Synthesis of the Selective D2 Receptor Agonist PNU-95666E from D-Phenylalanine Using a Sequential Oxidative Cyclization Strategy

Article abstract of DOI:10.1021/jo970526a

Compound 1 (PNU-95666E) is a selective and high-affinity agonist at the dopamine D₂ receptor subtype and is of interest as a potential agent for the treatment of Parkinson's disease. Requiring a synthetic route amenable to scale-up, a synthesis

Full text of DOI:10.1021/jo970526a

6582
J . Org. Chem. 1997, 62, 6582-6587
Syn th esis of th e Selective D2 Recep tor Agon ist P NU-95666E fr om
D-P h en yla la n in e Usin g a Sequ en tia l Oxid a tive Cycliza tion
Str a tegy
Arthur G. Romero,*,† William H. Darlington,† and Moses W. McMillan‡
Medicinal Chemistry Research and Chemical Research Preparations, Pharmacia & Upjohn, Inc.,
Kalamazoo, Michigan 49001
Received March 21, 1997X
Compound 1 (PNU-95666E) is a selective and high-affinity agonist at the dopamine D2 receptor
subtype and is of interest as a potential agent for the treatment of Parkinson’s disease. Requiring
a synthetic route amenable to scale-up, a synthesis of this enantiomerically pure tricyclic compound
was developed, starting from D-phenylalanine. Critical to the success of this synthesis were two
oxidative nitrogen annulations to provide the tricyclic ring system. A highly efficient reduction
with borane-methyl sulfide was used to reduce three different functional groups, a total of six
hydrides transferred, with no concomitant racemization, contributing to the synthesis of 1 in eight
steps with an overall yield of 26%. The utility of this synthetic route has been demonstrated by
the completion this synthesis on multikilogram scale.
In tr od u ction
Parkinson’s disease, a neurodegenetive disease char-
acterized by deteriorating motor function, is brought
about by the loss of cells in the brain responsible for
synthesizing the neurotransmitter dopamine. The stan-
dard treatment for this condition is L-DOPA, which is
converted to dopamine in situ by L-DOPA decarboxylase.
However, after a period of time L-DOPA loses its ef-
fectiveness, requiring adjunct therapy with dopamine
agonists to attempt to re-establish the necessary level of
dopamine receptor activation; adverse side effects such
as psychiatric disorders are also common. Evidence
suggests that it is the activation of D2 receptors in the
striatum, an area of the brain associated with motor
function, that is responsible for drug efficacy.1 However,
currently available dopamine agonist drugs are not
selective and possess high affinity for other dopamine
receptor subtypes, possibly contributing to the etiology
of drug-induced psychosis.2 Compound 1 (PNU-95666E)
is a potent and highly selective agonist at the dopamine
D2 receptor subtype and may possess potential as a
treatment for Parkinson’s disease with greatly dimin-
ished side-effect liability.3 Such a drug may also have
potential as a treatment for early stage Parkinson’s
disease in place of L-DOPA, possibly preventing the
development of response fluctuations seen with long term
L-DOPA therapy. This approach is not as feasible with
currently available, nonselective dopamine agonists due
to their undesirable side-effect profiles.
Requiring a synthetic route amenable to large-scale
synthesis,4 we sought an efficient procedure to synthesize
1, characterized structurally by three mutually fused
rings, as well as exhibiting (R)-chirality at the methyl-
amine-substituted position. Retrosynthetic analysis (Fig-
ure 1) indicated that D-phenylalanine (4) could potentially
provide a highly efficient starting material, already
containing the requisite stereochemistry as well as all
of the ring carbons except for the urea carbonyl group.
To successfully take advantage of these features would
require the attachment of a nitrogen onto the benzene
ring at the ortho-position to ultimately obtain a tetrahy-
droquinoline ring (e.g., 2), followed by the further sub-
stitution of another adjacent nitrogen to provide the
essential elements needed to generate the cyclic urea.
These requirements were realized using a pair of elec-
trophilic oxidative cyclizations to sequentially attach
these two nitrogen atoms.
Discu ssion
Examination of the retrosynthetic analysis (Figure 1)
indicated that the tetrahydroquinoline ring construction
would be a key step in the synthesis of 1 from D-
phenylalanine; we were therefore delighted to find that
Kikugawa had previously performed this type of trans-
formation (Scheme 1), where N-methoxyamide 5 was
oxidized to the N-chloro adduct, subsequently treating
this with a silver(I) salt in trifluoroacetic acid solvent to
generate an electrophilic intermediate (6), which cyclized
to afford 7.5 However, the necessity of working with
insoluble silver chloride salts and large amounts of
trifluoroacetic acid made this two-step procedure rela-
† Medicinal Chemistry Research.
‡ Chemical Research Preparations.
X Abstract published in Advance ACS Abstracts, August 15, 1997.
(1) Seeman, P.; Niznik, H. B. Dopamine receptors and transporters
in Parkinson's disease and schizophrenia. FASEB J . 1990, 2737-2744.
(2) Coleman, R. J . Current drug therapy for Parkinson’s disease.
A review. Drugs and Aging 1992, 2, 112-1124. Collier, D. S.; Berg,
M. J .; Fincham, R. W. Parkinsonism treatment: Part III-Update. Ann.
Pharmacother. 1992, 26, 227-233.
(3) Compound 1 possesses a Ki of 9 nM for the D2 receptor subtype,
while the affinity of 1 for the D1, D3, D4, and R1 receptors is greater
than 2300 nM and at the R2 and cholinergic receptors is greater than
1100 nM. Heier, R. F.; Dolak, L. A.: Duncan, J . N.; Hyslop, D. K.;
Lipton, M. F.; Martin, I. J .; Mauragis, M. A.; Piercey, M. F.; Nichols,
N. F.; Schreur, P. J . K. D.; Smith, M. W.; and Moon, M. W. Synthesis
and biological activities of (R)-5,6-dihydro-N,N-dimethyl-4H-imida-
zo[4,5,1-ij]quinolin-5-amine and its metabolites. J . Med. Chem. 1997,
40, 639-646.
(4) The original synthesis was 12 steps, including a tartaric acid
resolution to obtain optically pure 1 (see ref 3). Also, the use of
potentially hazardous reagents such as p-toluenesulfonyl azide made
it not amenable to large scale use.
S0022-3263(97)00526-4 CCC: $14.00 © 1997 American Chemical Society

Synthesis of Selective D2 Receptor Agonist PNU-95666E
J . Org. Chem., Vol. 62, No. 19, 1997 6583
Sch em e 2. P r ep a r a tion of th e Op tica lly P u r e
Tetr a h yd r oqu in olin e Rin g System fr om
D-P h en yla la n in e (4)
F igu r e 1. Retrosynthetic analysis of 1.
tively unattractive on a large scale. An alternative one-
step approach was examined, using the powerful oxidant
bis(trifluoroacetoxy)iodobenzene which was also reported
to induce the oxidative N-methoxyamide substitution of
aromatic rings.6 It was found that treatment of 5 with
this reagent in dichloromethane at 0 °C gave 7 in good
yield without any detectible racemization.7
described above, N-methoxyamides also react in the
presence of this reagent.
Sch em e 1. Kik u ga w a ’s Oxid a tive Cycliza tion of 5
To Obta in th e Tetr a h yd r oqu in olin e Sk eleton 7
To investigate this approach, the nitrogen of D-pheny-
lalanine was protected under Schotten-Baumann condi-
tions, affording the methyl carbamate as a thick liquid
(8) which was carried on without purification (Scheme
2). The free carboxylic acid was converted to N-meth-
oxyamide 9 by coupling with methoxylamine using
1-ethyl-3-(3-(dimethylamino)propyl)carbodiimide hydro-
chloride (EDC).9 The choice of EDC as the coupling agent
was 2-fold, permitting water to be used as solvent and
allowing for the removal of coupling reagents by aqueous
extraction. The resulting N-methoxyamide was purified
by crystallization from ethyl acetate, affording 9 in 61%
overall yield from D-phenylalanine.10
When N-methoxyamide 9 was treated with bis(trifluo-
roacetoxy)iodobenzene in dichloromethane at 0 °C, it
cyclized to lactam 10 in disappointingly low yield (40%).
A survey of experimental conditions was undertaken in
an attempt to improve the yield of 10. While changing
solvents (acetonitrile, toluene, acetone, chloroform) dem-
onstrated that halogenated solvents afforded cleaner
product; it did not improve the yield of 10. Addition of
sodium carbonate to scavenge trifluoroacetic acid only
further reduced the yield. However, this result suggested
that the presence of trifluoroacetic acid was beneficial;
to examine this possibility the cyclization was run in the
presence of 3 equiv of trifluoroacetic acid. This modifica-
tion dramatically increased the yield of 10 to 85%. It is
likely that the added trifluoroacetic acid serves to in-
crease the electrophilicity of the N-methoxy-N-acylnitro-
Although encouraged by this result, for a chemoeffi-
cient synthesis it was desirable to avoid the use of the
phthalimide protecting group due to its sizable molecular
weight and deprotection step. As an alternative, it was
envisioned that a simple carbamate group could serve
to protect the nitrogen of D-phenylalanine, serving as a
latent N-methyl group by eventual reduction after the
oxidative cyclization. However, it was not clear whether
a secondary carbamoyl nitrogen would tolerate exposure
to bis(trifluoroacetoxy)iodobenzene, since structurally
related primary amides undergo a Hoffmann-type rear-
rangement in the presence of this oxidant8 and, as
(8) (a) Louden, M. G.; Radhakrishna, A. S.; Almond, M. R.; Blodgett,
J . K.; Boutin, R. H. Conversion of aliphatic amides into amines with
[I,I-bis(trifluoroacetoxy)iodo]benzene. 1. Scope of the reaction. J . Org.
Chem. 1984, 49, 4272-4276. (b) Using bis(acetoxy)iodobenzene:
Moriarty, R. M.; Chany, C. J ., II; Vaid, R. K; Prakash, O.; Tuladhar,
S. M. Preparation of methyl carbamates from primary alkyl- and
arylcarboxamides using hypervalent iodine. J . Org. Chem. 1993, 58,
2478-2482.
(9) The coupling reagent EDC is a water soluble form of dicyclo-
hexylcarbodimide (DCC): Sheehan, J . C.; Preston, J .; Cruickshank,
P. A. A rapid synthesis of oligopeptide derivatives without isolation of
intermediates. J . Am. Chem. Soc. 1965, 87, 2492-2493.
(10) The ethyl carbamate version of 9 was not crystalline and
required silica gel chromatography for purification.
(5) Kawase, M.; Kitamura, T.; Kikugawa, Y. Electrophilic aromatic
substitution with N-methoxy-N-acylnitrenium ions generated from
N-chloro-N-methoxyamides: synthesis of nitrogen heterocyclic com-
pounds bearing a N-methoxyamide group. J . Org. Chem. 1989, 54,
3394-3403.
(6) Kikugawa, Y.; Kawase, M. An electrophilic aromatic substitution
by N-methoxyamides via hypervalent iodine intermediates, Chem. Lett.
1990, 581-582.
(7) Enantiomeric purity was determined by converting 1 to the BOC-
amine and analyzing this with a chiral HPLC column, where the two
enantiomers cleanly resolved, as determined by examination of racemic
1. See Experimental Section for details. We thank Malcolm W. Moon
for providing us with this procedure.

6584 J . Org. Chem., Vol. 62, No. 19, 1997
Romero et al.
nium ion intermediate, or possibly increases the reactiv-
ity of the bis(trifluoroacetoxy)iodobenzene reagent itself.
Interestingly, it was also found that N-methoxyamide 9
cyclized in good yield using the less powerful oxidant
bis(acetoxy)iodobenzene in the presence of 10 equiv of
trifluoroacetic acid, whereas only a small amount of
product is observed in the absence of trifluoroacetic acid.11
With a good synthesis of lactam 10 in hand, we were
now ready to attempt an exhaustive reduction to obtain
diamine 11. This was achieved with surprising ease
using an excess of borane-methyl sulfide complex in
refluxing tetrahydrofuran, representing a total of six
hydrides transferred to multiple reaction sites, including
the reduction of a nitrogen-oxygen bond. The liquid
diamine was of sufficient purity to be carried on to the
next step without purification; if desired, air sensitive
11 could be crystallized to purity as the maleic acid salt
(71% yield). Substituting lithium aluminum hydride for
borane-methyl sulfide gave much less satisfactory re-
sults.
Before we could attempt to synthesize the benzo-fused
urea (2-benzimidazolinone) portion of the molecule, it was
necessary to selectively protect the aliphatic methyl-
substituted amine of 11. The carbobenzyloxy (CBZ)
group was chosen for this protection because of the ability
to remove it by hydrogenolysis, the significance of which
will become apparent later (vide infra). Treatment of 11
with 1 equiv of benzyl chloroformate (CBZ-Cl) afforded
12 significantly contaminated with material where the
CBZ group was substituted on the arylamine, as well as
diprotected product.12 Decreasing the reaction temper-
ature from 0 to -30 °C during the CBZ-Cl addition
substantially decreased the amounts of these side prod-
ucts. Based upon this result, it was anticipated that
selectivity could be further improved by substituting the
less reactive CBZ transfer reagent N-(benzyloxycarbo-
nyloxy)succinimide (BCS). Indeed, when 11 was treated
with BCS at -40 °C excellent selectivity was demon-
strated for the aliphatic amine, with only a trace amount
of other CBZ-substituted products, by GLC analysis.13
Purification by silica gel chromatography gave 12 in 65%
overall yield from 10 (borane reduction and CBZ protec-
tion steps).
F igu r e 2. Model systems for synthesizing benzo-fused ureas
(2-benzimidazolinones) by oxidative cyclization of N-methox-
yureas (see text).
scale synthesis, or (2) the chlorination of 13 to give an
unstable N-chloro-N-methoxyurea species which was
induced to cyclize to 14 by treatment with strong base
to generate an anionic intermediate (Figure 2).15 Obvi-
ously this latter procedure which requires hydrogen
substitution on the aryl nitrogen of the urea (i.e., 13)
would not be compatible with our substrate (17).16
Requiring an alternative oxidant to activate the N-
methoxyurea group for intramolecular cyclization, we
quickly considered the possibility of once again using
bis(trifluoroacetoxy)iodobenzene. The demonstrated abil-
ity of this reagent to activate N-methoxyamide groups
for electrophilic aryl substitution suggested that it might
provide a superior method to induce oxidative cyclization
on the analogous N-methoxyurea groups. However, we
were disappointed to find that when model system 13
was treated with this reagent a mixture of unidentifiable
products was obtained (Figure 2). While at first puzzled
by this result, we speculated that the aryl-substituted
nitrogen could be serving as an alternative site for
oxidation by the bis(trifluoroacetoxy)iodobenzene reagent.
To examine this question, the N-methyl-substituted
model system 15, which more accurately represents the
aryl N-substitution of our actual substrate (17), was
treated with the oxidant and found to afford 16 in 79%
yield. Indeed, examination of this bis(trifluoroacetoxy)-
iodobenzene-mediated annulation procedure with other
N-methoxyurea substrates indicated that carbon substi-
tution at the aryl N-position of the N-methoxyurea was
critical,17 constituting a procedure complimentary to the
chlorination/sodium hydride method. However, our op-
timism for the application of the bis(trifluoroacetoxy)-
iodobenzene procedure to 17 was attenuated by the
consideration that an attempt to cyclize N-methoxy-3N-
benzyl-3N-methylurea (benzyl substituted for the phenyl
At this point there remained the critical introduction
of the second aryl nitrogen onto 12 for incorporation into
the benzo-fused urea. An examination of the mechanism
of the oxidative cyclization of N-methoxyamides (see
Scheme 1) suggested that this reaction may be general
to other types of N-methoxyacyl groups and that acyclic
N-methoxyurea 17 could potentially serve as a substrate;
substantiating this were two reports describing proce-
dures for the oxidative cyclization of 13 to obtain 14.
These employed (1) lead tetraacetate under dilute condi-
tions to aid in recovering 14 from the intractable lead
residue,14 making this procedure unappealing for large
(15) Perronnet, J .; and Demoute, J .-P. Approach to the 1-methoxy-
2-benzimidazolinones. Gazz. Chim. Ital. 1982, 112, 507-511.
(16) The N-chlorination/sodium hydride procedure requires a sub-
strate containing a hydrogen on the aryl-substituted nitrogen; an
attempt to perform the cyclization without such substitution failed,
demonstrating the need to enhance the reactivity of the aryl ring by
converting the aryl nitrogen into its anionic form. See ref 15.
(17) The absence of a substituent on the aryl-substituted nitrogen
of the N-methoxyurea apparently open up other avenues for bis(trif-
luoroacetoxy)iodobenzene oxidation, leading to decomposition. This
does not appear to be a problem when this nitrogen is substituted.
For further discussion, see: Romero, A. G.; Darlington, W. H.;
J acobsen, E. J .; Mickelson, J . W. Oxidative cyclization of acyclic ureas
with bis(triflyoroacetoxy)iodobenzene to generate N-substituted 2-ben-
zimidazolinones. Tetrahedron Lett. 1996, 37, 2361-2364.
(11) For another example of in situ generation of the more reactive
bis(trifluoroacetoxy)iodobenzene reagent, see ref 6 in the following:
Radhakrishna, A. S.; Parham, M. E.; Riggs, R. M.; Louden, G. M. New
method for direct conversion of amides to amines. J . Org. Chem. 1979,
44, 1746-1747.
(12) Not surprisingly, selective protection with the more reactive
trifluoroacetic anhydride totally failed, giving a statistical mixture of
products.
(13) This selectivity, the relative stability of this crystalline reagent
to storage, and its non-lachrymator properties also made it more
desirable than CBZ-Cl for the multikilo-scale reaction.
(14) Cooley, J . H.; and J acobs, P. T. Oxidative ring closure of
1-benzyloxy-3-arylureas to 1-benzyloxybenzimidazolones. J . Org. Chem.
1975, 40, 552-557.

Synthesis of Selective D2 Receptor Agonist PNU-95666E
J . Org. Chem., Vol. 62, No. 19, 1997 6585
Sch em e 3. In tr od u ction of th e Ben zo-F u sed Ur ea
atom onto a benzene ring; traditional approaches fre-
quently focus on nitration chemistry, a hazard on large
scale and not always amenable to ortho substitution. This
sequence of tandem oxidative cyclizations was instru-
mental in providing an eight-step synthesis of optically
pure 1 (PNU-95666E) from D-phenylalanine in 26%
overall yield and has been performed successfully on a
multikilo scale.20
via Oxid a tive Cycliza tion
Exp er im en ta l Section
Gen er a l. Methoxyamine hydrochloride, D-phenylalanine,
diacetoxyiodobenzene, and Pearlman’s catalyst were obtained
from Aldrich; phosgene in toluene was obtained from Fluka;
and 1-(3-(dimethylamino)propyl)-3-ethylcarbodiimide hydro-
chloride (EDC) was obtained from Sigma. Both commercial
bis(trifluoroacetoxy)iodobenzene and reagent prepared from
diacetoxyiodobenzene as described by Loudon21 was used. All
reactions were run under a nitrogen atmosphere. Proton NMR
were run on a Bruker Model AM-300 spectrometer at 300 MHz
field strength and are reported in ppm relative to TMS.
J
in 15) resulted in decomposition, presumably by oxidation
of the activated benzylic methylene. Consequently we
maintained a concern about the stability of the benzylic
methylene portion of 17’s CBZ protecting group to oxida-
tion by bis(trifluoroacetoxy)iodobenzene.18
vaues are in hertz. Mass spectrometry was performed at 70
eV ionization energy with ions reported as m/e. Melting points
are uncorrected. GLC conditions are as follows: Hewlett-
Packard Model 5890A capillary gas chromatograph, J & W
Scientific Inc., DB-5.5% phenylmethyl silicone column (15 m
× 0.53 mm × 1.5 µm film thickness), helium carrier gas (100
mL/min.), hydrogen flame ionization detector. Programmed:
100 °C, 1 min; increasing 20 °C/min to 250 °C; 250 °C, 10 min.
HPLC conditions are as follows: Zorbax Rx-C8 column (4.6
mm × 25 cm), solvent A, 10% CH3CN and 90% H2O (pH 3
phosphate buffer), solvent B, 85% CH3CN and 15% H2O (pH
3 phosphate buffer), programmed gradient, 90% A/10% B to
5% A/95% B over 12 min, maintain at 5% A/95% B for 4 to 10
min; λ ) 215 nm, flow ) 2 mL/min.
(R)-2-((Met h oxyca r b on yl)a m in o)-3-p h en ylp r op a n oic
Acid (8). A solution of D-phenylalanine (25.00 g, 0.151 mol)
and sodium hydroxide (6.05 g, 0.151 mol) in water (170 mL)
and tetrahydrofuran (225 mL) was cooled to -15 °C, and a
solution of methyl chloroformate (18.6 g, 0.197 mol) in tet-
rahydrofuran (50 mL) was added dropwise. When the one-
half of the methyl chloroformate had been added, a solution
of sodium hydroxide (9.10 g, 0.227 mol) in water (20 mL) was
added. When the addition was complete, the solution was
stirred at 25 °C for an additional 2 h and acidified with 10%
hydrochloric acid to pH 2. The solution was extracted twice
with diethyl ether, and the extracts were washed with brine
and dried (MgSO4). The solvent was removed under vacuum
to leave a clear oil (34 g): 1H NMR (CDCl3) δ 3.09 (dd, 1H, J
) 6.4, 14.0), 3.19 (dd, 1H, J ) 5.6, 13.8), 3.65 (s, 3H), 4.66
(two d, 1H, J ) 6), 5.25 (d, 1H, J ) 8.2), 7.15-7.31 (m, 5H),
8.22 (br s, 1 H); IR (thin film) 1726, 1498, 1455, 1448, 1377
cm-1; exact mass calcd for C11H13NO4 224.0923, found 224.0921.
(R)-N-Meth oxy-2-((m eth oxyca r bon yl)a m in o)-3-p h en yl-
p r op a n a m id e (9). A solution of sodium carbonate (10.20 g,
96.2 mmol) in water (170 mL) was added to a solution of 8
(∼0.148 mol crude) in methylene chloride. Methoxyamine
hydrochloride (14.2 g, 0.170 mol) and 1-(3-(dimethylamino-
)propyl)-3-ethylcarbodiimide hydrochloride (EDC) (31.21 g,
0.163 mol) were added, and the mixture was stirred at room
temperature for 22 h. The mixture was diluted with tetrahy-
drofuran (to dissolve the precipitate), and the layers were
separated. The aqueous layer was extracted with 1:1 tetrahy-
drofuran/diethyl ether, and the combined organic extracts were
washed with 10% hydrochloric acid and a saturated sodium
bicarbonate solution. The solution was dried (MgSO4) and
filtered, and the solvent was removed under vacuum to leave
a white solid (34.2 g). Crystallization from ethyl acetate gave
22.6 g of colorless crystals as a single crop, mp 154-155 °C
Having thus determined with the aid of these model
compounds the manner in which we were going to
attempt to carry out the urea annulation on the authentic
substrate (17), we turned to the actual preparation of this
compound (Scheme 3), easily accomplished by the addi-
tion of aryl amine 12 to a THF solution of phosgene,
subsequently adding methoxylamine to the carbamoyl
chloride adduct to afford 17.19 In the critical step,
treatment of N-methoxyurea 17 with bis(trifluoroac-
etoxy)iodobenzene proved our concerns regarding the
stability of the CBZ group to be unfounded; the reaction
smoothly afforded the desired tricyclic N-methoxyurea
18 in 78% overall yield from 12.
All that remained to complete the synthesis of 1 was
to concurrently remove the CBZ protecting group and
cleave the nitrogen-oxygen bond of N-methoxyurea 18
by means of hydrogenation over palladium hydroxide on
carbon (Pearlman’s catalyst). While the CBZ group came
off quickly, cleavage of the nitrogen-oxygen bond was
quite sluggish, requiring 2-3 days at 25 °C with 50 psi
of hydrogen pressure. To increase the rate of this
conversion, the ethanol slurry was heated to 50 °C,
causing the nitrogen-oxygen bond reduction to proceed
to completion in several hours. Verification of the
enantiomeric purity of 1 by chiral HPLC demonstrated
that it was greater than 99% optically pure.7 Since 1 was
very water soluble, it was isolated directly as the maleic
acid salt (84% yield of pure, crystalline material), after
first removing the hydrogenation catalyst by filtration.
Con clu sion
The utility of oxidative cyclization of N-methoxyacyl
groups, both amides and ureas, has been demonstrated
to be a powerful tool for the organic chemist to synthesize
heterocyclic rings in complex molecules. In part this is
because few procedures exist to directly attach a nitrogen
(18) For example, treatment of the N-methoxyurea obtained from
N-benzyl-N-methylamine with bis(trifluoroacetoxy)iodobenzene led to
decomposition. We postulate that this is due to the enhanced sensitiv-
ity of the benzylic methylene group to oxidation.
(19) Phosgene was replaced with triphosgene, affording a similar
yield when the reaction was scaled-up. While also effective, substitu-
tion of carbonyldiimidazole (CDI) was less satisfactory.
(20) This synthesis was presented at the Technical Achievement
Symposium at the 210th National Meeting of the American Chemical
Society, Chicago, IL, August 20-24, 1995.
(21) For a preparation of bis(trifluoroacetoxy)iodobenzene, see ref
8a.

6586 J . Org. Chem., Vol. 62, No. 19, 1997
Romero et al.
(61% overall from D-phenylalanine): 1H NMR (CDCl3) δ 3.05
(d, 2H, J ) 7.4), 3.58 (s, 3H), 3.61 (s, 3H), 4.34 (q, 1H, J )
7.9), 5.66 (br d, 1H, J ) 8.2), 7.15-7.31 (m, 5H), 9.44 (s, 1H);
C18H20N2O2: C, 72.95; H, 6.80; N, 9.45. Found: C, 72.87; H,
6.84; N, 9.32.
(R)-Meth yl[1,2,3,4-tetr a h yd r o-1-[(m eth oxya m in o)ca r -
bon yl]-3-qu in olin yl]ca r ba m ic Acid , P h en ylm eth yl Ester
(17). A solution of (R)-methyl(1,2,3,4-tetrahydro-3-quinoli-
nyl)carbamic acid, phenylmethyl ester (12) (3.81 g, 12.86
mmol) and triethylamine (3.9 g, 39 mmol) in dry tetrahydro-
furan (50 mL) was added with stirring to a solution of
phosgene (7.1 mL of a 1.93 M toluene solution) in tetrahydro-
furan (100 mL) at 0 °C. After 1 h, methoxyamine hydrochlo-
ride (2.15 g, 25.7 mmol) and triethylamine (3.9 g, 39 mmol)
were added, and the mixture was stirred at room temperature
for 2 days. The solution was diluted with diethyl ether and
washed with water and brine. The organic layer was dried
(MgSO4), and the solvent was removed under vacuum to leave
an oil (5.13 g, >100% crude yield) which was sufficiently pure
for the next step. An analytical sample was purified via flash
chromatography (230-400 mesh silica gel; 50% ethyl acetate/
hexane) to give the title compound as an oil: 1H NMR (CDCl3)
δ 2.88 (s, 3H), 2.77-2.97 (m, 2H), 3.75 (s, 3H), 3.52-4.08 (m,
2H), 4.54 (m, 1H), 5.13 (s, 2H), 7.10-7.27 (m, 4H), 7.35 (br s,
IR (mineral oil) 1694, 1668 cm-1; [R]25 ) +5.2° (CH3OH, c )
D
1.045). Anal. Calcd for C12H16N2O4: C, 57.13; H, 6.39; N,
11.10. Found: C, 57.39; H, 6.41; N, 11.09.
Meth yl (R)-N-(1,2,3,4-Tetr a h yd r o-1-m eth oxy-2-oxo-3-
qu in olin yl)ca r ba m a te (10). A suspension of (R)-N-methoxy-
2-((methoxycarbonyl)amino)-3-phenylpropanamide (9) (11.25
g, 44.6 mmol) in dichloromethane (170 mL) was cooled in an
ice bath, and trifluoroacetic acid (9.25 mL, 13.7 g, 0.120 mol)
was added. Bis(trifluoroacetoxy)iodobenzene (19.78 g, 0.046
mol) was added portion wise over 10 min at 0 °C, and the
mixture was stirred at this temperature for 1 h. The mixture
was washed with a 10% sodium carbonate solution and dried
(MgSO4). The solvent was removed under vacuum to leave
an amber oil (19.58 g). Purification by flash chromatography
(230-400 mesh silica gel, 40-50% ethyl acetate/hexane) gave
the desired product as an amber oil which solidified (9.45 g,
85% yield). An analytical sample (1.5 g) was crystallized from
ethyl acetate/hexane to give white crystals (1.36 g, mp 117-
119 °C): 1H NMR (CDCl3) δ 2.85 (t, 1H, J ) 14.7), 3.44 (dd,
1H, J ) 15.0, 5.8), 3.72 (s, 3H), 3.93 (s, 3H), 4.42 (dt, 1H, J )
14.2, 5.6), 5.82 (br s, 1H), 7.09 (dt, 1H, J ) 7.4, 1.2), 7.33 (t,
1H, J ) 7.7), 7.22 (d, 2H, J ) 7.6); IR (mineral oil) 1722, 1703
5H), 7.76 (br s, 1H); IR (thin film) 1734, 1697, 1605 cm-1; [R]25
D
) +38° (CH3OH, c ) 0.980); exact mass calcd for C20H23N3O4
369.1688, found 369.1682.
(R )-Me t h yl(1,2,5,6-t e t r a h yd r o-1-m e t h oxy-2-oxo-4H -
im id a zo[4,5,1-ij]q u in olin -5-yl)ca r b a m ic Acid , P h en yl-
m et h yl E st er (18). A solution of (R)-methyl[1,2,3,4-tetra-
hydro-1-[(methoxyamino)carbonyl]-3-quinolinyl]carbamic acid,
phenylmethyl ester (17) (7.26 g, 19.7 mmol) in chloroform (150
mL) was cooled to -5 °C in an ice-salt bath. Bis(trifluoro-
acetoxy)iodobenzene (10.14 g, 23.6 mmol) was added, and the
mixture was stirred at -5 to 0 °C for 4 h and then at 25 °C for
2 h more at which time the reaction was complete by TLC.
The reaction mixture was washed with 10% aqueous sodium
carbonate, back-extracting the aqueous fractions with diethyl
ether. The combined organic layers were dried (MgSO4), and
the solvent was removed under vacuum to leave a brown oil
(10.7 g). Purification by flash chromatography (230-400 mesh
silica gel, 50% ethyl acetate/hexane) gave an amber oil which
slowly solidified (5.67 g, 78% yield). HPLC analysis indicated
two peaks, 10.79 (97.4%) and 11.95 min (2.6%). An analytical
sample (0.54 g) was crystallized from ethyl acetate/hexane to
give off-white crystals (0.41 g, mp 105-106.5 °C): 1H NMR
(CDCl3) δ 2.93 (s, 3H), 2.90-3.30 (m, 1H), 3.14 (dd, 1H, J )
15.5, 11.0), 3.68 (m, 1H), 4.07 (s, 3H), 4.11 (dd, 1H, J ) 11.9,
4.8 Hz), 4.65 (m, 1H), 5.16 (m, 2H), 6.88 (d, 1H, J ) 7.5), 6.96
(d, 1H, J ) 7.5), 7.04 (t, 1H, J ) 7.7), 7.36 (m, 5H); IR (mineral
oil) 1725, 1717, 1694 cm-1; [R]25D ) +46.8° (CH3OH, c ) 0.731).
Anal. Calcd for C20H21N3O4: C, 65.38; H, 5.76; N, 11.44.
Found: C, 65.41; H, 5.77; N, 11.42.
cm-1; [R]25 ) +34.2° (CH3OH, c ) 0.927). Anal. Calcd for
D
C12H14N2O4: C, 57.59; H, 5.64; N, 11.19. Found: C, 57.55; H,
5.64. N, 11.32.
(R)-3-(Meth yla m in o)-1,2,3,4-tetr a h yd r oqu in olin e Ma le-
a te Sa lt (11). A solution of (R)-N-(1,2,3,4-tetrahydro-1-
methoxy-2-oxo-3-quinolinyl)carbamate (10) (29.1 g, 116.4 mmol)
in dry tetrahydrofuran (400 mL) was cooled to 0 °C, and
borane-methyl sulfide (10.0 M solution, 70 mL, 6.0 equiv) was
slowly added. The solution was allowed to warm to 25 °C and
stirred for 2.5 h. The solution was then refluxed on a steam
bath for 30 h and then cooled to 0 °C, quenching dropwise
(careful hydrogen evolution) with 10% hydrochloric acid (160
mL). This solution was refluxed on the steam bath for 1.5 h,
cooled in ice, and made basic with 12 N aqueous sodium
hydroxide. The mixture was extracted twice with diethyl
ether, and the combined extracts were washed with brine and
dried (MgSO4). The solvent was removed under vacuum to
leave a clear oil (19.6 g, approximately 100% crude yield) which
was carried on without further purification. Examination of
crude diamine 11 by GLC shows peaks at 5.15 min (2%), 5.46
min (11, 85%), 5.83 min (3%), and 7.39 min (10%). To obtain
an analytical sample, an aliquot of crude 11 was crystallized
as its maleate salt in methanol/ether (71% yield); mp 175 °C;
1H NMR of the maleic acid salt (CDCl3) δ 2.64 (s, 3H), 2.80
(dd, 1H, J ) 17.0, 5.3), 3.11 (dd, 1H, J ) 16.9, 4.9), 3.20-3.52
(m, 3H), 3.55 (m, 1H), 5.92 (s, 1H), 6.03 (s, 2H, maleic acid
CHdCH), 6.53-6.58 (m, 2H), 6.927-6.97 (m, 2H), 8.48 (br s,
(R )-5,6-Dih yd r o-5-(m e t h yla m in o)-4H -im id a zo[4,5,1-
ij]qu in olin -2(1H)-on e Ma lea te Sa lt (1; P NU-95666E). A
mixture of methyl(1,2,5,6-tetrahydro-1-methoxy-2-oxo-4H-imi-
dazo[4,5,1-ij]quinolin-5-yl)carbamic acid, phenylmethyl ester
(18) (3.87 g, 10.5 mmol) and 20% palladium hydroxide on
carbon (1.0 g) in absolute ethanol (100 mL) was shaken in a
Parr apparatus with an initial hydrogen pressure of 50 psi
for 19 h. The mixture was filtered through diatomaceous
earth, and the catalyst was washed with ethanol, and the
solvent was removed under vacuum to leave a thick oil (2.2
g). This was dissolved in methanol (25 mL) and added to a
solution of maleic acid (1.20 g, 10.3 mmol) in methanol (25
mL). Crystallization gave an off-white solid (2.55 g, mp 211
°C). A second crop was obtained by adding diethyl ether (0.29
g) obtaining a total of 2.84 g of analytically pure material (84%
yield). HPLC analysis of this salt indicated it to be greater
than 99% pure: 1H NMR of the maleic acid salt (DMSO) δ
2.68 (s, 3H), 3.05 (dd, 1H, J ) 16.6, 5.7), 3.22 (dd, 1H, J ) 3.3,
16.3), 3.90-4.06 (m, 3H), 6.05 (s, 2H, maleic acid CHdCH),
6.85-6.97 (m, 3H), 8.74 (br s, 1H), 10.83 (s,1H); IR (mineral
1H); [R]25 ) +19.0° (CH3OH, c ) 1.01); IR (thin film) 1638,
D
1608 cm-1
. Anal. Calcd for C10H14N2‚C4H4O4: C, 60.42; H,
6.52; N, 10.07. Found: C, 60.51; H, 6.53; N, 10.06.
(R)-Met h yl(1,2,3,4-t et r a h yd r o-3-q u in olin yl)ca r b a m ic
Acid , P h en ylm eth yl Ester (12). A solution of (R)-1,2,3,4-
tetrahydro-N-methyl-3-quinolinamine (11) (15.0 g of crude
material prepared above, approximately 84.4 mmol) in toluene
(50 mL) was stirred at -40 °C while N-(benzyloxycarbonylox-
y)succinimide (24.2 g, 97.1 mmol) in toluene (150 mL) was
added over 1 h. After 30 min at -40 °C, GLC analysis
indicated that all of 11 had been consumed. The solution was
quenched by the addition of sodium bicarbonate (300 mL of a
10% aqueous solution) and warmed to 0 °C, followed by the
addition of methanol (100 mL). This was stirred overnight
and then extracted with ethyl acetate. Drying over MgSO4
and solvent removal afforded a liquid which was purified by
flash chromatography (230-400 mesh silica gel; 4:1 hexane/
ethyl acetate) to give a liquid (17.2 g, 65% overall yield from
10) which crystallized from ethyl acetate/hexane to afford
white crystals (mp 80 °C): 1H NMR (CDCl3) δ 2.88 (s, 3H,),
2.80-3.04 (m, 2H), 3.30 (d, 2H, J ) 6.9), 3.83 (s, 1H), 4.57 (m,
1H), 5.16 (s, 2H), 6.51 (d, 1H, J ) 7.9), 6.64 (t, 1H, J ) 7.4),
6.96-7.02 (m, 2H), 7.35 (m, 5H); [R]25D ) -50.1° (CH3OH, c )
oil) 1696, 1638 cm-1; [R]25 ) -26.3° (H2O, c ) 0.836). Anal.
D
Calcd for C11H13N3O‚C4H4O4: C, 56.42; H, 5.37; N, 13.16.
Found: C, 56.45; H, 5.17; N, 13.13.
P r ep a r a tion of BOC-P r otected 1. Sodium hydroxide (2.0
N, 78 µL, 0.16 mmol) was added to a mixture of 1 (32.7 mg,
0.816); IR (mineral oil) 1680, 1606 cm-1
. Anal. Calcd for

Synthesis of Selective D2 Receptor Agonist PNU-95666E
J . Org. Chem., Vol. 62, No. 19, 1997 6587
0.136 mmol) and di-tert-butyl dicarbonate (0.034 g, 0.155
mmol) in dry tetrahydrofuran (1.0 mL) at room temperature-
,and the mixture was heated at reflux for 20 min. The mixture
was diluted with diethyl ether, washed with brine, and dried
(MgSO4). The solvent was removed under vacuum to leave
an oil. Purification by flash chromatography (230-400 mesh
silica gel, 50% ethyl acetate in hexane to pure ethyl acetate)
gave BOC-protected 1: 1H NMR (CDCl3) δ 1.48 (s, 9H), 2.88
(s, 3H), 2.93 (m, 1H), 3.13 (dd, 1H, J ) 15.9, 11.1), 3.68 (t, 1H,
J ) 11.2), 4.12 (dd, 1H, J ) 11.8, 4.8), 4.58 (m, 1H), 6.87 (d,
1H, J ) 6.9), 6.94-7.02 (m, 2H). Analysis with a chiral HPLC
column (Regis (R,R)-WHELK-O 1 column; 25 cm × 4.6 mm
i.d.; 5 µm; 50% 2-propanol in hexane; flow 1 mL/min; λ 215
nm) showed a single peak at 12.7 min (>99.5%). Previously
prepared racemic BOC-protected 1 displayed peaks at 8.0 and
12.7 min, corresponding to the two enantiomers.
J O970526A