Eukaryote-Like Serine/Threonine Kinases and Phosphatases in Bacteria

Abstract

Summary: Genomic studies have revealed the presence of Ser/Thr kinases and phosphatases in many bacterial species, although their physiological roles have largely been unclear. Here we review bacterial Ser/Thr kinases (eSTKs) that show homology in their catalytic domains to eukaryotic Ser/Thr kinases and their partner phosphatases (eSTPs) that are homologous to eukaryotic phosphatases. We first discuss insights into the enzymatic mechanism of eSTK activation derived from structural studies on both the ligand-binding and catalytic domains. We then turn our attention to the identified substrates of eSTKs and eSTPs for a number of species and to the implications of these findings for understanding their physiological roles in these organisms.

INTRODUCTION

The survival of an organism relies on its capacity to quickly respond and adapt to a constantly changing environment. Underlying this adaptive potential is the ability of cells to sense and transduce external and internal signals. Protein kinases, together with their cognate phosphatases, play a central role in signal transduction by catalyzing reversible protein phosphorylation. Upon sensing external stimuli, kinases undergo autophosphorylation and proceed to transphosphorylate substrate proteins. Phosphorylation on specific amino acid residues, most commonly serine (Ser), threonine (Thr), tyrosine (Tyr), histidine (His), and aspartate (Asp), can control the activity of target proteins, either directly, for example, by inducing conformational changes in the active site, or indirectly, by regulating protein-protein interactions. While most bacterial kinases have been thought to target only His/Asp residues, increasing attention is being paid to the Ser/Thr kinases and their partner phosphatases. This review focuses primarily on bacterial kinases that show homology in their catalytic domains to eukaryotic Ser/Thr kinases (eSTKs) and, to a lesser extent, on their partner eukaryote-like phosphatases (eSTPs).

Ser/Thr PROTEIN KINASES

Protein phosphorylation on Ser, Thr, and Tyr residues was first described for eukaryotes (36, 90, 94, 176). After these initial studies, the functional relevance of these modifications was soon realized, and it quickly became clear that protein phosphorylation is a key mechanism to regulate protein activity and, consequently, to control cellular functions. However, Ser, Thr, and Tyr phosphorylation was assumed for a long time to occur exclusively in eukaryotes (8). Eukaryotic Ser/Thr and Tyr kinases are grouped together in the eukaryotic protein kinase superfamily based on sequence homology between their kinase domains (63). These domains are typically organized into 12 subdomains that fold in a characteristic two-lobed catalytic core structure, with the catalytic active site lying in a deep cleft formed between the two lobes (63, 79) (Fig. 1A). The smaller, N-terminal lobe is involved primarily in binding and orienting the phospho-donor ATP molecule, whereas the larger, C-terminal lobe binds the protein substrate and initiates the transfer of the phosphate group. The structural conservation of the catalytic domain between different kinases is remarkable and is maintained across kingdoms (Fig. 1B) (also see “Structural and functional studies of bacterial eSTKs”). While subdomains vary in size, there is little sequence homology between members of the superfamily. However, the kinase catalytic domain can be defined by the presence of specific conserved motifs and 12 nearly invariant residues which are directly or indirectly involved in positioning the phosphate donor ATP molecule and the protein substrate for catalysis (63) (Fig. 1C and D).

FIG. 1.

Structure of the Ser/Thr kinase catalytic domain. (A) Crystal structure of the mouse PKA catalytic domain in complex with an ATP molecule and an inhibitor peptide (Protein Data Bank [PDB] accession number 1ATP). The PKA N-terminal lobe is shown in gray, and the C-terminal lobe is shown in blue. ATP is represented as sticks, with two manganese ions shown as spheres, and the inhibitor peptide is shown as a red line. (B) Superimposition of tertiary structures of PKA and the M. tuberculosis eSTK PknB (PDB accession number 1MRU). PKA is shown in blue, and PknB is shown in yellow. (C) The regulatory elements that comprise the catalytic cleft formed between the N- and C-terminal lobes of the Ser/Thr kinase catalytic domain are indicated in the structure of PKA as follows: green, P loop; yellow, catalytic loop with the catalytic Asp residue; magenta, magnesium-binding loop; orange, activation loop with the phosphorylated Thr residue; and cyan, P+1 loop. ATP is represented as sticks, with two manganese ions shown as spheres, and the inhibitor peptide is shown as a red line. (D) Primary sequence alignment between the PKA (residues 33 to 283) and PknB (residues 1 to 266) catalytic domains. The N- and C-terminal lobes of PKA are shown in gray and blue, respectively. Conserved motifs are shown in boxes, and the invariant residues are depicted in black. Other important residues are highlighted and/or shown in bold. Red and orange asterisks indicate the catalytic Asp and phosphorylated Thr residues, respectively.

Kinases are molecular switches that exist in either an “off,” inactive state or an “on,” active state (67). The transition to the activated state is tightly controlled by diverse mechanisms, including the binding of allosteric effectors and the subcellular localization. Both the αC helix in the N-terminal lobe and the activation loop in the C-terminal lobe undergo extensive conformational changes essential for this transition (Fig. 1C and D). When the αC helix is directed toward the active site, it allows multiple interactions between the two lobes that are essential for the kinase activity. The activation loop is part of the most important regulatory element of kinase activity, the so-called activation segment, which is defined as the region between and including the conserved DFG and APE motifs (122) (Fig. 1C and D). This segment includes the magnesium-binding loop, the activation loop, and the P+1 loop. The activation loop is involved in determining substrate specificity and is the most variable region of the activation segment. Many kinases are activated by phosphorylation on at least one Ser/Thr (Tyr for tyrosine kinases) residue in the activation loop, by either autophosphorylation or transphosphorylation by another kinase. This modification promotes several interactions that stabilize the loop in a conformation that allows substrate binding and catalysis (67, 122). These include the critical interaction between the phosphorylated Ser/Thr residue in the activation loop and the conserved arginine that precedes the catalytic aspartate in the catalytic loop of the so-called arginine/aspartate (RD) kinases. The phosphorylated Ser/Thr residue can also interact with the αC helix, bringing the two lobes together in an active conformation. The activation loop is also a site of protein-protein interaction for activity modulators in many kinases. The P+1 loop is a critical point of contact between the kinase and its substrate and is a major determinant of the distinct substrate specificity between Ser/Thr and Tyr kinases. In the Ser/Thr kinases, the loop includes a conserved Ser or Thr residue that interacts with the catalytic loop. Also of relevance is the glycine-rich consensus motif, called the P loop, since it covers both the β- and γ-phosphates and plays an important role in both phosphoryl transfer and ATP/ADP exchange during the catalytic cycle (Fig. 1C and D). All of these conformational changes bring the ATP γ-phosphate, the kinase catalytic Asp residue, and the substrate phospho-acceptor residue together, allowing the transfer of γ-phosphate from ATP to the phospho-acceptor Ser or Thr residue (Tyr in tyrosine kinases) in the substrate protein.

The two main groups of the superfamily, the Ser/Thr kinases and the Tyr kinases, can be subdivided further into smaller families composed of enzymes that show similar substrate specificities and modes of regulation (63). This review focuses on bacterial kinases with catalytic domains that share structural and functional homology with eukaryotic Ser/Thr kinases. These kinases are referred to herein as eSTKs to distinguish them from other prokaryotic enzymes that can also phosphorylate Ser/Thr residues (see the next section). Although outside the scope of this review, phosphorylation in bacteria also occurs at Tyr residues, and it should be noted that most bacterial tyrosine kinases belong to the BY kinase family and do not resemble eukaryotic enzymes (reviewed in references 57 and 58).

eSTKs in Bacteria

In prokaryotes, protein phosphorylation was assumed for a long time to occur only on histidine and aspartic acid residues involving two-component systems (TCS) (64, 172). The Escherichia coli tricarboxylic acid cycle enzyme isocitrate dehydrogenase (IDH) was the first example of Ser/Thr phosphorylation described for bacteria (53). However, IDH was phosphorylated on a Ser by a bifunctional kinase/phosphatase which lacked sequence homology with eSTKs (86). The first characterized eSTK in bacteria was Myxococcus xanthus Pkn1 (114). When Pkn1 was overexpressed in E. coli, it underwent autophosphorylation on Ser and Thr residues. Although the gene was not essential, a pkn1 deletion strain showed premature differentiation, indicating that Pkn1 is required for normal M. xanthus development. The identification of a transmembrane eSTK, M. xanthus Pkn2, suggested that membrane receptor kinases, such as the transforming growth factor beta (TGF-β) receptor kinases that play a central role in eukaryotic cellular regulation, also exist in bacteria (183; see the next section). The weak phenotypes of pnk1 and pkn2 mutants suggested the presence of additional kinases with redundant functions. This was confirmed in subsequent studies that revealed a large number of eSTKs interacting in intricate signaling pathways involved in the control of the complex M. xanthus life cycle (118-120; also see “Physiological roles and targets of eSTKs in bacteria”).

The availability of prokaryotic genomic sequencing data led to the rapid identification of numerous bacterial eSTKs containing characteristic signature sequence motifs (78, 83, 89, 138, 165) (see Ser/Thr Protein Kinases). The number of eSTKs identified in prokaryotes has continued to increase since those initial efforts, and these enzymes are now considered ubiquitous. It has also become clear that prokaryotes show a remarkable degree of diversity in signal transduction pathways (51, 78). A comprehensive and regularly updated list of eSTKs and other prokaryotic signal transduction molecules can be found at http://www.ncbi.nlm.nih.gov/Complete_Genomes/SignalCensus.html (51). Variability exists at the level of the modular organization of prokaryotic eSTKs (83; also see “Structural and functional studies of bacterial eSTKs”) as well as in the catalytic kinase core domain itself, as exemplified by the E. coli YihE kinase (200). YihE belongs to the category of so-called “atypical” kinases, which share homology with the catalytic core of eSTKs but do not conserve all of the usual kinase motifs (152). It should also be emphasized that not all bacterial Ser/Thr kinases are eSTKs. In addition to the IDH kinase/phosphatase previously mentioned and the HPr kinase/phosphorylase (81), several other non-eSTK Ser/Thr kinases have been described, including Bacillus subtilis SpoIIAB, RsbT, and RsbW. SpoIIAB phosphorylates the anti-anti-sigma factor SpoIIAA on a Ser as part of a regulatory circuit that controls the activity of the sigma factor σ^F in sporulation (116). Although SpoIIAB is not an eSTK, its phosphatase partner SpoIIE is a eukaryote-like Ser/Thr phosphatase (40). RsbT and RsbW, which are members of the ATPase/kinase superfamily (41), phosphorylate their substrates on a Ser as part of a complex regulatory network that controls the activity of σ^B, the major sigma factor involved in the stress response (193).

Cross talk between eSTKs and TCS can also occur. The M. xanthus transcription factor Mrp is under the control of both the MrpA/MrpB TCS and the eSTKs Pkn8 and Pkn14 (119, 173; also see “Physiological roles and targets of eSTKs in bacteria. (i) Developmental processes”). Similarly, the CovR/CovS TCS and the eSTK Stk1 both regulate expression of the β-hemolysin/cytolysin, which is critical for survival of group B streptococci in the bloodstream and for resistance to oxidative stress (145). The Mycobacterium tuberculosis response regulator DosR of the DosS/DosR TCS, which controls the DosR regulon involved in hypoxia- and nitric oxide-induced dormancy, is phosphorylated on Thr198 and Thr205 by PknH, and in conjuction with phosphorylation on Asp54 by DosS, this cooperatively enhances DosR DNA affinity for target genes (25).

Structural and functional studies of bacterial eSTKs.

The first bacterial eSTK structure described was that of M. tuberculosis PknB (130, 195). The cocrystal of the PknB catalytic domain in complex with an ATP analog revealed that PknB is structurally very similar to the mouse cyclic AMP (cAMP)-dependent protein kinase (PKA) in the activated state (100, 130, 195) (Fig. 1B). The PknB catalytic domain exhibits the typical two-lobed structure, with the nucleotide analog tightly bound in the catalytic cleft between the two lobes and in contact with the regulatory structural elements (see Ser/Thr Protein Kinases and Fig. 1). Both the overall conformation of PknB and the conformation of the majority of the regulatory regions resemble those for the activated state of eukaryotic Ser/Thr kinases (67, 100). Nevertheless, two important exceptions were observed. First, the PknB αC helix is oriented away from the active site, in a position characteristic of the inactive state (67). Despite this orientation, the essential interaction between the conserved Lys40 residue (Lys72 in PKA) in the β3 strand and the conserved Glu59 residue (Glu91 in PKA) in the αC helix is still maintained, as well as the interaction with the α- and β-phosphates of the nucleotide (67) (Fig. 1C). Second, the PknB activation loop was found to be disordered, which is also characteristic of an inactive kinase. However, this might be explained by the phosphorylation state of the activation loop or the absence of the protein substrate, both of which are known to stabilize this region (67, 122).

Overexpression of the PknB kinase domain in E. coli resulted in the phosphorylation of four residues in the activation loop (Ser166, Ser169, Thr171, and Thr173) (195). Interestingly, PknB Thr171 corresponds to the phospho-receptor Thr197 residue in the PKA activation loop, which undergoes autophosphorylation during the activation of the kinase (100) (Fig. 1C). In addition, when the sequences of 22 PknB orthologs were mapped on the surface of the PknB structure, extensive conservation was found both in the ATP binding cleft, in residues adjacent to the ATP γ-phosphate binding site, and in the surface analogous to the PKA substrate-binding groove (195). This suggests a common activation mechanism shared by homologous eukaryotic and prokaryotic Ser/Thr kinases.

In both structures, PknB crystallized as a dimer, indicating interactions between the opposite or “back” sides of the N-terminal lobes of two catalytic domains (130, 195) (Fig. 2A). This dimerization interface was conserved among PknB orthologs (195) and was also found in the structure of M. tuberculosis PknE, despite the low level of sequence conservation between the two kinases in this region (54). Consistent with a common activation mechanism between eukaryotic and prokaryotic Ser/Thr kinases, the eukaryotic double-stranded RNA-activated protein kinase PKR also crystallized as a back-to-back dimer (30). Moreover, PKR was shown to be activated by a ligand-induced dimerization mechanism (for a review, see reference 28). Experimental data support a similar model of activation for prokaryotic eSTKs (Fig. 3A). The replacement of the M. tuberculosis PknD ligand-binding domain (see below) by rapamycin-induced dimerization domains resulted in activation of unphosphorylated kinase (59). The importance of dimerization in kinase activation was further documented by mutagenesis studies in which the replacement of conserved residues in the PknD back-to-back N-terminal lobe dimerization interface reduced autophosphorylation and altered substrate specificity (59). Similar observations were also reported for Pseudomonas aeruginosa PpkA (66).

FIG. 2.

Dimerization interfaces involved in bacterial eSTK activation. (A) Back-to-back dimer revealed in the crystal structure of the M. tuberculosis PknB kinase domain (PDB accession number 1MRU). Two PknB monomers interact through a dimerization interface located in the back sides of the N-terminal lobes (130, 195). (B) Asymmetric front-to-front dimer found in the cocrystal of a mutant PknB kinase domain in complex with an ATP competitive inhibitor (PDB accession number 3F69). This structure resembles an activation complex involving the contact between the αG helixes of two monomers (103). One of the monomers (blue) shows an ordered activation loop (red), characteristic of the active state, whereas the other monomer (green) shows a disordered activation loop (represented by a dashed line). (C) Dimerization of two M. tuberculosis PknG kinase monomers. The complete structure of PknG is shown (PDB accession number 2PZI), including the N-terminal rubredoxin domain (RD) and the C-terminal tetratricopeptide repeat domain (TPRD) that surround the kinase catalytic domain (KD). In contrast with the case for PknB, dimerization of PknG occurs through interaction between the TPRDs of two monomers (153).

FIG. 3.

Activation model for eSTKs. M. tuberculosis PknB was chosen for the purpose of illustrating the hypothesized eSTK activation pathways. (A) In the presence of a ligand, two or more PknB monomers bind to a single ligand molecule through their extracellular domains. This brings the intracellular catalytic domains closer, resulting in the formation of a symmetric back-to-back dimer and in the consequent activation of the kinases by autophosphorylation (see text for details). (B) An activated kinase can directly phosphorylate a downstream protein target or activate a soluble kinase through the formation of an asymmetric front-to-front dimer, which will then phosphorylate downstream targets as part of a signaling pathway.

While these observations clearly support a role for dimerization in the activation of eSTKs, the mechanism by which dimerization results in autophosphorylation remains unknown. It is unlikely that it occurs through intermolecular phosphorylation, since in a back-to-back dimer the active sites of the kinase domains are oriented away from each other (Fig. 2A and 3A). In addition, a heterodimer composed of an inactive PknD catalytic site mutant and a wild-type monomer still activated the wild-type catalytic domain (59). These results are consistent with allosteric activation of the kinase upon dimerization, but it is also possible that activation can occur through transphosphorylation by another monomeric or dimeric activated kinase (Fig. 3).

Ligand-promoted dimerization is unlikely to be the only activation mechanism for bacterial eSTKs. First, the phosphorylated PknD kinase domain is fully active even in the absence of dimerization (59). Second, this mechanism does not explain the activation of eSTKs that lack a ligand-binding domain. Consistently, the structure of a PknB mutant kinase cocrystallized with an inhibitor molecule revealed a second dimerization interface in which the two monomers bound to the inhibitor form an asymmetric front-to-front dimer (103) (Fig. 2B). This dimerization interface occurs through interactions between the αG helix and the ordered activation loop of one kinase domain and the αG helix of the second kinase domain. The conformations of the two proteins suggest that one monomer functions as an activator and the other functions as a substrate. In agreement, mutagenesis studies showed that the asymmetric dimer interface mimics a trans-autophosphorylation complex (103). This dimerization provides a new mechanism by which an allosterically activated kinase could phosphorylate and thereby activate other kinases that are not associated with a receptor domain (Fig. 3B).

The M. tuberculosis soluble kinase PknG has a unique modular organization, with the kinase domain sandwiched between an N-terminal rubredoxin domain and a C-terminal tetratricopeptide repeat domain (TPRD) (153) (Fig. 2C). While PknG crystallizes as a dimer, in contrast with PknB and PknD, dimerization does not occur through the kinase domain but via the TPRD (Fig. 2C). Also, the mechanism of PknG activation seems to differ from that of PknB and most kinases. First, the activation loop of PknG is not phosphorylated, but it is fully ordered and stabilized despite being in an open and extended conformation. Second, the conserved Arg immediately preceding the invariant catalytic Asp residue is absent in the PknG catalytic loop (see above and Ser/Thr Protein Kinases). Activation of PknG is controlled by its rubredoxin domain, which is often found in electron transfer proteins, suggesting that the redox status of the environment may regulate PknG (153).

The catalytic domains of many bacterial eSTKs are associated with an additional domain(s) (83). The diversity in this modular organization is well illustrated in organisms that express a large number of these regulatory proteins, such as in myxobacterial species and Streptomyces coelicolor (138, 140). While these extra domains can be enzymatic, they most commonly mediate ligand binding or protein-protein interactions. For example, virtually all Gram-positive bacteria contain at least one transmembrane eSTK composed of a cytoplasmic catalytic kinase domain linked by a transmembrane segment to an extracellular domain consisting of a variable number of PASTA (penicillin-binding protein and Ser/Thr kinase-associated) repeats (194). PASTA repeats were first observed in the crystal structure of Streptococcus pneumoniae PBP2X, where two approximately 70-residue-long repeats adopted a unique β₃α structural topology (35). However, the PASTA repeats of M. tuberculosis PknB and its Staphylococcus aureus homolog adopt a linear organization (12, 134) (Fig. 4A). The PBP2X structure includes a molecule of the β-lactam antibiotic cefuroxime interacting with one of the PASTA repeats. The portion of cefuroxime bound to the PASTA repeat is structurally analogous to unlinked peptidoglycan, suggesting that these repeats bind peptidoglycan (194). This hypothesis was confirmed by demonstrating that peptidoglycan bound the PASTA repeats of B. subtilis PrkC, providing the first evidence for a ligand of a bacterial receptor eSTK (160; also see “Physiological roles and targets of eSTKs in bacteria. (i) Developmental processes”). This observation, together with the structural studies of PknB proteins from M. tuberculosis and S. aureus, supports a ligand-induced dimerization model (see above and Fig. 3) in which the activation of these eSTKs results from dimerization of two kinase molecules that bind peptidoglycan through their extracellular PASTA repeats (12, 160). M. tuberculosis PknD is also a transmembrane eSTK, but its catalytic domain is linked to an extracellular domain consisting of a rigid β propeller with six blades symmetrically arranged around a central pore (56) (Fig. 4B). While the ligand(s) of PknD remains unknown, this motif is found in eukaryotic proteins with a wide variety of functions, and among bacteria, it appears to be present only in pathogenic mycobacteria.

FIG. 4.

Structures of M. tuberculosis PknB and PknG sensor domains. (A) The extracellular domain of M. tuberculosis PknB is composed of four PASTA repeats organized linearly (PDB accession number 2KUI) (12). (B) The highly symmetric six-bladed β propeller formed by the extracellular domain of M. tuberculosis PknG kinase (PDB accession number 1RWI) (56). The PknB catalytic domain is shown to represent the intracellular catalytic kinase domain.

Physiological roles and targets of eSTKs in bacteria.

Open reading frames encoding eSTKs are found in many sequenced microbial genomes (51). However, complete characterization of their essentiality and the identification of their specific substrates have been difficult to achieve, likely resulting from functional redundancy and/or substrate promiscuity. Table 1 lists the identified substrates of eSTKs, but it should be emphasized that the majority of these substrates were identified by phosphoproteomic approaches and/or in vitro kinase assays and site-directed mutagenesis and lack in vivo confirmation. In addition, only limited information exists concerning the effect of phosphorylation on the activity of substrates.

TABLE 1.

eSTKs in bacteria^a

Species	eSTK	Substrate	Function	Phospho-residue(s)	Methodology	Reference(s)
Bacillus subtilis	PrkC	AlsD	α-Acetolactate decarboxylase; central metabolism	S88	In vitro kinase assay, mass spectrometry	1a, 141
		CpgA	GTPase; peptidoglycan deposition	Thr(s)/Ser(s)	In vitro kinase assay, site-directed mutagenesis	1
		EF-G	Elongation factor; protein translation	Thr(s)	In vivo immunoprecipitation, in vitro kinase assay	50, 159, 160
		EF-Tu	Elongation factor; protein translation	T384	In vitro kinase assay, site-directed mutagenesis	1
		GlnA	Glutamine synthetase; central metabolism	S207, T26, T147, T286	In vitro kinase assay, mass spectrometry	141
		Icd	Central metabolism	T138, T14, T396	In vitro kinase assay, mass spectrometry	141
		HPr	Kinase, phosphotransferase system	S12	In vitro kinase assay, mass spectrometry	141
		YezB	Stress	Thr(s)	In vitro kinase assay, phospho-amino acid analysis, 2D electrophoresis	1
		YwjH	Transaldolase; central metabolism	S55, T26, T82, T125, T159, T184	In vitro kinase assay, mass spectrometry	141
Chlamydia trachomatis	Pkn1	IncG	Pathogenesis	NA	Bacterial two-hybrid assay, in vitro kinase assay	186
	PknD	Pkn1	eSTK	NA	Bacterial two-hybrid assay, immunoprecipitation, in vitro kinase assay	186
Corynebacterium glutamicum	PknA	FtsZ	Cell division	T108	In vitro kinase assay, 2D electrophoresis, mass spectrometry	157
		MurC	Cell wall synthesis	T51, T120, T133, T16, T362, T365	In vitro kinase assay, mass spectrometry, site-directed mutagenesis	48
		PknG	Soluble eSTK	T451, T787	In vitro kinase assay	49
		OdhI	Glutamate catabolism	T15	In vitro kinase assay, mass spectrometry	13, 49, 157
	PknB	FtsZ	Cell division	NA	In vitro kinase assay	157
		OdhI	Glutamate catabolism	T15	In vivo and in vitro kinase assays	13, 49, 157
	PknG	OdhI	Glutamate catabolism	T14	In vivo 2D electrophoresis, in vitro kinase assay, mass spectrometry, site-directed mutagenesis	49, 121, 157
	PknL	FtsZ	Cell division	T63, S353, T388	In vitro kinase assay, 2D electrophoresis, mass spectrometry	157
Mycobacterium tuberculosis	PknA	EmbR	Arabinan synthesis, cell wall	NA	In vitro kinase assay	162
		FadD	Mycolic acid biosynthesis	Thr(s)	In vitro kinase assay, phospho-amino acid analysis	109
		FabH	Mycolic acid pathway; cell wall biosynthesis	T45	In vitro kinase assay, mass spectrometry, site-directed mutagenesis	187
		FipA	Cell division under oxidative stress	T77	In vivo and in vitro kinase assays, mass spectrometry	175
		FtsZ	Cell division	T343	In vitro kinase assay, mass spectrometry	175, 179
		GlmU	Cell wall synthesis	Thr(s)	In vitro kinase assay	135
		GroEL1	Heat shock protein	NA	In vitro kinase assays, mass spectrometry	23
		KasA	Mycolic acid biosynthesis	Thr(s)	In vitro kinase assay, phospho-amino acid analysis	109
		KasB	Mycolic acid biosynthesis	Thr(s)	In vitro kinase assay, phospho-amino acid analysis	109
		MabA	Mycolic acid biosynthesis	T21, T114, T191	In vitro kinase assay, mass spectrometry, site-directed mutagenesis	188
		MurD	Cell division	NA	In vitro kinase assay	178
		PknB	eSTK	NA	In vitro kinase assay	75
		Rv1422	NA	T325	In vivo and in vitro kinase assays, mass spectrometry, site-directed mutagenesis	75
		Wag31	Cell division	T73	In vivo and in vitro kinase assays, mass spectrometry, site-directed mutagenesis	75
	PknB	EmbR	Arabinan synthesis, cell wall	NA	In vitro kinase assay	162
		FadD	Mycolic acid biosynthesis	NA	In vitro kinase assay	109
		GarA	Glycogen recycling, tricarboxylic acid cycle	T22	Phosphoproteomic kinase assays, site-directed mutagenesis	189
		GlmU	Cell wall synthesis	Thr(s)	In vitro kinase assay	135
		GroEL1	Heat shock protein	NA	In vitro kinase assays, mass spectrometry	23
		KasA	Mycolic acid biosynthesis	NA	In vitro kinase assay	109
		KasB	Mycolic acid biosynthesis	NA	In vitro kinase assay	109
		MabA	Mycolic acid biosynthesis	T21, T114, T191	In vitro kinase assay, mass spectrometry, site-directed mutagenesis	188
		PBPA	Cell wall synthesis	T437, T362	In vitro kinase assay, mass spectrometry, site-directed mutagenesis	32
		PknA	eSTK	NA	In vitro kinase assay	75
		RshA	SigH anti-sigma factor, oxidative stress	T94	In vivo and in vitro kinase assays, mass spectrometry, site-directed mutagenesis	136
		Rv0020c	FHA-containing protein	NA	In vitro kinase assay	60
		Rv1422	NA	T325	In vitro kinase assay, mass spectrometry, site-directed mutagenesis	75
		Rv1747	Putative ABC transporter	NA	In vitro kinase assay	60
		SigH	Alternate sigma factor; oxidative stress	T26, T106	In vivo and in vitro kinase assays, mass spectrometry, site-directed mutagenesis	136
	PknD	FadD	Mycolic acid biosynthesis	NA	In vitro kinase assay	109
		FabH	Mycolic acid pathway, cell wall biosynthesis	T45	In vitro kinase assay, mass spectrometry, site-directed mutagenesis	187
		GarA	Glycogen recycling, tricarboxylic acid cycle	NA	In vitro kinase assays	189
		GroEL1	Heat shock protein	NA	In vitro kinase assays, mass spectrometry	23
		KasA	Mycolic acid biosynthesis	NA	In vitro kinase assay	109
		KasB	Mycolic acid biosynthesis	NA	In vitro kinase assay	109
		MabA	Mycolic acid biosynthesis	T21, T114, T191	In vitro kinase assay, mass spectrometry, site-directed mutagenesis	188
		Mmpl7	Membrane transporter; resistance, nodulation, and cell division family	NA	2D electrophoresis, mass spectrometry	139
		Rv0516c	Anti-anti-sigma factor	T2	In vivo and in vitro kinase assays, mass spectrometry, site-directed mutagenesis	59a
		Rv1747	Putative ABC transporter	NA	In vitro kinase assay	60
	PknE	FadD	Mycolic acid biosynthesis	NA	In vitro kinase assay	109
		FabH	Mycolic acid pathway, cell wall biosynthesis	T45	In vitro kinase assay, mass spectrometry, site-directed mutagenesis	187
		GarA	Glycogen recycling, tricarboxylic acid cycle	NA	In vitro kinase assays	189
		GroEL1	Heat shock protein	NA	In vitro kinase assays, mass spectrometry	23
		KasA	Mycolic acid biosynthesis	NA	In vitro kinase assay	109
		KAsB	Mycolic acid biosynthesis	NA	In vitro kinase assay	109
		MabA	Mycolic acid biosynthesis	T21, T114, T191	In vitro kinase assay, mass spectrometry, site-directed mutagenesis	188
		Rv1747	Putative ABC transporter	NA	In vitro kinase assay	60
	PknF	FadD	Mycolic acid biosynthesis	NA	In vitro kinase assay	112
		FabH	Mycolic acid synthesis, cell wall	T45	In vivo and in vitro kinase assays, mass spectrometry, site-directed mutagenesis	187
		GarA	Glycogen recycling, tricarboxylic acid cycle	NA	In vitro kinase assays	189
		GroEL1	Heat shock protein	T25, T54	In vivo and in vitro kinase assays, mass spectrometry, site-directed mutagenesis	23
		KasA	Mycolic acid biosynthesis	NA	In vitro kinase assay	109
		KasB	Mycolic acid biosynthesis	NA	In vitro kinase assay	109
		Rv0020c	FHA-containing protein	NA	In vitro kinase assay	60
		Rv1747	Putative ABC transporter	NA	In vitro kinase assay	112
	PknG	GarA	Glycogen recycling, tricarboxylic acid cycle	T21	Kinase assays, mass spectrometry, site-directed mutagenesis, in vivo studies	128
	PknH	DacB1	Penicillin-binding protein, cell wall	Thr(s)	In vitro kinase assay, site-directed mutagenesis	201
		DosS	Dormancy	T198, T205	In vitro kinase assays, mass spectrometry, site-directed mutagenesis	25
		EmbR	Arabinan synthesis, cell wall	Thr(s)	In vitro kinase assay	111
		FadD	Mycolic acid biosynthesis	NA	In vitro kinase assay	109
		FabH	Mycolic acid pathway, cell wall biosynthesis	T45	In vitro kinase assay, mass spectrometry, site-directed mutagenesis	187
		GroEL1	Heat shock protein	NA	In vitro kinase assays, mass spectrometry	23
		KasA	Mycolic acid biosynthesis	NA	In vitro kinase assay	109
		KasB	Mycolic acid biosynthesis	NA	In vitro kinase assay	109
		Rv0681	TetR class transcription factor	T170	In vitro kinase assay, site-directed mutagenesis	201
	PknI	FadD	Mycolic acid biosynthesis	NA	In vitro kinase assay	109
	PknJ	Pyruvate kinase A	Glycolysis	S37	In vitro kinase assay	6
		EmbR	Arabinan synthesis, cell wall	NA	In vitro kinase assay	71
		MmaA4	Mycolic acid biosynthesis	NA	In vitro kinase assay	71
		PepE	Dipeptidase	NA	In vitro kinase assay	71
	PknK	FadD	Mycolic acid biosynthesis	NA	In vitro kinase assay	109
		VirS	Transcriptional regulator	Thr(s)	In vitro kinase assay, phospho-amino acid analysis	84
	PknL	FadD	Mycolic acid biosynthesis	NA	In vitro kinase assay	109
		GroEL1	Heat shock protein	NA	In vitro kinase assays, mass spectrometry	23
		KasA	Mycolic acid biosynthesis	NA	In vitro kinase assay	109
		KasB	Mycolic acid biosynthesis	NA	In vitro kinase assay	109
		MabA	Mycolic acid biosynthesis	T21, T114, T191	In vitro kinase assay, mass spectrometry, site-directed mutagenesis	188
		Rv2175c	DNA-binding protein, putative cell wall/cell division protein	T9	In vivo and in vitro kinase assays, mass spectrometry, site-directed mutagenesis, phospho-amino acid analysis, 2D electrophoresis	24, 27
Mycoplasma pneumoniae	PrkC	HMW1-3	Cytadherence	NA	Pro-Q Diamond staining of cell lysates	155
		MPN474	Surface protein	NA	Pro-Q Diamond staining of cell lysates	155
		P1	Adhesin	NA	Pro-Q Diamond staining of cell lysates	155
Myxococcus xanthus	Pkn2	HUα	Histone-like protein	T59	Suppressor screen in E. coli, in vitro kinase assay, site-directed mutagenesis	182
		β-Lactamase	Ampicillin resistance	Thr(s)	Phospho-amino acid analysis	183
	Pkn4	PFK	Glycolysis	T226	In vitro kinase assay, site-directed mutagenesis	117
	Pkn8	Pkn14	Soluble eSTK	Thr(s)	In vitro kinase assay	119
	Pkn14	MrpC	Transcription factor; fruiting body development	Thr(s)	In vitro kinase assay	119
		PFK	Glycolysis	T226	In vitro kinase assay, site-directed mutagenesis, phospho-amino acid analysis	117
Pseudomonas aeruginosa	PpkA	Fha1	Protein secretion, virulence	T362	In vitro kinase assay, mass spectrometry, site-directed mutagenesis	66, 113
Staphylococcus aureus	PknB	AFT-2	Human transcription factor	Thr73	Peptide arrays, in vitro kinase assay, mass spectrometry	107
		MgrA	Global transcriptional regulator	NA	In vitro kinase assay	181
		PurA	Purine biosynthesis	NA	In vitro kinase assay	39
		SarA	Transcription regulator; virulence	Thr(s)	In vitro kinase assay, phospho-amino acid analysis	38
		SA0498	Ribosomal protein L7/L12; central metabolism	Thr(s), Ser(s)	In vitro kinase assay, phospho-amino acid analysis	96
		SA0545	Phosphate acetyltransferase; central metabolism	Thr(s), Ser(s)	In vitro kinase assay, phospho-amino acid analysis	96
		SA0729	Triose isomerase; central metabolism	Thr(s)	In vitro kinase assay, phospho-amino acid analysis	96
		SA0731	Enolase; central metabolism	Thr(s), Ser(s)	In vitro kinase assay, phospho-amino acid analysis	96
		SA0944	Pyruvate dehydrogenase; central metabolism	Thr(s)	In vitro kinase assay, phospho-amino acid analysis	96
		SA1359	Elongation factor P; central metabolism	Thr(s), Ser(s)	In vitro kinase assay, phospho-amino acid analysis	96
		SA1499	Trigger factor; central metabolism	Thr(s), Ser(s)	In vitro kinase assay, phospho-amino acid analysis	96
		SA2340	Glyoxalase; central metabolism	Thr(s), Ser(s)	In vitro kinase assay, phospho-amino acid analysis	96
		SA2399	Fructose biphosphate aldolase; central metabolism	Thr(s)	In vitro kinase assay, phospho-amino acid analysis	96
Streptococcus agalactiae	Stp1	CovR	Response regulator; toxin expression	T65	In vitro kinase assay	93, 145
		DivIA	Cell division	NA	In vivo phosphopeptide enrichment approach, in vitro kinase assay	168
		PpaC	Mn-dependent inorganic pyrophosphatase; virulence	Ser(s)	In vitro kinase assay, phospho-amino acid analysis	144
		PurA	Purine biosynthesis	NA	In vitro kinase assay	146
Streptococcus pneumoniae	StkP	DivIVA	Cell division	Thr(s)	In vivo 2D electrophoresis, mass spectrometry, in vitro kinase assay	124
		FtsZ	Cell division	Thr(s)	In vitro kinase assay	55
		GlmM	Cell wall	NA	2D electrophoresis, in vitro kinase assay	125
		PpaC	Mn-dependent inorganic pyrophosphatase; virulence	Thr(s)	In vivo 2D electrophoresis, mass spectrometry, in vitro kinase assay	124
		RitR	Transcriptional regulator; iron transport	NA	In vitro kinase assay	184
		Spr0334	NA	Thr(s)	In vivo 2D electrophoresis, mass spectrometry, in vitro kinase assay	124
Streptococcus pyogenes	SP-STK	SP-HLP	Histone-like protein	Thr(s)	In vitro kinase assay	72
Streptomyces coelicolor	AfsK	AfsR	Transcriptional factor; secondary metabolism	Ser(s), Thr(s)	In vitro kinase assay	102
Yersinia spp.	YpkA	Eukaryotic actin	Cytoskeleton	NA	In vitro kinase assay	74

^aNA, not available.

Eukaryotic Ser/Thr kinases form complex signaling networks that involve phospho-dependent protein-protein interactions mediated by conserved protein modules or domains (137). One example found in some bacterial proteins is the FHA (forkhead-associated) domain, which interacts preferably with phosphorylated Thr residues (62). Proteins containing FHA domains include M. tuberculosis GarA, which is involved in glycogen recycling and the tricarboxylic acid cycle (14, 121). GarA is the substrate of several kinases, including PknB and PknG (128, 189; also see “Physiological roles and targets of eSTKs in bacteria. (i) Developmental processes”), and its FHA domain interacts with the phosphorylated activation loop of PknB. This interaction led to a model of substrate recruitment in which the PknB activation loop provides a secondary docking site for FHA-mediated binding (189). Alternatively, this interaction may stabilize the PknB activation loop and thereby activate the kinase, reminiscent of the case for many eukaryotic kinases. Regardless of the exact activation mechanism, the PknG/PknB-GarA interaction results in the phosphorylation of GarA at Thr21 (PknG) or Thr22 (PknB and other kinases) (128, 189), and this modification inhibits its function (44, 123).

The M. tuberculosis Ser/Thr phosphoproteome was recently analyzed, and the phosphorylation site motif for the M. tuberculosis eSTKs was determined (142). The identified motif, XααααTX(X/V)φ(P/R)I, in which T corresponds to the phosphorylated Thr residue, α corresponds to an acidic residue, and φ corresponds to a large hydrophobic residue, is shared by 6 of the 11 M. tuberculosis kinases (PknA, PknB, PknD, PknE, PknF, and PknH). There are, however, differences in the optimal substrate sequences for each kinase, including the hydrophobic residue at the third position after the phospho-Thr residue, the positions of acidic residues preceding the phospho-Thr residue, and the preference for Pro/Arg at the fourth position after the phospho-Thr residue. The inability of a Ser residue to substitute for a Thr residue as the phospho-acceptor for any kinase demonstrates the specificity of the kinase-substrate interaction. The phosphorylation sites of several substrates of S. aureus PknB were also analyzed, and although the most frequently found residues were serine and threonine, tyrosine phosphorylation was also identified (107). In addition, a proline residue was found next to the phosphorylated residue, indicating that PknB is the first identified prokaryotic representative of the proline-directed kinases, a major family of eukaryotic regulators involved in cellular physiology (also see “Physiological roles and targets of eSTKs in bacteria. (iv) Essentiality, virulence, and central metabolism”).

In the following sections, we discuss some examples that illustrate the diversity of biological processes and the complexity of signaling pathways in which bacterial eSTKs are involved. We describe the identified substrates and, if known, how phosphorylation affects their activity.

(i) Developmental processes.

M. xanthus grows unicellularly and forms large swarms of cells in the presence of nutrients. However, cells aggregate upon nutrient depletion and form multicellular fruiting bodies composed of around 100,000 cells. If unfavorable conditions persist, the aggregated cells can form metabolically dormant myxospores until nutrients become available again. The tight regulation of this complex developmental cycle has been reviewed in detail (69). Briefly, FruA is a key transcription factor that regulates the expression of a large number of genes involved in both fruiting body formation and sporulation. Another transcription factor, MrpC, controls fruA expression and is itself under a dual regulatory mechanism involving both a TCS and an eSTK. Transcription of mrpC is controlled by the MrpA/MrpB TCS, encoded on the same locus. In response to starvation signals, the sensor His kinase MrpA is activated and phosphorylates the response regulator MrpB, which in turn positively regulates mrpC expression, leading to aggregation. The eSTK Pkn14 also regulates MrpC by phosphorylating it on a Thr residue(s), reducing its affinity for both its own promoter and the fruA promoter (119, 120). The transmembrane eSTK Pkn8 phosphorylates Pkn14, and it has been suggested that Pkn8 and -14 form an eSTK cascade that inhibits the activity of MrpC during vegetative growth (69, 119) (Fig. 5). Consistent with this model, MrpC accumulated in Δpkn8 and Δpkn14 deletion strains, leading to fruA expression during early stationary phase and, consequently, to a faster developmental phenotype (119). Similar observations were reported for other organisms, such as cyanobacteria (197, 198) and streptomycetes (115, 190).

FIG. 5.

Regulation of the developmental cycle of Myxococcus xanthus. The M. xanthus transcriptional activator MrpC controls the expression of fruA, which encodes a major regulator of fruiting body formation and sporulation. MrpC is itself under dual regulation of an eSTK signaling cascade and a TCS. During vegetative growth, phosphorylation by the eSTK PknB14/8 signaling cascade decreases MrpC's affinity for the fruA promoter and its own promoter. In response to starvation signals, the MrpA/MrpB TCS activates the transcription of mrpC as well as itself, leading to the expression of fruA and the consequent developmental commitment.

Bacillus subtilis forms heat- and desiccation-resistant spores under conditions of nutrient limitation. These dormant spores germinate when stimulated by various nutrient sources, including specific amino acids. Peptidoglycan fragments released by growing cells can also induce germination, and the transmembrane eSTK PrkC is necessary for this response (160). Peptidoglycan fragments bind the PASTA repeats in the extracellular domain of PrkC and lead to kinase activation and subsequent germination. The essential translational GTPase elongation factor EF-G is phosphorylated in this PrkC-dependent germination pathway (160). PrkC-dependent phosphorylation of EF-G on a Thr residue(s) in vivo has also been described for vegetative B. subtilis cells (45, 50, 91, 98).

In addition to the phylogenetically conserved phosphorylation of EF-G identified in the phosphoproteomes of Corynebacterium glutamicum (17), E. coli (97), and Mycoplasma pneumoniae (156), other proteins involved in translation have been implicated as targets of eSTKs. Ser/Thr phosphorylation of another elongation factor, EF-Tu, has been reported for E. coli (158), Thermus thermophilus (95), Streptomyces coelicolor (65, 106), B. subtilis (1, 45), and S. pneumoniae. Importantly, EF-Tu phosphorylation inhibited binding to amino-acylated tRNA and to kirromycin, an antibiotic that inhibits the elongation step of translation by preventing the release of EF-Tu-GDP from the ribosome (5). For Streptomyces collinus, a ribosome-associated kinase was reported to phosphorylate several ribosomal proteins on Ser and Thr residues, which led to a significant reduction in protein synthesis (104, 105). Finally, Ser/Thr phosphorylation of a number of other ribosomal and ribosome-associated proteins has been observed in phylogenetically diverse bacteria by use of phosphoproteomic methodologies (17, 97, 125, 142, 156, 171, 174).

Regulation of translational initiation and elongation by phosphorylation in response to diverse stimuli is well established for eukaryotes (21). A bacteriophage T7 kinase expressed during E. coli infection phosphorylates EF-G and ribosomal protein S6, suggesting that this modification is important for the specific translation of phage mRNAs (148). An appealing but as yet untested hypothesis is that eSTKs direct the protein translation machinery to specific mRNAs that are required in the transition between different growth stages or in response to particular growth conditions.

(ii) Secondary metabolism.

Streptomyces spp. are Gram-positive filamentous soil bacteria that undergo complex morphological differentiation and synthesize a wide variety of small molecules, including many clinically relevant antibiotics. The production of actinorhodin by S. coelicolor is under the control of the regulator AfsR, which itself is a target of the membrane-associated eSTK AfsK (102). AfsK phosphorylates AfsR in vitro on a Ser and/or Thr residue(s), and the production of actinorhodin is reduced in a ΔafsK strain. Also, the transcription of afaS, a gene located immediately downstream of afsR that encodes a small protein involved in secondary metabolism, is delayed in a ΔafsK mutant. Phosphorylation of AfsK was shown to both increase its affinity for the afsS promoter and modulate AfsR ATPase activity, which is essential to activate the expression of afsS, resulting in the production of actinorhodin (87). The activity of AfsK itself is under regulation. The protein KbpA interacts with the AfsK kinase domain in its unphosphorylated form and subsequently inhibits AfsK autophosphorylation and, consequently, its activation (185). KbpA is encoded by a gene located upstream of afsK, and its transcription increases in response to the production of actinorhodin, suggesting that KbpA controls the production of secondary metabolites by modulating AfsK activity. In addition to AfsK, AfsR is also phosphorylated in vitro by at least two other eSTKs, AfsL and PkaG (151).

(iii) Cell division and cell wall synthesis.

A large number of eSTK substrates involved in cell division or cell shape have been reported. Here we focus on proteins with either mutant/overexpression phenotypes or direct in vivo phosphorylation. Those for which only in vitro data support phosphorylation are listed in Table 1. Given the presence of numerous eSTKs in bacteria such as M. tuberculosis (110) and the potential for in vitro artifacts (101), caution must be used in interpreting in vitro phosphorylation data to make definitive assignments of physiologically relevant kinase-substrate interactions. For example, PknG is the primary kinase responsible for phosphorylating M. tuberculosis GarA (128); however, PknB, PknD, PknE, and PknF also phosphorylate GarA in vitro (189).

Bacterial operons often contain functionally related genes, and there is a well-established relationship between bacterial morphology and the order of genes in a cluster involved in cell division (177). Thus, the presence of eSTK genes such as those for M. tuberculosis PknA and PknB in the operon containing the genes encoding penicillin-binding protein A (PBPA) and RodA suggests a role for these eSTKs in cell shape. Both pknA and pknB are expressed preferentially during exponential growth phase, and their overexpression results in long, broad, and in some cases branched cells, whereas partial depletion results in long cells (75). PknB phosphorylated PBPA on Thr362 and Thr437 when the proteins were coexpressed in E. coli (32). Moreover, a PBPA Thr437Ala mutant did not complement a pbpA deletion in a heterologous system and led to an increased number of nucleoids per cell and to mislocation of the protein. Another in vivo substrate of PknB is Wag31, a homolog of the essential cell shape/cell division protein DivIVA, and overexpression of a mutated form of Wag31 in which the Thr residue identified as the phospho-acceptor was replaced by a Glu residue resulted in morphological defects (75). PknA phosphorylates FtsZ, a homolog of eukaryotic tubulin and a major component of the division septum, on a Thr residue(s) in vitro (179). The coexpression of both proteins in E. coli resulted in significantly reduced FtsZ GTPase and polymerization activities. The S. pneumoniae eSTK StkP colocalized with FtsZ at midcell (55) and interacted with and phosphorylated FtsZ in vitro, but this modification was not confirmed in vivo. In the actinobacterium Corynebacterium glutamicum, overexpression of the PknA and PknB kinases also resulted in growth and morphological defects (49). In addition, PknA inhibits the activity of the muropeptide ligase MurC by phosphorylating it on multiple Thr residues in vivo (48). The enzyme GlmM, which catalyzes the interconversion of N-acetylglucosamine-6-phosphate into GlcN-1-P, which serves as a precursor for peptidoglycan synthesis, is activated by phosphorylation in E. coli (73) and is also a substrate of S. pneumoniae StkP (125).

The involvement of eSTKs in cell wall synthesis and division has been reported for other Gram-positive pathogens. For example, a Staphylococcus aureus ΔpknB strain grew slower and exhibited severe cell division defects, including multiple and incomplete septa, bulging, and irregular cell size (16). Consistent with this phenotype, a ΔpknB mutant also revealed a strong regulatory impact on the expression of genes encoding proteins involved in cell wall metabolism and autolysis (39), and this strain showed increased susceptibility to antimicrobials that target the cell wall (16, 39). The importance of PknB and its cognate phosphatase in the synthesis of peptidoglycan and wall teichoic acid was further supported by the impact of the ΔpknB mutation on S. aureus metabolism (92). Similar division and growth defects, as well as antibiotic susceptibility phenotypes, were described for Streptococcus pyogenes, Streptococcus agalactiae, S. pneumoniae, and Streptococcus mutans (9, 37, 72, 132, 144, 149, 174), as well as Enterococcus faecalis (82).

(iv) Essentiality, virulence, and central metabolism.

While eSTKs regulate many essential bacterial processes, most eSTKs are not themselves essential, with M. tuberculosis PknA and PknB being prominent exceptions (47, 75, 150). The reported essentiality of the C. glutamicum PknA and PknB homologs appears to be strain specific (49, 157). While mutations in B. subtilis PrkC (50, 159) or in the eSTKs of several streptococcal species (68, 72, 144) revealed that these proteins are not, strictly speaking, essential, they led to significant stationary-phase survival defects. Two plausible explanations for this discrepancy are the existence of one or more kinases with redundant functions and the fact that the in vitro conditions in which the majority of the studies were carried out differed considerably from the in vivo conditions.

Several eSTKs appear to have important roles in bacterial pathogenesis. An M. tuberculosis ΔpknG strain exhibited delayed mortality in SCID mice and reduced viability in a BALB/c mouse infection model, with significant reductions in the bacillary loads in the lungs, spleen, and liver (29). Similarily, an M. tuberculosis ΔpknH mutation increased the bacillary load in a BALB/c mouse infection model (133). PknH phosphorylates the transcription activator EmbR (involved in antibiotic production in S. coelicolor [see above]) through interaction with its FHA domain (111; also see “Structural and functional studies of bacterial eSTKs”). Phosphorylation of EmbR activates the expression of the embCAB operon, which encodes arabinosyltransferases involved in the metabolism of arabinan, an essential component of the mycobacterial cell wall, and resistance to ethambutol (163). Since EmbC catalyzes the conversion of lipomannan into lipoarabinomannan, a cell wall lipoglycan that prevents the host proinflammatory response (20), PknH may modulate the host response during infection by controlling the ratio of these two cell wall components (163). PknH also phosphorylates DosR on two Thr residues, and DosR together with DosS controls the DosRS regulon responsible for the nitric oxide and hypoxia response, an important evasive response to host macrophages (25).

The Yersinia pseudotuberculosis eSTK YpkA is absolutely essential for virulence (52). YpkA is produced in an inactive form that is activated upon secretion into the host cell and interaction with the host actin (74). Activation of YpkA results in disruption of the host actin cytoskeleton, presumably leading to the inhibition of macrophage function during infection. S. aureus PknB was also found to be released into the external milieu (107), with the first report of a prokaryotic proline-directed kinase with the potential to phosphorylate host target proteins involved in essential processes such as signal transduction, cell communication, gene regulation, and immune responses (also see “Physiological roles and targets of eSTKs in bacteria” and Table 1). The eSTK PrkC is required for M. pneumoniae adhesive growth and cytotoxicity (155) (Table 1). PrkC phosphorylates several cytadherence proteins as well as the major adhesin P1 and the surface protein MPN474, and it was observed that these modifications affect the accumulation of these proteins. Homologs of PrkC have also been implicated in S. pyogenes adherence and invasion and in E. faecalis persistence (72, 82). The involvement of eSTKs in virulence has also been reported for several other pathogens, such as Streptococcus pneumoniae, E. faecalis, Pseudomonas aeruginosa, and Salmonella enterica serovar Typhi (33, 43, 46, 144-146, 191).

M. tuberculosis PknG exemplifies the involvement of eSTKs in central metabolism. A ΔpknG mutant showed higher intracellular levels of glutamate, glutamine, and derivatives, suggesting that PknG controls glutamate synthesis in response to nutritional stress signals (29). Consistent with this hypothesis, deletion of the C. glutamicum pknG homolog impaired glutamine utilization. OdhI, an FHA-containing protein that inhibits the 2-oxoglutarate dehydrogenase (ODH) complex of the tricarboxylic acid cycle, was inhibited by phosphorylation and identified as an in vivo PknG substrate (121). Similarly, M. tuberculosis PknG phosphorylates GarA, an OdhI homolog, on Thr21 in vivo (128).

Ser/Thr PROTEIN PHOSPHATASES

Reversible phosphorylation occurs in organisms from bacteria to humans (166). In prokaryotes, the need for dedicated phosphatases was not initially appreciated in the context of two-component systems and phosphorelay signal transduction (64), since both phosphohistidine and aspartyl-phosphate residues undergo relatively rapid hydrolysis (167, 172, 196), and therefore a dedicated enzyme to remove the phosphate groups was not thought to be necessary. In contrast, phosphorylated Ser, Thr, and Tyr residues are not as labile, and therefore cognate phosphatases can be necessary in order to quench signaling cascades (3). Homologs of four eukaryotic protein phosphatase superfamilies are present in archaea and bacteria, including conventional and low-molecular-weight protein tyrosine phosphatases, phosphoprotein phosphatases (PPPs), and metal-dependent phosphatases (PPMs) (78, 165). We focus here on the last two classes (for a review of tyrosine phosphatases, see reference 10). Eukaryote-like serine/threonine phosphatases (eSTPs) belong to two structurally distinct phosphatase families: PPPs and PPMs. The PPP family members are formally described as serine/threonine phosphatases but can also dephosphorylate phosphohistidine and phosphotyrosine residues (10). For example, PrpE, a B. subtilis PPP family member, exclusively removes phosphate groups from phosphotyrosines in vitro (70). Other PPP family members have been annotated for bacteria and archaea, and in vitro, most are dual-specificity phosphatases that dephosphorylate tyrosine residues in addition to serine and threonine residues (85, 88, 108, 164, 170).

The serine/threonine Mg²⁺- or Mn²⁺-dependent phosphatases of the PPM family share a conserved catalytic domain with eukaryotic PP2C that contains 11 to 13 signature motifs containing eight conserved amino acids (2, 19, 78, 165, 199) (Fig. 6A). Bacterial PPMs can be divided into two subfamilies depending on the presence of motifs 5a and 5b (199). One subfamily, which includes the B. subtilis sporulation-specific phosphatase SpoIIE (40) and the stress response phosphatases RsbU and RsbX (34, 42, 76, 126, 193), lacks the 5a and 5b catalytic domain motifs (77, 199). This subfamily also includes phosphatases that are cognate to non-eukaryote-like serine/threonine kinases and is not discussed further. Members of the second subfamily (eSTPs) contain all 11 signature motifs and are cognate phosphatases to the eSTKs discussed above. This subfamily is discussed in detail in the next section and is of particular biological importance, as there are no known inhibitors of these proteins.

FIG. 6.

Ser/Thr phosphatase structures. (A) Primary sequence alignment of human PP2C (residues 1 to 299) and PphA (residues 1 to 241) catalytic domains. Conserved amino acids are indicated in bold, and those that also form part of the metal-binding pocket are shown in red. Motifs 1 to 5, 5a, 5b, and 6 to 11, as defined by Bork et al. (19), are indicated in boxes. (B) Crystal structure of the human PP2C structure (PDB accession number 1A6Q). The β-sandwich is represented in yellow, α-helices are represented in blue, the large irregular loop is shown in red, and two manganese ions are shown in gray. His62 (orange) is predicted to be a general acid. The C-terminal domain (α7, α8, and α9) is characteristic of mammalian PP2C and is absent from prokaryotic family members. (C) Crystal structure of the Thermosynechococcus elongatus eSTP PphA (PDB accession number 2J86). The β-sandwich is shown in yellow, the α-helices are shown in blue, the three magnesium ions are shown in gray, and the large irregular loop is shown in red. The flap subdomain (red) appears flexible and is thought to be involved in substrate binding and catalytic activity. Instead of His62, Met62 (orange) occupies the homologous position.

PPM Family eSTPs

eSTPs have been characterized for both Gram-negative and Gram-positive bacteria (Table 2). While typically there appears to be one cognate phosphatase per kinase and phosphatases are often encoded in the same operon as their partner eSTKs, M. tuberculosis has 11 eSTKs and a single eSTP (7, 80). The physiological relevance of this discrepancy has yet to be ascertained, but it suggests that, at least in M. tuberculosis, phosphatases are subject to complex regulation.

TABLE 2.

eSTPs in bacteria^a

Species	eSTP	Type	Partner kinase	Substrate	Function	Phospho-residue(s)	Methodology	Reference
Bacillus anthracis	Ba-Stp1	PP2C	Ba-Stk1	Ba-Stk1-P	eSTK	NA	In vitro phosphatase assay	161
Bacillus subtilis	PrpC	PP2C	PrkC	EF-G	Translation factor	NA	In vitro phosphatase assay	50
				CpG	Small ribosome-associated GTPase	NA	In vitro phosphatase assay	1
				EF-Tu	Translation factor	NA	In vitro phosphatase assay	1
				HPr kinase	Phosphotransferase system	S46	In vitro phosphatase assay	169
				PrkC-P	eSTK	NA	In vitro phosphatase assay	129
				YezB	Stress response	NA	In vitro phosphatase assay	1
	SpoIIE	PPM		SpoIIAA	Phosphoprotein anti-anti-sigma factor	NA	In vitro phosphatase assay	127
	RsbU	PPM	RsbV-P	RsbV-P	Regulation of sigma B	NA	In vitro phosphatase assay	193
	RsbX	PPM	RsbS-P	RsbS	Regulation of sigma B	NA	In vitro phosphatase assay	193
Listeria monocytogenes	Stp	PPM		EF-Tu	Translation factor	NA	2D phosphoprotein gel electrophoresis, in vitro phosphatase assay, decreased kirromycin sensitivity	5
Mycobacterium tuberculosis	PstP	PPM	PknB	PBPA	PBP, cell wall	NA	In vitro phosphatase assay	32
				PknB-P	eSTK	T171/T173	Mass spectrometry	18
Mycoplasma pneumoniae	PrpC	PP2C		HPr	Phosphocarrier protein of phosphotransfer system	S46	In vitro phosphatase assay	61
Myxococcus xanthus	Pph1	PP2C	Pkn5	Pkn5	Negative effector of development	NA	Yeast two-hybrid interaction	180
Staphylococcus aureus	Stp1	PP2C	Stk1	Stk1-P	eSTK	NA	In vitro phosphatase assay	16
Streptococcus agalactiae	SaSTP	PP2C		PpaC	Family II inorganic pyrophosphatase		In vitro phosphatase assay	144
				PurA	Purine biosynthesis	NA	In vitro phosphatase assay
Streptococcus pneumoniae	PhpP	PP2C	StkP	StkP-P	eSTK		Coimmunoprecipitation	131
				RitR	Transcriptional regulator		In vitro phosphatase assay	184
Streptococcus pyogenes	SP-STP	PP2C	SP-STK	SP-HLP	Histone-like protein	Thr(s)	In vitro phosphatase assay	72
							Phospho-amino acid analysis
				SP-STK-P	eSTK	NA	In vitro phosphatase assay	72

^aNA, not available.

Structure and function of eSTPs.

The conserved catalytic core domain of human PP2C contains a central β-sandwich comprised of two five-stranded antiparallel β sheets, each flanked by a pair of antiparallel α helices (31) (Fig. 6B). A binuclear metal center is located within the channel of the β-sandwich, with the two metal ions located at the base of the cleft and hexa-coordinated with water and amino acids. The mechanism of dephosphorylation is thought to occur by metal-activated water nucleophilic attack of the phosphorus atom. This mechanism is supported by the high level of conservation of active site residues among eukaryotic and bacterial homologs. There are three additional α helices (helices 7 to 9) that associate with the core domain and are also hypothesized to be involved in substrate specificity and/or regulation, but these are absent in bacterial PP2C family members (Fig. 6B). In the human PP2C structure, a 60-residue segment between β7 and β8 contacts a large flap comprised of a helix followed by an irregular loop. This loop does not appear to be involved in enzyme catalysis but may assist with protein substrate specificity (31).

The bacterial eSTPs M. tuberculosis PstP (192), Mycobacterium smegmatis MspP (192), Streptococcus agalactiae STP (147), and Thermosynechococcus elongatus PphA (154) have been crystallized. Their catalytic domains are structurally nearly identical to that of human PP2C and contain the highly conserved active site residues (166) (Fig. 6A). However, there are several key differences in the structures, namely, the coordination of a third metal ion, the lack of the His62 residue shown to function in splitting of the P-O bond, and the repositioning of the flap region (11, 15, 143, 147, 166) (Fig. 6C). While metal ions 1 and 2 in the bacterial structures are positioned exactly at the locations of the two metal ions present in the human PP2C structure, bacterial enzymes have an additional metal ion in the active site. Likely due to different crystallization buffers, the third ion is magnesium in the PphA and STP structures and manganese in the PstP and MspP structures. Another important difference is the absence of the His62 residue in the bacterial structures. This residue has been shown to function as an acid splitting the phosphate oxygen bond in human PP2C (31). His62 is replaced with a methionine residue in PphA, PstP, and STP and with a phenylalanine residue in MspP. The most striking structural difference between the bacterial and human structures corresponds to the flap subdomain. In the bacterial structures, this region is shifted further away from the active site, being located comparatively closer in PstP and MspP than in PphA and STP. Mutational analysis of the flap region of PphA revealed that Arg169, which seems to be involved in structural organization of the flap domain, causes a pronounced effect on enzymatic activity (154). This suggests that the flap region serves as a mobile element that may facilitate substrate binding and turnover and, furthermore, may introduce specificity to the dephosphorylation of substrates.

Physiological roles of eSTPs.

The physiological functions of eSTPs are not as well understood as those of eSTKs, and Table 2 is a comprehensive list of their identified substrates. S. pyogenes eSTP and S. pneumoniae PhpP deletion mutants are nonviable, highlighting the importance of eSTPs in bacterial physiology (72, 131). However, a deletion of S. pneumoniae PhpP could be obtained when the eSTK (StkP) located in the same operon was also deleted, suggesting that an important role of these phosphatases is to modulate the autophosphorylation state of their partner kinases (131). For example, autophosphorylated StkP is a substrate for PhpP (125), the kinase activity of autophosphorylated M. tuberculosis PknB is reduced by PstP-mediated dephosphorylation (18, 26), B. subtilis PrpC dephosphorylates PrkC in vitro (127), the enzymatic function of the Bacillus anthracis eSTK Ba-Stk1 is reduced after desphosphorylation by the eSTP Ba-Stp1 (161), and a strong interaction was reported between the M. xanthus eSTP Pph1 and the eSTK Pkn5 (180). In some cases, however, the eSTP target is a noncognate kinase. S. pneumoniae RitR is an orphan TCS response regulator that is phosphorylated by StkP and forms a ternary complex with the eSTP PhpR that is necessary for regulation of RitR activity (184).

Since eSTKs regulate cell division and peptidoglycan synthesis in Mycobacteria, Corynebacterium, and Streptomyces (110), it is not surprising that deletions in eSTPs also alter normal cell division or growth in diverse bacteria. For example, deletion of S. aureus STP results in thickened cell walls and increased resistance to lysostaphin (16) and in other cell wall defects (129), suggesting alterations in peptidoglycan composition and/or structure. Deletion of the Streptococcus mutans phosphatase PppL, like its cognate eSTK PknB, results in abnormal cell shape (9). Deletion of B. subtilis PrpC also results in prolonged stationary-phase survival, which is the opposite of the case for the kinase null mutant, which has a strong stationary-phase defect, suggesting the importance of modulating the cognate kinase PrkC (50). A number of in vitro targets of PrpC are proteins involved in translation, including EF-Tu, EF-G, and CpgA (1, 50). EF-Tu is also a substrate of the eSTP Stp in Listeria monocytogenes (4, 5). Thus, while eSTPs appear to modulate processes involved in growth, the mechanistic basis of this regulation is not well understood.

A null mutant of the M. xanthus eSTP Pph1 exhibited defects during late vegetative growth, swarming, and glycerol spore formation (167). Under starvation-induced developmental conditions, the Pph1 mutant showed reduced aggregation and failure to form fruiting bodies with viable spores (180). A mutant in the B. subtilis eSTP PrpC resulted in defects in spore and biofilm formation (99), although unlike its cognate kinase PrkC, it was not necessary for spore germination in response to peptidoglycan fragments (160).

Although the role of eSTPs in virulence has largely not been examined, deletion of an L. monocytogenes STP inhibited growth in a murine model of infection (5). A B. anthracis STK/STP double mutant had an impaired ability to survive within macrophages (161), and an S. agalactiae STK/STP double mutant was severely attenuated in a neonatal rat sepsis model (144). However, these defects could be due to the absence of both proteins or just of the kinase. Finally, exotoxins, including the hemolysins, play an important role in the pathogenesis of S. aureus infections, and inactivation of the eSTP Stp1 reduces hemolysin expression (22).

CONCLUDING REMARKS

The studies described here and listed in Tables 1 and 2 illustrate the important roles that eSTKs and eSTPs play in the physiology of diverse bacteria, as they regulate central processes such as cell division and translation. The mechanistic basis underlying their roles is still far from clear, and most of the evidence for specific substrates is based on in vitro phosphorylation experiments that are notoriously subject to false-positive results. However, in the near future, efforts aimed at confirming that these largely in vitro observations accurately reflect the in vivo situation should bear fruit, and we should be in a position to understand this regulation in mechanistic detail, as well as the physiological conditions responsible for its activation.

Biography

Sandro F. F. Pereira received his Ph.D. in biology from the Instituto de Tecnologiaq Uímica e Biológia at the Universidade Nova de Lisboa, Lisbon, Portugal, in 2008. The last 3 years of his thesis research were conducted in the Laboratory of Microbiology at The Rockefeller University in New York, NY. His thesis focused on cell division and cell wall synthesis in the human pathogen Staphylococcus aureus. He is currently a postdoctoral fellow in the Department of Microbiology and Immunology, Columbia University, New York, NY. His postdoctoral research explores the exit from bacterial dormancy, particularly the signaling pathway mediated by PrkC, a serine/threonine kinase in Bacillus subtilis.

Lindsie Goss earned her undergraduate degree in molecular and cell biology, with an emphasis in genetics, from the University of California, Berkeley. Before attending graduate school, she was a research assistant for 2 years at the University of California, San Francisco, in the laboratory of Matthias Wabl. She is currently a Ph.D. candidate in Jonathan Dworkin's laboratory at Columbia University in the Department of Microbiology and Immunology. She has been working on her thesis research for 2 years to understand the role of PrkC, one of the eukaryote-like Ser/Thr kinases in Bacillus subtilis.

Jonathan Dworkin studied physics as an undergraduate at Swarthmore College and received his Ph.D. with Peter Model at The Rockefeller University in 1998. His postdoctoral training was with Richard Losick at Harvard University, where he worked on aspects of bacterial development and chromosome segregation. He started his independent position in 2004 in the Department of Microbiology and Immunology, Columbia University, and his lab presently examines a range of topics, including exit from dormancy, cell biological and genetic characterization of peptidoglycan synthesis, and the role of the microbiota in the physiology of the bacteriovorous nematode Caenorhabditis elegans.