Elsevier

Geoderma

Volume 345, 1 July 2019, Pages 31-37
Geoderma

The structural composition of soil phosphomonoesters as determined by solution 31P NMR spectroscopy and transverse relaxation (T2) experiments

https://doi.org/10.1016/j.geoderma.2019.03.015Get rights and content

Highlights

  • Phosphomonoesters as a broad NMR signal is a major pool of soil organic P.

  • The structural composition of the broad signal is due to ‘homogenous’ broadening.

  • The T2 times of the broad signal were shorter than that of sharp signals (e.g. myo-IHP).

  • The broad signal is likely comprised of more than one component.

Abstract

In terrestrial ecosystems, a large proportion of the phosphorus (P) in soil is often found within soil organic matter. However, the majority of organic P in soil remains ‘unresolved’ and is largely observed as a ‘broad’ signal within the phosphomonoester region of a solution 31P nuclear magnetic resonance (NMR) spectrum on soil extracts. Our aim was to gain insight into the composition of four soils using the transverse relaxation (T2) time of the magnetisation in solution 31P NMR spectroscopy as a probe of their structure. We found the broad signal within the phosphomonoester region rapidly decayed compared to the sharp signals (i.e. myo- and scyllo-inositol hexakisphosphate) across all soils, which corresponded to the former having a shorter T2 time than the latter, and supports the existence of a broad signal due to supra-/macro-molecular structures. Furthermore, measures of the broad signal's line-width at half peak intensity based on T2 times were found to be less than that obtained from spectral deconvolution fitting. Therefore, our results strongly suggest that the broad signal is itself comprised of more than one component. The significance of this is that the chemical nature of a large proportion of soil organic P appears to be structurally complex.

Introduction

Soil organic matter is an essential component of terrestrial ecosystems because of its critical role in providing ecosystem services (Schmidt et al., 2011). Phosphorus (P) is an innate constituent of soil organic matter (Kirkby et al., 2011), however, a large proportion (typically >50%) of the organic P in soil remains ‘unresolved’ (Jarosch et al., 2015; McLaren et al., 2015a). This pool has so far not been directly attributed to recognisable biomolecules (Bünemann et al., 2008a; Noack et al., 2012), but largely described as comprising of phosphomonoesters (McLaren et al., 2015a), associated with large molecular weight material (Jarosch et al., 2015; McLaren et al., 2015a; Steward and Tate, 1971), ‘humic acid’ fractions (Bedrock et al., 1994; He et al., 2006), and soil organic matter (Condron and Goh, 1989; McLaren et al., 2014; Swift and Posner, 1972). New insight is needed on the chemical nature of this unresolved pool of soil organic P, which will improve our understanding of the P cycle in soil-plant systems (George et al., 2018).

Solution 31P nuclear magnetic resonance (NMR) spectroscopy on alkaline extracts has been the predominant method used to identify the chemical nature of soil organic P since 1980 (Cade-Menun and Liu, 2014; Newman and Tate, 1980). The main classes of organic P detected in soil extracts are phosphomonoesters (ROPO32−: where R is an organic moiety), phosphodiesters (R1OR2OPO2: where R1 and R2 are organic moieties), and phosphonates (e.g. RPO(OH)2, where R is an organic moiety) (Condron et al., 2005). In general, the majority of organic P (typically >80%) occurs as phosphomonoesters in soil extracts (Condron et al., 2005). However, due to a high degree of signal overlap within this region of a NMR spectrum, spectral deconvolution fitting procedures are required to partition the NMR signal and assign peaks to various compounds of organic P (Doolette and Smernik, 2015). The problem with this is that there are differing views on how to apply spectral deconvolution fitting procedures to solution 31P NMR spectra on soil extracts (Doolette and Smernik, 2015), which results in considerable differences in the reported distribution and concentrations of ‘identified’ soil organic P.

There are two main approaches of applying spectral deconvolution fitting procedures to the phosphomonoester region. These are: 1) the deconvolution fitting procedure of Turner et al. (2003) or modifications therefore (i.e. Hill and Cade-Menun (2009) and Vincent et al. (2012)), which involve fitting a series of ‘sharp’ peaks from the peak maxima to the baseline of the spectra within the phosphomonoester region; and 2) the deconvolution fitting procedure of Bünemann et al. (2008b) or modifications therefore (i.e. McLaren et al. (2015b)), which involves fitting a ‘broad’ (or ‘background’) peak and then any ‘sharp’ overlying peaks. There is strong evidence to include a broad peak when carrying out spectral deconvolution fitting based on (i) an overestimation of concentrations of myo-inositol hexakisphosphate (myo-IHP) if a broad signal is not included (Doolette et al., 2010; Doolette et al., 2011), (ii) preferential extraction of myo-IHP with hydrofluoric acid, which reveals NMR signal at the chemical shifts of myo-IHP but within an underlying broad signal (Dougherty et al., 2007), (iii) concentrations of organic P that were not hydrolysable by enzymes in NaOH-EDTA extracts were related to that of the broad signal and also organic P in large molecular weight (>5 kDa) fractions (Jarosch et al., 2015), and (iv) the isolation of the broad signal in large molecular weight fractions (>10 kDa) (McLaren et al., 2015a). However, the structural composition of the broad signal is itself not known and could be described as (i) a series of neighbouring sharp peaks that merge to visually appear as a broad peak or (ii) one (or a few) large and complex supra−/macro-molecules that exhibit a broad peak (Levitt, 2008).

Compounds have several different types of relaxation properties under NMR conditions, which have previously been explained in detail (Bloembergen et al., 1948; Claridge, 2016; Keeler, 2010; Levitt, 2008). Transverse relaxation (T2) experiments can probe the underlying structure of a NMR signal as comprising of ‘broad’ and ‘sharp’ components (Claridge, 2016; Keeler, 2010; Levitt, 2008; Meiboom and Gill, 1958; Schmidt-Rohr and Spiess, 1994). This is based on probing their T2 times, which are generally related to a compound's line-width at half peak intensity (Bloembergen et al., 1948; Claridge, 2016). Essentially, T2 relaxation provides information on the dynamics of molecules as probed by the coherence lifetime of their spin magnetisation in the transverse plane (x-y plane). This distinguishes the effects of line-broadening due to variations within the magnetic field caused by instrumental imperfections or slight conformational variations, in which case a broad signal would comprise a series of sharp neighbouring peaks (i.e. ‘inhomogeneous’ broadening) (Levitt, 2008). In contrast, variations within the magnetic field caused by differences in the molecular dynamics of compounds can be quantified based on their T2 properties, in which case a broad signal would comprise of a single (or a few) compounds (i.e. ‘homogeneous’ broadening) (Levitt, 2008). See Fig. 3.24 in Schmidt-Rohr and Spiess (1994) for a graphical representation on the structural composition of line-broadening. In general, compounds with relatively slow molecular motion tend to be of larger molecular weight and have shorter T2 times (i.e. ‘broad’ line-widths), whereas compounds with relatively fast molecular motion tend to be of smaller molecular weight and have longer T2 times (i.e. ‘sharp’ line-widths) (Bloembergen et al., 1948; Keeler, 2010). Therefore, T2 experiments can preferentially ‘filter’ out NMR signal from ‘broad’ components to that of ‘sharp’ components based on differences in their T2 times (Keeler, 2010).

Here, for the first time, we carry out solution 31P NMR spectroscopy with T2 experiments on soil extracts to determine the structural composition of soil phosphomonoesters. We hypothesised that a broad signal would preferentially decay to that of sharp signals (i.e. myo-IHP and scyllo-inositol hexakisphosphate (scyllo-IHP)) in the phosphomonoester region based on differences in their T2 times.

Section snippets

Soil information

Two soils of the mineral (Ah) horizon under forest were sourced from the “Ecosystem Nutrition: Forest Strategies for limited Phosphorous Resources (SPP 1685)” project in Germany (Lang et al., 2017). The two soils collected in 2015 from the Bad Brückenau (BBR) and Vessertal (VES) sites are classified as Cambisols (IUSS Working Group WRB, 2015), which were developed on basalt and trachyandesite parent material, respectively. At each site, the litter layer was removed and approximately 5 cores

Soil extraction and solution phosphorus-31 nuclear magnetic resonance spectra

The four soils used in this study were chosen for differences in their chemical properties and distribution of soil organic P (Table 1). Concentrations of total P ranged from 1202 and 3793 mg kg−1 across all soils based on laboratory X-ray fluorescence (LXRF). Pools of total P in NaOH-EDTA extracts ranged from 309 to 1612 mg kg−1 across all soils. The extraction efficiency of total P by NaOH-EDTA ranged from 26 to 42% of the total soil P by LXRF, which is consistent with previous studies (

Acknowledgements

Financial support from the Swiss National Science Foundation is gratefully acknowledged (Project: 200021_169256). The authors would like to thank Dr. Laurie Paule Schönholzer and Ms. Doris Sutter for technical assistance. The authors are grateful to Dr. Federica Tamburini, Dr. Chiara Pistocchi, and Dr. Astrid Oberson for providing soil samples. Lastly, we thank the anonymous reviewers who provided constructive feedback on the manuscript.

References (61)

  • K. Kouno et al.

    Measurement of soil microbial biomass phosphorus by an anion exchange membrane method

    Soil Biol. Biochem.

    (1995)
  • T.I. McLaren et al.

    Spectral sensitivity of solution 31P NMR spectroscopy is improved by narrowing the soil to solution ratio to 1:4 for pasture soils of low organic P content

    Geoderma

    (2015)
  • K. Schmidt-Rohr et al.

    Chapter 3 - high-resolution NMR techniques for solids

  • J.H. Steward et al.

    Gel chromatography of soil organic phosphorus

    J. Chromatogr. A

    (1971)
  • N. Bloembergen et al.

    Relaxation effects in nuclear magnetic resonance absorption

    Phys. Rev.

    (1948)
  • B.J. Cade-Menun et al.

    Solution phosphorus-31 nuclear magnetic resonance spectroscopy of soils from 2005 to 2013: a review of sample preparation and experimental parameters

    Soil Sci. Soc. Am. J.

    (2014)
  • B.J. Cade-Menun et al.

    Soil and litter phosphorus-31 nuclear magnetic resonance spectroscopy

    J. Environ. Qual.

    (2002)
  • B.J. Cade-Menun et al.

    Characterizing the phosphorus forms extracted from soil by the Mehlich III soil test

    Geochem. Trans.

    (2018)
  • L.M. Condron et al.

    Molecular weight distribution of soil organic phosphorus under irrigated pasture in New Zealand

    J. Soil Sci.

    (1989)
  • L.M. Condron et al.

    Chemistry and dynamics of soil organic phosphorus

  • R.C. Dalal

    Soil organic phosphorus

    Adv. Agron.

    (1977)
  • A.L. Doolette et al.

    Quantitative analysis of 31P NMR spectra of soil extracts - dealing with overlap of broad and sharp signals

    Magn. Reson. Chem.

    (2015)
  • A.L. Doolette et al.

    Spiking improved solution phosphorus-31 nuclear magnetic resonance identification of soil phosphorus compounds

    Soil Sci. Soc. Am. J.

    (2009)
  • A.L. Doolette et al.

    Rapid decomposition of phytate applied to a calcareous soil demonstrated by a solution 31P NMR study

    Eur. J. Soil Sci.

    (2010)
  • A.L. Doolette et al.

    The composition of organic phosphorus in soils of the Snowy Mountains region of South-Eastern Australia

    Soil Res.

    (2017)
  • W.J. Dougherty et al.

    On the use of hydrofluoric acid pretreatment of soils for phosphorus-31 nuclear magnetic resonance analyses

    Soil Sci. Soc. Am. J.

    (2007)
  • T.S. George et al.

    Organic phosphorus in the terrestrial environment: a perspective on the state of the art and future priorities

    Plant Soil

    (2018)
  • Z. He et al.

    Spectral and chemical characterization of phosphates associated with humic substances

    Soil Sci. Soc. Am. J.

    (2006)
  • J.E. Hill et al.

    Phosphorus-31 nuclear magnetic resonance spectroscopy transect study of poultry operations on the Delmarva peninsula

    J. Environ. Qual.

    (2009)
  • IHSS

    Source Materials for IHSS Samples

    (2013)
  • Cited by (17)

    • Phosphorus speciation along a soil to kettle hole transect: Sequential P fractionation, P XANES, and <sup>31</sup>P NMR spectroscopy

      2023, Geoderma
      Citation Excerpt :

      Average concentrations of assigned P compounds within the broad signal were more than two times higher in the sediments compared to the soil samples (Table 5). Compounds forming this broad signal were previously assigned to complex forms of Po in apparent high molecular weight material especially present in macro-molecular structures and associated with the soil OM (McLaren et al., 2015b; McLaren et al., 2019; Reusser et al., 2020a; Alewell et al., 2021). The predominance of phosphomonoesters in the soil as well as sediment extracts (Table 5) coincides with several investigations of soils, where phosphomonoesters also made up major proportions of Po in NaOH-EDTA extracts (e.g. Turner et al., 2003b; Alewell et al., 2020; McLaren et al., 2020).

    • Phosphorus species in sequentially extracted soil organic matter fractions

      2023, Geoderma
      Citation Excerpt :

      The molecules comprising the unresolved phosphomonoesters are of apparent large molecular weight, presumably complex in structure, and resistant to enzymatic hydrolysis (He et al., 2006; Jarosch et al., 2015; McLaren et al., 2020; McLaren et al., 2015). Furthermore, there is increasing evidence this pool contains several components of varying molecular weights rather than a single, large and polymeric macromolecule (McLaren et al., 2019). The occurrence of unresolved phosphomonoesters in large molecular weight factions of alkaline (NaOH) soil extracts suggests a close association with apparently large SOM structures, because by definition, humic and fulvic acids are extracted with alkali from soil (Stevenson, 1994).

    • The molecular size continuum of soil organic phosphorus and its chemical associations

      2022, Geoderma
      Citation Excerpt :

      This was similarly the case for T2 experiments carried out on soil extracts following hypobromite oxidation (Reusser et al. (2020b). McLaren et al. (2019) compared the line width of the broad signal at half peak intensity based on spectral deconvolution with that derived from T2 experiments. The authors reported that the line width at half peak intensity of the broad signal based on the former approach was significantly larger than that based on the latter approach, suggesting the broad signal itself is derived from multiple components.

    • Is the enzymatic hydrolysis of soil organic phosphorus compounds limited by enzyme or substrate availability?

      2019, Soil Biology and Biochemistry
      Citation Excerpt :

      The soils studied here were previously characterised using 31P nuclear magnetic resonance (NMR) spectroscopy and adding enzymes to alkaline soil extracts (Jarosch et al., 2015). This identified inositol phosphate P as the largest identifiable organic P class in most soils, followed by non-phytate phosphomonoesters (“monoester-like P″, excluding yet unidentified monoesters (McLaren et al., 2019)). We hypothesized that the hydrolysis of phytate as a typically dominant soil organic P class is enzyme-limited, whereas the hydrolysis of less abundant organic P classes (non-phytate phosphomonoesters and phosphodiesters) is substrate-limited.

    • Influence of long-term phosphorus fertilisation history on the availability and chemical nature of soil phosphorus

      2019, Geoderma
      Citation Excerpt :

      See Suppl. Fig. 1 for a detailed view of the orthophosphate and phosphomonoester region. Spectral deconvolution fitting was carried out on the orthophosphate and phosphomonoester region due to overlapping NMR signals, which was carried out using Matlab (The MathWorks Inc.; Natick, MA), as described in McLaren et al. (2019). The following compounds arising from sharp signals were identified: orthophosphate (δ 5.34 ppm), myo-inositol hexakisphosphate (δ 5.04, 4.09, 3.71 and 3.60 ppm), scyllo-inositol hexakisphosphate (δ 3.24 ppm), neoinositol hexakisphosphate in the 4-equatorial/2-axial (δ 5.93 and 3.76 ppm) and 2-equatorial/4-axial (δ 4.93 ppm) conformations, chiro-inositol hexakisphosphate in the 2-equatorial/4-axial (δ 5.69, 4.28 and 3.86 ppm) conformation, and several unidentified sharp signals.

    View all citing articles on Scopus
    View full text