintaining NDP/NTP levels, the phosphoryl group is transferred from NTP to NDP via phosphorylation of a conserved histidine residue in the active site of the enzyme (His118 in human NDPK A and B) . The question that is as yet unanswered is how this phosphoryl group is then transferred to histidine residues on other proteins. One possibility is that there is direct transfer of the phosphoryl group between the two histidine residues. This type of transfer occurs between the phosphorylated His717 of the Hpt domain of the histidine kinase ArcB and His8 of the histidine phosphatase SixA . Another possibility is that there is an initial phosphotransfer from the phosphohistidine residue of NDPK to an aspartate residue on an intermediary protein with a domain analogous to a receiver domain. As we have seen, there are likely to be a number of proteins with receiver domains in higher eukaryotes. There would then follow another phosphotransfer from the phosphoaspartate to the histidine residue on the target protein. This mechanism is similar to that which occurs in phosphorelay histidine kinases, and the phosphotransfer reactions would be catalysed by the receiver domain-containing protein. There is evidence that NDPK does require an additional protein component(s) in order to phosphorylate Gβ , although the identity of this protein(s) is not known.
Higher eukaryotes: the role of protein histidine phosphorylation
As we have seen in the two-component histidine kinases, the propensity of phosphohistidine to transfer its phosphoryl group is made use of to phosphorylate aspartate residues on proteins that contain a receiver domain, ultimately to an aspartate residue on a response regulator protein where the phosphorylation of the aspartate residue modulates the protein's biological activity. Thus, in these systems, protein histidine phosphorylation plays an intermediary role and does not in itself modulate protein activity. Although, as we have seen, similar systems may exist in higher eukaryotes, there is evidence in the case of the phosphorylation of KCa3.1 that histidine phosphorylation results in activation of the channel. This is suggestive that phosphorylation of histidine residues may directly trigger conformational changes. Owing to the lability of phosphohistidine, demonstrating that histidine phosphorylation can induce structural conformational changes directly is likely to be difficult. However, recent development of stable analogues of phosphohistidine [47–49] and the use of click chemistry to insert them in a site-specific way into peptides  is likely to prove very useful in this regard.
The system involving NDPK-mediated phosphorylation of Gβ and the subsequent phosphotransfer to GDP so as to result in the activation of Gsα is a more complex puzzle. As remarked on elsewhere , direct transfer between the phosphorylated histidine residue on Gβ and GDP bound to the nucleotide-binding site of Gsα in the GsαGβγ trimer is highly unlikely owing to the long distance between the GDP-binding site on Gsα and His266 on Gβ in the trimeric G-protein complex. Hippe and Wieland  considered the possibility that, in the trimeric G-protein complex, the proximity of Gsα and Gβ would allow the GDP to dissociate from Gsα and be phosphorylated by the phosphohistidine residue on Gβ and rebind to Gsα; however, how this reaction would be catalysed is not clear. In this system, the phosphohistidine residue in Gβ may act as an intermediate for the direct phosphorylation of GDP or perhaps in a storage role to phosphorylate His118 in NDPK B, which in turn phosphorylates GDP.
It is likely that there are two-component histidine kinase systems in higher eukaryotes, although not as prevalent as in bacteria and lower eukaryotes. However, there also appear to be other histidine kinases in higher eukaryotes that do not act in the same way as in the two-component systems and where, it appears, that histidine phosphorylation can trigger protein conformational changes in an analogous manner to serine/threonine or tyrosine phosphorylation. The investigation of histidine kinases in higher eukaryotes remains in its infancy, and we may discover more histidine kinases that are not directly related to the two-component enzymes. In considering the biological roles of these enzymes and the resultant histidine phosphorylations, we should not forget the fundamental chemical rationale of the operation of the two-component signalling systems. This is that histidine phosphorylation is an intermediate process with the final response resulting from the transfer of the phosphoryl group to a different amino acid.
Exploring Kinomes: Pseudokinases and Beyond: A Biochemical Society Focused Meeting held at Robinson College, Cambridge, U.K., 24–26 March 2013. Organized by Dario Alessi (Dundee, U.K.), Patrick Eyers (Liverpool, U.K.) and James Murphy (Walter and Eliza Hall Institute of Medical Research, Australia). Edited by Patrick Eyers and James Murphy.