Synthesis of injectable antifungal Sch 59884

Article abstract of DOI:10.1021/op010212w

A synthesis for the preparation of multikilogram quantities of the injectable antifungal Sch 59884 is described.

Full text of DOI:10.1021/op010212w

Organic Process Research & Development 2001, 5, 622−629
Synthesis of Injectable Antifungal Sch 59884
Gary M. Lee, Jefrey Eckert, Dinesh Gala,* Martin Schwartz, Paul Renton, Edward Pergamen, Michael Whttington,
Doris Schumacher, Larry Heimark, and Petia Shipkova
Chemical Process Research and DeVelopment, Schering-Plough Research Institute, 1011 Morris AVenue,
Union, New Jersey 07083, U.S.A.
Abstract:
unstable to the oxidation conditions, the use of these reagents
was deemed incompatible for the preparation of 1. Either
the cost of the commercially available pentavalent phospho-
rylating reagents or the conditions required for the removal
of the protecting groups inherent in them (incompatible with
Sch 59884 stability) deterred us from using them. Since
hydrogenolysis was known to be compatible with Sch
56592,2 we chose dibenzyl phosphate as the preferred
phosphate moiety for the preparation of 1. Initially direct
phosphorylation of 3 with dibenzylchlorophosphate (DBCP,
its preparation described later in this contribution) was
attempted with the expectation that the mixed anhydride
formed from the reaction of 3 and DBCP could eventually
be hydrolyzed to the desired phosphoester 6, or ideally, the
mixed anhydride could be used to for the direct synthesis of
7. In reality these attempts led to a mixture of inseparable
products, and this approach was abandoned. Next, a literature
survey revealed no example for the selective protection of
alcohol function of 3. At this stage conditions reported3 for
the carboxylic acid moiety protection of amino acids were
applied to 3. Thus, reaction of p-nitrobenzylbromide with 3
resulted in the formation of 4. Although the major product
of this reaction was the desired product 4, several intensely
colored byproducts also formed, and purification of 4 from
this reaction mixture via a plant-friendly procedure proved
difficult. Thus, it was decided to progress with crude 4 for
its subsequent coupling with DBCP.
A synthesis for the preparation of multikilogram quantities of
the injectable antifungal Sch 59884 is described.
Opportunistic fungal infections in immune compromised
individuals have resulted in a renewed interest in potent
antifungals. These infections can cause swallowing difficul-
ties in these individuals, and therefore injectable antifungals
are desirable. Sch 59884, 1, is one such potent azole
injectable antifungal,1 and large amounts of this phosphate
were needed to further evaluate its potential as an injectable
antifungal. We describe here our work towards the develop-
ment of a practical synthetic route which allowed for the
preparation of multikilogram batches 1.
Sch 59884 is derived from the oral antifungal Sch 56592,
2.2 This compound is comprised of four stereogenic centers,
two of which emanate from the synthetically challenging 2,4-
substituted tetrahydrofuran moiety. As such, its preparation
is lengthy and expensive. To minimize the economic burden
of the preparation of 1, it was necessary that incorporation
of 2 should come at a late stage, and that the subsequent
synthetic transformations should be very high-yielding. This
reasoning led to the convergent synthesis of 1 which is shown
in Scheme 1.
The availability of 4-hydroxy sodium butyrate, 3, made
it a suitable starting material for the preparation of 1. For
the phosphate moiety, many commercially viable trivalent
as well as pentavalent phosphorylating reagents were con-
sidered. The trivalent reagents phosphorylate well, and they
would require subsequent oxidation. Since Sch 56592 is
DBCP is a commercially unavailable, reactive, unstable
reagent.4,5 Several methods for its preparation have been
reported.4a-e On the basis of our laboratory work it was
concluded that the method described in ref 4a could be scaled
up, it would not require purification of unstable DBCP, and
it did not generate hazardous waste. Hence, this procedure
was optimized for a scale-up in the plant. Calorimetric
evaluation of the literature procedure for its preparation via
the addition of N-chlorosuccinamide (NCS) to a cold (0 °C)
(3) Perich, J. W.; Alewood, P. F.; Johns, R. B. Aust. J. Chem. 1991, 44, 253.
(4) (a) Atherton, F. R. Biochem. Prep. 1957, 5, 1. (b) Asai, S.; Nakamura, H.;
Tanabe, M.; Sakamoto, K. Ind. Eng. Chem. Res. 1994, 33, 1687. (c)
Silverberg, L. J.; Dillon, J. L.; Vemishetti, P.; Sleezer, P. D.; Discordia, R.
P.; Hartung, K. B.; Gao, Q. Org. Process Res. DeV. 2000, 4, 34. (d)
Silverberg, L. J.; Dillon, J. L.; Vemishetti, P. Tetrahedron Lett. 1996, 37,
771. (e) Oza V. B.; Corcoran, R. C. J. Org. Chem. 1995, 60, 3680.
(5) (a) Encyclopedia of Reagents for Organic Synthesis; Paquette, L. A., Ed.;
John Wiley and Sons: West Sussex, England, 1995; Vol. 3, p 1536. (b)
Kenner, G. W.; Todd, A. R.; Weymouth, F. J. J. Chem. Soc. 1952, 3675.
(c) Atherton, F. R.; Howard, H. T.; Todd, A. R. J. Chem. Soc. 1948, 1106.
(d) Personal communication with manufacturer of phosphorous reagents
suggested that traces of benzyl alcohol can trigger an autodecomposition
of DBCP. Thus, it was deemed safe to consume this reagent immediately
after its preparation.
(1) (a) Bennett, F.; Ganguly, A.; Girijavallabhan, V. M.; Patel, N. M.; Saksena,
A. K. New Piperazinyl-phenyl Triazolyl-4-hydroxybutanoate Compound
Useful for Treating Systemic Fungal Infections. U.S. Patent 6,043,245,
March 28, 2000. (b) Girijavalabhan, V. M.; Bennett, F.; Saksena, A. K.;
Lovey, R. G.; Wang, H.; Pike, R. E.; Patel, N.; Loebennerg, D.; Nomier,
A. A.; Lin, C. C.; Ganguly, A. K. Sch 59884, A Water-Soluble Prodrug of
Sch 56592 for Intravenous Formulations. Interscience Conference on
Antimicrobial Agents and Chemotherapy; American Society for Microbi-
ology: Washington, DC, 1999, Paper 1932.
(2) Andrews, D. R.; Gala, D.; Gosteli, J.; Gunter, F.; Mergelsberg, I.; Sudhakar,
A. Process for the Preparation of Triazolones. U.S. Patent 5,625,064, April
29, 1997.
622
•
Vol. 5, No. 6, 2001 / Organic Process Research & Development
10.1021/op010212w CCC: $20.00 © 2001 American Chemical Society and The Royal Society of Chemistry
Published on Web 10/12/2001

Scheme 1. Synthesis of Sch 59884a
a (a) DMF, p-nitrobenzylbromide.( b) Toluene, pyridine, -20 °C. (c) Na2S/THF/H2O/0-5 °C or THF/Zn/HCl/rt. (d) THF/DMAP/TSCl. (e) THF/Pd/C/HCOOH or
THF/Pd/C/H2 60-100 psi, 50-60 °C.
Figure 1. Literature preparation of DBCP.
solution of dibenzyl phosphite in toluene (in place of benzene
which is an undesirable solvent due to its toxicity) showed
that this procedure was unsuitable for large-scale work. This
reaction was quite exothermic with an adiabatic temperature
rise of 56.5 °C, and at this temperature a large amount of
heat was released from this reaction mixture over several
hours past the addition of NCS (Figure 1). In view of this
the reagent was prepared by the controlled addition of
commercially available dibenzyl phosphite to a suspension
of NCS in toluene at 20-25 °C. Under the warmer conditions
the reaction was rapid, and no heat accumulation was seen
in laboratory experiment with single reagent addition (Figure
2). Since the instability of this reagent prevents its purifica-
tion via distillation, it was desirable that the preparation of
this reagent be as clean and complete under the modified
conditions as it was under the literature conditions.
The laboratory reaction was conveniently monitored with
ReactIR (Figure 3) for the consumption of the phosphite as
well as the formation of the phosphate. As shown in the
figure, ReactIR confirmed that the product was stable under
the reaction conditions. Alternatively, 31P NMR can also be
utilized to ascertain the progress of this reaction. This
technique confirmed that the conversion of the phosphite to
the phosphate was quantitative in both cases,8 and that the
latter technique was more convenient for monitoring of the
reaction; hence, it was practiced during the scale-up. In
(6) (a) Na2S: Lamment, S. R.; Ellis, A. I.; Chauvette, R. R.; Kukolja, S. J.
Org. Chem. 1978, 43, 1243. (b) Na2S2O5: Perich, J. W.; Reynolds, E. C.
Synlett 1991, 577. (c) NaS2O4: Guibe-Jample, E.; Wakselman, M. Synth.
Commun. 1982, 12, 219.
(7) (a) Chauvette, R. R.; Pennigton, P. A. J. Med. Chem. 1975, 18, 403. (b)
Pless, J.; Guttmann, S. In Peptides 1966; H.-C. Beyermann, Ed.; North-
Holland: Amsterdam, 1967; p 50.
Vol. 5, No. 6, 2001 / Organic Process Research & Development
•
623

Figure 2. Modified preparation of DBCP.
Figure 3. Laboratory analysis of DBCP formation via ReactIR.
practice this reagent was prepared in situ and consumed
promptly after its preparation. Coupling of DBCP with 4 in
the presence of pyridine progressed well. It was noticed that
the residual NCS, if present, led to chlorinated 7 (absolute
structure unknown). On the other hand, process stressing
showed that residual dibenzyl phosphite, did not generate
new impurities in the subsequent steps. Hence, as noted in
the Experimental Section, for practical purposes NCS was
the limiting reagent in the preparation of DBCP. Regardless
of the method of preparation of 5, a convenient, nonchro-
matographic purification of compound 5 proved difficult, as
it carried the impurities formed during the conversion of 3
to 4, as well as the ones emanating from their reaction with
DBCP. Realizing that compound 6, a carboxylic acid, could
allow for removal of neutral organic impurities by acid/base
treatments, it was decided to use crude 5 for further reaction
even though the purification burden was likely to be
significantly large at this stage. In the early stages of this
work, the sodium sulfide, NaS2O4, and Na2S2O8 procedures6
as well as the Zn/HCl procedure reported in the literature7
were applied for the removal of the p-nitrobenzyl group. On
a small scale the Na2S procedure quickly led to the desired
product, whereas the Zn/HCl procedure appeared irrepro-
ducible, primarily due to the heterogeneity of the reaction
mixture. To quickly meet the initial urgent kilogram need
of the drug substance the former procedure was piloted.
Scale-up of this procedure in the pilot plant required a larger
excess of sodium sulfide compared to that needed for
laboratory results. This resulted in a significant amount of
(8) Under moisture-free conditions, a small amount of the heterogeneous reaction
mixture in toluene was cooled to room temperature and added to d6-DMSO
and subjected to 31P NMR analysis. For this mixture a 10 ppm shift (from
ca. 34-24 ppm) was seen upon conversion of the phosphite to the phosphate.
Typically, the reaction was considered complete when no signal for the
phosphite could be detected (implying >98% conversion).
624
•
Vol. 5, No. 6, 2001 / Organic Process Research & Development

decomposition of the product due to longer than expected
exposure to basic conditions and necessitated a partial
chromatographic purification of 5 prior to its use. It was also
noticed that the removal of the benzyl groups of 7 required
prolonged hydrogenation time when Na2S was used for the
removal of the p-nitrobenzyl group of 5. This was attributed
to residual sulfur-containing impurities that poisoned the
catalyst. Thus, an alternate to Na2S was much needed for a
viable larger-scale process.
THF). This conversion was slow even at elevated temper-
atures (∼60 °C and up to 100 psi H2 pressure). This
conversion proceeded smoothly and reproducibly in the
laboratory via transfer hydrogenation (toluene, 50% water
wet 5% Pd/C, HCOOH). When scaled up in the plant, the
reaction proceeded smoothly; however, the isolated product
consistently showed a high amount of residual sodium.
Control experiments in the laboratory followed by testing
of the catalyst established that sodium salts in the catalyst
were converted to sodium salt of 1 during debenzylation,
resulting in the high sodium content of crude 1.
Thus, the removal of bis-butyrophosphate 9, resulting
from debenzylation of 8, and the removal of sodium in crude
1 was necessary in the plant batches to obtain high-quality
1. It was clear that the removal of sodium would be possible
either via the use of a mineral acid or via the use of a large
excess of organic acids (e.g., sulfonic or phosphoryl acids,
by driving the equilibration of sodium towards the added
acid). During this removal the hydrolysis of 1 back to 2 had
to be rigorously contained, as 2 could interfere with the
dissolution of 1 to a clear injectable solution. After much
work, an aqueous acid slurry procedure was developed which
allowed for the removal of sodium (as sodium chloride) with
minimal hydrolysis of 2 and an excellent recovery of 2.
Similarly, several dozen conditions were evaluated prior
to identifying the aqueous EtOH slurry procedure which
allowed for the removal the bis-phosphate resulting from the
debenzylation of 8. Removal of the above impurities resulted
in >98% crude 1, which at this stage was highly insoluble,
and had 2 as a major impurity. This made the final
crystallization of 1 extremely challenging. Again, a program
had to be undertaken to identify a plant-friendly procedure
for crystallization of 1 which would result in a reproducible
polymorph of 1. A THF/DME procedure was identified,
successfully implemented in the plant to prepare several
multikilogram batches of 1, and is described in the Experi-
mental Section. Compound 1 is a stable solid which can be
stored at room temperature for months.
Ongoing laboratory worked suggested that both com-
pounds 5 and 6 were stable to acidic conditions. This
prompted us to reevaluate the removal of the p-nitrobenzyl
group of 5 via the use of a heterogeneous Zn/HCl procedure.
In fact, under these conditions, crude 5 was converted to 6
in reasonably good yields. Initial addition of HCl to the
reaction mixture generated a large amount of heat and
hydrogen gas. This reaction also produced some polymeric
beads. The initial rapid heat and gas generation issues were
overcome in the plant by efficient heat removal and slow
charging of HCl. The size of the polymeric beads was
controlled by rapid agitation of the reaction mixture.
Although highly pure 6 could not be obtained from this
reaction, suitably pure (92%) 6 was obtained after acid/base
workup, and this was then reacted with 2 to form 7.
Compound 7 is a solid, and it was readily purified by i-PrAc/
heptane or i-PrOH/water precipitation. For the coupling of
compounds similar to acid 6 with 2° alcohol 2 it was known
in our laboratory that the carbodiimide reagents worked well,
but the removal of the byproducts was difficult. Conversion
of the acids to their acid chlorides (with thionyl chloride, or
oxalyl chloride) led to highly colored products, which too
were difficult to purify. In some cases the contaminant
interfered with the removal of the benzyl groups. On the
other hand, the mixed anhydride procedure via the use of
DMAP and TsCl typically led to good quality product in
high yields, and hence this procedure was chosen for the
coupling of 6 with 2.9 This procedure worked in high (90%)
yield both in the laboratory and in the plant. Interestingly
enough, during the scale-up in the plant, up to 3% of
compound 8 also formed at this stage. As the subsequent
publications will show, the formation of this compound
occurs during the esterification stage, and it is not due to a
byproduct that contaminated 6 in pilot plant. Precipitation
of 7 removed most of the impurities that were carried over
in crude 6, but it did not remove 8. This was removed as
the diphosphate 9 after deprotection of the benzyl groups as
discussed below.
Last, it was decided to develop laboratory procedures that
can allow for preparation of water soluble forms of this
compound preferably without the use of organic solvents in
which 1 was soluble (refluxing THF/iPrOH, hot MeOH, CH2-
Cl2).10 Two salts, the disodium and the di-NMG (N-methyl-
D-glucamine) salts were of interest. It was discovered that
free acid 1 could be suspended in water at room temperature
with no detectable degradation for several hours. This fact
was used to make the above salts, whereby 1 and the base
were suspended in water and vigorously stirred. Under these
conditions 1 gradually was converted to its water soluble
salt. The end point was a clear aqueous solution of salts of
1 allowing for the formation of homogeneous salts. Lyo-
(10) Compound 1 was fairly soluble in methylene chloride (an environmentally
restricted solvent), whereas it required large quantities of refluxing THF/
i-PrOH. It was reasonably soluble in warm MeOH; however, solvolysis to
2 was a problem. Furthermore, a strict limit on residual solvents in the
drug substance was anticipated for an injectable product. Thus, water would
be an ideal solvent.
(11) HPLC conditions: Phenomenex Ultracarb ODS (5 µ), 0.02 M aqueous
ammonium acetate:CH3CN::35:65 to 65:35 over 33 min at 1 mL/min;
detection at 254 nm.
The deprotection of 7 in the laboratory was initially
attempted via hydrogenation (50% water wet 5% Pd/C,
(9) This procedure, developed for other Schering compounds, was forwarded
by Drs. W. Leong and D. Andrews. Initial work can be found in: (a)
Brewster, J. H.; Ciotti, C. J. J. Am. Chem. Soc. 1955, 77, 6214. (b) Adams,
W.; Baeza, J.; Liu, J.-C. J. Am. Chem. Soc. 1972, 99, 2000.
Vol. 5, No. 6, 2001 / Organic Process Research & Development
•
625

philization of this solution led to the solid salts. Since the
sodium salt was hygroscopic whereas the di-NMG salt as a
hydrate was stable, the latter was selected as the salt of
choice. For practical purposes 1 was delivered as a crystal-
line, stable, solid drug substance. The preparation of the di-
NMG salt, along with the necessary sterile filtration and
formulation required for an injectable pharmaceutical, was
incorporated in the formulation of 1 by the formulation
group.
In summary, a plant-suitable synthesis of injectable
antifungal Sch 59884 was devised. Unanticipated issues that
surfaced during its scale-up were resolved, and several
kilograms of this compound were prepared.
temperature was maintained below 0 °C. The reaction
mixture was stirred at 0 °C for about 4 h, and then it was
warmed to room temperature over about 8 h and stirred for
an additional 8 h. The reaction was monitored by HPLC,
and when the reaction was complete, the mixture was washed
with 2 × 11 L of 10% sodium chloride solution, 2 × 7.5 L
of 10% sodium bicarbonate solution, and 1.1 L of 10%
sodium chloride solution. The toluene layer was concentrated
in vacuo to about 2 L (batch temperature was not allowed
to exceed 40 °C). The solution of 5 in toluene under nitrogen
atmosphere was stored at room temperature prior to its use
for the subsequent step. Typical yield for the two steps was
70% (1.6 kg, 3.2 mol active 5 in 4.6 kg solution as
determined by HPLC vs a working reference standard) and
was used without purification for subsequent reactions.
1H NMR (CDCl3) δ ) 8.15 (2H, d, J ) 8.8 Hz), 7.45
(2H, d, J ) 8.8 Hz), 5.15 (2H, S), 3.64, t, J ) 9.6 Hz), 2.48
(2H, t, J ) 7.8 Hz), 1.9-1.8 (2H, m), 1.64(1H, s).
Experimental Section
Preparation of 4-O-(Dibenzyl)phosphoryl]-butyric Acid
(6). Preparation of this compound was carried out in four
stages.
Stage A: Preparation of (4-Hydroxybutyric acid)p-
nitrobenzoate (4). A solution of 1.0 kg (4.6 mol) of
4-nitrobenzyl bromide in 3.0 L of dimethylformamide (DMF)
was added slowly to a suspension of 0.7 kg (5.6 mol) of
4-hydroxybutyric acid sodium salt in 2.0 L of DMF while
maintaining a temperature <30 °C. The reaction was heated
to 35-40 °C and the temperature maintained for about 2 h.
The reaction was monitored by TLC (silica gel, 1:1 hexane/
ethyl acetate), and when complete the reaction was cooled
to room temperature and partitioned between 4 L of toluene
and 10 L of water. The addition of water was exothermic,
and the temperature of the reaction mixture was held below
30 °C during addition. After separation of the layers, the
aqueous layer was extracted with an additional 4 L of toluene.
The toluene layers were combined and washed with 2 L of
10% sodium chloride and concentrated in vacuo to about 2
L (the batch temperature was not allowed to exceed 40 °C).
The solution of 4 in toluene was stored at room temperature
under nitrogen for use in Stage C. Typical yield was 0.8 kg
(3.3 mol) of active per 2 L of solvent as determined by
HPLC.
Stage D: Preparation of the Title Compound 6. The
removal of p-nitrobenzyl group from 5 was accomplished
via either procedure (a), which was less desirable for plant
scale, or procedure (b).
(a) Na2S Procedure. A solution of 5 (1.0 kg active, 2.0
mol) in toluene as prepared above was concentrated to
remove most of the solvent via vacuum distillation. The
resultant solution was dissolved in 21 L of tetrahydrofuran
and cooled to 5 °C. A solution of 1.4 kg (6.0 mol) of sodium
sulfide nonahydrate in 12 L of water was added to the batch
while a batch temperature of <10 °C was maintained. The
batch was allowed to react for 1 h and was checked for
completion by TLC. Upon reaction completion, 11 L of
TBME was added, and after thorough agitation the layers
were settled and split. The aqueous layer was washed with
an additional 9 L of TBME. The pH of the aqueous layer
was adjusted to 4.7-4.8 with 5% aqueous HCl while a
temperature of <5 °C was maintained. The aqueous layer
was then extracted twice with 11 L of TBME, and the two
organic washes were combined and distilled to 5 L. The
solution was used “as-is” in the next step.
Stage B: Preparation of Dibenzylchlorophosphate
(DBCP). Two liters (9 mol) of dibenzyl phosphite in 9.4 L
of toluene was added to a suspension of 1.1 kg (8.2 mol) of
N-chlorosuccinimide in 9.4 L of toluene at 25 °C over 6 h
while the temperature was maintained between 20 and 25
°C. The reaction mixture was held between 20 and 25 °C
for 3 h, sampled, and analyzed by 31P NMR. Upon reaction
completion, the solution was filtered under closed, inert
conditions to remove the solid byproduct succinimide and
washed with 3 L of toluene. The resultant filtrate was cooled
to -10 °C and used soon after; its preparation is described
below in Stage C. For its use it was assumed that the yield
for this step was quantitative (2.4 kg, 8.1 mol) in 21 L of
toluene.
Stage C: Preparation of {4-O-[(Dibenzyl)phosphoryl]-
butyric acid}p-nitrobenzyl. (5). A solution of 2.4 kg (8.1
mol) of dibenzyl chlorophosphate in 21 L of toluene
(prepared in Stage B) was cooled to -15 to -20 °C, and
the solution of 4 (prepared in Stage A) followed by 0.67 L
(8.7 mol) of pyridine was added to it while the reaction
(b) Zn/HCl Procedure. Five liters of 4 M HCl was added
slowly to an ice-cooled, stirred solution of 8 L of DMF, 1.0
kg (2.0 mol) of “active” 5, and 0.56 kg of zinc (8.6 mol,
ACS grade 99.8%+, -10+50 mesh). The HCl was added
at a rate such that the temperature of the reaction was e20
°C (note: initial addition of the HCl solution produces the
greatest temperature increase.). Upon reaction completion,
a cloudy yellow solution was produced. Cooling was
maintained for an additional 10 min, and the reaction was
allowed to warm to ambient temperature over 1-4 h. The
reaction was generally allowed to continue overnight, at
which time all of the zinc had dissolved, and the reaction
mixture was a clear yellow solution. Upon reaction comple-
tion as determined by HPLC, 10 L of TBME and 0.5 kg of
Celite was added to the reaction. Sixteen liters of water was
then slowly added to the reaction mixture and stirred for
10-15 min. The resulting suspension was filtered through
a sparkler containing a pad of Celite (0.5 kg) and washed
with an additional 5 L of TBME. The resultant solution was
626
•
Vol. 5, No. 6, 2001 / Organic Process Research & Development

Figure 4.
split, and the aqueous layer was washed with an additional
10 L of TBME. The pH of the aqueous was approximately
1.5-2.5. The organic layers were combined and washed with
10 and 5 L of 10% NaHCO3 respectively. The pH of the
aqueous extracts was between 8.5 and 8.0. The basic aqueous
layers were combined with 5 L of TBME and slowly brought
to pH 4.8-5.0 with 4 N HCl (approximately 4 L). The
solution was stirred 10-15 min while the pH was maintained
at 4.8-5.0. The layers were split, and the aqueous layer was
washed with an additional 5 L of TBME. The final two
organic layers were combined and distilled to 4 L of volume.
The yield of 6 was 66% corrected for purity, the purity was
92% versus working reference standard 9 (which was
prepared by preparative flash chromatography: SiO2 and
gradient 1-2% MeOH in CH2Cl2). This was used without
purification for the next step.
temperature between 0 and 3 °C. The PTS-Cl solution was
added at a rate to maintain the reaction temperature between
0 and 3 °C. The reaction was allowed to warm slowly over
5 h to room temperature, upon which it showed consumption
of 2 as determined by HPLC. Upon reaction completion 7
L of water and 24 L of isopropyl acetate were added to the
reaction mixture. The aqueous layer was brought to a pH of
0.9-1.0 with aqueous 5%HCl, and the layers were separated.
The acidic aqueous layer was extracted with an additional 7
L of isopropyl acetate. The combined organic layers were
washed sequentially with 8 L of a 10% sodium bicarbonate
and 8 L of sodium chloride solution. The resulting organic
solution was then concentrated to approximately 10 L and
treated with 0.1 kg of Darco. The resultant mixture was
refluxed for 10 min, cooled to room temperature, and filtered.
The Darco cake was washed with additional 6 L of isopropyl
acetate to minimize product loss. The organic solution was
reheated to reflux, and 8.6 L of hexanes was added slowly
while a temperature between 85 and 89 °C was maintained.
The reaction was cooled to a temperature between 50 and
55 °C over about 4 h and finally cooled to ambient
temperature over an additional 4 h. The resultant solids were
filtered and washed with 1:1 isopropyl acetate/hexanes and
dried in vacuo at room temperature to give product as an
off-white solid. The isolated yield of 7 was 90% corrected
for purity which was 97.5% as measured against reference
standard.
1H NMR (d6-DMSO) δ ) 7.4-7.3 (10H, m), 5.03 (4H,
d, J ) 8.2 Hz), 3.99 (2H, q, J ) 6.5, 13.6 Hz), 2.27 (2H, t,
J ) 7.3 Hz), 1.8 (2H, m); HRMS FAB MS m/z 365.1154
([M + H]+, calcd 365.11).
Preparation of (-)-2(S)-4[-4[-4[-4[[(R-cis)-5,2,4-
Difluorophenyl)tetrahydro-5-(1H-1,2,4-triazol-1-yl-
methyl)-3-furanyl]methoxy]phenyl]1-piperazinyl]-
phenyl-4,5-dihydro-5-oxo-1H-1,2,4-triazol-1yl]-
1(S)-methylbutyl[[(phenylmethoxy)carbonyl]amino]-
acetate (7). p-Toluene sulfonyl chloride (PTS-Cl) [0.43 kg
(1.85 mequiv, molar equivalent compared to 2)] dissolved
in 3.5 L of tetrahydrofuran was added to a solution of 1.0
kg (1.0 mequiv) of 2, 0.73 kg (1.40 mequiv) of 6 (its solution
weight was 7.4 kg), 0.56 kg (4.0 mequiv) of of (dimethyl-
amino)pyridine in 11 L of tetrahydrofuran cooled to a
1H NMR (DMSO) δ ) 8.41 (1H, s), 83.8 (1H, s), 7.84
(1H, s), 7.53 (2H, d, J ) 9.0 Hz), 7.5-7.3 (12H, m), 7.11
(2H, d, J ) 9.2 Hz), 7.05 (1H, dt, J ) 2.3, 8.5 Hz), 6.99
(2H, d, J ) 9.2 Hz), 6.86 (2H, d, J ) 9.1 Hz), 4.86 (2H, dd,
Vol. 5, No. 6, 2001 / Organic Process Research & Development
•
627

Figure 5.
J ) 14.6, 21.6 Hz), 4.2-4.1 (2H, m), 3.96 (2H, q, J ) 6.5,
13.3 Hz), 3.8 (2H, m), 3.7 (1H, m), 3.39 (3H, s), 3.3 (3H,
m), 3.2 (3H, m), 2.6 (1H, m), 2.5-2.4 (1H, m), 2.3 (2H,
m), 2.19 (1H, dd, J ) 8.0, 13.1 Hz), 1.8-1.7 (3H, m), 1.27
(3H, d, J ) 6.4 Hz), 0.83 (3H, t, J ) 7.2 Hz). Anal. Calcd
for (C55H61O9N8F2P): C, 63.09; H, 5.87; N, 10.70. Found:
C, 63.24; H, 5.94; N, 10.73. MS (FAB) 1047 (M + H)+.
Preparation of 2,5,-Anhydro-1,3,4-trideoxy-2-C-(2,4-
difluorophenyl)-4-[[4-[4-[4-[1-[(1S,2S)-1-ethyl-2-[1-oxo-4-
(phosphonooxy)butoxy]propyl]-1,5-dihydro-5-oxo-4H-
1,2,4-triazol-4-yl]phenyl]-1-piperazinyl]phenoxy]methyl]-
1-(1H-1,2,4-triazol-1yl)-D-threopentitol (1). Under nitrogen,
1.0 kg of 7, and 0.5 kg of 50% water wet 5% Pd/C were
suspended in 10 L of toluene and cooled to 15 °C. One liter
of formic acid was added to the reaction mixture over 15-
20 min while the reaction temperature was maintained
between 15 and 20 °C. The biphasic mixture was vigorously
agitated for an additional 3 h (NOTE: The reaction time
was heavily dependent upon which method was used to
deprotect the PNB ester. If the Na2S procedure was used,
the reaction must be stirred for 18 h to ensure completion).
After confirming reaction completion via HPLC analysis,
10 L of i-PrOH and 0.2 kg of siliceous earth were added to
the reaction mixture, which was then filtered through 0.45
µm filter to obtain a clear solution. The solution was next
filtered through a 0.2 µm pore-sized filter. Both filters were
washed with a solution of 0.05 L of formic acid in 6 L of
i-PrOH to minimize product loss. This solution was con-
centrated via distillation under atmospheric pressure until the
internal temperature reached 100 °C (the volume was
approximately 3 L), and then it was cooled to 75 °C. A
solution of 0.1 L of THF in 10L of i-PrOH was separately
warmed to 65 °C and transferred to the hot reaction solution.
The resulting solution was held at 65 °C for 30 min and
then cooled to 55 °C and held for 1 h. The solution was
cooled to 0 °C slowly over 3 h, during which the product
crystallized. The batch was collected by filtration. This was
washed with 5 L of cold i-PrOH. The wet cake was sucked
dry and checked for sodium content. If high sodium levels
were found, the product was reslurried in 5 L of dilute
aqueous HCl (to pH 2.0) at 5-10 °C for 30 min, refiltered,
and washed sequentially with water and IPA. The wet cake
was sucked dry, transferred to a vacuum oven, and dried
until the i-PrOH content was <0.2% (LOD). Isolated yields
ranged from 0.65 to 0.67 kg (79-81%). See below for
analytical data.
Removal of Na and the Diphosphate impurity 9. Fifteen
liters of water heated to 65 ( 5 °C was added to a suspension
of 1.0 kg of 1 in 10.4 L of 2B ethanol heated to reflux. The
temperature of the mixture was maintained at 70 °C or above
throughout this addition. The addition of water resulted in a
clear solution. The solution was cooled to 50-55 °C over
approximately 1.5 h, held at this temperature for 1.5 h, and
seeded if necessary for crystallization. Upon crystallization
the mixture was cooled to room temperature and stirred for
12 h. The solids were filtered and washed with water and
dried under reduced pressure at 40-50 °C until the ethanol
content was less than 0.2% w/w. The solids were stirred as
a slurry in 20 L of water, and the pH was adjusted to 2.9 (
0.1 with 5% HCl. The slurry was agitated slowly overnight,
628
•
Vol. 5, No. 6, 2001 / Organic Process Research & Development

filtered, and washed with water until a pH of >4.0 was
reached for the wash. The wet cake was transferred to a
vacuum oven and dried at 40-45 °C.
filtered through a 0.45 µm pore-sized filter. The clear solution
was then flash frozen (dry ice/2-propanol) and lyophylized
overnight to afford 143 g (113 mmol, 99%) of the di-NMG
salt of 59884. MS (calcd for M + H): 1258.3; (obsd):
1258.4 (main fragment at 867 for parent acid). Elem. Anal.
(calcd for monohydrate): 51.80 C, 6.72 H, 10.98 N; (obsd):
51.51 C, 6.70 H, 10.83 N.
Recrystallization of 1. Under nitrogen, for 1.0 kg of 1,
20 L of THF was used. The mixture of 1 and THF was stirred
and heated to reflux for 30 min to ensure complete dissolu-
tion. The mixture was recooled to ambient temperature, and
0.05 kg of activated carbon and 0.75 kg of siliceous earth
were added. The mixture was reheated to reflux for 30 min.
The mixture was cooled to 40 °C and filtered through 0.45
µm followed by 0.2 µm pore-sized filters. The reaction vessel
and the filters were washed with 5 L of THF to maximize
recovery. The combined solution was concentrated at atmo-
spheric pressure to a volume of approximately 5 L. Sepa-
rately, 30 L of 1,2-dimethoxyethane (DME) was heated to
55 °C, passed through a 0.2 µm pore size filter, and added
to the above solution. The resultant solution was reheated
to reflux for 30 to 60 min, during which the product
crystallized. The mixture was then slowly cooled to room
temperature over 6-8 h, followed by cooling to 0-5 °C for
1 h. The product was collected by filtration. The reaction
mixture and the filter cake were rinsed with an additional 5
L of DME. The filter cake was transferred to a draft oven
where it was dried under reduced pressure at 30 °C until the
residual solvent analysis was <0.2% (GC). Isolated yields
ranged from 0.76 to 0.82 kg (76-82%); mp 175-176 °C.
[RD]) -54.2° (c)10 mg/mL, CH2Cl2); 1H NMR (d6-
DMSO): 8.35 (s, 1H); 8.32 (s, 1H); 7.79 (s, 1H); 7.45 (m,
2H); 7.30 (m, 2H); 7.10 (m, 2H); 6.95 (m, 3H); 6.80 (m,
2H); 5.08 (m, 1H); 4.60 (m, 2H); 4.05 (m, 2H); 3.70 (m,
5H); 3.30 (m, 4H); 3.15 (m, 4H); 2.55 (m, 1H); 2.40 (m,
1H); 2.30 (m, 2H); 2.15 (m, 1H); 1.73 (m, 4H); 1.25 (m
3H); 0.80 (m, 3H). HRMS (calcd for M + H): 867.3406;
(obsd): 867.3428; Elem. Anal. (calcd): 56.81 C, 5.70 H,
12.93 N; (obsd): 56.55 C, 5.73 H, 12.70 N.
Identification of Compounds 8 and 9. Liquid chroma-
tography/mass spectrometry (LC/MS) analyses to identify
process-related impurities were performed on a Sciex API3
mass spectrometer operated in the positive ion electrospray
mode. For all LC/MS analyses, the orifice potential was set
to 80 v. Tandem MS spectra were acquired using argon as
the collision gas at CGT ) 170 and the orifice potential
set to 60 v. Liquid chromatography was performed using
Shimadzu 10-AD HPLC pumps and a YMC ODS-A 4.6 mm
× 150 mm column. Mobile phase A was 20 mM ammonium
acetate, and mobile phase B was acetonitrile. A linear
gradient from 30 to 65% B in 15 min at a flow rate of 1
mL/min was used to elute components off the column.
LC/MS analysis of 7 (tR 29.0 min, MW 1046) detected 8
(Figure 4) as a significant impurity eluting at 33.1 min. The
molecular weight of 1392 Da recorded for 8 is 346 Da greater
than that for 7, suggesting an additional butyrodibenzyl
phosphate group. The MS/MS fragmentation spectrum shows
loss of two dibenzyl phosphate groups to give an ion at m/z
837. Further fragmentation of m/z 837 at the ester bond
generates ion m/z 701 representing the alcohol 2. The
difference of 136 mass units between ion m/z 837 and 701
corresponds to the mass expected for the bis-butyro group.
Debenzylation of 8 generated the bis-butyrophosphate 9 (see
Figure 5) with a molecular weight of 1032 Da. The MS/MS
fragmentation spectrum showing sequential loss of two
phosphate groups to give ions m/z 935 and 837 followed by
loss of the bis-butyro group to give ion m/z 701 confirmed
9 as the bis-butyrophosphate.
Preparation of Water-Soluble Salt of 1: Sch 59884‚2
NMG. In a clean, dry flask, 100 g of 1 (115 mmol) and 45
g of N-methyl-D-glucamine (NMG, 231 mmol) were com-
bined and diluted with 1.45 L of H2O. The contents were
agitated until all solids dissolved (30 min). The solution was
Received for review April 23, 2001.
OP010212W
Vol. 5, No. 6, 2001 / Organic Process Research & Development
•
629