Assignment of nitrogen-15 NMR signals in bovine pancreatic trypsin

Determination of the secondary structure of the DNA binding protein Ner from phage .mu. using proton homonuclear and nitrogen-15-proton heteronuclear ...
0 downloads 0 Views 378KB Size
7879

J . Am. Chem. SOC.1987, 109, 7879-7881 Table I. Rate and Equilibrium Constants for Binding Chloride Ion bv Hosts 1" host temp ("C) k, (s-') k f (M-I s - ' ) ~ Kq (M-I) lb'

-50 -40 -30

(2.0 f 0.8) (3.6 0.4j (8.4 f 0.8) (2.0 i 0.3) 4 x 103 3 x 104 6 X lo4 3 x 105 3 x 106

*

X

** 4)I ) xx lo'104 (3.7 f 0.8) IO4

lo2 (9

x io2 (2

John Glushka and David Cowburn*

45

56

X X IO2 44 x 103 (7 2) x 104 -20 35 1 x 10s -10 25 1 x 106 IC -20 33 5 3 x 105 Id -50 7 2 x 106 -20 3 I x 107 20 OResults from line shape analyses of spectra of CDC13 solutions containing hosts and tetrahexylammonium chloride. Error limits, when given, are l a for multiple determinations and do not contain an estimate of systematic errors. Single determination values are given without error limits; these values are believed to be accurate to better than f 2 5 % for k , and 150% for kf. kf was calculated from the k[ of the simulation and the concentration of free chloride. 'Averages of four values for -50 to -30 OC and three values for -20 O C .

*

of the slower exchanging C-8 host l b (with 0.5 equiv of C1-) contained two sharp signals at -50 OC that broadened on warming (Figure 1). HOST

+ C1-

(HOSTC1)-

Assignment of lSN NMR Signals in Bovine Pancreatic Trypsin Inhibitor

(1)

The binding of chloride by lb-d is described by the simple model in eq 1. Line shape analyses of the spectra containing hosts lb-d and chloride gave kinetic results: some of which are listed in Table I. The rates of complexation and decomplexation were slowed appreciably as the chain length decreased; we presume that this reflects steric interactions as the chloride squeezes between two chains to enter or exit the cavities. For host lb, studies of solutions of varying concentration indicated that decomplexation ( k , ) was unaffected by concentration; an Arrhenius treatment for this first-order decomplexation of chloride from the (lbC1)- complex gave an E, of 9.1 i 0.5 kcal/mol (two solutions, 15 measurements between -50 and -20 OC, the error limit is ln). The binding constants of hosts l b and ICat -20 O C were substantially greater than that for host Id; apparently the best cavity size for chloride occurs between hosts l b and IC. A more dramatic demonstration of size selectivity was found when the C-6 bicycle l a was studied. At room temperature the l19Sn N M R spectrum of l a ($148 ppm) was unaltered by the addition of excess chloride. The absence of line broadening or a second signal in the +30 to +40 ppm region in the spectra of l a showed that no complexation occurred to the limit of our detection capabilities. With the conservative estimate that even in the most difficult to detect case (slow exchange limit) we would have observed 5% complex if it was present, we can set a limit on Kq for l a of