Improved techniques for the characterisation of soil organic phosphorus using ³¹P nuclear magnetic resonance spectroscopy and their application to Australian soils.

Abstract

Organic phosphorus is potentially an important source of phosphorus (P) for agriculture, although it is not directly available for plant or microbial uptake. However, organic P can be converted into available inorganic P though hydrolysis or mineralisation. The rate of P release from organic P forms depends partly on the specific organic P compounds present in the soil. Until recently characterising soil organic P has been limited by the lack of appropriate analytic techniques. Consequently, organic P dynamics remains poorly understood. In this thesis, the focus was on improving techniques for the characterisation of soil organic P using solution ³¹P nuclear magnetic resonance (NMR) spectroscopy, applying these techniques to characterise a range of Australian soils and developing a better understanding of the cycling and potential bioavailability of soil organic P. The characterisation of soil organic P relies on the correct identification of resonances. Orthophosphate monoester peaks were identified by spiking model organic P compounds into NaOH- EDTA soil extracts. In this way, seven major resonances that were common to most of the NMR spectra were assigned to adenosine-monophosphate (AMP), scyllo-inositol hexakisphosphate, α- and β-glycerophosphate and myo-inositol hexakisphosphate (phytate). More importantly, spiking highlighted the similarly in appearance and chemical shift of some of the orthophosphate monoester resonances, particularly those of phytate and α- and β-glycerophosphate. This may have resulted in the misidentification and over-estimation of the concentrations of these species in previous studies. To provide a detailed quantitative assessment of soil organic P using ³¹P NMR spectroscopy, a modified method of spectral deconvolution, which included using an internal standard (methylenediphosphonic acid; MDP), was developed. The method of deconvolution implemented in this thesis considered P contained in larger humic molecules. A broad signal, in addition to the routinely fitted sharp peaks, was fitted to the orthophosphate monoester region of the NMR spectrum. A large proportion of monoester P (32–78%) could be assigned to this signal. When the broad signal was not taken into account phytate concentrations were over-estimated by 54%. It is likely that the concentrations of other specific orthophosphate monoester compounds were also over-estimated. The potential over-estimation of phytate concentrations has implication for the understanding of phytate stability in soils. High phytate concentrations in soils are usually explained by the stability of phytate in soils or the limited presence or activity of specialised enzymes (phytase). Lower phytate concentrations suggest phytate maybe less stable in soils than previously supposed. Therefore, the rate of phytate degradation in a calcareous soil was investigated. Phytate was applied to a calcareous soil at four different concentrations (ranging from 58–730 mg kg⁻¹) and the effect of wheat straw as an additional source of carbon was also examined. Regardless of treatment, phytate concentrations decreased over the 13-week incubation period and were adequately fitted to a first order decay model. There was no clear trend in the rate of phytate loss with treatment and the half life of phytate ranged from 4 to 8 weeks. The loss of phytate coincided with an increase in orthophosphate concentration, that in some cases more than doubled the native soil P concentrations, and there was very little variation in extraction efficiency. This result provided evidence for the microbial degradation of phytate. It demonstrated that in the calcareous soil examined, phytate was not highly stable, but a bioavailable source of organic P. The composition of soil P in 18 diverse Australian soils was also examined. Across all NaOH-EDTA soil extracts analysed, phytate comprised up to 9%, but averaged only 3% of total extractable P. Two other resonances that were also prominent in all the ³¹P NMR spectra and comprised a similar proportion of total organic P were due to α- and β-glycerophosphate. By examining the alkaline hydrolysis of a phospholipid (phosphatidlycholine), the potential source of α- and β-glycerophosphate was identified. Although α- and β-glycerophosphate and phyate gave rise to the most intense peaks, the broad signal, which was attributed to humic P, represented the most abundant form of soil organic P (27–72% of total extractable organic P). Therefore, it was suggested that the development of methods that aim to increase the availability of stabilised forms of organic P should give preference to increasing the availability of P contained in humic P complexes. Understanding P cycling not only relies on analytical methods that enable the accurate identification and quantification of soil organic P but also requires methods that can gauge the susceptibility of different organic P species to enzymatic hydrolysis. Therefore, enzymatic hydrolysis was combined with ³¹P NMR spectroscopy to identify and compare the specific organic P species in the enzyme labile and non-enzyme labile fractions of a range of NaOH-EDTA soil extracts. Phosphorus-31 NMR analysis of NaOH-EDTA soil extracts treated with active and inactivated phytase enzyme preparations showed that phytase hydrolysed the majority of the small, orthophosphate monoester compounds (α- and β-glycerophosphate, phytate, scyllo-inositol hexakisphosphate) and pyrophosphate, but orthophosphate diesters (DNA) and humic P were generally unaffected. The ³¹P NMR spectra revealed that not only was organic P hydrolysed but new orthophosphate monoester species were formed, possibly as a result of enzymatic phosphorylation. Although combining enzymatic hydrolysis and ³¹P NMR spectroscopy enabled the identification of individual organic P species that were susceptible or resistant to enzyme hydrolysis, there is still a need for further improvement and refinement of the technique in order to provide an accurate estimate of the potentially available fraction of soil organic P.