Pure In-Phase Heteronuclear Correlation NMR Experiments

Laura Castañar1,2, Josep Saurí1,3, Robert Thomas Williamson3, Albert Virgili2 and Teodor Parella1,2

1. Servei de RMN, Universitat Autònoma de Barcelona, Bellaterra, Barcelona. 2. Departament de Química, Universitat Autònoma de Barcelona, Bellaterra, Barcelona.

3. NMR Structure Elucidation, Process & Analytical Chemistry, Merck & Co. Inc., Rahway, NJ.

A general NMR approach to provide pure in-phase (PIP) multiplets in heteronuclear correlation experiments is described1. The implementation of a zero-quantum filter efficiently suppresses any unwanted anti-phase contributions that usually distort the multiplet pattern of cross-peaks and can hamper their analysis. The clean pattern obtained in PIP-HSQMBC experiments is suitable for a direct extraction of coupling constants in resolved signals, for a peak-fitting process from a reference signal, and for the application of the IPAP technique in non-resolved multiplets.

Introduction

SeRMN – UAB blog: http://sermn.uab.cat SMASH NMR Conference. September 7th – 10th 2014 Atlanta, Georgia, USA

Fig. 1 Pulse sequence to obtain pure in-phase (PIP) cross-peaks in heteronuclear correlation experiments. The interpulse delay Δ is set to 1/[2*1J(CH)] or 1/[2*nJ(CH)] in PIP-HSQC or PIP-HSQMBC experiments, respectively. The ZQF filter includes a chirped adiabatic 180º 1H pulse applied simultaneously with a purging G0 gradient. Broadband heteronuclear decoupling during proton acquisition is optional. The IPAP technique can be applied by recording separately two complementary IP (ε=on; Ψ=y) and AP (ε =off; Ψ=x) data sets, which are further added/subtracted in the time domain followed by conventional processing to give two separate α-β spectra.

Methodology

Experimental

H1x c2s‘2

- 2H1yCz c2s‘c‘

+ 2H1yH2z css‘2 + 4H1xH2zCz csc‘s

- H1z c2s‘2

- 2H1yCz c2s‘c‘

+ 2H1yH2x css‘2 + 4H1zH2xCz csc‘s‘

Fig. 2 A) 1H NMR spectrum of strychnine; B-E) 1D traces extracted at the C12 chemical shift showing the signal distortions in B) HSQMBC, C) CLIP-HSQMBC, D) z-filtered HSQMBC, and E) PIP-HSQMBC spectra (all experiments were optimized to 8 Hz (Δ=62.5 ms)).

Fig. 3 A) 8-Hz optimized PIP-HSQMBC spectrum of strychnine acquired with a BIRD cluster in the initial INEPT period to minimize direct responses; B) 1D row slices taken at different 13C frequencies showing in-phase multiplet patterns for all observed cross-peaks; C) Expanded area showing how the magnitude of nJ(CH) can be easily determined from direct analysis of the undistorted and resolved IP peaks.

Measurement of RDC’s

Experimental

Conclusions

The performance of the proposed PIP methods has been also verified under anisotropic conditions, using a sample of strychnine weakly aligned in a reversible compressed poly(methylmethacrylate) (PMMA) gel swollen in CDCl3. In Particular, a special focus is made for the analysis and precise determination of 1D(CH) and 2D(HH) RDCs from undistorted cross-peaks belonging to diasterotopic CH2 groups. The H11a/H11b protons of strychnine are a good example to evaluate the more accurate measurement of their homonuclear (2D(HH) =12.5 Hz) and heteronuclear (1D(C11H11a)= +7.7 Hz and 1D(C11H11b) = -18.2 Hz) RDCs, facilitating determination of their unequivocal orientation and stereo-selective assignment. It is shown that errors in the measurement of up to 10% Hz can be introduced from distorted cross-peaks in conventional F2-coupled CLIP-HSQC and F2-decoupled HSQC spectra. These errors are avoided in the undistorted PIP-HSQC cross-peaks.

Fig. 5 1D traces extracted at the C11 chemical shift of strychnine in isotropic and anisotropic conditions showing the signal distortions originated in (A) conventional HSQC, (B) CLIP-HSQC, and (C) PIP-HSQC spectra recorded with Δ set to 3.6 ms (1JCH = 140 Hz).

The extraction of nJ(CH) in more complex or non-resolved multiplets can be performed with several established methods: 1) measuring overall multiplet widths, 2) fitting/matching them to an external reference cross-peak obtained from the same sequence with broadband 13C-decouplingduring acquisition, or 3) from the internal satellite lines corresponding to the direct 1J(CH) responses, if available, without need to acquire a second reference spectrum. Alternatively, a simple approach relies on the implementation of the IPAP technique, which is achieved by recording two separate IP and AP datasets as a function of the last 180º 13C pulse (marked with ε in Fig. 1)

Fig. 4 Comparison of methods for the measuring of small heteronuclear coupling constants taking some multiplet patterns from an 8 Hz optimized PIP-HSQMBC spectrum: A) direct extraction, B) fitting from the internal satellite lines and, C) IPAP methodology. The resolution along the detected F2 dimension was of 0.4Hz.

• ZQF is an extremely efficient tool to suppress unwanted homo- and heteronuclear AP contributions in HSQC-like experiments, providing undistorted in-phase cross-peaks that are amenable for the easy extraction of coupling constant values. • The major advantage is that IP signals remove any doubts about possible J(HH) interferences in the resulting multiplet phases. • The method can be recorded in full automation mode without any prior calibration and we anticipate that the proposed PIP technique offers a general implementation on a large variety of inverse-detected NMR experiments and applicability to other nuceli than carbon, as well as for the measurement of RDCs.

In HSQMBC-like experiments, the interference of J(HH) effects is more obvious because the size of J(CH) and J(HH) are of the same order. In such cases the importance of the role of the ZQF becomes more prominent as is illustrated with the superior IP performance of the 8-Hz PIP-HSQMBC experiment over conventional, CLIP and z-filtered HSQMBC experiments acquired under the same conditions (Fig. 2). As pointed out in Fig. 1, only the first term representing the desired IP magnetization remains detectable after the ZQF (point b in Fig. 1). To maintain the pure IP character during detection, the classical δ-180ºx(

1H)-G2 block has been replaced by a perfect echo gradient element, in the form of δ-180ºy(

1H)-δ-90ºy(1H)-δ-180ºy(

1H)-G2, where J(HH) should be fully refocused.

c = cos(πJ(H1H2)Δ), s = sin(πJ(H1H2)Δ), c’ = cos(πJ(CH1)Δ), s’ = sin(πJ(CH1)Δ), and Δ = J evolution period

Before ZQF (a point), the magnetization of a given H1 spin is described as a mixture of IP and AP components

After ZQF (b point) only the first IP component remains

Measurement of nJ(CH)

Measurement of nJ(CH)

1 L. Castañar, J. Saurí, R. T. Williamson, A. Virgili and T. Parella. Angewandte Chemie International Edition (2014), DOI: 10.1002/anie.201404136

Acknowledgements: Financial support for this research provided by MICINN (project CTQ2012-32436) and Bruker Española S.A. are gratefully acknowledged. We also thank to the Servei de Ressonància Magnètica Nuclear, Universitat Autònoma de Barcelona, for allocating instrument time to this project.

The inspection of some selected traces along the F2 dimension belonging to the 8 Hz PIP-HSQMBC spectrum of strychnine clearly reveals that all cross-peaks display a clean IP character. Note the rough proportionality between signal intensity and nJ(CH) values. Under these conditions, even a small coupling value of 2.3 Hz can be directly measured for the two-bond C7/H8 correlation.