PPIP5Ks (E.C. 220.127.116.11) are ATP-dependent, small molecule kinases that synthesize inositol pyrophosphates (PP-InsPs), which are intracellular signals with functionally significant and “energetic” phosphoanhydride bonds. The purification and cloning of these enzymes was first described in a body of work that was performed independently by two groups and published in 2007 (Choi et al. 2007; Fridy et al. 2007; Mulugu et al. 2007). PPIP5Ks are distributed throughout eukaryotic kingdoms; the orthologs have aliases: Asp1 in Schizosaccharomyces pombe, Vip1 in Saccharomyces cerevisiae, VIH in Arabidopsis thaliana. Since yeasts and plants diverged from the animal kingdom 1.3 and 1.5 billion years ago, respectively (see www.timetree.org), it is clear that PPIP5Ks are evolutionarily ancient enzymes. Two PPIP5K genes are expressed in both mammals and plants (Choi et al. 2007; Fridy et al. 2007; Laha et al. 2015; Mulugu et al. 2007).
The Biological Functions of the Kinase Activity
A kinetic analysis of human PPIP5K2 has determined that the Km value for ATP is 22 and 37 μM for the phosphorylation of InsP6 and 5-InsP7, respectively (Weaver et al. 2013). Physiological levels of ATP lie in the 1–5 mM range, well above the level that ensures the kinase is operating at Vmax (i.e., under zero-order conditions) with respect to the ATP. Zero-order conditions for the inositol phosphate substrates are also expected, since the Km values for InsP6 and 5-InsP7 (0.4 and 0.06 μM, respectively) are at least 15-fold below their apparent cellular levels (Weaver et al. 2013). In other words, the rates of the kinase activities of PPIP5Ks are kinetically insulated from any physiological fluctuations in the cellular concentrations of both ATP and substrates.
At the time of writing, there are no published phenotypes for knock-out of either PPIP5K1 or PPIP5K2 in any animal model. Such information is likely to be available soon, since a PPIP5K2 knock-out mouse strain has been produced, according to the International Mouse Phenotyping Consortium (http://www.mousephenotype.org). The siRNA-mediated knock-down of PPIP5K expression in cultured mammalian cells has phenotypic consequences: decreased cell proliferation (Gokhale et al. 2011), inhibition of the PtdIns(3,4,5)P3-AKT signaling cascade (Gokhale et al. 2013), reduced cell migration (Machkalyan et al. 2016), and inhibition of certain innate immune responses – specifically, RIG-1 dependent transcription of interferon-β in response to viral invasion (Pulloor et al. 2014). In the latter study, further evidence was obtained that reinforces the idea that it is the catalytic activity of PPIP5K that regulates immune responses: ectopic expression of PPIP5K enhanced interferon-β transcription. Moreover, no such response was observed following expression of a kinase-dead mutant version of PPIP5K (Pulloor et al. 2014). Evidence has also been provided that 1-InsP7 and InsP8 act to enhance the degree of phosphorylation of IRF3, which promotes its dimerization and entry into the nucleus, where it stimulates interferon-β transcription (Pulloor et al. 2014). It is intriguing that 1-InsP7 and/or InsP8 also appear to regulate innate immune responses in A. thaliana (Laha et al. 2015); here, cellular levels of InsP8 are elevated by methyl-jasmonate, one of a group of plant hormones that helps protect plants from herbivorous insects and necrotrophic fungi (Laha et al. 2015). This protective response is impaired in lines of A. thaliana in which the kinase domain of VIH2 is disrupted by insertional mutagenesis (Laha et al. 2015). It has been speculated that InsP8 may augment transcriptional activation by methyl-jasmonate (Laha et al. 2015).
Generally speaking, a case can be made that PPIP5Ks and their orthologs are important for regulating homeostatic responses to fluctuating environmental conditions. In addition to the roles of PPIP5K in immunity described above, experiments with S. cerevisiae have indicated that 1-InsP7 activates transcriptional adaptations to inorganic phosphate deprivation (Lee et al. 2007). Other studies have revealed that the kinase activities of Asp1 modulate the environmentally controlled switch to an alternative morphology – pseudohyphal invasive growth (Pohlmann and Fleig 2010). The latter appears to involve regulation by 1-InsP7 and/or InsP8 of microtubule dynamics, which is also a presage of further molecular actions of these PP-InsPs upon chromosome segregation and intracellular trafficking (Pohlmann et al. 2014). Finally, InsP8 has been implicated as being a sensor of bioenergetic health in mammalian cells (Choi et al. 2008).
Naturally, PPIP5Ks should be considered as signaling “on-switches,” by virtue of their synthesizing the intracellular signals 1-InsP7 and 1,5-InsP8. Additionally, as there is a significant body of work devoted to the many signaling actions of 5-InsP7 (Thota and Bhandari 2015), PPIP5Ks should also be considered as “off-switches,” through their metabolism and hence deactivation of 5-InsP7.
Motifs and Domains in the PPIP5Ks
The type 1 and 2 forms of PPIP5Ks are relatively large proteins, approximately 160 and 140 kDa, respectively; they may oligomerize into even larger complexes (Choi et al. 2007). The kinase domain is self-contained at the N-terminal one-third of these proteins (Fridy et al. 2007; Mulugu et al. 2007; Pohlmann et al. 2014; Wang et al. 2012). A particularly remarkable feature of the PPIP5Ks, which was originally recognized by York’s laboratory, is that they also contain separate 1-phosphatase domains (Fig. 2), in other words, a rare example of a single protein that catalyzes a “futile” cycle (Fridy et al. 2007; Mulugu et al. 2007; Pohlmann et al. 2014; Wang et al. 2015). The two phosphatase domains of PPIP5K1 and PPIP5K2 share 77% sequence identity.
There is a considerable general interest in understanding what are the particular advantages of this specialized kinase/phosphatase bifunctionality. It is intuitive that if there were to be coordinated reciprocal regulation of the two competing catalytic activities in PPIP5Ks, this could form the basis for exquisitely sensitive mechanisms of control over PP-InsP metabolism and function. However, as yet, no such regulatory processes have been demonstrated.
Nevertheless, a mechanism for regulating the phosphatase activity has been described, at least for Asp1; this domain of the protein hosts a Fe-S cluster that inhibits phosphatase activity and thereby promotes net kinase activity (Wang et al. 2015). This may be a unique role for a Fe-S cluster: despite their being widely distributed throughout Nature, there are no precedents for any having a direct catalytic influence upon a phosphatase activity. It has been proposed that the role of the cluster in Asp1 may be to couple PP-InsP turnover to iron availability (Wang et al. 2015).
Data obtained by electron paramagnetic resonance indicate the cluster has a [2Fe-2S]2+ arrangement that apparently is not redox active. It remains to be established whether or not orthologs of Asp1 – particularly mammalian PPIP5Ks – also contain Fe-S clusters. Reliable predictions have not been obtained from multiple sequence alignments, in part because Cys-based consensus motifs for [2Fe-2S] centers are so diverse (Wang et al. 2015). Also, towards the C-terminus of Asp1, where the Fe-S cluster is most likely to be hosted, the amino-acid sequence diverges substantially from those in PPIP5K1 and PPIP5K2, which in turn diverge considerably from each other at their own C-termini (Wang et al. 2015). Moreover, the standard biophysical techniques that can unequivocally detect Fe-S clusters require considerable (i.e., mg) quantities of highly pure enzyme; it has not yet proved possible to satisfy that requirement, due to difficulties expressing and purifying such large proteins that are also rather labile (Choi et al. 2007).
Within the phosphatase domain of the PPIP5Ks is a cryptic phosphoinositide-binding domain (PBD) (Gokhale et al. 2013; Gokhale et al. 2011). In the latter studies, individual phosphoinositides were imbedded in immobilized vesicles, and binding of PBDs was monitored by surface plasmon resonance. The preferred ligand is PtdIns(3,4,5)P3: Kd = 0.1 and 0.7 μM for the PBDs of PPIP5K1 and PPIP5K2, respectively (Gokhale et al. 2013). PtdIns(4,5)P2 also binds, but less strongly (0.6–1.3 μM, respectively). Nevertheless, since PtdIns(4,5)P2 is the more abundant of the two phosphoinositides in vivo, both should be considered as potential physiological ligands. There is also evidence that in some cell types, stimulus-dependent elevations in PtdIns(3,4,5)P3 levels promote translocation of PPIP5K1 to the plasma membrane (Gokhale et al. 2013; Gokhale et al. 2011). The functional significance of this regulated intracellular compartmentation may be to locally deplete levels of 5-InsP7, which can inhibit PtdIns(3,4,5)P3-signaling cascades (Chakraborty et al. 2010; Gokhale et al. 2013).
The C-termini of both PPIP5Ks are unusually long, intrinsically disordered regions (IRD), which are typically considered as molecular scaffolds for promoting binding of other proteins (Machkalyan et al. 2016). A proteomic-based search for proteins that bind to PPIP5K1 yielded many that participate in trafficking of intracellular vesicles, particularly those in the exocyst complex that tether post-Golgi vesicles to plasma membranes (Machkalyan et al. 2016). Additionally, there is evidence that PPIP5K1 interacts with proteins that participate in lipid metabolism and cytoskeletal organization (Machkalyan et al. 2016). Thus, it appears likely that further delving into scaffolding functions for PPIP5Ks will be a productive future research direction.
The NCBI database lists multiple transcripts for the mammalian forms of both PPIP5K1 (human GeneID:9677) and PPIP5K2 (human GeneID: 23,262), suggestive of the occurrence of alternatively spliced mRNAs. The extent to which different protein isoforms are expressed has not yet been investigated. With regard to PPIP5K2, some of its putative isoforms contain, within the IDR, a so-called CPR region (for “contains penta-arginine”) (Yong et al. 2015). This penta-arginine functions as a nuclear localization sequence (NLS). Indeed, importin-5 and importin-8, which both participate in shuttling proteins into the nucleus, were found to be PPIP5K2 binding partners by subtractive proteomic analysis of PPIP5K2 pull-downs (Yong et al. 2015). Furthermore, when a CPR-containing version of GFP-tagged, human PPIP5K2 was expressed in HEK cells, a distinct nuclear pool of this protein was observed (at a concentration that was approximately 30% of that found in the cytoplasm (Yong et al. 2015)). Mutagenesis of the NLS (RRRRR to RAAAR) reduced the degree of nuclear localization almost 4-fold (Yong et al. 2015).
A Ser residue that lies immediately to the C-terminus of the penta-arginine is phosphorylated by an as yet unknown protein kinase (Yong et al. 2015); candidates include PKA, CaMK2, and AKT (Phosida database; http://18.104.22.168/phosida/index.aspx). The mutation of this Ser residue to Ala increases the nuclear concentration of the PPIP5K2 by 40% (Yong et al. 2015). These data indicate that phosphorylation of the native Ser inhibits the functionality of the NLS. Thus, there appears to be regulation of the degree to which this particular isoform of PPIP5K2 is distributed between the cytosol and the nucleus. A role for nuclear PPIP5K activity has not yet been established. PPIP5K1 does not contain the penta-arginine motif, nor, as far as is known, any other NLS. Indeed, PPIP5K1 is entirely cytosolic (Yong et al. 2015).
The Phosida database lists several other Ser and Thr residues on mammalian PPIP5Ks that can be phosphorylated; the functional significance of these events remains to be explored.
Structure of the PPIP5K Kinase Domain
The kinase domains of the PPIP5Ks have evolved to overcome considerable catalytic hurdles: the accommodation into the active site of the steric bulk and intense electronegativity of Nature’s most concentrated three-dimensional array of phosphate groups; the maintenance of stringent ligand specificity; and the ability to overcome a significant energy barrier to the transition state to synthesize “high-energy” phosphoanhydride bonds. To gain an atomic-level understanding of how these acute challenges have been met, the structure of the kinase domain of human PPIP5K2 was solved using X-ray crystallography (Wang et al. 2012). It is believed the structural information that was obtained applies equally to human PPIP5K1: the kinase domains of the two enzymes share 86% sequence identity, and residues that are dissimilar are confined to the protein surface, distal to the active site (Wang et al. 2012).
The analysis of the crystal structure of the human PPIP5K2 kinase domain revealed that ligand specificity is in part enforced by the geometry of the substrate-binding pocket, which comprises two near-parallel grooves that form a staggered “H”-shape. This architecture makes a perfectly tailored aperture for accommodating six phosphate/pyrophosphate groups attached to an inositol ring in its chair conformation (Wang et al. 2012). Specificity is additionally imposed by an array of Lys and Arg residues that make polar contacts with every phosphate group in both InsP6 and 5-InsP7; Mg2+ ions also contribute to specificity by electrostatic bridging between certain phosphate groups (Wang et al. 2012). Substrate binding to the PPIP5K-nucleotide complex is accompanied by induced fit motion of side chains of three active site residues: Arg262, Arg281, and Lys329; such conformational dynamics within the active-site of PPIP5K help limit the degree of free energy of activation. Were there instead to be backbone rearrangements, the energetic demand would have been higher (Wang et al. 2012). Induced fit motion of amino acid side chains is also known to help define ligand specificity. Finally, in the active site of the kinase, the inositol ring of InsP6 and 5-InsP7 is presented perpendicular to the plane of the nucleotide’s β-phosphates, which avoids steric and electrostatic clashing between the nucleotide and the nonreacting oxygens on the 1-phosphate of the substrate (Wang et al. 2012). This innovative topology distinguishes PPIP5Ks from all of the other inositol phosphate kinases that phosphorylate hydroxyl groups (Wang et al. 2012).
As for understanding how the catalytic cycle moves from the substrate-bound ground state to the transition state, a crystal structures that were complexed with a MgF3− transition-state mimic revealed that conformational dynamics of both enzyme and substrate are involved (Wang et al. 2012). In particular, there is exploitation of a specific property of a phosphoester bond – its ability to rotate. Finally, an extensive negative charge balance by positively charged amino acid residues facilitates a partly associative, in-line phosphoryl transfer mechanism (Wang et al. 2012). That is, the new P-O bond can be formed by nucleophilic attack of the acceptor oxygen, before the original P-O bond with the donor oxygen is broken.
The structural data also beg some questions: ATP is enveloped between two sets of antiparallel β-sheets, so that less than 10% of the nucleotide is predicted to be solvent accessible. Thus, it can be anticipated that the enzyme must undergo considerable conformational changes when admitting ATP and releasing ADP. Unfortunately, the nature of these rearrangements is unknown, as the structure of the apo-enzyme has not yet been solved. A second issue is, perhaps, more subtle: considering how highly constraining is the structure of the substrate-binding pocket that will only admit ligand presented in the appropriate orientation – that is, a direct hit – could that limit the enzyme’s ability to attract randomly diffusing ligand from the bulk phase? To illustrate that point, consider the usual outcome of a classic carnival game as an analogy: the high failure rate when attempting to throw a table tennis ball into the narrow neck of a distant goldfish bowl.
Further work demonstrated how PPIP5Ks appear to circumvent the possibility of such a “high failure rate.” Through X-ray analysis of several synthetic PP-InsP analogs in crystal complexes with the PPIP5K2 kinase domain, together with biochemical assays and mutagenesis, it has been shown that this enzyme utilizes a surface-mounted, ligand-binding site that is adjacent to the catalytic pocket (Wang et al. 2014). It has been concluded that this unusual ligand-binding site facilitates substrate capture from the bulk phase, prior to its transfer into the catalytic pocket, a “catch-and-pass” reaction mechanism (Wang et al. 2014).
Once again, a structural rationalization of one question raises another: occupation of the substrate-capture site stimulates an intrinsic, nonproductive ATPase activity (Wang et al. 2014). Intrinsic ATPase – reactivity of the gamma phosphate of ATP with water instead of substrate – appears to be an inherent property of most, if not all, kinases. However, the catalytic cycle of the kinase domain of PPIP5K is apparently unique in that the ATPase activity is stimulated by the substrate itself, specifically, during its occupation of the capture site (Wang et al. 2014; Weaver et al. 2013). It has been estimated that during the phosphorylation of 5-InsP7 by human PPIP5K2, an accompanying substrate-stimulated ATPase activity accounts for 17% of the total ATP that is consumed (Weaver, Wang, and Shears 2013). The proportion is much higher during InsP6 phosphorylation: 50% (Weaver et al. 2013). It seems reasonable to speculate that this ATP hydrolysis is not simply “wasted,” but instead has some significance for the catalytic cycle.
In view of evidence that PPIP5Ks regulate cell biology through both catalytic and scaffolding effects (see above), it would be analytically useful to have access to cell-permeable inhibitors of the kinase and phosphatase activities of these signaling enzymes. Such chemical tools can also augment experiments that involve genetic manipulation of PPIP5K expression; altering the expression of a PPIP5K may lead to many secondary genetic changes that complicate efforts to obtain a molecular understanding of the contribution of PPIP5K to the resulting phenotype. The development of PPIP5K inhibitors could therefore prove to be a useful future research direction. Beyond the application of inhibitors as research tools, there is the holy grail: development of therapeutically beneficial drugs that improve human health. The kinase activities of PPIP5Ks have been found to activate PtdIns(3,4,5)P3-AKT signaling pathways (Gokhale et al. 2013) and inflammatory responses to viral infection (Pulloor et al. 2014). Kinase inhibitors might be useful as anticancer agents and also for reducing inflammation.