- Email: [email protected]
Interaction between asymmetric solutes and solvents
~OKrzzQ! of fzzfol?sur~~Q..hy,123 (1976) ~49-166 0 Elseviec ScientificPublishingCompany, _Amsterdam-
INTERACTION DIAMIDES GAS-LIQUID
BET?WEN
ASYMMETRIC
Printed in The Netherlands
SOLUTES
DERIVED FROM L-VALINE AS PARTITION CHROMATOGRAPHY
UZX BEITLER and BINYAMIN Depurrmenr of Organic Chemistry,
AND
STATIONARY
SOLVENTS
PHASES
IN
FElBUSH’ The Weizmann Institr;re of Science, Rehvot
(Isrrzel)
(Received Oerember 24th, 1975)
SUMMARY
The resolution of enantiomers of (-J-)-a-amino acid derivatives on optically active diamides derived from L-valine, RICONHCH[CH(CH3)2]C0NHR,, as gas chromatographic stationary phases was studied. The effects of the structure of R, and R2 in the stationary phases, the structure of the racemic a-amino acid derivatives, temperatnre, etc., on the separation factors are reported. In the Iight of these and related results, a more probable mechanism of resolution is presented.
mTR0mJcn0N
The gas chromatographic (GC) resolution of N-TFA”-a-amino acid esters on optically active N-acyl-dipeptide 0-alkyl esters was first reported in the period 19674970’~. Later workers tried mainly to achieve the separation of all naturally occurring ammo acids?, but none used chemical variations of the stationary phases other than to use different analogous amino acids as components of the dipeptide moiety. It was found that the selectivity of the dipeptide staticnary phases is due essentially to the N-terminal amino acid, while the C-terminal group has only a minor effect on resob.~tion~~‘~. In agreement with this observation, it was shown” that N-lauroyl-L-valine tert.-butylamide, a stationary phase that contains only one ammo acid, but includes the necessary two amide functional groups and no other polar substituents, is superior to all previously reported dipeptide stationary phases. soluteThese findings were also deducible from the “I : 1 three-hydrogen-bonded solvent associate”, regarded as the model for resolutions4*6, which will be discussed later and modified to give 2 more appropriate model. The GC resolution of enantiomers, beside its obvious practical value, elucidates stereochemical structural aspects. The resolution factors are directly related to the difference in the stereospecific interaction of the corresponding antipodic solutes * Presentadress: lntergen Research Ltd., 31 Bayit-Vegan Str., Jerusalem, Lsraei. -* TFA = trifiuoroaosyi.
tith the chirai stationary phases. Their de_oendence on structural changes mzy lead to a better understanding of the mechanism of resolution, the conformation of the interacting molecules and the nature of the aggregates in the rnelr_These aspects, in the particular system studied, involve the fit of antipodic c-amido ester solutes into the netwoik of chiralic diamide solvents, i.e., the spatial correlation of the different polar of the solute and and apolar groups -interacting through attraction or rep&ionsolvent molecules. EXPERIMENTAL
Instruments The IR spectra were measured with a Perk&Elmer
Infracord instrument, the NMR data with a Variarz A/60 instrument and the optical rotation with a PerkinElmer Model I41 polarimeter and a Schmidt & Haensch, Berlin S. GC was carried out on a Perkin-Elmer 801 chromatograph and a Varian Aerograph 1200chromatograph adjusted to operate with capillary columns. The chromatographic columns were stainless-steel capillaries, 150ft. x 0.02 in. I.D., coated with the appropriate stationary phases by the plug method with 5 ‘A dichloromethane solutions_ The temperatures used were: columns, 130” (except in a few instances specified in the text); injector and detector, 200”. The carrier gas was helium at a pressure of 10 p.s.i. Materials
Compounds I-XIX (Table I) were commercial products. The melting and boiling points given below are not corrected. Cyclooctanecarboxylic acid. I-(N-Pyrrolidyl)-l-cyclohexene was prepared according to Szmuszkovicz12in 75 ‘4 yield, b.p. 95-97” (3.5 mm). This compound was treated with acrolein to form 2-(N-pyrrolidyl)bicyclo-(3.3.1)-nonan-g-one in 55 % yield, b-p. 125-127” (0.6 mm), which was converted by reaction with methyl iodide into 2-(N-pyrrolidyl)bicyclo-(3.3.1)~nonan-g-one methiodidez3, crystallized [email protected]. 40-60°)in 80% yield, m.p. 216-217”. IJnderrefuxin20”/, sodium hydroxide solution,a40 % yield of 4-cyclooctenecarboxylic acid, b.p. 11%120”(0.5 mm), was obtained. This compound was hydrogenated in ethanol in the presence of 10% PdfC to the corresponding cycfooctanecarboxylic acid, b.p. 98-100” (0.1 mm). DipentyZacetic acid. Diethyl malonate, under reflux in methanol in the presence of 2 equiv. of &odium methoxide and 2 equiv. of 1-bromopentane, gave dimethyl dipentylmalonate, b-p. 104-105” (0.15 ITJI), which, after saponfrcation and acid%cation, gave dipentylmalonic acid, m-p. 111.5-113”. This acid was decarboxylated to yield dipentylacetic acid, b-p. 117-I 19” (O-5mm)_ 6-Un&cyZamine. This compound was prepared by the Leuckart method’j from 6-undecanone (Fluka, Buchs, Switzerland). N-Suc&jmido alkanoates (XX). A solution of 0.1 mole of the appropriate carboxylic acid and 0.1 mole of N-hydroxysuccinimide in 200 ml of dry chloroform was cooled to IO” and 0.1 mole of dicyclohexyl carbodiimide was added. The mixture was stirred for 12 h, filtered and the miltratewas concentrated under reduced pressure and diluted with dry diethyl ether. The mixture was Altered, the solvent removed under reduced pressure and the residue dispersed in light petroleum (b-p. 40&W) for 3 h.
IIWERA~ON
l3EIWEEN ASYMMETRIC SOLUTES AND SOLVENTS
151
The _precipitatewas fiBered and dried in a desiccator. The following N-succinimido
compounds were prepared in 70-90x yield: isobutyrate, m.p. t35”; pivaloate, m-p. 69-70”; tert.-butylacetate, m.p. 11 l-112”; laurate, m.p. 7_L74”; dipentylacetate, oil; cyclohexanecarboxylate, m.p. 83.5-S5.5”; cycloheptanecarboxylate, m-p. 9S-99”; cyclooctanecarboxylate, m-p. 87-89”; and I-adamantanecarboxylate, m-p. 172-174”. N-Cbz*-wdize N-succinimido ester (XXI). Was prepared in a similar manner from N-Cbz-L-valine (Miles-Yeda, Rehovot, Israel) and crystallked from diethyl ether, m-p. 113-114”, [ag -17.9” (cS in CHCL,)“. N-Cbz-L-V&E crikylomide (xxlr). A solution of 0.1 mole of XXI in 500 ml of dry chloroform was cooled in an ice-bath and 0.15 mole of triethylamine and 0.1 mole of the appropriate allqlamine were added. The mixture was stirred for 36 h at 5” and the solvent removed under reduced pressure. The residue was dissolved in 300 ml of hot ethanol and 150 ml of water were slowly added. The mixture was stirred at ream temperature for 3 h, filtered and the precipitate was washed with ethanolwater (2 :l), dispersed in water for 2 h and air dried. When necessary, the product was further purified by chromatography on silica gel (Merck, Darmstadt, G.F.R.) CORtaining an additional 6% of water, starting with n-hexane as eluent and increasing the polarity with ethyl acetate. The following N-Cbz-L-valine compounds were obtained in 75-85 % yield: isopropylamide, m-p. 180-18 1 O,[a]g - 12.2” (cS in CHCl,)**; tert.-butylamide, m-p. lOSS-109.5”, [a]$ -S-5” (cl0 in CHCI,)“; neopentylamide, m.p. 126.5-127.5”, m.p. 116.5-l 17-Y, [a]:: -21.6” (cl0 in CHC!,)“;iz-dodecylamide, CaE -13.1” (cl0 in CHC13”; 6_undecylamide, m-p. 142-143”, [a]2 -5.8” (cS in CHC&)**; cyclohexylamide, m.p. 202-204”, [ccJ,W-6.5” (cS in CHCl,)**; cycloheptylamide, m.p. 184-lS5”, [a]g - 11.6” (cS in CHCI,) l *; and cyclooctylamide, m-p. 152.5 153.5”, ]a],” -15.2” (cl0 in CHCl,)“. L-Vdike uikyhmide (XXIII). Hydrogen gas was bubbled at a low flow-rate for 4 h through an alcoholic solution of 0.2 mole of XXII containing 1.6 g of 10 % Pd/C. The cafs?lyst was filtered off and the solvent removed from the filtrate under reduced pressure to give XXIII quantitatively. It should be noted that before and after the use of hydrogen, the reaction mixture should be flushed with nitrogen. &kyk-vdine alkyhmide. A solution of 0.02 mole of XXIII in 300 ml of dry chloroform was cooled in an ice-bath, stirred and 0.025 mole of triethylamine and 0.025 mole of the appropriate XX were added. The mixture was stirred at 5” for 36 h. The solvent was removed under reduced pressure, the residue dissolved in 200 ml of ethanol and 12 ml of concentrated ammonia solution were added to destroy the excess of XX. After stirring for l/2 h, the precipitate was filtered, washed with water, dissolved in chloroform, extracted with 2 % hydrochloric acid and 5 % sodium hydrogen carbonate solution and dried on anhydrous magnesium sulphate. The compound that was recovered from an ether filtrate was crystallized from light petroleum (b-p. 40-60“). When it showed Impurities, it was chromatographed under the conditions described previously. The following compounds were prepared (see Table I). N-LuuroyEL-valine (XXrV). This compound was prepared by the following
* a2 = czsrbobenzoxy_ ** Opticalpurity ~2s not determined.
Con~pormrl
IX
VIII
VII
VI
V
1v
III
11
N-Lauroyl-L-vuline isopropylflmidc N-Lauroyl-bvnlinc rwf,.butylnmide N-Lnuroyl-bvnline neopentylumidc N-Lauroyl-t-valinc 6andccylamido N-Luuroyl-L-vnlinc cyclohexylamidc N-Lauroyl-bvnlinc cyclohcptylamido N.Lauroybvaline cyclooctylumidc N-Jauroyl-wnlinc 1-adamantylomidc
_____________ ______ ___. _.__ -. ___ N-Lauroyl-wnline amide I
phI?
StRtiONNrJ’
(“C) **
WC/3
(“/o)
pwity
oh.
[a/b
Opticd
Ct
-_- ____- --._ _-_--- -_-------.- _ . -.- -------I - 26°C IGS-167(d)” 70.1 -I- 3.6” (c,)‘*’ 126.5-128 t = 25°C 81.1 -23,8” (~10) <51 75,G t = 24°C -21,s” (c4) t = 26°C 71*s-73.5 73.2 -3&l” (ClO) Bl- 83 67.4 I c= 24°C -28.4” (c4) 161.5-162.5 73.5 I ^ 24°C -32” (CA) Mb144 77.1 t = 24°C -28.5” (c4) 113-114.5 80.2 t = 25°C -26.1” (~9) 130-144 85.0 t = 25°C -21.1” (ClO)
M.p.
_-.---
.- ._._ --_.-
Culc1r1mri
74.75
11.15
11.90
1I.60
72.80 73.33
12.25 11.85
74.31 72.58
12.25
11.91
71032 71.53
11.57
69.99
6.31
6.89
6.90
6.10 7,38
7,87
7.83
7.89
74.95
73.47
73.04
74.28 72.58
71.68
71.13
70.54
11.18
11.84
11.75
12.47 Il.65
12.03
11.94
11.84
6.48
6.86
7.10’
‘6.19 7.36
7.60
7.90
ii.23
G
F
!!!
pl #
. 1 . j/j
_-..._.-_ -_--_._ __._____ -__..___.__ _____
--__-.._.-__-.- ..-_..-_ .-_._.._
- _._.._ .____ _ _._.-____-__ _..______... _ (7%) W%) N’%) C(%) Hf%) Nf%) .. .. _.__-___-_-_____._-__^_C” -_-__-_l_._ 68.01 11.21 9.48 68.41 11.48 9.39
FO:owlfl .__” _---.
RNRlJuiS
Eh?lllelrluI
PHYSICAL PROPERTIES OF SOME OF THE STATIONARY PHASES STUDIED ,..-- _”.--..-- ....-. ----._--_-_ - _.__._--_. _.-_- ....._._-.-- --_----.--._.- ..-_._-_-._-..-___.____ _- ._-._----.- ._____.__
TABLE I
I
___.____
120-121.5
88-89.5
142-144
165.5-167
100-104
72.5-7305
77.5-78.5
124-125
113-l 14.5
-_-___,___-_--.-_-
N-Louroyl-t-vnlinc n.dodccylnmidc N”llsobutyryl-r,-vnline r~dodccylnmidc N.Pivnloyl-wnline rr-dodccylamidc N-rcrt.-Butylacetyl.Lvnlincndodccylamidc N-Dipcntylocctyl-Lvnline tr-dodecylamidc N-Cyclohcxunccarbonyltwline n-dodccylnmidc NXyclohcptunccurbonylL-vnlinewdodecylamidc N-CyclooctnuecarbonylL-vnlinen-clodccylnmide N-Adnmunlanc-l~carbonylL-vnlinen-dodccylamidc
74.78 t = 22°C --20.7” (~5) 70.96 I I- 24°C 95.8 -29.7” @IO) 71.71 94.5 I - 22°C -23” (~10) 72.02 I = 26°C 743 --38.9” (~10) 74.57 ti8.3 t = 26°C -28.9” (~10) 73.28 94.1 t = 22°C -24” (~3) 73,56 91.0 I = 22°C -22.2” (ClO) 73.G4 85.4 t = 22°C -18.9” (~10) 75.33 85.3 r = 22% - 20.4”(~6) _-.-- _ -.- _ _._II____- - __---__-. 90.1
’ (d) = decomposition. ‘* The optical purity was mcnsurcdby GC (see Experimental). “0°Solvent = CFJCO&-I
--_-
XIX
XVIII
XVII
XVI
xv
XIV
XIII
XII
x
73.88
6.40 6.18
11.75 11.12
_-_--_----_
73.47
I I.GO 6.75
11.28
11,92
II.84
11.75
12,53
12.12
12.03
Il.94
12,52
6.27
6.G3
6.86
7.10
6,(x)
7.32
7,6O
7.90
6.00
_-__---_ --- --- -_...- . . .
75.28
73.04
7.29
Il.50
74.62
72,20
71.68
71.13
74.62
6,27
7.68
7040
7,8G
5.82
12.32
Il.95
11.83
11.96
12.33
u_B_
154
B. FEEHJsl3
method15. Sodium hydroxide (0.2 mole) and L-valine (0.1 molz) wqre d&solved ih 60 ml of water, cooled to CQ. S’, stirred vigorously and 0.11 mole of iauroyl chloride added dropwise, with further stirring for 112 h. The solution was diIured with 900 til of water and acidified to congo red with 6 N hydrochloric acid. The precipitate was filtered, washed with water, air dried and recrystallized by diluting a concentrated chloroform solution with 300 ml of light petroleum (b.p. 40-60”), .rn.p: 103-104”, [c]E f 22.4” (cl0 in CHCl,)‘. iV&zuroyl-L-vai&e N-succinimido ester (XXV). This compound was obtained from XXIV as described previously; m-p. 94-95.5”, la]E -21.4” (ci0 in CHQ)*. S’;ationary phases I and IX were obtained by treatment of XXV with aminonia and adamantylamine, res_pectively. N-Lawoyl-L-@z?nino-o-piperidone (XXVI). L-#?-Amino-a-pip&done was prepared accortiing to Fischer and ZemplerP from L-ornithine, and without further purification was treated with N-succinimido laurate according to the previous procedure elemental analysis reto give XXVI, m-p. 114-115”, [Q]E + 55.6” cc5 in CHQ)‘; q*zired for C,,EI,,N,O,, C 68.87 %, I-I 10.88% and N 9.45%; found, C 68.88x, I-I 11 .ClO% and N 9.44 %_ N-TFA-(f)-aminoacidesters. These esters were prepared according to Gil-Av ei al,“. Determination of the optical purity of the diamide prodzrcts
H-jdrolysis of 50 mg of N-acylvaline alkylamide was carried out by refluxing with 20 ml of 6 N hydrochloric acid for 4 h. The hydrochloric acid was removed under reduced pressure and the residue converted into its N-TFA isopropyl ester”. This derivative was gas chromatographed on stationary phase III and the optical parity was calculated from the areas of the peaks corresponding to the derivatives of D- and L-valine. XESUJLTS AND DISCUSSION
The chiral diamide compounds
I
R*
C--N-CH*-C-N
7
----I II
T=
contain a strongly polar asymmetric centre surrounded by three alkyl substituents: an isopropyl radical and two alkyl groups, Rx and Rt (their shape and size tin be readily varied). These compounds, when used as gas chromatographic stationary phases, resolve racemic a-amino acid esters
l
Optical
purity was
rootdetermined.
INTERACITON BETWEEN ASYMMETBIC SOLUTES AND SOLVENTS
155
which have suitably disposed functional groups that are capable of interacting with . the diamide by hydrogen bonding. In this study, separation factors of the above solutes were measured on diamides derived from L-valine. This amino acid was chosen because of the high resolution factors achieved with L-valyl-L-valine derivatives, which are the analogous dipeptide stationary phases’. In order to compare the steric effect of the substituent of the N-acyl groups (RI) and the substituent at the N-amide side (Ra on resolution factors, two series of diamide stationary phases were studied (Table I): (i) N-lauroylL-vzzlinealkylamides and (ii) N-acyl-L-valine dodecylsmides. In the first series, R2 = H (I), isopropyl (II), tert.-butyl (III), necpentyl (IV), 6-undecyl (V), cyclohexyl (VI), cycloheptyl (WI), cyclooctyl (VILI), I-adamantyl (IX) and n-dodecyl (X); in the second series, RI = isopropyl (XII), tert.-butyl (XIII), neopentyl (XIV), 6-undecyl (XV), _ cyclohexyl (XVI), cycloheptyl (XVII), cyclooctyl (XVIII) and 1-adamantyl (XIX). Calcdation of separation factors on optically pure statioimry phases from separatiofz fccfors obtained with partially active stationary phases
The separation factors of racemic solutes on a given stationary phase, at a certain temperature, depend on its optical purity. The conditions used to prepare the diamide stationary phases (see experimental) were such that different degrees of racemization occurred. A comparison of the efficiencies of these stationary phases in resolving racemic solutes could not be made unless the separation factors were measured or calculated for a certain opticai purity.
Let r,(P) and r=(P) be the corrected retention times of enantiomers D and L stationary phase of optical purity” P, respectively, let Po and FL be the respective values at 100°? optical purity of the stationary phase under the same chromatographic conditions and let 4r(P) and dr* be the differences in the corrected retention times of the enantiomers on a stationary phase with optical purity P% and lCKl%, respectively. OR
a
_____ lp=
L
.-100 (XL - XD) > whereXL 2nd d’, are the appropriatemolar fractions of enantiomers
XL t XD 2nd D, respectiveiy.
If the resolution of the enantiomers on-the chiral stationary phase satisfks the following two assumptions: (a} dr(P) is-linearly dependent on P, and (b) the vah~ F(P). = l/2 [Q_(P) + r,(P)] is independent of P, then the resolution factors at 109 % optical purity, PL/ D, can be calculated by means of eqn. I (see below). These two asi sumptions are reasonable approximations in instances when the diEerenCe in the total interactions of the D and L enantiomers ~5% the solvent molecules results of all I : 1 solute-solvent bimofezxdar interaceions in an additive way; where the free energy of salvation of the L-solute resuhs from all possible L-L' and L-D’ associates * and that of the D-solute from the D-L’ and D-D’ pairs, While usu&y, interaCtiOn of Lr-E)’eqtEIk that of D-L’ associate and that of L-L’ equals that of D-D’ associate.
(a):
From
dr(P) = /Ii-* P/100 Therefore,
From (b):
Rearranging these equations into
and substituting rL[u for rL(p)/fr,(p) and ‘*L/o for r*&*n, r*L,D = ?*&*D
eqn. 1 is derived:
=
Two columns With stationary phase III of optical purities 59.3 % and 75.6 %, respectively, were prepared. Separation factor% of N-TFA-a-amino acid methyl and
l
L.’ and D’ are each
a soIveHnolewle of L-or D-co&guration, respcctive~y.
mid
h&r P L= 59.3%
Methyl ester ._.-.--_--_-----rLl,jn
rl./u
- ___.._-.-----.----A43
-----.---
Isopropyl ester ---. -._--_,_____.__-___. r* L/D rL/D for
____-.._---...... h/D for
--_
Ain cnlcularcd
___. __-
-_._ - __” _--m.-_-----
calcrrlare(l P = 75.G% ccrlct4lrre~l P = 59.3% ctrlcrrlured P = 75.6% ___ __._._ _ _ .- -_----. -_._--.- -_.-- ____-.-. . ..-_ _-_____ ___-_ _- _-._-.-_Almine 1.175 1.313 1.235 1.323 1,194 1,351 1.255 1,351 Aspartic acid 1.044 1.076 1.062 1.080 1.046 1.079 1.062 1.083 Leucinc 1.228 1.420 1.301 1.419 1.265 1.490 1.356 1.490 Mcthioninc 1.183 1.329 1.230 1.319 1.205 1,372 1.276 1.382 Phcnylali~ainc 1.170 1.304 1.222 1.304 1.193 1,349 1.265 1.347 Wine 1.129 1a233 1.171 1.232 1.172 1.309 1.225 1.308 . . -.- .__. - _----_- -.-. __ .__ .__ __._. - -.-.-I. . . -.--- .--- _ ___ . - ___-.._-.- .. -~.. .- --_ --- . . -. .- - - ______.._ .___
a-hino
-__.-.
--.-_....
.-
MEASURED AND CALCULATED CORRECXD SEPARATION FACTORS (r ,,,” AND reLIo, RESPECTIVELY) OF N.TFA-a-AMINO ACID ESTERS ON STATIONARY PHASE III AT 130” _______ ____.____ - _____ -_-I.-_--__ .--_. -- ._-_ ...-.__.-_-_- .__.-.. . -_. _.. _ .__._ .-_._--._ _____-_ _-_- __.____-__ ---*-
. TABLE II
U. BEiTLER. B. JZIBUSH
158
isopropyl esters were measured on these columns and corrected for lOO~/oopti~cxd purity of the phase according to eqn. 1. The results, which are summarized in Table II, are in good agreement and prove the usefulness of eqn. 1. Dependence of ?he separation factors, rX,jD, on the a&y1 ,tistituents RI and R2 of the chiral diamiiie solvent The nature of the alkyl substituents Rx and R2 has a~ iduence on the sepa-
ration factors, through their bulkiness and shape, which also afkct the configuration of the two amide groups, as will be discussed later. The correlation of the last aspect can elucidate the mechanism of resolution. Two series of diamides were studied: (i) stationary phases I-X, where I& is a straight-chain Cl&& group and R2 has various shapes, and (ii) stationary phases XII-XIX, where R2 is a straight-chaiu group and RI has the structures of R2 ia series (i). Using these stationary phases, resolution of N-TFA-a-amino acid esters and other compoundsi was achieved. When L stationary phases were used, the L enantiomers of ail the a-amino acid derivatives emerged last from the column. Tables III and IV summarize the correct& resolution factors, T*L/~, of the above methyl and isopropyl ester derivatives. In general, it is observed that series (i), where the branching is on the N-amide radical, leads to higher resolution factors than those obtained on series (ii); e.g., the reL,o values of N-TFA-leucine isopropyl ester on different phases (given in pareotheses) at 130” are: 1.499 (III) > 1.376 (VIII) > TABLE
III
RESOLUTION FACTORS (r* & OF N-TFA-Q-AMINO OPTICAL PURITY OF THE STATIONARY PHASES
a-Amzkoacid
Stationary phase -_ __ III IV v
~
vizz
AIanine
1.323
1.221
1.123
acid a-Amiuobutyiic acid Giutamic acid a-Aminohexzmoic acid Letxine ierr_-Leucine Methionine a-AminocxZanoic acid Phenylaianine Prolioe Serine ‘Ihreonice a-AminovzIerfc acid L Valisie
1.080
1.057
1”
l-302 1.264
1.2u 1.141
1.~ l-218 1.109 1.182
Aspvtic
1.221 1.051
.-_ IX’ 1.165 1”
ACID METHYL AT 130”
1.227
3.‘45
1.333 1.231 1.113 i-304 1.2225 1.137 1.071 1.056 1” 1.167 1.133 1.078 1.175 1.146 1.135 1.34: 1.247 1.141 -1.232 1.176 1”
l At 150c. -- No l-solution. l *- Shoulders
CALCULATED
FOR 100%
--__
x
xzz
xzzz
XfV
xv
___-_ xvzzz xzx
1.136 1.036
1.125 1”
1.130 1.044
1.262 1.033
1.098 1”
i.166 1.057
1_ISS l-112 1.142 1.123
1.114 1.090
1.344 1.236 1.132 1.248 1.175 1.13s 1.121 1.419 1.287 1.167 1.320 1.223 1.167 1.163 1” 1” 1.116 1.086 1” 1.089 1” I.319
ESTERS
kT l.li9 l-202 I.108 I.160
1.121 1”
1.152 1” 1.094 1.149
1.084 1.098
1.129 1.155 1.049
1.232 1.073 1.176 1.270 1.098 1.200 1.081 1” 1”
1.087 1.121 1”
1.206
1.096
1.233
1.176
1.132
1.123
1.135
1.249 1.249 1”’ 1.156 1.156
1.174 1.17s 1” 1.077 1”’
1.134 1.138 1” 1.080 1.086
1.128 l-122 I” 1.057 1”
1.124 1.229 1.063 1.131 l-142 l-232 l-058 l-190 1.034 1.061 1” 1.043 1.089 1.098 1.047 1.096 1.107 1.121 1.058 1.120
I.084 l-107 1” I.087 1.093
1.145 1”
1.129 1.082
1.093 1”
1.241 1.180 l-i56 1.171 1.122 I.086
1.240 1.139
1.076
1.06S 1”
1.173
1.162 1.119
INl-ERAaON
BETWJZN
ASY1MMETRlC SOLUTES AND
SOLVENTS
159
TABLE N RESOLUTION FACl-ORS (rfL,o) OF N-TFA-a-AMINO ACID ISOPROPYL 100% OPTlCAL PURIM OF THE STATIONARY PHASES AT 130” a-kfmino afid
ZZZ
ZY
v
VZZZ
zx=
1.323 1.083
1.270 1.070
1.174 1.036
1.282 1.097
1.206 1.044
1.350 1.285
1.259 1.234
1.105 1.153
1.268 1.231
1.395 1.499 1.136 1.382
1.299 1.375 1.109 1.301
1.180 1.246 1” 1.189
1.307 1.376 1.110 1.282
1.405 1.347 1.210
1.282 1.180 1.293 1.170 1.118 1.078
1_os5
Se&e 1.19a Threonine 1.239 a-Arninovaleric acid 1.391 &dine 1.308 -_ l At 150”. ** No xparztion. *** Shoulder.
x
XZZ
xzzz
XIV
xv
xvzzz
XIX
1.162 1.048
1.174 1.045
1.150 1.051
1.270 1.051
1.142 1”’
1.228 1.063
1.145 1.*3
1.179 1.168
1.143 1.137
1.144 1.136
1.126 1.122
1.212 1.169
1.068 1.106
1.181 1.181
1.119 1.109
1.206 1.251 1.079 1.199
1.154 1.216 1.050 1.148
1.167 1.223
1.144 1.190 1.064 1.146
1.237 I.329 1.110 1.218
1.102 1.172 1” 1.137
1.186 1.280 1.054 1.187
1.127 1.166 1.oss 1.181
1.337 1.203 1.304 1.201 1.121 l-OS4
I-153 1.149 1.060
1.171 1.172 1.0&t
1”
1.178
1”’
1”
I”
1”
1”’
1”
1.183 1.203
1.107 1.162
1.183 1.218
1.122 1.126
1.098 1.119
1.102 1.141
1.135 1.246 1.110 1.181 1.118 1.143 1.222 1.165 1.205 1.125 1.068 1@93 1.062 1.076 1.039 1.031 1.038 I=* 1”’ 1”’ 1.112 1.117 1.122 1.140 1.110 1.1X 1.170 1.134 1.147 1.143
1.295 1.223
1.175 I.067
1.294 1.252
l.204 1.163
1.173 1.125
1.161 1.114
1.145 1.127
1.234 1.099 1.201 1.175 1” 1.152
1.375 (IV) > 1.329 (XIV) > 1.280 (XVIII) > 1.246 (v) > 1.233 (XIX) > 1.216 (X) > 1.190 (XID) > 1.172 (XV) > 1.166 (XIX); and 1.251 (IX) at 150”. For the same derivative of methionine at 130”, the values are: 1.382 (III) > 1.301 (IV) > 1.282 (VIII) > 1.218 (XV) > 1.189 (v) = 1.187 (XVIII) > 1.181 (XIX) > 1.178 (XII) > 1.148 (X) > 1.146 (XlII) > 1.137 (XV); and 1.199 (EC) at 150“. The best resolution of the above and similar racemic solutes (see Tables IIIV), was given by stationary phase III, N-lauroyl-L-valine tert.-butylamide, which has a combination of R1 = n-undecyl for lowering its vapour pressure and melting point and R2 = be&-butyl for improved stereoselectivity. The correlation between the selectivity 2nd the type of R, and R, substituents can be used in preparing GC stationary phases that are capable of operating over a wider range of temperature without losing their ability to resolve the appropriate antipodes. Moreover, a combination of diamide stationary phases at adjusted optical purities should he considered for the complete separation of all natural and “unnatural” enantiomers of the protein amino acids. Dependknce
FOR
-Staz&nwy pFzzse
Alanine Aspartic acid a-Aminobutyric acid Glutamk acid a-Plminohexanoic acid Leucine ?ert.-Leucine Methiooine a-Aminoocfanoic acid Phenylalanine Phenylglycine PrOliX
ESTERS CALCULATED
of separatim factors, rfLID, on the substituents R3 and R4 in the racemic
sohtes In gemmxl, the a-amino acid solutes behave on the diamide stationary phases in a similar manner to that on the dipeptide stationary phases: (a) the order of emer-
1.127 1.113
160
U. BESLER,
v
TABLE
RESOLUTION
FACTORS
100% OPTICAL z-Amino acid
AI&e
Leilsinc
l
B. FEIBUSH
PURlTY
(r*uo)
-
OF
Pi-3TA-ALANWJ
A&T
-LEUCINE
ESTERS CALCULATED
FOR
AT 130”
Ester
Stationaryphase
Methyl Ethyl rr-Propyl n-B utyi Isopropyl 3-Pentyl CycIopentyl
1.323 1.221 1.123 1.331. 1.233 1.141 1.314 1.254 1.150 1.313 1.258 1.148 1.351 1.270 1.174 1.402 1.313 l.i92 1321 1.249 i-128
1.136 1.151 I.134 1.153 1.162 1.197 1.141
1.125 1.129 1.154 1.160 1.174 1.184 1.138
1.130 1.137 1.137 1.141 1.150 1.177 1.121
1.262 1.268 1.210 1.230 1.278 1.293 1.218
I.098 1.106 1.131 1.106 1.142 1.155 1.089
I.166 1.121 1.2G3 1.127 1.201 1.116 1.207 1.120 1.228 1.145 1.252 1.166 1.188 1.117
Methyl Ethyl n-Propyl n-Butyl Isoixopyi 3-i&_yl Cyclopentyl
1.419 1.287 1.167 1.320 1.223 1.167 1.431 1.323 1.198 1.337 1.235 1.181 l-433 1.329 1.2G6 1.323 !.217 1.193 1.433 1.308 1.206 1.339 1.231 1.183 1.499 1.375 1.246 l-376 i-251 1.2i6 1.576 1.422 1.264 1.435 1.298 1.246 1.467 1.341 1.200 1.337 1.223 1.195
1.163 1.175 1.1~ 1.174 l-223 1.235 1.189
1.i5.5 1.174 1.176 l.i74 I.190 1.217 1.i4i
1.270 1.293 1.~6 1.296 i-329 1.393 1.283
1.098 1.104 I.091 1.090 l-172 1.184 I.115
I.200 1.231 1.233 1.220 1.280 1.336 1.233
1.221 1.258 1.261 1.268 1.282 1.327 1.257
1.168 1.181 1.186 1.181 1.206 1.216 1.185
1.121 1.140 1.126 1.121 1.166 1.200 1.139
At 150’.
gence on the L stationary phases is L after the D enantiomers; (b) branching of the a&y1 ester group, RS, improves the separation factors (isopropyl and 3-pentyl esters in comparison with n-all@ derivatives; see Table V); the best separation factors are obtained for 3-per@, poorer for isopropyl (shortening of the side chains) and worst for cyclopentyl (rigidity); (c) branching of the a&y1 substituent, R4, on the #I-carbon lowers the separation factors; but (d) branching of the y-carbon improves the separation factors, Le., leucine derivatives have the highest rsLfD values, phenylalanine is one of the best resolved compounds, whereas phenylglycme has one of the poorest
resolution factors, etc.; (e) a decrease in r*L/,, is observed when R., includes a polar carbonyl group, in particular at the /&position, which competes for the hydrogen bonding with the a-carbonyl ester radical participating in the “selective associate” discussed later. From a study on the behavior of b- and y-amino acidsI on stationary phase V, it was found that the D enantiomers emerge after ffie L enantiomers, wh_$h is the reverse order of emergence of a-amino acid derivatives. The separation factors of y-ammo acids on stationary phase V are CCI.1.1, whereas those of #I-amino acids are Ody
CQ.
l.&%.
Association of L-aspartic and L-glutamic acids with the molecules of the stationary phase through the ccarbonyl is analogous to that of regular apolar L-aamino acids (see Fig. 1, left) and association through the @- or y-carbonyl is analogous to that of D-B- or D-y-do acids, resi>ectively, as shown in Fig. 1 (right). From the
relative disposition of the groups, it happeti that the L-&amide stationary phases interact more strongly with the miizcids-in both associative forms than with those of their D-antipodes. However, as reported above, the sekctivity of the diamide stationary phase towards the &carbonyl is small and this additional associative form strongly decreases the separation factors of aspartic acid, while the y-carboryi competing as-
161
=3
P”
-o=c H
I_-Aspartic
D-$-AA
acid
FH
>
4
“t, cl
R
OR
c
-o=
=o
3
L-~-AA*
R
m<;COa A
-H<
7
L_GMamicocid
-n
=o
D-Y-AA
Fig 1. Schematic& spatial disposition of L-aspartic and L-glutamic acid derivatives in comparison with those of L-a-AA and D-B- or D-Y-AA, respectively. The groups participating in a particularbimolecular 1:l solute-soIventassociationare underlined.AA = amino acid. sociative form, being of high stereoselectivity, does not iower the resolution factors of glutamic acid (or, similarly, the separation factors of se&e and threonine derivatives)_ It is of interest that the proline derivative, which lacks the amide hydrogen atom that participates in the “selective associate” (see the following section), is resolvable. This is also the only amino acid whose methyl ester is better resolved than its isopropyl ester derivative.
Mechanism of resolution The high stereoseiectivity
of the active diamides tow@s the resolution of racemic amino acid derivatives, in comparison with the less effictive stationary phases, N-TFA-L-c-amino acid estersI, was presumed to result from the spatial requirements of the “three-hydrogen-bonded associate”J, represented schematically ss
and its similar opposite oriented associa@. The C-N amide bond in RECONRzR’, has z-bond character, leading to separate c&- and ?rms-isomers. The activation ener,gy of their interconversion is 16-20
162
u; BEITLER, 3. FEmJSH
kcal/mcie for compounds where R1 f HZ’. It was found by NMR measurements that the &m-isomer is the exclusive form (>990/,) for compounds in which R,, RZ = alkyl and K, = 33. On ;be other hand, similar compounds in which RZ is an aromatic substituent have a great proportion of the c&isomer, which in certain instances becomes the predominant form 20_Another example is the N-alkylformamides (RI, R’, = H and RZ = alkyl), in which the &-form is present in increasing percentages in cumpound‘s that have bulky R2 alkyl groups, e.g., it reaches 20 % for R, =. fert.-buty12’. This last case is unique for the homologous series, where a hydrogen replaces the R1 group and the steric factors work in an opposite direction, in favouc of the less hindered &-isomer.
The stationary phase N-lauroyl-r-$-amino-cr-piperidone
which has the corresponding amide group only in the c&conformation, was synthesized, and was found to be incapable of resolving racetic a-amino acids. Moreover, as mentioned previously, the order of increasing seIectivity of the phases in series (i) is: X (RZ = n-dodecyl) < V (RZ = Gundecyl) < IV (R, = neopentyl] < VIII (RZ = cyciooctyl) < III (Rr = rert.-butyl); where at least III is exclusively in the transamide conformation. The difficulties in building the above “three-hydrogen-bonded selective associate” with C.P.R. (Corey, Pauling, Koltun) models confirms the experimental results that the corresponding amide bond of the solvent in the selective associate is in the trans-conformation:
Diamides of this type, when dissolved in carbon tetrachloride at high dilution (CCLlWJ M), form mixtures of two possible intramolecular hydrogen-bonded ritigs, and the folded “G’ ring, which has twq adj+& planes the planar TS” ring”” intersecting at 115” (refs. 26, 27) as well as the unassociated open structureZ-29. The
last form becomes predominant at higher temperatures27-2g. The “C5” and “C,” forms disappear at higher concentrations, where intermolecular hydrogen-bonded associative oligomers of the extended structure are formedB*30. Mizushima and co-workerP*3i studied the IR spectra of the crystalline structure of different diamides with polarized light and concluded that, for example, Nacetyf-L-leucine methyfamide (see Fig. 2.a) possesses the crystalline structure of poIyL-alaaineU, keratin-, etc. Ichikawa and Iitaka~” analyzed the X-ray pattern of the above racemic mod&z&ion, N-acetyl-n,L-leucine methylamide (see Fig. Zb) and found that the “G” site was not associated as su_gested above, but involved bonding of three mo!ecules. This structure is not possible for diamides that have large alkyl substituents, e.g., with the N-tert.-butylamide derivative. For such derivatives, the structure represented in Fig. Za is more probable.
(0)
(bl
Fig. 2 (a) The structure of N-xetyl-L-Ieucine methylamide derived from IR measurements30*3’ (R = isobutyl). (b) The structure of N-acetyl-D,L-leucine methylamide derived from X-ray measurements% (R = isobutyl).
The ass;unption that over a short range (a distance of a few moiecules) an associative oligomeric structure, similar to that in Fig. 2a, exists to a considerable extent in the melt at 130” is reasonable and agrees with ali known experimental results. This structure gives a good fit wi’h the soiutes studied, as can be seen on C.P.K.models, and explains the stereoselectivity via association, represented in Fig. 3. The substituent Rs of the solute is above the plane of the paper and syfz to the isopropyl group of the L solvent for the L' enantiomer and below the plane and anti for the D' antipode.
b4
L -D’
associate
associate
Fig. 3. The stereoselective hydrogen bonds associationof the L- and IXUXG~Oacid derivativessolvated in the active diamide phase.
The experimental results indicate that the L-L' combination, in which theisopropyl a;ld I& groups are in close proximity, have a larger difference in free energy of solvation t&n
the L-d
sistem.
The diflerence in G,,, B,, and SO qf the emnhmers in the chiral dimnide solvent The resolution factors are a direct measure of the difference in the free energy of solution : AC,
=
G,,(L) -
t&(D)
=
-RZln
r&,
=
A&-,-
TAS,,
and from its temperature dependence the other quantities are obtainable:
In PLOD=
A&
1
-R-T
E;rom the experiznenti results summarized in Table VI, the slopes [Z(ln P&/3(1/T)] for some compounds were czlculati, taking i&o account tie I&t-square deviations. - It is noticeable that AS, is independent of temperature (See Tablet WI) which
INTERA~ON
BETHEEN
ASYMMETRIC
SOLLJ
AND SOLVENTS
165
TABLE VI CORRECTED RESOLXJTXON FACTORS (r*,_,& OF N-TFA ESTERS 3F ALANINE LEUCINE &S A FUNCIlON OF TEMFERATWRE ON STATIONARY PHASE IV
AND
Leu.&iz ester
Temperature l”C)
Afmine ester Methyl
Ethyl
I..opropYr 3-Petztyl
Methyl
Ethyl
Isapropyl
3-Pen@
loo Ii0 120 130 140 150
1.339 1.293 1.254 I.221 1.179 1.139
1.3@ 1.316 1.274 1.233 1.198 1.157
1.413 1.362 1.315 1.270 1.223 1.179
I.463 1.417 i.gx5 I.287 1.236 1.190
1.510 I.446 1.391 I.318 1.265 1.X8
1.572 1.506 1.441 1.375 1.302 I.245
1.560 1.5@? 1.422 1.343 1.278
TABLE
1.525 I.4445 1.391 1.316 1.270 1.217
w
DIFFERENCE IN ENTHAL PY (&& kcaljmole), ENTROPY (A& &/mole- deg) AND FREE ENERGY OF SOLVATION (&,, kcal/mo!e)OF N-TFA ESTERS OF ALAMNE AND LEUCINE ON STATXONARY PKASE iV dg0 = rz, (L)- R. (D), d& = so (L) - & (0) and AG,, = G,J(L) - f?x,(D).
a-Amino fiter -Alan&
Leucine
MO”
110”
acid
AC,
AS,
150”
A&
AG -___
AS, ___-.
-0.995 -1.016 -1.130 -1.363
-0.196 -0.209 -0.235 -0.280
-2.iI9 -2.il -2.33 -2.S3
-0.160 -0.168 -0.192 -a.220
-2.07 -2.10 -2.33 -2.83
-0.109
Ethyl Isopropyl 3-Pen@
-0.123 -0.138 -0.165
-2.09 -2.11 -2.34 -2.83
Methyl Ethyl rsopropyl 3-Pentyl
-1.331 -1.404 - 1.474 --i.Sli
-0.265 -0.281 -0.312 -0.338
-2.80 -2.93 -3.03 -3.32
-0.202 -0.221 -0.255 -0.282
-2.80 -2.93 -3.02 -3.30
-0.146 -0.160 -0.184 -0.206
-2.80 -2.95 -3.05 -3.32
Methyl
strengthens the model of the selective association discussed previousJy, in which no steric hindrance or strain in forming the associate is observed, but is mainly due to the digerence between the disposition of the a&yl group, R3, of the antipodes relsrtive to the isopropyl radical of &e stztionary phase (see Fig. 3). REFERENCES 1 E. Gil-Av, B. Feibush and R Cbaries-Sigler, in A. B. Littlewood (Editor), Gas Cfrrornatogrqhy 1966, Institute of Petroleum, London. 1967,p. 227. 2 E. Gil-Av, B. Feibusb and R. CbariwSigler, in A. B. Littlewood (Eidtor), Gas Chromatography 1965, Institute of Petroleum, London, 1967,p. 235. 3 B. Feibush and E. Gil-Av, Tetrdedron Lett., (1967) 3345. 4 3. Feiiusfi and E. Gil-Av, Tetrakcdwn, 26 (1970) 1361. 5 S. Nakaparaslkin, P. BirreZ, E. Gil--4~ md J. Orb, /. C/w~nzzf~gr_Sci., 8 (1970) 177. 6 W. Parr, C. Yug, E. Bayer and E. Gil-Av, J. Chrornrrrogr.SC& 8 (1970) 591. 7 W. Pzrr and P. Howard, Ckromdographia, 4 (1971) 162 8 W. A. Koenig, W. Parr, H. A. Lichtinstein,E. Gayer and J. Or&,/. Chromatugr. Sci., 8 (2970) 183. 9 J. A. Carbin, I_ E. Rhoad and L. l3. Rogers, And. Chem., 43 (19771) 327_ 10 S. Wein&%, G. Jmg and E. GEAv, Proceedings of the 4Ist Anmtd Meeting of the Zsrae! Ciremical Skiety, 1971,p_ 202. 11 B. Feibus~ C&m. Cbmmun., (1971) 544.
166 _
IJ. BEITLER,
B. FETBtiSH
12 J. Stinl;szkovicz, in R. A. Raphael, E. C Taylor and H. yyaberg (Editors), Ac&znceSin Organic Chemisry, Merhodr and Reszz.& Vol. 4, Interscience,New York, London, f963, p_ 98. 13 6. Stork and H. K. Landcsmao, .F. Amer. C&/n. Sac., 78 (?956) 5129. 14 A. I. Vogel, A Tear-Bookof Organic Chemtirry, Lo-, Green, Loncbn, 3rd ed., 1956, p_ 558. 15 K. Gmhma~ and W. Parr, C&onzzrogrqxZz, 5 (1972) 18. 16 E. Fischer and G. Zemplen, Chem. i3er., 42 (1909) 4878. 17 E. Gil-Av, R. Chades-SiJer, G. Fischer and D. Nurok, J. Gus Chromcrtogr.,4 (1966) 51. 18 S. Nakapamskin, E. Gil-Av and J. Orb, /z!zzzi.B&.&vn_, 33 (1970) 374_ 19 B. Feibush and E Gil-Av, to be published. 20 W_ E. Stewart and T. EL Siddall, Chem. Rev., 70 (1970) 517. 21 L. A. Laplanche and M. T. Rogers, J. Anrer. Cfrem. Sue., 86 (1964) 337. 22 M. Tsu’wi, 23 S. M&&ha,
24 25 26 27 28 29
T. Shirnanouchi and S. Miiushi~
.i. Panzer. C&n.
Sot., 8i (1959) 14d6.
T. Shimanouchi,M. Tsuboi, T. Sugita, E. Kate and E. Kondo,J. Anzer_Chem. Sue_ 73 (i951) 1330. S. Mizushima, T. Shiox~~ouchi,M. Tsuboi andT. Arakawa,J. A.-. C&/n. Sue., 79 (1957) 5357. S. M&&ma, T_ Shimanouchi, K. Ku&a& T. Sugita, I. Nakagawa and K. Kurosaki. Nature (b&n), 169 (1952) 1058. M. T. Cung, M. Marraud md J. N&l, in E. D. Bergmano and B. Pullman(Editors), 2’7ze Jerusalem Symposium on @anrum Chemisrry and Biochemisrry, Vol. 5, israel Academy of !%zienceand Humanities, 1973, p. 69. M. Avignon and J. Lascombe, Ref. 26. p_ 97. J. Smoiikova, A. Vitek and K. Btaha, Collect. Czedr. Chem. Commwz., 36 (1971) 2474. S. Mizwhins~, T. Shimz~ouchi, M. Tsuboi, T. Sugita, K. Kurosaki, N. Mataga and R Souda, 3. Amer. Chem. Sot., 74 (1952) 439.
30 S. Mizwshixx~, T. Shimanouchi, M_ Tsuboi, K. Kurztani, T. Sugh, M. Matags uld R Souda, L Amer. Chem. Sac., 75 (1953) 1863. 31 T. hforiwaki, M. Tsuboi, T. Sbimanouchi and S. Mizushima, I. Amer. Chem_ Sot., Sl (1959) 5914. 32 C. H. Ramford, -4. Elliott, W. Hanby and I. Trotter, Narure (Loruion), 173 (i954) 27. 33 W. T. Astbury, Advan. Enzymol., 3 (1943) 63. 34 T. Ichikawa and Y. fitaka, Acra Crysraflogr., 3325(1969) 1824.
INTERACTION DIAMIDES GAS-LIQUID
BET?WEN
ASYMMETRIC
Printed in The Netherlands
SOLUTES
DERIVED FROM L-VALINE AS PARTITION CHROMATOGRAPHY
UZX BEITLER and BINYAMIN Depurrmenr of Organic Chemistry,
AND
STATIONARY
SOLVENTS
PHASES
IN
FElBUSH’ The Weizmann Institr;re of Science, Rehvot
(Isrrzel)
(Received Oerember 24th, 1975)
SUMMARY
The resolution of enantiomers of (-J-)-a-amino acid derivatives on optically active diamides derived from L-valine, RICONHCH[CH(CH3)2]C0NHR,, as gas chromatographic stationary phases was studied. The effects of the structure of R, and R2 in the stationary phases, the structure of the racemic a-amino acid derivatives, temperatnre, etc., on the separation factors are reported. In the Iight of these and related results, a more probable mechanism of resolution is presented.
mTR0mJcn0N
The gas chromatographic (GC) resolution of N-TFA”-a-amino acid esters on optically active N-acyl-dipeptide 0-alkyl esters was first reported in the period 19674970’~. Later workers tried mainly to achieve the separation of all naturally occurring ammo acids?, but none used chemical variations of the stationary phases other than to use different analogous amino acids as components of the dipeptide moiety. It was found that the selectivity of the dipeptide staticnary phases is due essentially to the N-terminal amino acid, while the C-terminal group has only a minor effect on resob.~tion~~‘~. In agreement with this observation, it was shown” that N-lauroyl-L-valine tert.-butylamide, a stationary phase that contains only one ammo acid, but includes the necessary two amide functional groups and no other polar substituents, is superior to all previously reported dipeptide stationary phases. soluteThese findings were also deducible from the “I : 1 three-hydrogen-bonded solvent associate”, regarded as the model for resolutions4*6, which will be discussed later and modified to give 2 more appropriate model. The GC resolution of enantiomers, beside its obvious practical value, elucidates stereochemical structural aspects. The resolution factors are directly related to the difference in the stereospecific interaction of the corresponding antipodic solutes * Presentadress: lntergen Research Ltd., 31 Bayit-Vegan Str., Jerusalem, Lsraei. -* TFA = trifiuoroaosyi.
tith the chirai stationary phases. Their de_oendence on structural changes mzy lead to a better understanding of the mechanism of resolution, the conformation of the interacting molecules and the nature of the aggregates in the rnelr_These aspects, in the particular system studied, involve the fit of antipodic c-amido ester solutes into the netwoik of chiralic diamide solvents, i.e., the spatial correlation of the different polar of the solute and and apolar groups -interacting through attraction or rep&ionsolvent molecules. EXPERIMENTAL
Instruments The IR spectra were measured with a Perk&Elmer
Infracord instrument, the NMR data with a Variarz A/60 instrument and the optical rotation with a PerkinElmer Model I41 polarimeter and a Schmidt & Haensch, Berlin S. GC was carried out on a Perkin-Elmer 801 chromatograph and a Varian Aerograph 1200chromatograph adjusted to operate with capillary columns. The chromatographic columns were stainless-steel capillaries, 150ft. x 0.02 in. I.D., coated with the appropriate stationary phases by the plug method with 5 ‘A dichloromethane solutions_ The temperatures used were: columns, 130” (except in a few instances specified in the text); injector and detector, 200”. The carrier gas was helium at a pressure of 10 p.s.i. Materials
Compounds I-XIX (Table I) were commercial products. The melting and boiling points given below are not corrected. Cyclooctanecarboxylic acid. I-(N-Pyrrolidyl)-l-cyclohexene was prepared according to Szmuszkovicz12in 75 ‘4 yield, b.p. 95-97” (3.5 mm). This compound was treated with acrolein to form 2-(N-pyrrolidyl)bicyclo-(3.3.1)-nonan-g-one in 55 % yield, b-p. 125-127” (0.6 mm), which was converted by reaction with methyl iodide into 2-(N-pyrrolidyl)bicyclo-(3.3.1)~nonan-g-one methiodidez3, crystallized [email protected]. 40-60°)in 80% yield, m.p. 216-217”. IJnderrefuxin20”/, sodium hydroxide solution,a40 % yield of 4-cyclooctenecarboxylic acid, b.p. 11%120”(0.5 mm), was obtained. This compound was hydrogenated in ethanol in the presence of 10% PdfC to the corresponding cycfooctanecarboxylic acid, b.p. 98-100” (0.1 mm). DipentyZacetic acid. Diethyl malonate, under reflux in methanol in the presence of 2 equiv. of &odium methoxide and 2 equiv. of 1-bromopentane, gave dimethyl dipentylmalonate, b-p. 104-105” (0.15 ITJI), which, after saponfrcation and acid%cation, gave dipentylmalonic acid, m-p. 111.5-113”. This acid was decarboxylated to yield dipentylacetic acid, b-p. 117-I 19” (O-5mm)_ 6-Un&cyZamine. This compound was prepared by the Leuckart method’j from 6-undecanone (Fluka, Buchs, Switzerland). N-Suc&jmido alkanoates (XX). A solution of 0.1 mole of the appropriate carboxylic acid and 0.1 mole of N-hydroxysuccinimide in 200 ml of dry chloroform was cooled to IO” and 0.1 mole of dicyclohexyl carbodiimide was added. The mixture was stirred for 12 h, filtered and the miltratewas concentrated under reduced pressure and diluted with dry diethyl ether. The mixture was Altered, the solvent removed under reduced pressure and the residue dispersed in light petroleum (b-p. 40&W) for 3 h.
IIWERA~ON
l3EIWEEN ASYMMETRIC SOLUTES AND SOLVENTS
151
The _precipitatewas fiBered and dried in a desiccator. The following N-succinimido
compounds were prepared in 70-90x yield: isobutyrate, m.p. t35”; pivaloate, m-p. 69-70”; tert.-butylacetate, m.p. 11 l-112”; laurate, m.p. 7_L74”; dipentylacetate, oil; cyclohexanecarboxylate, m.p. 83.5-S5.5”; cycloheptanecarboxylate, m-p. 9S-99”; cyclooctanecarboxylate, m-p. 87-89”; and I-adamantanecarboxylate, m-p. 172-174”. N-Cbz*-wdize N-succinimido ester (XXI). Was prepared in a similar manner from N-Cbz-L-valine (Miles-Yeda, Rehovot, Israel) and crystallked from diethyl ether, m-p. 113-114”, [ag -17.9” (cS in CHCL,)“. N-Cbz-L-V&E crikylomide (xxlr). A solution of 0.1 mole of XXI in 500 ml of dry chloroform was cooled in an ice-bath and 0.15 mole of triethylamine and 0.1 mole of the appropriate allqlamine were added. The mixture was stirred for 36 h at 5” and the solvent removed under reduced pressure. The residue was dissolved in 300 ml of hot ethanol and 150 ml of water were slowly added. The mixture was stirred at ream temperature for 3 h, filtered and the precipitate was washed with ethanolwater (2 :l), dispersed in water for 2 h and air dried. When necessary, the product was further purified by chromatography on silica gel (Merck, Darmstadt, G.F.R.) CORtaining an additional 6% of water, starting with n-hexane as eluent and increasing the polarity with ethyl acetate. The following N-Cbz-L-valine compounds were obtained in 75-85 % yield: isopropylamide, m-p. 180-18 1 O,[a]g - 12.2” (cS in CHCl,)**; tert.-butylamide, m-p. lOSS-109.5”, [a]$ -S-5” (cl0 in CHCI,)“; neopentylamide, m.p. 126.5-127.5”, m.p. 116.5-l 17-Y, [a]:: -21.6” (cl0 in CHC!,)“;iz-dodecylamide, CaE -13.1” (cl0 in CHC13”; 6_undecylamide, m-p. 142-143”, [a]2 -5.8” (cS in CHC&)**; cyclohexylamide, m.p. 202-204”, [ccJ,W-6.5” (cS in CHCl,)**; cycloheptylamide, m.p. 184-lS5”, [a]g - 11.6” (cS in CHCI,) l *; and cyclooctylamide, m-p. 152.5 153.5”, ]a],” -15.2” (cl0 in CHCl,)“. L-Vdike uikyhmide (XXIII). Hydrogen gas was bubbled at a low flow-rate for 4 h through an alcoholic solution of 0.2 mole of XXII containing 1.6 g of 10 % Pd/C. The cafs?lyst was filtered off and the solvent removed from the filtrate under reduced pressure to give XXIII quantitatively. It should be noted that before and after the use of hydrogen, the reaction mixture should be flushed with nitrogen. &kyk-vdine alkyhmide. A solution of 0.02 mole of XXIII in 300 ml of dry chloroform was cooled in an ice-bath, stirred and 0.025 mole of triethylamine and 0.025 mole of the appropriate XX were added. The mixture was stirred at 5” for 36 h. The solvent was removed under reduced pressure, the residue dissolved in 200 ml of ethanol and 12 ml of concentrated ammonia solution were added to destroy the excess of XX. After stirring for l/2 h, the precipitate was filtered, washed with water, dissolved in chloroform, extracted with 2 % hydrochloric acid and 5 % sodium hydrogen carbonate solution and dried on anhydrous magnesium sulphate. The compound that was recovered from an ether filtrate was crystallized from light petroleum (b-p. 40-60“). When it showed Impurities, it was chromatographed under the conditions described previously. The following compounds were prepared (see Table I). N-LuuroyEL-valine (XXrV). This compound was prepared by the following
* a2 = czsrbobenzoxy_ ** Opticalpurity ~2s not determined.
Con~pormrl
IX
VIII
VII
VI
V
1v
III
11
N-Lauroyl-L-vuline isopropylflmidc N-Lauroyl-bvnlinc rwf,.butylnmide N-Lnuroyl-bvnline neopentylumidc N-Lauroyl-t-valinc 6andccylamido N-Luuroyl-L-vnlinc cyclohexylamidc N-Lauroyl-bvnlinc cyclohcptylamido N.Lauroybvaline cyclooctylumidc N-Jauroyl-wnlinc 1-adamantylomidc
_____________ ______ ___. _.__ -. ___ N-Lauroyl-wnline amide I
phI?
StRtiONNrJ’
(“C) **
WC/3
(“/o)
pwity
oh.
[a/b
Opticd
Ct
-_- ____- --._ _-_--- -_-------.- _ . -.- -------I - 26°C IGS-167(d)” 70.1 -I- 3.6” (c,)‘*’ 126.5-128 t = 25°C 81.1 -23,8” (~10) <51 75,G t = 24°C -21,s” (c4) t = 26°C 71*s-73.5 73.2 -3&l” (ClO) Bl- 83 67.4 I c= 24°C -28.4” (c4) 161.5-162.5 73.5 I ^ 24°C -32” (CA) Mb144 77.1 t = 24°C -28.5” (c4) 113-114.5 80.2 t = 25°C -26.1” (~9) 130-144 85.0 t = 25°C -21.1” (ClO)
M.p.
_-.---
.- ._._ --_.-
Culc1r1mri
74.75
11.15
11.90
1I.60
72.80 73.33
12.25 11.85
74.31 72.58
12.25
11.91
71032 71.53
11.57
69.99
6.31
6.89
6.90
6.10 7,38
7,87
7.83
7.89
74.95
73.47
73.04
74.28 72.58
71.68
71.13
70.54
11.18
11.84
11.75
12.47 Il.65
12.03
11.94
11.84
6.48
6.86
7.10’
‘6.19 7.36
7.60
7.90
ii.23
G
F
!!!
pl #
. 1 . j/j
_-..._.-_ -_--_._ __._____ -__..___.__ _____
--__-.._.-__-.- ..-_..-_ .-_._.._
- _._.._ .____ _ _._.-____-__ _..______... _ (7%) W%) N’%) C(%) Hf%) Nf%) .. .. _.__-___-_-_____._-__^_C” -_-__-_l_._ 68.01 11.21 9.48 68.41 11.48 9.39
FO:owlfl .__” _---.
RNRlJuiS
Eh?lllelrluI
PHYSICAL PROPERTIES OF SOME OF THE STATIONARY PHASES STUDIED ,..-- _”.--..-- ....-. ----._--_-_ - _.__._--_. _.-_- ....._._-.-- --_----.--._.- ..-_._-_-._-..-___.____ _- ._-._----.- ._____.__
TABLE I
I
___.____
120-121.5
88-89.5
142-144
165.5-167
100-104
72.5-7305
77.5-78.5
124-125
113-l 14.5
-_-___,___-_--.-_-
N-Louroyl-t-vnlinc n.dodccylnmidc N”llsobutyryl-r,-vnline r~dodccylnmidc N.Pivnloyl-wnline rr-dodccylamidc N-rcrt.-Butylacetyl.Lvnlincndodccylamidc N-Dipcntylocctyl-Lvnline tr-dodecylamidc N-Cyclohcxunccarbonyltwline n-dodccylnmidc NXyclohcptunccurbonylL-vnlinewdodecylamidc N-CyclooctnuecarbonylL-vnlinen-clodccylnmide N-Adnmunlanc-l~carbonylL-vnlinen-dodccylamidc
74.78 t = 22°C --20.7” (~5) 70.96 I I- 24°C 95.8 -29.7” @IO) 71.71 94.5 I - 22°C -23” (~10) 72.02 I = 26°C 743 --38.9” (~10) 74.57 ti8.3 t = 26°C -28.9” (~10) 73.28 94.1 t = 22°C -24” (~3) 73,56 91.0 I = 22°C -22.2” (ClO) 73.G4 85.4 t = 22°C -18.9” (~10) 75.33 85.3 r = 22% - 20.4”(~6) _-.-- _ -.- _ _._II____- - __---__-. 90.1
’ (d) = decomposition. ‘* The optical purity was mcnsurcdby GC (see Experimental). “0°Solvent = CFJCO&-I
--_-
XIX
XVIII
XVII
XVI
xv
XIV
XIII
XII
x
73.88
6.40 6.18
11.75 11.12
_-_--_----_
73.47
I I.GO 6.75
11.28
11,92
II.84
11.75
12,53
12.12
12.03
Il.94
12,52
6.27
6.G3
6.86
7.10
6,(x)
7.32
7,6O
7.90
6.00
_-__---_ --- --- -_...- . . .
75.28
73.04
7.29
Il.50
74.62
72,20
71.68
71.13
74.62
6,27
7.68
7040
7,8G
5.82
12.32
Il.95
11.83
11.96
12.33
u_B_
154
B. FEEHJsl3
method15. Sodium hydroxide (0.2 mole) and L-valine (0.1 molz) wqre d&solved ih 60 ml of water, cooled to CQ. S’, stirred vigorously and 0.11 mole of iauroyl chloride added dropwise, with further stirring for 112 h. The solution was diIured with 900 til of water and acidified to congo red with 6 N hydrochloric acid. The precipitate was filtered, washed with water, air dried and recrystallized by diluting a concentrated chloroform solution with 300 ml of light petroleum (b.p. 40-60”), .rn.p: 103-104”, [c]E f 22.4” (cl0 in CHCl,)‘. iV&zuroyl-L-vai&e N-succinimido ester (XXV). This compound was obtained from XXIV as described previously; m-p. 94-95.5”, la]E -21.4” (ci0 in CHQ)*. S’;ationary phases I and IX were obtained by treatment of XXV with aminonia and adamantylamine, res_pectively. N-Lawoyl-L-@z?nino-o-piperidone (XXVI). L-#?-Amino-a-pip&done was prepared accortiing to Fischer and ZemplerP from L-ornithine, and without further purification was treated with N-succinimido laurate according to the previous procedure elemental analysis reto give XXVI, m-p. 114-115”, [Q]E + 55.6” cc5 in CHQ)‘; q*zired for C,,EI,,N,O,, C 68.87 %, I-I 10.88% and N 9.45%; found, C 68.88x, I-I 11 .ClO% and N 9.44 %_ N-TFA-(f)-aminoacidesters. These esters were prepared according to Gil-Av ei al,“. Determination of the optical purity of the diamide prodzrcts
H-jdrolysis of 50 mg of N-acylvaline alkylamide was carried out by refluxing with 20 ml of 6 N hydrochloric acid for 4 h. The hydrochloric acid was removed under reduced pressure and the residue converted into its N-TFA isopropyl ester”. This derivative was gas chromatographed on stationary phase III and the optical parity was calculated from the areas of the peaks corresponding to the derivatives of D- and L-valine. XESUJLTS AND DISCUSSION
The chiral diamide compounds
I
R*
C--N-CH*-C-N
7
----I II
T=
contain a strongly polar asymmetric centre surrounded by three alkyl substituents: an isopropyl radical and two alkyl groups, Rx and Rt (their shape and size tin be readily varied). These compounds, when used as gas chromatographic stationary phases, resolve racemic a-amino acid esters
l
Optical
purity was
rootdetermined.
INTERACITON BETWEEN ASYMMETBIC SOLUTES AND SOLVENTS
155
which have suitably disposed functional groups that are capable of interacting with . the diamide by hydrogen bonding. In this study, separation factors of the above solutes were measured on diamides derived from L-valine. This amino acid was chosen because of the high resolution factors achieved with L-valyl-L-valine derivatives, which are the analogous dipeptide stationary phases’. In order to compare the steric effect of the substituent of the N-acyl groups (RI) and the substituent at the N-amide side (Ra on resolution factors, two series of diamide stationary phases were studied (Table I): (i) N-lauroylL-vzzlinealkylamides and (ii) N-acyl-L-valine dodecylsmides. In the first series, R2 = H (I), isopropyl (II), tert.-butyl (III), necpentyl (IV), 6-undecyl (V), cyclohexyl (VI), cycloheptyl (WI), cyclooctyl (VILI), I-adamantyl (IX) and n-dodecyl (X); in the second series, RI = isopropyl (XII), tert.-butyl (XIII), neopentyl (XIV), 6-undecyl (XV), _ cyclohexyl (XVI), cycloheptyl (XVII), cyclooctyl (XVIII) and 1-adamantyl (XIX). Calcdation of separation factors on optically pure statioimry phases from separatiofz fccfors obtained with partially active stationary phases
The separation factors of racemic solutes on a given stationary phase, at a certain temperature, depend on its optical purity. The conditions used to prepare the diamide stationary phases (see experimental) were such that different degrees of racemization occurred. A comparison of the efficiencies of these stationary phases in resolving racemic solutes could not be made unless the separation factors were measured or calculated for a certain opticai purity.
Let r,(P) and r=(P) be the corrected retention times of enantiomers D and L stationary phase of optical purity” P, respectively, let Po and FL be the respective values at 100°? optical purity of the stationary phase under the same chromatographic conditions and let 4r(P) and dr* be the differences in the corrected retention times of the enantiomers on a stationary phase with optical purity P% and lCKl%, respectively. OR
a
_____ lp=
L
.-100 (XL - XD) > whereXL 2nd d’, are the appropriatemolar fractions of enantiomers
XL t XD 2nd D, respectiveiy.
If the resolution of the enantiomers on-the chiral stationary phase satisfks the following two assumptions: (a} dr(P) is-linearly dependent on P, and (b) the vah~ F(P). = l/2 [Q_(P) + r,(P)] is independent of P, then the resolution factors at 109 % optical purity, PL/ D, can be calculated by means of eqn. I (see below). These two asi sumptions are reasonable approximations in instances when the diEerenCe in the total interactions of the D and L enantiomers ~5% the solvent molecules results of all I : 1 solute-solvent bimofezxdar interaceions in an additive way; where the free energy of salvation of the L-solute resuhs from all possible L-L' and L-D’ associates * and that of the D-solute from the D-L’ and D-D’ pairs, While usu&y, interaCtiOn of Lr-E)’eqtEIk that of D-L’ associate and that of L-L’ equals that of D-D’ associate.
(a):
From
dr(P) = /Ii-* P/100 Therefore,
From (b):
Rearranging these equations into
and substituting rL[u for rL(p)/fr,(p) and ‘*L/o for r*&*n, r*L,D = ?*&*D
eqn. 1 is derived:
=
Two columns With stationary phase III of optical purities 59.3 % and 75.6 %, respectively, were prepared. Separation factor% of N-TFA-a-amino acid methyl and
l
L.’ and D’ are each
a soIveHnolewle of L-or D-co&guration, respcctive~y.
mid
h&r P L= 59.3%
Methyl ester ._.-.--_--_-----rLl,jn
rl./u
- ___.._-.-----.----A43
-----.---
Isopropyl ester ---. -._--_,_____.__-___. r* L/D rL/D for
____-.._---...... h/D for
--_
Ain cnlcularcd
___. __-
-_._ - __” _--m.-_-----
calcrrlare(l P = 75.G% ccrlct4lrre~l P = 59.3% ctrlcrrlured P = 75.6% ___ __._._ _ _ .- -_----. -_._--.- -_.-- ____-.-. . ..-_ _-_____ ___-_ _- _-._-.-_Almine 1.175 1.313 1.235 1.323 1,194 1,351 1.255 1,351 Aspartic acid 1.044 1.076 1.062 1.080 1.046 1.079 1.062 1.083 Leucinc 1.228 1.420 1.301 1.419 1.265 1.490 1.356 1.490 Mcthioninc 1.183 1.329 1.230 1.319 1.205 1,372 1.276 1.382 Phcnylali~ainc 1.170 1.304 1.222 1.304 1.193 1,349 1.265 1.347 Wine 1.129 1a233 1.171 1.232 1.172 1.309 1.225 1.308 . . -.- .__. - _----_- -.-. __ .__ .__ __._. - -.-.-I. . . -.--- .--- _ ___ . - ___-.._-.- .. -~.. .- --_ --- . . -. .- - - ______.._ .___
a-hino
-__.-.
--.-_....
.-
MEASURED AND CALCULATED CORRECXD SEPARATION FACTORS (r ,,,” AND reLIo, RESPECTIVELY) OF N.TFA-a-AMINO ACID ESTERS ON STATIONARY PHASE III AT 130” _______ ____.____ - _____ -_-I.-_--__ .--_. -- ._-_ ...-.__.-_-_- .__.-.. . -_. _.. _ .__._ .-_._--._ _____-_ _-_- __.____-__ ---*-
. TABLE II
U. BEiTLER. B. JZIBUSH
158
isopropyl esters were measured on these columns and corrected for lOO~/oopti~cxd purity of the phase according to eqn. 1. The results, which are summarized in Table II, are in good agreement and prove the usefulness of eqn. 1. Dependence of ?he separation factors, rX,jD, on the a&y1 ,tistituents RI and R2 of the chiral diamiiie solvent The nature of the alkyl substituents Rx and R2 has a~ iduence on the sepa-
ration factors, through their bulkiness and shape, which also afkct the configuration of the two amide groups, as will be discussed later. The correlation of the last aspect can elucidate the mechanism of resolution. Two series of diamides were studied: (i) stationary phases I-X, where I& is a straight-chain Cl&& group and R2 has various shapes, and (ii) stationary phases XII-XIX, where R2 is a straight-chaiu group and RI has the structures of R2 ia series (i). Using these stationary phases, resolution of N-TFA-a-amino acid esters and other compoundsi was achieved. When L stationary phases were used, the L enantiomers of ail the a-amino acid derivatives emerged last from the column. Tables III and IV summarize the correct& resolution factors, T*L/~, of the above methyl and isopropyl ester derivatives. In general, it is observed that series (i), where the branching is on the N-amide radical, leads to higher resolution factors than those obtained on series (ii); e.g., the reL,o values of N-TFA-leucine isopropyl ester on different phases (given in pareotheses) at 130” are: 1.499 (III) > 1.376 (VIII) > TABLE
III
RESOLUTION FACTORS (r* & OF N-TFA-Q-AMINO OPTICAL PURITY OF THE STATIONARY PHASES
a-Amzkoacid
Stationary phase -_ __ III IV v
~
vizz
AIanine
1.323
1.221
1.123
acid a-Amiuobutyiic acid Giutamic acid a-Aminohexzmoic acid Letxine ierr_-Leucine Methionine a-AminocxZanoic acid Phenylaianine Prolioe Serine ‘Ihreonice a-AminovzIerfc acid L Valisie
1.080
1.057
1”
l-302 1.264
1.2u 1.141
1.~ l-218 1.109 1.182
Aspvtic
1.221 1.051
.-_ IX’ 1.165 1”
ACID METHYL AT 130”
1.227
3.‘45
1.333 1.231 1.113 i-304 1.2225 1.137 1.071 1.056 1” 1.167 1.133 1.078 1.175 1.146 1.135 1.34: 1.247 1.141 -1.232 1.176 1”
l At 150c. -- No l-solution. l *- Shoulders
CALCULATED
FOR 100%
--__
x
xzz
xzzz
XfV
xv
___-_ xvzzz xzx
1.136 1.036
1.125 1”
1.130 1.044
1.262 1.033
1.098 1”
i.166 1.057
1_ISS l-112 1.142 1.123
1.114 1.090
1.344 1.236 1.132 1.248 1.175 1.13s 1.121 1.419 1.287 1.167 1.320 1.223 1.167 1.163 1” 1” 1.116 1.086 1” 1.089 1” I.319
ESTERS
kT l.li9 l-202 I.108 I.160
1.121 1”
1.152 1” 1.094 1.149
1.084 1.098
1.129 1.155 1.049
1.232 1.073 1.176 1.270 1.098 1.200 1.081 1” 1”
1.087 1.121 1”
1.206
1.096
1.233
1.176
1.132
1.123
1.135
1.249 1.249 1”’ 1.156 1.156
1.174 1.17s 1” 1.077 1”’
1.134 1.138 1” 1.080 1.086
1.128 l-122 I” 1.057 1”
1.124 1.229 1.063 1.131 l-142 l-232 l-058 l-190 1.034 1.061 1” 1.043 1.089 1.098 1.047 1.096 1.107 1.121 1.058 1.120
I.084 l-107 1” I.087 1.093
1.145 1”
1.129 1.082
1.093 1”
1.241 1.180 l-i56 1.171 1.122 I.086
1.240 1.139
1.076
1.06S 1”
1.173
1.162 1.119
INl-ERAaON
BETWJZN
ASY1MMETRlC SOLUTES AND
SOLVENTS
159
TABLE N RESOLUTION FACl-ORS (rfL,o) OF N-TFA-a-AMINO ACID ISOPROPYL 100% OPTlCAL PURIM OF THE STATIONARY PHASES AT 130” a-kfmino afid
ZZZ
ZY
v
VZZZ
zx=
1.323 1.083
1.270 1.070
1.174 1.036
1.282 1.097
1.206 1.044
1.350 1.285
1.259 1.234
1.105 1.153
1.268 1.231
1.395 1.499 1.136 1.382
1.299 1.375 1.109 1.301
1.180 1.246 1” 1.189
1.307 1.376 1.110 1.282
1.405 1.347 1.210
1.282 1.180 1.293 1.170 1.118 1.078
1_os5
Se&e 1.19a Threonine 1.239 a-Arninovaleric acid 1.391 &dine 1.308 -_ l At 150”. ** No xparztion. *** Shoulder.
x
XZZ
xzzz
XIV
xv
xvzzz
XIX
1.162 1.048
1.174 1.045
1.150 1.051
1.270 1.051
1.142 1”’
1.228 1.063
1.145 1.*3
1.179 1.168
1.143 1.137
1.144 1.136
1.126 1.122
1.212 1.169
1.068 1.106
1.181 1.181
1.119 1.109
1.206 1.251 1.079 1.199
1.154 1.216 1.050 1.148
1.167 1.223
1.144 1.190 1.064 1.146
1.237 I.329 1.110 1.218
1.102 1.172 1” 1.137
1.186 1.280 1.054 1.187
1.127 1.166 1.oss 1.181
1.337 1.203 1.304 1.201 1.121 l-OS4
I-153 1.149 1.060
1.171 1.172 1.0&t
1”
1.178
1”’
1”
I”
1”
1”’
1”
1.183 1.203
1.107 1.162
1.183 1.218
1.122 1.126
1.098 1.119
1.102 1.141
1.135 1.246 1.110 1.181 1.118 1.143 1.222 1.165 1.205 1.125 1.068 1@93 1.062 1.076 1.039 1.031 1.038 I=* 1”’ 1”’ 1.112 1.117 1.122 1.140 1.110 1.1X 1.170 1.134 1.147 1.143
1.295 1.223
1.175 I.067
1.294 1.252
l.204 1.163
1.173 1.125
1.161 1.114
1.145 1.127
1.234 1.099 1.201 1.175 1” 1.152
1.375 (IV) > 1.329 (XIV) > 1.280 (XVIII) > 1.246 (v) > 1.233 (XIX) > 1.216 (X) > 1.190 (XID) > 1.172 (XV) > 1.166 (XIX); and 1.251 (IX) at 150”. For the same derivative of methionine at 130”, the values are: 1.382 (III) > 1.301 (IV) > 1.282 (VIII) > 1.218 (XV) > 1.189 (v) = 1.187 (XVIII) > 1.181 (XIX) > 1.178 (XII) > 1.148 (X) > 1.146 (XlII) > 1.137 (XV); and 1.199 (EC) at 150“. The best resolution of the above and similar racemic solutes (see Tables IIIV), was given by stationary phase III, N-lauroyl-L-valine tert.-butylamide, which has a combination of R1 = n-undecyl for lowering its vapour pressure and melting point and R2 = be&-butyl for improved stereoselectivity. The correlation between the selectivity 2nd the type of R, and R, substituents can be used in preparing GC stationary phases that are capable of operating over a wider range of temperature without losing their ability to resolve the appropriate antipodes. Moreover, a combination of diamide stationary phases at adjusted optical purities should he considered for the complete separation of all natural and “unnatural” enantiomers of the protein amino acids. Dependknce
FOR
-Staz&nwy pFzzse
Alanine Aspartic acid a-Aminobutyric acid Glutamk acid a-Plminohexanoic acid Leucine ?ert.-Leucine Methiooine a-Aminoocfanoic acid Phenylalanine Phenylglycine PrOliX
ESTERS CALCULATED
of separatim factors, rfLID, on the substituents R3 and R4 in the racemic
sohtes In gemmxl, the a-amino acid solutes behave on the diamide stationary phases in a similar manner to that on the dipeptide stationary phases: (a) the order of emer-
1.127 1.113
160
U. BESLER,
v
TABLE
RESOLUTION
FACTORS
100% OPTICAL z-Amino acid
AI&e
Leilsinc
l
B. FEIBUSH
PURlTY
(r*uo)
-
OF
Pi-3TA-ALANWJ
A&T
-LEUCINE
ESTERS CALCULATED
FOR
AT 130”
Ester
Stationaryphase
Methyl Ethyl rr-Propyl n-B utyi Isopropyl 3-Pentyl CycIopentyl
1.323 1.221 1.123 1.331. 1.233 1.141 1.314 1.254 1.150 1.313 1.258 1.148 1.351 1.270 1.174 1.402 1.313 l.i92 1321 1.249 i-128
1.136 1.151 I.134 1.153 1.162 1.197 1.141
1.125 1.129 1.154 1.160 1.174 1.184 1.138
1.130 1.137 1.137 1.141 1.150 1.177 1.121
1.262 1.268 1.210 1.230 1.278 1.293 1.218
I.098 1.106 1.131 1.106 1.142 1.155 1.089
I.166 1.121 1.2G3 1.127 1.201 1.116 1.207 1.120 1.228 1.145 1.252 1.166 1.188 1.117
Methyl Ethyl n-Propyl n-Butyl Isoixopyi 3-i&_yl Cyclopentyl
1.419 1.287 1.167 1.320 1.223 1.167 1.431 1.323 1.198 1.337 1.235 1.181 l-433 1.329 1.2G6 1.323 !.217 1.193 1.433 1.308 1.206 1.339 1.231 1.183 1.499 1.375 1.246 l-376 i-251 1.2i6 1.576 1.422 1.264 1.435 1.298 1.246 1.467 1.341 1.200 1.337 1.223 1.195
1.163 1.175 1.1~ 1.174 l-223 1.235 1.189
1.i5.5 1.174 1.176 l.i74 I.190 1.217 1.i4i
1.270 1.293 1.~6 1.296 i-329 1.393 1.283
1.098 1.104 I.091 1.090 l-172 1.184 I.115
I.200 1.231 1.233 1.220 1.280 1.336 1.233
1.221 1.258 1.261 1.268 1.282 1.327 1.257
1.168 1.181 1.186 1.181 1.206 1.216 1.185
1.121 1.140 1.126 1.121 1.166 1.200 1.139
At 150’.
gence on the L stationary phases is L after the D enantiomers; (b) branching of the a&y1 ester group, RS, improves the separation factors (isopropyl and 3-pentyl esters in comparison with n-all@ derivatives; see Table V); the best separation factors are obtained for 3-per@, poorer for isopropyl (shortening of the side chains) and worst for cyclopentyl (rigidity); (c) branching of the a&y1 substituent, R4, on the #I-carbon lowers the separation factors; but (d) branching of the y-carbon improves the separation factors, Le., leucine derivatives have the highest rsLfD values, phenylalanine is one of the best resolved compounds, whereas phenylglycme has one of the poorest
resolution factors, etc.; (e) a decrease in r*L/,, is observed when R., includes a polar carbonyl group, in particular at the /&position, which competes for the hydrogen bonding with the a-carbonyl ester radical participating in the “selective associate” discussed later. From a study on the behavior of b- and y-amino acidsI on stationary phase V, it was found that the D enantiomers emerge after ffie L enantiomers, wh_$h is the reverse order of emergence of a-amino acid derivatives. The separation factors of y-ammo acids on stationary phase V are CCI.1.1, whereas those of #I-amino acids are Ody
CQ.
l.&%.
Association of L-aspartic and L-glutamic acids with the molecules of the stationary phase through the ccarbonyl is analogous to that of regular apolar L-aamino acids (see Fig. 1, left) and association through the @- or y-carbonyl is analogous to that of D-B- or D-y-do acids, resi>ectively, as shown in Fig. 1 (right). From the
relative disposition of the groups, it happeti that the L-&amide stationary phases interact more strongly with the miizcids-in both associative forms than with those of their D-antipodes. However, as reported above, the sekctivity of the diamide stationary phase towards the &carbonyl is small and this additional associative form strongly decreases the separation factors of aspartic acid, while the y-carboryi competing as-
161
=3
P”
-o=c H
I_-Aspartic
D-$-AA
acid
FH
>
4
“t, cl
R
OR
c
-o=
=o
3
L-~-AA*
R
m<;COa A
-H<
7
L_GMamicocid
-n
=o
D-Y-AA
Fig 1. Schematic& spatial disposition of L-aspartic and L-glutamic acid derivatives in comparison with those of L-a-AA and D-B- or D-Y-AA, respectively. The groups participating in a particularbimolecular 1:l solute-soIventassociationare underlined.AA = amino acid. sociative form, being of high stereoselectivity, does not iower the resolution factors of glutamic acid (or, similarly, the separation factors of se&e and threonine derivatives)_ It is of interest that the proline derivative, which lacks the amide hydrogen atom that participates in the “selective associate” (see the following section), is resolvable. This is also the only amino acid whose methyl ester is better resolved than its isopropyl ester derivative.
Mechanism of resolution The high stereoseiectivity
of the active diamides tow@s the resolution of racemic amino acid derivatives, in comparison with the less effictive stationary phases, N-TFA-L-c-amino acid estersI, was presumed to result from the spatial requirements of the “three-hydrogen-bonded associate”J, represented schematically ss
and its similar opposite oriented associa@. The C-N amide bond in RECONRzR’, has z-bond character, leading to separate c&- and ?rms-isomers. The activation ener,gy of their interconversion is 16-20
162
u; BEITLER, 3. FEmJSH
kcal/mcie for compounds where R1 f HZ’. It was found by NMR measurements that the &m-isomer is the exclusive form (>990/,) for compounds in which R,, RZ = alkyl and K, = 33. On ;be other hand, similar compounds in which RZ is an aromatic substituent have a great proportion of the c&isomer, which in certain instances becomes the predominant form 20_Another example is the N-alkylformamides (RI, R’, = H and RZ = alkyl), in which the &-form is present in increasing percentages in cumpound‘s that have bulky R2 alkyl groups, e.g., it reaches 20 % for R, =. fert.-buty12’. This last case is unique for the homologous series, where a hydrogen replaces the R1 group and the steric factors work in an opposite direction, in favouc of the less hindered &-isomer.
The stationary phase N-lauroyl-r-$-amino-cr-piperidone
which has the corresponding amide group only in the c&conformation, was synthesized, and was found to be incapable of resolving racetic a-amino acids. Moreover, as mentioned previously, the order of increasing seIectivity of the phases in series (i) is: X (RZ = n-dodecyl) < V (RZ = Gundecyl) < IV (R, = neopentyl] < VIII (RZ = cyciooctyl) < III (Rr = rert.-butyl); where at least III is exclusively in the transamide conformation. The difficulties in building the above “three-hydrogen-bonded selective associate” with C.P.R. (Corey, Pauling, Koltun) models confirms the experimental results that the corresponding amide bond of the solvent in the selective associate is in the trans-conformation:
Diamides of this type, when dissolved in carbon tetrachloride at high dilution (CCLlWJ M), form mixtures of two possible intramolecular hydrogen-bonded ritigs, and the folded “G’ ring, which has twq adj+& planes the planar TS” ring”” intersecting at 115” (refs. 26, 27) as well as the unassociated open structureZ-29. The
last form becomes predominant at higher temperatures27-2g. The “C5” and “C,” forms disappear at higher concentrations, where intermolecular hydrogen-bonded associative oligomers of the extended structure are formedB*30. Mizushima and co-workerP*3i studied the IR spectra of the crystalline structure of different diamides with polarized light and concluded that, for example, Nacetyf-L-leucine methyfamide (see Fig. 2.a) possesses the crystalline structure of poIyL-alaaineU, keratin-, etc. Ichikawa and Iitaka~” analyzed the X-ray pattern of the above racemic mod&z&ion, N-acetyl-n,L-leucine methylamide (see Fig. Zb) and found that the “G” site was not associated as su_gested above, but involved bonding of three mo!ecules. This structure is not possible for diamides that have large alkyl substituents, e.g., with the N-tert.-butylamide derivative. For such derivatives, the structure represented in Fig. Za is more probable.
(0)
(bl
Fig. 2 (a) The structure of N-xetyl-L-Ieucine methylamide derived from IR measurements30*3’ (R = isobutyl). (b) The structure of N-acetyl-D,L-leucine methylamide derived from X-ray measurements% (R = isobutyl).
The ass;unption that over a short range (a distance of a few moiecules) an associative oligomeric structure, similar to that in Fig. 2a, exists to a considerable extent in the melt at 130” is reasonable and agrees with ali known experimental results. This structure gives a good fit wi’h the soiutes studied, as can be seen on C.P.K.models, and explains the stereoselectivity via association, represented in Fig. 3. The substituent Rs of the solute is above the plane of the paper and syfz to the isopropyl group of the L solvent for the L' enantiomer and below the plane and anti for the D' antipode.
b4
L -D’
associate
associate
Fig. 3. The stereoselective hydrogen bonds associationof the L- and IXUXG~Oacid derivativessolvated in the active diamide phase.
The experimental results indicate that the L-L' combination, in which theisopropyl a;ld I& groups are in close proximity, have a larger difference in free energy of solvation t&n
the L-d
sistem.
The diflerence in G,,, B,, and SO qf the emnhmers in the chiral dimnide solvent The resolution factors are a direct measure of the difference in the free energy of solution : AC,
=
G,,(L) -
t&(D)
=
-RZln
r&,
=
A&-,-
TAS,,
and from its temperature dependence the other quantities are obtainable:
In PLOD=
A&
1
-R-T
E;rom the experiznenti results summarized in Table VI, the slopes [Z(ln P&/3(1/T)] for some compounds were czlculati, taking i&o account tie I&t-square deviations. - It is noticeable that AS, is independent of temperature (See Tablet WI) which
INTERA~ON
BETHEEN
ASYMMETRIC
SOLLJ
AND SOLVENTS
165
TABLE VI CORRECTED RESOLXJTXON FACTORS (r*,_,& OF N-TFA ESTERS 3F ALANINE LEUCINE &S A FUNCIlON OF TEMFERATWRE ON STATIONARY PHASE IV
AND
Leu.&iz ester
Temperature l”C)
Afmine ester Methyl
Ethyl
I..opropYr 3-Petztyl
Methyl
Ethyl
Isapropyl
3-Pen@
loo Ii0 120 130 140 150
1.339 1.293 1.254 I.221 1.179 1.139
1.3@ 1.316 1.274 1.233 1.198 1.157
1.413 1.362 1.315 1.270 1.223 1.179
I.463 1.417 i.gx5 I.287 1.236 1.190
1.510 I.446 1.391 I.318 1.265 1.X8
1.572 1.506 1.441 1.375 1.302 I.245
1.560 1.5@? 1.422 1.343 1.278
TABLE
1.525 I.4445 1.391 1.316 1.270 1.217
w
DIFFERENCE IN ENTHAL PY (&& kcaljmole), ENTROPY (A& &/mole- deg) AND FREE ENERGY OF SOLVATION (&,, kcal/mo!e)OF N-TFA ESTERS OF ALAMNE AND LEUCINE ON STATXONARY PKASE iV dg0 = rz, (L)- R. (D), d& = so (L) - & (0) and AG,, = G,J(L) - f?x,(D).
a-Amino fiter -Alan&
Leucine
MO”
110”
acid
AC,
AS,
150”
A&
AG -___
AS, ___-.
-0.995 -1.016 -1.130 -1.363
-0.196 -0.209 -0.235 -0.280
-2.iI9 -2.il -2.33 -2.S3
-0.160 -0.168 -0.192 -a.220
-2.07 -2.10 -2.33 -2.83
-0.109
Ethyl Isopropyl 3-Pen@
-0.123 -0.138 -0.165
-2.09 -2.11 -2.34 -2.83
Methyl Ethyl rsopropyl 3-Pentyl
-1.331 -1.404 - 1.474 --i.Sli
-0.265 -0.281 -0.312 -0.338
-2.80 -2.93 -3.03 -3.32
-0.202 -0.221 -0.255 -0.282
-2.80 -2.93 -3.02 -3.30
-0.146 -0.160 -0.184 -0.206
-2.80 -2.95 -3.05 -3.32
Methyl
strengthens the model of the selective association discussed previousJy, in which no steric hindrance or strain in forming the associate is observed, but is mainly due to the digerence between the disposition of the a&yl group, R3, of the antipodes relsrtive to the isopropyl radical of &e stztionary phase (see Fig. 3). REFERENCES 1 E. Gil-Av, B. Feibush and R Cbaries-Sigler, in A. B. Littlewood (Editor), Gas Cfrrornatogrqhy 1966, Institute of Petroleum, London. 1967,p. 227. 2 E. Gil-Av, B. Feibusb and R. CbariwSigler, in A. B. Littlewood (Eidtor), Gas Chromatography 1965, Institute of Petroleum, London, 1967,p. 235. 3 B. Feibush and E. Gil-Av, Tetrdedron Lett., (1967) 3345. 4 3. Feiiusfi and E. Gil-Av, Tetrakcdwn, 26 (1970) 1361. 5 S. Nakaparaslkin, P. BirreZ, E. Gil--4~ md J. Orb, /. C/w~nzzf~gr_Sci., 8 (1970) 177. 6 W. Parr, C. Yug, E. Bayer and E. Gil-Av, J. Chrornrrrogr.SC& 8 (1970) 591. 7 W. Pzrr and P. Howard, Ckromdographia, 4 (1971) 162 8 W. A. Koenig, W. Parr, H. A. Lichtinstein,E. Gayer and J. Or&,/. Chromatugr. Sci., 8 (2970) 183. 9 J. A. Carbin, I_ E. Rhoad and L. l3. Rogers, And. Chem., 43 (19771) 327_ 10 S. Wein&%, G. Jmg and E. GEAv, Proceedings of the 4Ist Anmtd Meeting of the Zsrae! Ciremical Skiety, 1971,p_ 202. 11 B. Feibus~ C&m. Cbmmun., (1971) 544.
166 _
IJ. BEITLER,
B. FETBtiSH
12 J. Stinl;szkovicz, in R. A. Raphael, E. C Taylor and H. yyaberg (Editors), Ac&znceSin Organic Chemisry, Merhodr and Reszz.& Vol. 4, Interscience,New York, London, f963, p_ 98. 13 6. Stork and H. K. Landcsmao, .F. Amer. C&/n. Sac., 78 (?956) 5129. 14 A. I. Vogel, A Tear-Bookof Organic Chemtirry, Lo-, Green, Loncbn, 3rd ed., 1956, p_ 558. 15 K. Gmhma~ and W. Parr, C&onzzrogrqxZz, 5 (1972) 18. 16 E. Fischer and G. Zemplen, Chem. i3er., 42 (1909) 4878. 17 E. Gil-Av, R. Chades-SiJer, G. Fischer and D. Nurok, J. Gus Chromcrtogr.,4 (1966) 51. 18 S. Nakapamskin, E. Gil-Av and J. Orb, /z!zzzi.B&.&vn_, 33 (1970) 374_ 19 B. Feibush and E Gil-Av, to be published. 20 W_ E. Stewart and T. EL Siddall, Chem. Rev., 70 (1970) 517. 21 L. A. Laplanche and M. T. Rogers, J. Anrer. Cfrem. Sue., 86 (1964) 337. 22 M. Tsu’wi, 23 S. M&&ha,
24 25 26 27 28 29
T. Shirnanouchi and S. Miiushi~
.i. Panzer. C&n.
Sot., 8i (1959) 14d6.
T. Shimanouchi,M. Tsuboi, T. Sugita, E. Kate and E. Kondo,J. Anzer_Chem. Sue_ 73 (i951) 1330. S. Mizushima, T. Shiox~~ouchi,M. Tsuboi andT. Arakawa,J. A.-. C&/n. Sue., 79 (1957) 5357. S. M&&ma, T_ Shimanouchi, K. Ku&a& T. Sugita, I. Nakagawa and K. Kurosaki. Nature (b&n), 169 (1952) 1058. M. T. Cung, M. Marraud md J. N&l, in E. D. Bergmano and B. Pullman(Editors), 2’7ze Jerusalem Symposium on @anrum Chemisrry and Biochemisrry, Vol. 5, israel Academy of !%zienceand Humanities, 1973, p. 69. M. Avignon and J. Lascombe, Ref. 26. p_ 97. J. Smoiikova, A. Vitek and K. Btaha, Collect. Czedr. Chem. Commwz., 36 (1971) 2474. S. Mizwhins~, T. Shimz~ouchi, M. Tsuboi, T. Sugita, K. Kurosaki, N. Mataga and R Souda, 3. Amer. Chem. Sot., 74 (1952) 439.
30 S. Mizwshixx~, T. Shimanouchi, M_ Tsuboi, K. Kurztani, T. Sugh, M. Matags uld R Souda, L Amer. Chem. Sac., 75 (1953) 1863. 31 T. hforiwaki, M. Tsuboi, T. Sbimanouchi and S. Mizushima, I. Amer. Chem_ Sot., Sl (1959) 5914. 32 C. H. Ramford, -4. Elliott, W. Hanby and I. Trotter, Narure (Loruion), 173 (i954) 27. 33 W. T. Astbury, Advan. Enzymol., 3 (1943) 63. 34 T. Ichikawa and Y. fitaka, Acra Crysraflogr., 3325(1969) 1824.