Crystal Structure and Enzymology of Solanum tuberosum Inositol Tris/Tetrakisphosphate Kinase 1 (StITPK1)

Inositol phosphates and their pyrophosphorylated derivatives are responsive to the phosphate supply and are agents of phosphate homeostasis and other aspects of physiology. It seems likely that the enzymes that interconvert these signals work against the prevailing milieu of mixed populations of competing substrates and products. The synthesis of inositol pyrophosphates is mediated in plants by two classes of ATP-grasp fold kinase: PPIP5 kinases, known as VIH, and members of the inositol tris/tetrakisphosphate kinase (ITPK) family, specifically ITPK1/2. A molecular explanation of the contribution of ITPK1/2 to inositol pyrophosphate synthesis and turnover in plants is incomplete: the absence of nucleotide in published crystal structures limits the explanation of phosphotransfer reactions, and little is known of the affinity of potential substrates and competitors for ITPK1. Herein, we describe a complex of ADP and StITPK1 at 2.26 Å resolution and use a simple fluorescence polarization approach to compare the affinity of binding of diverse inositol phosphates, inositol pyrophosphates, and analogues. By simple HPLC, we reveal the novel catalytic capability of ITPK1 for different inositol pyrophosphates and show Ins(3,4,5,6)P4 to be a potent inhibitor of the inositol pyrophosphate-synthesizing activity of ITPK1. We further describe the exquisite specificity of ITPK1 for the myo-isomer among naturally occurring inositol hexakisphosphates.


■ INTRODUCTION
−15 The claim of the contribution of inositol pyrophosphates to aspects of plant biology rests heavily on the characterization of enzymes and the analysis of mutants thereof.ITPKs and PPIP5Ks, variously referred to as VIH or Vip1 in plants and yeast, respectively, contribute to inositol pyrophosphate synthesis.Both possess the ATP-grasp fold.PPIP5Ks contain an additional histidine acid phosphatase domain. 16he catalytic potential of plant ITPKs is particularly diverse 17−22 when compared with the kinase activity of VIH1/2, considered only to act on InsP 6 and 5-InsP 7 , 5-PP-Ins(1,2,3,4,6)P 5 , hereafter 5-PP-InsP 5 .For plants, VIH1/2 analysis has been restricted to isolated domains. 23,24Like other inositol phosphate kinases, IP6K 25 and IP5K (IPK1), 26 ITPKs are reversible phosphotransferases 21,22,27 as is the ATP-grasp kinase domain of PPIP5K. 28Consideration of reversibility under prevailing physiological conditions, with usually poorly defined nucleotide status, has the consequence that the causative signaling species among ITPK and VIH substrates and products are difficult to decipher.
−32 Herein, we have solved a crystal structure for a potato enzyme, StITPK1, in complex with ADP.We show the enzyme's preference for InsP 4 over InsP 6 and PP-InsP substrates and describe a simple, yet powerful, fluorescence polarization approach that could advance the study of other ATP-grasp kinases.

■ MATERIALS AND METHODS
Details of protein purification, enzyme assays, HPLC analysis of reaction products, ligand-binding assays, and X-ray crystallography can be found in the Supporting Information.

Biochemistry
scyllo-InsP 6 have been identified. 33Their presence is unexplained, but because plant matter is a major input to the soil we tested whether they might be substrates for StITPK1.Among isomers of inositol hexakisphosphate, StITPK1 phosphorylates the myo-isomer only (Figure S2A,D−F).The lack of activity toward other inositol hexakisphosphates described in the soil makes it unlikely that this ancestral plant enzyme, present in liverworts, bryophytes, and early vascular aquatic plants, 29 could contribute to the presence of noncanonical inositol pyrophosphates in soil, should they be found.

Biochemistry
as a master regulator of phosphate starvation response. 15,27he physiological balance of such reactions will be dependent on the concentrations of nucleotides and inositol pyrophosphosphates.

Ligand-Binding Assay Allows Comparison of Binding of Diverse Inositol Phosphates and Inositol Pyrophosphates to ATP-Grasp Kinases.
To determine the relative strengths of binding of different inositol phosphate and inositol pyrophosphate substrates of StITPK1, we undertook fluorescence polarization experiments with 2-FAM-InsP 5. 35 This molecule has proved a useful probe of the active and/or inositol phosphate-binding sites of enzymes as diverse as SHIP2 36 and IP5K (AtIPK1), 37 and HDAC complexes. 35A saturation curve for the binding of 2-FAM-InsP 5 to StITPK1 is shown in Figure S5, and displacement curves are shown for diverse inositol phosphates in Figure 2. The structures of these compounds are shown in Figure S1.
We rationalize the failure of StITPK1 to phosphorylate neo-InsP 6 (Figure S2D), with its two axial phosphates in the plane of symmetry, equivalent to the axial 2-phosphate and equatorial 5-phosphate of myo-InsP 6 (Figure S1), as being consistent with the requirement for an equatorial 5-phosphate in the single myo-InsP 5 substrate, Ins(1,2,3,4,5)P 5 .The failure of StITPK1 to phosphorylate scyllo-InsP 6 (Figure S2F), a C2epimer of myo-InsP 6 with six equatorial phosphates (Figure S1), suggests a requirement for phosphokinase substrates to possess a single axial phosphate in the plane of symmetry (for myo-InsP 6 , in the 2-position).Consistent with this, D-chiro-InsP 6 , which has two axial phosphates in trans, was also not a substrate (Figure S2E).
Nucleotide-Liganded Structure of StITPK1 Allows Modeling of Phosphotransfer Reactions of Plant ITPK1.Crystal structures have been reported for two plant ITPKs, AtITPK4 (PDB: 7PUP) and ZmITPK1 (PDB: 7TN8). 22,38The former lacks bound inositol phosphate, and the latter lacks bound nucleotide.To explain the interaction of ITPK1 with nucleotide cosubstrate, the StITPK1 crystal structure (residues 8−320) was solved in space group C222 with a monomer of the enzyme in the asymmetric unit (Figure 3) (PDB 8OXE).Refined against all data to 2.26 Å resolution, the final structural model had an R-factor of 19.0% (Rfree 24.1%) (Table S1).−40 Relative to AtITPK4, ITPK1 possesses a "tether" insertion following the central subdomain, while AtITPK4 possesses a "tab" in the N-terminal subdomain. 22The tether comprises a polypeptide connection between the C-terminal β-strand of the central domain and a helix of the C-terminal domain running under the protein.This polypeptide lies across the top of the active site cavity, linking the two domains (Figure 3B,D).In all the ITPK1s of known molecular structure where this polypeptide is resolved, it provides residues that contribute to the ATP cofactor/substrate binding pocket.Consistent with the absence of the resolved nucleotide, this region is disordered in the crystal structure of ZmITPK1, 38 but in StITPK1, it is stabilized in most parts by adventitious interactions with a neighboring copy of the molecule in the crystal lattice.Both the tab and tether insertions help shape the active site cleft in plant ITPK1 and ITPK4, suggesting that they may contribute to differential substrate recognition (Figure 3B,C).As for other ITPK1 enzymes, the active site of StITPK1 is narrow, unlike the more open active site in AtITPK4 (Figure S7), and features a highly positively charged active site (Figure 3D).
To explain the preference of StITPK1 for its substrates we modeled the binding of the Ins(1,4,5,6)P 4 and Ins(3,4,5,6)P 4 enantiomers to StITPK1, adopting the consensus specificity subsite nomenclature. 39Briefly, subsite A is the site of phosphoryl transfer and constitutes the catalytic center.For Ins(1,4,5,6)P 4 (Figure 4A), substituents on locants 3, 2, 1, 6, 5, and 4 of the myo-inositol ring occupy sites A, B, C, D, E, and F, respectively, while for Ins(3,4,5,6)P 4 (Figure 4B), substituents on locants 1, 6, 5, 4, 3, and 2 occupy sites A, B, C, D, E, and F, respectively.Residues forming polar interactions with the hydroxyl and phosphate groups of the substrates in the relaxed models are summarized (Table S2).The predicted pose of the poor substrate, Ins(1,4,5,6)P 4 , lacks polar interactions in the Band C-subsites (Figure 4A), while the strong substrate, Ins(3,4,5,6)P 4 , enjoys polar interactions in all subsites except F, occupied by the 2-hydroxyl group (Figure 4B).In the Bpocket, the 6-phosphate of Ins(3,4,5,6)P 4 is predicted to interact with the side chain of Asn272.The substitution of a conserved glycine residue at this site in the ITPK4s (Gly437 in AtITPK4) may help explain the poor activity of AtITPK4 toward this potential substrate.If accurately predicted, these interactions in the B-and C-subsites are likely crucial for enantiospecific hydroxy-kinase activity by StITPK1 toward inositol tetrakisphosphates.
A model for a stereochemically productive complex of InsP 6 with StITPK1 derived by molecular docking is shown in Figure S8.All residues observed to interact with InsP 6 in its complex with the maize enzyme 38 are conserved in StITPK1.However, due to the lack of bound nucleotide in the crystal structure of ZmITPK1, the central domain is displaced relative to that seen in the potato enzyme structure and InsP 6 binds in such a way that in-line phosphoryl transfer from the γ-phosphate of ATP is implausible. 39,41It, therefore, appears that the absence of a nucleotide from the structure of the complex of ZmITPK1 with InsP 6 leads to a situation where an unproductive binding mode is stabilized.On the other hand, in the pose of InsP 6 predicted for StITPK1, InsP 6 binds with its "receiving" 5-phosphate in Biochemistry the F-specificity subsite (Figure S8), contrasting with the "receiving" 1-hydroxyl of Ins(3,4,5,6)P 4 which occupies subsite A (Figure 4).The minimum distance in this pose from the γphosphate phosphorus atom of ATP to the 5-phosphate oxygen of the substrate is a little over 3.5 Å, rendering plausible in-line phosphoryl transfer from the γ-phosphate of ATP to generate the observed 5-PPInsP 5 product.Again, polar contacts to the ligand are made by Q224 of the tether region; thus, the bound nucleotide may stabilize the central subdomain and a portion of the tether, enabling recognition of InsP 6 and catalysis.In ZmITPK1, the tether is part of a catalytic specificity element that sanctions phosphokinase activity against InsP 6 . 38The absence of the tether polypeptide in ITPK4 would then be consistent with its inability to synthesize inositol pyrophosphates. 20,22P-InsP Analogues Confirm the Stereospecificity of StITPK1 for Phosphorylation of PP-InsPs.For both StITPK1 and AtITPK1, InsP 6 was the strongest phosphokinase substrate, with 3-PP-InsP 5 being a stronger substrate than 1-PP-InsP 5 , whereas 5-PP-InsP 5 was not a substrate (Figures 1  and S4, Table S3).Nevertheless, the three PP-InsP 5 s tested showed similar K i values for displacement of 2-FAM-InsP 5 from StITPK1, in the range 88−178 nM (Figure S9A−C).Overall, PP-InsP 5 s displayed similar K i values to InsP 6 isomers (cf.Figures 2 and S9).We also tested a range of PP-InsP analogues as phosphokinase substrates (Figure 5).Interestingly, both 3-PCP-Ins(1,2,4,5,6)P 5 , hereafter 3-PCP-InsP 5 , and 1-PCP-Ins(2,3,4,5,6)P 5 , hereafter 1-PCP-InsP 5 , were substrates.They yielded products, with identical retention times, that we assume to be 3-PCP-, 5-PP-Ins(1,2,4,6)P 4 and 1-PCP-, 5-PP-Ins(2,3,4,6)P 4 respectively (Figure 5A,B).AtITPK1 also displays the same preference for enantiomers of PCP-InsP 5 analogues (Figure S4).

30,31
We further suggest that enhanced accumulation of Ins-(3,4,5,6)P 4 in itpk1 mutants 9,15,27 and ipk1 mutants 4,15 amplifies reductions in PP-InsP species 9,15,27 that are widely reported to be the agents of much physiology. 3Indeed, itpk1 and ipk1 mutants show a constitutive phosphate starvation response. 4,14,15DISCUSSION Our understanding of inositol pyrophosphate function rests heavily on the molecular genetic disruption of hydroxy-kinase and phosphokinase activities.These have pleiotropic influence on plant physiology, reflecting involvement in processes as diverse as pathogen resistance, symbiosis, phosphate starvation response, and the action of plant growth regulators auxin and jasmonate.Disruption also has a multifaceted effect on inositol phosphate metabolism, with impacts on "lower" and "higher" inositol phosphates and inositol pyrophosphates alike.This is particularly apparent for ITPKs 9,15,27 and IP5K (IPK1). 4,15blation of AtITPK1 increases lower inositol phosphates 9,15,27 and reduces PP-InsP 5 and [PP] 2 -InsP 4 alike, in plants.While it has been widely assumed that [PP] 2 -InsP 4synthesizing activity belongs exclusively to VIH1/2, 3,6,15,23,24,42 until the recent detailed description of noncanonical PP-InsP 5 species 9,21,27 little consideration had been given to other possibilities.Similarly, rather little consideration has been given

Biochemistry
to the possibility that different inositol phosphates and inositol pyrophosphates are active site competitors of ATP-grasp kinases.With respect to the latter premise, the data presented here suggests that competition could be significant, and the principle likely applies to other ATP-grasp kinases.In respect of the former premise, the null hypothesis, that ITPK1 does not contribute directly (i.e., other than by provision of the 5-PP-InsP 5 substrate to VIH1/2) to [PP] 2 -InsP 4 levels, is challenged.8,43 It is, perhaps, instructive, therefore, to compare the relative activities of different ATP-grasp kinase proteins of plants. AtVIH2 has been charactized as separate kinase and phosphatase domains.24 The former shows 1-phosphokinase activity against InsP 6 and 5-PP-InsP 5 with turnover numbers of ∼0.4 and ∼1 (min −1 ), respectively.24 These constants were derived under assay conditions very similar to those employed in the present and earlier studies of AtITPK1.21,27 K cat for AtITPK1's phosphokinase activity against InsP 6 , calculated from data, 20,21,27 is comparable, ∼1.5 min −1 , here, for StITPK1, K cat ∼1.14 min −1 at pH 6.5.K cat for 5-PP-InsP 5 -driven ATP synthesis is greater.27 Clearly, the energy charge acts through ITPK1 and other ATP-grasp kinases to modulate the levels of PP-InsP 5 and [PP] 2 -InsP 4 that are widely reported to influence plant physiology.Indeed, genetic evidence shows that despite these very low K cat values for both AtITPK1 and AtVIH2, ablation of AtITPK1 markedly reduces the levels of PP-InsP 5 and [PP] 2 -InsP 4 in plants, 9,27 while ablation of VIH1/2 markedly reduces [PP] 2 -InsP 4. 8,24 The low cellular levels of these species, relative to InsP 6 , typically less than 1%, perhaps implies that the bulk of cellular InsP 6 is accessible to AtITPK1. The dfference in rate constants for 5-pyrophosphorylation and hydroxy-kinase activities may arise from poorer geometry for the former, with the placement of the receiving 5phosphate or hydroxyl (for the latter) in different enzyme subsites.

■ CONCLUSIONS
The catalytic flexibility of the ATP-grasp fold kinase ITPK1 has been extended to include inositol pyrophosphates considered previously to be substrates/products of only VIH1/2, among kinases.Detailed analysis of the binding of diverse substrates and substrate analogues to StITPK1 was enabled by the use of fluorescence polarization assays.This assay may find use for other ATP-grasp kinases.This and the solution of a nucleotideliganded crystal structure for StITPK1 offer an opportunity for more precise elaboration of inositol pyrophosphate function in plants, one that accommodates competition by substrates/ inhibitors for cognate partners such as TIR1, COI1-ASK and COP signalosome components and enzymes alike.Plant ITPK1 also offers an opportunity for the study of ATP-grasp kinases in pyrophosphate biology contexts by provision of probes of phosphotransfer.One envisages the simple incorporation of 32 P or 33  The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.biochem.3c00404.

Figure 1 .
Figure 1.StITPK1 is a reversible inositol pyrophosphate-ADP phosphotransferase.HPLC resolution of products of 12h reaction of StITPK with ADP and (A) 1,5-[PP] 2 -InsP 4 ; (B) rac-1,5-[PP] 2 -InsP 4 .Products of the dephosphorylation of 1,5-[PP] 2 -InsP 4 coelute with 1-PP-InsP 5 , (C).Substrates are indicated, S and products, P. Chromatograms of reactions without enzyme are shown in red, with enzyme in black and chromatograms of standards are shown in blue.The position of elution of ATP formed by phosphotransfer to ADP is shown in panel A. ADP elutes in the solvent front.The standards showing the elution position of PP-InsP 5 in C contain ATP.The ATP peaks in all other panels are products of phosphotransfer from substrate to ADP.HPLC of products of reaction of StITPK1 with ATP and 1-PP-InsP 5 , 3-PP-InsP 5 , or 5-PP-InsP 5 are shown in D, E and F. Here, chromatograms of 3 h incubations with enzyme are shown in black, and 12 h incubations are shown in blue.The position of a 1,5-[PP] 2 -InsP 4 standard (trace offset on the y-axis) is shown in panel D. The HPLC column was eluted with a gradient of HCl.

Figure 3 .
Figure 3. Overview of the crystal structure of StITPK1.(A) Cartoon representation of the structure of StITPK1, colored by secondary structure (αhelix red, β-sheet yellow, and coil green).Broken lines in the backbone trace indicate residues that were unresolved in the model due to disorder.Bound nucleotide is shown in stick format with coloring as follows: carbon-green, oxygen-red, nitrogen-blue, and phosphorus-orange.(B) Molecular surface representation of the structure of StITPK1 colored by subdomain.Subdomains are the kinase N-terminal domain (light blue), kinase central domain (lime green) and kinase central domain (sand).Here and in panel D, the polypeptide connecting the central and C-terminal domains is colored dark blue.Bound ADP is shown in atom sphere format.(C) Molecular surface representation of the structure of AtITPK4 (PDB: 7PUP).Coloring as in panel C except that the additional HAD domain found in this enzyme is colored pink and the tab insertion unique to ITPK4s is colored magenta.(D) Molecular surface of StITPK1 colored by electrostatic potential (red-acidic, blue-basic).The orientation of the molecule is the same as that in panel A. In panels B−D, bound ADP is shown in atom sphere format and colored according to that in panel A.

Figure 4 .
Figure 4. Prediction of the binding modes of an enantiomeric pair of substrates in the active site of StITPK1.(A) Closeup view of the energy minimized predicted binding mode of the poor substrate, Ins(1,4,5,6)P 4 , in the kinase domain active site.The enzyme is shown in the cartoon format and colored green.The substrate and active site residues (labeled) with which it forms polar interactions are shown in the stick format with carbon colored green, oxygen red, nitrogen blue, and phosphorus orange.Magnesium ions are shown as dark green spheres.Polar interactions are indicated by black dashed lines.Specificity subsites are labeled A−F (magenta font) such that the hydroxyl group positioned to accept the γ-phosphate of ATP by inline transfer (the hydroxyl attached to carbon 3 of the inositol ring, in this case) occupies subsite A and the remaining subsites are arrayed in an anticlockwise sense when observed from the viewpoint adopted in this figure.(B) View of the energy minimized predicted binding mode of the good substrate, Ins(3,4,5,6)P 4 , in the kinase domain active site.The hydroxyl attached to carbon 1 of the inositol ring, in this case, occupies subsite A. Display format and coloring as in panel A.

Figure 6 .
Figure 6.Inhibition of StITPK1 phosphokinase activity by inositol pyrophosphate analogues.(A) Ins(1,2,3,4,5)P 5 5-phosphokinase activity; (B) InsP 6 5-phosphokinase activity.Reactions were performed for 2 h at 30 °C with 3 μM StITPK1, 0.5 mM ATP, and 1 mM inositol phosphate in the absence or presence of competitor.The extent of inhibition at inhibitor concentration (mM) indicated by number in each column was estimated from the integrated peak areas of substrate and product peaks resolved by HPLC (example HPLC traces are shown in Figure S10).Significant difference at p = 0.05 between inhibitor treatments (by one-way ANOVA and Tukey's multiple comparisons test) is indicated by the absence of a common letter.