Determination of neo- and D-chiro-inositol hexakisphosphate in soils by solution 31P NMR spectroscopy.

The inositol phosphates are an abundant but poorly understood group of organic phosphorus compounds found widely in the environment. Four stereoisomers of inositol hexakisphosphate (IP(6)) occur, although for three of these (scyllo, neo, and D-chiro) the origins, dynamics, and biological function remain unknown, due in large part to analytical limitations in their measurement in environmental samples. We synthesized authentic neo- and D-chiro-IP(6) and used them to identify signals from these compounds in three soils from the Falkland Islands. Both compounds resisted hypobromite oxidation and gave quantifiable (31)P NMR signals at δ = 6.67 ppm (equatorial phosphate groups of the 4-equatorial/2-axial conformer of neo-IP(6)) and δ = 6.48 ppm (equatorial phosphate groups of the 2-equatorial/4-axial conformer of D-chiro-IP(6)) in soil extracts. Inositol hexakisphosphate accounted for 46-54% of the soil organic phosphorus, of which the four stereoisomers constituted, on average, 55.9% (myo), 32.8% (scyllo), 6.1% (neo), and 5.2% (D-chiro). Reappraisal of the literature based on the new signal assignments revealed that neo- and D-chiro-IP(6) occur widely in both terrestrial and aquatic ecosystems. These results confirm that the inositol phosphates can constitute a considerable fraction of the organic phosphorus in soils and reveal the prevalence of neo- and D-chiro-IP(6) in the environment. The hypobromite oxidation and solution (31)P NMR spectroscopy procedure allows the simultaneous quantification of all four IP(6) stereoisomers in environmental samples and provides a platform for research into the origins and ecological significance of these enigmatic compounds.

Introduction

A considerable proportion of the phosphorus in many soils occurs as stereoisomers of inositol hexakisphosphate (IP6).[1,2] The most abundant of these is myo-IP6, which occurs as a phosphorus storage compound in seeds,[3] although there is evidence that three additional stereoisomers also occur in the environment.[4] In a remarkable series of papers published in the 1960s, Dennis Cosgrove identified scyllo-, neo-, and d-chiro-IP6 in soil extracts using a combination of alkaline extraction, column chromatography, dephosphorylation, and paper chromatography of the free inositols.[5−7] Subsequent studies have detected the stereoisomers of IP6 in a number of soils and sediments,[8−12] yet they are extremely rare elsewhere in nature: scyllo-IP6 has never been identified in an organism,[4] while neo- and d-chiro-IP6 have been identified only in human intestinal amoebae[13] and velvet mesquite leaves,[14] respectively. A report of muco-IP6[14] was criticized on analytical grounds,[1] and this compound does not appear to exist in nature. The remaining four inositol stereoisomers (allo, l-chiro, cis, epi) do not appear to occur naturally in phosphorylated forms.

Given that the origins of the stereoisomeric forms of IP6 remain unknown, it is perhaps not surprising that the pathways to their synthesis, biological function, and behavior in the environment are also unknown. Each of the three stereoisomers differs from the myo form in the orientation of a single phosphate group, so it is possible that they are synthesized by epimerization of myo-IP6 as well as by phosphorylation of the base isomer.[15] It is well-known that myo-IP6 reacts strongly with metal ions to form insoluble complexes,[1] and the other stereoisomeric forms presumably react in a similar manner. The inositol phosphates are therefore often considered to be relatively recalcitrant in the environment.[2] Despite this, a number of organisms can synthesize phytase enzymes that allow them to acquire phosphorus from myo-IP6,[16,17] although scyllo-, neo-, and d-chiro-IP6 are all more resistant to phytase hydrolysis than the myo isomer.[18] Taken together, these properties suggest that possible functions of the stereoisomers might include protection against metal toxicity, conservation and protection of phosphorus from competitors, or an abiotic role in soil structure.[4]

Despite the widespread occurrence of inositol phosphates in the environment, the stereoisomers of IP6 other than myo have been virtually unstudied for the previous 40 years in any field (but see refs (19 and 20)), including biomedicine, where the inositol phosphates are of particular interest for their role in cell signaling.[21] This is unsatisfactory given the importance of the inositol phosphates in the phosphorus cycle in both terrestrial and aquatic ecosystems but is due in large part to analytical limitations inherent in the measurement of inositol phosphates in environmental samples.[2] The study of organic phosphorus in such samples has been improved dramatically by the application of solution 31P NMR spectroscopy.[22] This technique has been adopted widely because it provides information on phosphorus compounds in alkaline extracts of soils or sediments without the need for the complex extraction and sample clean up procedures involved in conventional chromatography.[23] However, solution 31P NMR spectroscopy has been applied only recently to the quantification of IP6 in environmental samples: a technique using spectral deconvolution was used to quantify myo-IP6,[24] while a procedure involving hypobromite oxidation, which destroys all organic matter except the inositol phosphates, and spiking with authentic compounds was used to identify scyllo-IP6.[12]

Despite these advances, it has not been possible to identify neo- and d-chiro-IP6 by solution 31P NMR spectroscopy, due principally to the difficulty in obtaining authentic samples of these compounds. Here we report the chemical synthesis of neo- and d-chiro-IP6, the first efficient and high-yielding route to the latter, and their identification in soil extracts by hypobromite oxidation and solution 31P NMR spectroscopy. Our results confirm the presence of all four stereoisomers of IP6 in soils (i.e., myo, scyllo, neo, and d-chiro) and, by enabling a re-evaluation of the literature on solution 31P NMR spectroscopy of soils and sediments, reveal the widespread occurrence of neo- and d-chiro-IP6 in both terrestrial and aquatic ecosystems.

Materials and Methods

Synthesis of d-chiro-Inositol 1,2,3,4,5,6-Hexakisphosphate

Bis(cyanoethyl)(N,N-diisopropylamino)phosphine (2.40 g, 9.00 mmol) was added to a stirred suspension of d-chiro-inositol (180.2 mg, 1.00 mmol) and 5-phenyltetrazole (1.75 g, 12.00 mmol) in anhydrous dichloromethane (10 mL) under an argon atmosphere. Stirring continued for 20 h at room temperature. The reaction mixture was cooled to −40 °C, and tert-butyl hydroperoxide (70%, 1.72 mL, 12.00 mmol) was added portion-wise while stirring. The mixture was allowed to warm to room temperature and stirred for a further 1 h. After removal of solvent under reduced pressure, the residue was redissolved in dichloromethane (100 mL) and washed with a 10% sodium sulfite solution (2 × 100 mL), the organic phase was dried (MgSO4), and the solvent was concentrated under reduced pressure to afford the crude 1,2,3,4,5,6-hexakis-O-[bis(cyanoethyloxy)phosphoryl] d-chiro-inositol. Without further purification, the crude material was dissolved in concentrated aqueous ammonia solution (30 mL), and the mixture was heated at 60 °C overnight in a Pyrex pressure tube. After evaporation of the solution under vacuum, the residue was purified by ion exchange chromatography to afford the pure triethyl ammonium salt of d-chiro-inositol 1,2,3,4,5,6-hexakisphosphate (1.06 g, 91% from d-chiro-inositol) as a hygroscopic white solid[25] (analytical information presented in the Supporting Information).

Synthesis of neo-Inositol 1,2,3,4,5,6-Hexakisphosphate

Authentic neo-IP6 was synthesized by a similar procedure to that described above for d-chiro-IP6 (see the Supporting Information).

Preparation of Authentic Compounds for Solution 31P NMR Spectroscopy

To determine chemical shifts of neo- and d-chiro-IP6 in NaOH–EDTA solution by solution 31P NMR spectroscopy, compounds (5 mg) were dissolved in deionized water (4 mL) to yield solutions containing ∼200 μg P mL–1. Of this, 3 mL was adjusted to pH >13 by addition of 10 M NaOH, while 1 mL was pretreated by hypobromite oxidation (see below). Both samples were then analyzed by solution 31P NMR spectroscopy, and the unbrominated samples were used to prepare a spike solution (see below).

Soils and Phosphorus Extraction

We studied three soils from the Falkland Islands known to contain inositol phosphates (B. L. Turner, unpublished data). The soils were from three distinct locations and were all under pasture consisting mainly of white grass (Cortaderia pilosa) with some Christmas bush (Baccharis magellanica, an evergreen shrub). The soils were rich in organic matter (11–23% total C), moderately acidic (pH in water 5.2–5.4), and contained relatively high phosphorus concentrations (753–1107 mg P kg–1) (see Table S1 in the Supporting Information). Soils were extracted in a solution containing 0.25 M NaOH and 50 mM disodium EDTA (ethylenediaminetetraacetate) for 16 h in a 1:20 solid/solution ratio at 22 °C, centrifuged (8000 g, 30 min), lyophilized, and ground to a fine powder. An aliquot of each extract was taken prior to lyophilization for determination of total phosphorus by inductively coupled plasma optical-emission spectrometry (ICP–OES).

Hypobromite Oxidation

Solutions containing authentic inositol phosphates and soil NaOH–EDTA extracts were analyzed by solution 31P NMR spectroscopy before and after treatment by hypobromite oxidation (also known as alkaline bromination). The method was adapted from previous studies[12,26,27] and is reported in full in the Supporting Information. Briefly, samples were made strongly alkaline and treated with ice-cold pure bromine. After 1 h at room temperature the mixture was boiled (5 min), acidified to pH <3 (accompanied by a color change from yellow to orange), and reheated to dispel excess bromine (color change from orange to yellow), and the pH was adjusted to 8.5 (color change from yellow to colorless). Phosphates were precipitated with barium acetate, resuspended using cation exchange resin (Amberlite IR120, hydrogen form; Sigma-Aldrich, St Louis, MI), and the pH adjusted to >13 with 1 M NaOH. Solutions containing authentic compounds were analyzed directly, while soil extracts were frozen and lyophilized for later analysis.

Addition of Authentic Inositol Phosphates to Soil Extracts

Two solutions were prepared containing authentic inositol phosphates. These were added as spikes to soil extracts to identify individual signals. The first spike solution contained myo- and scyllo-IP6, while a second spike solution contained neo- and d-chiro-IP6. Spike solutions were prepared by combining 0.25 mL of each of the authentic compound solutions with 0.5 mL of a solution containing 1.0 M NaOH and 100 mM EDTA. For analysis, 0.25 mL of spike solution was added to 0.75 mL of reconstituted soil extract (see below) and 1 mL of 1.0 M NaOH and 100 mM EDTA solution. The spike was therefore added at relatively similar concentrations of both neo- and d-chiro-IP6 compared to soil extracts.

Solution 31P NMR Spectroscopy

Samples were analyzed by solution 31P NMR spectroscopy using a procedure similar to that described previously,[12,28] with details reported in the Supporting Information. Briefly, compounds were redissolved in D2O and the 1 M NaOH–100 mM EDTA solution, vortexed (1 min), and transferred to a 10 mm NMR tube. Soil extracts were filtered (1 μm GF-B filter, Whatman) prior to analysis. Spectra were acquired on a Bruker Avance 500 using a 30° pulse, 0.58 s acquisition time, and a 2 s T1 delay. These parameters yield quantitative spectra based on literature reports (e.g., ref (29)). The number of scans varied from ∼1000 for authentic compounds to ∼30,000 for soil extracts to ensure acceptable signal-to-noise ratios. Chemical shifts of signals were determined in parts per million (ppm) relative to an external orthophosphoric acid standard (85%) set to δ = 0 ppm. Signals were subsequently adjusted using the chemical shift of scyllo-IP6 (δ = 4.03 ppm) to facilitate signal identification among spectra and assigned based on literature reports of model compounds spiked into NaOH–EDTA soil extracts.[30] Signal areas were calculated by integration, and deconvolution was performed on the region between δ = 3.0 and 7.5 ppm to separate orthophosphate from phosphomonoesters and to quantify signals from inositol phosphates (see below). Finally, spectra were plotted with 1 Hz line broadening to show fine resolution in the phosphomonoester region.

Quantification of Inositol Hexakisphosphate Stereoisomers

The proportion of the total spectral area was determined for the four IP6 stereoisomers (myo, scyllo, neo, d-chiro). The proportion of scyllo-IP6 was determined from the strong signal at δ = 4.03 ppm. The proportion of myo-IP6 was determined by summing the four signals in a 1:2:2:1 pattern (see below for chemical shifts). The proportion of neo-IP6 was determined by multiplying the signal at δ = 6.67 ppm by 1.5 (i.e., the signal represents four of the six phosphates on the molecule). The proportion of d-chiro-IP6 was determined by multiplying the signal at δ = 6.48 ppm by three (i.e., it represents two of the six phosphates on the molecule). We assumed that all of the neo- and d-chiro-IP6 molecules were present as a single conformer in soil extracts, given the absence of detectable signals from the alternate conformers in spectra of brominated samples (see below).

Results

Signals from neo- and d-chiro-Inositol Hexakisphosphates in NaOH–EDTA

Both neo- and d-chiro-IP6 can occur in two conformational forms in solution (Figure 1), and both forms were present in preparations of the authentic samples in alkaline solution (Figure 2). The authentic sample of neo-IP6 in NaOH–EDTA gave four signals in solution 31P NMR spectroscopy (Figure 2A). Two signals at δ = 6.61 and δ = 4.54 ppm in a 4:2 ratio was assigned to the 4-equatorial/2-axial conformation, based on 1H, 13C, and 2-dimensional NMR experiments (see the Supporting Information). Two smaller signals at δ = 4.88 and δ = 5.14 ppm, also in a 4:2 ratio, were assigned to the 2-equatorial/4-axial conformation (Figure 2A). The ratio of the two conformers was 6:1. The chemical shifts of the signals were essentially unchanged by hypobromite oxidation (Figure 2B), but the ratio of the two conformers was reduced to 2:1.

Figure 1

Conformational changes in neo- and d-chiro-inositol hexakisphosphate (IP6) depending on solution pH. Phosphate groups are indicated by ‘P’, with the number of the associated carbon group indicated by the subscript numeral. At low pH both neo- and d-chiro-IP6 adopt the 4-equatorial/2-axial conformation. At high pH, the 2-equatorial/4-axial and 4-equatorial/2-axial forms exist in equilibrium, and the relative proportions of the two conformers might be influenced by factors other than pH (e.g., metal counterions). In 31P NMR spectra of alkaline soil extracts, neo-IP6 was observed (within the limits of detection) to be exclusively in the 4-equatorial/2-axial conformation, while d-chiro-IP6 occupied the 2-equatorial/4-axial conformation.

Figure 2

Solution 31P NMR spectra of authentic neo- and d-chiro-inositol hexakisphosphate dissolved in NaOH–EDTA. Spectra show samples before (A, C) and after (B, D) pretreatment by hypobromite oxidation. Spectra are plotted with 1 Hz (A–C) or 2 Hz (D) line broadening.

The authentic sample of d-chiro-IP6 gave six signals in solution 31P NMR spectroscopy (Figure 2C). Three large signals at δ = 4.62, 5.04, and 6.44 ppm in a 2:2:2 pattern were assigned to the 2-equatorial/4-axial conformation, based on 1H, 13C, and two-dimensional NMR experiments (see the Supporting Information). A second set of smaller signals at δ = 4.33, 5.61, and 6.85 ppm, also in a 2:2:2 pattern, were assigned to the 4-equatorial/2-axial conformation. The ratio of the two sets of signals was 11:1. The chemical shifts of signals from d-chiro-IP6 were essentially unchanged following hypobromite oxidation (Figure 2D), although poor resolution in the brominated sample due to the low phosphorus concentration prevented clear quantification of signals from the 4-equatorial/2-axial conformer.

In both cases, no inorganic phosphate was detected in brominated samples, confirming the resistance of neo- and d-chiro-IP6 to hypobromite oxidation.

Phosphorus Composition of Soil NaOH–EDTA Extracts and the Effect of Hypobromite Oxidation

Untreated NaOH–EDTA extracts of the three soils were similar in terms of their broad phosphorus composition (Table 1, Figure 3A). The NaOH–EDTA solution recovered ≥90% of the total phosphorus from the three soils (see Table S1 in the Supporting Information), consistent with the efficacy of this extraction solution for high latitude soils rich in carbon and phosphorus.[31−33] Most of the extracted phosphorus was in organic forms, including phosphomonoesters (66–70% extracted phosphorus), DNA (DNA; 5.2–7.5%), alkali-stable phospholipids (1.8–4.4%), and phosphonates (1.2–2.4%). Inorganic phosphorus included orthophosphate (17–20% extracted phosphorus) and pyrophosphate (1.4–2.4%).

Table 1

Phosphorus Composition of NaOH–EDTA Extracts of Three Soils Determined by Solution 31P NMR Spectroscopy with or without Pretreatment by Hypobromite Oxidationa

		% of total spectra area
		inorganic phosphorus		organic phosphorus
soil	pretreatment	orthophosphate	pyrophosphate	phosphomonoester	phospholipid	DNA	phosphonate
East 50	none	20.3	2.4	67.1	1.8	5.9	2.4
	bromination	42.3	0.7	57.0	0	0	0
West 18	none	18.8	1.4	66.4	4.4	7.5	1.5
	bromination	43.4	0.7	55.7	0	0	0
West 22	none	17.3	2.3	70.2	3.9	5.2	1.2
	bromination	47.2	2.6	50.2	0	0	0

Values are the proportion (%) of the total spectral area assigned to functional groups of phosphorus compounds.

Figure 3

Solution 31P NMR spectra of NaOH–EDTA extracts of a soil (East 50) from the Falkland Islands before (A) and after (B) pretreatment with hypobromite oxidation to destroy all organic phosphorus other than the higher-order inositol phosphates. The zoomed insets show the phosphonate signals between δ = 19 and 21 ppm. Spectra are plotted with 5 Hz line broadening. In (B) the spectrum is truncated vertically to show the scyllo-inositol hexakisphosphate signal at the same height in both spectra. Signal assignments are as follows: a, neo- and d-chiro-inositol hexakisphosphate; b, inorganic orthophosphate; c, phosphomonoesters other than the signals from neo- and d-chiro-inositol hexakisphosphate between δ = 6.4 and 6.9 ppm; d, alkali-stable phospholipids; e, DNA; f, pyrophosphate; g, phosphonates.

Hypobromite oxidation destroyed phospholipids, DNA, phosphonates, and some phosphomonoesters (Table 1, Figure 3B). Pyrophosphate also declined partially in two of the three soils. Inorganic orthophosphate increased to 42–47% of the total phosphorus following bromination, while phosphomonoesters declined to 50–57% (Table 1). The phosphomonoesters persisting after hypobromite oxidation presumably represent only inositol phosphates.[12,26]

Identification of Signals from Inositol Hexakisphosphate Stereoisomers in Soil Extracts

All four stereoisomers of IP6 were identified in extracts of the three soils by spiking solutions with the respective stereoisomers (Table 2, Figure 4). The presence of myo- and scyllo-IP6, their chemical shifts, and their resistance to hypobromite oxidation[12,26] was confirmed by spiking with authentic compounds (data not shown). The large stable signal from scyllo-IP6 was adjusted to δ = 4.03 ppm (the chemical shift of this compound in the unbrominated East 50 soil) in all spectra, to facilitate comparison among spectra with small differences in pH and salt content. Based on this, signals from myo-IP6 occurred at δ = 4.38, 4.53, 4.89, and 5.79 ppm in a 1:2:2:1 pattern (Figure 4, Figure S1 in the Supporting Information). The chemical shift of inorganic orthophosphate varied slightly among samples, presumably due to variation in salt content,[34] but was typically within 0.05 ppm of δ = 5.95 ppm (Figure 4). Chemical shifts of the remaining signals in the phosphomonoester region were remarkably consistent across soils and treatments.

Table 2

Solution 31P NMR Spectroscopy Chemical Shifts (δ ppm) of Phosphates in Inositol Hexakisphosphate (IP6) Stereoisomers in Alkaline Soil Extracts (pH > 13)a

stereoisomer	configuration	contributing P groups	orientation (ax/eq)	chemical shift (δ ppm)
scyllo-IP6	6 ax	C1, C2, C3, C4, C5, C6	ax	4.03
myo-IP6	1-eq/5-ax	C5	ax	4.38
		C1, C3	ax	4.53
		C4, C6	ax	4.89
		C2	eq	5.79
neo-IP6	4-eq/2-ax	C2, C5	ax	4.58
		C1, C3, C4, C6	eq	6.67
	2-eq/4-ax	C1, C3, C4, C6	ax	4.93
		C2, C5	eq	5.17
d-chiro-IP6	4-eq/2-ax	C1, C6	ax	4.33
		C2, C5	eq	5.66
		C3, C4	eq	6.90
	2-eq/4-ax	C3, C4	ax	4.66
		C2, C5	ax	5.08
		C1, C6	eq	6.48

Signals are scaled relative to the large stable signal from scyllo-IP6 at δ = 4.03 ppm. Values varied by ± 0.01 ppm or less among samples and treatments (unbrominated, brominated, spiked). Eq, equatorial; ax, axial.

Figure 4

Solution 31P NMR spectra of NaOH–EDTA extracts of a soil from the Falkland Islands (East 50) showing the phosphomonoester region in detail (δ = 3.0 to 7.5 ppm). The following treatments are shown: untreated (A), untreated and spiked with a mixture of neo- and d-chiro-inositol hexakisphosphate (IP6) (B), pretreated by hypobromite oxidation (C), pretreated with hypobromite oxidation and spiked with a mixture of neo- and d-chiro-IP6 (D). Signal assignments are from the untreated spiked sample and assigned to myo (m), scyllo (s), d-chiro (c), or neo (n) IP6 stereoisomers. Inorganic orthophosphate (Ortho-P) is the large signal close to δ = 6.00 ppm. Spectra are plotted with 1 Hz line broadening and referenced to the chemical shift of scyllo-IP6 in spectrum A (δ = 4.03 ppm). The signal at δ = 4.58 ppm from the two axial phosphates of the 4-equatorial/2-axial conformer of neo-IP6 were small in the spike solution and were not detectable in soil extracts. Spectra for two additional soils showing the same treatments are in the Supporting Information (Figure S1).

The second spike solution contained neo-IP6 conformers in a 1:4 ratio (spectrum not shown), which did not appear to change after addition of the spike to the soils extracts. The ratio was different in the soil extracts themselves, because the largest neo-IP6 signal was at δ = 6.67 ppm (i.e., from the four equatorial phosphates of the 4-equatorial/2-axial conformer), with no similar signal detected at δ = 4.93 ppm (i.e., from the four axial phosphates of the 2-equatorial/4-axial conformer) in any of the brominated soil extracts. For quantification, we therefore assumed that all neo-IP6 in soil extracts was in the 4-equatorial/2-axial conformation.

Signals at δ = 6.48 and 6.67 ppm downfield of orthophosphate in all soils were identified as being from d-chiro- and neo-IP6, respectively. These signals were present in both brominated and untreated soil extracts. A signal at δ = 6.90 ppm from the 4-equatorial/2-axial conformer of d-chiro-IP6 present in the spike was not detected in any soil extract (untreated or brominated). Other signals from d-chiro- and neo-IP6 were not identified clearly in soil extracts either with or without pretreatment by hypobromite oxidation. A signal at δ = 5.08 ppm in untreated extracts coincided with one of the three signals from the 2-equatorial/4-axial conformer of d-chiro-IP6, as shown by spiking, but was considerably reduced by hypobromite oxidation. This indicates that the signal was from an organic phosphorus compound other than an inositol phosphate, probably α-glycerophosphate.[35] It should be noted that the signals assigned to d-chiro-IP6 could also arise from its enantiomer l-chiro-IP6, although we consider this unlikely given that the latter has never been identified in nature (see Discussion). A small signal at δ = 4.32 ppm increased in extracts of all soils after hypobromite oxidation. This remains unidentified but appears commonly following pretreatment by this procedure (B. L. Turner, unpublished observations).

Total IP6 in the three soils was between 36.7 and 41.4% of the extracted phosphorus (mean 38.6 ± 2.6%) with little variation between values calculated from untreated and brominated extracts (Table 3). Inositol hexakisphosphate therefore accounted for an average of 48.9 ± 4.2% of the soil organic phosphorus and 57.0 ± 4.7% of the phosphomonoesters (Table S2 in Supporting Information). These values are representative of the total soil phosphorus given that NaOH–EDTA extracted ≥90% of the total soil phosphorus. Thus, about half of the organic phosphorus in the three soils was IP6. The remaining phosphomonoesters resisting hypobromite oxidation presumably represented lower-order inositol phosphates and constituted an additional 18–21% of the soil organic phosphorus (see Table S2 in the Supporting Information).

Table 3

Inositol Hexakisphosphate (IP6) Stereoisomers in NaOH–EDTA Extracts of Three Soils from the Falkland Islands Determined by Solution 31P NMR Spectroscopy with or without Pretreatment by Hypobromite Oxidation (Bromination)a

	inositol hexakisphosphate stereoisomer
	% of spectra area					% of total IP6
soil/treatment	myo	scyllo	neo	d-chiro	total IP6	myo	scyllo	neo	d-chiro
East 50
none	22.4	15.1	2.50	1.69	41.7	53.8	36.2	6.0	4.1
bromination	22.7	14.7	2.37	1.37	41.1	55.1	35.8	5.8	3.3
average	22.5	14.9	2.44	1.53	41.4	54.5	36.0	5.9	3.7
West 18
none	21.4	12.4	1.83	0.81	36.4	58.7	34.1	5.0	2.2
bromination	21.6	12.6	2.19	2.88	39.3	54.9	32.1	5.6	7.3
average	21.5	12.5	2.01	1.84	37.8	56.8	33.1	5.3	4.8
West 22
none	21.0	10.9	2.63	2.84	37.4	56.3	29.1	7.0	7.6
bromination	20.4	10.6	2.55	2.38	36.0	56.8	29.5	7.1	6.6
average	20.7	10.7	2.59	2.61	36.7	56.5	29.3	7.1	7.1
overall average	21.6 ± 0.9	12.7 ± 2.1	2.35 ± 0.30	1.99 ± 0.56	38.6 ± 2.5	55.9 ± 1.3	32.8 ± 3.3	6.1 ± 0.9	5.2 ± 1.7

Values are expressed as both the proportion (%) of total spectral area and the proportion (%) of the total IP6. The overall average is the mean ± standard deviation of the three averages for each soil.

Most of the IP6 was in the form of myo-IP6 (55.9 ± 1.3% of the total IP6) and scyllo-IP6 (32.8 ± 3.3%), with smaller proportions of the neo (6.1 ± 0.9%) and d-chiro (5.2 ± 1.7%) stereoisomers (Table 3). Values were remarkably consistent across the three soils, and there was little variation between untreated and brominated extracts for all four stereoisomers in almost all cases. Overall, the stereoisomers were present in the following order: myo > scyllo > neo > d-chiro. The ratio of the four stereoisomers was 10.78:6.32:1.17:1.

Discussion

Solution 31P NMR spectroscopy is the most widely used procedure for determining the chemical nature of organic phosphorus in soils and sediments.[23] Our results extend the scope of this technique to include identification of all four stereoisomeric forms of IP6 present in nature. Both neo- and d-chiro-IP6 can be identified readily by signals downfield of orthophosphate between δ = 6.40 and 6.70 ppm (relative to scyllo-IP6 at δ = 4.03 ppm) and are detectable at relatively low concentrations given that the signals are well-resolved from the large orthophosphate signal close to δ = 6.00 ppm.

Importantly, both neo- and d-chiro-IP6 resist hypobromite oxidation. Our results indicate that it is possible to quantify the IP6 in well-resolved spectra containing high concentrations of inositol phosphates without brominating samples. However, bromination is clearly a useful pretreatment for estimation of the IP6 in environmental samples by solution 31P NMR spectroscopy, as indicated by its ability to distinguish between α-glycerophosphate and one of the three signals from the 2-equatorial/4-axial conformer of d-chiro-IP6 at δ = 5.08 ppm. The impact of this procedure on lower-order esters of all four stereoisomers remains unknown.

Both myo- and scyllo-IP6 can occur in two conformations depending on solution pH,[20,36,37] although both appear in a single conformation at the high pH of the extracts used in solution 31P NMR spectroscopy of soils. This yields consistent and stable signals and facilitates their identification and quantification.[12,24] In contrast, both conformers of neo- and d-chiro-IP6 can be present in strongly alkaline extracts, as indicated by spectra of authentic compounds shown here. In soil extracts, however, only a single conformer of each stereoisomer was detected. Specifically, no signals were detected at δ = 4.93 or 5.18 ppm from the 2-equatorial/4-axial conformer of neo-IP6, nor at δ = 6.90 ppm from two equatorial phosphates of the 4-equatorial/2-axial conformer of d-chiro-IP6. This simplifies quantification of these compounds in soil extracts, which can be achieved using the two signals at δ = 6.67 and 6.48 ppm for neo- and d-chiro-IP6, respectively. The possible presence of other conformers is indicated by signals at δ = 6.90 ppm (d-chiro-IP6) or δ = 5.18 ppm (neo-IP6; the signal at δ = 4.92 ppm from this conformer may be obscured by a signal from myo-IP6). It remains unclear why both conformers were present in the authentic samples at high pH, especially as signals from these compounds in the spike solution were unchanged following addition to soil extracts. Solution pH was strongly alkaline in all cases, so it seems possible that conformers might be stabilized by interaction with metals.[38] Further experiments are required to understand why the ratio of the two neo-IP6 conformers varies at strongly alkaline pH.

A caveat to the identification of d-chiro-IP6 is that identical signals in solution 31P NMR spectroscopy would be expected from its enantiomer l-chiro-IP6, so the possibility remains that the form detected in soil extracts is the l-, rather than the d-, enantiomer. We assume that the signals are from d-chiro-IP6, however, given that chiro-IP6 in soils was previously identified as the d-enantiomer[39,40] and that the l-enantiomer has never been detected in nature in phosphorylated form.[1,4]

Very few studies have quantified neo- or d-chiro-IP6 in soils or sediments since the early work of Cosgrove.[5,6] For the soils studied here, both compounds were present in much lower concentrations than either the myo or scyllo stereoisomers, which is in agreement with the older literature.[4] However, the concentration of the neo isomer appears to be greater than or similar to the d-chiro isomer in all three soils studied here. This contrasts with early reports that the d-chiro isomer occurred at greater concentration[8,9] but is consistent with the suggestion that the neo isomer might have been underestimated in studies that did not include pretreatment by hypobromite oxidation.[26]

The identification of neo- and d-chiro-IP6 reported here allows a reappraisal of publications that have applied solution 31P NMR spectroscopy to soils and aquatic sediments. A number of studies reported signals downfield of orthophosphate in solution 31P NMR spectra of soil and sediment extracts but did not identify them as neo- or d-chiro-IP6. It was initially speculated that the signals represented aromatic phosphodiesters based on their similarity to compounds such as binaphthyl diylhydrogen phosphate,[41,42] although the resistance of the signals to hypobromite oxidation subsequently indicated the likelihood that they represented inositol phosphates.[12] We can now assign these signals to neo- and d-chiro-IP6.

A series of temperate pasture soils yielded signals that we can now assign to neo- and d-chiro-IP6 at concentrations of 8–51 mg P kg–1, which represented 2–5% of the extracted phosphorus.[42] The same signals were also reported in Madagascan rice soils at concentrations of 0.8–4.9 mg P kg–1 (0.5–1.3% of extracted phosphorus)[43] (note that these values are 6-fold lower than the erroneous values in the original article). Similar signals were also reported in extracts of Russian grassland soils,[44] Scottish mineral soil,[41] sewage sludge,[45] and humic acids extracted from an agricultural soil.[46] However, only traces were detected in subarctic tundra and lowland temperate and tropical rain forests.[32,47,48] The humic acid study also detected scyllo-IP6 but not the more widespread myo isomer, suggesting a difference in the mechanisms by which the various stereoisomeric forms of IP6 become stabilized in soils.[46]

Importantly, we can now extend the identification of neo- and d-chiro-IP6 to aquatic systems. Previously, the only report of either of these stereoisomeric forms was the measurement of d-chiro-IP6 by gas chromatography in riverine, estuarine, and marine sediments of Tokyo Bay and the surrounding catchment.[11] In that study, most of the IP6 was in the myo form, with 10–15% as scyllo and 1–5% as d-chiro (neo-IP6 was not detected). Elsewhere, signals that can now be assigned to neo- and d-chiro-IP6 were reported in sediments from a wide range of Danish lakes,[49] settling seston and sediments from a eutrophic lake in China,[50,51] and sediments from a brackish embayment in Helsinki.[52] In the latter study, the signal corresponding to neo-IP6 constituted 0.3–3.0% of the organic phosphate and was greatest in sediment layers dated to periods of high pollutant inputs into the embayment. No signals corresponding to neo- and d-chiro-IP6 were reported in NMR studies of marine sediments,[53,54] although d-chiro-IP6 was detected in sediments from Toyko Bay.[27]

Although it has been known for some time that inositol phosphate concentrations vary widely among soils (from none to all of the organic phosphorus, reviewed in ref (4)), it has been suggested recently that the amount of inositol phosphate in soil has been overestimated and is quantitatively unimportant.[34,35] Here, IP6 accounted for half the soil organic phosphorus extractable in NaOH–EDTA, which in turn represented ≥90% of total soil phosphorus. These results demonstrate unequivocally that IP6 can constitute a considerable proportion of the organic phosphorus in some soils.

Despite the apparent abundance of IP6 stereoisomers in terrestrial and aquatic ecosystems, their origins, dynamics, and biological function remain unknown. Given the importance of phosphorus as both a nutrient and a pollutant, it is remarkable that so little is known about one of the most widespread groups of organic phosphates in the environment. Results presented here confirm the prevalence of the IP6 stereoisomers in soils and lake sediments and support previous reports that the relative abundance of the four stereoisomeric forms of IP6 follows the pattern myo >scyllo >neo >d-chiro (although the scyllo isomer can be the dominant form in some environments; e.g. ref (47)). The investment in energy and phosphorus involved in the synthesis of the stereoisomeric forms of IP6 suggests that they have some important yet currently unknown biological function. Research on this topic has been restricted by the difficulty in identifying the inositol phosphate stereoisomers in environmental samples, but the methodology presented here allows the simultaneous determination of all four stereoisomeric forms of IP6 in a single analysis. This opens up new possibilities for studying the origins and ecological significance of these enigmatic compounds.