Joyce A. Wilde

A method for the observation of selected proton NMR resonances of proteins

The determination of protein structures in aqueous solution represents a formidable problem in biochemistry and mo lecular biology. Although NMR spectroscopy has the potential for providing structural information through proton chemical shifts, proton-proton couplings, and proton-proton nuclear Overhauser effects, the extent of overlap in the proton spectra, whether one or two dimensional, is often too great to allow detailed analysis. A number of approaches have been used to allow selective detection of only a particular subset of protons to ease assignment and overlap problems (I-4); however, all of these methods have significant shortcomings. We have devised a solution to this problem which relies upon both the high sensitivity selective observation of the NMR spectral properties of protons directly bonded to 13C and the use of recombinant plasmids which facilitate the economical preparation of the necessary isotopically labeled protein mo lecules. We have chosen staphylococcal nuclease, Nase, as our experimental system since the structural gene was recently c loned and sequenced (5). The availability of this gene has permitted the construction of a recombinant plasmid that permits the facile preparation of Nase samples that are labeled with r3C in a specific site of a given am ino acid; in addition, site directed mutagenesis can be used to generate mutant versions of the enzyme that may prove useful in the assignment of the ‘H NMR resonances of selected am ino acid residues. Such isotopically labeled Nase samples can be studied using NMR techniques that allow observation of only those protons directly bonded to a heteronucleus (neighbor protons) (6, 7); a two-dimensional NMR experiment can be used to observe those protons which are coupled to the neighbor protons (remote protons) (6). Selective detection of the neighbors was performed using the pulse sequence ‘H: 90”-r/2-180”-~/2-r-acquisition

Isotopic labeling with hydrogen-2 and carbon-13 to compare conformations of proteins and mutants generated by site-directed mutagenesis, II

Publisher Summary This chapter discusses the use of isotopic labeling to assign the aromatic partners of cross-peaks of both wild-type and mutant proteins through the use of deuteration of both aliphatic and aromatic amino acid side chains. Comparison of the data for isotopically labeled wild-type and mutant proteins allow independent assignment of the cross-peaks in these proteins. This contribution will conclude with a description of how the side chain assignments can be extended to site-specific assignments. For a protein the size of Snase, 149 amino acids plus a hexapeptide tail in the form one study, methods that depend on resolved scalar couplings or on sufficient resolution of the two-dimensional data will typically not be entirely adequate. Although isotopic labeling is both expensive and labor intensive, it does yield unambiguous information about both wild-type and mutant proteins. In favorable cases assignments can be made by comparing the NMR data of wild-type and mutant proteins. However, comparative methods are limited to instances of conformationally silent amino acids, or when there is a localized probe of the protein structure.

Suppression of couplings in homonuclear multiple-quantum spectroscopy

Multiple-quantum spectroscopy is potentially a very powerful method for the correlation of the chemical shifts of homonuclei (1). The double-quantum experiment is the version most commonly used for correlating the chemical shifts of rare nuclei such as carbon13 (2). However, in spectroscopy of protons, multiplequantum spectroscopy is rarely used. There are apparently several reasons for this. One is that the multiplets along the F, axis obtained in multiple-quantum experiments do not correspond to the multiplets along F2 and hence may be unfamiliar to many NMR spectroscopists. Another obstacle is the appearance in multiple-quantum spectroscopy of signals which have no analog in conventional chemical-shift correlation experiments. For example, in double-quantum spectroscopy a signal at F, frequency @A + wx can be obtained even when there is no coupling between A and X as long as they share a common coupling partner. In this communication we will demonstrate a method which suppresses the homonuclear couplings along F, such that the unusual multiplets typically associated with multiplequantum spectroscopy are not obtained. This approach has the additional advantages of effectively enhancing the resolution along F, since all signals will be singlets along this axis and the sensitivity also improves since all of the intensity is present in the single line. In addition, this method removes the signals arising from class II and class III coherences and hence only signals from direct connectivity are detected. These features in combination with the intrinsic lack of autocorrelation signals increase the attractiveness of double-quantum spectroscopy as a means of investigating, in particular, the connectivities of homonuclear spin systems. The approach used here is an adaptation of the constant-time experiment first proposed by Bax and Freeman (3) and later expanded on by Rance et al. (4). The basic constant-time experiment is