2902 COMMUNICATIONS TO THE EDITOR Vol.

80
Ethyl acetate was then added dropwise until the unreacted
lithium aluminum hydride was decomposed. Several
milliliters of 50% NaOH was added, and then water was
added dropivise until the solid hydroxides became granular
in appearance. The ether Iayer was poured off and the
solid residue was washed with ether. The combined ethe-
real solution was evaporated and the residue was sublimed
on the steam-bath. White needles of 4-( l-hydroxy-l-
methylethyl)-piperidine were obtained in about 32 G/o yield.
Analyses and physical constants are given in Table 11,
item 5 .
Products from d,Z-Phonic Acid and d,Z-Ethyl Pinonate.-
The d,l-pinonic acid, m.p. 105-106, used in the reaction is
reported to be the cis isomer.' This material was used to
prepare d,Z-ethyl pinonate, b.p. 129' ( 5 mm.), %*OD 1.4532,
by direct esterification in carbon tetrachloride using p-tolu-
enesulfonic acid as a catalyst. The ester is believed to be
a mixture of cis and Imns isomers, although this was not
established (see Argus3)).
The material from the reaction of ethyl pinonate was an
amber colored liquid after removal of the solvent in vacuum.
Distillation usually gave a small forecut of unreacted ester
with little still residue. The main fraction was a colorless
liquid, dZo 1.0120, @D 1.4701; ~VD found 60.9, calcd. 61.11.
Hydrolysis of Materials from d,l-Phonic Acid or its Ester.
-Fifty grams of distilled material from the ester (a) or an
equivalent amount of material, based on organic acidity,
from pinonic acid (b) were placed in a flask with 45 ml. of
concentrated hydrochloric acid and 150 ml. of water. The
mixture was refluxed 16 lir. and extracted with ether to re-
move any non-basic material. The aqueous phase was then
evaporated to near dryness under aspirator vacuum. Dilu-
tion of the residue with acetone resulted in the precipitation
of a slightly less than quantitative yield, taking into account
the S-methylamide originally present, of the hydrochloride
of 2,2-dimethy1-3-amiuocyclobutaneacetic acid and methyl-
amine hydrochloride which can be purified by dissolving in a
small amount of water or aqueous 50y0 alcohol and repre-
cipitating with acetone. Physical constants and analyses
are listed in Table 11, item 6.
Evaporation of the ether extracts from three hydrolyses
of (a) and distillation of the residue gave three fractions.
The first, b.p. 77", was characterized by odor as ethyl ace-
tate, the second was principally acetic acid; and a third,
11 g., b.p. 185-198" (1 mm.), neut. equiv. 97. The viscous
distillate was slow to crystallize and was believed to be a
mixture of cis- and trans-d,l-pinic acids. A few drops dis-
solved in acetone gave a colorless precipitate, m.p. 151-
153,5', by the addition of a few drops of dicyclohexy1a:nine.
A salt, m.p. 162.6-153.8', was prepared similarly fron an
authentic sample of a mixture of cis- and frum-d.l-pinic
acid, neut. equiv. 97, obtained by the hypohalitc oxidation
of a pure cis-d,l-pinonic acid. A mixture of the salts melted
at 152-153.8'. From this evidence the third fraction must
be principally pinic acid and the yield is approxilmately 9
mole per cent. The ether extracts from hydrolyses of (b)
gave the same results except no ethyl acetate was obtained.
Saponification of either (a) or (b) with an excess of 6 N
sodium hydroxide liberated 10 to 12 mole per cent. of meth-
ylamine which was recovered by steam distillation and
characterized by its p-nitrobenzamide. Acidification of the
hydrolysate with sulfuric acid and distilling liberated 85 to
90 mole per cent. of acetic acid which was converted to
acetanilide for identification. With both the amine and
acetic acid dry salts were prepared in order to concentrate
the materials for use in making the derivatives. The
yields, however, are based upon titration of di quots of each
of the distillates.
Hydrolyses of both (a) and (b) also were carried out using
an equivalent amount of 3 N sulfuric acid in place of the
hydrochloric acid. The resulting hydrolysate; after ether
extraction, contained the sulfates of the amino acid and
methylamine. However, these could not be isolated in the
solid form.
d,l-2,2-Dimethyl-3-aminocyclobutaneacetic Acid.-The
amino acid d,l~2,2-dimethyl-3-aminocyclobutaneacetic acid
was prepared fr ~m the crude hydrochloride by use of an ion
exchange resin." Evaporation of the eluent resulted in a
thick paste having a slight odor of methylamine. Dilution
with acetone brought about precipitation of crystalline d, l -
2,2-dimethyl-3-aminocyclobutaneacetic acid in 90% yield.
This is presumably the cis isomer. No evidence of a second
isomer has been observed.
IVhen sulfuric acid was used for hydrolysis the aqueous
solution was treated with sufficient barium hydroxide to
precipitate the sulfate ion present. After removal of the
barium sulfate by filtration, the amino acid was isolated as
above.
Purification of the acid can be achieved by dissolving in a
small amount of water, treating with activated charcoal and
reprecipitating with acetone or by sublimation at the melt-
ing point. At this temperature, however, polymerization
also occurs.
The amino acid reacts with nitrous acid at room tempera-
ture, liberating a gas which is undoubtedly nitrogen. No
attempt was made to characterize the products from this
reaction.
The a-naphthylurea and acetamide of the amino acid were
prepared by treating with the appropriate reagent. The
urea was filtered to remove insoliibles and was isolated from
the filtrate by acidification. Recrystallization from 50%
aqueous ethanol gave a neutral equivalent of 326 which
is the theoretical value. The 9-bromophenacyl ester was
prepared from the acetamide by treating an alkaline soh-
tion with p-bromophenacyl bromide. This product sepa-
rated out as formed and proved to be identical to the p-
bromophenacyl ester formed directly from the crude liquid
residue isolated from the reaction of d,l-pinonic acid and
hydrazoic acid. The physical constants and analyses of the
amino acid and its derivatives are given in Table 11, itenis
7 , 8 and 9.
(11) C Y. hleyers and L. E. Miller, Org. Syi i t heses, 32, 13 (19529.
OLUSTEE, FLA.
COMMUNI CATI ONS TO THE EDI TOR
RACEMIZATION BY THE DICYCLOHEXYLCARBO-
DI I MI DE METHOD OF PEPTI DE SYNTHESI S
Sir :
N,N'-Dicyclohexylcarbodiniide was introduced
by Sheehan and Hess' in 1955 as a peptide-forming
reagent. I ts ready availability, simplicity of use
under mild conditions with frequent good yields of
peptides, and the apparent lack of racemizatioii i n
the formation of optically active peptide derivatix es
has resulted in widespread use. lVe have found
that racemization can occur, and recommend cau-
tion in the use of this reagent. I n harmony with
our results, which are based on actual separation of
(I ) J . C Sheehan and G. P. Hess. THIS JOURNAL, 77, 1007 (1955).
isomers, enzymatic evidence has recently been given
for racemization in the synthesis of another pep-
tide.2
I n 1952, a sensitive test for racemizatioti was de-
vised in our l ab~ratori es.~ This involved the re-
action of carbobenzoxyglycyl-L-phenylalanine with
ethyl glycinate, and subsequent fractional crystal-
lization of the tripeptide product from a 2% solu-
tion in ethanol. The racemic form crystallizes
first, and careful removal of fractions from time to
(2) K, Hofmann, M. E. Woolner, G. Spiihler and E. T. Schwartr,
ibid., 80, 1486 (1958).
(3) G. W. Anderson and R. W. Young, i bi d. , 74, 5307 (1952);
G. W. Andersen, J. Blodinger and 4. D. Welcher, i bi d. , 74, 5309
1952); J . R. Vaughsn, J r., i bi d. , 74, 0137 (1932).
J une 5, 1958 COMMUNICATIONS TO THE EDITOR 2903
time can detect the presence of less than 1%.
Some of the L-form will then subsequently crystal-
lize, the remainder on concentration. The melting
point and optical rotation of this material are not
absolute criteria of optical purity, since other im-
purities may be present. This may account for
the lack of detection of racemization in the syn-
thesis of the same peptide by Sheehan and Hess.
Following the conditions of Sheehan and Hess'
as closely as possible, we treated 0.010 molar quan-
tities of carbobenzoxygl ycyl - L- phenyl al ani ne (care-
fully recrystallized from water; m.p. 127.5'
(sharp), [ a I z 4 ~ +38.8" * 0.5" ( c , 5; ethano1)j4
and ethyl glycinate (freshly distilled) in the pres-
ence of 0.01 1 mole of N,N'-dicydohexylcarbodi-
imideb in 50 ml. of dry tetrahydrofuran at room
temperature. The temperature of the mixture
spontaneously rose to 37", then fell. After four
hours, acetic acid was added to decompose excess
reagent. Dicyclohexylurea was removed here and
after concentrating the filtrate in vacuo to about 10
ml. (89.4y0 yield). Addition of 50 ml. of water
and chilling precipitated the tripeptide; it was
washed with 2 X 10 ml. of water, 10 ml. of 5%
potassium bicarbonate solution, then 2 X 10 ml.
of water. The crude dry yield was quantitative,
and the m.p. 106-109. A 2% solution of the
product in absolute alcohol gave crystallization on
refrigeration (0' ). During several hours, fractions
of the DL-tripeptide, m.p. in the 129-133" range
(the pure compound melts at 132-133') amounting
to 6.6% were collected. After a fraction of a few
mg. in the 120-130" range, the L-form began to
appear ; concentration of the filtrates gave a total of
76% yield, n1.p. 116.5-119.5, [ a I z 6 D - 11.5", ( c, 2,
ethan~l ).~.' Two repetitions of the synthesis gave
yields of 7.6% DL, 74% L and 8.2% DL, 69% L.
For comparison, the reaction was performed in 7 ml.
of tetrahydrofuran with tetraethyl pyrophosphite
as the reagent, and at reflux temperature for 30
minutes, giving 4.39 g. of crude product, m.p.
118.3-119.5'. Crystallization from solution
in ethanol gave no DL form, and 4.24 g. (96y0 yield)
of L-, m.p. 120-120.5, [cu]~'D - 13.2" ( c, 2, ethanol).
An experiment with dicyclohexylcarbodiimide
at -5" for 48 hours gave 0.5% DL and 75% L of the
tripeptide, and another in methylene chloride at
room temperature gave 12% DL and 75% L.
Thus, temperature and solvent are factors in race-
mization.
(4) K. Hofmann and M. Bergmann. J . Biol. Chem. , 134, 225
( 5 ) Purchased from Aldrich Chemical Co.
( 6) Rotations reported here were obtained by W. Fulmor and
(7) Ref. (1) gives m.p. 118-110, [CT]~'D -13.5' [ethanol].
(1940), give m.p. 125-126O, [ a ] * ' D +38.5" (c. 5, EtOH).
staff. The statistical deviation for the tripeptide is &l o .
ORGANIC CHEMICAL RESEARCH DEPARTMENT
RESEARCH DIVISION GEORGE W. ANDERSON
AMERICAN CYANAMID COMPANY FRANCIS M. CALLAHAN
PEARL RIVER, NEW YORR
RECEIVED APRIL 25, 1958
THE SYNTHESIS OF DIHYDROCONESSINE. A
METHOD FOR FUNCTIONALIZING STEROIDS
AT Cia
Sir:
The selective introduction of a functional group
in place of hydrogen at a carbon atom not directly
joined to labilizing centers such as C=O, C=C,
etc., poses an interesting challenge in synthetic
chemistry, which is magnified by the rife occur-
rence of such transformations in Nature under the
influence of enzymes. I n the field of steroids the
problem assumess pecial trenchency with regard to
the CIS (C/D fusion) angular methyl group be-
cause functionality at this position is a special
feature of the important hormone, aldosterone, and
of the Holarrhena alkaloids. ' This communication
describes an efficient and selective method for
functionalizing C18, the free radical chain decompo-
sition of an N-chloro-20-aminosteroid in acid solu-
tion, and the illustrative synthesis of dihydro-
conessine (IV).
3p-Acetoxy-20a-aminopregnene-5 acetate2 (I,
R =AcO, Rz =NHYHOAC) was formylated to
give 3,R-acetoxy-20a-formamidopregnene-5 (I, R
-66', found: C: 74.32; HI 9.87, which was hy-
drolyzed to 3,R-hydroxy-20a-formamidopregnene-5,
m.p. 229-230' (dec.), [ a I z 4 ~ -70.4', found: C,
76.31; H, 10.08; N, 4.32. Treatment of the
hydroxyformamide with p-toluenesulfonyl chlo-
ride-pyridine yielded crude 3,R-tosyloxy-2Oa-iso-
cyanopregnene-5, which readily was hydrated to
3~-tosyloxy-20a-formamidopregnene-5 (I , R =
found: C, 69.92; HI 8.32. 3,R-Dimethylamino-20ar-
formamidopregnene-5, m.p. 226-230' (dec.), [aI z7~
=AcO, Rs =NHCHO), m.p. 191-193', [CL]'~D
OTS, RS =NHCHO), m.p. 132-133', [aI z7D -47.4',
I II
m r?
-52', found: C, 77.31; H, 10.90; N, 7.23, ob-
tained from the tosylate and dimethylamine, was re-
duced (LiAlH4) to 3,R-dimethylamino-20a-methyl-
aminopregnene-5, m.p. 123-124.5', [aI z7~ -37',
found: C, 80.23; H, 11.84; N, 7.78, which on hy-
drogenation gave 3,R-dimethylarnino-20au-methyl-
aminoallopregnane (11) , m.p. 103.5-1 04.5', [a] 2 7 ~
+27.5', found: C, 79.23; H, 12.OG; N, 7.80.
20-N-Chloro-11, prepared using N-chlorosuccini-
mide in ether,4 upon irradiation (in 90% H2S04)
with ultraviolet light was converted to 3,R-dimethyl-
amino- 18- chloro- 20a-methylaminoallo -pregnane5
(1) See R. Tschesche and A. C. Roy, Ber., 89, 1288 (1956).
(2) P. L. Julian, E. W. Meyer and H. C. Printy, THIS JOURNAL, 70,
887 (1948).
(3) See W. R. Hertler and E. J . Corey, J. Org. Chem. , in press, for
other formamide - isocyanide conversions.
(4) H. Ruschig and J. Schmidt-Thome, U. S. Patent 2,697,107
(1954>.
(5) The general intermediacy of 8-chloro-secondary amines in the
N-chloroamine - pyrrolidine process has been demonstrated by
experiments which will be published separately. Cf. W. R. Hertler,
Ph.D. Thesis, University of Illinois, 1958, and S . Wawzonek, M. F.
Nelson, Jr., and P. J. Thelen, THIS JOURNAL, 7S, 2806 (1951).