Sept. 5 , 1961
~ G G R E G A T I O NOF
ACRIDINEORAGE BOCNDTO POLYAKIONS
3627
ORGANIC AND BIOLOGICAL CHEMISTRY [COSTRIBUTION FROM
THE
SECTION O N PHYSICAL CHEMISTRY,KATIOSALINSTITUTEOF MENTALHEALTH,BETHESDA, MARYLAND]
Aggregation of Acridine Orange Bound to Polyanions : The Stacking Tendency of Deoxyribonucleic Acids BY A. L. STONE AND D. F. BRADLEY RECEIVEDMAY2, 1960 The strength of the dl-e-dye interaction (stacking tendency) between acridine orange (AO) cations bound t o neighboring sites on the polyanion deoxyribonucleic acid (DNA) has been measured experimentally in terms of a stacking coefficient and several other parameters derived from the spectrum of the bound dye. The stacking coefficient was found to be uniformly low for native DSAs obtained from a variety of sources ( K = 1.25 f 0.07) and to increase upon denaturation ( K = 2.66 f 0.34 for half-denatured DNAs and K = 6.30 + 0.50 for 100°-heated DNAs). These results together with previous findings demonstrate that the strength of the dye-dye interaction is a function of the conformation of the polymer to which the dye is bound.
Introduction Many organic dyes which do not obey Beer's law have long been thought to form dimers and higher aggregates in solution. I n certain cases (e.g., thionine,' methylene blue1 and acridine orange2) this theory has been rigorously tested by studies of the variation of absorption with concentration and temperature and has been validated. Color changes similar to those accompanying aggregation in solution also occur when certain of these dyes are bound to a number of different polyelectrolytes. Michaelis3 and others4-' have proposed that the dye molecules are held sufficiently closely to one another on the surface of the polymer to allow them to interact to form aggregates similar to those found in solution. In recent publicat i o n ~ ,this ~ , ~theory has been extended by the discovery that although these dyes aggregate on many polyions, the strength of the dye-dye interaction between neighboring dye molecules depends markedly on the particular polymer to which the dye is bound. The terms stacking and stacking tendency were introduced both to indicate that the interaction is maximal when dyes are ordered in card-pack fashion and that the strength of the interaction varies with the dye and polymer involved. If the dyes are bound to fixed sites on the polyions, the differences in the stacking tendencies of a dye bound to different polymers could be explained readily in terms of differences in the distance and angle between neighboring binding sites. Evidence for this interpretation was obtainedlO by showing that the stacking tendency (1) E. Kabinowitch and L. I ? . tSpstein. . I . A m Chew. S o r . , 63, 69 (l9tl). (2) V. Zanker. %. g h y s i k . Chein , 199, 2'25 (1952). (3) I.. Michaelis, Cold S p v i n g Harbor .Yplnposianr oii l J u o i i i i l o f i w fiiology, XII, 131 (1947); J . P h y s . Colioid C h e w . . 64, 1 (1950) (1) P. D. I,umley, Riochim Biophys .,lrfci, 19, 328 (18.56). (5) U '. Appel and G . Scheihe, Z. S a t u r f o ~ s r h u n g 13b, , 359 ( 1 9 d X ~ . ( t i ) A . R. Peacocke and J. N. H. Sherrett, Tvoirs Faraday SOL.,62, 2tj1 (1956).
(7) hl. D. Schoenberg, C. N . Loeser and 1. J,, Orhison, personal communication; for a review of t h e subject of metachromasia in isolated dye-polymer systems see M. Schubert and D. Hamerman, J . Hislochem. Cytochem., 4, 168 (1956). (8) D. F. Bradley and M . K. Wolf, Proc. S a t l A c a d . Sci. 0'.S., 46, 944 (1959). (9) D. F Bradley and hI. K. Wolf, in "Neurochemistry and Nucleotides and Amino Acids." R. 0. Rrady and D. B. Tower, John Wiley and Sons, Inc., New York, N. Y . , 1960, p. 89-108. (10) D. F. Bradley and G . Felsenfeld, Nolure, 186, 1920 (1959).
of A 0 on DNA increased markedly with heat treatment of native DNA which causes the polymer to undergo a helix + coil conformational transition. Further evidence for this interpretation could be obtained by demonstrating that all polymers with identical conformations have the same stacking tendency with a given dye. This communication gives the results of studies of the stacking tendencies of A 0 on 24 DXA preparations from various sources isolated by a variety of methods. Experimental and theoretical methods for evaluating the stacking tendency are described and discussed in detail.
Materials and Methads Dyestuff .-Acridine orange (AO) obtained from National Aniline Co. was 2 X recrystallized as the free base from ethanol-water by dropwise addition of NaOH less than 0.1 M in concentration. The precipitate was washed with water and dried under vacuum (m.p. 180-181.5", Beilstein, 181-182"). The equivalent weight was determined by potentiometric titration with 0.100 M HCl in 1:1 methanol: HzO. The average of 5 determinations was 264 4= 3 (f.w. = 266). Molar extinction coefficients were measured in a Cary Model 14 Spectrophotometer on solutions prepared b y weighing aliquots of the dye, adding sufficient 0.100 M HC1 t o neutralize and filling t o volume with standard buffer Msodium cacodylate ((CH8)pAsOzNa) 2 X N HC1 pH 6.7). Experiments were carried out in the region of M where the extinction coefficient of A 0 varies with concentration. The relation between the optical density (OD) and molarity of A 0 was found t o be OD (492 mp) Molarity -40= --- ________56000 - 4300 OD (492 r n p ) ( 1 ) The molar extinction a t the absorption maximum (492 m p ) extrapolated to zero concentration is 56000. Although Zanker2 has reported a higher value, 61000, a t zero concentration in PH 6.0 acetate buffer a t the maximum (20400 cm.-' = 490 mk), the extinction coefficient from the present work (55700) is in good agreement with Zanker's observed value (57600) a t the lowest concentration he employed ( 1 X 10-0 M). Stock solutions of A 0 were spectrophotornetricall! stable over periods of many months in the dark a t 4'. Solutions were centrifuged a t 3000 r.p.m. in the International Clinical Centrifuge (Model CL, f213 head, Will Co.) for several minutes before each experiment. Deoxyribonucleic Acids. A. Normal Mammalian Deoxyribonuc1eates.-Calf thymus DNA samples C. T . M . (Mann Research Laboratories, M.A., highly polymerized), C. T.S. (Sigma Chemical Company, highly polymerized) and C. T. W .( Worthington Biochemicals, highly polymerized) were prepared by the method of Mirsky and Po1listerl1using
+
(11) A E Mirsky and A W Pollister, J (1946).
Gen P h y s t o l , SO, 117
A. L. STONE AND D. F. BRADLEY
:$'LS
Vol. s3
E. Bacteriophage Deoxyribonuc!eates.--Sam$es 'r,w3s prepared from Tc bacteriophage by one of us (DFB) using n phenol extraction procedure essentially that described bv Gicrer atid Schramm.le Sample #X from +X 174 b?ct e r i o p h a ~ e land ~ stored in physiological saline was a gift of Dr. Robert Sinsheimer, California Institute of Techriology. Preparation of DNA Solutions.-One nd. of standarc! buffer was added to 5 mq. of DXA and allowed t o stand a t 4 for 16-20 hr. The resulting transparent gel was further diluted by the stepwise addition of buffer over a period of several hours with gentle manual shaking. Finaily, t1:r solution was subjected t o mechanical shaking (16 hr. a t 3 " ) . In most cases this procedure resulted in apparently hnmogeneous solutions of 1-2 mg./ml. DSA. With - o m r tumor DXAs (RCS, RCJ and RCJI) the gel. did not disperse upon addition of buffer. In these cases, in order to bring tile samples into solution, the system was subjected t o high hydrodynamic shear by passing the gel suspension rapidly through a #30 gauge syringe needle. All stock DXA solutions were centrifuged a t 31100 r.p.rn. in the Interiiationrll L Clinical Centrifuge t o remove any insolubie matter and stored i n the dark at 4'. J I 1 I--] I 10 1 2 3 4 5 6 7 8 9 Samples Th, Pn mid Qx which were obtained in salt solutions were dialyzed against standard buffer belore use. D N A ml , In the heating experiments, samples of the stock DXA soluFig. I.-Spc~troi)lir~ttI;i:etric titration of A 0 with nativc tiorx were heated a t 62" and at 100" for 15 minute perirda J)KA: ._.__., iioriiial titration curve obtained under nnd then cooled rapidly. Previous worklo had established standard conditions; stepwise addition of DKA stock solu - that Sigma calf thymus DNA is half denatured at 62" ( = Tm! under the conditions of these experiments. tion, 0.04 mg./ml., to the initial dye solution containiiig Viscosity.-~iscosities were measured with n single-bull) 4.6 X 10" m o l e of A 0 in 2.4 ml. of buffer (see text for furmodified Ostwald Viscometer. The sample volume was one ther details); 0, optical density of A0:DNA mixtures at ml. with an outflow time for water at 25" of 57 sec. and :in sverage shear of 209 sec. -1. Solutions of DXA at concentraselected points in the titration before ultrafiltration; A, tions of about 0.04 mg./m!. in M/1000 cacodylate buffer optical density of the free A 0 in the ultrafiltrate of the were run a t 25" t o determine the low salt viscosity. Fiftv corresponding AO: DXA mixtures. PI. of 5 M NaCl in the stsndarci buffer then was added t o the sample and the viscosity a t high salt (0.25 M )determined. Sedimentation.-Sedimentation studies mere made with a mo6ified Sevag technique; sample C. T . D . - I I ~was ~ prepared by one of us (.4LS) using the method of Kay. P t aZ.Ib the Model E Spinco Ultracentrifuge equipped with ultraKat intestine DX.4, sample K . I . (prepared by ALS) and violet optics. Optical density tracings of the films were guinea pig testis DNA, sample G.P.7'. (gift of Dr. Sanford made either with a Spinco Analytrol rnicrodensitotneter or a n enlarger-Densichron Densitometer arrangement. Dye Stone, National Institute of Allergy and Infectious Diseases) were prepared hy the Inethod of Kay, ~t nl., modified by the adsorption o n the cell walls was redirced by use of anodized center pieces. use of 0.001 M NaCl instead of distilled water.lab Ultrafiltration.-Dye and dye-DSA solutions were filtered B. Tumor Deoxyribonucleates f previously described12J4). --Sample SlS04 was obtained from the Crocker Lahs., through 50 mp Millipore filters (VM) using a Swiiiny hypoMouse Sarcoma. 180; sample SIROR? was obtained from a 6- dermic adapter and her-lock syringe. Under these condimercaptopuriue resistant strain and sample SlRORtl was tions the dye: DNA complexes were retnined quantitatively rlbtairied from the resistant strain after treatment of the mice by the filter and the filtrate consisted of a solution of free dye. Free dye adsorbed onto the syriuge surfaces and millipore with f i - m e r c a ~ t ~ - , p u r i i i eAn . ~ ~aliquot ~ of SIPOR2 fibers was filter. Therefore, the system was pre-equilibrated with dye taken from the ethanol wash without subsequent acetone washes yielding sample S189Rsx: sample S180Rt, flor was so that the amount of adsorbed dye remained relatively conolit.ained from the first supernatant as a flocculent precipi- stant during the filtration of mixtiires containing both free dye and iiye:DNA complexes. The concentration of t!ie tate. Sample R C was obtained from R C Rdenocarrinom.1 mouse free dye ii! the filtrate was corrected for the small changes i i i the amount of dye bound t o the filter and tiimorsl"'; liCsl arid RCJI were two fractionq (fibrous and flocculent, respectively) of DNA isolatediab from the same rurred using control solutions containing s tinns of free dye. tumor *at a later time when growth potential was declining. Spectrophotometric Titrations.-Spectrophotometric tiAfter tumor bearing mice were treated with 6-mercaptopurine, D S A samples RCSS? i f r i m grnwing tumors) and KCS trations of A 0 with DNA were carried out (;ls described previouslylO) in stoppered silica cuvette-. The starting ( from regressiiig, sensitive tumors) were obtained.'*a Sample KC.7 soZiddP was a readily soluble fraction of RCS using solution contained 4.6 X mole of A 0 in 2.4 ml nf standard buffer and had an optical density at 592 mp of the standnrd buffer system. Sample Th (gift of Dr. James Rose. Kational Cancer Institute) was obtained from a mouse approximately 1.0. Aliquots of the stock DNA solutions were added stepwise to this solution with calibrated Hamilton thyroid and qtored in physioloqical saline solution. C. Salmon Sperm Deoxyribonuc1eates.-Samples S.Sp.1 microliter syringes. The delivery of the pipets was checked by weighing the cuvette before and after each addition and !California Cnrporation for Biochemical Research) and S.was generally within =!= 0.1 p1. of the rated value. The C O R Sp.2 (Mmin IZrsearch Laboratories, M.A., highly polymerized) were prepared by themethod of Emanueland Chaikoffl6; tents of the cuvette were then stirred by means of a small sample SP w:i- rlisciissed in previouq puhlications.8,10 magnetic stirring bar's which remained in the cuvette througliD. Pneunionococcus Deoxyribonuc1eate.-Sample Pn, out the titration. The stirring time ranged from 90-180 ? transfnrminq D S X containing 4.97, RN.4 and stored seconds depending on the voluiiie and viscosity of the solution. The spectrum of the sdution was then measured in p:~v~iolo~ic saline, al wap a gift of Dr. Aaron Bendich, with a Carp 14 Spectrophotometer. Tests were made t o Sloan-Kettering Institute for Cancer Research. establish that the spectral changes which occurred after each increment of DNA were complete and showed no further (12) f,3! .4. I, Stone, P!r.D Thesis. Graduate School of Medicnl change with time. A t the end of the titration the PH and Science-. Corne:l U n i v e r s i t y . New York, N Y . , 1959; (b) i h i r l , p. 4:3. the temperature of the solution were measured (PH 6.6-6.7, (13) l a ) E . K a y N S i m m o n s and A. Douncr, J . A m . C h c m . S o c , T := 27 xk 1'). 74, 1721 f195'2), (h) i h t d , modified by the w e of 0.001 A4 NaCl inr
1
T
8 r
-
T
I
I
1
r - ,
-Noirnol Tifrolion C u r v e o Optical Denslty of fotol A O ( 8 e f o r e f l l f r o t i o n l A Optical Denslty of free A 0 (Ultrafiltraie 1
i
8tead of distilled wiiter T h c P r r ! r r ! i - i l A ' r c r o r t r , 2 0 , 1 I57 (19.79). C IC. liinaniicl and 1 . L. Chiiikuff, J . B i d . Chcm., 203, 167
11.1) A I.. S t o n r ,
11 !s) ( 1 Y5:i).
Sept. 5, 1961
-4GGREGhTION O F
ACRIDINEO R A N G E BOUNDTO
As seen in Fig. 1 titrations show a n end-point indicated by a minimum value in the volume-corrected optical density of the dye a t 504 m p . With 1.0 mg./ml. DNA solutions this eud-point occurs a t approximately 14 A. The titrations are continued beyond this point until 1300 pl. of DXA have been added, that is, to approximately 100 times the amount needed t o reach the end-point. With some samples of DNA a red precipitate appeared before the end-point. To avoid precipitation a dilute solution (0.04 mg./ml.) of DNA was used in the titration. The end-point was reached a t about 500 pl., and the titration was carried out to about 3 times the DS.4 end-point. A separate titration with the concentrated solution (1.0 mg./ml.) was then performed using a sufficiently large initial aliquot (ca. 25 pl.) to overstep the region of A 0 excess and thereby eliminate precipitation (an alternate procedure reduces molecular volume and weight of the 1 nig./ml. DNA sample by hydrodynamic shearing). Stoichiometry.-The titration end-point was defined as the intersection of the two linear limbs of the experimental curves obtained by plotting the volume corrected optical densities or E, vs. ml. DNA (upper curve, Fig. 1). The titration data were treated as follows: the moles of dye initially present were computed from the optical density and vulume of the iuitial solution using eq. 1; molar extinction coefficients (E,) of the dye a t each point in the titration were computed from the moles OF dye and the optical density and volume of each solution. From the volume of DNA added to reach the end-point, and the moles of dye, the molarity of DNA phosphates in the titrating solution was calculated assuming a 1:l A 0 : D N A complex (see below). The molar extinction of the DNA in standard buffer ( E p ) was calculated from the optical density of the DNA stock solution a t 260 mp and the Dh'A phosphate molarity computed above. This molarity was also used t o compute the ratio of DNA phosphate ( P ) t o A 0 ( D ) , P / D , for each point 011 the titration curve. Spectra of A0:DNA Complexes.-During the course of the titration the visible absorption spectra of the solutions vary with the relative amounts of DNA and dye. The 1: 1 A0:DKA complex spectra correspond t o the point in the titration a t which Emc is a minimum. The 1:100 AO: DNA complex spectra are computed from the final experimental point in the titration, corresponding t o P / D = 100. a/P Ratio.-The free monomeric A 0 absorption maximum is a t 492 m p . The maximum for the 1: 1 AO: DNA complex spectrum occurs a t 361 mp. These absorption bands have been designated as the CY and 6 bands, respectively,3 and we shall refer to the ratio of their molar extinction coefficients ratio is a minimum near the titraas the a/@ ratio. The CY/@ tion end-point and increases with P / D . As shown in Fig. 2 a plot of CY/@ vs. P / D is linear and the parameters ( ? / ~ ) P - D (P/D),5 and d(a/B)/d(P/D) = s can be computed. dE/d(P/D).-The extinction coefficient a t 504 mp rises linearly with P / D from the end-point t o approximately P / D = 3 (Fig. 1). The slope of the best straight line through these points is termed d E / d ( P / D ) . F Values.-As the P / D ratio increases indefinitely the AO: DNA complex spectrum asymptotically reaches a limiting curve, the 1: m A 0 : D N A complex spectrum. At intermediate P / D the AO:DN.4 complex spectra lie between the l :l and the l : m spectra. The fraction, F , of the total change from 1:l t o 1: m a t any P / D is given by the exmession ( c f . ref. 8) where E, E,:] and E m are the extinction coefficients a t a given wave length et any P / D > 1, a t P / D = 1 and as P/D- m , respectively. Unless otherwise indicated the F values are computed at the wave length maximum of the 1: m complex (504 mp) where E , - E l : , and E,/EI:I are maximal. El:] and E, are obtained from observed E values by extrapolation t o P / D = 1 and P / D + m using a plot of E US. (1 - D/P)O. Stacking Coefiicients.-The variation of F with P/D can be expressed in terms of a parameter, K , the stacking coefficient,by a n equation previously presented* P I D = (1 F / z ) - ~ (K 1)(1 F/z)-l (1 F F'/g)F1/t ( 3 ) or
-
+
-
+ -
i
I
I
I
3639
POLYANIOXS
'/"
I
/
1
4
J
1
NATIVF
d J
i
I
I
1.5
--~J
--I__-J_I--L-..b-!
20
25
3b
35
40
45
50
P/9
Fig. 2.-Plot of a l p ratio vs. P / D : o-0-0, native calf 62O-heated SlSOKl DNA; thymus DNA; A-A-A, 100O-heated calf thymus DNA. The plot yields the following parameters listed on Tables I1 and 111: the slope, s = d(a/a)/d(P/D); the reciprocal of the intercept = @/a a t P/D = 1; the value of P / D a t a/,8 = 1.
K is computed from F and P / D values from a single titration and averaged over the range of F between 0.2 and 0.8. Outside of these limits the computed value of K is sensitive to small errors in the determination of El: 1 and E,.
Results and Discussion Formation of an A 0 :DNA Complex.-A typical spectrophotometric titration of A 0 with DNA is shown in Fig. 1 (upper curve). The linearity of the titration curve to the end-point indicates that the equilibrium is far on the side of complex formation. This indication was confirmed by ultrafiltration experiments. No free A 0 was present in ultrafiltrates of titration mixtures a t the end-point (Fig. 1, lower curve). Additional evidence was obtained from ultracentrifuge studies which showed that there was no significant amount of non-sedimentable, ultraviolet-absorbing material in end-point titration mixtures (Table I). These data indicate that complex formation is >95% complete a t the end-point . TABLE I SEDIMENTATION OF AO-DNA COMPLEXES Sedimentable component in D and E had 3 = 22.3 and 24.0, respectively, corresponding to native DNA. In C sedimentation of moving component was complete by the time rotor reached set maximum speed (59780 r.p.m.). A l * is the monomolar concentration. L e . . the concentration of the monomer.
Solution
A 10-8 M Cacodylate buffer B 2 X 10-'M A 0 c B x 10-5 M A O 0.5 x 1 0 - 8 A i * DNA D 2 x 10-5 M AO 2.0 x 10-6 M* DNA E Same as D -t 8 X 10-0 Mcacodylate 100 X 10-6 AI* F 2 X 10-e .if A 0
+ +
+
DXA
Film optical density suln.
Film optical den sit y non-sedimsnting component
0 0.26
0 0.26
.45
.19
.45 ,53
.21
>>1
.05
.01
A. I,. STONE AND D. F.BRADLEY
ITOl. s3
TABLE I1 STACKISG PARAMETERS OF NATIVEDEOXYRIBONUCLEIC ACIDS o,ool __
I)
P/D at
Sample
K
C.T.S.d C.T.M. C.T.W. C.T.D.-I S.Sp.1 S.Sp.2 R.I.
1.16 f 0.14 1.13 2Z . l l 1.11 i .15 1.15 2Z .12
G.P.T.
@/aa
1 . 1 8 3 ~ .11 1.25 f .18 1 . 2 1 f .10 1.37 f .16
Th
S1804 S180Rz S180K2, S1801tt2 RCI KCSS2 RCS T 4
Pn Mean Mean dev.
1.26 i 1.19 f 1.23 + 1.24f 1.27 f 1.19 i 1.29 f 1.33 f 1.54 f
.07 .12 .09
.11 .21 .17 .14
.11 .07
1.25
p
= a
Extrapolated
AP/D
sb
1.31 1.91 0.259 1.25 1.89 ,253 1.30 1.91 ,225 1.29 1.94 ,239 1.39 2.08 ,245 1.35 2.15 ,224 1.36 2.00 ,262 1.32 2.11 ,221 1.27 1.85 ,253 1.33 2.04 ,260 1.32 2.02 ,237 1.32 2.00 ,247 1.33 2.15 .. 1.35 2.18 .232 1.33 2.04 ,241 1.31 2.04 ,230 1.43 2.26 ,239 1.37 2.22 ,224 - - - _ _ _ _ 1.33 2.04 .242
ei:i
9200 9200 9200 9200 8300 7800 8900 7600 8600 8600 8800 8600
.. 8000 8100 8000 8300 7400 - 8500
14000 13900 13500 15000 13500 13500 13600 14000 12800 13200 13500 13300 14500 14300 15000 14000 13600 14300 _ _ 13900
€a
eD
54700 55000 55800 54000 55000 54000 54500 54000
6000 6200 6200 6100 7000 6300 6200 6300
P/D at F = 0.5
7)
phoshfo!. Dhate wt. A,C 0.25 analysis X 10-6 %
3.72 3.66 3.62 3.70 2 . 4 4 3.75 2 . 9 7 3.89 3.89 4.12 2.24
...
..
..
54000 55000 54800 54000 54000 54000 54500 55000 55500
6300 6200 6000 6100 6400 6100 6300 6000 6400
3.91 3.77 3.85 3.87 3.93 3.77 3.93 4.04 4.45
54600
6200
3.88
cp from
6400
5 . 4 30.1* 29.7f 29.7'
2.65
6300 6200
10.9 31.ze 1 0 . 5 30.V'
1.48 2.54 2.21 2.42 1.98
6600 6100 5800
4 . 6 30.8" 4 . 3 29.8" 30.;' 31.6e 32 21 29. SJ
__ 6200
1 0 . 0 3 1 0 . 0 9 f O . l l i 5 0 0 f 5 0 0 f500 f l O O f O . 1 4 f200 1.60 3.68 0.154 4700 8750 Prep / a extrapolated t o P / D = 1 from a / p vs. P I D plot, Slope of a / @vs. P / D plot. c Adenineltotal base X 100. viously reported.10 e Ref.12 f E. Chargaff, "The Nucleic Acids," hcademic Press, Inc., Yew York, N. y., 1955, pp. 307371.
+X
f0.07
.........
Sample
Heated 100" C.T.S.
s.sp.* S180R2 R.I.
Mean Mean dev. Heated 62' C.T.S.c S180R2 S180Rt2
TABLE I11 STACKING PARAMETERS OF DENATURED DEOXYRIBONUCLEIC ACIDS P/D at Lk Extrapolated K
6.92 A 3.50 5.40 i2.43 6.68 A 3 . 2 5 6.20 f 2 . 9 8 6.30
f0.50 2.15 i 0.07 3 . 0 5 f .42 2 . Z zk .34
a =
@/a5
p
Sb
AP/D
5200 0.179 .161 4600 ,170 4900 .155 5200 - - - - _ _ _ 1.71 4.14 ,166 5000 10.008 f 0 . 4 0 f 0 . 0 0 8 f 2 0 0 3.75 4.29 4.79 3.74
1.72 1.71 1.71 1.69
1.48 1.52 1.55
2.61 0.202 2.86 .183 2.88 ,187 _ _ 2.78 .191 f0.12 f0.008
Mean 2.66 1.52 Mean dev. 2Z0.34 f0.02 Heated 100": Native ( 1 : l ) R . I . 2.36 zk 0.13 1.51 Other DNA SS DNAC 3.49 & 0.49 1.54 S180Rt2 floc. 1.99 f 0.26 1.46 RC3I ......... 1.39 RCJI ......... 1.39 RCS soluble ......... 1.45 a p / a extrapolated t o P / D = 1 from CY/^ z's. P / D
2.64 3.68 2.75 2.39 2.54 3.8 plot.
0.207
b
Stoichiometry of Complex Formation.-The ( E p ) ' sof the DN-4 solutions calculated from the titration data (Table 11, col. 9) agree with those calculated from phosphate analyses and are within the expected range for native DNA (Table 11, col. 12). This agreement shows that the end-point corresponds to the formation of the 1:l complex. The titration may be used to measure DNA mono-
ei:i
12100 12300 12000 12700 - - _ 12300 f300
5500 13700 5400 13700 5500 13400 ~ _ _ 5500 13600
em
49300 47800 47300 50300 _ 48700 iz1100 54300 56000 55400 _ _ 55200
f O f100 i600 6000
0.260 4500 ,180 6100 ,200 6100 ,183 5800 .111 4200 Slope of a/p vs.
14300
55000
P / D at CI)
F = 0.5
8400 8300 7700 8300
15.8 12.2 19.4 7.8
8100 +300
13.8 *3.8
~.!.OO! 7j 0.23
98.0
'7100 5.61 7300 7.33 6.00 7500 6.80 4.63 _ _ _ _ 7300 6.58 &lo0 1 0 . 6 5
7300
6.02
13700 53700 7500 13000 55600 5700 13500 ... 6600 13500 ... 6700 13300 ... 5900 Previously P / D plot.
8 . I7 5.31
9.85
2.22
molar concentrations to within a few per cent. with small quantities of DNA (ca. 20 X g.). The number of titratable sites does not change upon heat denaturation. The E,'s increase upon denaturation (Table 111), in agreement with previous work.I0 Single-stranded 4x174 DNA exhibits many of the properties of heat-denatured DNAs, These data show that E,, vatues calculated
Sept. 5, 1961 235000-'
-4GGREGATION OF A C R I D I N E OR.lNGE
"
'
I
'
I
'
3 30000 0
a
k
25000
X W 1
I 20000 00LL
15000
-
I
0
z
0
5
10000
z
X +
a
a
i 0
z
5000
I
540
1
"
0
0
363 1
POLYANIONS
POLYMER S I T E / D Y E RATIO, P/D.
-
I
I-Unheated DNA 2-Heated DNA 0-&.174 DNA
I
k
I
BOUNDTO
1
0
55
50000
49 -*
IO
5
100
CnLF THYMUS DNA o Native A 62OHeaIed
45000
240000
2
I
2
IOO'Heated
35000
I
520
500
480
460
440
420
I
400
1
IOOOO~
I
02
01
'
04
1
1
I
I
1
05
06
07
OB
0.9
I
LO
(l-D/p)z
WAVELENGTH IN rnp.
Fig. 3.-Spectra of 1 : 1 complexes of A 0 with native and denatured D x - 4 ~ . Molar extinctions of A 0 (abscissa) at various wave lengths (ordinate) were calculated for the titration solution with minimum molar extinction at 504 nip using standard titration procedure of stepwise addition of D S A solution t o initial dye solution containing 4.6 X lo-* mole of A0 in 2.5 ml. of buffer (see text for further details). Curve 1, native calf thymus D X 4 ; curve 2, 100"heated calf thymus D S A ; 0 , 4x174 DNA.
1
03
Fig. 4.-The unstacking of DNA-bound A 0 with increasing polymer site/dye ratio: _ _ , theoretical curves for K = 1 and K = 2.26; .-.-. , experimental curve for 100'heated calf thymus DNA. Open circles are experimental points for native calf thymus DNA. Open triangles are experimental points for 62O-heated calf thymus DXB.
with a maximum a t 504 mp and extinction of 55000 (@/cy = 0.47) (Table IV). Because the spectrum of the 1:100complex is similar to that of the freefrom the A 0 titration may be used to determine A 0 monomer except for a 12 mp red shift, the the extent of denaturation of DNAs (e.g., see new band is referred to as the bound-A0 monomer band ( a ' ) . The difference in wave length between Table 111, "other DNA"). All of the E,'s reported have been calculated a and p bands in bound-A0 (40 m p ) is almost from titration curves a t 504 m p . Titration end- identical with that in free-A0 (39 mp).2 The bound-monomer spectra of all native DNAs points calculated a t other wave lengths, e . g . , 464 or 430 mp, agree with these values within a few per are the same as that shown in Table IV. The mean deviation of the extinction coefficient (E,) a t cent. Spectrum of the 1 :1 A 0 :DNA Complex.-Under the band maximum is 1% (Table 11, col. 8). The the standard conditions of the titrations of the A 0 E , values reported for 100O-heated DNAs are spectrum has a maximum a t 492 mp with a molar lower (Table 111). This may result from the for deextinction of 52000 and a shoulder a t 464 mp with difficulty in extrapolating to P I D an extinction of 39000 ( i e . , @ / a ratio = 0.75). natured DNA since the spectrum is changing apThe 1:I complex with native DNAs has a shoulder preciably a t the highest P / D examined (Fig. 4). a t 492 m p with extinction of 17000 and a maximum Experiments with a limited quantity of +X DNA 5 ; therefore, a value for E , could a t 464 mp with extinction of 21600 ( @ / aratio = ended a t P / D 1.27) (Table IV). Thus complex formation is ac- not be obtained. Variation of A 0 :DNA Complex Spectra with companied by a large hypochromic effect with a relative enhancement of the 464 mp band. P/D.-A number of parameters have been derived I t can be seen that all native DNAs have the to measure the variation of the complex spectra same 1:l -4O:DNA complex spectrum (mean de- with PID: the rate, dE/d(P/D), a t which the exviation 3%). A convenient parameter for com- tinction a t a given wave length changes; the rate, paring these spectra is (p/cy), = D (mean deviation da/P/d(P/D) = s, a t which the ratio of the heights 3%, Table 11). Heat denatured and +X DNA of the a and p bands change; the P / D a t which show different 1:1 spectra and @ / aratios indicating the two bands are equal, (P/D), = 6; the P / D a t that the spectrum of the 1:l complex is a function which F = 1/2; and the stacking coefficient, K . of the degree of intactness of the DNA helical All of these parameters are the same for native structure (Fig. 3) and that disruption of the DNAs (mean deviation 4-670) (Table 11). These structure causes a relative enhancement of the (p) parameters are also identical for denatured DNAs band. (Table 111), although they differ from those of the Spectrum of 1 :lo0 AD :DNA Complex.-Titranative samples. They indicate that the rate a t tion of A 0 with DNA in the region of DNA excess which the spectrum changes with P/D is less for causes spectral changes which are complete within denatured than native DNAs. Values of these parameters for partially deapproximately 15 seconds. *4s (P/D) is increased indefinitely (to ca. lOO:l), the p band present in natured DNAs lie between those of the native and the 1:1 complex a t 464 mp becomes a shoulder with denatured forms, and it was found that the values an extinction of 20000 while a new band appears for a 1 : l mixture of native and 100O-heated DNA f:
(w)
1'01.
A. L. STONEAND D. F. B~LIDLEY
3632
s:;
TABLE IV SPECTRA O F .kCRIDINE ORAXGE A N D ACRIDINEORANGE: DEOXYRIBONUCLEATE C O l ? P L E X I ? S Extinction coefficient X 10-2 360 380 400 420 430 440 450 460 464 470 474 480 485 402 ,')ilO 504 .?IO 51,; 16 25
26 27
45 33
77 47
Acridine otange (2 X M) 109 156 229 343 387 429 440 457 486
516 4-13 3V2
Acridine orange :deoxyribonucleate complex (1: 77) 58 78 106 165 198 249 273 296 331 426
226
137
541 549
448 303
151 159 155 150 155 151 153 148 161 170 154 157 155 1.4 153
106 109 109 110 108 108 111 106 111 118 109 115
Acridine orange :native deoxyribonucleate 1 : 1 complexes Sample
C.T.S. C.T.M. C.T.W. s.sp.1 s.sp.2 R.I. s1m4 S180R2, RCI
RCNSi RCS T4
24 22 26 25 25 25 26 25 27 26 30
-
Mean 26 Meandev.fl 4x 27
31 30 27 32 30 31 31 28
42 43 41 46 44 45 44 43
. . . . 33 50 32 45 38 53 - 32 45 f 2 f 2 31 46
82 83 82 8; 84 85 81 84
117 121 118 124 121 121
162 165 161 167 165 164 . . 158 121 164 . . 124 . . 92 132 181 83 118 161 93 132 178 - __ 85 123 166 1 3 f 4 3Z5 87 125 175
195 201 198 202 199 157 193 196
208 214 209 216 216 209 211 210
. . . . 218 233 194 209 215 229
- 200 215 f6 1 6 227 261
208 213 209 216 216 209 211 210 222 233 211 229 216 3Z7 267
199 207 203 208 210 202 206 201
191 197 195 197 199 153 196 191
176 184 179 181 183 179 183 176
.. 178 172 170 175 171
.. 167
. . . . . . . . 226 216 199 191 205 196 182 17.5 219 209 192 180
-- - 208
198 183 175 1 3 f 5 260 246 218 196
f 6 zk5
163 173 168 162 169 166 If23 le1 178 187 I69 171 170 1 5 173
135 141 139 136 139 136 139 134 145 152 138 143 140 +4 139
82 82 83 80 83 83
..
80 84 88 83 88 -15 110 84 36 1 3 f2 1 3 1 2 18 113 92 43
Acridine orange : heat denatured deoxyribonucleate 1:1 complexes 62
40 81 119 167 209 21 27 .. 24 . . . . . . 120 . . 28 43 85 I24 175 220 24 43 85 123 174 217 24 29 - - - - _ - 84 122 172 215 Mean 23 28 42 Mean dev. f l 3 Z l f l f 2 1 2 1 3 f 4
C.T.S. S180, S180Rz 180Rti
234 .. 245 240 240 f 4
237 229 218 198 184 171 251 . . . . . . . . 170 248 241 228 205 186 170 243 236 223 201 184 169 - -- - - 245 235 223 201 185 170 1 5 f4 f3 f 2 i l f l
155 149 153 152 152 f 2
138 136 139 137 138
108 83 34 109 85 35 112 8'1 . . 111 88 . . - --110 86 33 f l f 2 A2 f l
.. .
10 17
--
17 f1
1000
31 45 86 24 81 40 21 26 44 85 S180Rz 23 29 - - - 84 Mean 23 29 43 Mean dev. f l 2% 1 2 f 2 C.T.S.
s.sp.*
124 177 224 118 167 213 123 175 223 - - 122 173 220 f 2 f 4 f 5
252 255 244 229 199 178 156 139 125 102 242 244 235 221 191 170 150 131 120 97 254 257 248 232 203 181 161 142 129 105
-
-
37 35 37
15
..
13 - - - - - - - - - - -
249 2 52 242 227 f 5 f5 5 5 1 4
are identical with those of 50% denatured (62'heated) DNA (Table 111). Values of some parameters for 4x174 DNA indicate an even less rapid rate of change of spectrum than for 100O-heated DNA. Stacking Coefficient.-The stacking coefficients for native DNAs are nearly identical (mean deviation 6y0)(Table TI) and are close to unity. The small variation in K between the lowest and highest values is reproducible and may signify real differences. The average K values will vary slightly with the range of F values included in the average since there is a slight bias in the deviation of the data from a fit to equation 4, e.g., although the average value of K is 1.25 as F + 1 the individual values of K -+1.0. The numerical value of K differs somewhat with the wave length a t which F values are computed. F values are several per cent. lower computed on the long wave length side of the bound monomer band so that K values are correspondingly higher: = 1.25 f 0.07, Ksio = 1.62 f 0.10, Kau KM4 = 1.74 f 0.10 for native DN-4 and K w = ~ 6.22
82 78 83
198 176 156 137 125 101 81 36 f 4 f 4 54 f 4 f 3 1 3 f 2 f l
14 11
f 0.67, KE15 = 7.38 f 0.44for 100O-heated DNA. This discrepancy could be due to the small but unequal contributions of dimers and larger aggregates to the spectrum a t these wave lengths. K values computed a t the band maximum (504 mp) rather than on the steep shoulder of the band provide a more reproducible standard. According to the model of Bradley and Wolf8dyes bound to specific polyanionic sites in the 1:l complex interact with dyes on neighboring sites through the same electronic forces which cause the dyes to aggregate in solution. This interaction can alter the energy levels of either the ground or excited states of the dye, or both, causing a spectral shift in the dye (change in energy difference between ground and excited states) and/or a nonrandom distribution of dyes among sites (change in ground state), As excess sites are added to the system containing the 1:I complex, the dyes, in dynamic equilibrium with a small pool of free dye in the solution, rapidly distribute among all available sites. A certain fraction, F, of these dyes will by chance
Sept. 5, 1961
AGGREGATION O F A C R I D I N E O R A N G E
BOUNDTO
POLYANIONS
3633
The average K value for native DNA (1.25) occupy sites with no nearest neighbors. As P/D increases indehitely the fraction of such isolated corresponds to a AF = -0.065 kcal./mole of dyes increases continuously from 0 to 1 and the single dye molecules. The spectral shift from 504 spectrum changes continuously from the 1:1 spec- mp for isolated dye to 464 mp for aggregated dye corresponds to a shift in relative energy levels of trum to the 1: spectrum. Differences between the 1:l and l:m spectra ground and excited states of +4.8 kcal. If the result from dye-dye interactions which affect the entropy change is zero upon aggregation, dye-dye ground and excited states of the dye to a different interaction lowers the ground state by 0.065 kcal. extent. The relatively smaller difference between and raises the first excited state by 4.7 kcal. the spectrum of the 1:m complex and the mono- These relative changes are approximately what meric dye in solution (Ama, 504 and 492 mp, respec- would be predicted on the basis of recent theories of tively) as compared with the 1:1 complex and ag- exciton interaction.20-22 The stacking coefficients for DNA heated to 62’, gregated dye in solution (Amax 464 and 451 mp, respectively’) can be ascribed to the binding proc- the T m for calf thymus DNA, under these condiess. Differences between the 1:1 spectra of native tions varies from 2.2 to 3.1 (mean 2.66) correspondand denatured DNAs result from differential ing to a AF = -0.24 to -0.34 kcal./mole A 0 effects of dye-dye interaction on ground and ex- (mean -0.29). The variation may arise from difcited states of the dye bound to polymers with clif- ferences in the T m values for different samples.28 ferent conformations of binding sites. DNA denatured in this way behaves like an equiGround state dye-dye interaction determines molar mixture of native and denatured DNA, the distribution of the dyes (on the sites) among which has a K = 2.36 0.13 giving further supisolated and interacting dyes. If there is no ground port for the conclusions of Hall and Litt.24 state interaction the dyes will distribute randomly, The stacking coefficient of 100O-heated DNA and the fraction of isolates, F , will vary with P / D varies from 5.4 to 6.9 (mean 6.30) corresponding to as shown in eq. 3 with K = 1. For a randoin dis- a AF = -0.50 to -0.58 kcal./mole A 0 (mean tribution F = l/* when P / D = 3.41. Equating -0.55). These values are not as reliable as for the fraction of isolates with the fractional change native and half-denatured DNA because of the in absorption spectrum defined by eq. 2, we find difficulty encountered in extrapolating reliably to that (rf. Tables I1 and 111) F = l / 2 when P / D = 3.9 E , and the deviation of the experimental curve for native DNA, 6 6 for 62O-heated DNA and 13.8 from the theoretical equation. Using these values for 100O-heated DNA. Therefore native DNA and AF for native DNA the predicted A F for a 50y0 nearly satisfies this criterion of randomness, while mixture of native and denatured DNAs is (-0.07 denatured DN.4s are non-random. There are -0.55)/2 = -0.31 kcal./mole A 0 compared fewer isolates in these cases suggesting an attrac- with the observed value of -0.29 for 62O-heated tive ground-state interaction. Differences ob- and -0.26 for the synthetic mixture of native and served in the rate of variation in spectrum with denatured DNA. P / D using other parameters such as d(cu/P)/dConclusions (PID), dE/d(P/D), (PID), 8 for native and Stacking Tendency.-The term stacking tenddenatured DNAs are explained by a larger ground- ency refers to all effects of dye-dye interaction and state dye-dye attraction in the case of the de- the parameters such as K , dE/d(P/D), ( c Y / B ) P - D , natured DNAs. etc., by which they are measured. The stacking Lawley treated the problem of a dye (rosaniline) tendency will vary with the particular dye and distributing randomly among sites (DNA) when polymer in the complex. Differences in the stackboth free and bound dye are present in the ~ y s t e m . ~ing tendency of a particular dye on a series of His data fit the random case with reasonable agree- polymers is related to differences in the polymers. ment. The treatment of this problem when all the The terms “stacking tendency of a polymer” or dye is bound and there is non-randomness in the “stacking coefficient of a polymer” is used to refer distribution has been carried out by Bradley and to the dependence of the dye-dye interaction on Wolf The degree of non-randomness is expressed the polymer to which the dye is bound. as the parameter K , equal to e-IIFIkT, where Acridine orange has been shown to exhibit a AF is the ground state free energy of interaction of wide range of stacking tendencies with different a pair of neighboring bound dyes. polymers: DNA (K = 1.16),1° ribonucleic acid A probability interpretation of K may also be (K = 2.9), acid polyadenylic acid ( K = 12.3), given. For example, for K = 1.25, a dye will be polyuridylic acid ( K = log), basic polyadenylic 1.25 times as likely to bind to any given empty site acid (K = 161), heparin (K = 787) and polyphoswith one dye-filled neighboring site as any given phate (K = S27).8 In addition the changes in empty site with no dye-filled neighbors. The several measures of stacking tendency] ( a l p ) , D, average K values reported herein have been cal- (PID), 6 and dE/d(P/D) follow the helix culated from eq. 4 using F values computed from coil transition in calf thymus DNA sufficiently eq. 2. Because both of these equations must be well to be used as measures of the extent of deconsidered as approximations,1g these K values naturation in a partially denatured sample. cannot be considered t o correspond exactly with (20) 0.S. Levinson, W. T. Simpson and W. Curtis, J. A m . Chem the definition of K in terms of the free energy of SOL.,79, 4314 (1957). (21) E. G. McRae and M. Kasha, J . Chsm. Phys., 88, 721 (1958); dye-dye interaction although they may be used to Rasha, Revs. Mod. Phys., 31, 162 (1959) characterize polymers and to compute approximate M.(22) I. Tinoco. J. A m . Chcm. SOL, 81, 4785 (1960). values for AF. (23) J. Marmur and P. Doty, Nature, 183, 1427 (1959)
*
E
-
(1%
D F. Bradley and
S Gcisrer, in preparation.
-
(24) C. Hall and M. Litt, J . Bmphys. Bioclrem. Cytoi., 4 , 1 (1958).
3634
OSCARGAWROIY, A. J. GLAID,111, .ITD T. P. FONDY
We have measured the stacking tendencies of twenty-four native DNA samples prepared from various biological sources by different methods and having different nucleotide sequences and compositions, molecular weights and extensions and traces of bound ions, proteins, etc. All of these samples, with the exception of the singlestranded +X DNA, show the same stacking tendency regardless of the parameter chosen as a basis of comparison. Using X-ray diffraction techniques, Langridge, et al.,25have shown that all two-stranded native DNA samples have the same molecular conformation. These data provide further support for the theory that the stacking tendency depends upon the molecular structure of the polyanion to which the dye is bound and in this sense is an intrinsic characteristic of the polyanion. Recent theories of aggregation20-22express the interaction as a function of the distance and angle between the interacting chromophores. Detailed application of these theories to the spectra of the (25) R. Langridge, W. E. Seeds, H . R. Wilson, C. W. Hooper, M. H. P. WilkinsandL. D. Hamilton, J . Binphys. Bdochem. Cytnl., 3 , 7 6 7 11957)
[CONTRIBUTION FROM THE
Vol. 83
bound dye should permit the calculation of the distance between binding sites in polymer systems where, unlike the case of DNA, this information is not otherwise available. Such calculations will of course determine the distance between sites in the dye-polymer complex which may differ to a limited extent from that of the free polymer in solution. It should be possible to use the dye method described above to determine the extent of denaturation in any given DNA sample. Under identical conditions the parameters would be expected to be within the range described above. To determine the numerical values of some of the stacking parameters of DNAs under different conditions (small changes in ionic strength, $H, temperature, etc., or with a different dye), those parameters for known native and fully denatured samples should be measured and used as standards. Acknowledgments.-The authors with to express their gratitude to Drs. A. Bendich, J. Rose, R. Sinsheimer and S. H. Stone for their gifts of DNA samples and to Drs. M. K. ITolf and 11.Kasha for helpful comments and discussions.
DEPARTMENT O F CHEMISTRY, DUQCESNE UNIVERSITY,
PITTSBURGH, PENSSYLVASIA]
Stereochemistry of Krebs' Cycle Hydrations and Related Reactions BY OSCAR GAWRON, ANDREW J. GL.\ID, 111, AND THOXAS P. F O N D Y ' RECEIVED DECEMBER 23, 1960 3-Deuterio-~-malicacid obtained from the fumarase catalyzed hydration of fumaric acid in DzO is shown to have the erythro configuration by n.m.r. comparison with stereospecifically synthesized threo-3-deuterio-~L-malicacid. The fumarase and aspartase systems thus operate by a trans mechanism as do the P-methylaspartase and cis-aconitase systems. Brewster's rules are applied to the problem of the stereochemistry of citric acid synthesized in the Krebs cycle and a configuration related to D-malic acid is arrived at. This configuration plus the ~D.,PL.configuration of d-isocitric acid permits depiction of the stereochemical pathway of the cis-aconitase system and of the reactions of the Krebs cycle. The possibility that cisaconitase exhibits a preferred direction for addition of OH, with concomitant trans addition of H, is discussed.
cycle hydrations and the related aspartase and prnethylaspartaseG reactions. The stereospecific synthesis of threo-3-deuterioDL-malic acid (11) was accomplished (Fig. 1) by a COOH trans opening7s8with lithium aluminum deuteride I HO-C-H of the oxide ring of 3,4-epoxy-2,5-dimethoxy-tetraI hydrofuran (111) followed by acid hydrolysis to the D-C-H I dialdehyde and oxidation of the dialdehyde to I COOH threo-3-deuterio-~~-mak acidg (11). stereospecifically synthesized threo-3-deuterio-~~Experimental malic acid (11). With this result the stereochemi3-Deuterio-4-hydroxy-2,5-dimethoxy-tetrahydrofwan cal mechanism of the fumarase and aspartase sys- (IV).-The deuterated tetrahydrofuran derivative was obtems was unequivocally demonstrated to be trans tained from 3,4-epoxy-2,5-dimethoxy-tetrahydrofuran7by and, in conjunction with previously obtained re- the procedure of Sheehan and Bloom7 utilizing lithium deuteride'o in place of lithium aluminum hydride. s u l t ~on~ the ~ ~ cis-aconitase system, the stereo- aluminum ___chemistry of the Krebs cycle from fumaric acid to ( 6 ) H. A. Barker, R. D. Smyth, R. X.Wilson and H. Weissbach, a-keto-glutaric acid was elucidated. J . Bioi. Chem., 284, 320 (1959). (7) J. C. Sheehan and B. >T. Bloom, J . Am. Chem. Soc., 74, 3825 It is the purpose of this paper to detail proof of the configuration of natural 3-deuterio-~-rnalic (1952). (8) P. A. Plattner, H. Heuser and M. Feurer, Helv. Chim. Acta, 31, acid and to discuss the stereochemistry of Krebs 587 (1949); L. W. Trevoy and W. G. Brown, J . Am. Chem. soc., 71, In a previous communication2 natural3 3deuterio-L-malic acid (I) was shown to have the erythro configuration by n.m.r. comparison with
(1) National Science Foundation Cooperative Graduate Fellow. (2) 0. Gawron and T. P. Fondy, J . A m . Clzem. S O L .81, , 6333 (195P). (3) Obtained by the fumarase catalyzed hydration ot fumaric acid in DnO. (4) 0. Gawron and A. J. Glaid, 111, J. A m . Chem. SOL.,77, 6638 (19,551 (5) 0. Gawron, A. J. Glaid, 111, A. LoMonte and S. Gary, ibid., 8 0 , 6866 ( l Q 5 8 ) .
1675 (1949); W. G. Dauben, R. C. Tweit and R. L. McLean, ibid., 77, 48 (1955); A. Streitweiser, Jr., R. H. Jagow, R. C. Fahey and S. Suzuki, ibid., 80, 2326 (1958). (9) Subsequent to this work, i t was found t h a t palladium catalyzed hydrogenation of cis-12-dicarboxy-ethylene oxide proceeded in a stereospecific fashion yielding Ihreo-3-deuterio-~~-malic acid by a lvans opening. 0. Gawron and T. P. Fondy, unpublished work. (10) Metal Hydrides, Inc.
