After a brief theoretical description, new gradient-selected, proton-detected heteronuclear correlation sequences are introduced. The gs-HMBC and gs-Relayed-HMQC are closely related to the original gs-HMQC proposed by Hurd and John. A new approach to obtain pure absorption line shapes in gradient selected spectroscopy is used to measure phase-sensitive gs-HMQC spectra, to carry out multiplicity editing in HSQC spectra and to distinguish direct and long-range correlations in HMQC/HSQC-TOCSY spectra.

A gradient-selected J-HMBC experiment is presented. Coherence selection with magnetic field gradients results in HMBC spectra in which t , noise is very low or completely eliminated. This makes it possible to detect unequivocally very small ( < t Hz) long-range H,X coupling constants. With a slight modification, the HMBC experiment can be used for the quantitative measurement of these couplings. The signal amplitude of spectra with increasing coupling evolution time z shows the characteristic sin (zJXH t) dependence. By fitting a sine curve to the experimental data or by performing a 3D J-resolved HMBC experiment, accurate "JXH coupling constants are obtained. This method is especially useful for estimating coupling constants between protons and unprotonated heteronuclei, e.g. quaternary carbons, "N nuclei in labelled prolines or "P in organophosphates. These experiments offer the coupling information independent of the linewidth, because the coupling constant need not be extracted from the signal fine structure but rather is available from the displacement of two signal elements. All experiments of this kind have one common shortcoming: they require a proton-bound heteronucleus. Coupling constants to unprotonated heteronuclei are not accessible in this way. Furthermore, signals arising from TOCSY or NOESY transfers may appear that are not related to a long-range J coupling. Additionally, the signal splitting by the 'J,, coupling in F, reduces the intensity, and the signal separation is limited because the experiment is performed as an H,H correlation. For * Author to whom correspondence should be addressed. a sufficient separation these experiments are often extended into three dimensions. Another technique, i.e. the quantitative J correlation, measures the intensity ratio of cross peaks to reference peaks. One version has been implemented as a carbon-excited 3D experiment".' whereas another uses a refocussed HMBC experiment." The HMBC experiment was also used by Bermel et al. 13 and Titman et al.14, Bermel et al. compared the total multiplet width of an HMBC signal (including the long-range coupling) with that from a COSY-type experiment (without the long-range coupling). Titman et al. fitted the multiplet with the aid of reference multiplets.We used the HMBC experiment to measure the longrange H,X coupling constants by stepwise incrementation of the coupling evolution time in a constant-time manner. The HMBC experiment is in principle the most suitable experiment to obtain information on heteronuclear long-range couplings, because all long-range correlations are accessible, even those to unprotonated heteronuclei. However, the HMBC experiment in the phase cycled version" often suffers from unacceptably intense t , noise. This artifact can be removed almost completely by using magnetic field gradients to select the coherence pathway.I6 We use a slight modification of the previously presented gradientselected (gs) HMBC,'77'8 which might be called gs-J-HMBC (see Fig. 1). The long-range evolution delay t is varied...

SUMMARYIn degenerative brain disorders characterized by neuronal cell loss, a decrease of N-acetyl aspartate (NAA) is frequently detected by noninvasive proton MRspectroscopy ( 1 H-MRS). Because NAA is almost exclusively found in neurons but not glial cells, this decrease is considered a marker for the neuronal loss (Simmons

The metabolism of [2-13C]glycine in astrogliarich primary cultures obtained from brains of neonatal Wistar rats was investigated using 13c NMR spectroscopy.After a 24-h incubation of the cells in a medium containing glucose, glutamate, cysteine, and [2-13C]glycine, cell extracts and incubation media were analyzed for 13C-labeled compounds. Labeled creatine, serine, and glutathione were identified in the cell extracts. If arginine and methionine were present during the incubation with [2-13C]glycine, the amount of de novo synthesized [2-13C]creatine was twofold increased, and in addition, 13C-labeled guanidinoacetate was found in cell extracts and in the media after 24 h of incubation. A major part of the [2-13C]glycine was utilized for the synthesis of glutathione in astroglial cells. 13Clabeled glutathione was found in the cell extracts as well as in the incubation medium. The presence of newly synthesized [2~13C] serine, [3-13C] serine, and [2,3-13C] serine in the cell extracts and the incubation medium proves the capability of astroglial cells to synthesize serine out of glycine and to release serine. Therefore, astroglial cells are able to utilize glycine as a precursor for the synthesis of creatine and serine. This proves that at least one cell type of the brain is able to synthesize creatine. In addition, guanidinoacetate, the intermediate of creatine synthesis, is released by astrocytes and may be used for creatine synthesis by other cells, i.e., neurons. Key Words: Creatine-Glial metabolism-Glutathione-Glycine-NM R spectroscopy-Serine. Abbreviations used: DMEM, Dulbecco's modified Eagle's medium; GSx, amount of GSH plus twice the amount of GSSG; HSQC, heteronuclear single quantum coherence; MM, minimal medium; SHMT, serine hydroxymethyltransferase.

Amyotrophic lateral sclerosis (ALS) is a fatal neurodegenerative disease characterized by a progressive loss of motor neurons. Visualizing corresponding metabolic changes in the brain of patients with ALS with proton magnetic resonance spectroscopy ((1)H-MRS) may provide surrogate markers for an early disease detection, for monitoring the progression and for evaluating a treatment response. The primary objective of our study was to evaluate whether modifications in MR metabolite levels occur before clinical disease onset, and whether these changes are directly linked to a distinct spatial progression pattern in the CNS. Therefore, age-dependent alterations in the cerebral and spinal metabolic profile in the mouse model of ALS overexpressing the mutated human G93A-superoxide dismutase 1 (G93A-SOD1) were determined by high-resolution MRS of tissue extracts at 14.1 Tesla. Both non-transgenic mice (control mice) and transgenic mice overexpressing the non-mutated human SOD1 (tg-SOD1) served as controls. In the spinal cord of G93A-SOD1 mice significantly decreased levels of N-acetyl aspartate were already detected 34 days postpartum, i.e. about 60 days before the average disease onset caused by motor neuron decline. In addition, glutamine and gamma-aminobutyric acid concentrations were significantly diminished at Day 75, which is still in the presymptomatic phase of the disease. These metabolic changes were further progressive in the course of the disease and started to involve the brainstem at Day 75. Overall, high-resolution (1)H-MRS allows a sensitive spatial and temporal metabolite profiling in the presymptomatic phase of ALS even before significant neuronal cell loss occurs.