Phosphorus NMR in biology: phospholipids, phosphorylated proteins, oligonucleotide drugs and phosphates in food
Share
Phosphorus NMR in biology: phospholipids, phosphorylated proteins, oligonucleotide drugs and phosphates in food
What a 31P spectrum tells about a biological sample, which questions it answers better than chromatography or mass spectrometry, and which it does not
Phospholipids, nucleotides, nucleic acids, phosphorylated proteins and the phosphates added to food all contain phosphorus, an element whose only natural isotope, 31P, is observed directly by NMR. The 31P spectrum of a biological sample therefore contains the signals of its phosphorus compounds and of nothing else; triglycerides, sugars and solvents do not appear in it.
Each kind of phosphorus atom gives one signal, and in a spectrum recorded under quantitative conditions the area of a signal is proportional to the number of nuclei that give it.[1] A single spectrum thus shows which phosphorus compounds a sample contains and how much of each, without a separation and without a reference sample of every compound.[2] In a comparison on eight commercial lecithins, 31P NMR quantified all phospholipids and lysophospholipids, whereas high-performance liquid chromatography (HPLC) failed to determine all but one of the lysophospholipids.[3] The method has also found what other methods lose or do not see: phosphate on histidine, most of which is lost in the sample preparation for mass spectrometry,[4] and differences between batches of an oligonucleotide drug which HPLC and mass spectrometry could not distinguish.[5] Its weakness is sensitivity: in the lecithin study the detection limits of NMR were 10 to 70 times higher than those of the chromatographic methods.[3]
Five terms used in this article. The chemical shift is the position of a signal in the spectrum, given in parts per million (ppm); it depends on the chemical surroundings of the phosphorus atom. A phospholipid class consists of all phospholipids that share one head group, for example phosphatidylcholine, whatever their fatty acids (Figure 2). A lysophospholipid is a phospholipid from which one of the two fatty acids has been removed. In a phosphorothioate linkage, sulfur replaces one oxygen atom of the phosphate group that joins two nucleosides.[5] Diastereomers are molecules with the same atoms and the same bonds that differ in the spatial arrangement at one or more atoms and are not mirror images of each other.
At a glance: the question and the answer from the 31P spectrum
| Question | What the spectrum tells | Limit |
|---|---|---|
| Which phospholipids does a lecithin, a milk fraction or a krill oil contain? | Every class and its lyso forms, each as a molar amount, against one internal standard | The classes, not the individual fatty acids; detection limits 10 to 70 times higher than those of HPLC |
| How much phosphate do the proteins of a cell carry on histidine? | Phosphohistidine and phosphoserine with phosphothreonine, in micromoles per gram of protein | Not which protein or which site; phosphotyrosine was not observed; 14 to 15 hours of measurement |
| Which shape does a nucleotide-binding protein adopt in solution? | Two conformations of the protein Ras in equilibrium, from the signals of the bound nucleotide | A purified protein is required; shown for Ras and three related proteins |
| Are two batches of a phosphorothioate oligonucleotide the same? | Differences in the signals at 54 to 56 ppm between batches that HPLC and mass spectrometry found identical | The cause of the differences remained tentative |
| Was polyphosphate added to prawns or meat? | Diphosphate and triphosphate beside the natural orthophosphate | Added polyphosphates are hydrolysed rapidly in the product |
| What is the pH inside a living cell? | The pH from the position of the phosphate signal, without destroying the cell | Requires a calibration for the temperature and the salt content of the sample[6] |
Where the signals appear
The signals discussed in this article occupy three regions of the spectrum (Figure 1). Nucleoside phosphoramidites, the starting materials of oligonucleotide synthesis, appear near 150 ppm.[7] The phosphorothioate linkages of oligonucleotides appear between 54 and 56 ppm, and ordinary phosphodiester linkages near 0 ppm.[5] Phospholipids, orthophosphate and pyrophosphate are found between 6 and −11 ppm.[6, 8] Phosphohistidine in proteins appears in the same region, near −5 ppm at pH 10.8.[4]
Figure 1. Positions of the 31P signals discussed in this article. Our drawing of published values: phosphoramidites;[7] phosphorothioate and phosphodiester linkages, mongersen in D2O;[5] phospholipid classes in a detergent solution at pH 7.7;[8] orthophosphate and pyrophosphate, limiting values of the pH titration at 300 K;[6] triphenyl phosphate, the position required by a monograph of the United States Pharmacopeia (USP).[9] The sources differ in medium and in shift reference: trimethyl phosphate appears at 2.98 ppm in [8] and is set to 3.80 ppm in [6].
Within one group of compounds the signals lie close together. The phospholipid classes of Figure 1 extend over 4.3 ppm, and several pairs of classes are separated by less than 0.1 ppm.[8] The signals of orthophosphate and pyrophosphate move by several ppm when the pH changes.[6] A laboratory therefore keeps the solvent, the pH and the temperature of the sample constant.
Phospholipids: which classes, and how much of each
A phospholipid class is defined by its head group (Figure 2). Every class shown contains one phosphorus atom per molecule, so that the area of a signal is a direct measure of the number of molecules of that class.
Figure 2. Six phospholipid classes. Phosphatidylcholine, phosphatidylethanolamine, phosphatidylserine and phosphatidylinositol are esters of glycerol with two fatty acids and differ in the head group. Lysophosphatidylcholine has lost one fatty acid; one of its two positional isomers is shown. Sphingomyelin is built on sphingosine in place of glycerol. Our drawings; the configuration of the inositol ring is not shown.
Lecithins: the lyso forms that HPLC does not separate
Lecithins are commercial mixtures of phospholipids, obtained for example from soya beans or egg yolk.[10] In 2003 Helmerich and Koehler, of the German Research Centre for Food Chemistry in Garching, determined the phospholipid classes of eight commercial lecithins and three flour improvers by three methods: thin-layer chromatography (TLC), HPLC and 31P NMR. The methods gave comparable total amounts and comparable amounts of the individual classes, although the values depended on the method used.[3]
Three methods on eight lecithins and three flour improvers
| Method | Classes not quantified | Detection limit; variation |
|---|---|---|
| TLC | Lysophosphatidylcholine | 20–170 μg/mL; 5.5 % |
| HPLC | All lysophospholipids except lysophosphatidylethanolamine | 20–170 μg/mL; 6.8 % |
| 31P NMR | None | 10 to 70 times higher; 12.8 % |
Values from the abstract of [3], which gives one range of detection limits for HPLC and TLC together. Variation: the coefficient of variation, which is the standard deviation divided by the mean.
The authors recommended 31P NMR for the quantification of phospholipids, TLC where little equipment is available, and HPLC where high sensitivity is required.[3]
The international standard for the HPLC analysis of lecithins acknowledges the same limitation. ISO 11701:2009 specifies HPLC with a light-scattering detector for lecithins from vegetable fats and oils. It excludes animal and ruminant lecithins and enzymatically hydrolysed lecithins, on the grounds that three lyso forms are insufficiently separated.[11]
The lyso forms are products of hydrolysis (Figure 3). A lecithin that has been treated with a phospholipase contains them in larger amounts, and it is such samples that the HPLC standard excludes. In the 31P spectrum a lyso form gives a signal of its own beside that of its parent class: a table published in 2008 by a US supplier of lipids places phosphatidylcholine at −0.82 ppm and two of its lyso forms at −0.35 and −0.55 ppm (Figure 4).[8]
Figure 3. Hydrolysis of phosphatidylcholine by phospholipase A1, the enzyme used in [12]. The enzyme removes the fatty acid at position 1 of glycerol, so the product is the positional isomer that Figure 2 does not show. Our drawing.
In 2017 Yang, Hiserodt and Li, of International Flavors & Fragrances, hydrolysed six phospholipid classes separately with phospholipase A1, followed each reaction by 31P NMR, and assigned 26 phosphorus-containing products. These assignments made it possible to determine the composition of commercial lecithins routinely and without phospholipid standards.[12]
Figure 4. Positions of phospholipid classes in a detergent solution at pH 7.7 and 400 MHz. Our drawing of published values: the positions are those of a table that a US supplier of lipids published in 2008;[8] the heights and widths are arbitrary. PC, PE, PS and PI are phosphatidylcholine, phosphatidylethanolamine, phosphatidylserine and phosphatidylinositol; SM is sphingomyelin. Sphingomyelin and phosphatidylethanolamine are 0.04 ppm apart, which is 6.5 Hz at the 31P frequency of 162 MHz (our calculation).
Milk phospholipids: no suitable calibration standards
Milk phospholipids are rich in sphingomyelin and phosphatidylserine, and their fatty acids differ from those of vegetable lecithins. In 2009 MacKenzie, Vyssotski and Nekrasov, of Industrial Research Limited in Lower Hutt, New Zealand, set out the consequences for HPLC. With ultraviolet (UV) detection the response depends on the degree of unsaturation of the fatty acids, which makes standards prepared from soya lecithin unsuitable for dairy samples. Light-scattering detection, which the International Lecithin and Phospholipid Society had recommended after its trials, proved useful for soya lecithin but failed for more complex lecithins, those from milk among them.[2]
The group dispersed the samples in a detergent solution, recorded 31P spectra and validated the results against quantitative two-dimensional TLC. For the soya lecithin standard of the society, their values agreed closely with the certificate of analysis.[2]
Krill oil: a pharmacopoeial test
The USP monograph Krill Oil prescribes 31P NMR for the content of total phospholipids. In the text of USP 38–NF 33 it requires 28 to 52 % (w/w) of total phospholipids, of which 60 to 90 % must be phosphatidylcholine.[9] The definition of the current monograph gives other limits: 30 to 59 % (w/w) of total phospholipids, of which 60 to 96 % are phosphatidylcholine together with its two lyso forms.[13] The lyso forms, the hydrolysis products of phosphatidylcholine, are thus now counted with it. In 2018 Zailer, Monakhova and Diehl, of a German NMR contract laboratory, reported a ring trial of a 31P method for krill oil, in which fourteen laboratories in five countries measured five oils and returned 17 data sets. The relative standard deviation between the laboratories was 2.1 to 3.3 % for the content of phosphatidylcholine and 2.8 to 12.3 % for the sum of the other phospholipids. The principal source of the scatter was the hydrolysis of phosphatidylcholine during the 14 weeks that preceded the measurements.[14]
A 31P integral counts phosphorus atoms, and the result for each class is therefore a molar amount. Its conversion into a mass fraction requires an average molecular mass for the class. Kato and co-workers at a Japanese food testing laboratory (2018) assumed that every molecule contains two stearic acid residues,[10] whereas a US patent on krill oil derived the masses from the fatty acid profile of the sample.[15]
Tissues: lipid classes for lipidomics
Lipidomics is the analysis of all lipids of a sample, and it is mostly done by mass spectrometry. In 2021 Furse and co-workers, from the University of Cambridge, the Royal Botanic Gardens at Kew, the University of Nottingham and King's College London, described the complete 31P NMR procedure from the extraction of a tissue to the analysis of the spectrum. Their aim was to make the method accessible to laboratories that are new to it. They analysed liver, heart and brain of the mouse and soya lecithin. The lipids were dissolved in a mixture of dimethylformamide, triethylamine and guanidinium chloride, and the spectra were measured on an 800 MHz spectrometer with 1,312 scans. Among the classes assigned is cardiolipin, a lipid of mitochondria with two phosphorus atoms per molecule. The result is the molar percentage of each class in the total phospholipid.[16]
The authors name two reasons to measure 31P beside mass spectrometry. NMR does not require the molecules to be ionised, whereas every method based on mass spectrometry depends on how efficiently each lipid is ionised; and the class composition from NMR supports the interpretation of the mass spectrometry data. They also name the limits: at least 4 mg of lipid from a biological extract is needed, and phosphatidylcholine is hydrolysed as the dissolved sample ages.[16]
Proteins: phosphate that mass spectrometry loses
Cells regulate proteins by attaching phosphate groups to them. On serine, threonine and tyrosine the phosphate is bound to oxygen and withstands acid. On histidine it is bound to nitrogen, and this bond is unstable at neutral and acid pH (Figure 5).[4] The usual analysis of phosphorylated proteins by mass spectrometry digests the proteins with trypsin for many hours at pH 8 and then desalts the peptides.[4]
Figure 5. Four phosphorylated amino acids. In phosphoserine, phosphothreonine and phosphotyrosine the phosphate is bound to oxygen; in phosphohistidine it is bound to nitrogen. The isomer shown, 3-phosphohistidine, is the one that dominated in the cell lysate of [4]. Our drawings; the amino acids are drawn in the neutral form.
In 2022 Makwana, Williamson, Jackson and Muimo, of the University of Sheffield, measured phosphohistidine in the proteins of a human cell line without exposing it to these conditions. They lysed bronchial epithelial cells directly in a carbonate buffer of pH 10.8 that contained 7 M urea, removed free phosphate and other small molecules through a membrane with a cut-off of 3 kDa, and recorded the 31P spectrum of the proteins.[4]
Phosphate on the proteins of a human cell line
| Phosphorylated amino acid | Per gram of protein | Per cell |
|---|---|---|
| Phosphohistidine | 23 μmol | 1.8 fmol |
| Phosphoserine and phosphothreonine together | 68 μmol | 5.8 fmol |
Values of [4] for a lysate of human bronchial epithelial cells (16HBE14o-).
Phosphohistidine was thus about one third as abundant as phosphoserine and phosphothreonine combined, and by the estimate of the authors roughly 15 times more abundant than phosphotyrosine.[4] The authors also showed why mass spectrometry reports less. When myoglobin phosphorylated on histidine was digested with trypsin under typical conditions (pH 8.0, 37 °C, 16 hours), most of the phosphohistidine was hydrolysed: the ratio of its signal to that of free phosphate fell from 15.8 to 0.97.[4]
The spectrum gives the total amount of each kind of phosphorylated amino acid in the proteins. It does not identify the proteins or the sites, which remains the task of mass spectrometry. Phosphotyrosine could not be observed, because its signal overlaps with those of orthophosphate, and the study covers one cell line. Each measurement took 14 to 15 hours on 0.3 mL of the protein solution.[4]
Ras: the bound nucleotide reports the shape of the protein
A protein need not be phosphorylated to be studied by 31P NMR; a bound nucleotide is sufficient. Ras is a protein that binds guanosine triphosphate (GTP) and, while GTP is bound, passes a signal on to other proteins, its effectors. In 2001 Spoerner, Herrmann, Vetter, Kalbitzer and Wittinghofer, of the University of Regensburg and the Max Planck Institute of Molecular Physiology in Dortmund, studied Ras with GppNHp, an analogue of GTP. Earlier 31P spectra of the bound nucleotide had indicated two conformations of the protein in rapid equilibrium. One of them, state 2, corresponds to the structure found in complexes with effectors. Two mutants with a reduced affinity for effectors occurred predominantly in the other conformation, state 1. The addition of an effector moved the signals of one of the two mutants to the positions of state 2. In the crystal structure of this mutant the region that binds the effector was invisible, which indicates that it is mobile.[17]
In 2010 the Regensburg group found the same equilibrium for Ras with GTP itself. State 1 is weakly populated and has only a very low affinity for effectors; state 2 dominates and binds them with high affinity. For a shortened form of the protein at 5 °C the equilibrium constant between the two states was 11.3, which corresponds to about 8 % of state 1. Two states had been shown earlier for three related proteins, Ran, Cdc42 and Ral.[18]
Oligonucleotide drugs: batches that HPLC and mass spectrometry did not distinguish
Many antisense drugs are oligonucleotides in which phosphorothioate linkages replace the natural phosphodiester linkages. The phosphorus atom of a phosphorothioate linkage carries four different groups, so that the linkage exists in two spatial arrangements, designated RP and SP (Figure 6). Because common synthetic methods do not control the arrangement, each linkage is formed in both,[5] and an oligonucleotide with n such linkages consists of up to 2n diastereomers.[5]
Figure 6. Top: the two arrangements of a phosphorothioate linkage, drawn for two thymidines. Bottom: the backbone of mongersen. Our drawings; the descriptors RP and SP were assigned by us with the Cahn–Ingold–Prelog rules.
Mongersen is an antisense oligonucleotide of 21 nucleotides and 20 phosphorothioate linkages, which corresponds to 1,048,576 possible diastereomers. It was developed as an oral drug for Crohn's disease and acts by lowering the amount of the protein SMAD7. A phase 3 study was discontinued in 2017 for lack of efficacy.[5]
In 2022 a group from the University of Pisa and the University of Rome Tor Vergata, with co-authors from Nogra Pharma, the company that funded the study, and from two service companies, examined 27 batches, most of which had been used in clinical trials. The chemical purity was 92.1 ± 0.7 %. Reversed-phase HPLC, liquid chromatography with mass spectrometry (LC-MS), X-ray powder diffraction, infrared and Raman spectroscopy, calorimetry and circular dichroism revealed no significant differences between the batches; the 31P spectra in solution did.[5]
Each phosphorus atom of the oligonucleotide gives a single line between 54 and 56 ppm. The lines lie close together, and the pattern they form depends, among other effects, on the distribution of the RP and SP arrangements along the chain. Since most chemical impurities give no signal in this narrow range, the differences between the batches can arise from the stereochemistry.[5] The authors cite earlier work according to which the formation of the linkage is not random, and a synthesis yields the same mixture of diastereomers only under rigorously identical conditions and with the same starting materials.[5]
The batches differed in a cell assay as well. Cells of the line HCT-116 were transfected with each batch, and SMAD7 was measured by Western blot. A principal component analysis of the 31P spectra separated the batches that inhibited SMAD7 weakly from those that resembled the reference batch.[5] The authors attribute the differences in activity tentatively to the diastereomeric composition and state that this attribution is difficult to confirm.[5]
31P NMR is not the usual method for the ratio of phosphorothioate to phosphodiester linkages. At a USP workshop on therapeutic peptides and oligonucleotides in March 2021, a speaker from Roche described NMR as not the method of choice for this ratio, but as useful for the detection of duplexes and of interactions between the drug substance and excipients or metal ions.[19]
Food: was phosphate added?
Diphosphates, triphosphates and polyphosphates are additives allowed by law, which increase the water-binding capacity of meat and fish products.[20] The analyst has to establish whether they have been added. A UK food laboratory offers an NMR test for prawns that distinguishes the forms of phosphate (Figure 7). The test relies on the fact that orthophosphate, which the laboratory calls monophosphate, occurs naturally in prawns, whereas diphosphate and triphosphate do not; their detection confirms the use of polyphosphates.[20]
Figure 7. Orthophosphate and three condensed phosphates. The condensed phosphates are hydrolysed in the product, and the end product is orthophosphate.[21] Our drawings.
Between 2009 and 2011 Jastrzębska, Hrynczyszyn and Szłyk, of Nicolaus Copernicus University in Toruń, Poland, published three studies of 31P NMR for added phosphates in pork, meat products and seafood. They recovered at least 95 % of the added phosphorus compounds, with a coefficient of variation of at most 5 %.[22, 23] They also named the principal difficulty of the analysis: polyphosphates added to meat and seafood are rapidly hydrolysed to phosphates, and in their comparison capillary isotachophoresis, an electrophoretic method, was simpler, faster and more economical than NMR.[21] A treated product may therefore contain only orthophosphate after storage.
Living cells: the pH from the phosphate signal
The position of a phosphate signal depends on the pH, because the signal moves when the phosphate group gains or loses a proton (Figure 8). In 2024 Carta and Scorciapino, of the University of Cagliari, fitted the titration curves of orthophosphate and pyrophosphate at 300 K. The signal of orthophosphate moves from 0.86 ppm for H2PO4− to 3.34 ppm for HPO42−, and the pKa of this step is 6.76.[6] Near this pH a change of 0.1 pH unit displaces the signal by 0.14 ppm (our calculation).
Figure 8. Chemical shift of orthophosphate and pyrophosphate against pH, with the two forms of orthophosphate that exchange near pH 7. Our calculation from the dissociation constants and limiting shifts of [6] (300 K, ionic strength 0.15 to 0.20 M).
The dependence turns the phosphate naturally present in a cell into a pH indicator. In 1973 Moon and Richards, of the California Institute of Technology, recorded 31P spectra of red blood cells and determined the pH inside the cells from the positions of the signals of inorganic phosphate and 2,3-diphosphoglycerate, without destroying the cells.[24] From the shifts of orthophosphate and pyrophosphate together, Carta and Scorciapino obtained the pH of solutions between pH 5 and 12 with an average deviation of 0.09 pH unit from the glass electrode.[6]
Phosphorus spectra are also measured on people, in a magnetic resonance scanner. A consensus paper of 2021 by fifteen specialists describes the method for skeletal muscle and calls it the oldest method of magnetic resonance spectroscopy in metabolic research on living subjects. The pH inside the muscle is calculated from the distance between the signals of inorganic phosphate and phosphocreatine, the energy reserve of muscle. For the resting calf muscle of healthy people the paper lists, as means of published studies, a pH of 7.03 and a signal ratio of phosphocreatine to adenosine triphosphate (ATP) of 4.23. After exercise the phosphocreatine signal recovers with a time constant of about 41 s. This recovery is analysed as a measure of the function of the mitochondria, and the authors point out pitfalls of the analysis.[25]
Nucleotides give 31P signals as well. A nucleoside monophosphate such as β-nicotinamide mononucleotide contains one phosphorus atom, whereas ATP contains three in different environments and gives three signals (Figure 9). We found no published comparison of 31P NMR with HPLC for the purity of nucleotides, and we do not claim that NMR is the preferable method for this purpose.
Figure 9. Adenosine 5′-triphosphate and β-nicotinamide mononucleotide. The phosphorus atoms of ATP are named α, β and γ from the ribose outwards. Our drawings; ATP is drawn as the free acid.
Available from our catalogue: β-nicotinamide mononucleotide, CAS 1094-61-7.
How much sample, how long, and what the method cannot do
The amounts of sample and the measuring times in the published work differ widely, because they depend on the concentration of phosphorus in the sample.
Sample and time in the cited studies
| Sample | Amount | Measurement |
|---|---|---|
| Krill oil, USP monograph[9] | 300–350 mg | At least 512 scans at 300 MHz, at least 128 minutes (our arithmetic) |
| Krill oil, detergent solution[15] | 20–40 mg | 576 scans at 121.5 MHz for 31P |
| Lecithin in a detergent solution[8] | Classes above 2 % (w/w) of the lecithin are quantified | 512 scans in 34 minutes at 400 MHz |
| Egg lecithin, 1986[26] | 100 mg in 3 mL | 10 mm tube, 161.7 MHz for 31P |
| Lipid extract of a tissue, 1988[27] | 0.01–100 mg of lipid extract in 2 mL of chloroform | About 15 minutes per sample at 11.75 T |
| Lipid extract of mouse tissue, 2021[16] | 5–10 mg of phospholipid | 1,312 scans on an 800 MHz spectrometer; the time is not stated |
| Oligonucleotide drug[5] | About 9 mg in 0.5 mL | 2,000 scans at 14.1 T |
| Proteins of a cell lysate[4] | 0.3 mL of protein solution | 14 to 15 hours at 162 or 202.5 MHz for 31P |
161.7 and 162 MHz are the 31P frequencies of instruments with a 1H frequency of 400 MHz; 121.5 MHz and 202.5 MHz are those of 300 MHz and 500 MHz instruments; 11.75 T and 14.1 T correspond to 1H frequencies of 500 and 600 MHz; the 800 MHz spectrometer observes 31P at 324 MHz.
For the amounts to be correct, the laboratory measures against a known quantity of a standard, which is added to the sample or kept in a sealed capillary inside the tube,[4, 9] keeps the pH constant and allows every signal to recover fully between two pulses; Part 9 of this series explains the last requirement. The procedure of Kato and co-workers, for example, keeps the pH within ± 0.04.[10]
Questions that 31P NMR answers less well
| Question | Reason |
|---|---|
| A phosphorus compound at low concentration | The detection limits of NMR were 10 to 70 times higher than those of HPLC and TLC[3] |
| Which protein is phosphorylated, and at which site | The spectrum gives the total amount of each phosphorylated amino acid[4] |
| Ratio of phosphorothioate to phosphodiester linkages | NMR is not the method of choice[19] |
| Routine control of added polyphosphate | Capillary isotachophoresis was simpler, faster and more economical[21] |
Summary
- A 31P spectrum contains the signals of the phosphorus compounds of a sample and of nothing else, and the area of each signal gives the amount.
- In lecithins, 31P NMR quantified all phospholipid classes and their lyso forms. HPLC did not determine most lysophospholipids, and the ISO method by HPLC excludes animal and enzymatically hydrolysed lecithins.
- The USP monograph for krill oil, in the text of USP 38–NF 33, prescribes 31P NMR; in a ring trial of fourteen laboratories the relative standard deviation for phosphatidylcholine was 2.1 to 3.3 %.
- In extracts of mouse liver, heart and brain, one spectrum gives the molar percentage of each phospholipid class; this composition supports the mass spectrometry of the same lipids.
- In the proteins of a human cell line, 31P NMR found 23 μmol of phosphohistidine per gram of protein, about one third of the amount of phosphoserine and phosphothreonine. Most of the phosphohistidine is lost in the usual sample preparation for mass spectrometry.
- The 31P signals of the nucleotide bound to the protein Ras show two conformations of the protein in equilibrium; only one of them binds effector proteins with high affinity.
- Twenty-seven batches of the oligonucleotide drug mongersen were indistinguishable by HPLC and LC-MS but differed in their 31P spectra and in a cell assay.
- In prawns, the detection of diphosphate and triphosphate shows that polyphosphate was added, although added polyphosphates are rapidly hydrolysed in the product.
- The position of the phosphate signal gives the pH, also inside intact cells and in human muscle: it moves by 0.14 ppm for each 0.1 pH unit near pH 6.8.
- The method gives total amounts and needs milligram quantities: it does not identify phosphorylation sites, and its detection limits are 10 to 70 times higher than those of HPLC.
Phosphorus NMR from NorrChemica
Our NMR service records 1H and 31P spectra of one sample as Package II, in solution at 400 MHz and priced per sample; a quantitative measurement against a calibrated standard is Package V. The service does not accept biohazardous samples or material of animal origin. It accepts samples of plant or synthetic origin, such as soya or sunflower lecithin, synthetic phospholipids, nucleotides and synthetic oligonucleotides. Please describe the sample and the question on the request form. Research use only.
- β-Nicotinamide mononucleotidenucleotide with one phosphorus atom
- 5 mm NMR tubes for quantitative NMRfor integrals against an internal standard
- 10 mm NMR tubesthe tube diameter of the 1986 phospholipid method
- 5 mm precision NMR tubes, 300–800 MHz1H and 31P spectra
- 3 mm NMR tubesfor small amounts of sample
NMR tubes made in Germany by Deutero and shipped from Finland. See the NMR tube guide →
References
- Khoury S, Canlet C, Lacroix MZ, Berdeaux O, Jouhet J, Bertrand-Michel J. Quantification of lipids: model, reality, and compromise. Biomolecules 2018, 8, 174 (extracts). DOI
- MacKenzie A, Vyssotski M, Nekrasov E. Quantitative analysis of dairy phospholipids by 31P NMR. J. Am. Oil Chem. Soc. 2009, 86, 757–763 (passages of the introduction and of the results). DOI
- Helmerich G, Koehler P. Comparison of methods for the quantitative determination of phospholipids in lecithins and flour improvers. J. Agric. Food Chem. 2003, 51, 6645–6651 (abstract). DOI
- Makwana MV, Williamson MP, Jackson RFW, Muimo R. Quantitation of phosphohistidine in proteins in a mammalian cell line by 31P NMR. PLoS ONE 2022, 17, e0273797. DOI
- Arrico L, Stolfi C, Marafini I, Monteleone G, Demartis S, Bellinvia S, Viti F, McNulty M, Cabani I, Falezza A, Di Bari L. Inhomogeneous diastereomeric composition of mongersen antisense phosphorothioate oligonucleotide preparations and related pharmacological activity impairment. Nucleic Acid Ther. 2022, 32, 312–320. DOI
- Carta P, Scorciapino MA. The combination of inorganic phosphate and pyrophosphate 31P-NMR for the electrodeless pH determination in the 5–12 range. Magn. Reson. Chem. 2024, 62, 101–113 (accepted manuscript). DOI
- Oxford Instruments. Analysing phosphorus containing compounds using 31P benchtop NMR. Application Note 28 (extracts). Link
- Lanier KL, Moore JD, Smith D, Li S, Davis B, Shaw WA. Quantitative phospholipid analysis of soy lecithin and krill lecithin by 31P NMR. Avanti Polar Lipids, Alabaster, Alabama, September 2008 (archived copy of 28 December 2024). Link
- United States Pharmacopeia. Krill Oil. USP 38–NF 33, Second Supplement; test “Content of Total Phospholipids” (a copy of the monograph text). Link
- Kato T, Nishimiya M, Kawata A, Kishida K, Suzuri K, Saito M, Fujita K, Igarashi T, Inagaki M. Quantitative 31P NMR method for individual and concomitant determination of phospholipid classes in polar lipid samples. J. Oleo Sci. 2018, 67, 1279–1289 (abstract). DOI
- ISO 11701:2009. Vegetable fats and oils — Determination of phospholipids content in lecithins by HPLC using a light-scattering detector (abstract on the ISO website). Link
- Yang Y, Hiserodt R, Li J. Rapid identification and relative quantification of the phospholipid composition in commercial lecithins by 31P NMR. J. Am. Oil Chem. Soc. 2017, 94, 885–892 (abstract). DOI
- United States Pharmacopeia. Krill Oil. Dietary Supplement Monographs, USP–NF, 2022 (public preview: definition only). Link
- Zailer E, Monakhova YB, Diehl BWK. 31P NMR method for phospholipid analysis in krill oil: proficiency testing — a step toward becoming an official method. J. Am. Oil Chem. Soc. 2018, 95, 1467–1474 (abstract). DOI
- Bioeffective krill oil compositions. US Patent 9,644,169, Example 7.
- Furse S, Williams HEL, Watkins AJ, Virtue S, Vidal-Puig A, Amarsi R, Charalambous M, Koulman A. A pipeline for making 31P NMR accessible for small- and large-scale lipidomics studies. Anal. Bioanal. Chem. 2021, 413, 4763–4773. DOI
- Spoerner M, Herrmann C, Vetter IR, Kalbitzer HR, Wittinghofer A. Dynamic properties of the Ras switch I region and its importance for binding to effectors. Proc. Natl. Acad. Sci. USA 2001, 98, 4944–4949. DOI
- Spoerner M, Hozsa C, Poetzl JA, Reiss K, Ganser P, Geyer M, Kalbitzer HR. Conformational states of human rat sarcoma (Ras) protein complexed with its natural ligand GTP and their role for effector interaction and GTP hydrolysis. J. Biol. Chem. 2010, 285, 39768–39778. DOI
- Kibbey M, Healy N. Synthetic oligonucleotides: regulatory, analytical and manufacturing considerations. Report on the USP Workshop on Therapeutic Peptides and Oligonucleotides, March 2021. Bioprocess Online, 28 May 2021. Link
- Fera Science. Polyphosphates testing (web page of the test). Link
- Jastrzębska A, Hrynczyszyn P, Szłyk E. Comparison of cITP and 31P NMR methods for determination of polyphosphates in meat product and sea fruits. Czech J. Food Sci. 2009, 27 (Special Issue 1), S340 (abstract). DOI
- Jastrzębska A, Szłyk E. Application of 31P NMR for added polyphosphate determination in pork meat. Chem. Pap. 2009, 63, 414–419 (abstract). DOI
- Szłyk E, Hrynczyszyn P. Phosphate additives determination in meat products by 31-phosphorus nuclear magnetic resonance using new internal reference standard: hexamethylphosphoroamide. Talanta 2011, 84, 199–203 (abstract). DOI
- Moon RB, Richards JH. Determination of intracellular pH by 31P magnetic resonance. J. Biol. Chem. 1973, 248, 7276–7278 (abstract). DOI
- Meyerspeer M, Boesch C, Cameron D, Dezortová M, Forbes SC, Heerschap A, Jeneson JAL, Kan HE, Kent J, Layec G, Prompers JJ, Reyngoudt H, Sleigh A, Valkovič L, Kemp GJ. 31P magnetic resonance spectroscopy in skeletal muscle: experts' consensus recommendations. NMR Biomed. 2021, 34, e4246. DOI
- Sotirhos N, Herslöf B, Kenne L. Quantitative analysis of phospholipids by 31P-NMR. J. Lipid Res. 1986, 27, 386–392 (abstract and extracts of the experimental part). DOI
- Meneses P, Glonek T. High resolution 31P NMR of extracted phospholipids. J. Lipid Res. 1988, 29, 679–689 (abstract). DOI
About this article. For sources 4, 5, 16, 17, 18 and 25 and for Example 7 of source 15 we read the passages that carry the cited statements and numbers in the open full text in the browser. We compared the numbers of source 9 with a copy of the monograph text on a third-party site. Of source 13, the current monograph, only the definition is public; we have not read its test, and the description of the test in this article follows source 9. Sources 6, 8, 19 and 20 were read through a retrieval tool that returned extracts and summaries of the open full text; the values taken from 6 and 8 were confirmed in a second query. The original address of source 8 no longer opens, and the link leads to an archived copy. Sources 3, 10, 12, 14, 21, 22, 23, 24 and 27 could not be opened in full, because they are behind a paywall or are available as PDF files only; they are known to us from their abstracts or bibliographic records. Source 1 is known to us from extracts of the open full text shown by a search engine. For source 26 we read the abstract and extracts of the experimental part, and for source 2 passages of the introduction and of the results. Source 11 is the abstract of the standard on the ISO website. Source 7 was read as extracts. The year 1986 of source 26 follows from the volume number; two databases give 1988. The web page of source 20 names the method of its test as NMR and does not name the nucleus. Our own work: all structure drawings and the descriptors of Figure 6; the reaction scheme of Figure 3 and the statement on phospholipase A1, which is textbook chemistry; the calculated curves of Figure 8 and the value of 0.14 ppm for 0.1 pH unit; the number of diastereomers; the share of about 8 % of state 1, from the equilibrium constant of source 18; the time of 128 minutes; the conversion of 0.04 ppm to 6.5 Hz; the assignment of the run time of 34 minutes in source 8 to a delay of 3 s; the reading of the abbreviation TMP in source 8 as trimethyl phosphate, and of its entry LPC-2 as a second lyso form of phosphatidylcholine; the note on the two shift references in Figure 1; the statement that the protein spectrum does not identify proteins or sites, and that the phospholipid spectrum gives classes and not individual fatty acids; the remark on stored products; the remark that the current USP limit counts the lyso forms with phosphatidylcholine; the remark on why sample amounts and measuring times differ; the remark on nucleotides; the remark that the measurement on Ras requires the purified protein; and the descriptions of cardiolipin and of phosphocreatine, which are textbook knowledge. The statement that phosphate on serine, threonine and tyrosine withstands acid is taken from the preprint of source 4 (bioRxiv, 2020). All sources were read on 15 and 16 November 2026.
NMR Series
- Part 1 — Reading 1H and 13C NMR
- Part 2 — Reading 19F, 31P and 11B NMR
- Part 3 — Reading 2D NMR: COSY, HSQC and HMBC
- Part 4 — Quantitative NMR: how a purity figure is made
- Part 5 — 15N NMR: which nitrogen is alkylated, protonated or coordinated?
- Part 6 — NMR tubes by experiment: which tube for which measurement
- Part 7 — Quadrupolar nuclei in NMR: broad lines and missing carbons
- Part 8 — Two sets of NMR signals: rotamers, diastereomers or an impurity
- Part 9 — NMR yield and isomer ratios: relaxation and the delay between pulses
- Part 10 — Silicon NMR: the glass signal, the negative NOE and M, D, T, Q units
- Part 11 — Solid-state NMR: magic-angle spinning, cross-polarisation and what a CP-MAS spectrum counts
- Part 12 — Enantiomeric excess by NMR: when NMR is the best method, the chiral auxiliaries and the detection limit for the minor enantiomer
- Part 13 — Predicted NMR shifts: how accurate ChemDraw, MestReNova and other predictors are
- Part 14 — Phosphorus NMR in biology: phospholipids, phosphorylated proteins, oligonucleotide drugs and phosphates in food (this article)
- Part 15 — Multiplets in ¹H NMR: coupling trees, second-order patterns and two signals from one CH₂
- Part 16 — NMR for cosmetics: adulterated essential oils, free formaldehyde, aloe vera, rancid lipsticks and mineral oil in lip products
- Part 17 — Polymers by NMR: composition, sequence, chain ends and branches, and what a spectrum of many chains can and cannot measure
- Part 18 — Benchtop NMR: what a permanent-magnet spectrometer does well, and when a sample needs high field
- Part 19 — NOESY, ROESY and TOCSY: regioisomers, E and Z isomers, anomers, and the signals of one compound in a mixture
- Part 20 — Rapid-injection NMR: starting a reaction inside the spectrometer, how fast the spectra can follow it, and the intermediates it has revealed
- Part 21 — Fluorine NMR for PFAS: total organofluorine, what targeted LC-MS/MS does not detect, and the detection limit
- Part 22 — Polysiloxanes by proton NMR: silicone oil in medicines, food and cosmetics, the reactive groups before curing, and the rubber network after it
© 2026 NorrChemica™ · Content may not be reproduced without written permission