Simple synthesis of 32P-labelled inositol hexakisphosphates for study of phosphate transformations

Background and aims In many soils inositol hexakisphosphate in its various forms is as abundant as inorganic phosphate. The organismal and geochemical processes that exchange phosphate between inositol hexakisphosphate and other pools of soil phosphate are poorly defined, as are the organisms and enzymes involved. We rationalized that simple enzymic synthesis of inositol hexakisphosphate labeled with 32P would greatly enable study of transformation of soil inositol phosphates when combined with robust HPLC separations of different inositol phosphates. Methods We employed the enzyme inositol pentakisphosphate 2-kinase, IP5 2-K, to transfer phosphate from [γ-32P]ATP to axial hydroxyl(s) of myo-, neo- and 1D-chiro-inositol phosphate substrates. Results 32P-labeled inositol phosphates were separated by anion exchange HPLC with phosphate eluents. Additional HPLC methods were developed to allow facile separation of myo-, neo-, 1D-chiro- and scyllo-inositol hexakisphosphate on acid gradients. Conclusions We developed enzymic approaches that allow the synthesis of labeled myo-inositol 1,[32P]2,3,4,5,6-hexakisphosphate; neo-inositol 1,[32P]2,3,4,[32P]5,6–hexakisphosphate and 1D-chiro-inositol [32P]1,2,3,4,5,[32P]6-hexakisphosphate. Additionally, we describe HPLC separations of all inositol hexakisphosphates yet identified in soils, using a collection of soil inositol phosphates described in the seminal historic studies of Cosgrove, Tate and coworkers. Our study will enable others to perform radiotracer experiments to analyze fluxes of phosphate to/from inositol hexakisphosphates in different soils. Electronic supplementary material The online version of this article (doi:10.1007/s11104-017-3315-9) contains supplementary material, which is available to authorized users.


Introduction
In consideration of the different forms of inositol hexakisphosphate identified in soils: 1D-chiro-, myo-, neo-and scyllo-( (Anderson 1964;Anderson and Malcolm 1974;Baker 1974, cited in Turner et al. 2002Cosgrove 1962Cosgrove , 1963Cosgrove , 1966Cosgrove , 1969aCosgrove and Tate 1963;Halstead and Anderson 1970;L'Annunziata 1975;L'Annunziata and Fuller 1971;L'Annunziata et al. 1972;reviewed, Cosgrove 1980); Irving and Cosgrove 1982), it remains unclear what the biotic or abiotic origins of D-chiro-, neoand scyllo-inositol phosphates are (L'Annunziata 2007; Turner and Richardson 2004;Turner et al. 2002). A limited number of studies have shown biotic contribution to the epimerization of unsubstituted inositols (Cosgrove 1969b;L'Annunziata 1975;Gonzalez 1977 (reviewed L'Annunziata 2007)), or have shown that chemical, and hence, geochemical, epimerization of myo-inositol pentakisphosphate is possible (Cosgrove 1972). It is clear from the foregoing that studies of soil phosphate transformations, particularly those arising from input of myo-inositol hexakisphosphate from plant sources, would be greatly enabled by the provision of 32 P or 33 P-labelled myo-inositol hexakisphosphate, and, indeed, of other inositol hexakisphosphates. Such materials would, with established extraction and separation techniques, allow facile determination of the exchange of phosphate between organic and inorganic pools and, with appropriate separation techniques, would allow study of transformations of different isomers of inositol hexakisphosphate. With labelled inositol hexakisphosphates and an increasing literature on the 'pathways' of myo-inositol hexakisphosphate degradation by phytases of different classes; cysteine phytase, histidine acid phytase, purple-acid phytase, βpropeller phytase (Konietzny and Greiner 2002), it would be possible to begin to describe 'pathways' of inositol hexakisphosphate turnover in soils and the contribution of different organisms to that turnover.
With these thoughts in mind, we have sought to synthesize 32 P-labeled inositol hexakisphosphates by enzymic means. We have taken opportunity of a recombinant inositol pentakisphosphate 2-kinase (IP5 2-K, also known as IPK1) characterized (Banos-Sanz et al. 2012;Gonzalez et al. 2010;Gosein and Miller 2013;Sweetman et al. 2006). This enzyme transfers the gamma-phosphate) from ATP to the sole axial hydroxyl on carbon 2 of myo-inositol phosphates, and is believed to be the enzyme responsible for synthesis of myoinositol hexakisphosphate in all kingdoms that make this molecule. We reasoned that the enzyme might be capable of transferring phosphate from ATP to the axial hydroxyl(s) of other inositols bearing equatorial phosphates.
The structures of the different inositol 'parents' of the inositol phosphates used in this study are shown in Fig. 1. The rules for numbering of carbon atoms, and hence of inositol phosphates bearing phosphate substituents on particular carbons are given in the IUPAC-IUB rules (IUPAC-IUB 1973, 1977. For myo-inositol phosphates, only, a relaxation of the rules (NC-IUB 1989) allows numbering of carbons by the D-(1D) or L-(1L) nomenclature, e.g. to assist in delineating metabolic sequences. The 1D-and 1L-numbering of myo-inositol is shown in Fig. 1. Supplemental Fig. 1 shows symmetry aspects of the 'parent' inositols of the myo-, neo-, 1D-chiro-and scyllo-inositol phosphates discussed hereafter, while Supplemental Fig. 2 shows the structures of the substrates and products obtained therefrom.
For the assistance of researchers wishing to adopt these methods, we provide (Table 1) a summary of the methodologies used with an indication of their applicability and a reference to prior use of the method. We also list the figures in this manuscript in which the method is applied.

Cloning and expression of AtIPK1
Arabidopsis thaliana IPK1 (AGI number): At5G42810, was cloned from Arabidopsis thaliana Col-0 cDNA with forward primer AAGTTCGTTTTCAGGGCCCG ATGGAGATGATTTTGGAGGAGAA and reverse p r i m e r A T G G T C TA G A A A G C T T TA G C T G TGGGAAGGTTTTG (vector specific sequence in italics) using Phusion High Fidelity Polymerase (Thermo Scientific). Purified product (Wizard SV Gel and PCR Cleanup System, Promega) was inserted into pOPINF linearized with HindIII and KpnI (Berrow et al. 2007) by ligation independent cloning using In Fusion HD enzyme kit (Clontech). The vector adds a Nterminal hexahistidine tag to the recombinant protein.
Recombinant plasmid was transformed into E coli Rosetta (DE3) (Novagen) and protein production induced from an overnight culture grown in LB containing 0.5% (w/v) glucose and ampicillin by transfer to LB containing 0.5 mM IPTG and ampicillin with further growth for 7 h at 25°C.

Enzyme assays
For assays without radiolabel, inositol phosphatedependent conversion of ATP to ADP was followed by reverse-phase ion pair HPLC and subsequent detection of nucleotides at 260 nm (Caddick et al. 2008). Briefly, 1 μg enzyme was incubated with 500 μM inositol phosphate and 50 μM ATP in 20 mM HEPES, pH 7.3, 1 mM MgCl 2 for 2 h at 25°C, the volume of the assay was 20 μL. Reactions were terminated by the addition of 1 μL of conc. HCl, followed after 5 min on ice by the addition of 50 μL water. Aliquots (50 μL) were analysed by HPLC. Fig. 1 The structures of myo-, neo-scyllo-and 1D-chiroinositols. The projections shown left to right are a Mills projection, a Haworth projection and a three-dimensional structure. Carbon atoms in the ring are numbered according to the IUPAC/IUPAC-IUB recommendations (1973,1977). For myo-inositol, both 1L-(upper) and 1D-(lower) notation is shown  This manuscript The structures of substrates and products of AtIP5 2 K Suppl. Fig. 1 This manuscript For assays with radiolabel, ATP-dependent conversion of inositol phosphates to higher (more phosphorylated) species was followed by anion-exchange HPLC with on-line detection of 32 P. Briefly, 1 μg enzyme was incubated with 500 μM inositol phosphate and 5 μM ATP in the presence of 0.37 MBq [γ-32 P]ATP in 20 mM HEPES, pH 7.3, 1 mM MgCl 2 for 1 h at 25°C, the volume of the assay was 20 μL. The reaction products were diluted with water and approximately 1-5% of the products were analysed by HPLC.

HPLC separation of inositol phosphates
Radiolabeled inositol phosphates were separated by anion exchange HPLC on Partisphere SAX columns eluted at a flow rate of 1 mL min −1 with a gradient derived by mixing solvent from reservoirs containing (A) water and (B) 1.25 M (NH 4 ) 2 HPO 4 , adjusted to pH 3.8 with H 3 PO 4 , according to the following schedule: time (min), % B; 0, 0; 5, 0; 65, 100. Radioactivity was detected by Cerenkov counting in a Radiomatic A500 Series Flo Detector (Canberra Packard, Pangbourne, Bucks, UK) fitted with a 0.5 ml flow cell using an integration interval of 12 s (Hanke et al. 2012).

HPLC separation of nucleotides
Aliquots of the products of enzyme assays were analysed according to (Caddick et al. 2008).

Nomenclature
For the purpose of this article, the term 'Ins' with prefix 1D-chiro-, myo-, neo-or scyllo-is used as an abbreviation of the described inositol phosphate. Hence, neoinositol 1,3,4,6-tetrakisphosphate is abbreviated neo-Ins(1,3,4,6)P 4 . It should be noted, however, that the numbering of phosphate substituents (of the carbon atoms to which they are attached) is not necessarily the same for different stereoisomers of inositol. The reader is referred to Shears and Turner (2007) for a concise description of terminology and to (Thomas et al. 2016) for a comprehensive review of inositol and inositol phosphate nomenclature and terminology.
AtIP5 2-K can be used to synthesize a range of 32 P-labeled inositol phosphates Having established that neo-Ins(1,3,4,6)P 4 and 1Dchiro-Ins(2,3,4,5)P 4 are novel substrates of AtIP5 2 K, we performed enzyme assays to produce 32 P-labelled inositol phosphates. The substrate concentrations used were 500 μM inositol phosphate and 5 μM ATP. Reactions were terminated and the products spiked with additional ATP to allow online tandem UVradioactivity monitoring of the chromatography (Fig. 3). We included myo-Ins(1,3,4,5,6)P 5 as the canonical substrate and observed (Fig. 3a) in addition to a major peak of myo-InsP 6 eluting at a retention time of 55 min, a small peak of unidentified material eluting at 38.8 min, a major peak of unreacted ATP at 25.2 min and a peak of inorganic phosphate at 14.6 min. We note that others have reported impurities in commercial [ 32 P]ATP that elute on Partisphere SAX columns with similar chromatographic mobility to myo-InsP 4 s (Stephens 1990).

Separation of inositol hexakisphosphates identified in soils
Using the post-column complexation method (Phillippy and Bland 1988) and a CarboPac PA-200 column eluted with a gradient of HCl, we were able to resolve in order of increasing retention time, neo-InsP 6 , muco-InsP 6 , 1D-chiro-InsP 6 , myo-InsP 6 and scyllo-InsP 6 (Fig. 5a), all samples obtained from the laboratory of the late Dennis Cosgrove. We assume that the muco-InsP 6 was that made by chemical phosphorylation of the inositol (Cosgrove 1975). The gradient shows a strongly sloping baseline arising from the acid eluent and this UV  BtMinpp and c A ficuum phytase were separated by Partisphere Strong Anion Exchange HPLC and radioactivity estimated by online Cerenkov counting. The positions of elution of representative classes of myo-inositol phosphate products and of contaminating ATP in the myo-[ 32 P]InsP 6 preparation are indicated. The difference in retention time of inositol hexakisphosphate between different panels reflects the use of different Partisphere SAX columns for the separations absorbance can be negated when using methanesulfonic acid as the eluent (Blaabjerg et al. 2010); however, at least up to 0.6 M, methanesulfonic acid was not a strong enough eluent to elute scyllo-InsP 6 from this column.

Separation of neo-inositol phosphates
Methanesulfonic acid was, however, suitable for separation of neo-InsPs: neo-InsP 4 s, neo-InsP 5 s and neo-InsP 6 (Fig. 5b), again, these compounds were likely produced by chemical phosphorylation of neo-inositol with sodium trimetaphosphate (Cosgrove 1969a). Irving (1980) reported that neo-InsP 6 was a good substrate of both the Pseudomonas (sp. unknown) bacterium (SB 2 ) phytase of Cosgrove (Cosgrove et al. 1970) and A. ficuum phytase (Irving and Cosgrove 1974). We note that there are nine possible neo-InsP 4 s, comprised of three pairs of enantiomers and three meso-compounds and there are three possible neo-InsP 5 s (Thomas et al. 2 0 1 6 ) . O f t h e n e o -I n s P 5 s , t w o : 1 L -n e o -Ins(1,2,3,4,5)P 5 = 1D-neo-Ins(1,2,3,5,6)P 5 and 1Dneo-Ins(1,2,3,4,5)P 5 = 1L-neo-Ins(1,2,3,5,6)P 5 are a pair of enantiomers, the third is the meso-compound neo-Ins(1,2,3,4,6)P 5 . In the absence of chiral HPLC methods for separating enantiomers of inositol phosphates, the separation of two peaks of neo-InsP 5 from the Cosgrove samples (Fig. 5b) is all that is achievable, but nevertheless could be diagnostic in studies of neo-InsP 6 transformation in soils. Separation of inositol pentakisphosphates; the initial products of phytase action on inositol hexakisphosphate Finally, we show (Fig. 6) the separation of a variety of InsP 5 s including the four separable peaks of myo-InsP 5 , identified in Fig. 6b by the position of the single hydroxyl; the two enantiomeric pairs 1D-1/3-OH and 1D-4/6-OH, indicated [1/3-OH] and [4/6-OH] on the figure, cannot be separated into individual enantiomers on nonchiral HPLC. Figure 6b also shows the single scyllo-InsP 5 and several InsP 5 s present in a 1L-chiro-InsP 5 sample. There are three possible 1L-chiro-InsP 5 s and, similarly, three possible 1D-chiro-InsP 5 s. Minor chiro-InsP 5 peaks with common retention times were observed in the 1D-chiro-InsP 6 sample (Fig. 6a) and the 1L-chiro-InsP 5 sample (Fig. 6b), with two of the three co-eluting precisely with peaks in a 1L-chiro-InsP 5 sample (Fig. 6a). Clearly, this column/eluent combination has great resolving power for all the InsP 5 s expected of soil samples.

Discussion
Inositol phosphates are major components of total soil phosphate and commonly the dominant organic phosphates in these environments (Turner et al. 2002). Despite their abundance, the origins of neo-, 1D-chiro-and scyllo-inositol hexakisphosphates in soils are poorly defined. It is plausible that they derive from the considerable inputs to soil of plant-derived myo-inositol Fig. 6 HPLC separation of soil-representative inositol pentakisphosphates. A Dionex CarboPac PA200 column was eluted with a gradient of methanesulfonic acid. Inositol phosphates were obtained from the laboratory of the late DJ Cosgrove. a The traces from individual injections of different inositol phosphate samples: a 1D-chiro-InsP 6 sample with three 1D-chiro-InsP 5 s, a 1L-chiro-InsP 5 sample with two 1L-chiro-InsP 5 s and a neo-InsP 5 sample. b The traces from individual injections of different inositol phosphate samples: a 1L-chiro-InsP 6 sample with three 1L-chiro-InsP 5 s, a scyllo-InsP 5 sample and an acid hydrolysate of myo-InsP 6 with all four resolvable myo-InsP 5 . a, b Traces are offset on the Y-scale (left or right) to aid visualisation hexakisphosphate, representing up to several percent of seed dry weight (Raboy 2003). It has been estimated that the sequestration of phosphorus in the myo-inositol hexakisphosphate, phytate, reserves of cropped organs of plants is equivalent to the per annum application of phosphorus as fertilizer to soils (Lott et al. 2000). It is remarkable therefore that we do not know, how plantderived phytate is, likely, converted to other forms (epimers) of inositol phosphates, nor whether the processes are organismal or geochemical.
The epimerization of unsubstituted inositols by unsterilized soil is well described (L'Annunziata and Gonzalez 1977) and pathways by which myo-inositol and scyllo-inositol are converted have been engineered in Bacillus subtilis (Kang et al. 2017;Tanaka et al. 2013;Yamaoka et al. 2011). Similarly, Geobacillus kaustophilus HTA426, has three dehydrogenases that are capable of acting as epimerases to interconvert myo-, scyllo-, and 1D-chiro-inositols (Yoshida et al. 2012). The American cockroach, Periplaneta americana, was reported to possess epimerase activity producing neo-inositol. (Hipps et al. 1973). Plants have the ability to epimerize a number of methylated inositols, reviewed (Thomas et al. 2016). These studies seem to suggest that epimerase activity is restricted to inositols lacking phosphate substituents. It is intriguing therefore that scyllo-phosphoinositides were detected in barley (Narasimhan et al. 1997) and that highly phosphorylated inositol phosphates and diphosphoinositol phosphates containing neo-rather than myo-inositol are the major form of inositol phosphate in Entamoeba histolytica (Martin et al. 2000), despite earlier suggestion to the contrary (Martin et al. 1993).
Because the axial 2-phosphate is the last phosphate added in the synthesis of myo-inositol hexakisphosphate in plants (Brearley and Hanke 1996) and Dictyostelium discoideum (Stephens and Irvine 1990) it is possible to use the enzyme catalyzing this step to make myo-inositol 1,[ 32 P]2,3,4,5,6-hexakisphosphate. Moreover, because most phytases of plant, bacterial or fungal origin do not remove this phosphate until very late in the sequence of myo-inositol hexakisphosphate degradation, if at all (Konietzny and Greiner 2002), the 32 P label will be retained in successive myo-InsP 5 , InsP 4 , InsP 3 and InsP 2 products of dephosphorylation. Consequently, addition of 32 P-labeled myo-InsP 6 to soils will, by simple chromatography using the methods elaborated here, allow researchers to study myo-inositol hexakisphosphate turnover in their soil of choice.
From a practical perspective, the use of a high energy β emitter such as 32 P allows facile detection of radioactivity in column eluates by on-line Cerenkov counting, obviating the requirement for the addition of scintillation fluid. The use of flow-detectors therefore allows for simple collection of radiolabelled fractions, albeit in high salt concentrations required to elute highly polar inositol phosphates. For phosphate eluents, it is a simple exercise to desalt the collected fractions on Dowex AG1 X8 resin with volatile ammonium formate / formic acid mixtures, subsequently removed by freeze-drying (Stephens 1990;Brearley et al. 1997), while for HCl eluents the HCl can be removed directly by freezedrying.
From a diagnostic perspective, because different phytases produce different myo-InsP 5 products (after all, this is the explanation of their classification e.g. as D3, Aspergillus; D6, E.coli or D5, lily pollen alkaline phytases (Konietzny and Greiner 2002)), simple analysis of products at the level of myo-InsP 5 will allow identification of the likely class of enzyme predominantly responsible for initial degradation of myo-InsP 6 in different soils. Indeed, it will be fascinating to correlate myo-inositol hexakisphosphate degradation products with metagenomic characterization of phytase and micoorganism abundance in different soils of the sort recently described (Neal et al. 2017).
In consideration of the routes of degradation of other inositol hexakisphosphates, the other labeled isomers that we describe will be of particular value. Perhaps unsurprisingly, we note the seminal work of Cosgrove (1969aCosgrove ( , 1970 and Irving and Cosgrove (1971) in characterization of products of dephosphorylation of myo-inositol-, scyllo-inositol-, and 1D-chiroinositol hexakisphosphate by a bacterial phytase, and similar studies of wheat phytase by Tate (1971, 1973) following the methods of Tomlinson and Ballou (1962). These works, reviewed (Irving 1980), show that the axial phosphates of myo-and 1D-chiro-inositol hexakisphosphates are the last to be removed by plant, bacterial and fungal phytases where tested. Similar conclusions can be drawn for fungal phytase action on 1L-chiro-inositol hexakisphosphate (Adelt et al. 2003). With these observations in mind, the use of the labeled InsP 5 and InsP 6 species and the powerful separation approaches identified in the foregoing should allow for facile assessment of the exchange of phosphate between different inositol phosphates in soil contexts.