31 I t '
11.C. O T L Y ,J . P. GRISENSTEIN, bI. \\\'INITZ
ANL)
SANFOKU A I . I~IRNILIL:X
[COSTRIBUTION FROM THE LABORATORY O F BIOCHEMISTRY, NATIONAL CANCER IXSTITUTE, NATIONAL INSTITUTES UF
i T 0 1 . 77 HEALTH
Studies on Diastereoisomeric a-Amino Acids and Corresponding a-Hydroxy Acids. IV. Rotatory Dispersion of the Asymmetric a- and @-CarbonAtoms of Several Diastereoisomeric Amino Acids BY M. CLYDEOTEY,JESSE P. GREENSTEIN, MILTONWINITZAND SANFORD bl. BIRNBAUM RECEIVED DECEMBER 22, 1954 The optical rotatory dispersiori of 42 L-amino acids was studied a t 589, 578, 546, 433, 405 aiid 365 in*, and describctl f o r the most part in terms of a two constant Drude equation. The empirical relation between the rotatory characteristics of the a-amino acids and the magnitude of XO proposed by Patterson and Brode was tested in this large series of compouiids and, in modified form, was found t o be generally applicable t o amino acids with a single wasymmetric center as well as to the N asymmetric center of diastereoisomeric amino acids.
In an earlier paper in this series, the contribution of the individual asymmetric a- and w-carbon atoms to the total optical rotation a t 589 mp of several diasymmetric amino acids was determined. The present communication extends these calculations to the results of rotatory measurements at wave lengths of 589, 5TS, 546, 435, 405 and 365 trip, and, by the use of the Drude equation or appropriate modifications thereof, presents a graphical analysis of the data obtained over this wave length range. The Drude equation expresses the relation between optical rotation and wave length as a = Z k , / X ? - Xi, where a represents the rotation, k is a constant, X is the wave length of the light employed for the measurements, and X, is a series of wave lengths associated with the spectral absorption hands controlling the dispersion.? n'hen a graphical plot of 1,'a against X 2 in the visible region of the spectrum yields a straight line, the equation is ixxpressed iii its simplest or one-term form, a = k / X ? - o. If the resultant graph is not a straight line, additional terms must generally be added to express the dispersion linearly to a t least a first approximation. The a-amino acids comprise a particularly appropriate series of compounds for the study of rotatory dispersion for the rotation of the asymmetric carbon atom may be expected to be governed appreciably by the position of the anisotropic bands of the a-amino and a-carboxyl groups common t o all of these c o m p o ~ n d s .T~ h a t certain regularities emerge from the study of the optical dispersion of the a-amino acids was shown by Patterson and Brode who examined some 14 compounds a t wave lengths between 440 and 660 ~11l.l." On the basis of the graphical use of the single tertii Drude equation, these authors suggested that the a-amino acids possessed an L,-configuration if the following conditions were met: (a) the dispersion was normal, positive, and the value of XOwas above 205 mp, (b) the dispersion was normal, negative, and the value of An was below 140 mp, and (C'I the dispersion was anomalous (produced by partial rotations of opposite sign) and the sign of rotation changed from negative to positive with decreasing wave length. More recently, the late
Dr. Erwin Brand and his co-workers published rotatory dispersion data on five L-amino acids between 250 and 750 mp employing a two-term Drude e q ~ a t i o n . They ~ found in all cases that the first term of this equation was positive, and that the XO values calculated were between 220 and 290 m u . Xeither of these two groups of investigators i n cluded in their studies any of the diasymmetric CYamino acids. Before undertaking the study of the rotatory dispersion of each of the asymmetric centers of the diasymmetric a-amino acids, it was considered desirable to first examine the dispersion of the large variety of amino acids available in this Laboratory as a result of the development of enzymatic methods for the resolution of the corresponding racemate^.^ Some 42 L-amino acids of tested 01)tical purity (>99.9yo) were studied a t wave l e n g t h between 365 and 589 nip, in part to see whether the Patterson and Rrode suggestions were applicable to a wider variety of amino acids than that hitherto studied, and in part to provide a basis of comparison for the analysis of the behavior of the niorc. complex diasymmetric amino acids which was to follow. I n most cases the use of the single term Drude equation was sufficient to represent the data obtained, and the observed values of [ a ] n t 589 mp, and the calculated values of Xa from the plot of lOO/[a]against X?,are given in Table I. If the values for the diasymmetric amino acids, isoleucine, threonine, hydroxyproline, phenylserine and aminotricarballylic acid, and their respective all0 stereomers, are excepted from present consideration, i t would appear that for positive rot:itions the values of Xo range in water solution from 111 8 to 348 and in HCl solutions froni 199 to 314. For tiegative rotations in water the values of Xn include l;il for histidine, 146 for phenylalanine, and 105 for proline; in HC1 solutiori such values are 198 for cystine and 151 for proline. Several instances where Xi is negative and A, therefore imaginary occur with negatively rotating amino acids in water (leucine, t-leucine, serine, homoserine, methionine, ethionine and tryptophan). Ai similar situation occurs in the case of the positively rotating d-hydroxynorvaline in water and tryptophan in HC1. Anomalous dispersion was noted with S-benzylcysteine and phenylalanine in €&CI solution. I n nearly
(1) M.Winitz, S. M. Birnbaum and J. P. Greenstein. THISJOURNAL, 77, 716 (1955). (4) E. Brand, E. Washburn, B. F. Erlanger, E. Ellenbogen, J. ( 2 ) P. A. Levene and A. Rothen in 11. Gilman's "Organic ChemisDaniel, F. Lippmann and M.Scheu, THISJOURNAL, 76, 5037 (1964). try," Vol. 2, John Wiley and Sons, Inc., New York, N. Y., 1938, pp. 1779- 1850. (5) J. P. Greenstein, S. M. Birnbaum and M. C. Otey, J . B i d . 13) J W. 1';itterson a n d W. R . R r o d e , Arrifif, H i o r / i e ) i i , 2 , 247 1 'h?ui 204, 3118 (1953); 1. €7 C r w n ~ t e i n.I d w r i r p r i n Proteiiz Chemis, l!443) i l l 9 . 12:' ( l % i l
June 5 , 1955
ROTATORY DISPERSION OF ASYMMETRICCARRQN ATONSOF AMINO h c m s
3113
TABLE I ROTATORY DISPERSION DATAFOR L-CY-AMINO ACIDS" Amino acid
Alanine Butyrine Norvaline Xorleucine Aminoheptylic acid Aminocaprylic acid Valine Isovaline Leucine t-Leucine Isoleucine Alloisoleucitie Serine Homoserine &Hydroxynorvaline e-Hydroxyncrleucine Threonine Allothreonine Methionine Ethionine Cystine
H20 as solvent la]D inc
-t 1.8 348
+ 7.9 282 + 7.0 293 + 4.7 307 + 6.7 242 .... ... + 5.6 298 +11.2 -10.7 - 9.1 4-12.4 $15.9 - 7.8 - 8.8 6.2 3.6 -28.5 9.6 - 8.4 - 6.8
218 hag. Imag.
HCl as solvent b [a]D
be
H20 a s solvent Amino acid
[a]D
Homocystine S-Benzylcysteine S-Benzylhomocysteine Aspartic acid Glutamic acid Aminoadipic acid b-Aminoalanine 7-Aminobutyrine Ornithine Lysine Arginine Histidine Phenylalanine .... . . . Tryptophan . . . Proline y-Hydroxyproline +23.5 244 -14.5 h a g . 7-Allohydroxyproline @-Phenylserine $31.7 246 p- Allophenylserine $23.9 236 Aminotricarballylic acid +23.4 240 Alloaminotricarballylic acid -212d 198
+14.1 +20.1 +24.1 +24.2 +23.3 +23.5 $26.5 6.6 $15.6 7.4 $39.5 $39.6 +14.5
+ +
258 261 251 244 247 237 253 204 274 314 231 246 271
.... .... ....
+
AnC
...
...
...
5.2 257 +11.9 241
.... ....
...
...
HC1 as solvent b D I . [ ho c
+77.1d - 4.0d +24.8d +24.9 +31.5 +24.1 +33.1 $31.7 +27.7 $25.9 $26.9 $13.8 - 5.9
221 Anom.
226 214 228 233 199 223 226 231 203 270 Anom.
.... ... -39.2 151 -34.0 146 hag. 6.4d h a g . -33.3 h a g . hag. -54.1 151 -85.1 105 hag. -46.7 167 -76.7 192 296 -20.6 106 -60.0 181 126 -50.3 112 -33.1 68 299 $82.2 237 8.4 294 hag. -32.8 207 -48.0 204 Imag. +36.4 205 7.5 276 ,... ... Optical rotations determined with a photoelectric polarimeter a t 589, 578, 546, 435,405 and 365 mp with 0.5-2.0c/; solutions at temperaturcs between 24 and 28". Values given in the table refer only t o the D or 589 mp line of sodium. * Except where noted concentration of HCl was 5 N. c XO values determined from the dispersion data by the method of
+ + +
248 270
.
I
.
+
.
least squares. Where Xu2 was negative and XOimaginary the fact was noted by Imag. 1 N HCl. wave length fact noted by Anom. (anomalous dispersion).
+ +
Where sign of rotation changed with
every case, the XO value of a positively rotating amino acid in water is greater than that for the same positively rotating amino acid in HCl solution. Only proline furnished an appropriate example of an amino acid with a negative rotation in both water and HC1, and in this instance the XO in water is less than that in HC1. It is difficult to assign a particular physical significance to values of XO when the experimental data are obtained a t wave lengths so relatively far from the region of absorption, but i t is of interest to observe the apparent regularities associated with this figure. Thus, in accord with the suggestion of Patterson and Brode, the positively rotating L-amino acids possess XOvalues of 200 mp or more. The relatively few appropriate examples of the negatively rotating L-amino acids indicate that their values of A,, are 200 mp or less. It appears reasonable therefore to adopt a slightly modified form of the Patterson and Brode suggestion as a means of determining the optical configurations of a-amino acids, whereby the Xo value of 200 mp is set as the dividing line between positively rotating L-amino acids with XO >200 and negatively rotating L-amino acids with XO x i j!
H.0 as solvent a-Carbon atom w-Carbon atom xo Sloped An Sloped
Isoleuciite 263k 0.025 Threonine hag.' . ..., -rHydroxyyroline 189" -0,005 B-Phenylserine hag.' - ,019 An~ittotric:irl~all~lic :wit1 163' - ,024 " I h t a calculated from rl~:il)leI. 1, Positive rotation. experiincnt nl d a t a
216' 217"
-0.219
- ,018 230" - ,041 183 - ,Ol,? 218" - ,016 c Negative rotation.
ters did not follow the Lutz-Jirgensons rule. Anticipating data to be given below, it is of interest to note that these pairs also may be eliminated from consideration for their a-centers do not follow the Patterson-Brode rule either. The choice between the 1,d-allo and d,d-allo pairs of diastereomers was made tentatively on the basis of the optical rotatory behavior of the a-asymmetric center of natural d-isocitric acid which had been obtained by deamination of I-aminotricarballylic acid.* The diastereomeric pair, Z,d-allo, was that finally chosen.' However, as noted in Table I, the optical behavior of I-aminotricarballylic acid follows neither the Lutz-Jirgensons nor the Patterson-Brode rule. When the optical rotation value for this compound is referred to the partial rotations of its a - and pasynimetric centers, i t is found that not only does the a-center follow the Lutz-Jirgensons rule as noted before,' but the Patterson-Brode rule as well. The pertinent data are given in Table 11. The partial rotations of each asymmetric center of aminotricarballylic acid, isoleucine, threonine, hydroxyproline and phenylserine were calculated from the experimental values a t each wave length by the method previously described,' and XO calculated for each center from least-square slopes of the Drude equations. With one exception, the oneterm form of this equation was sufficient to represent the data. The rotation of the a-carbon atom of isoleucine in HC1 solution appeared to be best described by a two term equation cy = -1.992/X2 0.04654 1.676/X2. The partial rotations of the d-carbon atoms of isoleucine in HC1 a t the three highest wave lengths were too small in magnitude to be included in an accurate calculation of XO, and only the values obtained a t the three lowest wave lengths were therefore employed. I n addition t o values of Xo, the slopes of the Drude plots also are given in Table 11. Inspection of Table 11 indicates that the Patterson-Brode generalization is applicable to the determination of the configuration of the a-asymmetric center of the diastereoisomeric amino acids, for with the L-series studied values of XO for positively rotating a-centers were above 200 mp and those with negatively rotating centers possessed XO values less than 200 mb. In some cases, however, as in that of aminotricarballylic acid, additional evidence is necessary t o assign unequivocally an L- or
+
(8) J P Greenstein, N. Izumiva, M. W i n k and S. M. Birnbaum, THISJ O U R N A L , 7 7 , 707 (1955).
.$CiUS"
HC1 as solvent "-Carbon a t o m a-Carbon atom Sloped A0 Sloped
xo
243"
0.009 ,046 - ,009
333' 185"
-0.827 ,014
198' ,024 209" 198' - ,061 203" Least square slopes calculated
307* 150' PFI?
+
,025 ,005 ,008 from the
D-configuration to the a-center of a diastereomeric amino acid. Thus, although the 1,d-allo diastereomeric pair of aminotricarballylic acid follows in the optical behavior of its a-asymmetric center both Lutz-Jirgensons' and Patterson-Brode (Table 11) rules, so too does the d,d-all0 combination, for the XO value of the positively rotating a-center of these compounds in water and in HC1, respectively, is 218 and 203. The Xo values of the a-centers of the 1,l-allo and d,l-allo pairs in water are, respectively 218 for a negative rotation and 163 for a positive rotation, in HCI solution they are 205 for a negative rotation and 198 for a positive rotation, and thus in no case is the Patterson-Brode rule followed. All of the w-asymmetric centers described for the compounds in Table I1 possess a negative rotation, and the values of XO calculated range from 18.5 to 333. The generalization noted does not apply to optically active centers other than the a in amino acids. Consideration of the slopes described in Table I1 reveals the strikingly greater order of magnitude associated with the p-carbon of isoleucine compared with that of either a - or w-centers in the other diasymmetric amino acids studied. This is a reflection in part of the much smaller magnitude of the rotation values of the @-center in isoleucine, which a t ,589r n p are - 1 .So in water and -0.2" in HCI.
Experimental The forty-two L-amino acids used in the present investigation were all prepared in this labor at or^..^ The instrument employed was a Rudolph photoelectric polarimeter of high precision equipped with glass filters for isolating the sodium D line at 589 mp, and lines of the mercury arc spectrum at 578, 546, 435, 405 and 3fi5 mp. The actual rotations were observed by the method of symmetrical angles, with the aid of a photomultiplier tube connected to a photometer. Calibration of the instrument was effected by means of two quartz control plates, with a maximum error a t a n y of these n-ave lengths of 0.22%. Replicate readings were reproducible to f 0 . 0 0 3 " . Measurements were made on 0.5 to 2.0% solutions of the compounds in water and in HC1, and a t temperatures between 24 and 28". Solutions were prepared of the carefully dried materials in 5-ml. volumetric flasks. The readings which were taken in a 2-dm. polarimeter tube were corrected with respect to zero points obtained with solvent blanks. The rotatory dispersion of each compound required several hours of investigation, and hence such compounds as homoserine which rapidly lactonize in acid solution could not be studied in this medium. For nearly all of the compounds studied, except those noted whose dispersion was anomalous, the plot of lOO/[a]against A* was quite linear, and XOvalues could be calculated to within 1 %. BETHESDA,MARYLAND
