Pulsed selective excitation theory and design in multiphoton MRI

Excitation in MRI is traditionally done at the Larmor frequency, where the energy of each radiofrequency photon corresponds to the energy difference between two spin states. However, if multiple radiofrequencies are employed, then multiphoton excitation can also occur when the sum or difference of multiple photon frequencies equals the Larmor frequency. Although multiphoton excitation has been known since the early days of NMR, it has been relatively unexplored in MRI. In this work, equations and principles for multiphoton selective RF pulse design in imaging are presented and experimentally demonstrated. In particular, the case where there are radiofrequency fields in both the traditional xy-direction and non-traditional z-direction is considered. To produce the z-direction radiofrequency field, an additional uniform coil was added to a clinical MRI scanner. Using this coil, two-photon slice-selective pulses were designed to be equivalent to traditional pulses, producing similar excitation, slice profiles, and in vivo images. Being the result of a combination of multiple radiofrequency fields instead of just one, two-photon pulses have more flexibility in how their parameters can be changed. Although individual multiphoton excitations are less efficient than their traditional counterparts, when the z-direction radiofrequency field is spatially non-uniform, multiple multiphoton resonances can be simultaneously used at different locations to produce simultaneous multislice excitation with the same pulse duration but less tissue heating than a naive implementation. In particular, non-uniform z-direction radiofrequency fields with negligible added tissue heating provided by oscillating the MRI scanner's gradient fields at kilohertz frequencies were used to excite multiple slices simultaneously with less high-frequency xy-direction radiofrequency power. For an example three-slice excitation, we achieve half the xy-direction radiofrequency power compared to the naïve approach of adding three single-slice pulses. For conventional or unconventional applications, multiphoton excitation may be of interest when designing new MRI systems.

Analysis of HMQC experiments applied to a spin ½ nucleus subject to very large CSA

The acquisition of solid-state NMR spectra of "heavy" spin I = 1/2 nuclei, such as ¹¹⁹Sn, ¹⁹⁵Pt, ¹⁹⁹Hg or ²⁰⁷Pb can often prove challenging due to the presence of large chemical shift anisotropy (CSA), which can cause significant broadening of spectral lines. However, previous publications have shown that well-resolved spectra can be obtained via inverse ¹H detection using HMQC experiments in combination with fast magic angle spinning. In this work, the efficiencies of different ¹⁹⁵Pt excitation schemes are analyzed using SIMPSON numerical simulations and experiments performed on cis- and transplatin samples. These schemes include: hard pulses (HP), selective long pulses (SLP) and rotor-synchronized DANTE trains of pulses. The results show that for spectra of species with very large CSA, HP is little efficient, but that both DANTE and SLP provide efficient excitation profiles over a wide range of CSA values. In particular, it is revealed that the SLP scheme is highly robust to offset, pulse amplitude and length, and is simple to set up. These factors make SLP ideally suited to widespread use by "non-experts" for carrying out analyses of materials containing "heavy" spin I = 1/2 nuclei that are subject to very large CSAs. Finally, the existence of an "intermediate" excitation regime, with an rf-field strength in between those of HP and SLP, which is effective for large CSA, is demonstrated. It must be noted that in some samples, multiple sites may exist with very different CSAs. This is the case for ¹⁹⁵Pt species with either square-planar or octahedral structures, with large or small CSA, respectively. These two types of CSAs can only be excited simultaneously with DANTE trains, which scale up the effective rf-field. Another way to obtain all the information is to perform two different experiments: one with SLP and the second with HP to excite the sites with moderate/large and small/moderate CSAs, respectively. These two complementary experiments, recorded with two different spinning speeds, can also be used to discriminate the center-band resonances from the spinning sidebands.

Evaluation of excitation schemes for indirect detection of 14N via solid-state HMQC NMR experiments

It has previously been shown that ¹⁴N NMR spectra can be reliably obtained through indirect detection via HMQC experiments. This method exploits the transfer of coherence between single-(SQ) or double-quantum (DQ) ¹⁴N coherences, and SQ coherences of a suitable spin-1/2 'spy' nucleus, e.g., ¹H. It must be noted that SQ-SQ methods require a carefully optimized setup to minimize the broadening related to the first-order quadrupole interaction (i.e., an extremely well-adjusted magic angle and a highly stable spinning speed), whereas DQ-SQ ones do not. In this work, the efficiencies of four ¹⁴N excitation schemes (DANTE, XiX, Hard Pulse (HP), and Selective Long Pulse (SLP)) are compared using J-HMQC based numerical simulations and either SQ-SQ or DQ-SQ ¹H-{¹⁴N} D-HMQC experiments on l-histidine HCl and N-acetyl-l-valine at 18.8 T and 62.5 kHz MAS. The results demonstrate that both DANTE and SLP provide a more efficient ¹⁴N excitation profile than XiX and HP. Furthermore, it is shown that the SLP scheme: (i) is efficient over a large range of quadrupole interaction, (ii) is highly robust to offset and rf-pulse length and amplitude, and (iii) is very simple to set up. These factors make SLP ideally suited to widespread, non-specialist use in solid-state NMR analyses of nitrogen-containing materials.

Managing emerging mutagenicity risks: Late stage mutagenic impurity control within the atovaquone second generation synthesis

The mutagenic-impurity control strategy for a second generation manufacturing route to the non-mutagenic antipneumocystic agent atovaquone (2-((1R,4R)-4-(4-chlorophenyl)cyclohexyl)-3-hydroxynaphthalene-1,4-dione) 1 is described. Preliminary assessment highlighted multiple materials of concern which were largely discharged either through returning a negative bacterial mutagenicity assay or through confidence that the impurity would be purged during the downstream processing from when it was first introduced. Additional genotoxicity testing highlighted two materials of concern where initial assessment suggested that testing for these impurities at trace levels within the drug substance would be required. Following a thorough review of process purging detail, spiking and purging experimentation, and an understanding of the process parameters to which they were exposed an ICH M7 Option 4 approach could be justified for their control. The development of two 1H NMR spectroscopy methods for measurement of these impurities is also described as well as a proposed summary table for describing the underlying rationale for ICH M7 control rationales to regulators. This manuscript demonstrates that process purging of potential mutagenic impurities can be realised even when they are introduced in the later stages of a process and highlights the importance of scientific understanding rather than relying on a stage-counting approach.

Exploring methods to expedite the recording of CEST datasets using selective pulse excitation

Chemical Exchange Saturation Transfer (CEST) has emerged as a powerful tool for studies of biomolecular conformational exchange involving the interconversion between a major, visible conformer and one or more minor, invisible states. Applications typically entail recording a large number of 2D datasets, each of which differs in the position of a weak radio frequency field, so as to generate a CEST profile for each nucleus from which the chemical shifts of spins in the invisible state(s) are obtained. Here we compare a number of band-selective CEST schemes for speeding up the process using either DANTE or cosine-modulated excitation approaches. We show that while both are essentially identical for applications such as ¹⁵N CEST, in cases where the probed spins are dipolar or scalar coupled to other like spins there can be advantages for the cosine-excitation scheme.

F 1 F 2-selective NMR spectroscopy

Fourier transform NMR spectroscopy has provided unprecedented insight into the structure, interaction and dynamic motion of proteins and nucleic acids. Conventional biomolecular NMR relies on the acquisition of three-dimensional and four-dimensional (4D) data matrices to establish correlations between chemical shifts in the frequency domains F ₁, F ₂, F ₃ and F ₁, F ₂, F ₃, F ₄ respectively. While rich in information, these datasets require a substantial amount of acquisition time, are visually highly unintuitive, require expert knowledge to process, and sample dark and bright regions of the frequency domains equally. Here, we present an alternative approach to obtain multidimensional chemical shift correlations for biomolecules. This strategy focuses on one narrow frequency range, F ₁ F ₂, at a time and records the resulting F ₃ F ₄ correlation spectrum by two-dimensional NMR. As a result, only regions of the frequency domain that contain signals in F ₁ F ₂ ("bright regions") are sampled. F ₁ F ₂ selection is achieved by Hartmann-Hahn cross-polarization using weak radio frequency fields. This approach reveals information equivalent to that of a conventional 4D experiment, while the dimensional reduction may shorten the total acquisition time and simplifies spectral processing, interpretation and comparative analysis. Potential applicability of the F ₁ F ₂-selective approach is illustrated by de novo assignment, structural and dynamics studies of ubiquitin and fatty-acid binding protein 4 (FABP4). Further extension of this concept may spawn new selective NMR experiments to aid studies of site-specific structural dynamics, protein-protein interactions and allosteric modulation of protein structure.

Selective excitation for spectral editing and assignment in separated local field experiments of oriented membrane proteins

Spectroscopic assignment of NMR spectra for oriented uniformly labeled membrane proteins embedded in their native-like bilayer environment is essential for their structure determination. However, sequence-specific assignment in oriented-sample (OS) NMR is often complicated by insufficient resolution and spectral crowding. Therefore, the assignment process is usually done by a laborious and expensive "shotgun" method involving multiple selective labeling of amino acid residues. Presented here is a strategy to overcome poor spectral resolution in crowded regions of 2D spectra by selecting resolved "seed" residues via soft Gaussian pulses inserted into spin-exchange separated local-field experiments. The Gaussian pulse places the selected polarization along the z-axis while dephasing the other signals before the evolution of the ¹H-¹⁵N dipolar couplings. The transfer of magnetization is accomplished via mismatched Hartmann-Hahn conditions to the nearest-neighbor peaks via the proton bath. By optimizing the length and amplitude of the Gaussian pulse, one can also achieve a phase inversion of the closest peaks, thus providing an additional phase contrast. From the superposition of the selective spin-exchanged SAMPI4 onto the fully excited SAMPI4 spectrum, the ¹⁵N sites that are directly adjacent to the selectively excited residues can be easily identified, thereby providing a straightforward method for initiating the assignment process in oriented membrane proteins.

Solid-state NMR indirect detection of nuclei experiencing large anisotropic interactions using spinning sideband-selective pulses

Under Magic-Angle Spinning (MAS), a long radio-frequency (rf) pulse applied on resonance achieves the selective excitation of the center-band of a wide NMR spectrum. We show herein that these rf pulses can be applied on the indirect channel of Hetero-nuclear Multiple-Quantum Correlation (HMQC) sequences, which facilitate the indirect detection via spin-1/2 isotopes of nuclei exhibiting wide spectra. Numerical simulations show that this indirect excitation method is applicable to spin-1/2 nuclei experiencing a large chemical shift anisotropy, as well as to spin-1 isotopes subject to a large quadrupole interaction, such as (14)N. The performances of the long pulses are analyzed by the numerical simulations of scalar-mediated HMQC (J-HMQC) experiments indirectly detecting spin-1/2 or spin-1 nuclei, as well as by dipolar-mediated HMQC (D-HMQC) experiments achieving indirect detection of (14)N nuclei via (1)H in crystalline γ-glycine and N-acetyl-valine samples at a MAS frequency of 60kHz. We show on these solids that for the acquisition of D-HMQC spectra between (1)H and (14)N nuclei, the efficiency of selective moderate excitation with long-pulses at the (14)N Larmor frequency, ν0((14)N), is comparable to those with strong excitation pulses at ν0((14)N) or 2ν0((14)N) frequencies, given the rf field delivered by common solid-state NMR probes. Furthermore, the D-HMQC experiments also demonstrate that the use of long pulses does not produce significant spectral distortions along the (14)N dimension. In summary, the use of center-band selective weak pulses is advantageous for HMQC experiments achieving the indirect detection of wide spectra since it (i) requires a moderate rf field, (ii) can be easily optimized, (iii) displays a high robustness to CSAs, offsets, rf-field inhomogeneities, and fluctuations in MAS frequency, and (iv) is little dependent on the quadrupolar coupling constant.

Conditional rotations of heteronuclear coupled spins

We present a new pulse sequence that conditionally excites I spin magnetization only in the presence of a nonzero heteronuclear coupling to an S spin. The pulse sequence, referred to as the reverse INEPT pathway selective pulse or RIPSP, generates a pure I spin rotation by an angle that depends upon the heteronuclear coupling constant in InS spin systems. Experimental demonstrations are shown in (13)C labeled chloroform, dichloromethane, and toluene samples and in unlabeled 2,3-dibromopropionic acid and brucine samples.

|B1(+)|-selective excitation pulse design using the Shinnar-Le Roux algorithm

A new mathematical treatment and algorithm for the design of |B1(+)|-selective RF excitation pulses is presented and validated. The algorithm is based on a rotated Shinnar-Le Roux pulse design algorithm, wherein the pulse's frequency modulation waveform is directly designed by the algorithm, and its amplitude and sign modulation waveform takes the place of the gradient field. A new pulse configuration is described that enables excitation of large tip-angle slice-selective profiles. Experiments were performed to validate the pulses, and simulations were performed to characterize the pulses' sensitivity to off-resonance, and to compare them to adiabatic (BIR-4) pulses.