+ All Categories
Home > Documents > Influence of hydrogen bonding on the rotamer distribution of the histidine side chain in peptides:...

Influence of hydrogen bonding on the rotamer distribution of the histidine side chain in peptides:...

Date post: 27-Aug-2016
Category:
Upload: toan-tran
View: 215 times
Download: 2 times
Share this document with a friend
9
Biochimica et Biophysica Acta, 492 (1977) 245-253 © Elsevier/North-Holland Biomedical Press BBA 37667 INFLUENCE OF HYDROGEN BONDING ON THE ROTAMER DISTRIBU- TION OF THE H1ST1D1NE SIDE CHAIN IN PEPTIDES: 1H NMR AND CD STUDIES TOAN TRAN, KARL LINTNER, FLAVIO TOMA and SERGE FERMANDJIAN* Service de Biochirnie, D~partement de Biologic, Centre d'dtudes nucl~aires de Saclay. B.P. No. 2, 91190 Gif-sur- Yvette (France) (Received November 9th, 1976) SUMMARY Both 1H NMR and circular dichroism pH titration studies on histidine, His- Gly, Gly-His and Gly-His-Gly indicate that the side-chain spatial orientation depends strongly on the vicinal charges. The arrangement of the imidazole side-chain (rotamer population) is shown by the histidine fl and fl' and the glycine methylene proton chemical shifts as well as the vicinal XH-1H coupling constants 3Jca_n.~.H. ~'-u. For His-Gly and Gly-His-Gly a good correlation can be found between the ionization of the glycine COOH group and the increase of rotamer III (g-g) which is also visualized by circular dichroism through an enhancement of the ellipticity at 212 rim. In these two peptides a hydrogen bond between the imidazolium and the carboxylate group is supposed to stabilize rotamer III at pH 4-5. INTRODUCTION Our interest in histidine was aroused by the fact that it enters the composition of many natural peptides studied in this laboratory, such as thyroliberin, angio- tensin 11 and corticotropin, in which it plays an important role [I-3]. Although histidine has been studied in many ways [4, 5], recent developments in high field NMR spectroscopy now allow a more detailed analysis of the aliphatic region of the spectrum. Free histidine, and small peptides with a histidine residue in the sequence: L-histidylglycine (His-Gly), glycyl-L-histidine (Gly-His) and glycyl-L-histidylglycine (Gly-His-Gly) were studied. Full analysis of the proton NMR spectra gave, among other data, the vicinal coupling constants between the a and fl, fl' protons of the histidine residue. Internal chemical shifts for the fl and fl' and glycine methylene protons (nonequivalences) were also obtained. As the histidine side chain is a chromophore of its own and gives rise to circular dichroism (CD) signals, it was interesting to study the four compounds mentioned by this technique, too. In these simple model peptides the side chain * To whom correspondence should be addressed.
Transcript
Page 1: Influence of hydrogen bonding on the rotamer distribution of the histidine side chain in peptides: 1H NMR and CD studies

Biochimica et Biophysica Acta, 492 (1977) 245-253 © Elsevier/North-Holland Biomedical Press

BBA 37667

INF LUENC E OF HYDROGEN BONDING ON THE ROTAMER DISTRIBU- TION OF THE H1ST1D1NE SIDE CHAIN IN PEPTIDES: 1H NMR AND CD STUDIES

TOAN TRAN, KARL LINTNER, FLAVIO TOMA and SERGE FERMANDJIAN* Service de Biochirnie, D~partement de Biologic, Centre d'dtudes nucl~aires de Saclay. B.P. No. 2, 91190 Gif-sur- Yvette (France)

(Received November 9th, 1976)

SUMMARY

Both 1H NMR and circular dichroism pH titration studies on histidine, His- Gly, Gly-His and Gly-His-Gly indicate that the side-chain spatial orientation depends strongly on the vicinal charges. The arrangement of the imidazole side-chain (rotamer population) is shown by the histidine fl and fl' and the glycine methylene proton chemical shifts as well as the vicinal XH-1H coupling constants 3Jca_n.~.H. ~'-u. For His-Gly and Gly-His-Gly a good correlation can be found between the ionization of the glycine COOH group and the increase of rotamer III (g-g) which is also visualized by circular dichroism through an enhancement of the ellipticity at 212 rim. In these two peptides a hydrogen bond between the imidazolium and the carboxylate group is supposed to stabilize rotamer III at pH 4-5.

INTRODUCTION

Our interest in histidine was aroused by the fact that it enters the composition of many natural peptides studied in this laboratory, such as thyroliberin, angio- tensin 11 and corticotropin, in which it plays an important role [I-3].

Although histidine has been studied in many ways [4, 5], recent developments in high field NMR spectroscopy now allow a more detailed analysis of the aliphatic region of the spectrum.

Free histidine, and small peptides with a histidine residue in the sequence: L-histidylglycine (His-Gly), glycyl-L-histidine (Gly-His) and glycyl-L-histidylglycine (Gly-His-Gly) were studied. Full analysis of the proton NMR spectra gave, among other data, the vicinal coupling constants between the a and fl, fl' protons of the histidine residue. Internal chemical shifts for the fl and fl' and glycine methylene protons (nonequivalences) were also obtained.

As the histidine side chain is a chromophore of its own and gives rise to circular dichroism (CD) signals, it was interesting to study the four compounds mentioned by this technique, too. In these simple model peptides the side chain

* To whom correspondence should be addressed.

Page 2: Influence of hydrogen bonding on the rotamer distribution of the histidine side chain in peptides: 1H NMR and CD studies

246

chromophore contribution is relatively strong and can be followed on its own as a function of conformational parameters.

EXPERIMENTAL

The compounds used (Sigma Chem. Comp.) were of the highest purity and in the form of the free base. Solutions were prepared in deuterium oxide (Com- missariat /~ l'Energie Atomique, France; isotopic enrichment 99.8~) at concen- trations ranging from 0.10 to 0.15 M, except for Gly-His-Gly which was at a concen- tration close to its isoionic point, 0.03 M. The pH was adjusted with 2HCI or NaOZH solutions (Commissariat ~ l'Energie Atomique, France; isotopic enrichment superior to 97 ~o). None of the measurements was corrected for p2H. Spectra were obtained at 90 MHz on a Bruker HX 90 spectrometer equipped with a deuterium lock or at 250 MHz on a Cameca TSN 250 spectrometer using a homolock on the tertiary butanol signal. Frequency readings were accurate to ± 0.05 Hz. "[he choice of spectrometer depended on the observability of the nonequivalence of the /4-CH2 protons of histidine. In almost all the experiments the proton C-5-H of the imidazole ring was decoupled during observation of the fl-CH2 protons. All spectra were analysed with the LAOCN3 program [6].

Circular dichroism measurements were made on a Jobin Yvon Dichrograph I l l model using fused quartz cells of 0.1 mm path length. Concentrations used were in the order of I mg/ml, pH adjustments were effected with concentrated HCI and N a O H solutions so that the peptide concentration did not change measurably. Results are expressed in molar ellipticity [0].

RESULTS AND DISCUSSION

N M R Fig. 1 shows the spectrum of His-Gly at 250 MHz, obtained in 2H20 at

pH 4.2, as an example. It displays the following characteristics:

Protons Structure Remarks

C-2-H doublet coupled with C-5-H C-5-H unresolved quintuplet ~t-C-H quadruplet /

form an ABX system fl-CH2 multiplet / CH2 (glycine) quadruplet AB system, well separated

The assignments of the fl and/3' protons of histidine in the three rotamers are represented in Table I. They are based on generally accepted arguments, e.g. rotamer l which presents the least steric hindrance is almost always preferred over the two others [5-7, 8].

This finds support in the 3ff13co.13c4 values in histidine and histidine derivatives [9] which allow the determination of the rotamer I fraction, and in studies of phenyl- alanine with a selectively deuterated fl-carbon [10]. All the proton chemical shifts as well as signal aspects and coupling constants vary greatly depending on the pH; thus

Page 3: Influence of hydrogen bonding on the rotamer distribution of the histidine side chain in peptides: 1H NMR and CD studies

247

i + r i ~

• 6;0 ' 6~0' 660 6;0 6+0 ++0 ' +70' ++0 +~0 Hz Hz

j • 5Hz ~ ~ . ~

t,,

CH. C-2-H C--5-H C-a-H ~ C-#-H a

. . . . . . . 7 6 5 ' ~, ' 3 ' ppm / t-SuON

Fig. 1. 250 MHz ~H NMR spectra of L-histidylglycine in 2 H 2 0 at pH 4.2. All parts of the survey spectrum (bottom) are presented expanded. At this pH the AB quadruplet of glycine methylene protons exhibits the greatest nonequivalence (0.226 ppm).

the determination o f chemical shifts and coupling constants for each pH value allows the plotting o f the curves shown in Figs. 2, 3 and 4.

None o f these curves shows monotonous variation. All represent, to a varying extent, the ionization states o f the different groups with privileged experimental points, e.g. the p H values for which the monocat ions are in a pure state; isoionic

Page 4: Influence of hydrogen bonding on the rotamer distribution of the histidine side chain in peptides: 1H NMR and CD studies

248

T A B L E I

R O T A M E R P O P U L A T I O N S OF T H E H I S T I D I N E SIDE C H A I N O B T A I N E D AT pH V A L U E S C O R R E S P O N D I N G T O P U R E I O N I C SPECIES

C o m p o u n d Ionic species " p H " c o co co

N r ~1 ~ "Ha Im Ha H a '

I I I I11

Histidine N H ; ~ l m H + C O O H 0.5 0.38 0.36 0.26 N H ~ l m H + C O O - 3.7 0.40 0.32 0.28 NH3 + Im C O 0 - 7.8 0.53 0.16 0.31 NHz lm C O O - 12.0 0.48 0.22 0.301

His-Gly NH~ l m H + C O O H 1.3 0.49 0.29 0.22 N H ~ I m H ÷ C O 0 - 4.2 0.37 0.15 0.48 N H ~ lm C O 0 - 6.9 0.34 0.35 0.3 I NHz Im C O O - 12.0 0.33 0.39 0.28

Gly-His NH;- l m H + C O O H 1.3 0.52 0.22 , 0.26 NH3 + h n H ÷ C O O - 5.0 0.52 0.22 I 0.26£ NH3 ~ l m C O O - 7.5 0.56 0.18 , 0.26 NHz lm C O 0 - 10.0 0.56 0.16 t 0.28

Gly-His-Gly NH + l m H + C O O H 1.0 0.50 0.33 0.17 i NH~ ImH ÷ C O 0 - 4.9 0.32 0.33 0.35 N H ; Im C O 0 - 8.0 0.52 0.26 0.22' NH2 l m C O O - 11.0 0.55 0.23 0.22 ~

i

+

D

- f l U +

I : 1 z

!

.... y i J ........ +:++ +:J

\ X+

• Hi+ +

" \L t H,S - Gly my - Hi+ ÷' ,~' }~ ,_

, J , ~, i L, i l, ,i 1 s 6

" p W

Fig. 2. Chemical shift nonequivalence curves o f hist idine side chain fl and ~' protons. As an example of an a romat ic amino acid wi thout ionizable func t ions the curve o f phenyla lanine is also presented. Note tha t in the case of His-Gly the ~ da, nonequiva lence is zero at p H 6.4 where the glycine methylene proton 6 nonequivalence is also zero and Jaa -- Jaa" = 6.3 Hz (see Figs. 3 and 4). For such a coupl ing cons tan t the fractions of ro tamers I, II, III are 0.33.

Page 5: Influence of hydrogen bonding on the rotamer distribution of the histidine side chain in peptides: 1H NMR and CD studies

249

0,2~ I

0.20

/

/

/ s÷ /

O.

'~ a05

0.00

i

- o,o~ P i i

-o,0 J i 1 z 3

/ /

J

, - ~+%

\

\

I

I I t

. _o_ .o=. - . o -

, i i s

\ \ \ - -

\ I ..... " I

" p H "

i

+ His - Gly

o Gly - His

• ~ - His - G~y

o GLy - His - Gly

i I i

~,O-- 0-O --IL. .e

i

i - i I 11 12

Fig. 3. Chemical shift nonequivalence curves of glycine methylene protons.

points, where neutral zwitterions constitute the majority if not all the molecules. Comparison of the nonequivalence curves of the glycine methylene protons

(Fig. 3) with those of the fl and fl' protons of the histidine residue (Fig. 2) reveals a more or less pronounced analogy between their respective behaviour depending on the position of the glycine residue in the peptide (N- or C-terminal). Primarily for His-Gly, the two curves translate the course of the imidazole and carboxyl ionizations

7.7

L3 . . . . . .

03

0.1

0

~.÷

s7 ! \ N

o

o't J a P ' i

' \

\ J~p

/,3 ~.,

1 2 3

..~,/\,,

/ /

/

~pH"

+

o

. +

..o

I + His - G l y

+ -+'+

ooooi

Fig. 4. Jaa" and Jr~a in His-Gly as a function of pH.

Page 6: Influence of hydrogen bonding on the rotamer distribution of the histidine side chain in peptides: 1H NMR and CD studies

250

dramatically. Simultaneously the values of Ja~ and Ja~" vary significantly (Fig. 4). When the histidine residue is at the C-terminal, the nonequivalence varies little, even if the curves do show the titration of the ionizable groups of the molecule. Finally, a histidine residue in the middle of the peptide is submitted to the effects of glycine at the N- and C-ends; the resulting curves have several of the characteristics and exhibit some of the trends of the curves of both His-Gly and Gly-His.

From the coupling constants Ja~ and Ja~', we determined the rotamer popu- lations for histidine at several of these selected points (Table 1). Among the existing methods, we chose the one of Pachler [7], taking Jg - 3.25 Hz and Jt - 12.40 Hz, which are the values resulting from the calibration curve by Kopple et al. [11].

The peptides whose rotamer distributions show the highest pH sensitivity are His-Gly and Gly-His-Gly (in both, a glycine residue is found at the C-terminal) con- cerning the ionization of the carboxyl and the imidazole groups. On the other hand, only small changes are observed during the titration of free histidine and the peptide Gly-His.

Several kinds of interactions such as hydrogen bonding, van der Waal's forces and salt bridges are generally considered to be present in amino acids and peptides. In addition, although neglected in this study, the role of water molecules that may solvate and bridge the different groups, is probably important. The influence of hydrogen bonds seems, nevertheless, to emerge clearly, as follows from the close observation of the variations of rotamer populations versus pH. For His-Gly the rotamer population changes are particularly evident during the deprotonation of the COOH- and the imidazolium groups. If we assume the two groups, C O 0 - and imidazolium to interact at pH 4.2 (Table 1), then the preponderance of the retainer Ill over the other two could be explained. In the proposed interaction, the CH2COO- part of glycine is "frozen"; this finds support in the nonequivalence of the glycine methylene protons (0.266 ppm at pH 4.2), which is very similar to that found in cyclo-Gly-His [12]. On the other hand, 13C NM R measurements of His-Gly have shown that with increasing pH it is almost exclusively the imidazole N~ (bl) site which is deprotonated, rather than N~ [13]. Since for sterical reasons it can be assumed that it is the N.~-H which is hydrogen bonded to the carboxylate, the important loss of rotamer IlI with increasing pH above 4.5 would be a logical consequence.

Whereas in His-Gly we found a strong parallelism of the h-nonequivalence curves of the histidine /3, /3' and glycine methylene protons, these curves vary in directions opposed to each other in Gly-His-Gly. Notably at pH 5 the nonequivalence of the/~,/3' protons is at a minimum; the nonequivalence of the C-terminal glycine protons, however, mimicking well the one found in His-Gly, is at a maximum. Still, as in the case of the dipeptide, one notes an increase in rotamer llI between p H I and 5, although at pH 4 and 5 the rotamer distribution which in the dipeptide was very uneven, is almost statistical in the tripeptide: 0.37, 0.15, 0.48 and 0.32, 0.33, 0.35 respectively (Table l). Consequently whereas the nonequivalence of the methylene protons of the C-terminal glycine can be explained here, too, by the interaction of its carboxylate group with the imidazolium group in the rotamer I l l state, the decrease of nonequivalence of the/3 and/3' protons is explained by the statistical distribution of the three rotamers.

For His-Gly at the isoionic point (pH 6.9) all rotamers are about equally populated which contrasts with the findings for histidine and the other two peptides,

Page 7: Influence of hydrogen bonding on the rotamer distribution of the histidine side chain in peptides: 1H NMR and CD studies

251

this shows that steric factors are compensated by interacting forces in the rotamers, ~.s found in Gly-His-Gly at pH 5. During the deprotonation of the NH + group the distribution of the rotamer population changes little in all compounds examined, which indicates that the NH + group can be substituted by NH2 without affecting the interactions with the side chain.

In summary we find that (a) nonequivalence of C-terminal glycine methylene protons is greatest when there is an extra stabilization of the side chain in the rotamer llI conformation as shown by His-Gly (pH 4.2) and to a lesser degree by Gly-His-Gly (pH 5). It becomes minimal when the fractions of rotamer populations tend to be equal (His-Gly at pH 6.5-6.9).

(b) Nonequivalence of N-terminal glycine methylene protons is always weak and remains almost constant for Gly-His, where practically no change of rotamer population is observed during titration. The situation is about the same for Gly- His-Gly.

(c) Nonequivalence of the histidine fl and/3' protons depends on whether or not a rotamer is preferred over the other two. This is the case for Gly-His for which the unvarying nonequivalence corresponds strongly to the almost total pH inde- pendence of the preferred rotamer I. For His-Gly the nonequivalence reaches its maximum when rotamer III becomes preponderant (pH 4.2); when the fractions of rotamer populations tend to be equal (His-Gly at pH 6.5-7 and Gly-His-Gly at pH 5) the nonequivalence is minimal.

We can therefore conclude that nonequivalence of the glycine methylene as well as the histidine/3 and/3' protons can be attributed largely to the restriction of rotation around the bonds of the molecule [14]. Anisotropy of the peptide bond, orientation of the carboxylic or carboxylate group and anisotropy of the imidazole or the imidazolium ring are all factors influencing the observed nonequivalence. In a molecular model of His-Gly showing the histidine side chain in the rotamer III conformation several of the observed effects can be visualized: the magnetically anisotropic imidazolium ring, hydrogen bonded to the carboxylate group, contributes to the nonequivalence of the vicinal fl and /3' and, to a lesser degree, the glycine methylene protons. Each of them is differently influenced by the carbonyl group of the peptide linkage and the carboxylate group. The deprotonation of the imidazolium group induces the rupture of the carboxylate-imidazolium hydrogen bond, thus leading to concerted changes in the molecule.

Circular dichroism To the NMR studies of these peptides we have associated circular dichroism

measurements of their aqueous solutions as a function of pH. These studies allow, by the parallelism observed to the NMR results, the corroboration of the conclusions that have been drawn.

Circular dichroism spectra of histidine and its derivatives consist of the peptide chromophore contributions and the imidazole side-chain contribution. The latter has only recently been studied in relation to its conformation, and efforts were concen- trated on cyclo-dipeptides [15, 16]. Theoretical interpretation of the side chain con- tribution has been restricted to the 207 nm (lowest energy ~t--~*) transition. In the study of His, Gly-His, His-Gly and Gly-His-Gly by CD we, too, mainly followed this wavelength region.

Page 8: Influence of hydrogen bonding on the rotamer distribution of the histidine side chain in peptides: 1H NMR and CD studies

252

[0]2,~ 10 -~ I I I

-10

I I I L i I

/ o . ~ . , . o - - • - -

i . . . . "\'x /

"\.. .... i12 -.

,. ..... Y \', . ! \+÷ \ i + / ~.I)~D'--I)--()

o -~-~-~,\~ /" - -

!

\ i _,_ \ i -

\ i \ i \ / \ /

- \,._ .... i I I I t I I t t i 1 2 3 t. 6 6 7 8 9

pH

Fig. 5. Ti t ra t ion curves: Ellipticities 0212 . . . . plotted as a funct ion of pH. Histidine (F7 -iT3), Gly-His (O---O), His-Gly ( L [ ), Gly-His-Gly (tD--ID).

At pH 4.8 (NH~, ImH +, COO-) the side chain contribution of the imidazole ring appears positive in the free amino acid and in Gly-His, whereas it is strongly negative in His-Gly and almost annihilated in the tripeptide. Fig. 5 shows the titration curves of these four compounds, obtained at 212 nm. We notice that all compounds translate the titration of the carboxyl and the imidazole group through spectral changes, whereas only the NH~- group attached to histidine (free or in His-Gly) influences the spectrum by its deprotonation. More important, however, are the relative differences observed from one titration curve to another and the spectral changes they reflect. Indeed, the positive band of histidine and of Gly-His undergoes only small intensity changes during variations of pH, much as the already very weak band of Gly-His-Gly at 212 nm decreases in intensity to almost zero at either pH 1 or pH 8. The spectacular spectral modifications during the titration of His-Gly present a strong contrast to the former 3 compounds. Although the imidazole chromophore is now removed from the C-terminal carboxyl group by an optically inactive glycine residue, the deprotonation of this group between pH 1 and pH 4 changes the positive band at 212 nm to a strong, negative one at pH 4-5. In the same way, this band changes back to positive during the imidazole titration, before under- going the influence of the transition NH~ . . . . . . NH2. It is known that rotational power increases with increased asymmetry of a chromophore, induced, for instance,

Page 9: Influence of hydrogen bonding on the rotamer distribution of the histidine side chain in peptides: 1H NMR and CD studies

253

by a decrease o f ro ta t iona l f reedom a round dihedral angles. F r o m the intensit ies o f the C D bands at 212 nm where we expect the main cont r ibu t ion o f the imidazole ch romophore , we find the s trongest signal for His -Gly at p H 4.5.

Thus the spectral changes are in full agreement with the N M R results showing that a change in p ro tona t i on state o f ei ther o f the groups l m H + -+ Im, C O O - -7 C O O H destroys the p roposed interact ion and leads to radical changes in the ro tamer popu- lat ions.

CONCLUSION

Thus, even if ro t amer popula t ions of the hist idine side chain are not the only factors influencing the intensit ies o f the C D band at 212 nm, the var ia t ions of N M R paramete r s such as coupl ing constants , nonequivalence and ampl i tudes o f chemical shift var ia t ions can be corre la ted well with the changes in the CD spectra as reflected in Fig. 5. I t appears clearly tha t the restr ict ion of ro ta t ion a round the bonds o f this side chain, which is magnet ical ly an iso t ropic as well as an ul t raviole t ch romophore , const i tutes the main fac tor influencing nonequivalence and opt ical activity. In these experiments , unfor tunate ly , we did not encounter pept ides whose side chain is s tabil ized in one single confo rma t ion ; this would have al lowed us to bet ter correlate our results and perhaps even to quant i fy them. Studies, however, on several pept ides conta in ing a romat ic side chains with restr icted ro ta t ion a round the bonds , are in progress.

ACKNOWLEDGMENT

We are grateful to Dr. J. Thi6ry for valuable discussions.

REFERENCES

1 Haar, W., Fermandjian, S., Vicar, J., Blaha, K. and Fromageot, P. (1975) Prcc. l~'atl. Acad. Sci. U.S. 72, 4948-4952

2 Fermandjian, S., Lintner, K., Haar, W., Fromageot, P., Khosla, M. C., SnTeby, R. R. and Bumpus, F. M. (1976) in: Proc. 14 Eur. Peptide Symp. (Loffet, A., ed.), Editions de l'Universit6 de Bruxelles, Belgium

3 Greff, D., Toma, F., Fermandjian, S., L6w, M. and Kisfaludy, L. (1976) Biochim. Biophys. Acta 439, 219-231

4 McDonald, C. C. and Phillips, W. D. (1963) J. Am. Chem. Soc. 85, 3736-3742 5 Weinkam, R. J. and Jorgensen, E. C. (1973) J. Am. Chem. Soc. 95, 6084-6090 6 Bothner-By, A. A. and Castellano, S. M. (1968) in: Computer Programs for Chemistry (DeTar,

D. F., ed.), Vol. 1, pp. 10-53, W. A. Benjamin, New York 7 Pachler, K. G. R. (1964) Spectrochim. Acta 20, 581-587 8 Feeney, J. (1975) Proc. R. Soc. London A 345, 61-72 9 Fermandjian, S. (1975) in: Abstracts of Soviet-French Symposium on the Physical Chemistry of

Proteins and Peptides, Pushchino-on-Oka, Sept. 9-12 10 Kainosho, M. and Ajisaka, K. (1975) J. Am. Chem. Soc. 97, 5630-5631 11 Kopple, K. D., Wiley, G. R. and Tauke, R. (1973) Biopolymers 12, 627-636 12 Kopple, K. D. and Ohnishi, M. (1969) J. Am. Chem. Soc. 91,962-970 13 Reynolds, W. F., Peat, I. R., Freedman, M, H. and Lyerla, J. R. Jr. (1973) J. Am. Chem. Soc. 95,

328-331 14 Nakamura, A. and Jardetzky, O. (1967) Proc. Natl. Acad. Sci. U.S. 58, 2212-2219 15 Grebow, P. E. and Hooker, Jr., T. M. (1975) Biopolymers 14, 871-881 16 Grebow, P. E. and Hooker, Jr., T. M. (1975) Biopolymers 14, 1863-1883


Recommended