Preparation of guanine PDE inhibitors: Development of the common synthetic route strategy. A case study

Article abstract of DOI:10.1021/op030212r

A single synthetic route, called the chloropurine route, capable of quickly delivering initial kilogram quantities of several chiral as well as achiral guanine phosphodiesterase inhibitors of increasing complexity is described. During the course of this w

Full text of DOI:10.1021/op030212r

Organic Process Research & Development 2004, 8, 396−400
Lecture Transcripts
Preparation of Guanine PDE Inhibitors: Development of the Common Synthetic
Route Strategy. A Case Study
D. Gala,* D. DiBenedetto, G. Gloor, J. Jenkins, M. Kugelman, D. Maloney, and A. Miller
Synthetic Chemistry-Chemical DeVelopment, Schering-Plough Research Institute, 1011 Morris AVenue,
Union, New Jersey 07083, U.S.A.
Abstract:
Scheme 1. Advanced activated pyrimidine route to PDE
inhibitors
A single synthetic route, called the chloropurine route, capable
of quickly delivering initial kilogram quantities of several chiral
as well as achiral guanine phosphodiesterase inhibitors of
increasing complexity is described. During the course of this
work, unraveling the formation mechanism of chloropurines
allowed for the scale-up of the key intermediates. The mecha-
nism for the cyclization of the chiral five-membered amino
alcohols and its implication on the enantiomeric purity needed
for this step are also described.
The timely availability of large quantities of a newly
identified pharmaceutically active compound in early stages
of its life cycle (Tox, Phase I) is important in its expedited
development. This stage can often slow the development of
an active pharmaceutical ingredient (API) if large quantities
(kilograms) of this candidate are unavailable due to a lack
of appropriate intermediates and/or synthetic methodologies.
To overcome such delays in the development of the
phosphodiesterase (PDE) inhibitors 1-3 we envisaged two
As an alternate to the above approach we were also
developing a common synthetic route strategy. Although such
approach is common practice for medicinal chemists for a
small-scale preparation of closely related drug analogues,
significant development is required to make such anapproach
viable on scale even for structurally related compounds. This
manuscript describes the challenges encountered and the res-
olution of these challenges, leading to a speedy delivery of
a diverse set of compounds. The approach described in this
contribution was based on the route used by the medicinal
chemists, but without the protecting groups (benzyl, TES)
1
7,18
and chromatographic separations/purifications.
In this
approaches. The first involved use of an advanced intermedi-
ate, such as activated pyrimidine 4, which can be converted
to any of the above desired PDE inhibitors of varying com-
plexities (Scheme 1). This work is described in the earlier
publications. This approach was well suited for the prepa-
ration of the PDE inhibitors 1-2 derived from a secondary
amino alcohol 5. However, the activated intermediate 4
proved resistant towards its reaction with tertiary amino
alcohol cycloleucine 8, limiting its utility for the synthesis
of 3.
series of compounds the absence of protecting groups not
only led to solubility difficulties, but it also led to previously
unanticipated intermediates/reaction mechanisms and stereo-
17,18
chemical outcomes compared to the above work.
This
1
,2
contribution describes the findings and the resolution of the
3
above issues leading to the “chloropurine” common syn-
thetic pathway. This was the preferred route for an expedited
delivery of kilogram quantities of all of the above PDE
inhibitors.
The development of the chloropurine route started with
Sch 47687 (1, Scheme 2), the first PDE inhibitor of interest
to us. The conversion of commercially available 6-amino
(
1) Gala, D.; DiBenedetto, D. J.; Kugleman, M.; Puar, M. S. Tetrahedron Lett.
003, 44, 2717.
2
(2) Gala, D.; DiBenedetto, D. J.; Kugelman, M.; Mitchell, M. B. Tetrahedron
Lett. 2003, 44, 2721.
(3) 2-Chloropurine is the key intermediate pathway approach.
3
96
•
Vol. 8, No. 3, 2004 / Organic Process Research & Development
10.1021/op030212r CCC: $27.50 © 2004 American Chemical Society
Published on Web 04/03/2004

Scheme 2. Synthesis of Sch 47687 via the chloropurine
route
complications (dimerization, polymerization, N-oxidations,
oiling of free base, etc.), allowing for good isolated yields
of high-melting crude 14. Crude 14 was then reacted with a
stoichiometric amount of chiral amino alcohol 5 in refluxing
acetonitrile containing an excess of TEA as an HCl scav-
enger. Typically, most of 14 was consumed after 72 h at
which stage the reaction mixture was filtered to remove TEA‚
HCl as well as residual 13. Addition of the reaction mixture
to water allowed for precipitation of 15. Crude 15, after
drying, was dissolved in methylene chloride and treated with
a slight excess of thionyl chloride. A gentle warming (35-
4
0 °C) of the reaction mixture led to essentially complete
cyclization of 15 to 1 as its HCl salt. Neutralization of this
reaction mixture with NH OH allowed for precipitation of
4
free base 1 which was recrystallized from acetonitrile. The
two-step yield, although good, was slightly variable. This
was traced to the degree of neutralization as well as the
uracil to caffeine metabolite A
1
, i.e., 12, via nitrosopyrimidine
7
11 developed and scaled up to kilogram scale in our labora-
4
tories and plant has been reported previously by us. Insolu-
bility of 12 in a variety of solvents posed a serious challenge
to our initial goal of converting it to the xanthine 13. It is
worth noting that the research synthesis1 utilized a benzyl
protecting group at the 5-acetamido nitrogen which mini-
mized the insolubility difficulties. It was also known that
the removal of this benzyl group at the end of the sequence
via catalytic hydrogenation was very difficult, often requiring
amount of water (introduced via aqueous NH
ficient neutralization caused the loss of 1‚HCl in water as
well as CH Cl , whereas an excess of water removed free
4
OH). Insuf-
7,18
2
2
base 1 due to its solubility in water. Since the above
preparation allowed for delivery of Tox and Phase I supplies
of 1, no further optimization work was done to improve the
last-step yield.
Sch 51866, 2, was chosen as the second PDE inhibitor
of interest. Application of the chloropurine route for its
preparation led to many dissimilar and some similar observa-
tions and required changes to the experimental conditions
at every step for a successful ending. Unlike acetic acid or
2
a large amount of 20% Pd(OH) /C. This was undesirable
from both the cost and “residual heavy metal (Pd) in the
API” perspective. Our initial desire was to “activate” 13 to
produce 14 by reacting it with an appropriate activating
reagent (sulfonyl halides, anhydrides, etc. leading to the
corresponding leaving group L in 14) such that the activated
purine 14 with a desirable leaving group L can then
efficiently react with expensive chiral amino alcohol 5. Our
continued work to form xanthine 13 led us to use POCl
When 12 was suspended in POCl and heated (to ensure
enough solubility for reasonable reaction rate), it led to the
formation of xanthine 13 as well as the chloropurine 14. We
further found that 14 is formed under these conditions. As
discussed later in this manuscript under the mechanism of
formation of chloropurines, an addition of excess NH
to the reaction mixture prior to heating 12 in neat POCl
a solvent and a subsequent workup with NH
for an efficient one-pot, two-step conversion of 12 to 14. In
a typical plant procedure, the batch was gradually heated to
reflux over 6-8 h and held at that temperature for 48 h for
complete consumption of 12 and 13. Attempts to minimize
3 2
acetic anhydride, p-CF -Ph-CH COOH is a solid. Thus, the
5
experimental conditions developed for the preparation of 12
via reductive amidation needed changes for the preparation
of 17. To this end it was found that 11 as an aqueous NaOH
solution can be completely hydrogenated to 16. This diamine
is air sensitive, and its isolation was undesirable. Again, it
was discovered that the treatment of a solution of this
diamine in aqueous NaOH with only a slight excess of
3
.
3
6
4
Cl
as
p-CF
3
-Ph-CH
2
COCl (prepared in a quantitative yield by
) led to
3
treatment of the corresponding acid with SOCl
2
4
OH allowed
precipitation of high-quality amide 17 in very good isolated
8
yield. Similar to amide 12, amide 17 could only be isolated
in its hydrated form. Initially this appeared counterproductive
for the next step where POCl was used as a solvent, and
3
hence, some effort was put into drying 17. As discussed later
in this contribution, difficulty in removing this water of
hydration from 17 proved fortuitous.
the use of POCl
lowered yields and/or increased impurities. Fortunately, after
neutralization of HCl and POCl , the isolation of 14 lacking
the protection of imidazole nitrogen(s) did not cause
3
via the use of high-boiling solvents typically
3
Unlike amide 12, initial attempts to convert 17 to xanthine
1
8 and chloropurine 19 led to irreproducible results. The rate
as well as extent (temperature and time) of heating the
suspension of 17 in POCl led to formation of one major to
(
(
(
4) Gala, D.; DiBenedetto, D.; G u¨ nter, F.; Kugelman, M.; Maloney, D.;
Cordero, M.; Mergelsberg, I. Org. Process Res. DeV. 1997, 1, 85.
5) (a) Barr, A.; Frencel, I.; Robinson, J. B. Can. J. Chem. 1977, 55, 4180. (b)
Overman, L. E.; Sugai, S. J. Org. Chem. 1985, 50, 4154.
6) Towards the development/optimization of 2-chloropurine formation/isolation
in high yields with the least number of impurities, several metal halides,
Vilsmeier reagents, phosphorous halides, phase-transfer reagents, amine
HCl salts, cosolvents (toluene, xylenes, chlorobenzene, dichlorobenzenes,
etc.), various workups (quenches with protic solvents, amines, inorganic
bases), and various combination of the above were evaluated. Of these,
3
many other products during its conversion to 19. This had a
significant impact on the yield as well as the quality of 19.
(7) The use of a mild base was desirable in view of the instability of these
PDE inhibitors to strong base such as aqueous NaOH. See Gala, D.; Puar,
M. S.; Czarniecki, M.; Das, P. R.; Kugelman, M.; Kaminiski, J. Tetrahedron
Lett. 2000, 41, 5025.
the use of POCl
3
and NH
4
Cl gave the best results. Since the latter is
(8) Such favorable faster amidation compared to the hydrolysis of this acid
chloride under such basic conditions was unexpected. This procedure was
applicable to several aliphatic as well as aromatic acid chlorides (unpub-
lished results).
inexpensive, appeared to improve solubility/stirability of the reaction
mixtures, and did not interfere with the workup, it was used in excess (3-5
mol).
Vol. 8, No. 3, 2004 / Organic Process Research & Development
•
397

Scheme 3. Preparation of Sch 51866 (2) via the
Scheme 4. Proposed chloropurine(s) formation mechanisma
chloropurine route
Conversion of 17 to 18, due to the insolubility of the latter,
typically resulted in a very thick, difficult-to-stir slurry. This
a
Conditions - a: Elimination. b: Addition:elimination to the nitriliums
necessitated use of more POCl
3
compared to 13 and required
followed by addition may be preferred; however, addition followed by elimination
cannot be ruled out.
longer reaction time for its conversion to 19. After unraveling
the mechanism of conversion of 17 to 19 via the intermediacy
of oxazole 21, a two-stage heating (75-80° for 24 h followed
by reflux for 3 days) led to 65% isolated yield of 19 (with
Several interesting mechanistic insights were made during
the discovery of the N-unprotected chloropurine route and
its application to the preparation of PDE inhibitors 1 and 2.
One pertained to the formation of chloropurines 14 and 19
from amides 12 and 17, respectively, and the other pertained
to cyclization of purine amino alcohols 15 and 20 to 1 and
5-10% 18 in it). An incomplete conversion of 18 to 19 and
some hydrolysis of 19 during neutralization/isolation were
the causes for contamination of 19 with 18. In the next set
of dissimilarities, it was discovered that unlike 14, the
reaction of chloropurine 19 with chiral 5 was very slug-
gish in refluxing acetonitrile, requiring over a week for
2
, respectively.
The conversion of 17 to 19 proved difficult compared to
that of 12 to 14. Had the latter conversion been difficult,
this route might never have materialized because trouble
shooting without a chromophore on 12-14 would have been
difficult! As stated above the one major and many minor
temperature-dependent compounds (in addition to 18) were
noticed during the conversion of 17 to 19. The presence of
the aromatic chromophore facilitated monitoring of the
reaction profile. The major new compound was isolated and
was determined (NMR, MS) to be oxazole 21. On the basis
of work done on isolated 21, the following conclusions were
3
made. Treatment of 17 with POCl leads to the formation
of 18 as well as 21. At temperatures of 80 °C or less, 21
converts to 18. In this case the amount of HCl and, hence,
water in 17 (i.e. water of hydration) is important. If the
∼
90% conversion. Development work for this reaction led
to the conditions described in Scheme 3 where the reaction
of 19 with (-)5 in N-methyl-2-pyrrolidinone (NMP) at ∼110
°
C was complete in ∼18 h. At the end of reaction, the
i
mixture was filtered to remove 18 and most of ( Pr)
2
Net‚
HCl salt. The reaction mixture was then added at room tem-
perature to water containing ∼3% MeOH. The presence of
MeOH allowed for the removal of traces of polar impurity
and led to a clean precipitation of 20 in 80% yields. As in
the case of 15, crude 20, after drying, was dissolved in meth-
ylene chloride (MC), treated with thionyl chloride, warmed
to ∼40 °C for its complete cyclization, and then treated with
9
4
aqueous NH OH, as above, to form free base 2. This free
base was not soluble in water; hence, it was isolated repro-
ducibly in 85-90% yield. This free base could be crystallized
from a variety of solvents. To improve the solubility and
bioavailability of this compound in water, a salt-selection
program ensued which identified the mono HCl salt as the
preferred salt. In the plant, free base 2 was suspended in
warm (60-65 °C) 2-propanol and treated with a stoichio-
metric amount of concentrated HCl at which stage the entire
(
9) As seen with the bicyclic azetidinone hemiketals,10 it is quite likely that
one or two of the of the compounds observed in this conversion are stable
bicyclic aminals 26. In view of facile 4-chlorination of pyrimidines,
1
4-chloropyrimidine derivative of 17 and or 18 can also be the sources of
the minor compounds.
mixture became a homogeneous solution from which the
.
2
HCl crystallized as a white solid when the solution was
(10) Kugelman, M.; Gala, D.; Jaret, R. S.; Nyce, P. L.; McPhail, A. T. Synlett
990, 431.
1
cooled. As noted in Scheme 3, in the initial stages of this
developmental work of 2 when analytical methods were
being developed, chemists monitored the progress of various
reactions by following the color changes and kept this
program on track.
398
•
Vol. 8, No. 3, 2004 / Organic Process Research & Development

Scheme 5. Mechanism of purine-amino alcohols cyclization to guanines
reaction mixture is rapidly heated to reflux, this HCl is
removed (presumably via N stream and/or azeotropic
distillation with POCl ) from the reaction mixture, resulting
this reaction “trans” five-membered PDE inhibitor products
or the “trans” chloro amines (via an intra- or intermolecular
displacement of OSO Cl group) were not detected. It appears
2
3
2
in degradation of 21 into a variety of compounds along with
formation of 18. This was proven in laboratory experiments
on isolated 21.
Thus, first-stage heating below 80 °C ensured a near
quantitative conversion of 17 to 18. In the case of 12 to 14,
less steric demand (methyl versus p-trifluoro-phenylmethyl)
in the corresponding oxazole for its conversion to 13 and
sufficient water of hydration in 12 probably made conversion
of 12 to 14 very facile and did not require a prolonged first-
stage heating.
that chlorosulfonylated 15 and 20 (27 and 28, respectively)
are well situated for a rapid, thermodynamically favored “cis”
five-membered ring cyclization, leading to the desired
products. Also noteworthy is that even though these “trans”-
substituted chlorosulfonylamines are well suited for the
formation of other compounds, such compounds remained
1
1
well below the detection limits in this series. As shown in
Scheme 5, when 5 was converted to “cis” amino alcohol
1
2
29 and subjected to the same reaction sequence, it too led
13
to the desired “cis” five-membered ring product. In this
case the chlorosulfonylated intermediate 31 does not undergo
thermodynamically less favored “trans” cyclization which
then allows for its conversion to the “trans” chloro amine
32 (a similar observation is reported by us in a previous
Isolated xanthines 13 and 18 are very stable and insoluble.
These properties required yet more energy for their conver-
sion to chloropurines 14 and 19, respectively. Our hypothesis
is that the enolization of insoluble 13/18 to 24 was the most
energy-demanding step. The trapped enolate 24, i.e., 25,
could then benefit from extra chloride ions for its conversion
to 19 (see Scheme 4). For this conversion (18, 24, 25, 19) a
variety of reaction conditions, cosolvents, metal chlorides,
amine hydrochloride salts, phase-transfer catalysts (PTC) in
combination with many divergent reaction-quenching and
1
article ). The latter is suitable for the favorable “cis” five-
1
4
membered ring closure. Thus, the chirality at the carbon
bearing the amine is very important, and the chirality of the
carbon bearing the hydroxyl group is unimportant for the
formation of the above PDEs. This was an important
observation for finding alternatives to expensive (-)15.
Finally, the lessons learned from the above two PDE
inhibitors allowed for a rapid application and scale-up of
the chloropurine route to the spirocyclic PDE inhibitor 3
(Scheme 6). One of the major issues with the unavailability
of the biphenyl acetic acid (33) was resolved by discovering
a novel, efficient, and scalable Pd/C coupling procedure as
an alternate to classical but laborious and impractical for
scale-up Wilger o¨ dt-Kindler biphenyl acetic acid synthesis.15
When the biphenyl acid chloride (via reaction of 33 with
6
isolation procedures were evaluated. Of these, the use of
NH
with NH
easy-to-remove “PTC” and actually improved the solubility
of 18 (contrary to expectation, the addition of NH Cl made
4
Cl as an additional chloride source and reaction quench
4
OH were found optimum. The former acted as an
4
the reaction mixture much easier to stir), lowered the reaction
time (to 3 days from 6 days), and posed no difficulties in
the workup and isolation of 19. The use of NH OH ensured
4
pH control (limited hydrolysis of chloropurines to xanthines),
and water solubility of its phosphate salts led to facile phase
splits during workup.
(
11) Trans-substituted chlorosulfonylamines are suitable intermediates to form
aziridines, but the latter were not detected in these reactions. In other
reactions we have noticed that tautomerism in imidazole ring led to the
reactions at either of the nitrogen atoms. In this series tautomers of 15 and
The second interesting observation was the stereochemical
outcome in the cyclization of 15 and 20. In a previous
20 (i.e. 39 and 40, respectively), if present, did not lead to 41 and 42. This
1
publication we have reported that the substrate in which
may be due to the fact that 39 and 40 would lead to seven-membered
strained ring products which are far less favorable than the desired five-
membered products, 1 and 2.
“
trans” amino alcohol (-)5 was incorporated led to the
formation of “cis” chloro amine when subjected to the SOCl
2
(
12) Enantiopure cis isomer was prepared by using the literature procedure for
the conversion of trans-R-amino alcohols to cis-R-amino alcohols: Barnard,
R. A. B.; Parkkari, J. H. Can. J. Chem. 1970, 48, 1377.
13) These conclusions were made after thorough H NMR, 13C NMR, HPLC,
and MS studies of the isolated compounds.
14) Incidentally, this observation can be used for the next generation of
inexpensive amino alcohol where the chirality at only the amino carbon
would be important, and chirality at the alcohol carbon would be less
important.
reaction. A formation of such “cis” chloro amine of 5 can
potentially lead to the “trans” five-membered cyclized isomer
of the PDE inhibitor. To determine the extent of this reaction
via a quantitative analytical method the following work was
undertaken. In our hands, cyclization of 15 and 20 remained
efficient and led only to the desired compounds (1, 2). In
1
(
(
Vol. 8, No. 3, 2004 / Organic Process Research & Development
•
399

Scheme 6. Preparation of Sch 59318 (3) via the chloropurine route
SOCl
2
, quantitative yield) was reacted with the diamine 16
treated with ∼5% MeOH/CH
2 2
Cl . This treatment allowed
prepared in aqueous NaOH, it resulted in the hydrolysis of
the acid chloride and recovery of 33. This was unanticipated
in view of the successful application of several other aliphatic
and aromatic acid chlorides for the preparation of PDE
inhibitors. Hence, aqueous NaOH-free 16 was needed. This
was achieved by neutralizing aqueous NaOH with 10%
for a recovery/removal of highly pure and insoluble 35 which
was then recycled, effectively improving the overall yield.
This recovery procedure proved useful, as in the next step
the reaction of cycloleucine (8) with chloropurine 36 was
very slow, requiring up to 72 h at 110 °C for only ∼35%
isolated yield of 37. In this reaction not only many highly
colored impurities formed, but also most of 36 hydrolyzed
to 35. It appears that the hindered tertiary amine group of 8
requires harsh reaction conditions during which time the
primary alcohol group in it displaces chlorine in 36. The
iminoether generated by the latter reaction then undergoes
ready hydrolysis during TLC, HPLC, or workup, resulting
in the formation of 35. Again, highly insoluble 35 was
recovered from this mixture as described above. Slurries in
THF followed by iPrOH were then necessary to remove
nonpolar and polar colored impurities, respectively, from 37.
8
2 4
aqueous H SO at 0-5 °C under inert atmosphere. This
resulted in a precipitation of 16 which contained ∼1.5 mol
of water. Even this amount of water competed with amida-
tion, causing hydrolysis of 33 acid chloride and resulting in
a mixture of 34 and 33. In the initial phase of the project,
treatment of 16 with 33 and EDCI resolved this issue, leading
to good yield of isolated 34.1
reaction, due to the higher steric demand of 34 (vs 19), each
stage required approximately double the reaction time for
its conversion to chloropuine 36. The conversion yields were
low as all of xanthine 35 could not be converted to 36. After
aqueous precipitation as described above, the mixture was
6-18
As expected, for the POCl
3
2 2
Compound 37 in a mixture of acetonitrile/CH Cl (to improve
solubility) was converted to 38 in high yield with the use of
thionyl chloride (35-40 °C, 4 h). This was then converted
to its mono HCl salt in iPrOH (as for 2) in an 80% yield.
In summary, a common synthetic route involving the
preparation and use of 2-chloropurines as important inter-
mediates was developed and utilized to deliver requisite
quantities of three PDE inhibitors of differing structural
complexity and of GMP quality. Interesting mechanistic and
stereochemical findings were made during the course of this
work.
(
15) Gala, D.; Stamford, A.; Jenkins, J.; Kugelman, M. Org. Process Res. DeV.
997, 1, 163.
16) Subsequently we discovered that treating 16 with the requisite quantity
necessary to consume water as determined by the Karl Fischer method)
of either bis-trimethylsilylacetamide (BSA) or mono-silylmethylacetamide
MSA) prior to addition of acid chloride of 33 allowed for a successful
1
(
(
(
preparation of 34 without the use of EDCI.
(
17) Ahn, H.-S.; Bercovici, A.; Boykow, G.; Bronnenkant, A.; Chackalamannil,
S.; Chow, J.; Cleven, R.; Cook, J.; Czarniecki, M.; Domalski, C.; Fawzi,
A.; Green, M.; Gundes, A.; Ho, G.; Laudicina, M.; Lindo, N.; Ma, K.;
Manna, M.; McKittrick, B.; Mirzai, B.; Nechuta, T.; Neustadt, B.; Puchalski,
C.; Pula, K.; Silverman, L.; Smith, E.; Stamford, A.; Tedesco, R. P.; Tsai,
H.; Tulshian, D.; Vaccaro, H.; Watkins, R. W.; Weng, X.; Witkowski, J.
T.; Xia, Y.; Zhang, H. J. Med. Chem. 1997, 40, 2196.
(
18) Vemulapalli, S.; Watkins, R. W.; Chintala, M.; Davis, H.; Ahn, H.-S.; Fawi,
A.; Tulshian, D.; Chiu, P.; Chatterjeee, M.; Lin, C.-C.; Sybertz, E. J. J.
CardioVasc. Pharmacol. 1996, 28, 862.
Received for review September 22, 2003.
OP030212R
400
•
Vol. 8, No. 3, 2004 / Organic Process Research & Development