JOURNAL OF MAGNETIC RESONANCE 125, 216–219 (1997) ARTICLE NO. MN961101 Ultra-High-Resolved HSQC Spectra of Multiple- 13 C-Labeled Biofluids WIELAND WILLKER,ULRICH FLO ¨ GEL, AND DIETER LEIBFRITZ Institut fu ¨r Organische Chemie, Fachbereich 2, Universita ¨t Bremen, 28334 Bremen, Germany Received December 17, 1996 The NMR analysis of cell extracts and culture media from multiplied by an exponential weighting function, giving an additional line broadening of 2 Hz. cells fed with multiple- 13 C-labeled substrates such as [U- 13 C 6 ]glucose gives access to various aspects of cellular me- F98 glioma cells were grown to confluency in 15 cm culture dishes in a humidified atmosphere of 10% CO 2 in tabolism. 13 C NMR allows the simultaneous identification of individual 13 C isotopomers in order to distinguish different air at 37°C in DMEM, supplemented with 5% FCS and penicillin/streptomycin (100 units/ml). For cell extracts, metabolic pathways. These measurements are commonly carried out with 1D carbon spectroscopy ( 1–6 ) . The advan- approximately 10 8 cells obtained from four culture dishes were incubated for 24 h in DMEM medium containing 8 m M tage of this method is that all labeled carbons are detected in one spectrum with very high spectral resolution. The dis- [U- 13 C 6 ]glucose. All samples were adjusted to pH values of 7–8. advantages, however, are the low sensitivity, the signal su- perpositions, and the need to assign unknown resonances Sensitivity-improved HSQC with gradient echo / antiecho selection is used for optimum sensitivity ( 10, 11 ) . There are based on chemical shifts only. In contrast, 2D inverse H,C spectroscopy offers better signal separation, very high sensi- several possibilities for introducing a selective pulse into this pulse sequence. The sequence in Fig. 1a replaces the tivity, and tools to confirm the assignment ( 7 ). A drawback is the low resolution in the carbon dimension. Especially at carbon 90° excitation pulse with a G4 Gaussian pulse cas- cade ( 12 ), and the sequence in Fig. 1b replaces the 180° high field ( 600 / 800 MHz ) , it is almost impossible to resolve C–C couplings with an acceptable number of t 1 increments pulse in the first INEPT step with a Gaussian 180° pulse. The pulse sequence in Fig. 1a is similar to the pulse sequence in a nonselective HSQC experiment. To overcome this limi- tation, we have recently proposed a region-selective HSQC presented in ( 8 ), but is now combined with sensitivity im- provement. The advantage of the new pulse sequence shown ( 8 ). Here we present an improved version and optimum combination with J scaling and constant-time evolution. Ap- in Fig. 1b is that it is of the same length as the nonselective HSQC with a signal intensity of approximately 120% com- plications to a multilabeled cell extract of F98 glioma cells are shown. pared to sequence 1a. However, the advantage of sequence 1a is an almost rectangular excitation profile of the G4 All experiments were performed on a Bruker DRX 600 MHz spectrometer. A 5 mm, H,C,N, inverse triple-resonance Gaussian pulse cascade. Sequence 1b requires a symmetrical pulse. We used a Gaussian 180° pulse, but other symmetrical probe with shielded gradients was used. Gradients were shaped by a waveform generator and amplified by a Bruker pulses like hyperbolic secant are also possible. A useful feature of 2D spectroscopy is J scaling ( 13 ). Acustar amplifier. Sinusoidal z gradients of duration 1 ms and recovery time 100 ms have been used for echo/antiecho The evolution of the chemical shift can be manipulated by inserting 180° pulses into the evolution time t 1 . In our case, gradient selection. Gradient fine adjustment ( 40:10.08 ) has been performed to get optimum intensity. Low-power adia- an upscaling of the 13 C– 13 C coupling constants was required. This can be obtained by introducing an additional 180° car- batic composite-pulse decoupling with WURST ( 9 ) has been used. A Gaussian 180° pulse of length 1 ms was used to bon pulse and adding several incremented d 0 delays (Fig. 1c). J CC evolution occurs during the full t 1 period, whereas excite a range of 15 ppm in F 1 ( 13 C). The selective HSQC experiments were acquired with 512 or 1024 t 1 increments, 13 C chemical shift evolves only during the last two d 0 delays. This version scales up the 13 C– 13 C coupling by a factor of for a spectral width of 15 ppm, to give a digital resolution in the carbon dimension of 4 or 2 Hz/pt. An acquisition 3. Additional d 0 delays cause an even greater upscaling but may cause signal loss due to relaxation. J scaling offers time of 285 ms has been used to acquire a spectral width of 3 ppm in the proton dimension using digital quadrature the opportunity to unravel superimposed lines by utilizing variable line splitting. This holds, for example, if one ele- detection. For the carbon spectra, an H,C dual probe was used with acquisition time 0.9 s and repetition time 2 s. This ment of a multiplet pattern is superimposed on a singlet signal of a monolabeled 13 C isotopomer. affords a digital resolution of 0.55 Hz/pt. The spectra were 216 1090-7807/97 $25.00 Copyright  1997 by Academic Press All rights of reproduction in any form reserved.