benzoic acid (HBABA) - American Chemical Society

The EHT programs are available for use on IBM. 7094 and IBM 360/50 systems. The modified CXDO programs have been used oil IBM 360/50 (convergence...
3 downloads 0 Views 887KB Size
September 1969

BIXDISG OF HBABA ASU CI’JIL’A

to the effective right-handed a-helical content of the peptide b a ~ k b o n e . ~ ~Such ~ O measurements, referred to as intrinsic rotatory dispersions,“ are extremely important when studying the conformational structure of the peptide, but provide no evidence concerning differences in amino acid sequence and in the peripheral arrangement of functional groups about the polypeptide backbone which are responsible for the binding of small molecules and ions. In order to probe into the species differences in arrangement of binding sites in serum albumin, extrinsic rotatory dispersionsg of albumins complexed with HBABA were investigated. Extrinsic Cotton effects refer to the rotatory dispersion of the peptide in combination with small molecules and ions, z.e., dyes, inhibitors, substrate analogs, metal ions, etc. The Cotton effects have been called extrinsic to distinguish them from the intrinsic ORD. H B h B h was chosen for these studies because (1) it has a well-defined affinity for serum albumin.12( 2 ) the reaction is reversible and competitive inhibition of the absorbance at 473480 mp (and therefore the HBABAprotein reaction) has been studied with many acidic drugs,13 and (3) because of the unique increase in absorbance observed a t 47.5 mp when various phenoxyacetic acids and indomethacin are added to a phosphate buffer solution of RS-4 and HBABA. Experimental Section All albumin preparations (fraction V) were obtained from Pentex, Inc., Kankakee, Ill., and used without further modification; fraction 1’preparations for bovine, guinea pig, rat, human, dog, sheep, rabbit, pig, and horse were studied. All of these preparations were described as >95% pure by electrophoresis a t pH 7.0. The rat serum albumin was purchased a t many different times throughout this study and a sample was also obtained from N a n n Laboratories, New York, N. Y. All rat serum albumin preparations afforded the same results. Crystallized samples of human, bovine, and rabbit serum albumin were also purchased from Pentex, Inc. Crystallized samples are described as 100% piire by electrophoresis a t pH 5.1 and gave the same results as their corresponding fraction 5’ preparations. A defatted sample of RSA was prepared according to G00dman.I~ This sample gave the same resrilts as the original fraction V preparation. Purified HBABA, mp 205-206”, was purchased from Sigma Chemical Co., St.Loids, N o . , Dajac Laboratories, Philadelphia, Pa., and The British I>rag Houses, Poole, Dorset, England, and was used wit,hout further purification. All HBABA samples produced the same general effects with RSA. ol-(4-Chlorophenoxy)-ol-methylpropionic acid (CPAIPA) was synthesized from p-chlorophenoL’5 Equilibrium Dialysis Experiments.-The procedure used to study the interaction of HBABA with serum albumin in the presence and absence of CPlZPA is similar to the method reported by Klotz and coworkers.16 Cellophane bags were made (9) E . R. Blout, J. P . Carver, a n d E. Schechter, “Optical Rotatory Dispersion and Circular Dichroism in Organic Chemistry,” G. Snatake. Ed., Sadler Research Laboratories, Ino., Philadelphia, Pa., 1967, Chapter 14. (10) (a) J. T. Yang, “Poly-a-Amino Acids: Protein Models for Conformational Studies,” G. D. Fasman, Ed., Marcel Dekker, Inc.. Xew York. K. Y., 1967, pp 239-291; (b) A. 3. Adler, R. Hoving, J. Potter, M. Wells, and G. D. Fasman, J. Amer. Chem. Soc., 90, 4736 (1968). (11) J. A. Schellman a n d C. Schellman, Proteins, 2, 90 (1964). (12) ( a ) D. D. Rutstein, E. F. Ingenito, JV. E. Reynolds, a n d J. 11.Burke, .I. Clin. I n r e s t . , 33, 211 (1954); (b) R. AI. Rosenberg and I. AI. Klotz, “-1 Laboratory Manual of Analytical Methods of Protein Chemistry,” P. Alexander and R. J. Block, Ed., Vol. 2, Pergamon Press, New York, 2;. Y . , 1960, PP 160-161; (c) -1.T. Ness, H. C. Dickerson, a n d J. V. Pastewka, Clin. Cham. A c t a , l a , 532 (1965). (13) (a) I. Aloriguchi, S. Wada, and T. Nishieawa, Chem. Phnrm. Bull. (Tokyo), 16, 601 (1968); (11) I. Xoriguclii, S.Wada, a n d H. Sano, ibid., 16, 592. 597 (1Y68). (14) 1). Y. Goodman, Science, l%S, 1296 (195i). (15) D. T. K i t i a k . T. C.-L. H o , R. E. Hackney, and TT-. E . Connor, J . J f e d . Chem.. 11, 1086 (1968). (16) I. RI. Klotz, F. 11. Walker, a n d R . B. Pivan, J . -4mer. Chem. Soc., 68, 1186 (1946).

TO

SERUM ALBUMIX

755

from 1-em diameter dialysis tubing (Union Carbide Corp., Chicago, Ill.) and were soaked in distilled H20 for 1 day prior to i~se. The bags were filled with exactly 4.0 ml of 1.0 X JI (assuming a molecular weight of 6i,000 for all albumins) serum albumin in 0.1 JI phosphat,e buffer, p H 7.4. The filled bag was immersed in a test tube of convenient diameter containing 4.0 ml of a solution of HBABA or HBABA pliis CPnIPA in the same buffer. Initial concentrations of HBABA outside the bag ranged from 1.2 x 10-5 to 2.0 x 10-3 M. CPMPA concentrations of 5.0 X 10-4 and 4.0 x 10-3 JI were employed. The test tubes were covered and kept, a t a temperature of 4”for a period of not less than 7 2 hr, an interval sufficient to attain equilibriiim.16 ,411 compounds were stable under t,hese conditions.6 The tubes were shaken periodically, the dialysis bags were removed, and the outside solutions were analyzed for HBABA content at 350 nip ( e 19,125) using a Beckman Model DU spectrophotometer. CPL\IP A shows no absorption of visible light at 350 mp. Initially, several controls were also rim in which only phosphate buffer and no albumin was placed inside the bag; other conditions remained the same. It was found that within an accuracy of about 5%, all of the HBABA could be accounted for. Since the experimental error is larger than this value, controls were n o t run in subsequent experiments. Knowing the amount of HBABA remaining outside the bag after equilibrium, the amount of dye boimd t o the albumin could be calculated. The Donnan effect on the ratio of monovalent ions inside and outside the bag has been calculated by AlcLean and Hastings‘? where the solution contained 1.2Yc total protein. In this calculation the conceiit,ration outaide the bag would be 0.5:; higher than inside. Since the concentration of albumin used in this study was even less (0.7%), any Donnan effect is iiegligible. The binding of HBABA t.o albumin was not corrected for protein volume because using the specific volume of 0.75 for serum proteins given by Svedberg and Sjogren’8 would mean that, at a concentration of 0.7% serum albumin, a maximum error of only 0.5% should result. The binding of HBABA4 in these studies is far greater than would be accounted for by protein volume alone. Spectrophotometric Methods.-In addition to the eqiiilibrium dialysis measurements binding of HBABA and CPAIP.4 wa:, studied spectrophotometrically in the region b e h e e n 330 and 600 mp using BSA and RSA for comparative purposes. Essentially following the method of Klotz, et al.,’* the e of HBABA bound to B 9 4 and RSA was determined by holding the HBABA concentration (2.4 X 10-5 111) constant and varying the concentration of albumins up to 5 . 2 X 10+ 111. Then, the albumin concentration (either 4.5 or 5.0 X 10-5 M ) and HBABA concentration (either 3.0 or 4.8 X 10-5 X )was held constant and the CPMPA concentration was varied from 1.3 X 10-5 to 18.0 X l o T 3M. The absorbance was determined at ambient temperature in a 1-em cell as a function of waveleiigt,h using the Carp Model 15 spectrophotometer. ORD Measurements.-The molar rotation as a fiinctioii of wavelength for various albumins in the presence or absence of HBABA or CPAlPA4or both was calculated from ORD spectra in the region between 350 and 650 mp using the Diirrum-Jasco ORD-CD/UV5 spectropolarimeter with a cell of I-em light path. All measurements were made at ambient temperatures (29’) with concentrations of 5.0 x 10-5 31 albumin, 3.0 x 10-6 JI HBABA, and 5.0 X M CPJIPA in 0.1 J1 sodium phosphate buffer, pH 7.4.

Results Equilibrium Dialysis.-The interactions observed between HBABA and bovine, human, rat, rabbit, and guinea pig serum albumins are typified by Figures 1 and 2 for bovine and rat serum albumins, respectively. The relationships between the moles of HBABA bound per mole of albumin and free HBABA concentration in the presence and absence of CPAIPA are plotted. Through use of a Scatchard plot,20z21applied to these (17) F. C. Mclean a n d A. B. Hastings. J . B i d . Chem., 108, 285 (1935). (18) T. Svedberg a n d I3. Sjogren, J . Amer. Chem. Soc., 60, 3318 (1928). (19) I. M . Klotz, iiiid., 68, 22YY (1946). (20) G. Scatchard, A n n . X. Y . A c n d . S c i . , 51, 660 (194Y). (21) I n the Scatchard treatment ;/A is plotted as a function of fi where i; is the number of moles of HBhB.1 per mole of albumin and d is the concentration of free H B h B h in equilibrium \vith hound dye. In the “Lang?a. 1/A is plotted where 9 and d are defined similarly. muir” treatment l/;

-{Ai i

BOVINE SERUM ALBUMIN

P

RAT SERUM ALBUMIN 2 7 -

z

3

m6 -1 U

0

HBABA ALONE

w 5 -I

4

HBABA

WITH

2 5 X t 0 - 4 M CPMPA

v

HBABA

WITH

2

0

z

0 x 1 6 M~

CPMPA

\ 4

U %3

m I

22 d

I, 0

I .o

0 5

5.0

IO

50

(LOG) CONCENTRATION FREE HBABA, M X IO E’1giirc

2.

’I’he Ielatioii Ixtweeu the

1110les

of HBABA bound per mole uf 1W.i n i i c l free IJBABA i ~ i i ugl i i i 1 i t ) i i i i i i i ( l i d \ -1preaeiice or a h w i r e of (’P1IP \.

111

the

BISDING OF HBABA AND CPMPA TO SERUM ALBUMIK

September 1969

TABLE I A COMPARISOK OF REPRESENTATIVE STRONGER BINDINGCONSTANTS A N D THE CORRESPONDING MAXIMVM MOLESBOUNDOF HBABA TO VARIOLXSERUMALBUMINS IN THE PRESENCE AND ABSENCEOF CPMPA .is CALCULATED STATISTICALLY FROM THE REGRESSION O F 115 OK 1j-A TVHERE 5 A S D 9 ARE OBTAINED FROM EQUILIBRIUM DIALYSIS EXPERIMENTS~ Regression slope & std errorb x 105

Representative binding constantC + max errord

757

9

AI ax no. of moles of H boun of albumin + max errord

System x 10-4 Bovine albumin X o CPbIPA 2 . 6 =t0 . 2 1 . 6 & 0.1 2.4 & 0.7 2 . 2 f 1 . 1 1 . 2 i0 . 7 Wit.h CPblPA 3 . 8 jz 1 . 6 Human albumin 2 . 4 i 0 . 5 2 . 2 i0 . 8 No CPRIPA 1.8f0.3 llat albumin 9.82~ 1 . 1 1 . 6 i0.6 KO CPMPA 0.63 f0.12 7 . 8 f8 . 1 1 . 4 f0 . 7 With C P l I P A 4 0 . 9 1 0 . 3 3 Rabbit albumin 17 f 11 1.2 + 1.0 No CPMPA 0.49 f 0.24 Guinea pig albumin X o CPRIPA 0 . 6 0 i 0 . 2 4 14 i 7 1.2 f0.7 a 8 = number of moles of HBABA bound/”ole of albumin and d = molar concentration of free HBABA. The regression analysis seems more appropriate to t,he Langmuir treatment, of binding data since regression assumes that error is due to only the error in the dependent variable. b Calculated a t a probThe binding constant tabulated here ability level of 0.05. reflects the t,endency of HBABA to bind to the stronger of two sites in serum albumin. The real binding constant would probably be slightly different from this number, but the relative strengths observed between the various albumins as reflected in the values given in this table should parallel the relative binding strengths between albumins using real binding constants. Primarily because of the small volumes (4.0 ml) used in these experiments (necessitated by the expense of some of the albumins) the resultant error does not permit a precise treatment of data according to Scat,chard to get real binding constants. d The error tabulated here was determined by calculating the representative binding constant or maximum number of moles of HBABAimole of albumin at maximum and minimum values of t,he regression slope. The average error between these two calculated values is given here. I n the systems bovine and rat serum albiimin (both in the presence of CPRIPA) the larger value of the regression slope led to a value for the maximum number of moles of HBABA bound/mole of albumin that were small negative numbers. Hence, in these two systems, the maximum number of moles of HB.4B.4 bound was calculated from the minimum value for the regression slope, and the difference between this calculated value and the average number (1.21 or 1.41) is reported as the error.

constant for RSA or BSA could be detected. However, an apparent decrease in the maximum moles of HBABA which could be bound to the stronger binding site was observed for BSA. For RSA no such decrease was observed a t this coricentration of CPUPA. When considering the effect of CPlIPA on the actual number of moles of HBABA bound to bovine arid rat serum albumin (rather than focusing 011 the maximum number that can be bound) it appears that CPMPA decreases the amount of HBABA bound a t both concenand 2.*5 X M for BSA, but trat,ioiis of 2.0 X that CI’.\IPA decreases HBABA bound to RSA only at its higher coiicetitratioii (Figures 1 atid 3, respectively). To btudy this relat ioriship further a considerably more sensitive spectrophotometric analysis was used. Spectrophotometric Analysis.-Comparative spectrophotometric analysis for BSA and RSA complexed

WAVELENGTH, nm.

Figure 3.--Absorbance as a function of wavelengths for HBABA (4.8 X 10-5 .TI) and BSrl (4.5 X 10-5 M) in the pres31 (- - -), 1.8 X lo-‘ Jf ( ’ ’ . ), and 1.8 X ence of 1.5 X Af (-) CPMPA in 0.1 Mphosphate buffer, p H 7.4.

with HBABA \$-as studied with and without added CPIIPA in the region 330-600 mp. The spectrum for BSA-HBABA is shown in Figure 3. Addition of CPllIPA causes a decrease in the absorbance at 477 mp with a proportional increase in the absorbance a t 345 mp. Since the absorption a t shorter wavelength is directly proportional to the amount of bound HBABA, the same number of moles of HBABA bound per mole of protein may be calculated through use of either eq 1 or 2.19

In thebe equations +;45 = fraction of HBABA free, = the observed absorbance divided by the total concentration of HBABA added, = extinction coefficient at 345 mb for HB-4B-4 in 0.1 19 sodium phosphate buffer, pH 7.4, and e2344 = extinction Coefficient for HBABA bound to protein at 345 mp. The latter constant was obtained from a study of the apparent extinction coefficient as a function of increasing albumin to HBABA concentration ratio (Figure 4). Analogously, +i77 = fraction of HBABA bound arid,;;:E and are similar to the definitions for e345 except that the absorabnce was determined a t 377 mp. I n this case, however, was determined from the maximum increase in absorbance due to HBABA bound (Figure 4). The results are shown in Table I1 and substantiate the predictability of reversible competition of CPAIPA for HBABA bound to the stronger site on BS-4. I n contrast to the excellent mathematical correlation obtained for BSA and the competitive binding of CPMPA and HBABA, an niiomalous effect was observed in a similar study with RSA. This is illustrated with Figures 5 and 6. At lower C P J I P a i coticeiitrations (0.15 X arid 1.8 X 11) the 477-mp peak increased and the Am,, at the shorter wavelength decreased in absorbance with increasing CPMPA coIiceiitration (Figure 5 ) . Paralleling the latter decrease iii 345

tapp

R A T SERUM ALBUMIN

'

0

x

BOVINE SERUM ALBUMIN m0

X

l3-

i

k 2

-

a

-

t-

i

11-

v)

m

-

s

9-

a

w

il

i! 4

1

'

1 2

1 4

1 6

1 8

1 10

1

1

12

[ALBUMIN]

I 14

/

" 16

' 18

l 20

22

[HBABA

Figiiic 4. The appaieiit absorptivity for HBABA az a fritiction of B8A 11)IIBABA ratio at wavelengths of 345 (0)arid 4 7 i tnp ( 0 ) .

4-

:ih*orbaiice a hypmchromic shift from 3% to 34h ink W : I ~ observed. Such :i shift indicates a decrease in aiiioriich rewiance coiitribution; conversely, the 355-mp peah scem,s to be a reflection of an interaction of RSA with the p-phenolic OH group of HBA4BAin which the OH ijrotoii i:, partly attracted to a binding site or1 IISA. The result is increased resoiiaiice contribution of the :iriioiiic form. This is analogous to the explanation for 1 he hithochromic shift observed when increasing the pH of :L solution of HBABA. In the absence of IiSA, ('1'1W-i has iio effect oii the HBABA absorption spec*It higher CPJII'A concentratioits (l.S X lo-' .I[) the 47i-mk peak decreased iri absorX h t i ~ i c caiid ~ the , , ,A at 350 nik increased as in the case wit h BSAi (1;igure 6). Howevcr, w e t i :it higher C'P3[l'L\ concentrations, i i o ni:ithem:iticiil correlaiiorih corre-poiitliiig t o oq 1 :iud 2 werc~applic~ablcwith RSAZwI.icrc~ value. of ,di" ~ I K cI 42 i 7 n(wcictc>rmiiie(it o i)c 14,200 aiid i0'10, rcy)c(*tivcl\ . Tlicsc cliffcwiiws iii tl:it:t l w \\(YTI ~ Bki :tiid IiSA n i : ~ 1 ) ~iiirther : i i i a l > x c d with the i i h ~of Figulc 7 . \\'ith 111-

I

RAT SERUM ALBUMIN

0

I

7 L

X

w

6 -

V

z

4

m

x

5 -

'"c;

/

RAT SERUM A L B U M I N

I I :

0 4 -

CA

AT 477 mp

m 3 2 -

I I

l

l

1 1 1 1 1 1 1 1 1 I 1 1 1 1 1 1 l I I I I 1 350 400 450 500 550

/

1

1

I

I

I

T

I 20

creasing concentration of CPL\IPAthere is an initial decrease in and then an increase which asymtoticallg approaches t':" (19,125). Alteriiatively, increasing concentrations of CI'lIPLi causes it11 initial iricreasc in E:: followed by a decrease, which, at a concentration of CPA\\II'A where 6;'' is approached, = 4.500. Actually, = 1300; this suggests that HBABA is interacting with IiSA even at high CPAII'A concentrat ions. ORD Studies.-Although there are small diff erenceh in molecular neight3" for serum albumins (fraction V) derived from different species, within the accuracy of our instrument, TI e observed the same negative anomalous optical rotatory dispersion (ORD) for all albumins at a concentration of 0.004y0 u i t h the same molar rotation as those reported for bovine and human serum albumin. All albumins showed the same intensity trough a t 233 mp. Addition of 1.33 X AI HBABA andlor 1.33 X 11 C P l I P h yielded no detectable change in the molar rotation at any wavelength. At higher protein concentration (5.0 X 10-5 Jl) only the region of the negative Cotton effect between 350 and 650 mp may be studied. For bovine, human, and sheep serum albumin, addition of HBABX (5.0 X 10-5 J I ) gave no change in the molar rotation in this region of the ORD spectrum (Figure S). For dog, horse, and pig

la

-80

400450

7 500 550 460460 500 55 0

Figure 8.--l\lolar rotatioil as a function of wavelength a t with or without 5.0 x albumin concentrations of 5.0 X 10-6 ~11 X HBABA in 0.1 M phosphate buffer, p H 7.4. 1, BSA with or without HBABA; 2, human SA with or without HBABA; 3, dog SA, (A) without HBABA, (B) with HBABA; 4, sheep SA with or without HBABA; 5, horse SA, (A) without HBABA, ( b ) with HBABS; 6 , pig SA, (A) without HBABA, (B) with HBABA.

serum albumins small differences (curves labeled B in Figure S) could be detected at these higher concentrations. However, with rat (Figure 9), rabbit (Figure lo), and guinea pig (Figure 11) serum albumins, a somewhat stronger induced anomalous optical rotatory dispersion (curves labeled C in the respective figures) was observed. This effect seems to parallel the equilibrium constant and therefore the free energy of binding; i.e., bovine and human serum albumins with small equi-

o

-20

nI

2 X

n .8

u

-140-

librium constants show no change in their ORD spectra upon addition of 5.0 X lop5 AI HBABA, while rat, rabbit, and guinea pig serum albumins with larger equilibrium constants afford similar ERD curves. With rabbit serum albumin and to a lesser extent with guinea pig serum albumin, addition of 5.0 X M CPJIPA blocks the ERD effect (curves B in Figures 10 and 11, respectively). This competitive block parallels the decrease in absorbance at 477 m p observed spectrophotometrically when CPlIPA is added to a solution of these albumins and HBABA. However, when CPJIPA is added to a solution of RSA and HBABA the induced anomalous ORD is not blocked, but rather shifts 26 mp to a higher wavelength (curve B in Figure 9). I n the absence of HBABA, CPlIPA has no effect on the ORD curve for RSA a t this concentration.

Discussion The spectrophotometric studies made at constant HBABA and BSA concentrations with varying CPAIPA indicate that with this albumin CPNPA competes with HBABA for an interaction site at all concentrations of CPXPA. From equilibrium dialysis studies the site involved a t the concentrations of HBABA used is most likely the stronger site ( K 2 X lo4)on the albumin. When, however, RSA is used under similar experimental conditions, low CPJIPA concentrations increase the absorbance at 477 m p and decrease the absorbance at 350 m p (indicating more HBXBA bound). Furthermore, at high CPA\IPL4concentrations (in the RSA system) the absorbance at 350 mp approaches that of

-

7tiU 0

0

-20

-20

-4 0

-40

-60

-6 0

-80

-8 0 n

n

I

I

P

o_ X

-100

-100

n

8

e

U

4

-120

-I 20

I

I

I

I

I

400

450

500

550

600

I~igiiw lO.-lIi~itt~ rotatii)ii itc a. f i i i i c t i o i i i i i waveleiigii ;it rabbit seriirn albmiiii concelitratioii> of 5.0 X .\I: b, without, HBARA, but with or withoiit CP3LP.A added; B, with .-LO X 10-5 J f HBABA and 5 . 0 x 10P .If CPSIPX added: C, with 5.0 X 1 0 F .TI H B h B h added, hilt 110 CPlIPA.

free HBXBX while the absorbance at 177 nip docs not. It should be noted that neither CPlIl’*\ alone nor Cl’lIl’A iii the presence of IiSA absorbs a t 477 nip. These observat,ions, iridicatiiig a11 iiicreasc i n binding of HBAHA upon addition of C P l l P X , suggest that there are at least three difl’erent sites of iiiteraction of HBABA 0 1 1 1iS.A: i~ strong arid weak biiidirig site (present i i t all species diff erent albumiiis studied :rh iridicatcd h\. equilibrium dialysis)! and :L third site liberated upon addition of Cl’:\II’:l: the c47i for the third site may nlso be greater than the t d i i for the stivngei- (first) site for RSX-HBABA. I.‘urthei, \ ~ o r l is i i i i progress in order t o determine to what cxtent increased binding of HBAUX contributen t o e4j7 i n the presence of C‘I’lll’X. The iricrezise iii 11inding of HB-\BX could he the result of i i sm:iil molecular perturbation on the ltSA itiolecule caused by the hiiidiiig of C”l’:\[l’.\ to a n allosteric site.’” 2 4 This results i i i the freeing of a rieir site to \vhich HBAH,\ iiiaj. :ilso bind. The IISX-HBXB.1 \\.stem ma\. ht. thought o f :w :milogous to :ui enzyme s>.?;teni w l i e ~ ~ t h r term allosteric site is used to differentiate betweeii t \vo diff’rrctiit sitw O I I i r i i e ~ i z y n i e . ? ~111 ~ the RS& tem HHXBA~Z may he thought of :is hinding to :in “iictive site” :md is conipetitivel\. inhibited k ) ~ . ~’I’lII’A. Ho\f-ever, (’1’lIP.l concomitantly binds r ~ versibly t o LLII “allosteric site” which releases n rien. “activtr site” f o r compt’titivc-’ hidirig of HEL\B.l and * I):\I l’;4.

-I4Q

I

400

4dO

5bO

550

6b0

;\IO CALCULATIONS OF ANTICONVULSANTS

September 1969

1 or 2, above) since (a) the induced anomalous ORD could only be detected in the region of maximum absorption for the HBABA-protein complex, (b) the E R D effect was small, and (c) no change was observed in the ORD spectrum a t 233 mp. These data, however, do not diff ereritiate between configurationally or conformationally induced rotatory activity. Irrespective of the mechanibm of induced optical rotatory activity, these data complement the results obtained bpect rophotometrically and by equilibrium dialysis and do buggest that if any structural change in the protein is involved, the change must be exceedingly small and/or occurring at the end of the polypeptide chain. Certainly, a large rearrangement of becondary structure is not involved. That a small

761

change in structure is occurring with RSA is suggested by the displacement of the induced ORD to a longer wavelength (a 26-mp shift) upon adding CPNPA. This shift in the induced ORD is not paralleled by a bathochromic shift of the A,, a t 477 mp; i.e., HBABA seems to be binding to a new site in the presence of CPXPA. Again, it is difficult to explain the unmasking of this new site without invoking the concept of allosteric transition. How general this phenomenon is and how important it may be in drug transport mechanisms remains to be seen. Acknowledgment.-We are grateful to the Sational Institutes of Health for support of this work through Grant HE 12740-01.

Molecular Orbital Calculations on Anticonvulsant Drugs P. R. ANDREW Chemistry Department, University of Jlelboume, Victoria, Australia

Received December 2, 1968 I\Iolecrilar orbital calciilations on a number of anticonvrilsant drugs and relat,ed compounds have been completed by two met,hods, extended Hiickel and complete neglect of differential overlap. Calculated dipole moments indicate that. the latter method is more suitable for assessing net atomic charges. The calculated atomic charges at a “biologically active center” proposed by Perkon, together with those a t atoms capable of forming hydrogen bonds, have been compared w-it,hobserved anticonvulsant. activity. The “biologically active center” does not appear to effect, activity, while the hydrogen-bonding at,oms, although common to all the driigs stndied, are not proved responsible for variations in activity.

Widespread research 011 anticonvulsant drugs has led to numerous theories’ which ascribe their CSS activity to a variety of simple physicochemical properties, but none appears to account satisfactorily for all the observed facts. A selection of anticonvulsant drugs which have proved useful clinically, together with some related compounds. is shown in Table I. These compounds all have a similar structure, and the presence of the grouping I appears to be a possible factor in their activity. Furthermore, it seem3 feasible that variation in the net atomic charges in this part of the molecule might change the C S S activity of the drugs by altering hydrogenbonding behavior.

complete neglect of differential overlap calculatio~i (CND0/2) devised by Pople and SegaL4 The original atomic parameters have been retained, except for some of the valence-state ionization potentials employed in the E H T calculations, which were averaged from atomic spectral data.; The values used were (in eV) HI,, 13.6; Czs,2023; C2p, 11.3; Nzs. 26.5; X2?, 13.6; 0 2 s , 33.0; OzP, 16.2. The calculation of atomic charges by both methods, and of dipole moments by the CNDO/? method, is described in the original papers. The E H T dipole moments, p , were evaluated from expressions 1 and 2 where Q.4 is the net charge on atom A, x.k is the atoms

p, =

4.80

atoms

&*x* - 7.337

P(2s,2p,)*/Z*k

(1)

A

PcL2=

H I

I1

Another hypothesis has been put forward by Perkow,2 who suggests that, the net charge at a biologically active center (BAC), starred in 11, is partly responsible for the type arid degree of CNS activity. I n this work both hypotheses have been tested by completing molecular orbital calculations on the compounds shown in Table I. Methods

The mo1ecul:ir orlital calculat~ionsused were the exteiided Hiickcl theory (EHT) of Hoffmnrui3 and the (1) T. C. 13utler, Pharmucol. R e v . . 2, 121 (lY50). (2) JV. Perkow, A r z n e i m i l t e l - F o r s c h . . 10, 284 (1960) (3) K . Hoffmann, J . Chem. Phy.., 39, 1397 (1963).

FZ2

+ + P,*

FcCz2

(2)

x coordinate, P(2s,2pZ)*is the bond order4between the 2s arid 2p, orbitals, and 2.i is the Slater exponent. The first term in eq 1 is the contribution from the net atomic charges, and the second is the atomic polarization contributio~i.~ Using Hoffmann’s program6 as a basis, a Fortran IV program6 has been written which does E H T calculatioris for systems involving up to 96 atomic orbitals. The C S D O I 2 calculations were done with a Fortran IV program6 written by Segd, which hnndles :L maximum of 72 (4) (a) .1. .A. Poyle and G . .\. Segal. i h i d . . 43, $130 ( l 9 G 5 ) ; i h j .I. . A . Pople and Q. .i. Seral, ihid., 44, :328Y (1Y66). ( 5 ) (a) H. A . Skinner and 11. 0. Pritcliard. Clrem. I l e i , . . 5 5 , T45 ( l Y 5 5 j : (11) G. Pilcher and H. A. Skinner. J . I n o r g . S u c l . Chem.. 24, H 3 i (19621. (6) Available from Quantum Chemistry Program Exchange. UniversitLof Indiana, Bloomington, Ind.