For the sequential assignment of the protein backbone, a number of powerful multidimensional triple-resonance experiments are available. With these experiments, stretches of sequentially neighboring amino acids are obtained which, in a subsequent step, must be mapped on the known amino acid sequence. For this purpose, the amino acid type must be determined for some of at least of the assigned peaks. This can be achieved through isotope labeling by residue type or by determining the spin systems using homonuclear or heteronuclear experiments. We present pulse sequences which combine the advantages of both methods, i.e., the direct identification of certain amino acid types with the need for only one 15N/13C-labeled sample. These pulse sequences select sequential connectivities to the adjacent NH group for amino acids which lack a 13Cbeta-13Cgamma coupling, i.e., glycines, alanines, cysteines, and serines. Furthermore, asparagines, aspartic acids, and aromatic residues can be selected since the resonance positions of their gamma carbons are well separated from the aliphatic 13C resonances.
V. Dötsch, R.E. Oswald & G. Wagner. Water suppression by coherence selection with absorptive lineshape without loss in sensitivity. J. Magnetic Res. Series B, 108, 285-288 (1995).
The use of pulsed field gradients to select coherence pathways has become part of many recently published triple resonance pulse sequences ( 1-6 ). In these sequences a gradient pulse is used to dephase the transverse magnetization according to its coherence order. After the magnetization has been transferred to another spin, the desired coherence is selectively refocused with a second gradient pulse. In addition to shortening the phase cycle, coherence selection in heteronuclear experiments can suppress the strong water signal very efficiently. The disadvantage of water suppression with heteronuclear coherence selection is a loss in signal-to-noise of a factor of two compared to other water suppression schemes ( 1, 5 ), e.g. presaturation. This sensitivity loss can be readily understood from the fact, that a sufficiently strong gradient pulse spreads out the magnetization over the whole x-y-plane. A projection of this magnetization on the two axis shows, that one half of the original amount is aligned along the x-axis, while the other half can be projected on the y-axis. A subsequent p/2-pulse as used for coherence transfer (e.g. along the x-axis) would therefore select only half of the magnetization without affecting the second half. Since the second gradient selectively refocuses magnetization, which is affected by the p/2-pulse, only one half of the original magnetization will be detected during acquisition. This loss in sensitivity can be circumvented by placing the first gradient in the evolution time and using a pulse sequence which refocuses both orthogonal magnetization components present at the end of the evolution time ( 2, 4, 6-10 ). These sensitivity-enhanced sequences can increase the signal-to-noise by a factor of up to . However, this method detects a phase-modulated signal, which cannot be processed in a standard way ( 2, 10 ). Instead, the normal amplitude-modulated signal has to be created by adding and subtracting two data sets, which are recorded with opposite signs of the refocusing gradient. In this communication, we describe a water-suppression scheme based on heteronuclear coherence selection, which results in a normal amplitude-modulated signal with the full sensitivity of methods which do not employ coherence selection. This method can easily be implemented in almost every heteronuclear multiple resonance experiment.
1. A.L. Davis, R. Boelens, and J. Kaptein. J. Biol. NMR. 2, 395 (1992).
2. D.R. Muhandiram, G.Y. Xu, and L.E. Kay. J. Biomol. NMR. 3, 463 (1993).
3. L.E. Kay, G. Yi Xu, and J. Yamazaki. J. Magn. Reson. 109 A, 129 (1994).
4. D.R. Muhandiram, and L.E. Kay. J. Magn. Reson. 103 B, 203 (1994).
5. E.T. Olejniczak, and S.W. Fesik. J. Am. Chem. Soc. 116, 2215 (1994).
6. J. Schleucher, M. Schwendinger, M. Sattler, P. Schmidt, O. Schedletzky, S.J. Glaser, O.W. S¿rensen, and C. Griesinger. J. Biomol. NMR. 4, 301 (1994).
7. J. Cavanagh, A.G. Palmer, P.E. Wright, and M. Rance. J. Magn. Reson. 91, 429 (1991).
8. A.G. Palmer, J. Cavanagh, P.E. Wright, and M. Rance. J. Magn. Reson. 93, 151 (1991).
9. A.G. Palmer, J. Cavanagh, R.A. Byrd, and M. Rance. J. Magn. Reson. 96, 416 (1992).
10. L.E. Kay, P. Keifer, and T. Saarinen. J. Am. Chem. Soc. 114, 10663 (1992).
Z.G. Wo, Z.C. Bian, & R.E. Oswald. Asn265 of frog kainate binding protein is a functional glycosylation site: Implications for the transmembrane topology of glutamate receptors. FEBS Lett., 368, 320-234 (1995).
Kainate binding proteins (KBPs) from frog and goldfish brain are glycosylated, integral membrane proteins. These KBPs are homologous (35-40%) to the C-terminal half of AMPA and kainate receptors which have been shown to form glutamate-gated ion channels. We report here that the frog KBP has three functional N-glycosylation sites. Of particular interest, Asn265, a residue located between two putative membrane spanning regions of the frog KBP, is a functional N glycosylation site. A mutation of Ser267 to Gly renders this site nonfunctional as shown using an in vitro translation system and by transient expression in human embryonic kidney (HEK 293) cells. The mutant receptor protein (S267G), when expressed in HEK cells, binds kainate with high affinity (KD = 16 nM). These results further support a topology with three transmembrane segments for KBPs and, by sequence homology, for glutamate-gated ion channels.
Z.G. Wo & R.E. Oswald. Unraveling the modular design of glutamate-gated ion channels. Trends in Neurosciences, 18, 161-168 (1995).
Glutamate receptors which function as ligand-gated ion channels are essential components of cell-cell communication in the nervous system. Despite a wealth of information concerning these receptors, details of their structure are just beginning to emerge. We propose that glutamate receptors are comprised of four modules: two related to bacterial periplasmic binding proteins, one related to the pore-forming region of K+ channels, and one regulatory module of unknown origin. A K+ channel-like domain inserted into a crucial region of a periplasmic binding protein-like domain suggests a mechanism for transduction of binding energy to channel opening. This modular design also suggests an evolutionary link between a ligand-gated ion channel family and voltage-gated ion channels.
Z.G. Wo & R.E. Oswald. A topological analysis of goldfish kainate receptors predicts three transmembrane segments. J. Biol. Chem., 270, 2000-2009 (1995).
Glutamate receptors are the most abundant excitatory neurotransmitter receptors in vertebrate brain. We have previously cloned cDNAs encoding two homologous kainate receptors (GFKARa, 45 kDa, and GFKARb, 41 kDa) from goldfish brain and proposed a topology with three transmembrane domains (Wo & Oswald, 1994, PNAS, 91, 7154-7158). These studies have been extended using an in vitro translation/translocation system in conjunction with site-specific antibodies and point and deletion mutations. We report here that the entire region between the previously proposed third and fourth transmembrane segments is translocated and likely to be extracellular in mature receptors. This was based on the following results: (1) The entire segment was protected from Proteinase K and trypsin digestion and could be immunoprecipitated by a site-specific antibody. (2) Functional sites for N-glycosylation are present in the C-terminal half of the segment, and (3) A mutation, constructed with an additional consensus site for N-glycosylation in the N-terminal half of the segment, was found to be glycosylated at that site. Given the fact that the N-terminal region of the protein is likely to be extracellular, this would place an even number of transmembrane segments between the extracellular N-terminus and the glycosylated segment. In addition, results of N-glycosylation and proteolysis protection assays of GFKARa mutations indicated that the previously proposed second transmembrane segment is not a true transmembrane domain. These results provide further evidence in support of a topology with three transmembrane domains which has important implications for the relationship of structure to function in ionotropic glutamate receptors.
A.A. Carter & R.E. Oswald. Linear prediction and single channel recording. J. Neurosci. Methods, 60, 69-78 (1995).
The measurement of individual single channel events arising from the gating of ion channels provides a detailed data set from which the kinetic mechanism of a channel can be deduced. In many cases, the pattern of dwells in the open and closed states is very complex, and the kinetic mechanism and parameters are not easily determined. Assuming a Markov model for channel kinetics, the probability density function for open and closed time dwells should consist of a sum of decaying exponentials. One method of approaching the kinetic analysis of such a system is to determine the number of exponentials and the corresponding parameters which comprise the open and closed dwell time distributions. These can then be compared to the relaxations predicted from the kinetic model to determine, where possible, the kinetic constants. We report here the use of a linear technique, linear prediction/singular value decomposition, to determine the number of exponentials and the exponential parameters. Using simulated distributions and comparing with standard maximum likelihood analysis, the singular value decomposition techniques provide advantages in some situations and are a useful adjunct to other single channel analysis techniques.
M.J. Sutcliffe, J. Feltham, R.A. Cerione, & R.E. Oswald. Model building predicts an additional conformational switch when GTP binds to the Cdc42Hs protein. Prot. Peptide Lett., 1, 84-91 (1994).
A total of 41 models of the human Cdc42Hs protein, a member of the rho subfamily of the ras superfamily, have been built to investigate GTP binding. These suggest that a large scale conformational change is induced in Cdc42Hs on GTP bindingÑa flap forms over the GTP which withdraws on GTP hydrolysis or the release of GDP. This region of structure is absent from the H-ras and elongation factor Tu proteins but is present in other members of the rho subfamily.