Eur. J. Biochem. Z02, 417-430 (1979)

Conformation of Cobrotoxin in Aqueous Solution as Studied by Nuclear Magnetic Resonance Toshiya ENDO, Fuyuhiko INAGAKI, Kyozo HAYASHI, and Tatsuo MIYAZAWA Department of Biophysics and Biochemistry, Faculty of Science, University of Tokyo, and Faculty of Pharmaceutical Sciences, Kyoto University (Received July 7, 1979)

The 270-MHz proton NMR spectra of cobrotoxin from Naja naja atra were observed in 2 H 2 0 solution. The pK, value (5.93) of His-32 is slightly lower than the p K value (6.65) of the reference model of N-acetylhistidine methylamide, because of the electrostatic interaction with Arg-33 and Asp-31. The pK, value (5.3-5.4) of His-4 is appreciably low, because of the interaction with the positively charged guanidino group possibly of Arg-59. The hydrogen-deuterium exchange rates in 'H20 solution were measured of cobrotoxin and imidazole-bearing models. The second-order rate constants of N-acetylhistidine methylamide, N-acetylhistidine and imidazole acetic acid satisfy the Br$nsted relation. With reference to this Br$nsted relation, the imidazole ring of His-32 is confirmed to be exposed. The imidazole ring of His-4 is also exposed and the exchange rate is excessively promoted by the presence possibly of Arg-59 in the proximity. All the methyl proton resonances are assigned to amino-acid types, by conventional double-resonance method and more effectively by the spin-echo double-resonance method. Eight methyl proton resonances are identified as due to the y and/or 6-methyl groups of Val-46, Leu-1, Ile-50 and Ile-52 residues. The proximity of aromatic ring protons and methyl protons is elucidated by the analyses of nuclear Overhauser effect enhancements. The aromatic proton resonances of Trp-29 are affected by the ionizable groups of Asp-31, His-32 and Tyr-35. The methyl groups of Ile-50 are in the proximity to the aromatic ring of Trp-29 and the methyl groups of Ile-52 are in the proximity to Tyr-25. The highest-field methyl proton resonance is due to a threonine residue in the proximity to His-4. The appreciable temperature-dependent chemical shift of this methyl proton resonance suggests a temperature-dependent local conformational equilibrium around the His-4 residue of the first loop of the cobrotoxin molecule. Cobrotoxin is a neurotoxic protein isolated from the venom of Taiwan cobra (Naja naju atra). The primary structure of cobrotoxin has been determined by Yang et al. [1,2] (Fig. 1); this protein consists of 62 amino acid residues with four disulfide bridges and accordingly belongs to the class of short-chain neurotoxins [3] (Table 1). Neurotoxic proteins from snake venoms block the neuromuscular transmission at the post-synaptic membrane by the specific binding to acetylcholine receptor protein [4]. The primary structures of more than 50 neurotoxins have been determined, and the concept of structurally invariant and functionally invariant residues has been proposed [5]. For elucidating the chain conformation and the roles of amino acid residues of neurotoxins in solution, nuclear magnetic resonance spectra [6 - 121 have been analyzed. X-ray diffraction analyses have also been Abbreviations. NMR, nuclear magnetic resonance; NOE, nuclear Overhauser effect; Tyr(NOz), nitrotyrosine.

made [13-151 of the crystal of a sea snake toxin erabutoxin b. In our previous study [8], the 270-MHz proton NMR spectra of erabutoxin a and b in 'H20 solution were analyzed. The microenvironments of Tyr-25, His-26 and Trp-29 residues and methyl proton signals of valine and isoleucine residues are consistent with the locations of a-carbon atoms as elucidated by X-ray crystal analyses [13- 151. However, the methyl proton signals of erabutoxin b in aqueous solution have been assigned only in part because of signal overlapping [8]. Cobrotoxin is homologous to erabutoxin (see Table 1) and the Tyr-35 residue of cobrotoxin may be chemically modified, without affecting the neurotoxicity [16]. The Tyr(N02)-35 residue of the modified cobrotoxin is expected to be a binding site for lanthanide ions [17]. Accordingly in the present study, the 270-MHz proton NMR spectra of cobrotoxin were analyzed in detail.

Conformation of Cobrotoxin in Solution

418 Table 1. Amino acid sequences of short chain neurotoxins [3/ C = Cys, D = Asp, E = Glu, F = Phe, G = Gly, H = His, I S = Ser, T = Thr, V = Val, W = Trp, Y = Tyr

Ile, K = Lys, L

=

Neurotoxin

=

Leu, N

=

Asn, P

=

Pro, Q

=

Gln, R

_ _

=

~

30

10 Cobrotoxin ( N n a p atra) Neurotoxin I ( N mommbrca m ) Neurotoxin I1 ( N n a p oxrana) Erdbutoxin b ( L wnrfasciata) ~

_

_

L E C H N Q Q S S Q T P T T T G C S G G E T N C Y K K R W R D

L E C H N Q Q S S E P P T T T R C S G G E T N C Y K K R W R D L E C H N Q Q S S Q P P T T K T C S G - E T N C Y K K W W S D R I C F N Q H S S Q P Q T T K T C P S G E S S C Y H K Q W S D

p

p

~

-

p

-

-

_

p

_

_

-

-

~

p

-

_

- -

p

_ _ ~ Cobrotoxin ( N . n a p atra) Neurotoxin I ( N mossamblta m ) Neurotoxin I1 ( N . n a p oxruna) Erabutoxin b ( L semlfa~crafa) _

p

p

~

-

-~

_

-

~

H H H F

R R R R

G G G G

Y Y T T

~

_

_

.

55

35 p

_

Residues

Neurotoxin

p -

Arg,

_

R R I I

_

T T I I

E E E E

-~

p

R R R R

G G G G

_

C C C C

G G G G

C C C C

P P P P

_

S T K T

V V V V

~

K K K K

p _

N K P P

_

G G G G

_

I I V I

-

E E N K

I L L L

_

N N S S

_

C C C C

_

C C C C

p

T T R E

~

T T T S

_

D D D E

__

_

R R R V

C C C C

N N N N

N N N N

previously [18]. Chemically modified cobrotoxin with the Tyr(N02)-35 residue was prepared, by treating cobrotoxin with tetranitromethane [16]. Deuterium Labelling

Fig. 1. Primary structure o/coblotoxin [1,2/

By a combined use of various methods, all of the methyl proton resonances of cobrotoxin in aqueous solution were classified in amino-acid types and half of the methyl resonances were assigned to individual residues. Furthermore, the microenvironments and dynamic properties of aromatic residues were elucidated. These observations will be discussed, in comparison with the molecular properties studied previously for neurotoxin I1 from Naja naja oxiana [6,12], neurotoxin I from Naja mossambica mossambica [7, lo], erabutoxins a and b from Laticauda semfasciata [8], and cobrotoxin [11]. MATERIALS AND METHODS Materials Cobrotoxin was isolated from the crude mixtures of toxins of Naja naja atra by the method described

For establishing the assignments of histidine C-2 proton resonances of proteins to individual residues, deuterium-labelling experiments are useful [19]. Cobrotoxin (6 mg) was dissolved in 2H20 (0.4 ml, in a 5-mm NMR tube) and the pH was adjusted to 9.0. The solution was incubated at 37 "C and the extent of hydrogen -+ deuterium exchange was monitored by the intensities of two histidine C-2 proton resonances. After the incubation for 165 h, one of the two C-2 protons was almost completely exchanged. Then this selectively-deuterated cobrotoxin was reduced and carboxymethylated, and was digested by trypsin (obtained from Sigma Chemical Company and recrystallized twice) for 6 h at room temperature at pH 8.2 (with 0.01 M NH4HC03). The digested protein was gel-filtered (Sephadex G-50, 1 x 116 cm; elution medium, 1 acetic acid) and the fraction I containing the large peptide fragments (Leu' - His4- L y ~ ~ ~ - L y s ~ ~ Arg2*, Leu' - Lys27 and Leu' - L Y S ~was ~ ) separated from the fraction I1 containing the tripeptide (Asp31H i ~ ~ ~ - Aand rg~ other ~ ) small peptide fragments. N M R Measurements 270-MHz proton NMR spectra were recorded on a Bruker WH-270 spectrometer, equipped with Nicolet 1180 computer system. pH values (direct pH-meter readings) were measured by a Radiometer PHM-26 pH meter. NMR samples of cobrotoxin were dissolved in 2H20 (with 0.1 M NaC1) at the concentration of 2-3 mM and the pH was adjusted by adding 2HC1

419

T. Endo, F. Inagaki, K. Hayashi, and T. Miyazdwa

or N a 0 2 H . Labile O H and N H hydrogen atoms of cobrotoxin were completely exchanged with deuterium atoms by incubating the samples in 2H20 solution for about 30 min at 50 "C. Chemical shifts were measured from the internal standard of sodium 2,2dimethyl-2-silapentane 5-sulfonate. The pH dependences of chemical shifts were analyzed by non-linear least-square simulation. Nuclear Ovevhauser Effect Enhancement For elucidating the spatial proximity of protons of different amino-acid residues of cobrotoxin, interproton NOE enhancements were measured. Several upfield-shifted methyl protons or aromatic protons were irradiated for 1 s before data acquisition and negative NOE enhancements were extracted by the use of NOE difference method [20]. Spin-Echo Double-Resonance Method For identifying the spin multiplicities, especially of methyl proton resonances of cobrotoxin, the spinecho double-resonance method [21,22] was applied. Spin-echo spectra may be obtained by the Fourier transformation of the spin echo after the 90"-~-180"-~ pulse sequence. By setting the delay time (22) equal to l / J ( J = spin-spin coupling constant of CH3 - C H < or CH3 - CH2 -), the doublet and triplet methyl resonances are observed as negative (180" out-ofphase) and positive (in-phase) signals, respectively. However for proteins, the spin multiplets of methyl proton resonances may not be readily identified because of signal overlappings. Accordingly in the present study, the spin-echo double-resonance method was used. By irradiation of the C H proton of a CH3 -

I

I

8.5

8.0

Fig. 2. 270-MHz proton N M R spectrum (aromatic region)

I

CH: group during the delay time ( 2 ~but ) not during the acquisition of spin echo, the doublet methyl proton resonance was converted to a positive signal. Therefore by taking the difference between the off-resonanceirradiated spin-echo spectrum and on-resonance-irradiated spin-echo spectrum, a negative doublet signal is extracted out of overlapping resonances and this doublet signal is assigned to the methyl protons which are spin-coupled with the C H proton subject to irradiation. On the other hand, by irradiation of one of the CH2 protons of a CH3 -CH2 - group during the delay time but not during the acquisition of spin echo, the triplet methyl proton resonance is converted to a negative signal. Therefore by taking the difference between the off-resonance and on-resonance-irradia ted spin-echo spectra, a positive triplet signal is extracted out of overlapping resonances and this triplet signal is assigned to the methyl protons which are spin-coupled with the C H proton subject to irradiation. This method was free from the difficulties such as those arising from the collapsing of resonance structures in normal spin-decoupled difference spectra.

RESULTS Assignments of Histidine Resonances The 270-MHz proton spectrum of cobrotoxin in 2H20 solution at pH 6.3 is shown in Fig.2 (aromatic region). The pH dependence of proton chemical shifts in the aromatic region is observed as shown in Fig. 3 and is treated by the nonlinear least-square method for calculating the chemical shift change and pKa values (Table 2).

I

7.5 70 Chemical shift (pprn) of

1

I

6.5

60

cobrotoxin in ' H z O solution at p H 6.3 and 23 'C

420

Conformation of Cobrotoxin in Solution

Table 2. pK, (and standard deviation CT)and protonation changes ( A S ) of' aromatic proton chemical shifts

PPm His-4 C2 c4

12 2

His-32 C2 c4

13 8

3.06 (0.49) 3.00 (0.11)

- 0.01

Tyr-25 C2,6 c3,5

3 1

3.85 (0.56) 3.23 (0.34)

- 0.01 - 0.04

Tyr-35 C2,6 c3,5

6 4

2.74 (0.26)

- 0.04

Trp-29 C4/7 C4/7 C5/6 c2

11 10 9 7

3.07 (0.13) 2.91 (0.18) 3.67 (0.06) 2.99 (0.05)

0.03 0.04 -0.11 0.08

- 0.04

5.33 (0.01) 5.41 (0.02)

PPm 1.32 0.23

5.93 (0.01) 5.93 (0.01)

0.97 0.37

5.95 (0.22)

PPm

0.03

-

5.87 (0.36)

0.01

5.97 (0.05)

0.06

11.75 (0.13) 12.10 (0.07)

0.21 0.33

9.67 (0.05) 9.81 (0.03)

0.13 0.27

9.63 (0.13) 9.48 (0.17) 10.02 (0.32) 9.74 (0.15)

--

0.03 0.03 0.02 0.02

75 9.0

Trp-29 C X 6 - t

-E i -._

E

His-32 C4-H Trp-2s C2-H

7.0

-

85

a

-

a

._ L

c

-

L Ln

.-"mE

-m" ._ 5

a

6 6.5

L

0

8.0

6.0

75

I

1

l

I

I

l

I

1

I

1

I

I

I

I

2

4

6

8

10

12

14

2

4

6

8

10

12

14

PH

PH

Fig.3. p H dependence ofthe proton chemical shifts (aromatic region) of cobrotoxin in ' H z 0 solution at 23 "C

Cobrotoxin has two histidine residues, His-4 and His-32. The singlet signals, peaks 13 and 8 in Fig.2, are assigned to the C-2 and C-4 protons of a histidine residue (His-a) because of the common pK, value (z5.9) (Fig. 3 and Table 2), while the peaks 12 and 2 (in Fig. 2) are assigned to the C-2 and C-4 protons of the other histidine residue (His-b). For identifying these histidine residues (a and b), the deuterium-labelling method [19] was used in the present study. The proton NMR spectrum of the deuterium-labelled cobrotoxin is shown in Fig. 4B, in comparison with the spectrum of unlabelled cobrotoxin (Fig.4A). For the labelled cobrotoxin, the C-2 proton resonance of His-a is clearly observed but the C-2 proton resonance of His-b is hardly observed.

Then after the trypsin digestion, the histidine C-2 proton resonance is clearly observed for the 2 H 2 0 solution (pH 3.0) of fraction I1 (with His-32) (see Fig. 4 D) but is only weakly observed for the solution of fraction I (with His-4) (see Fig.4C). Accordingly His-a and His-b are now unambiguously identified as His-32 and His-4, respectively. Previously, the same assignments of histidine C-2 protons were made on the assumption that the presence of Asp-31 near His-32 would raise the pK, value of His-32 above the pK, of His-4 (11). However, this assumption is not necessarily valid, since even the pK, value of His-32 (z5.9) of cobrotoxin is appreciably lower than the pK, value (6.6 at 23°C) of the model molecule of N-acetylhistidine methylamide [24].

T. Endo, F. Inagaki, K . Hayashi, and T. Miydzawa

42 1

His - a

cally modified cobrotoxin [16]; Tyr-a and Tyr-b are identified as Tyr-25 and Tyr-35, respectively. These assignments are confirmed, in the present study, by the observation of the effect of nitration on the aromatic proton resonances. The aromatic proton signals of Trp-29 have been assigned previously [ l l ] ; the doublet signals (peaks 11 and 10) are due to the C-4 and C-7 protons, the triplet signals (peaks 9 and 5) to the C-5 and C-6 protons, and the singlet signal (peak 7) to the C-2 proton.

c4

x

c ._ Lo

c4

I

a,

c ._ c

a, ._

His-4

c

m -

d C

D I

I

9

8

I

7 Chemical shift (ppm)

I

6

Fig.4. 270-MHz proton N M R spectra of ( A ) unluhellrd cobrotoxin and ( B ) deuterium-labelled cobrotoxin in 2H20solution at p H 9.0 and 23 ' C , ( C ) trypsin-digestedpeptidefraction I and (0) fraction I I in 'H2O solution at pH 3.0 and 23 "C

Only after the deuterium-labelling experiments in the present study can the assignments of histidine proton resonances of cobrotoxin now be established. Assignments of Aromatic Proton Resonances

In addition to histidine resonances, aromatic proton signals of Tyr-25, Tyr-35 and Trp-29 residues are expected to lie in the aromatic region. The doublet signals (peaks 6 and 4 in Fig.2), each corresponding to two protons, are appreciably shifted upfield as the pH is raised to around 9.7 (Fig. 3). The pH-dependent shift of the C-3,5 protons of a tyrosine residue is known to be larger than that of C-2,6 protons [25]. Accordingly, the peaks 6 and 4 are assigned to the C-2,6 and C-3,5 protons, respectively, of a tyrosine residue (Tyr-a). In addition, the broad signals (peaks 3 and 1) are also shifted upfield as the pH is raised to around 11.5 (Fig.3), and are assigned to the other tyrosine residue (Tyr-b). The identification of the two tyrosine residues (Tyr-a and Tyr-b) has been made previously [ l l ] on the basis of the pK, values of cobrotoxin and chemi-

Assignments of Methyl Proton Resonances to Amino Acid Types

The molecule of cobrotoxin has eight threonine (at positions 11, 13, 14, 15, 22, 37, 56 and 57), one valine (Val-46), one leucine (Leu-1), and two isoleucine residues (Ile-50 and Ile-52), and accordingly has 16 methyl groups. The resolution of methyl proton signals may be much improved by the use of the convolution difference technique [26]. As shown in Fig. 5, the methyl proton signals are well resolved in the convolution-difference proton NMR spectrum (270 MHz) of cobrotoxin, and the spin-spin coupling patterns are clearly seen for many signals. Complete classification of these methyl proton resonances to amino-acid residues has been accomplished, in the present study, by the double-resonance method and by the spin-echo method [21,22]. Cobrotoxin does not have methionine residues so that no singlet methyl proton resonances are expected to appear. The triplet signal of cobrotoxin at 0.56 ppm (Fig. 5, peak 2) is readily assigned to the &methyl group of an isoleucine residue. The assignments of doublet signals have been made with the same procedure as described previously [8]; the signals 16, 15, 14, 13, 12, 11, 10 and 1 (Fig.5) are from the y-methyl groups of threonine residues and the doublet signals (7 and 6) are from the y-methyl groups of a valine residue or from the &methyl groups of a leucine residue. Further assignments of methyl proton resonances to amino acid residues were made, in the present study, by the spin-echo double-resonance method. This method is especially useful for the assignments of methyl proton signals in complicated spectra. Representative examples for the comparison of the conventional double-resonance method and the spinecho double-resonance method is shown in Fig.6. The convolution-difference spectrum of cobrotoxin (Fig. 6A) may be compared to the convolution-difference spectrum with the irradiation at 1.42 ppm (Fig. 6B). However, the effect of the irradiation at 1.42ppm is not clear, in spite of the resolution enhancement by the convolution-difference technique. On the other hand, the on-resonance-irradiated (at

Conformation of Cobrotoxin in Solution

422

16 15 14 13 12 11

10 9 8 7 6 5 4 3 2 1

Thr-h y Thr-g y Thr-f y Thr-e y Thr-d p Thr-c y Thr-b y Ile-52 y Ile-50 p Val-46 y ' Val-46 y 11'6-52 6 Leu-1 6' Leu-1 6 Ile-50 6 Thr-a y

1.5

.o

1 Chemical shift (ppm)

0.5

Fig. 5. Methyl region of convolution-di~f.rmcr proton N M R spectrum (270 M H z ) of cobrotoxin in ' H z 0 solution at p H 6.3 crnd 33

c

A

2

C

2

rn

u u

4 I 1.0

3 I

I 1.0

0.5

I

0.5

Chemical shift (pprn)

Fig.6. Methyl proton resonances (at 270 M H z ) of cobrotoxin in 'H'O solution at p H 6.3 and 33 C. (A) Convolution-difference spectrum without irradiation, (6) convolution-difference spectrum with the irradiation at 1.42 ppm, (C) spin-echo spectrum (with the delay time of 140 ms) with off-resonance irradiation, (D) spin-echo spectrum wit.$ on-resonance irradiation at 1.42 ppm, (E) spin-echo difference spectrum (C-D) for the irradiation at 1.42 ppm, (F)spin-echo difference spectrum for the irradiation at 1.23 ppm

T. Endo, F. Inagdki, K . tlayashi, and T. Miyazawa

Table 3 . pK, (and standard deviation

Proton

Peak

Val-46 7' Leu-I 6' 6 Ile-50 6 Thr-a 7 Thr-c 7

7 4 3 8 2 1 11

0) and

423

protonation changes ( A S ) of methyl proton chemical shifts

PfG (4

A6

PKa

A6

(0)

PPm

);

3.14 (0.29)

0.02

3.06 (0.28) 3.91 (0.07)

0.02 0.10

1.42 ppm) spin-echo spectrum (with the delay time equal to l/J) is shown in Fig. 6D, in comparison with the off-resonance-irradiated spin-echo spectrum (Fig. 6C). By taking the difference between these spinecho spectra, three well defined signals are now extracted (Fig. 6E). The pair of negative doublet signals (4 and 3) are clearly assigned to y-methyl groups of a valine residue or to &methyl groups of a leucine residue. Further, the positive triplet signal (2) due to the 8-methyl group of an isoleucine residue is found to couple with one of the y-methylene protons at 1.42 ppm. The spin-echo difference spectrum is also observed for the irradiation at 1.23 ppm, and a positive triplet signal is clearly extracted as shown in Fig. 6 F . This triplet signal ( 5 ) is readily assigned to the &methyl group of the other isoleucine residue. Finally, the remaining two doublet signals (9 and 8 in Fig.5) are unambiguously assigned to the y-methyl groups of two isoleucine residues.

PfG (4

AS

10.34 (0.05) 9.35 (0.06) 9.45 (0.06) 10.62 (0.06) 10.64 (0.06)

0.03 0.06 0.06 0.06 0.10

PPm

5.42 (0.18)

- 0.02

5.29 (0.03) 4.84 (0.13)

0.20 0.02

PPm

10.25 (0.13)

- 0.02

16

12 15 13 14 11

g E 1.00

10 9

c ._

0.50t

I

213

1

I

I

2

4

6

I 8

I

I

10

12

1 14

PH

Fig.7. p H dependence of the proton chemical shifts (methyl proton region) of cobrotoxin in ' H z O solution at 23 "C

Methyl Proton Resonances of Individual Residues For the assignments of methyl proton resonances to individual residues, the pH dependence of chemical shifts is observed as shown in Fig. 7 and is treated by the nonlinear least-square method for calculating the chemical-shift changes and pK, values (Table 3). The pair of signals (7 and 6) and the other pair of signals (4 and 3) have already been found to be due to y-methyl groups of Val-46 or &methyl groups of Leu-1 (explained in the last section). The 8-methyl signals of Leu-1 at N-terminus are expected to shift upfield by about 0.04 ppm on deprotonation of the a-NH: group [25]. As shown in Fig.7 and Table 3, the titration shifts with A 6 = 0.06 ppm are observed around pH 9.4 for the signals 4 and 3, while the signal 7 shifts only by 0.03 ppm around pH 10.3 and the signal 6 does not shift at all in the alkaline pH region. Therefore the signals 4 and 3 are now assigned to the &methyl protons of the N-terminal residue of Leu-1 and subsequently the signals 7 and 6 are assigned to the y-methyl protons of Val-46.

The assignments of methyl proton signals due to the isoleucine residues (Ile-50 and Ile-52) of cobrotoxin are made, in the present study, by the comparison with the proton N M R spectrum of neurotoxin I (Naja mossambica mossambica) [lo]. The primary structures of cobrotoxin and neurotoxin I are the same except for six substitutions (Fig. 1): Gln-10, Thr-11, Gly-16, Ser-45, Asn-48 and Ile-52 in cobrotoxin are replaced by Glu-10, Pro-1 1, Arg-16, Thr-45, Lys-48 and Leu-52 in neurotoxin I. Thus, there are 13 homologous methyl groups for these two toxins and in fact the proton NMR spectra (and pH dependences) of these toxins in the methyl proton region are very similar, suggesting that these toxin molecules take homologous conformation in solution. However the triplet signal (5) of cobrotoxin is not observed for neurotoxin I [lo] and accordingly this peak ( 5 ) is now assigned to the &methyl group of Ile-52. The other triplet signal (2) of cobrotoxin is also observed for neurotoxin I [lo] and accordingly this triplet

Conformation of Cobrotoxin in Solution

424

Ile-50

G

His-4 0

'Ile-50

Trp-29

Thr-a

u c

His-4 0

c

.-c

His-4 0

a > , .c

B-

m -

$

h

I

I

I

I

I

8

7

6

1

0

Chemical shift (pprn)

Fig. 8. ( A ) The convolution-di~ferenc.e protori N M R spectrum (270 MH:) oJ cobrotoxin in ' H 2 0 solution at p H 6.0 and 23'C. Black arrows indicate the proton resonances to be irradiated (b, Thr-ay; c, Ile-506; d, Leu-I 6 ; e, Ile-52 6 and Leu-1 6'; f, His-4 C-2; g, Trp-29 C-4 and C-7; h, Tyr-25 C-2,6); (B-H) the difference spectra between off-resonance pre-irradiated (for 1 s) and on-resonance preirradiated (for 1 s at b-h, respectively) spectra; white arrows indicate the assignments of negatively enhanced proton resonances

signal is assigned to the S-methyl group of Ile-50. The triplet signal (2) and the doublet signal (8) of cobrotoxin are simultaneously shifted upfield as the pH is raised to around 10.6 (pKa, see Table 3) so that this doublet signal (8) is assigned to the y-methyl group of Ile-50. The doublet signal (9) is then assigned to the y-methyl group of Ile-52. Thus, all the methyl proton resonances of valine, leucine and isoleucine residues of cobrotoxin are now assigned to y or &methyl groups of individual residues.

NOE of Aromatic-Proton and Methyl-Proton Resonances

For elucidating the spatial arrangements of aromatic rings and methyl groups, NOE enhancements [27] were measured of cobrotoxin in 2H20 solution. Methyl proton resonances were irradiated as shown with black arrows (b- e) in Fig. 8 A and negative NOE enhancements of aromatic proton resonances were observed as shown in Fig. 8 B - E, respectively, by the NOE difference method. Furthermore, aromatic proton resonances were irradiated (f- h in Fig. 8A) and negative NOE enhancements of methyl proton resonances were observed as shown in Fig.8F-H.

Such NOE enhancement depends on the inverse sixth power of the interproton distance and also on the angle between the interproton vector and the internalrotation axis of the methyl group. Therefore, the observed NOE enhancements may not simply be related to the interproton distance. Nevertheless proximate proton pairs may be elucidated by the analyses of NOE enhancements. The irradiation of the y-methyl protons of Thr-a (black arrow b) causes NOE on C-2 and C-4 proton resonances of His-4 (Fig. 8 B) and on the fl-CH proton resonance (at 3.73 ppm) of Thr-a. Conversely, the irradiation of the C-2 proton of His-4 (black arrow f) causes NOE on the y-methyl proton resonance of Thr-a (Fig. 8 F). These observations indicate the proximity of His-4 and Thr-a residues of cobrotoxin. The irradiation of the S-methyl proton of Leu-1 causes NOE on the C-4 proton resonance of His-4 (Fig. 8 D), indicating the proximity of Leu-1 and His-4 residues. In fact for the chemical shift of the y-methyl proton resonance of Leu-I (peak 3 in Fig.7), an inflection is observed at pH z 5.4. This inflection corresponds to the pKa value of His-4 of cobrotoxin. The irradiation of the &methyl proton of Ile-50 causes NOE on the C-2 proton resonance of Trp-29 (Fig. 8C) and conversely the irradiation of the C-4/C-7 proton resonance of Trp-29 causes NOE on the y-methyl and &methyl proton resonances of Ile-50 (Fig. 8 G). These observations indicate the proximity of Trp-29 and Ile-50 residues. Finally, the simultaneous irradiation of the &methyl proton of Ile-52 and 6'methyl proton of Leu-1 causes NOE on the C-2,6 and C-3,5 protons of Tyr-25 (Fig. 8 E) and conversely the irradiation of the C-2,6 protons of Tyr-25 causes NOE on the y and &methyl proton resonances of Ile-52 (Fig. 8 H). These indicate the proximity of Tyr-25 and Ile-52 residues in the molecule of cobrotoxin. Hydrogen-Deuterium Exchange Rates qf' Histidine Residues

For hydrogen -,deuterium exchange of histidine C-2 protons of cobrotoxin, the sample solution in 2H20 was adjusted to pH 8.95 at 37 'C and was incubated for 250 h. The intensities of the C-2 proton signals of His-4 and His-32 were measured at appropriate intervals, where the C-2 proton signal of Trp-29 (at 6.99 ppm) was used as the internal intensity standard. The exchange reactions of the two histidine residues of cobrotoxin were found to be first-order and the rate constants ( k o b s d ) were obtained as shown in Table 4. For the analyses of the exchange rates in relation to the solvent accessibility (see Discussion), the pH-dependence of the C-2 proton chemical shifts were also measured at 37'C and pK, values were obtained as shown in Table 4. The hydrogen-+dew

425

T Endo, F. Inagdki. K . Hayashi. and T. Miyazawa Table 4. Psrudo-first-order exchange rate constant ( k o h 7 d ) . pK, and second-order rate constant (kh) at 37 ^C kobrdand pK, values (at 40°C) for ribonuclease A were from [30], pK, value for N-acetyl-L-histidine methylamide from 1241 and pK, values for the last five compounds from [23]

3.5 oHis-4

30

-histidine methyl ester

\ , '\\

Compound

10' x kobsd (pH)

pK,

IO-'x

kh

--

S-l

Cobrotoxin His-4 His-32 Ribonuclease A His-12 His-48 His-I05 His-I 19 N - Acetyl-L-histidine methylamide N- Acetyl-L-histidine Imidazole acetic acid Histamine L-Histidine L-Histidine methyl ester

'\

2.5

'\, .histidine

i

2.0 . -t"

His-320 His-12.

7-

5.72 (8.95) 1.97 (8.95)

5.11 5.72

1.73 0.146

1.75 0.19 5.56 4.00

(8.0) (8.0) (8.0) (8.0)

5.79 6.31 6.72 6.19

0.111 0.0037 0.043 0.103

3.89 (9.6) 5.56 (10.0) 6.94 (9.0) 5.28 (10.1) 5.44 (9.8) 6.22 (10.5)

6.42 7.07 7.55 6.14 5.91 5.33

0.058 0.0184 0.0079 0.149 0.260 1.131

-

'.

-histamine "..,oHis-119

Y-acetylhistidine

u' 9 1.5

.o

1

imidazole acetic ac& '\

*His-48

0.5

terium exchange rates (and pK, values) were also measured of imidazole-bearing model molecules in 2H20 solution (0.2 M NaCl) at 37"C, as shown in Table 4. DISCUSSION Microenvironments of Histidine Residues Cobrotoxin has two histidine residues, His-4 and His-32, homologous to the histidine residues of neurotoxin I (Nuju mossambicu mossambicu) and neurotoxin I1 (Nuju nuju oxianu). His-32 of cobrotoxin is more sensitive to photooxidation than His-4 and the lethal activity and antigenic activity of cobrotoxin are appreciably reduced upon photooxidation [28]. The pK, values of His-4 and His-32 residues are obtained as 5.33 and 5.93, respectively, from the nonlinear leastsquare treatments of the pH dependences of C-2 and C-4 proton chemical shifts (Table 2). These pK, values are appreciably lower than the pK, value (6.65) of N-acetylhistidine methylamide, a standard model for exposed histidine residues free from specific perturbations. Such low pK, values of the histidine residues of cobrotoxin may be due to the spatial proximity of positively charged group(s) [23], hydrogen-bond formation or hydrophobic environments [6]. The microenvironments of histidine residues of proteins may be studied by the analyses of the hydrogen + deuterium exchange reactions of C-2 protons. In the present experimental condition, the basecatalyzed exchange reaction is predominant and ac-

'.

1

5

I

I

I

6

I

I

7

P Ka

Fig. 9. BrOnstead plot f b r hydrogun --+ deuteriunz eschange reactiotis of model molecules (W), His-12, His-48, His-105 and Hi,s-I19 residues of rihonuclease A (@) and His-4 and His-32 residues flf c.ohrotoxin (OJ

cordingly the pseudo-first-order rate constant is given by the following equation [29]: kobsd

(kobsd)

= k b [Im(CH)'l [02H-I/[Im(CH)l~ = kbKw/(Ka [2H30+])

+

(1)

where [Im(CH)lo is the total concentration of imidazole ring with the C-2 hydrogen atom, [Im(CH)+]the concentration of imidazolium cation, k b the secondorder rate constant, K, the dissociation constant of imidazolium cation, and Kw the ion product of water (10-'3.6 at 37 "C). In this base-catalyzed reaction, the k b and K, values of imidazole-bearing molecules are expected to satisfy the following Br4nsted equation [30,31] : log k b = MpKa p (2) where a and fl are constants for the second-order reaction. This Br4nsted relation has, in fact, been found to hold for the hydrogen +tritium exchange reaction of trans-urocanic acid, N-acetylhistidine methylamide, N-acetylhistidine and imidazole propionic acid [31]. In the present study, the hydrogen + deuterium exchange measurements were made of N-acetylhistidine methylamide, N-acetylhistidine and imidazole acetic acid. As shown in Fig.9, the k b values for hydrogen + deuterium exchange and p& values of these three model molecules are also found to satisfy the Br4nsted relation (see Appendix). The Br4nsted relation applies for simple imidazole derivatives or histidine-bearing peptides but not to histidine residues buried in protein globule, and accordingly the deviation of k b values (lower values)

Conformation of Cobrotoxin in Solution

426

35

His - 32

w Arg - 33

Fig. 10. Sclzemtitic. rrprvsentution ofthe microenvironments of aminoarid residues r$t&rotouin in uqueous solution us elucidated by N M R analyses. Main-chain conformation is taken from the crystal structure of erabutoxin [14]. The proximity of amino-acid residues is confirmed by the pH-dependent chemical shifts (black arrows), NOE enhancements (white arrows) or perturbations on pKa values of ionizable groups (striped arrows)

from the Br$nsted relation is suggested to be a measure of the inaccessibility of histidine residues to aqueous solvent [31]. In fact the kb values (Table 4) for the exposed residues (His-105 and His-1 19) of ribonuclease A at 40 'C [30] lie close to the straight line for the Br4nsted relation (Fig.9) whereas the kb value for the deeply buried residue (His-48) lies appreciably below the straight line (Fig. 9). For cobrotoxin, the kb and pK, values of His-32 closely satisfy the Br4nsted relation, indicating that this residue is in fact exposed. The low pK, value of this histidine residue is probably due to the electrostatic interaction with a positively charged side chain of arginine or lysine residue. This is confirmed by the observation that the pK, value (5.75) of His-32 in the absence of NaCl is appreciably lower than the pK, value (5.93) in the presence of NaCl(O.1 M). The side chain of His-32 is not far from the positively charged guanidino group of Arg-33 because the main chain is folded around these residues (Fig. 10). Furthermore, the chemical shifts of the C-4 and C-2 proton resonances of His-32 are affected by an ionizable group with the pK, value of 3.0 (Fig.3 and Table 2). This

observation indicates that the imidazole ring of His-32 of cobrotoxin is affected by the positively charged group of Arg-33 and by the negatively charged carboxylate group, possibly of Asp-31. The second-order rate constant for the hydrogen + deuterium exchange of His-4 of cobrotoxin (see Fig. 9) is appreciably larger than expected from the Br4nsted relation. Similar upward deviations from the Br$nsted relation are also observed, in Fig. 9, for the kb values (Table 4) of histidine methyl ester and histidine which have an c[ NH; group. These observations indicate that the imidazole ring of His-4 of cobrotoxin is exposed and the hydrogen + deuterium exchange rate is enhanced, possibly by the proximate presence of a positively charged group. For a homologous neurotoxin, erabutoxin b, an appreciable NOE was observed between the aromatic ring protons of Phe-4 and y-methyl protons of Val-59 (Inagaki, F. et al., unpublished), indicating that these residues are in close proximity. Similarly for cobrotoxin, the imidazole ring of His-4 is expected to be close to the positively charged side chain of Arg-59. Microenvironments of Tyrosine Residues

The least-square treatment of the pH dependence of the chemical shifts of Tyr-25 yielded the pK, values of 21.8 (C-2,6 proton resonance) and 12.1 (C-3,s proton resonance) (Table 2). This unusually high pK, values of Tyr-25 is possibly due to the intramolecular hydrogen bonding (Fig. 10) with the carboxylate group of Glu-38 [8,9]. In fact, the intramolecular hydrogen bond between the hydroxyl group of Tyr-25 and carboxylate group of Glu-38 has recently been confirmed by the X-ray analysis of erabutoxin b in crystal (Petsko, G. A,, private communication). For studying the effect of pH on this intramolecular hydrogen bond of cobrotoxin in aqueous solution, the dependences of proton chemical shifts upon pH down to 1.3 were observed as shown in Fig. 3 . However, the chemical shifts of the C-2,6 and C-3,s proton resonances were affected little by pH down to 1.3 until finally the reversible acid denaturation of protein occurred. This observation suggests that the intramolecular hydrogen bond between Tyr-25 and Glu-38 is in fact important for maintaining the native conformation of cobrotoxin. For Tyr-35 of cobrotoxin, pK, is obtained as 9.7-9.8 from the pH dependences of the chemical shifts of aromatic proton signals (Table 1). However, the chemical shift of the C-2,6 proton signal of Tyr-35 is also affected by the pH change around 5.9.5 (pK, for the deprotonation of the imidazolium ring of His-32), indicating the proximity of the aromatic ring of Tyr-35 to the imidazole ring of His-32 (Fig.10). The C-2,6 proton chemical shift is further affected by the pH change around 2.7 (Fig. 3 and Table 2), suggesting

T. Endo, F. Inagaki, K . Hayashi, and T. Miyazawa

the proximity of a carboxylate group, possibly of ASP-31. Microenvironment of Tryptophan Residue Trp-29 of cobrotoxin is also found for homologous short-chain neurotoxins (Table 1). This residue is a functionally invariant residue of neurotoxins [32] and is supposed to play an essential role in the binding to the acetylcholine receptor protein. From the least-square treatment of the pH dependences of aromatic proton chemical shifts (Fig.3 and Table 2), the indole ring of Trp-29 of cobrotoxin is found to be affected by the ionizable groups with the pKa values of about 3.0, 5.9 and 9.7. The inflection of pH dependence of the C-4/C-7 proton chemical shift (peak 10) of Trp-29 around pH 9.7 is clearly due to the deprotonation of Tyr-35, since the corresponding inflection of the Trp-29 resonance of Tyr(N02)-35 cobrotoxin is found around pH 6.5. These observations indicate that the indole ring of Trp-29 is in proximity to the aromatic ring of Tyr-35. In fact, this pH dependence of Trp-29 of cobrotoxin is also observed for neurotoxin I (with Tyr-35) (see Fig.5 of [lo]) but not for erabutoxin b (with Thr-35 in place of Tyr-35). On the other hand, the inflections of tryptophan chemical shifts at about pH 3.0 and at 5.9 are commonly observed for cobrotoxin (Fig. 3) and neurotoxin I1 (Fig.5 of [lo]), indicating the proximity of Trp-29 of cobrotoxin to the carboxylate group possibly of Asp-31 as well as to the imidazole ring of His-32 (see Fig. 10). Thus, the positively charged groups of His-32 and Arg-33, the negatively charged group of Asp-31, and aromatic groups of Trp-29 and Tyr-35 residues are located around the tip of the central loop of cobrotoxin. Spatial Arrangements of Methyl Groups and Aromatic Rings Methyl groups of proteins are widely distributed over the protein globule and the differences between the methyl proton chemical shifts in the native state and in the denatured state are largely due to the ringcurrent effect of proximate aromatic rings in the native conformation. Such ring-current shifts depend upon the positions of methyl groups relative to aromatic ring and accordingly serve as sensitive probes for conformational changes of proteins. Furthermore, methyl proton resonances of proteins are well separated from other resonances due to methylene or methine groups. Accordingly, NOE measurements on methyl proton resonances of proteins are useful for conformation studies. In the present study, NOE enhancements were observed for elucidating the spatial proximity of methyl groups and aromatic rings of cobrotoxin in 'Hz0 solution (Fig. 8).

421

The NOE experiments (Fig.8) on the pair of the y-methyl proton resonance of Thr-a and C-2 (and C-4) proton resonance of His-4 clearly indicate that these protons are in close proximity in the molecule of cobrotoxin (Fig. 10). This is further supported by the pH dependence of the chemical shift of y-methyl proton signal of Thr-a. As shown in Fig. 7, this proton resonance (peak 1) is significantly shifted downfield as the pH is reduced to around 5, approximately corresponding to the pKa value of His-4. The appreciable upfield shift (by about 0.9 ppm) of this y-methyl proton signal from the normal position (about 1.3ppm) is at least in part due to the ring-current effect from the imidazole ring of His-4 residue. However, the least-square treatment of the chemical shifts of Thr-a with the Henderson-Hasselbach equation yielded the pKa value of 5.0 but the deviations of calculated chemical-shift values from the observed ones are appreciable and thus the Hill coefficient is found to be as low as 0.6. This pKa value is significantly lower than the pKa value (5.33) of His-4 as obtained directly from the C-2 proton chemical shift of His-4 (Table 2). Therefore, the least-square treatment of chemical shifts was made with two ionizable groups and the two pKa values were obtained as 5.29 0.03 and 3.91 ? 0.07. The first of these pKa values agrees with the pKa value of His-4, indicating the effect of the protonation of His-4 upon the chemical shift of Thr-a y-methyl protons. The chemical shift of this methyl proton is also affected by a carboxylate group with the pKa value of 3.9. Such a significant pH dependence of chemical shift is also observed for the methyl proton resonance of a threonine residue of neurotoxin I (see Fig. 8 of [lo]). As mentioned previously, the imidazole ring of His-4 of cobrotoxin lies in proximity to the guanidino group of Arg-59. The side-chain group of Arg-59, however, is much more bulky than the side-chain group of the corresponding Val-59 of erabutoxin b. On the basis of the atomic coordinates of erabutoxin b (Petsko, G. A., personal communication), the substitution of Val-59 with Arg-59 in cobrotoxin is expected to displace the imidazole group of His-4 closer to the first loop with three threonine residues. Thus, Thr-a is presumably one of the Thr-13, Thr-14 and Thr-15 residues of cobrotoxin. The y-methyl proton resonances of Leu-I (peaks 4 and 3 in Fig.7) are identified by the observation of the inflections of chemical shifts at pH 9.45 (Table 3) and are confirmed by the observation of NOE enhancements due to the dipolar interaction with His-4 protons. This inflection corresponds to the pKa for the deprotonation of NH: group of the N-terminal residue rather than lysine residues, since the lysine residues of cobrotoxin (Lys-26, Lys-27 and Lys-47) are not expected to lie in the proximity to Leu-1 (Fig. 1). However, this pKa value of Leu-l is appreciably higher

428

than the normal pK, value (zS.O), suggesting the electrostatic interaction of the N-terminal NH; group and a negatively charged carboxylate group, possibly of Glu-21 or Asp-58. These residues of Leu-I, Glu-21 and Asp-58 are homologous to cobrotoxin, neurotoxin I and neurotoxin I1 (Table 1). In fact, the inflection at pH z 9.5 is also observed for the Leu-I 7-methyl proton resonance of neurotoxin I (see Fig. 8 of [lo]). The 7 and &methyl groups of Ile-52 of cobrotoxin are in the proximity to the aromatic ring of Tyr-25 (Fig.10), as evidenced by the observation of NOE enhancements. Similarly, the y and &methyl groups of Ile-50 are found to be in proximity to the indole ring of Trp-29. The methyl proton resonances of He-50 are affected by the ionization of a lysine residue: the doublet signals (peaks 8 and 2) of Ile-50 are shifted upfield as the pH is raised across 10.6 (Fig.7 and Table 3). Similarly the side chain of Val-46 appears to be in the proximity to a lysine residue, since the chemical shift of the 7'-methyl resonance (peak 7) of Val-46 is affected by the pH change around 10.3 (Fig. 7).

Temperature-Dependent Local Conformation Finally for studying the microenvironment of the Thr-a residue, the pH dependence of proton chemical shifts of cobrotoxin was observed at 37" and 5 0 T as well as at 23 'C; the chemical shift of peak 1 due to Thr-a was found to vary appreciably with temperature (Fig.11). The pH dependence of the chemical shift of peak 1 could not be followed down to pH 1.7 at 37 or 5 0 T , because of enhanced denaturation. At 50"C, peak 1 overlapped with peaks 2 and 3 at pH above 7. Nevertheless, the pH dependence of the chemical shift of peak 1 was found to be significantly sensitive to temperature (Fig. 3 1). This observation may well be explained with the temperature-dependent local conformational equilibrium around the His-4 residue at pH above 6. In the pH region below 5, the imidazole ring of His-4 is protonated and accordingly is subject to the electrostatic interaction with the positively charged guanidino group of Arg-59. However, the chemical shift of peak 1 of Thr-a in the proximity of His-4 depends little on temperature. On the other hand, in the pH region above 6, the imidazole ring of His-4 is deprotonated (uncharged) but the chemical shift of peak 1 of Thr-a now depends sensitively on temperature. These observations suggest that, at pH above 6, the absence of the electrostatic interaction between His-4 and Arg-59 allows a temperature-dependent equilibrium among local conformations around His-4 residue. This local-conformation equilibrium is reflected in the significant temperature dependence of the chemical shift of peak 1 of Thr-a. This threonine residue is in the proximity to His-4 and the ring-current

Conformation of Cobrotoxln in Solulion

I

0.8

0.2

L 2

4

6

8

10

PH

Fig. 1 1 . p H dependence o f t h e chemical shijt of Thr-a y-rnethj71pmton resonance ( I t various temperatures

effect from His-4 is expected to be sensitive to variations in the spatial arrangement of the imidazole ring and observed methyl protons. The transitions among local conformations of His-4 will be enhanced at higher temperatures and will affect the proton resonance of Thr-a through the ring-current effect. In fact, the doublet peak 1 due to Thr-a of cobrotoxin becomes broad at pH above 7 and this line broadening is enhanced further at 37' and 50 "C, because of faster chemical exchange among various sites (corresponding to various local conformations around His-4). Thus, the ring-current-shifted methyl proton resonance of Thr-a was useful for elucidating the temperature-dependent local conformational equilibrium around His-4 residue in the first loop of cobrotoxin molecule. CONCLUSION The spatial arrangements of side-chain groups of cobrotoxin, as elucidated by the proton N M R analyses, are schematically shown in Fig. 10. The white arrows indicate the proximity of aromatic ring and methyl groups, as found by the observation of NOE enhancements. These NOE enhancements are also important for confirming the assignments, especially of methyl proton resonances. The black arrows indicate the proximity and resulting effect of the ionization of histidine, lysine, tyrosine, aspartic and glutamic acid residues and N-terminal a-amino group. Striped arrows indicate the effect of the electrostatic interaction of an ionizable group on the pK, value of an ionizable group of the pK, value of another ionizable group. The analyses of interactions such as those shown with arrows are useful for elucidating the micro-

T. Endo, F. Inagaki, K . Hayashi, and T. Miyazawa

429

environments of structurally or functionally invariant residues of neurotoxins. Furthermore in the present study on cobrotoxin, all the methyl proton resonances were completely assigned to amino-acid types and eight of them were identified as due to the y and &methyl groups of individual residues. These methyl groups are located all over this protein molecule and accordingly are useful for studying the conformational changes and mode of binding with other molecules, hopefully including acetylcholine receptor protein.

Since .f is as small as lo-'', the concentrations [Im(CH)+]and [OH-] stay constant and the concentration of the reaction product is finally given as [Im(C3H)'] =f'(kb/kb')[Im(CH)+] {I

-

exp (- khf [OH-] r ) ] .

Accordingly, in the hydrogen -+ tritium exchange reaction, the apparent second-order rate constant kh' is actually for the abstraction of the imidazolium C-2 tritium atom in the backward process. The authors are grateful to Professor G. A. Petsko for the atomic coordinates of erabutoxin b prior to publication and to Professor K. Wuthrich for the preprint of [lo].

APPENDIX This appendix is to point out that, in the hydrogen tritium exchange reaction, the second-order rate constant kh is for the abstraction of the imidazolium C-2 tritium atom by hydroxide ion but not for the abstraction of the imidazolium C-2 hydrogen atom. Accordingly, the kh value for the hydrogen -+ tritium exchange reaction may well be different from the kb value for the hydrogen + deuterium exchange reaction. In the hydrogen -+ deuterium exchange reaction, the sample is dissolved in pure deuterium oxide. Accordingly the forward process is predominant and the second-order rate constant ( k b ) is concerned with the abstraction of the imidazolium C-2 hydrogen atom by 0 2 H ion (the backward process is negligible). On the other hand, in the typical hydrogen -+ tritium exchange experiments [31], the concentration ratio .f'= [3HOH]/2[HOH] is as low as % Accordingly, the forward process is competed by the backward process until final equilibrium is reached.

REFERENCES

--+

-

In the forward process, the imidazolium C-2 hydrogen atom is abstracted by O H - ion, yielding the ylide intermediate. Then, only the small fraction lop1') of ylide reacts with 3HOH and is converted to the imidazolium ion [Im(C3H)'] with the C-2 3H atom, while the majority of the ylide reacts with 3HOH and HOH to yield Im(CH)+ (no net change). Accordingly the forward hydrogen +tritium reaction rate is given by,f. kh [Im(CH)+][OH-]. In the backward process, the C-2 3H atom of Im(C3H)+ is abstracted by an OH- ion, with the second-order rate constant khj. Then the ylide intermediate reacts with H O H and yields Im(CH)+. The backward reaction rate then is given by k v [Im(C3FI)+] [OH-]. The net hydrogen +tritium exchange reaction rate is now given by

vz

d [Im(C3H)']/dt = f k b [Im(CH)+][OH-] - kbf [Im(C3H)+][OH-] .

1. Yang, C. C., Yang, H. J . & Huang, J . S. (1969) Biochim. Biophjs. Acra, 188, 65-77. 2. Yang. C . C., Yang, H. J. Sr Chiu, R. H. C. (1970) Biwhim. Biophjs. Actu, 214, 355-363. 3. Yang, C. C. (1974) Touicon, 12, 1-43. 4. Changeux, J. P., Kasai, M. & Lee, C. Y. (1970) Proc.. N u t / A c ~ d Sci. . U . S . A . 67, 1241 -1247. 5. Ryden, L., Gabel, D. & Eaker. D. (1973) b i r . J . Peptit/e fI"Jt('iH R(7.s. 5 , 261 -273. 6. Arseniev, A. S., Balashova, T. A,, Utkin, Y. N., Tsetlin. V . 1.. Bystrov, V. F., Ivanov, V. T. Sr Ovchinnikov. Yu. A. (1976) Eur. J . Biochem. 71, 595-606. 7. Lauterwein, J., Wuthrich, K.. Schweitz, H., Vincent, J.-P. & Lazdunski, M. (1977) Biochem. Biciphjs. R(T. Conmuri. 76. 1071 -1078. 8. Inagaki, F., Miyazawa, T., Hori, H. & Tamiya. N. (1978) Eur. J . Biochem. 89. 433-442. 9. Bystrov. V. F.. Arseniev. A. S. Sr Gavrilov. Y. D. (1978) J . Mugti. Rrson. 30, 151 184. 10. Lauterwein, J., Lazdunski, M . & Wuthrich. K. (1978) Eur. J . Biochem. 92, 361 - 371. 11. Fung, C. H., Chang, C. C. & Gupta, R. K. (1979) Bioc/7cvni.slry. I # , 457 - 460. 12. Tsetlin, V. I., Arseniev. A. S., Utkin, Y. N.. Gurevich, A. Z.. Senyavina, L. B., Bystrov, V. F., Ivanov, V. T. & Ovchinnikov. Yu. A . (1979) Fur. J . Biochem. 94. 337-346. 13. Low, €3. W., Preston, H. S., Sato, A,, Rosen, L. S., Searl, J . E.. Rudko, A. D. & Richardson, J . S. (1976) Proc. Not/ Acatl. Sci. U.S.A. 73, 2991 -2994. 14. Tsernoglou, D. & Petsko, G . A. (1977) Pmc. Nut/ A w d . S c i . U.S.A. 74,971 -974. 15. Tsernoglou, D., Petsko, G. A. & Hudson. R . A. (1978) Mol. Pharmacol. 14, 710-716. 16. Chang, C. C., Yang, C. C.. Hamaguchi, K.. Nakai. K. & Hayashi, K. (1971) Biochim. Biophy.?. Actu, 236. 164- 173. 17. Marinetti, T. D., Snyder, G . H. & Sykes, B. D. (1976) Biochemistry, 15,4600- 4608. 18. Yang, C. C. (1965) J , Biol. Chem. 240, 1616- 1618. 19. Markley, J. & Kato, I. (1975) Biochrmistr.v. 14, 3234-3237. 20. Richarz, R. & Wuthrich, K. (1978) J . Mugn. Re.son. 30, 147150. 21. Campbell, E. D. & Dobson, C. M. ( I 975) J.C.S. Chem. Commun., 750-751. 22. Moore, G. R., Pitt, R. C. & Williams, R. J. P. (1977) Eur. J . Biochum. 77, 53-60. 23. Tanokura, M., Tasumi, M. & Miyazawa, T. (1976) Biopo/~~mcr.s, 15, 393-401. 24 Tanokura, M., Tasumi, M. s( Miyazawa,T. (1978) Chem. L,ett.. 739 - 742. -

430

T. Endo, F. Inagaki, K. Hayashi, and T. Miyazawa: Conformation of Cobrotoxin in Solution

25. Wuthrich, K. (1976) N M R in Biological Research, Peptides and Proteins, North Holland, Amsterdam. 26. Campbell, I. D., Dobson, C. M., Williams, R. J. P. & Xavier, A. V. (1973) J . Magn. Reson. 11, 172-181. 27. Noggle, J. H. & Schirmer, R. E. (1971) The Nuclear Overhauser Effect, Academic Press, New York. 28. Huang, J. S., Liu, S. S., Ling, K. H., Chang, C. C. & Yang, C. C. (1972) J . Formosan Med. Ass. 71, 383-388.

29. Markley, J. (1975) Biochemistry, 14, 3546- 3554. 30. Jencks, W. P. (1969) Catalysis in Chemistrj. and Enzymology, McGraw Hill Inc., New York. 31. Minamino, N., Matsuo, H. & Narita, K . (1978) in Peptide Chemistry 1977 (Shiba, T., ed.) pp. 85-90, Protein Research Foundation, Osaka. 32. Seto, A., Sato, S. & Tamiya, N. (1970) Biochim. Biophys. Acra, 214,483 - 489.

T. Endo, F. Inagaki, and T. Miyazawa*, Department of Biophysics and Biochemistry, Faculty of Science. University of Tokyo, Hongo, Bunkyo-ku, Tokyo 113, Japan K. Hayashi, Faculty of Pharmaceutical Sciences, Kyoto University, Sakyo-ku, Kyoto 606, Japan

* To whom correspondence should be addressed.