Multiple Parallel 2D NMR Acquisitions in a Single Scan

Abstract

Parallel receiving technologies have recently crossed the boundary separating magnetic resonance imaging from nuclear magnetic resonance (NMR) spectroscopy. In the latter case they promise significant time-saving advantages, by enabling the detection of multiple spectra simultaneously rather than in series. Moreover the full compatibility of parallel receiving with all other advances of contemporary NMR spectroscopy promises to open even further synergies in terms of speed and analytical capabilities. The present study shows one such instance, whereby the combination of parallel receiving multinuclear technologies is made with gradient-based spatial encoding methods, to yield multiple multidimensional NMR spectra in a single scan. The potential of this combination is demonstrated by the parallel acquisition of 2D H–H and H–S correlation spectra involving different S nuclei (F, P), within a single transient. Besides its potential conceptual message about what is nowadays within reach of NMR spectroscopy, the ensuing two-dimensional parallel ultrafast NMR spectroscopy (2D PUFSY) experiment carries new opportunities for high-throughput analyses, chemical kinetics, and fast experiments on metastable hyperpolarized solutions. Parallel receiving is an integral component of modern magnetic resonance; particularly in imaging applications where it can lead to substantial accelerating factors by scanning separate regions in space. The advent of multiple receivers is also beginning to influence NMR spectroscopy technologies; not by providing spatial multiplexing, but rather by enabling the simultaneous acquisition of two or more signals arising from different nuclei. Following the introduction of parallel NMR spectroscopy, it was shown that one of its main advantages results from its use to collect two or more different kinds of multidimensional correlation experiments, within the time duration that would normally entail to collect a single spectrum. This is the principle of parallel acquisition NMR spectroscopy (PANSY), which eventually evolved into more sophisticated pulse sequences capable of affording all the 2D correlation spectra necessary for a complete assignment of small molecules—within the timescale of the slowest experiment in the multiple set. These “parallel acquisition NMR, all-in-one combination of experimental applications” (PANACEA) strategies, have since been extended to systems of various heteronuclei and adapted to protein liquid-state and solid-state NMR experiments. While these multiple receiver techniques have demonstrated that substantial time savings are possible, they have still conformed to the classical means of indirect frequency encoding, whereby a series of independent scans are charged with encoding in a step-wise manner the evolution of the F1 indirect spectral domain. The incremented repetitions thus required to discretely sample the indirect time domain t1 implies that, even if sufficient sensitivity is available, sampling considerations associated with the slowest of all experiments still dictate the execution of all remaining 2D acquisitions. It was recently shown that sparse sampling coupled to non-Fourier processing techniques can alleviate this constraint, and break the Nyquist criteria without sacrifices in resolution or spectral bandwidth. Herein we present an alternative—and arguably ultimate—form of compressing multiple 2D experiments, involving their parallel implementation while following the spatially encoded protocol enabling the multiplexing of all the information involved in every indirect dimension, in a single scan. The spatiotemporal encoding principles underlying the acquisition of 2D NMR spectra/images in a single scan have been described elsewhere in detail, and hence they are only briefly described and within the context of the parallelized experiments presented here. “Ultrafast” NMR spectroscopy is based on endowing different z positions within a sample, with the different degrees of chemical shift evolution that would normally be associated with differing t1 values. If implemented in a one-to-one z–t1 fashion, this spatiotemporal encoding leads to a linear spatial winding of the magnetizations/coherences [Eq. (1)],