LC‐ICP‐MS analysis of inositol phosphate isomers in soil offers improved sensitivity and fine‐scale mapping of inositol phosphate distribution

Organic forms of phosphorus (P) prevail in soils and their quantification is vital to better understand global biogeochemical cycles. P speciation in soil is commonly assessed by 31P NMR spectroscopy of sodium hydroxide‐EDTA (NaOH‐EDTA) extracts. A liquid chromatography‐inductively coupled plasma‐mass spectrometry (LC‐ICP‐MS) method that employs NaOH‐EDTA is described. Comparison with 31P NMR shows that LC‐ICP‐MS is up to three orders of magnitude more sensitive. It allows measurement in samples as small as 1 mg. We reveal variation of inositol phosphate distribution in Swedish boreal forest soil and identify myo‐ and scyllo‐inositol hexakisphosphates and other isomers including scyllo‐inositol pentakisphosphate. Speciation of the major inositol phosphates was not altered by long‐term nitrogen fertilization.


| INTRODUC TI ON
Phosphorus (P) is a major nutrient that limits plant growth in diverse ecosystems, including tropical forests (Cunha et al., 2022), boreal forest (Giesler et al., 2012) and species-rich calcareous grassland (Johnson et al., 1999).Sustained input of nitrogen (N) from atmospheric deposition or fertilization increases P limitation of ecosystems globally (Chen et al., 2020) and often results in increased activity of enzymes implicated in organic P degradation (Johnson et al., 2005;Papanikolaou et al., 2010).The chemical diversity of soil P is vast, and resolving the various components of this, especially organic P, is vital to better understand biogeochemical cycling and ecological processes, such as coexistence and dominance of plants in communities (Turner, 2008).31 P NMR spectroscopy of sodium hydroxide-EDTA (NaOH-EDTA) extracts has allowed an analysis of P speciation in soil (Cade-Menun et al., 2002;Doolette et al., 2010;Newman & Tate, 1980;Reusser, Verel, Zindel, et al., 2020).The approach has the advantage of being able to quantify and assign P by speciation (Giles et al., 2011;Jarosch et al., 2015).Even so, assignment of resonances and identification of form of P within complex NMR spectra presented by soil matrices is most powerful when it is combined with spiking of extracts (Doolette et al., 2009;Liu et al., 2014;McLaren et al., 2022).This approach has been used widely to measure inositol phosphates, reported to be a major form of organic P in many soils (Giles et al., 2011;Turner et al., 2002) or undetectable in others (Doolette et al., 2010).
Here, we describe chromatographic separation and quantification of myo-, scyllo-, neo-and chiro-inositol hexakisphosphate isomers, besides 'lower' inositol phosphates, by direct analysis of NaOH-EDTA extracts on chromatography interfaced with inductively coupled plasma-mass spectrometry (ICP-MS).The approach is an extension of earlier ICP-MS work (Rugova et al., 2014), employing the considerable resolving power of anion chromatography on acid gradients for resolution of inositol hexakisphosphates (Whitfield et al., 2018).We use the method to characterize inositol phosphates of forest soils that have been fertilized for 38 years with N, as NH 4 NO 3 , and test the hypothesis that N addition leads to a reduction in inositol phosphate pools.

| Soil information
Soils were collected from an 87-year-old Pinus sylvestris forest stand in Åsele, Sweden (64°07′ N, 17°33′ E).The site is located 330 m above sea level.The soils are Podzols with a sandy sediment texture.The forest consists of six replicate 30 × 30 m 2 experimental plots, of which three have remained unfertilized and three have been subjected to N fertilization for 38 years at an equivalent to 75 kg N ha −1 year −1 in the form of NH 4 NO 3 (Jacobson & Pettersson, 2010).
Samples were stored at −20°C until processing.Soils taken for analysis of the properties shown in Table S1 were collected to a depth of 10 cm from 15 separate locations within each experimental plot, giving a total of 90 samples across all plots.

| Properties of the soils used in this study
Soil moisture was calculated by gravimetric weight.Organic content was determined by percentage loss on ignition at 450°C.Water extractable N, also known as dissolved organic N (DON), and water extractable inorganic N, also known as mineral or dissolved inorganic (plant available) N (DIN), were extracted from fresh soils (Bardgett et al., 2007) and analysed using a Seal AA3 flow multi-chemistry auto-analyser (Seal Analytical, UK).Dissolved organic and inorganic C (DOC, DIC) were extracted similarly and analysed on a 5000A TOC analyser (Shimadzu, Japan).Microbial C and N were extracted by fumigation-extraction (Vance et al., 1987) and analysed as above.
Total C and N (%) were determined by combustion analysis of oven-dried soils using Vario EL Cube (Elementar, Hanau, Germany).
Organic and inorganic P were extracted from dry soils and analysed by treating samples with the ascorbic acid method adapted from Kuo (1996), measuring absorbance at 880 nm with a CLARIOstar microplate reader (BMG Labtech, Ortenberg, Germany).

| NaOH-EDTA extraction
Samples (0.5 g) of each soil were weighed into a 50-mL Falcon tube in which 5 mL of 0.25 M NaOH/0.05M EDTA was added.Samples were shaken overnight (~16 h) at 180 rpm at room temperature.One millilitre from each sample was filtered into sampling vials using a 0.22μm filter.

| NaOH-EDTA extraction of Arabidopsis seed
Single seeds of Arabidopsis were ground by hand with a plastic pestle in 0.1 mL of 0.25 M NaOH/0.05M EDTA in a 1.5-mL microfuge tube and incubated overnight at room temperature before centrifugation at 22,000× g for 15 min at 4°C.An aliquot of the supernatant (80 μL) was removed to a Chromacol 03-FISV(A) glass autosampler vial.

| Microfractionation of soil
Replicate samples of 10 mg of soil (Sample 4A of Figure 1) were weighed into microfuge tubes.Separately, 2 g of the same soil was ground and sieved through a 1.4-mm stainless steel sieve.From this, replicate 10-mg portions were taken.All samples/portions were extracted with 100 μL of 0.25 M NaOH/0.05M EDTA with shaking overnight before centrifugation at 22,000× g for 15 min at 4°C.An aliquot of the supernatant (80 μL) was removed to an autosampler vial.

| Calibration
Influence of LC solvent on detector response to PO + at m/z 47 and P + at m/z 31 was tested in a matrix of KH 2 PO 4 versus HCl delivered at 0.4 mL min −1 (Table S2).Response to KH 2 PO 4 differed by ≤7% at the concentrations and flow rates used.Analyses were also made following chromatographic separation of inositol phosphates with the column eluate directed in toto to the nebulizer.A calibration curve for injections of myo-InsP 6 is shown (Figure S1).

| 31 P NMR spectrum acquisition and processing
NMR spectra were recorded, at 500 MHz for 1 H, with a Bruker AVANCE NEO spectrometer equipped with a 5-mm solution-state iProbe probe-head.Spectra were recorded at 31 P frequency of 202.12 MHz, with inverse gated decoupling sequence, applying 31 P 90° pulses (P1 calibrated for a length of 12 μs), with a recovery time

| Soil parameters
The soil parameters measured are shown in Table S1.Content of total P (tP) ranged from 171 to 1009 mg kg −1 dry wt with a median of 446 mg kg −1 .Of this, inorganic P (IP) ranged from 18 to 817 mg kg −1 dry wt with a median of 115 mg kg −1 and organic P (OP) ranged from 0 to 840 mg kg −1 dry wt with a median of 283 mg kg −1 .
Separate samples of soils from the same experimental plots, but not represented in the measurements of Table S1, were analysed for NaOH-EDTA-extractable total phosphorus tP NaOH-EDTA by ICP-AES at the University of East Anglia (Figure 1).Total P measured by ICP-AES in NaOH-EDTA-extracted samples fell in the range 6.85 mg P. kg −1 measured in subplot 2C to 66.9 mg Analysis by one-way ANOVA found significant differences in mean total P between plots (F 5,12 = 11.8, p = 0.0003), with post hoc Tukey's comparison finding Plot 1 to contain significantly higher total phosphorus compared to all other plots (Plot 1 vs. Plot 2, p = 0.001; vs. Plot 3, p = 0.0004, vs. Plot 4, p = 0.0008; vs. Plot 5, p = 0.0007; vs. Plot 6, p = 0.0014).All other plots were not significantly different from each other in respect to total P content as measured by ICP-AES by Tukey's comparison (for all, p > 0.05), and were found to be non-significant by one-way ANOVA when data for Plot 1 were removed from the model (F 4,10 = 0.6514, p = 0.6389).Removal of Plot 1 from the data set as identified by ROUT test gave a mean total P of 13.8 mg P. kg −1 across plots 2-6.
For the spiked sample (Figure 3b), the sharp C2, C1/3, C4/6 and C5 resonances at 5.21, 4.29, 3.93 and 3.80 ppm have a summed, integrated intensity ~10× that of the orthophosphate resonance at 5.34, most of that signal arising from the soil extract.Additionally, the extract which was mildly diluted by the addition of NMR reagents before analysis in Figure 3a was separately diluted 10-fold with NaOH-EDTA and an aliquot of 100 μL was injected (Figure S3).
The chromatograms generated from injections shown in Figure S3 and Figure 3c are more or less equivalent, both representing an extract of ~1 mg of soil on-column.These data highlight the difference in sensitivity of LC-ICP-MS and 31P NMR.
To assess the robustness of LC-ICP-MS with larger samples, three separate extractions of 0.5 g of a ground and sieved soil sample are shown (Figure 4).The coefficients of variation of peak area, a product of run-to-run instrument variability and extraction reproducibility, for myo-InsP 6 and scyllo-InsP 6 were 0.055 and 0.049, respectively.
With confidence in the sensitivity and reproducibility of LC-ICP-MS, the soil samples of Figure 1 were subjected to LC-ICP-MS.
We chose not to homogenize samples because we sought to elaborate the possibility of local (mm scale) variation in soil inositol phosphates.As example, the six samples of Figure 5 show striking variability in soil inositol phosphate content, predominant peaks of myo-and scyllo-InsP 6 in all samples and multiple peaks of less abundant isomers.

| Analysis of speciation of 'lower' inositol phosphates in soil by LC-ICP-MS
Spiking of samples with authentic standards is essential for identifica- experiments.Figure 6 compares the elution of a soil sample (Figure 6a) and a set of standards obtained by acid hydrolysis of myo-InsP 6 (Figure 6b).Peaks with identical retention times to inositol phosphates previously identified in myo-InsP 6 hydrolysates (Blaabjerg et al., 2010;Chen & Li, 2003;Whitfield et al., 2018Whitfield et al., , 2020) ) are numbered 1-10 and peaks in the soil sample that are not clearly represented in the hydrolysate are identified a-e, with e identified as scyllo-InsP 6 (see Figure 2, Figure S2).Notwithstanding the ability of chromatography (Figure 2, one that elutes among myo-inositol tetrakisphosphates (InsP 4 s) (cf. Figure 7a,b), while the d-chiro-InsP 6 standard was resolved from the peak of myo-InsP 6 (Figure 7a).The early elution of neo-InsP 6 (Figures 2 and 7a, Figure S2) and the increase of retention time through the series InsP 2 -InsP 6 for myo-inositol phosphates (Figure 6), together with previous descriptions of elution of neo-InsP 4 , neo-InsP 5 and neo-InsP 6 (Whitfield et al., 2018), reveal that 'lower than hexakisphosphate' esters of neo-inositol elute considerably earlier than myo-InsP 4 s.
The low abundance of InsP 5 s relative to InsP 6 isomers is consistent with the observations of Irving and Cosgrove (1982) with other soils.

| Lack of effect of N-fertilization on soil inositol phosphate profile
With identities assigned to various peaks of inositol phosphate in Åsele forest soil, a comparison was made of fertilized and unfertilized plots, example traces of which are shown in Figure 5.
The results of analysis of extractions of 18 soil samples 1A-C to 6A-C are shown in Table 2 with peaks assigned broadly to class by elution position, which is an excellent proxy for extent of phosphorylation.As described above, such a simplification risks misidentification of some of the minor components but, nevertheless, allows test of the effect of fertilization.The predominant inositol phosphate species identified were myo-InsP 6 and scyllo-InsP 6 (Table 2). 3 The 'Pi' peak coelutes with InsP 1 and likely includes other unknown organic phosphates. 4 The 'Pi' peak is summed with unknown compounds eluting before InsP 3 .
Analysis by one-way ANOVA showed that total NaOH-EDTAextractable P varied among the plots (Figure 1, p = 0.0003).Similarly, individual and total inositol phosphates also varied significantly between plots (Table 2), but only because of the departure of Plot 1 from all other plots.When plots were pooled within each treatment to give nine replicates and tested by a two-tailed Student's t-test with Welch's correction, there was no significant difference in total inositol phosphates with long-term N-fertilization (∑myo-[InsP 3 -InsP 6 ], scyllo-InsP 6 p = 0.1462).Thus, while significant difference was noted by ANOVA between non-fertilized soils 1 and 3 in respect of Pi, InsP 3 , myo-InsP 6 and tP NaOH-EDTA -this difference did not extend to other classes of inositol phosphates or to the aggregated total of inositol phosphates.
We anticipate further development of LC-ICP-MS will include tandem UV detection to identify UV-absorbing components in soil extracts.A recent study employing sequential chemical fractionation (Reusser et al., 2023) indicates that, in the soils studied, recovery of inositol phosphates was not complete in NaOH-EDTA.The LC-ICP-MS method offers opportunity to probe the distribution of inositol phosphate species between different fractions obtained by SCF.
Further analysis of Table 2, to compare the ratio myo-InsP 6 : scyllo-InsP 6 , is shown in Figure 8. Whether data were pooled or not, there was no difference in ratio between treatments, that is despite marked difference in total NaOH-EDTA-extractable P (10.2-53.3mg kg −1 , as plot means, Table 2), or 6.8-66.9mg P. kg −1 as individual subplots (Figure 1).The ratios of myo-InsP 6 : scyllo-InsP 6 were not different between control and N-fertilized plots (Welch's t-test for pooled treatment means, p = 0.8708).

| Accessing variation in soil organic phosphate distribution on mm scale
To explore the limits of sampling, we took nine individual 10-mg samples from soil 4C (plot 4, subplot C) and extracted each with 0.1 mL of NaOH-EDTA.In parallel, we took 2 g of the same sample and ground it to pass through a 1.4-mm sieve before taking nine individual 10-mg portions for extraction in the same volume.The LC traces of two of each of the dispersed samples and homogenized (ground) portions that represent the extreme outliers of detector response for each set are shown (Figure 9).

| DISCUSS ION
Considerable variation in organic P and phytate is reported in different soils (Smernik & Dougherty, 2007;Turner et al., 2002).Variation to be organic) of the parental material (organic layer) in most of more than 300 soils less than 1 mg kg −1 .
The data shown here confirm the resolving power of LC-ICP-MS for different inositol hexakisphosphates found in soil and for InsP 4 and InsP 5 species.One caveat is that the early (among inositol hexakisphosphates) elution of neo-InsP 6 places neo-InsP 6 in part of the chromatogram where myo-InsP 4 s elute, albeit as minor fractions.
Perhaps, more noteworthy is the gain in sensitivity obtained by LC-ICP-MS, one-two orders of magnitude more sensitive than UV detection of ferric complexes (Whitfield et al., 2020).LC-ICP-MS can measure less than a single pmol on column.To put this in context of demanded (Reusser, Verel, Zindel, et al., 2020).
While there has been little systematic research on the inositol phosphate profile of plant tissues other than storage organs (Raboy, 2003), myo-inositol hexakisphosphate is considered a major input to soil organic phosphate (Turner et al., 2002).Here, the ability of LC-ICP-MS and 31 P NMR (Reusser, Verel, Zindel, et al., 2020) to identify isomers bearing, or not, a 2-phosphate, is descriptive of the origin of these isomers: Those with a 2-phosphate are most likely products of phytate turnover, those without are synthetic precursors (Irvine & Schell, 2001).
The isomers measured by LC-ICP-MS in a single Arabidopsis seed is typical of bulk measurement of Arabidopsis (Bentsink et al., 2003), of seeds in general (Raboy, 2003) and highlights the use of LC-ICP-MS for analysis of local accretions of organic material.
While N inputs are reported to increase P limitation of ecosystems (Cunha et al., 2022) and increase soil enzymes associated with P cycling (Margalef et al., 2021;Papanikolaou et al., 2010), other studies report no effect on soil enzyme activities (Turner & Wright, 2014).We note that degradation of different phytates, d- chiro-, myo-, neo-and scyllo-, is not described in situ in soil at the level of speciation of 'lower' inositol phosphates.Consequently, the data do not allow comment on the enzymology of phytate turnover in control and N-fertilized plots.Nevertheless, there was no differential effect of fertilization on a specific isomer.
LC-ICP-MS offers analyses formerly only accessible to 31P NMR.It is quicker, less intensive in sample workup and more sensitive.It offers opportunity for two-and three-dimensional mapping of hydroxide-extractable organic and inorganic P in soil.Here, our analysis of Åsele soil can be placed in context of Scandinavian soils, for which landscape-scale surveys of soil P are reported (Ballabio et al., 2019;Spohn & Stendahl, 2022) and for which organic P content was shown to increase as particle size decreases (Spohn, 2020;Spohn & Stendahl, 2022).By microsampling, LC-ICP-MS allows the speciation of organic P on scales approaching single mm resolution.Microscale resolution of P-binding to soil mineral particles has been described (Adediran et al., 2020(Adediran et al., , 2022)).
Separately, LC-ICP-MS could be an ideal tool to study the P dynamics of experimental microcosms, such as mycorrhizal splitroot systems containing fungal and microbial partners (Schreider et al., 2022).

| CON CLUS IONS
As plant material is a major input to soil, with seeds especially enriched with myo-inositol hexakisphosphate, demonstration of how

(
D1) of 30 s, a spectral width of 40 ppm centred in 5 ppm (covering the region from 25 to −15 ppm) and an optimized acquisition time of 0.4 s.The number of scans was set to 8192.Spectra were processed with Topspin 4.1.3by applying an exponential window function with a line-broadening factor of 1 Hz before the Fourier transform to 32 K points.Phase and baseline correction were corrected with the automatic combined phase and baseline algorithm.Methylene diphosphonate (MDP) signal was used as chemical shift reference, set at 16.6 ppm.

F
I G U R E 1 NaOH-EDTA-extractable total P (tP NaOH-EDTA ), measured by ICP-AES.Fresh soil extracted with NaOH-EDTA at 1:10 mass:volume ratio was diluted and analysed by ICP-AES.TA B L E 1 Optimized instrumental parameters for LC-ICP-MS/ MS. −1 in subplot 1B.Mean total P was 20.35 mg P. kg −1 across all six plots.
From this, 500 μL was mixed with 25 μL of 30 mM methylenediphosphonic acid, 34 μL of NaOD and transferred to a 5-mm NMR tube.The sample, representing the NaOH-EDTA extract of 50 mg of soil, was subjected to NMR.The raw data, which are not fitted or deconvoluted, other than application of Fourier transform, phase and baseline correction, are shown (Figure3a).Resonances typical of inositol phosphates were barely detectable.Following NMR, an aliquot (105 μL) was removed and its volume replaced with 50 μL F I G U R E 2 LC-ICP-MS of a soil extract spiked with different inositol hexakisphosphate standards.(a) Sample 4A spiked with d-chiro-InsP 6 ; (b) sample 4A spiked with myo-InsP 6 ; (c) sample 4A spiked with neo-InsP 6 ; (d) sample 4A spiked with scyllo-InsP 6 ; (e) Sample 4A.For all panels, 10 μL was injected.
tion of inositol phosphates by peak position (chemical shift) in NMR F I G U R E 3 Comparison of methods of measurement of inositol phosphates in a soil extract.(a) 31P NMR of an NaOH-EDTA extract of soil (50 mg equivalent in 560 μL final volume).(b) 31P NMR of A after the removal of 100 μL, restoration to volume and spike with 1 μL 100 mM myo-InsP 6 .The resonances of individual P nuclei in myo-InsP 6 are identified by locant (C2, C1/3, C4/6 and C5), that of orthophosphate is indicated (orthoP).(c) LC-ICP-MS analysis of PO + at m/z 47 of 10 μL of the sample analysed in (a); (d) LC-ICP-MS analysis of 13 μL of the sample analysed in (b).The inset in (c) is a 10-fold expansion of the y-scale.

Figure S2 )
Figure S2) to separate all known inositol hexakisphosphates yet identified in soil nor the consistency of LC (retention time) and ICP-MS detector response (Figure 4), a separate sample of soil 4C, was spiked with d-chiro-InsP 6 , neo-InsP 6 and scyllo-InsP 5 before chromatography.The chromatograms from the original and spiked sample are shown in Figure 7a and that of the diluted (with myo-InsP 6 hydrolysate) sample is similarly overlaid with the original in Figure 7b.Precise co-elution of soil peaks with standards is revealed.The co-elution of peak d (Figure6a) with scyllo-InsP 5 (scyllo-Ins(1,2,3,4,5)P 5 ) (Figure7a) identifies 'd' as scyllo-InsP 5 , wholly separable from all myo-InsP 5 s(Whitfield et al., 2018(Whitfield et al., , 2020)).The neo-InsP 6 standard co-eluted precisely with
Figure 4, and its interaction with natural spatial variation (nonuniformity of distribution) of organic material within sample 4C.The coefficients of variation for myo-InsP 6 and scyllo-InsP 6 contents were 1.589 and 1.141, respectively.For the ground/ sieved portions (Figure 9c,d), the natural spatial variation has been removed by homogenization.Consequently, the coefficients of variation for myo-InsP 6 and scyllo-InsP 6 are smaller, 0.143 and 0.256, respectively.Here, the coefficients represent the product of run-to-run instrument variability and reproducibility of extraction of otherwise identical samples.These coefficients can be directly compared with those (0.055, myo-InsP 6 and 0.049, scyllo-InsP 6 ) from the larger scale (0.5 g) extraction of a ground/sieved soil (Figure 4).These comparisons define the boundaries of LC-ICP-MS for analysis of the soils studied here.Separately, they reveal non-uniformity of inositol phosphate content of soil on 10-mg scale or mm scale, and the suitability of LC-ICP-MS for investigation of this.

Figure 10
Figure 10 shows the inositol phosphate speciation of these nonsieved (NS) or sieved (S) samples, expressed per kg of soil.That the means of the individual inositol phosphates are numerically similar, is coincidental.There are 200 individual 10-mg samples in 2 g of the NS sample and we sampled nine of these.We have no reason to assume that, within the nine, we have found the outliers or the 'best' representatives of these.Here, the underlying extraction and run-to-run instrument variability of the 'measurement' of NS is represented by the error (sd) bars of S, the natural non-uniformity of the samples increases the error bars.The proportions of the different classes of inositol phosphates identified were approximately 1.3%, 2.5%, 7.6%, 73.3% and 15.1%, respectively, for myo-InsP 3 , myo-InsP 4 , myo-InsP 5 , myo-InsP 6 and scyllo-InsP 6 .To put local non-uniformity of organic P in context of inputs to soil, three extractions of single seeds of Arabidopsis are shown InsP 6 were detected, representing approximately 30% and 70% of total P, respectively.Minor peaks of D-and/or L-myo-Ins(1,2,3,4,5)P 5 and D-and/or L-myo-Ins(1,2,4,5,6)P 5 , representing less than 2% of total P were clearly detected (peaks b and c) and with data smoothing putative peaks of other myo-inositol pentakisphosphates (peaks a and d) could be discerned, but neither d-chiro-, neo-nor scylloinositol hexakisphosphate was detected.
may have arisen through the use of disparate methodologies, through mis-identification of phytate resonances in 31P NMR(Smernik & Dougherty, 2007) or because of spatial variation.Partly for these reasons, we have sought to develop a methodology that is compatible with the NaOH-EDTA extraction regimen that has been the foundation of 31P NMR analyses of many groups and which, by chromatography, separates inositol phosphates from the 'humic' fraction.Setting aside the general absence in the literature, as here, of study of the same soil by different groups, for the Åsele soil, total P (tP) of N-fertilized and unfertilized plots bound the values of ca.18-22 mmol/ kg (558-682 mg kg −1 ) reported for the organic layer of Swedish, Tärnsjö and Tönnersjöheden, forest soils (Figure1,Adediran et al., 2020).Of these, organic P represented the major P fraction at ~17-19 mmol kg −1 (Figure1,Adediran et al., 2020).Median values of 0.50-0.180mg kg −1 of organic P were measured for the 0-10 cm mineral soil layer of Scandinavian soil(Spohn & Stendahl, 2022), with total P (assumed all F I G U R E 9 LC-ICP-MS separation and detection of inositol phosphates in hydroxide-EDTA extracts of 10-mg soil samples.Individual samples representing the outliers of detector response for nine individual samples of non-sieved soil 4C (a, b) or portions of sieved soil 4C (c, d) were analysed by LC-ICP-MS at m/z 47.The panels on the right are expansions of those on the left.Classes of inositol phosphates common to all samples are labelled in (a and c).10-μL injections.

31P
NMR, which typically uses ~25%-50% of the sample extracted from 1 g equivalent of soil(Cade-Menun & Preston, 1996;Doolette et al., 2011), LC-ICP-MS measures multiple inositol phosphates in samples of 10 mg and can measure myo-and scyllo-InsP 6 in extracts equivalent of a single mg of soil.An added advantage is that detector response is directly proportional to P content, independent of degree of phosphorylation of separated species.The relative size of the peaks accurately reflects the amount of P in each species.Unlike for 31 P NMR, in which total signal for an individual inositol phosphate species is divided among the discrete resonances of individual P nuclei (the nuclei of the phosphate substituents on the inositol ring that are given individual locants, 1-, 2-.3-, 4-, 5-and -6), in LC-ICP-MS a single peak gives the total P signal for an individual inositol phosphate.Even so, distinct inositol phosphate species can co-elute in LC.Conveniently, LC-ICP-MS has a cycle time of 50 min, that could be shortened, contrasting with the 1-2 days required of 31P NMR when recycle delays of up to 30 s, to allow full T1 relaxation, and 1024 or 4096 scans are

F
Speciation of inositol phosphates in hydroxide-EDTA extracts of 10-mg soil samples.Nine individual samples of a soil (NS) were analysed by LC-ICP-MS at m/z 47.Separately, the soil was homogenized by grinding and sieving (S) and nine portions of this processed soil were analysed.Mean and standard deviation are shown.