Toward high-resolution NMR spectroscopy of microscopic liquid samples

Abstract

A longstanding limitation of high-resolution NMR spectroscopy is the requirement for samples to have macroscopic dimensions. Commercial probes, for example, are designed for volumes of at least 5 μL, in spite of decades of work directed toward the goal of miniaturization. Progress in miniaturizing inductive detectors has been limited by a perceived need to meet two technical requirements: (1) minimal separation between the sample and the detector, which is essential for sensitivity, and (2) near-perfect magnetic-field homogeneity at the sample, which is typically needed for spectral resolution. The first of these requirements is real, but the second can be relaxed, as we demonstrate here. By using pulse sequences that yield high-resolution spectra in an inhomogeneous field, we eliminate the need for near-perfect field homogeneity and the accompanying requirement for susceptibility matching of microfabricated detector components. With this requirement removed, typical imperfections in microfabricated components can be tolerated, and detector dimensions can be matched to those of the sample, even for samples of volume ≪5 μL. Pulse sequences that are robust to field inhomogeneity thus enable small-volume detection with optimal sensitivity. We illustrate the potential of this approach to miniaturization by presenting spectra acquired with a flat-wire detector that can easily be scaled to subnanoliter volumes. In particular, we report high-resolution NMR spectroscopy of an alanine sample of volume 500 pL.

Quantitative NMR-Based Biomedical Metabolomics: Current Status and Applications

“…Enzymatic activity has also been utilized to remove signal interference (e.g., removal of urea from urine with urease) [32,33]. An alternative approach is to chemically modify select chemical classes to introduce a distinct 15 N or 13 C chemical shift to circumvent the spectral overlap problem. For example, carboxyl and carbonyl groups may be selectively tagged with 15 N-cholamine [34] or N-(2-15 N-aminooxyethyl)-N,N-dimethyl-1-dodecylammonium [35].…”

“…Nevertheless, NMR methodologies continue to evolve and to improve and, accordingly, the application of NMR for the identification and quantification of metabolites has increased dramatically over the past 20 years (Figure 1). NMR spectroscopy is inherently quantitative, where signal intensity is directly proportional to the number or count of each individual nuclei (i.e., 1 H, 13 C, 15 N, etc.) present in a sample [3].…”

“…Deconvolution can also be achieved by multidimensional NMR approaches, in which metabolites are separated along a second dimension by means of through-bond (i.e., HSQC) or through-space (i.e., NOE) correlations (Section 3.1). NMR spectroscopy is inherently quantitative, where signal intensity is directly proportional to the number or count of each individual nuclei (i.e., 1 H, 13 C, 15 N, etc.) present in a sample [3].…”

“…Furthermore, the implementation of new internal standards and software-assisted metabolite deconvolution techniques has enabled an accurate and broader coverage of metabolites within complex biomedical samples [1,14]. The inclusion of multidimensional NMR, alternative NMR nuclei (e.g., 13 C, 15 N, and 31 P), and solid state NMR has further expanded the utility of qNMR to metabolomics studies [15][16][17]. Additionally, qNMR has found utility in combined analytical techniques, such as HPLC-NMR and NMR-MS [18][19][20][21].…”

“…Notably, a broader identification and quantification of the metabolome is achievable by combining multiple analytical techniques that are beyond the capabilities of each individual method. Overall, recent Advancements in high-field magnets, cryoprobe technology, and pulse sequence design have continued to push the limits of metabolite quantification below traditional µM concentrations [13]. More importantly, qNMR has been extensively utilized in pharmaceutical studies, where the chemical purity of organic compounds can be estimated with an accuracy and precision of ±1% and an uncertainty less than 0.1% [7].…”