From chemistry-request@ccl.net Wed Oct 9 12:10:54 1991 Date: Wed, 9 Oct 91 11:48:31 -0400 From: kxb3@po.CWRU.Edu (Kenneth Butenhof) To: chemistry@ccl.net Subject: NMR chemical shifts from structure -"summary" Status: R Netters, There have not been many responses concerning calculation of chemical shifts from structure. Below are the (edited) comments I received. Many of the studies referred to are still in press -- so perhaps interest in this area is picking up. From: FRAENKEL@MPS.OHIO-STATE.EDU Gideon Fraenkel, Chem. Dept., Ohio State There are two programs which can be used to calculate NMR chemical shifts starting with ab initio wavefunctions. They are based on much the same theory (Ramsay equation) and work in similar fashion. One is ARMPAC which comes from Jim Bowman, Northern Illinois, the other IGLO from Kutzelnig, Ruhr Universitaet. You probably need a 6-31G* calculation, fully optimized. For recent references I suggest you look at a current volume of Specialist Periodical Reports on NMR, Royal Society of Chemistry. Ditchfield at Dartmouth is someone who works in this field. Bill Kern and I wrote some papers in the seventies on ab initio calculations of chemical shifts. If by dynamic NMR you mean the effects of exchange and conformational interconversion on NMR spectra I refer you to the book by myself and J. I. Kaplan, NMR of Chemically Exchanging Systems, Academic Press, 1980, or the book edited by Jackman and Cotton,Dynamic NMR Spectroscopy, Academic Press, 1975. Gerhard Binsch wrote software, DNMR1 to DNMR6(?) which are distributed by QCPE,Indiana University. These only handle first order processes. We have all purpose software that let you calculate any exchanging system you want to imagine. However YOU have to derive the density matrix equations. Turning to proteins there are many contributions to chemical shifts which IGLO and ARMPAC do not take into account. These include electrostatic interactions and solvation. >From PCJ@PSUVM.PSU.EDU Peter Jurs, Chemistry Dept., Penn State We have been working in this area for several years. We work exclusively with small molecules. An overview of our approach is given in Analytical Chemistry, 61, 1115A (1989). We have published a number of papers in Analy. Chem. and Analy. Chim. Acta. We are now in the process of merging the database retrieval method and model building methods for prediction of CNMR of small molecules. >From case@scripps.edu David Case, Scripps Institute There is a great deal of current interest in looking at regularities in chemical shift distributions on protons and carbon-13 in proteins. In the realm of "semi-empirical" to "empirical" calculations, Klara Osapay and I have a paper in press in JACS that surveys proton chemical shifts in 21 proteins whose crystal structures are known. We find that with a combination of ring currents, peptide group anisotropy calculations, and electrostatic effects, that a reasonably good account of shifts of protons bonded to carbons can be obtained; root-mean-square deviations for about 5700 shifts are about 0.2 ppm for all shifts, and somewhat less, about 0.13 ppm, for methyl groups. The formulas used involve fitting parameters, but the best values match what is known, or thought to be known, about expected strengths of the various contributions. Our electrostatic model is very crude, and I know others are working on more complex models. As far as I know, out study is the first study of protein shifts at this level of empiricism since the early 80's, when complete assignments were first becoming available for proteins. Of course, there is a much literature on ring currents and magnetic anisotropy contributions to proton chemical shifts in organic systems.... My reading of the prevailing opinion amongst NMR people about similar calculations for nucleic acids is that that problem is much more difficult. For one thing, we don't know the detailed structure of oligonucleotides in solution (and crystal structures may be less relevant than for proteins); for another, ring current contributions from many rings are likely to be involved, which may make things difficult to untangle. It is sometimes possible to intepret particular (unusual) shifts in structural terms, but I have not seen a convincing general account. A preliminary account of some of our shift work appeared in Biochemical Pharmacology 40: 15-22 (1990). In terms of surveys, look for a paper soon to appear in J Mol Biol on proton shifts in proteins, and to the recent note in JACS by the NIH group on regularities in 13-C shifts in proteins. Q2) Is software currently available for STRUCTURE -> NMR chemical shifts The software we used to generate the data for our upcoming JACS paper (computing proton chemical shifts from protein structures using empirical formulas) is available via anonymous ftp from riscsm.scripps.edu. Look for the file pub/shifts.tar.Z >From kmoore@ncsc.org Kevin Moore, NCS My group at Duke University has been doing ab initio calculations for some time under the direction of Don Chesnut, and we have very good results with Carbon and Nitrogen (<5% error) and not so good with others (it is a mixed bag). Hydrogen shifts are good, but the range is small, and we have done a limited amount of large scale testing. Some recent reviews of the status of ab initio work in this area can be found in: (1)D.B. Chesnut, Ann. Rep. NMR Spect., 21, 51 (1989). (2)W. Kutzelnigg and M. Schindler, NMR Basic Principles and Progress, Springer-Verlag, Berlin, Vol. 23, 1990. >From Ken Butenhof, CWRU Those wanting an introduction to calculation of ring current effects may be interested in the (old) review: Progress in NMR spectroscopy, Vol 13 pp 303-344 by C.W. Haigh and R.B. Mellion -- Ken Butenhof Department of Pediatrics Case Western Reserve University Cleveland, Ohio 44106 -- Ken Butenhof Department of Pediatrics Case Western Reserve University Cleveland, Ohio 44106