Letter to the editor: "Interpretation of (31)P NMR saturation transfer experiments: do not forget the spin relaxation properties".

TO THE EDITOR: In an interesting Editorial Focus published last year in American Journal of Physiology-Cell Physiology, Balaban and Koretsky (1) drew attention to magnetization transfer (MT) effects that are observed in P magnetic resonance spectroscopy (MRS) experiments, in which the -ATP resonance is irradiated, leading to saturation transfer between -ATP and the and -ATP phosphates. The authors point out that the effects can be attributed to direct chemical exchange between small metabolite pools of ATP or ADP, which may be bound in enzyme-substrate complexes with resonances shifted to different frequencies and/or broad line shapes. Here we point out that magnetic interactions between phosphorous spins in metabolites have to be taken into account as well. Magnetization transfer effects do not only arise from chemical exchange, but also from cross-relaxation effects. The latter phenomenon has been largely neglected in in vivo NMR spectroscopy. A case in point is the MT between the phosphates of ATP. In a recent article (3) we describe saturation transfer studies on resting muscle in the hind legs of wild-type and double-mutant mice lacking the cytosolic muscle creatine kinase and adenylate kinase isoforms. We found saturation transfer of equal magnitude between the two combinations of nearest-neighbor phosphates in ATP, i.e., between and -ATP and between and -ATP. Not only were the transfer effects between phosphate combinations similar, also no differences between wild-type and mutant mice were found. This rules out involvement of the enzymatic reaction cascade that can catalyze -ADP↔ ATP phosphoryl exchange (and other reactions) usually invoked to explain the saturation transfer between and -ATP. In our explanation we attributed the magnetization transfer effects to so-called transferred NOEs, NOEs created in a pool of ATP in a bound state with a long rotational correlation time, which exchanges with the bulk cytosolic ATP. In their Editorial Focus, Balaban and Koretsky proposed an alternative mechanism, which also assumes involvement of ATP in enzyme complexes and bulk cytosolic ATP, but in which the saturation transfer follows from direct exchange of the small bound pool with bulk cytosolic ATP. For this scenario to occur, the resonances of the bound pools ( and -ATP) must be shifted near the bulk -ATP resonance so they get cosaturated when this resonance is irradiated. Then, via exchange between the bound and the bulk pool, saturation transfer can appear at the bulk or -ATP resonances. The problem with this proposed chain of events is that the shifts required are quite unlikely, e.g., 13 parts per million (ppm) for -ATP, while shifts of 0 to 3 ppm are normally observed for the systems considered (4 –7). One might argue that line broadening of the resonances in the bound pool can lead to overlap with -ATP in the bulk pool, but for the ATP complexes considered such a large line broadening is unlikely to occur (4 – 6) (see Fig. 1). The observation of magnetization transfer from -ATP to -ATP and the weaker transfer from -ATP to -ATP (3) is most curious because no exchange reactions are known to account for these transfers. Thus, we propose that cross-relaxation resulting in NOEs between phosphate spins in ATP, of which the rotational motion is restricted, provides a straightforward explanation for MT effects involving P resonances of ATP observed in MRS studies performed in vivo.