Process optimization in the synthesis of 9-[2-(diethylphosphonomethoxy)ethyl]adenine: replacement of sodium hydride with sodium tert-butoxide as the base for oxygen alkylation

Article abstract of DOI:10.1021/op980067v

9-[2-(Diethylphosphonomethoxy)ethyl]adenine (diethyl-PMEA), a key intermediate in the production of the antiviral drug adefovir dipivoxil, was originally produced via a process utilizing sodium hydride (NaH) to couple hydroxyethyl adenine with diethyl p-toluenesulfonyloxymethanephosphonate. The use of NaH presented safety and consistency problems. It was found that sodium tert-butoxide (NaOBu) was a suitable replacement for NaH as the base to effect the coupling reaction. Optimization of reagent stoichiometry and introduction of a simplified filtration workup procedure led to a robust process affording diethyl-PMEA in consistent yields and purities. The modifications and process improvements were scaled-up successfully to batch sizes of >100 kg.

Full text of DOI:10.1021/op980067v

Organic Process Research & Development 1999, 3, 53−55
Process Optimization in the Synthesis of
9-[2-(Diethylphosphonomethoxy)ethyl]adenine: Replacement of Sodium Hydride
with Sodium tert-Butoxide as the Base for Oxygen Alkylation
Richard H. Yu,*,† Lisa M. Schultze,† John C. Rohloff,† Pawel W. Dudzinski,‡,§ and Daphne E. Kelly†,
Process Research, Gilead Sciences, Inc., 353 Lakeside DriVe, Foster City, California 94404, and
Process Chemistry, Raylo Chemicals, Inc., 8045 Argyll Road, Edmonton, Alberta, Canada T6C4A9
Scheme 1. Synthesis of adefovir dipivoxil from adenine
Abstract:
9-[2-(Diethylphosphonomethoxy)ethyl]adenine (diethyl-PMEA),
a key intermediate in the production of the antiviral drug
adefovir dipivoxil, was originally produced via a process
utilizing sodium hydride (NaH) to couple hydroxyethyl adenine
with diethyl p-toluenesulfonyloxymethanephosphonate. The use
of NaH presented safety and consistency problems. It was found
that sodium tert-butoxide (NaOtBu) was a suitable replacement
for NaH as the base to effect the coupling reaction. Optimization
of reagent stoichiometry and introduction of a simplified
filtration workup procedure led to a robust process affording
diethyl-PMEA in consistent yields and purities. The modifica-
tions and process improvements were scaled-up successfully to
batch sizes of >100 kg.
Introduction
Adefovir dipivoxil (9-[2-bis(pivaloyloxymethyl)phospho-
nomethoxyethyl]adenine (8), Scheme 1)1,2a is an orally bio-
available prodrug of 9-[2-(phosphonomethoxy)ethyl]adenine
[PMEA (6)], a nucleotide analogue with activity against the
human immunodeficiency virus, hepatitis B virus, herpes
simplex virus, cytomegalovirus, Epstein-Barr virus, and
other DNA viruses.1b,2-4 Adefovir dipivoxil is currently in
late-phase clinical trials5 for the treatment of HIV and in
early-phase clinical trials for the treatment of hepatitis B
virus.
The supply of adefovir dipivoxil for these programs
depends on a four-step synthetic process (Scheme 1). In the
first step, adenine (1) is condensed with ethylene carbonate
(2) in hot DMF to afford the intermediate 9-(2-hydroxyethyl)-
adenine [HEA (3)] in 83-95% yield after crystallization from
toluene. This step scaled-up well and was reproducible at
the 100-200-kg scale. Similarly, the third step, phosphonate
ester cleavage with bromotrimethylsilane, worked well at
production scale to afford 6, as did the final esterification
of the phosphonate to append the pivaloyloxymethyl groups.
The second step, the synthesis of 9-[2-(diethylphospho-
nomethoxy)ethyl]adenine [diethyl-PMEA (5)] (3 f 5), which
at the outset of this investigation was derived from the
laboratory procedure initially described by Holy´ and Rosen-
berg,6 was problematic on large scale. In this step, the
alkylation of 3 was performed using diethyl p-toluenesulfo-
* To whom correspondence should be addressed. Tel.: (650) 572-6692.
Fax: (650) 573-4787. E-mail: Richard•Yu@gilead.com.
† Gilead Sciences, Inc.
‡ Raylo Chemicals, Inc.
§ Current address: Schering-Plough, Ltd., Singapore Branch, 50 Tuas West
Dr., Singapore 638408.
Current address: Axys Pharmaceuticals, Inc., South San Francisco, CA
94080.
(1) (a) Starrett, J. E., Jr.; Tortolani, D. R.; Russell, J.; Hitchcock, M. J. M.;
Whiterock, V.; Martin, J. C.; Mansuri, M. M. J. Med. Chem. 1994, 37,
1857-1864. (b) Starrett, J. E., Jr.; Tortolani, D. R.; Hitchcock, M. J. M.;
Martin, J. C.; Mansuri, M. M. AntiViral Res. 1992, 19, 267-273.
(2) (a) Benzaria, S.; Pe´licano, H.; Johnson, R.; Maury, G.; Imbach, J.-L.;
Aubertin, A.-M.; Obert, G.; Gosselin, G. J. Med. Chem. 1996, 39, 4958-
4965. (b) Bronson, J. J.; Ghazzouli, I.; Hitchcock, M. J. M.; Russell, J. W.;
Klunk, L. J.; Kern, E. R.; Martin, J. C. Abstracts of Papers, 5th International
Conference on AIDS, Montreal, Canada, 1989; Abstract MCP74. (c)
Naesens, L.; Neyts, J.; Balzarini, J.; Bischofberger, N.; De Clercq, E.
AntiViral Res. 1995, 26, A276. (d) Naesens, L.; Neyts, J.; Balzarini, J.;
Bischofberger, N.; De Clercq, E. AntiViral Res. 1995, 26, A277. (e) Naesens,
L.; Neyts, J.; Balzarini, J.; Bischofberger, N.; De Clercq, E. Nucleosides
Nucleotides 1995, 14, 767-770. (f) Naesens, L.; Balzarini, J.; Bischofberger,
N.; De Clercq, E. Antimicrob. Agents Chemother. 1996, 40, 22-28. (g)
Srinivas, R. V.; Robbins, B. L.; Connelly, M. C.; Gong, Y. F.; Bischofberger,
N.; Fridland, A. Antimicrob. Agents Chemother. 1993, 37, 2247-2250. (h)
Cundy, K. C.; Fishback, J. A.; Shaw, J.-P.; Lee, M. L.; Soike, K. F.; Visor,
G. C.; Lee, W. A. Pharm. Res. 1994, 11, 839-843.
(3) Naesens, L.; Balzarini, J.; De Clercq. E. ReV. Med. Virol. 1994, 4, 147-
159.
(4) De Clercq, E. Biochem. Pharmacol. 1991, 42, 963-972.
(5) (a) Walker, R. E.; Vogel, S. E.; Jaffe, H. S.; Polis, M. A.; Kovacs, J. A.;
Faloon, J.; Davey, R. T.; Ebeling, D.; Cundy, K.; Paar, D.; Markowitz, N.;
Masur, H.; Lane, H. C. 1st National Conference on Human Retroviruses
and Related Infections, Washington, DC, 1993; Abstract 522. (b) AntiViral
Agents Bull. 1994, 7, 232-233. (c) Barditch-Crovo, P. A.; Toole, J.; Burgee,
H.; Wachsman, M.; Ebeling, D.; Cundy, K. C.; Jaffe, H. S.; Lietmand, P.
S. AntiViral Res. 1995, 26, A229.
(6) Holy´, A.; Rosenberg, I. Collect. Czech. Chem. Commun. 1987, 52, 2801-
2809.
10.1021/op980067v CCC: $18.00 © 1999 American Chemical Society and Royal Society of Chemistry
Published on Web 12/17/1998
Vol. 3, No. 1, 1999 / Organic Process Research & Development
•
53

Table 1. Effects of stoichiometry on product purity and
reaction efficiency
nyloxymethanephosphonate (4) and 3.0 equiv of sodium
hydride in DMF. The heterogeneity of the reaction mixture,
in which both NaH and 3 were virtually insoluble, coupled
with a strong, delayed onset exotherm led to large variations
in product purities (83-96%) and yields (17-44%) at
production scale (Table 2). Furthermore, the hydrogen
evolution during the reaction and during the acetic acid
quench of the excess NaH at the end of the reaction caused
safety and operational difficulties. At 200-kg scale (1116 mol
of 3), the NaH process generated 7.51 × 104 L of hydrogen
gas. Dilution of this hydrogen to below the 4% hydrogen
flammability limit required 1.80 × 106 L of nitrogen (2.25
metric tons).7 Additionally, concentrated mixtures of NaH
in DMF can also result in runaway decomposition.8
HPLC assay
reaction
HEA + major
impurities
(area %)
scale NaOtBu tosylate time
entry (g) (equiv) (equiv) (h)
product
(area %)
1
2
3
4
5
6
7
8
9
35.8
35.8
50.0
100.0
35.8
35.8
239.0
35.8
35.8
1.10
1.30
1.50
1.50
1.50
1.75
1.75
2.00
3.00
1.20
1.20
1.40
1.40
1.20
1.25
1.25
1.25
1.25
26.0
23.0
20.0
20.0
11.0
7.5
7.5
10.0
30.0
21.7
17.5
18.5
17.8
14.7
13.4
12.9
11.8
>27.0
46.6
58.7
56.8
57.6
56.0
63.6
55.1
60.4
na
A continuous extraction procedure using methylene chlor-
ide to isolate the partially water soluble diethyl-PMEA (5)
worked well in the pilot plant facility; however, further scale-
up proved arduous, necessitating replacement with a tedious
batch extraction procedure. In addition, the removal of the
sodium tosylate byproduct was variable using this extractive
workup method. Finally, the mineral oil in which the NaH
was dispersed required an additional hexane extraction step
for its removal. Because of these process weaknesses and
concerns with the use of NaH, an effort was initiated to im-
prove the chemistry associated with this step.
Previously, NaH (3.0 equiv) was added portion-wise and
concurrently with a solution of tosylate 4 in DMF at less
than 10 °C to minimize the side reaction of NaH with 4.
Sodium tert-butoxide proved more selective and milder,
requiring only 1.5 equiv to effect the coupling reaction.
Furthermore, the entire amount of base could be charged in
one portion without degradation of 4.
Ranging studies (Table 1) established that the best
stoichiometry was 1.50-2.00 equiv of NaOtBu and 1.20-
1.25 equiv of 4. A final adjustment to 1.75 equiv of
NaOtBu and 1.25 equiv of 4 afforded a completed reaction
(<3% HEA) in under 8 h with yields and purities superior
to those of the NaH process. The use of NaOtBu in the
optimized stoichiometry eliminated the flammability hazards
and the mineral oil associated with the use of NaH, offered
a milder quench with acetic acid, and reduced the amount
of NaOTs and NaOAc generated, simplifying the purification
procedures.
Workup Optimization. The workup optimization cen-
tered on the removal of the salt byproducts in the product
mixture. In the NaH process workup procedure, the salt
byproducts were removed by dissolving the concentrated
reaction mixture in water, extracting with hexanes to remove
the mineral oil from NaH, followed by numerous batch
extractions (>12) with methylene chloride to recover the
product from the aqueous phase.
In the new workup, the use of NaOtBu eliminated the
need for a hexane extraction and an extensive aqueous
workup. After the initial concentration of the reaction mix-
ture, the semisolid product mixture was diluted with meth-
ylene chloride to precipitate the majority of the sodium tosy-
late and sodium acetate salts. Filtration of this mixture effi-
ciently removed the salts, provided that the DMF level after
distillation was <20 wt %10 to limit the solubility of the salts.
Above this level, the salts became soluble, and the filtration
was slowed considerably. A final 9:1 (CH2Cl2/H2O, w/w)
partition of the reconcentrated filtrate was incorporated to
entirely remove trace amounts of dissolved salts and residual
DMF. This ratio of solvents afforded the best separation with
the minimum amount of product retained in the water.
Results and Discussion
Reaction Modifications and Optimization. The modi-
fication of the diethyl-PMEA process focused on replacing
NaH with alternative bases while maintaining a reaction
profile similar to that of the NaH process. Lithium tert-
butoxide (LiOtBu), which previously was found to be
effective for the synthesis of the related compound, PMPA,9
afforded a thick suspension of the lithium salt of 3 in DMF
which hindered agitation and reaction progress. In contrast,
the related base, sodium tert-butoxide (NaOtBu), afforded a
homogeneous mixture on reaction with 3 in DMF. In some
instances, the sodium salt of 3 precipitated out of solution
as a viscous resin as the reaction progressed. Counterintu-
itively, the viscous resin did not dissolve as more solvent
was added. It was discovered that, at high concentration (1.5
M), the sodium salt of 3 was soluble, affording a homoge-
neous solution throughout the coupling reaction, whereas at
lower concentration (0.8-1.0 M), the mixture became
heterogeneous as the sodium salt of 3 precipitated. Aside
from this concentration issue, it was observed in some runs
that large clumps of starting 3 did not react with NaOtBu
and thus remained as floating particulates in the reaction
mixture. To circumvent this issue, an initial digestion at 125-
135 °C in DMF, followed by fast cooling, fully dissolved
and reprecipitated 3 as a fine solid. Once performed, addition
of NaOtBu reliably gave a homogeneous solution.
(7) Perry, R. H.; Green, D. W.; Maloney, J. O. Perry’s Chemical Engineers’
Handbook, 7th ed.; McGraw-Hill: New York, 1997; pp 11-107.
(8) (a) Buckley, J.; Lee, R.; Laird, T.; Ward, R. J. Chem. Eng. News 1982, 60
(28), 5. (b) DeWall, G. Chem. Eng. News 1982, 60 (37), 5 and 43.
(9) Schultze, L. M.; Chapman, H. H.; Dubree, N. J. P.; Jones, R. J.; Kent, K.
M.; Lee, T. T.; Louie, M. S.; Postich, M. J.; Prisbe, E. J.; Rohloff, J. C.;
Yu, R. H. Tetrahedron Lett. 1998, 39, 1853-1856.
(10) A representative sample was obtained after addition of methylene chloride.
The level of DMF was determined by 1H NMR comparing by wt % of
DMF versus NaOTs, NaOAc, and diethyl-PMEA.
54
•
Vol. 3, No. 1, 1999 / Organic Process Research & Development

Table 2. Comparison of large scale results using NaH and
1577 mol) in DMF (241 kg) was slowly added to the reaction
mixture with cooling so as to maintain the content temper-
ature at -10 to 0 °C. After completion of addition, the
mixture was stirred for an additional 1.5 h.
NaOtBu
results
reaction
purity
yield
Acetic acid (134.5 kg, 2240 mol) was charged to the
reaction mixture while the temperature was kept <20 °C.
After the mixture was stirred for 15 min, the contents were
concentrated in vacuo (maximum temperature of 80 °C with
minimum vacuum of 25 mmHg) to remove DMF and
residual solvents until the distillation stopped at a maximum
pot temperature of 80 °C. The concentrate was cooled to 40
°C, diluted with methylene chloride (3201 kg), and charged
with Celite (100.3 kg) before filtration. The solids containing
NaOTs and NaOAc salts were removed by filtration and
rinsed with methylene chloride (3 × 80 kg). The combined
filtrate was concentrated in vacuo until distillation halted at
a maximum pot temperature of 80 °C. The concentrate was
cooled to 40 °C, diluted with methylene chloride (1000 kg),
and reconcentrated in vacuo until distillation halted at a
maximum pot temperature of 80 °C. The concentrate was
cooled to 40 °C, diluted again with methylene chloride (1388
kg) to dissolve all solids, and washed once with water (163.3
kg). The aqueous washing was back-extracted twice with
methylene chloride (300 kg each back-extraction). The
methylene chloride phases were combined and concentrated
in vacuo until distillation halted at a maximum pot temper-
ature of 80 °C. The resulting oil was diluted with toluene
(600 kg) and reconcentrated in vacuo until distillation halted
at a maximum pot temperature of 80 °C. The concentrate
was again diluted with toluene (600 kg) and then warmed
to 80 °C for 30 min with moderate agitation. The solution
was cooled slowly to 30 °C over 90 min and then to 0 °C,
where it was maintained for 12 h with slow agitation. The
white solid product was isolated by filtration, washed with
cold toluene (3 × 35 kg), and dried in vacuo at 50 °C,
yielding 151.6 kg (41.52%) of 5. Purity by HPLC assay was
(HPLC area %)
(% theory)
base
no. of scale (kg
used batches of HEA, 3) av
range
av
range
NaH
10
4
5-249 92.6 82.7-96.1 29.0 17.5-43.7
200 95.0 94.6-95.3 41.0 39.4-43.7
NaOtBu
The isolation of 3 involved concentration of the methylene
chloride layer, a coevaporation with toluene to remove water,
and crystallization from toluene. The isolated laboratory
yields of the NaOtBu procedure were 35-45%, with purities
of 88-95%. On production scale, the yields and purities
(Table 2) were impressively consistent compared to the NaH
process.
Experimental Section
Starting material 3 was prepared according to the literature
procedure.11 Compound 4 was prepared using a method
similar to the literature description12 and was used as a ∼90%
concentrated toluene solution adjusted for purity. NaOtBu
and acetic acid were commercially available and were used
as received. All reactions were performed under a dry
nitrogen atmosphere. Proton nuclear magnetic resonance
spectra were obtained on a Varian spectrometer (300 MHz).
1
The H NMR chemical shifts are expressed in parts per
million (δ) relative to the protio form of the solvent used.
Cooling baths (for laboratory scale) used are as follows: -10
°C (ice/methanol), 0 °C (ice/water/salt). HPLC analyses were
performed on a Rainin Dynamax HPLC system using an
Alltech Alltima 5 µm, reversed-phase column with 5%
MeOH in water as eluent.
9-[2-(Diethylphosphonomethoxy)ethyl]adenine (5). A
suspension of 3 (200.2 kg, 1117 mol) in N,N-dimethylfor-
mamide (DMF) (960 kg, 1.10 M) was heated at 125-135
°C for 30 min. Upon reaching 130 °C, solution was achieved.
The contents were rapidly cooled to 20-30 °C to reprecipi-
tate 3 as a fine suspension in DMF. NaOtBu (188.0 kg, 1956
mol) was charged slowly to the suspension with cooling so
as to maintain the content temperature at 20-30 °C. The
resulting solution was then held at 20-30 °C for 30 min
before being cooled to -10 °C. A solution of 4 (508.4 kg,
1
93-96% by area. Characterization and H NMR data were
consistent with literature values.6
Acknowledgment
The authors gratefully thank Dr. Regan G. Shea of Gilead
Sciences, Inc., for his significant technical contributions and
suggestions in the scale-up production of the diethyl-PMEA
process, and Mr. Ernest J. Prisbe of Gilead Sciences, Inc.,
for his valuable comments and discussions in the reading of
this manuscript. We also thank Mr. Geoff McLennan, Mr.
Alkarim Ebrahim, and Mr. Gary Haesch of Raylo Chemicals,
Inc., for their technical assistance.
(11) (a) Ueda, N.; Kondo, K.; Kono, M.; Takemoto, K.; Imoto, M. Makromol.
Chem. 1968, 120, 13. (b) Raic, S.; Pongracic, M.; Vorkapic-Furac, J.; Vikic-
Topic, D.; Hergold-Brundic, A.; Nagl, A.; Mintas, M. Nucleosides Nucle-
otides 1996, 15, 937-960.
(12) Holy´, A.; Rosenberg, I. Collect. Czech. Chem. Commun. 1982, 47, 3447-
3463. Other methods: (a) Farrington, G. K.; Kumar, A.; Wedler, F. C. J.
Med. Chem. 1985, 28, 1668-1673. (b) Merlo, V.; Roberts, S. M.; Storer,
R.; Bethell, R. C. J. Chem. Soc., Perkin Trans. 1 1994, 11, 1477-1481.
Received for review August 18, 1998.
OP980067V
Vol. 3, No. 1, 1999 / Organic Process Research & Development
•
55