Urine is a readily-collected, information-rich biofluid that can provide insight into the metabolic state of an organism. As a result, urine is often a focus in metabolomics investigations using NMR and MRI spectroscopy, in both diagnostic and monitoring applications. Targeted profiling is a powerful tool that can drive such studies, providing direct identification and quantification of a variety of metabolites [11]. In this note we present a list of common metabolites found in many NMR spectra of human urine and several strategies for approaching targeted profiling of such spectra with Chenomx NMR Suite.
The inherent complexity of urine NMR spectra is a common barrier to applying targeted profiling; finding an entry point for an analysis is often difficult (see Figure 1). A good entry point for many analyses is a list of commonly-found metabolites, with which you can account for many of the most prominent signals in a spectrum. This initial fit can either act as a simple baseline for comparing a large number of samples, or as the first stage in a more in-depth analysis focusing on the less common metabolites.
When analyzing NMR spectra of urine samples via targeted profiling in Chenomx NMR Suite, certain compounds are likely to be detectable in normal samples (see Metabolite List for details). Some of these compounds are associated with key metabolic pathways, while others are associated with the consumption of particular food or drug products. This list of compounds provides a basis from which you can start any analysis of human urine spectra.
Fitting compounds to a urine spectrum can be difficult to generalize, since every spectrum can be affected by subtly different matrix effects, including varying pH, divalent cation concentration, dilution effects, and so on. However, some general strategies are available that can help you to get more accurate results from targeted profiling in Chenomx NMR Suite.
You should fit large, obvious signals quite early in your analysis. These signals tend to dominate the spectrum in the regions in which they appear, and fitting them first allows you to quickly fit a large amount of the total area of the spectrum with relatively little effort. Also, fitting the obvious compounds can help you to more confidently fit some of the more subtle patterns in the spectrum by effectively subtracting out the larger signals. Some of the compounds most likely to contribute large, obvious signals to a human urine spectrum include creatinine, urea and trimethylamine N-oxide. Compounds contributing smaller signals that should still be easy to pick out include citrate, hippurate, alanine, trigonelline and 1-methylnicotinamide.
Most methods of acquiring NMR spectra of aqueous samples such as urine involve some form of water suppression pulse. This has several effects, aside from the most apparent effect of reducing the size of the water peak. One of these is that clusters appearing close to the water peak, that is, those resonating at a similar frequency, also experience a damping effect. These clusters will appear smaller than you might expect, based on the number of protons responsible for the signal and the multiplicity of the signal. As a result, estimating the concentration of a multi-cluster compound based on clusters close to water can give artificially low results.
When you fit a compound with multiple clusters, you should use simple, isolated clusters, where available, to estimate the concentration of the compound. In any case, you should use clusters farther from the water peak in preference to those closer to the water peak. Clusters near water should be overfit slightly to compensate for the damping effect, while clusters farther away should be fit so that the sum line (appearing in red by default) best approximates the actual spectrum in the vicinity of the individual clusters.
Compounds that have only one signal can be difficult to fit, as it is common to see multiple signals of similar intensity in any given area in a urine NMR spectrum. Fitting single peaks to any one of these signals is simple enough, but determining which signal is the correct one for the compound in question poses an interesting challenge. In general, it can be helpful to fit these compounds later in your analysis. Many of the other compounds that you fit will have clusters nearby, and fitting these overlapping clusters will reduce the number of candidate signals for the single-peak compounds.
As you analyze a urine NMR spectrum, you will notice that a number of specific compounds have clusters that overlap. Naturally, this is most likely when there is a strong structural similarity among the compounds, as in the methyl groups of lactate and threonine, or the imidazole protons of histidine and the methylhistidines. However, coincidental overlap is also a distinct possibility (see Figure 1), as urine samples can contain several hundred metabolites detectable by 1H-NMR. These groups of compounds should generally be fit concurrently, as fitting them separately will result in inaccurate concentrations.
Exhaustively listing every possible combination of compounds that might overlap in a urine sample is well beyond the scope of this note. However, given a list of compounds commonly found in human urine, such as the list in Metabolite List, there are a number of combinations that should be considered in any analysis of a human urine sample.
Creatinine and creatine will always appear together in any aqueous sample containing either one (Figure 2), as they slowly interconvert in aqueous solution. They are structurally similar, and their methyl signals tend to overlap. In urine samples, creatinine will always be the larger of the two.
Lactate and threonine also show a degree of structural similarity, and their methyl signals tend to overlap. Relative concentrations will vary, so you should verify any concentration estimated using the methyl signals with other clusters for both compounds.
Figure 3. Overlap region of trimethylamine N-oxide (blue) and betaine (orange); several other compounds also overlap in the same region
Trimethylamine N-oxide and betaine each have a trimethylammonium group, so the large methyl signals for both appear very close together . Relative concentrations will vary, and several other compounds appear in the same region (Figure 3).
Histidine, π-methylhistidine and τ-methylhistidine pose an interesting challenge (Figure 4), as their aliphatic proton signals overlap heavily. All three are frequently present in human urine.
Tryptophan and 3-indoxylsulfate also have similar, indole-based ring systems, as do trigonelline and 1-methylnicotinamide, which are pyridinium-based. As a result, the associated signals from ring protons appear close together. Several clusters for each appear in otherwise quiet regions of the spectrum.
Citrate and dimethylamine appear frequently in human urine, and often appear together. Both have simple patterns, and appear in a relatively quiet region of the spectrum (Figure 5).
Succinate and pyroglutamate do not have any overt structural similarity, but the single peak of succinate overlaps one of the complex clusters of pyroglutamate. This can make arbitrarily identifying the signal associated with succinate difficult. Fitting pyroglutamate concurrently, using other clusters in the compound to aid in determining an appropriate concentration, can simplify the identification of the succinate signal.
Although applying targeted profiling to a urine spectrum appears complex, such analyses can be approached systematically. The list in Metabolite List, of common metabolites found in human urine, provides a handy starting point, and combined with the simple techniques outlined above, forms a basis for the systematic profiling of most human urine spectra.
Chenomx would like to thank Dr. Tom Marrie and Bruce Lix of the Magnetic Resonance Diagnostics Centre (MRDC) at the University of Alberta for providing data for this application note.
Table 1. Common Metabolites in Human Urine
| Amines and Derivatives | |
| 1-Methylnicotinamide | Hypoxanthine |
| Allantoin | Imidazole |
| Choline | sn-Glycero-3-phosphocholine |
| Creatine | Trimethylamine-N-oxide |
| Creatinine | Uracil |
| Dimethylamine | Urea |
| Ethanolamine | |
| Amino Acids and Derivatives | |
| 3-Indoxylsulfate | N,N-Dimethylglycine |
| Alanine | N-Acetylglycine |
| Arginine | O-Acetylcarnitine |
| Asparagine | O-Phosphocholine |
| Aspartate | Phenylacetylglycine |
| Betaine | Phenylalanine |
| Carnitine | Pyroglutamate |
| Glutamate | Salicylurate |
| Glutamine | Serine |
| Glycine | Taurine |
| Guanidoacetate | Threonine |
| Hippurate | Trigonelline |
| Histidine | Tryptophan |
| Indole-3-acetate | Tyrosine |
| Isoleucine | Valine |
| Leucine | π-Methylhistidine |
| Lysine | τ-Methylhistidine |
| Carbohydrates and Derivatives | |
| 1,6-Anhydro-β-D-glucose | Isopropanol |
| Acetone | Methanol |
| Ethanol | Propylene glycol |
| Fucose | Sucrose |
| Glucose | Xylose |
| Organic Acids | |
| 2-Hydroxyisobutyrate | Formate |
| 2-Oxoglutarate | Fumarate |
| 3-Hydroxybutyrate | Glycolate |
| 3-Hydroxyisovalerate | Lactate |
| 4-Hydroxyphenylacetate | Malonate |
| Acetate | Methylsuccinate |
| Acetoacetate | Pyruvate |
| Adipate | Suberate |
| Benzoate | Succinate |
| Butyrate | Tartrate |
| cis-Aconitate | trans-Aconitate |
| Citrate | |