We would like to comment upon the paper, recently published in this journal,' entitled 'The Complete, Unambiguous Assignment of the 13C NMR Spectrum of Erythromycin A' . We have completed extensive 'H and 13C NMR studies on 1 and a series of derivatives. Our 1985 unambiguous 13C NMR assignment of 1 in CDC1,' agrees with reference 1 except for the reverse ordering of C-11 and C-5' (A6 < 0.05 ppm).This change is probably accounted for by the high concentration (saturated), high temperature (55 "C) and apparently high water content (figure 1, reference 1) of the sample used in reference 1.We disagree with the conclusion', ' based solely on 13C 6 values, that the effect of solvent on the conformation of 1 seems to be negligible. Our detailed 'H and 13C NMR studies3z4 have Dear Madam
-Multinuclear N M R referencingFor 2 years the IUPAC 1-5 (Spectroscopy) Commission has been considering the need for updating the recommendations (1975)' on reference standards for NMR. The NMR press has echoed the preoccupation of the subgroup of this Commission, and some colleagues have sent us their contributions. As we are now entering an active period and wish to prepare the first proposal, it is necessary for more scientists to be involved in this project and send us their proposals with their comments.Apart from the few suggestions we have received, the Commission has to rely on scarce papers related to certain nuclei and the normal literature with the more or less accepted references, which are not necessarily the best ones as shown for instance for 70.2
6"1 shown that the conformations of the macrolide ring of 1 (especially the region C-6 to C-9) are sensitive to solvent. For example, 3J(78) -6 and -7 Hz in DMF-d, but 3J(78) -11.0 and -2.7 Hz in CDCl,. Moreover in solvents like CD30D, formation of the 6,9-epoxy tautomers of 1 is p r~m o t e d .~.~.~ Finally we have also shown that, contrary to expectations, the macrolide rings of sever-As a first step the easiest referencing mode is to use a definite compound under well defined conditions of solvent and concentration. I write this letter in order to have the maximum contact with spectroscopists who have good experience with referencing on nuclei which they use routinely, and who can help us with our choices.As referencing also depends on many other parameters3 and on the methods used, the second step will be to discuss the referencing methods themselves and to propose more precise procedures. This early stage of very high resolution NMR4 will lose part of its interest if no precise measurements of chemical shifts are available.Referencing is not a secondary problem in NMR, and we are sometimes faced with difficulties when we wish to compare results from different papers where referencing is unusual, or poorly a1 erythromycin A derivatives are unstable and in fast conformational exchange in CDCl, solvent at ambient t e m p e r a t~r e .~.~.~ Yours faithfully
