Development of a second-generation, highly efficient manufacturing route for the HIV integrase inhibitor raltegravir potassium

Article abstract of DOI:10.1021/op100257r

A manufacturing route for the synthesis of raltegravir potassium 1 was developed via a thermal rearrangement of amidoxime DMAD adducts 6 to construct the key, highly functionalized hydroxypyrimidinone core 7. Utilizing this route 1 was prepared in nine linear chemical steps with 22% overall yield. A second-generation synthesis was subsequently developed that solved the key chemical, productivity, and environmental impact issues of the initial synthesis. Highlights of the new synthesis include a highly selective methylation, 3-4-fold higher productivity, and a 65% reduction of combined organic and aqueous waste produced. The efficient second-generation manufacturing route provides raltegravir potassium 1 in 35% overall yield.

Full text of DOI:10.1021/op100257r

Organic Process Research & Development 2011, 15, 73–83
Development of a Second-Generation, Highly Efficient Manufacturing Route for the
HIV Integrase Inhibitor Raltegravir Potassium
Guy R. Humphrey,* Philip J. Pye,*,§ Yong-Li Zhong,* Remy Angelaud,† David Askin,† Kevin M. Belyk, Peter E. Maligres,
Danny E. Mancheno,‡ Ross A. Miller, Robert A. Reamer, and Steven A. Weissman⊥
Department of Process Research, Merck Research Laboratories, P.O. Box 2000, Rahway, New Jersey 07065, United States
Abstract:
raltegravir combination to the preferred regimens for treatment-
naive HIV patients.4
A manufacturing route for the synthesis of raltegravir potassium
1 was developed via a thermal rearrangement of amidoxime
DMAD adducts 6 to construct the key, highly functionalized
hydroxypyrimidinone core 7. Utilizing this route 1 was prepared
in nine linear chemical steps with 22% overall yield. A second-
generation synthesis was subsequently developed that solved the
key chemical, productivity, and environmental impact issues of
the initial synthesis. Highlights of the new synthesis include a highly
selective methylation, 3-4-fold higher productivity, and a 65%
reduction of combined organic and aqueous waste produced. The
efficient second-generation manufacturing route provides ralte-
gravir potassium 1 in 35% overall yield.
The medicinal chemistry synthesis of free phenol 1 was
carried out in 10 linear steps in 3% overall yield (Scheme 1).5
Strecker reaction of acetone cyanohydrin 2 followed by Cbz-
protection of 3 and hydroxylamine addition to the nitrile group
gave amidoxime 5. The subsequent two-component coupling
reaction between 5 and dimethyl acetylenedicarboxylate (DMAD)
afforded a Z/E mixture of adducts 6. Thermal rearrangement
of 6 provided the key hydroxypyrimidinone intermediate 7.
Selective hydroxyl group protection, N-methylation using
dimethylsulfate and lithium hydride then hydrogenation to
cleave the Cbz-group, gave free amine 10. Amide coupling of
10 with oxadiazole carbonyl chloride 11 provided intermediate
12, which was converted to raltegravir free phenol 1 by
amidation with 4-fluorobenzylamine (4-FBA).
The synthesis utilizes inexpensive starting materials and
reagents and constructs the highly functionalized hydroxypy-
rimidinone core 7 in an efficient and atom-economical fashion.
However, several steps, including the Strecker reaction, the
thermal rearrangement, the unselective N-methylation, and the
final amidation gave low yields, were therefore unproductive,
and required substantial optimization for a multikilogram route.
Repeated use of solvents such as chloroform, dichloromethane,
and 1,4-dioxane was undesirable for large-scale use due to
environmental concerns, toxicity, and cost.
Introduction
HIV/AIDS remains a significant cause of morbidity and
mortality worldwide. The Joint United Nations Program on
HIV/AIDS estimated in 2007 that more than 33 million people
lived with the disease worldwide and that AIDS killed more
than 2 million people, including 330,000 children.1,2 In the
United States alone, the Centers for Disease Control estimated
in 2006 that more than a million people were living with HIV
infection.3 Raltegravir potassium, or Isentress (1), was the first
FDA-approved inhibitor of HIV integrase. Launched in 2007,
raltegravir was originally indicated for combination therapy with
other antiretroviral agents in treatment-experienced adults with
evidence of viral replication and multidrug-resistant HIV-1
strains. In July 2009, the FDA approved an expanded indication
for raltegravir to include treatment-naive adult patients, and in
December 2009, the United States Department of Health and
Human Services revised its HIV treatment guidelines to add a
While suitable, with minor modifications, for a small drug
delivery to enable the start of initial safety assessment work,
we desired a higher-yielding, more environmentally friendly
and robust route for large-scale manufacturing purposes.
Results and Discussion
First-Generation Process for the Synthesis of 1. To
address the aforementioned chemical and environmental issues,
we began by investigating the synthesis of the key hydroxy-
pyrimidinone 7. The synthesis of amidoxime 5 was optimized
* Authorsforcorrespondence.E-mail:guy_humphrey@merck.com;yongli_zhong@
merck.com.
(4) Guidelines for the use of antiretroviral agents in HIV-1-infected adults
and adolescents. Panel on Antiretroviral Guidelines for Adults and
Adolescents; U.S. Department of Health and Human Services.
December 1, 2009; pp 1-161. Available at http://www.aidsinfo.nih.
gov/ContentFiles/AdultandAdolescentGL.pdf.
(5) (a) Summa, V.; Petrocchi, A.; Bonelli, F.; Crescenzi, B.; Donghi, M.;
Ferrara, M.; Fiore, F.; Gardelli, C.; Gonzalez Paz, O.; Hazuda, D. J.;
Jones, P.; Kinzel, O.; Laufer, R.; Monteagudo, E.; Muraglia, E.; Nizi,
E.; Orvieto, F.; Pace, P.; Pescatore, G.; Scarpelli, R.; Stillmock, K.;
Witmer, M. V.; Rowley, M. J. J. Med. Chem. 2008, 51, 5843–5855.
(b) Belyk, K. M.; Morrison, H. G.; Jones, P.; Summa, V. Preparation
of N-(4-fluorobenzyl)-5-hydroxy-1-methyl-2-(1-methyl-1-{[(5-methyl-
1,3,4-oxadiazol-2-yl)carbonyl]amino}ethyl)-6-oxo-1,6-dihydropyrimi-
dine-4-carboxamide potassium salts as HIV integrase inhibitors. PCT
Int. Appl. WO/2006/060712, 2006.
† Current address: Department of Process Chemistry, Genentech MS 432a, 1
DNA Way, South San Francisco, CA 94080-4990, United States.
‡ Current address: Department of Chemistry, Emory University, 1515 Dickey
Drive, Atlanta, GA 30329, United States.
§ Current address: Johnson and Johnson PRD, Turnhoutseweg 30, B-2340
Beerse, Belgium.
⊥ Current address: Concert Pharmaceuticals, Inc., 99 Hayden Avenue, Suite
500, Lexington, MA 02421, United States.
(1) Kallings, L. O. J. Intern. Med. 2008, 263 (3), 218–243.
(2) AIDS epidemic update: December 2007 Joint United Nations Program
on HIV/AIDS (UNAIDS) WHO; UNAIDS/07.27E/JC1322E; ISBN: 978
92 9 173621 8, 2007
(3) HIV/AIDS SurVeillance Report - 2006; Centers for Disease Control
and Prevention, 2008; Vol. 18.
10.1021/op100257r  2011 American Chemical Society
Published on Web 11/16/2010
Vol. 15, No. 1, 2011 / Organic Process Research & Development
•
73

Scheme 1. Medicinal chemistry route for the synthesis of free phenol 1
Scheme 2. Optimized synthesis of amidoxime 5
Scheme 3. Synthetic methods for the construction of hydroxypyrimidinone 7
from acetone cyanohydrin 2. Under the new conditions, the
Strecker reaction was achieved in 99% yield using 1.5 equiv
of neat liquid ammonia at 30 psi and 15 °C. Aminonitrile 3
was converted to the N-Cbz-protected intermediate 4 by
treatment with benzyl chloroformate in 90% assay yield.
Hydroxylamine addition to the nitrile 4 provided amidoxime 5
as a crystalline solid in 91% isolated yield. The overall isolated
yield for the three-step sequence to 5 was increased from 37%
to 81% (Scheme 2).
With the amidoxime 5 in hand, the synthesis of hydroxy-
pyrimidinone 7 was investigated. A survey of the literature
showed two ways of constructing the hydroxypyrimidinone
core, both methods starting from an amidoxime intermediate
(Scheme 3). Route-a involved a three-step sequence that
included hydrogenation of 13 to prepare amidine 14. Claisen
condensation of commercially available R-benzyloxy acetate
and methyl tert-butyl oxalate provided the dihydroxyfumarate
derivative 15, and subsequent condensation between 14 and 15
provided the benzyl-protected pyrimidinone 17.6 However, the
moderate yield for this sequence, the stability concerns with
15, and the overall poor atom-economy were not suitable for
our long-term synthesis. On the other hand, route-b starts with
a Michael reaction between 13 and DMAD, followed by a
thermal rearrangement to hydroxypyrimidinone 18.7 Although
the initial yield for the thermal rearrangement step was only
41%, the high atom-economy and simplicity were very attractive
74
•
Vol. 15, No. 1, 2011 / Organic Process Research & Development

Table 1. Optimization of stereoselective amidoxime 5
addition to DMAD
additive
(mol %)
conversion
(%)
entry solvent
T (°C)
Z/E ratioa
1
2
3
4
5
6
7
8
9
DMF
N/A
rt
100
100
100
70
100
100
100
30
2:98
1:99
DMSO
MeCN
DCM
N/A
N/A
N/A
rt
rt
rt
80
rt
-10
rt
20:80
31:69
21:79
65:35
75:25
50:50
79:21
74:26
Xylenes N/A
MeOH
MeOH
AcOH
THF
N/A
N/A
N/A
DABCO (15) -70
100
100
10 1,2-DCE DABCO (5)
rt
Figure 1
a Z/E ratio determined by NMR and HPLC analysis.
from a long-term route development perspective and warranted
further investigation.
such as DMF and DMSO (Table 1, entries 1 and 2), 6Z was
obtained as the major component in methanol (entry 6). The
stereoselectivity of 6Z vs 6E was improved in MeOH at lower
temperature (entry 7). A similar result was observed when the
reaction was run in THF or dichloroethane in the presence of
catalytic amount DABCO (entries 9 and 10). Isomerization of
6E to 6Z was unsuccessful under typical double bond isomer-
ization conditions such as heating with catalytic acids, bases,
or iodine/trialkylphosphines.10 The Z/E stereoselectivity was
lower using diethyl acetylenedicarboxylate instead of DMAD.
Although the highest ratio of 6Z to 6E was obtained in THF
at -70 °C with a catalytic amount of DABCO, further
experiments showed that DABCO must be removed prior to
the thermal rearrangement.11 Due to the minimal impact on
overall yield and ease of processing, the Michael addition was
run in methanol at 15-20 °C to afford a mixture of the Z- and
E-adducts 6 in quantitative yield. The resulting solution of
adducts 6 was solvent-switched to xylenes, then heated at 125
°C for 2 h, and at 135 °C for 4 h to give a 62% assay yield of
desired product 7. The reaction mixture was concentrated and
Addition of amidoxime 5 to DMAD typically afforded a
mixture of E/Z adducts 6. A solvent screen showed that, with
nonpolar solvents such as o-xylene or xylenes,8 the rearrange-
ment reaction provided a higher assay yield and proved to be
a better solvent choice. Initial experiments for the thermal
rearrangement of a mixture of E- and Z-adducts 6 in o-xylene
at 125 °C indicated that one of the isomers was consumed faster
than the other. In order to understand the chemistry, the two
isomers were separated using chromatography and the structure
of Z-adduct 6Z determined by the X-ray crystallography (Figure
1). The E- and Z-adducts were separately subjected to the
thermal rearrangement conditions (Scheme 4). The thermal
rearrangement of 6Z was completed at a lower temperature (125
°C) and gave a 72% assay yield of product 7. On the other
hand, thermal rearrangement of 6E required higher temperature
(135 °C) to reach full conversion and afforded a lower assay
yield.9
In order to improve the yield and minimize the reaction time
and temperature, Z-selective amidoxime addition was investi-
gated. While 6E was favored in strongly aprotic polar solvents
Scheme 4. Thermal rearrangement of E- and Z-amidoxime-DMAD adducts 6E and 6Z
Vol. 15, No. 1, 2011 / Organic Process Research & Development
•
75

Scheme 5. Synthesis of hydroxypyrimidinone core 7
Scheme 6. Direct N-methylation of hydropyrimidinone 7
hydroxypyrimidinone 7 obtained by direct crystallization in 54%
isolated yield as a white crystalline solid (Scheme 5).12
With the key hydroxypyrimidinone 7 in hand, optimization
of the N-methylation was performed. In the initial medicinal
chemistry route the phenolic OH was protected as a benzoate
ester prior to N-methylation. In order to minimize the use of
protecting groups, the direct N-methylation of hydroxypyrimi-
dinone 7 was investigated. Unexpectedly, methylation of 7 gave
a clean mixture of desired N-methyl product 22 and undesired
O-methyl byproduct 23 as a 70:30 mixture (path-b, Scheme
6). HPLC and NMR analysis revealed that methyl ethers 19-21
were not formed in this reaction (path-a, Scheme 6).
The optimization of the methylation reaction was further
investigated to improve the regioselectivity and yield and to
eliminate the use of 1,4-dioxane and LiH (Table 2). Interest-
ingly, the selectivity was significantly improved by addition of
1 equiv of MgBr2 ·OEt2 to the reaction in DMF (entry 6). The
study showed the best performance came from the use of methyl
iodide and Mg(OMe)2 in DMSO at 60 °C (entry 13). Under
these conditions, 22 was obtained in high purity (<1% 23) in a
70% isolated yield, after isolation by crystallization from
aqueous DMSO. A possible mechanism involving chelation
between the hydroxypyrimidinone 7 and Mg(OMe)2 was
proposed (24 and 25 in Scheme 7).
(6) (a) Johnson, T. B.; Caldwell, W. T. J. Am. Chem. Soc. 1929, 51, 873–
880. (b) Budesinsky, I.; Jelinek, V.; Prikryl, J. J. Collect. Czech. Chem.
Commun. 1962, 27, 2550–2560. (c) Sunderland, C. J.; Botta, M.; Aime,
S.; Raymond, K. N. Inorg. Chem. 2001, 40, 6746–6756. (d) Dreher,
S. D.; Ikemoto, N.; Gresham, V.; Liu, J.; Dormer, P. G.; Balsells, J.;
Mathre, D.; Novak, T. J.; Armstrong, J. D., III Tetrahedron Lett. 2004,
45, 6023–6025. (e) Employed this three-step sequence, the hydroxy-
pyrimidinones were prepared in 38-42% overall yields from 3 as
shown below:
The initial medicinal chemistry route installed the oxadiazole
fragment prior to introduction of the 4-fluorobenzylamine (4-
FBA). Chromatography was required, and the product was
obtained in a low 37% overall yield. In an attempt to improve
the chemical yield, amidation of 22 with 4-FBA prior to
(7) (a) Culbertson, T. P. J. Heterocycl. Chem. 1979, 16, 1423. (b) Summa,
V.; Petrocchi, A.; Matassa, V. G.; Taliani, M.; Laufer, R.; Franasco,
R. D.; Altamura, S.; Pace, P. J. Med. Chem. 2004, 47, 5336–5339.
(c) Stansfield, I.; Avolio, S.; Colarusso, S.; Gennari, N.; Narjes, F.;
Pacini, B.; Ponzi, S.; Harper, S. Bioorg. Med. Chem. Lett. 2004, 14,
5085–5088. (d) Wagner, E.; Becan, L.; Nowakowska, E. Bioorg. Med.
Chem. Lett. 2004, 12, 265–272. (e) Zhong, Y.-L.; Zhou, H.; Gauthier,
D. R., Jr.; Askin, D. Tetrahedron Lett. 2006, 47, 1315. (f) Colarusso,
S.; Attenni, B.; Avolio, S.; Malancona, S.; Harper, S.; Altamura, S.;
Koch, U.; Narjes, F. ARKIVOC 2006, (vii), 479. (g) Koch, U.; Attenni,
B.; Malancona, S.; Colarusso, S.; Conte, I.; Di Filippo, M.; Harper,
S.; Pacini, B.; Giomini, C.; Thomas, S.; Incitti, I.; Tomei, L.; De
Francesco, R.; Altamura, S.; Matassa, V. G.; Narjes, F. J. Med. Chem.
2006, 49, 1693. (h) Ferrara, M.; Crescenzi, B.; Donghi, M.; Muraglia,
E.; Nizi, E.; Pesci, S.; Summa, V.; Gardelli, C. Tetrahedron Lett. 2007,
48, 8379. (i) Zhong, Y.-L.; Pipik, B.; Lee, J.; Kohmura, Y.; Okada,
S.; Igawa, K.; Kadowaki, C.; Takezawa, A.; Kato, S.; Conlon, D.;
Zhou, H.; King, A. O.; Reamer, R. A.; Gauthier, D. R., Jr.; Askin, D.
Org. Process Res. DeV. 2008, 12, 1245–1252. (j) Naidu, B. N. Synlett.
2008, 547–550. (k) Pacini, B.; Avolio, S.; Ercolani, C.; Koch, U.;
Migliaccio, G.; Narjes, F.; Pacini, L.; Tomei, L.; Harper, S. Bioorg.
Med. Chem. Lett. 2009, 19, 6245–6249.
(10) Treatment of 6 with sodium methoxide afforded a Z/E mixture of
methyl [3-(2-{[(benzyloxy)carbonyl]amino}propan-2-yl)-5-oxo-4,5-
dihydro-6H-1,2,4-oxadiazin-6-ylidene]ethanoate, which could not be
converted to hydroxypyrimidinone 7 in the typical thermal rearrange-
ment conditions and decomposed at higher temperature.
(11) In one instance the thermal isomerization in the presence of DABCO
resulted in an uncontrolled exotherm causing product decomposition
and low yield.
(12) About 5% of the imidazole was generated from the reaction.
(8) For solvent screening on the reaction, please see: reference 7e.
(9) For details of the thermal rearrangement mechanism of 6, please see:
Pye, P. J.; Zhong, Y.-L.; Jones, G. O.; Reamer, R. A.; Houk, K. N.;
Askin, D. Angew. Chem., Int. Ed. 2008, 47, 4134.
76
•
Vol. 15, No. 1, 2011 / Organic Process Research & Development

Table 2. Optimization of direct methylation of hydroxypyrimidinone 7
entry
reagent
base
solvent
additive
T (°C)
conversion (%)
22:23 ratioa
1
Me2SO4
Me2SO4
Me2SO4
Me2SO4
Me2SO4
Me2SO4
MeI
LiOBut
MeOH
Toluene
THF
N/A
50
50
50
50
80
80
80
20
35
20
20
35
60
25
40
74
16:84
48:52
47:53
12:88
9:91
48:52
36:64
52:48
64:36
63:37
61:39
78:22
78:22
2
LiOBut
N/A
3
LiOBut
N/A
4
LiOBut
DMSO
DMF
N/A
100
100
86
5
LiOBut
N/A
6
LiOBut
DMF
MgBr2 ·OEt2
7
LiOBut
DMF
N/A
65
8
MeI
Mg(OMe)2
Mg(OMe)2
Mg(OMe)2
Mg(OMe)2
Mg(OMe)2
Mg(OMe)2
THF
N/A
16
9
MeI
DMF
N/A
58
10
11
12
13
MeI
NMP
N/A
70
MeI
DMAc
DMSO
DMSO
N/A
68
MeI
N/A
56
MeI
N/A
100
a Determined by 1H NMR and HPLC analysis.
Scheme 7. Proposed chelation of 7 with Mg2+
Scheme 8. Synthesis of free amine 27
Scheme 9. Preparation of oxadiazole carbonyl chloride fragment 11
installing the oxadiazole moiety was investigated. Thus, treat-
ment of intermediate 22 with 2.2 equiv of 4-FBA in EtOH at
72 °C afforded 26 in 90% isolated yield (Scheme 8). Hydro-
genation of 26 in the presence of 5% of Pd/C and 1 equiv of
MsOH, to efficiently remove the Cbz-protected group, followed
by neutralization of the crude reaction mixture with NaOH,
afforded crystalline free amine 27 in 99% isolated yield as a
monohydrate. Unfortunately, the hydrate was very stable and
conventional oven-drying techniques failed to provide anhy-
drous amine suitable for the acylation reaction. Thus, azeotropic
water removal using THF was used to dry the amine prior to
use in the acylation reaction.
Modification of a literature procedure13 provided an efficient
process for the oxadiazole fragment (Scheme 9). Acylation of
tetrazole 28 with ethyl oxalyl chloride afforded intermediate
29, which was heated at 70 °C to form the oxadiazole ester.
Direct KOH-mediated hydrolysis afforded the corresponding
potassium salt 30 as a crystalline solid, in 91% overall yield.
Treatment of 30 with oxalyl chloride in the presence of a
catalytic amount of DMF provided 5-methyl-1,3,4-oxadiazole-
2-carbonyl chloride 11 in quantitative yield.
Finally, coupling reaction of free amine 27 with acid chloride
11 was conducted in the presence of N-methylmorpholine (4-
NMM) to generate a mixture of the desired free phenol 1 and
the ester 31 in 10:1 ratio (Scheme 10). Treatment of the crude
reaction mixture with aqueous methylamine to cleave the ester
linkage in 31, followed by acidification with HCl provided free
phenol 1 in 88% isolated yield. Raltegravir 1 was isolated as a
crystalline potassium salt (1:1 molar ratio) by addition of KOEt
to a solution of the free phenol 1 in a 93% yield and 99.5%
purity (Scheme 10). This process for the manufacture of 1 was
used to produce API through launch.
Second-Generation Process for the Synthesis of 1. The
construction of the hydroxypyrimidinone 7 via the cyclo-
isomerization chemistry used inexpensive and readily available
raw materials, provided the complex core in a single step with
high efficiency, and was robust and simple to perform with a
direct crystallization to afford the product 7 with good purity.
Thus, the majority of the synthesis redesign was targeted at the
(13) Ogilvie, W.; Rank, W. Can. J. Chem. 1987, 65, 166.
Vol. 15, No. 1, 2011 / Organic Process Research & Development
•
77

Scheme 10. Complete synthesis of raltegravir potassium 1
Scheme 11. Preparation of amide 36
screening was utilized. DOE evaluation of a number of variables
including reaction concentration, temperature, time, and equiva-
lents of reagents showed that higher reaction temperature, higher
concentration, and higher equivalents of reagents all lead to
higher N-Me selectivity. Under optimum conditions [4 equiv
of Mg(OMe)2, 4 equiv of MeI, 20 h, 65 °C, 1.0 M] N-methyl
amide 26 was obtained in 90% assay yield in a 99:1, N-Me vs
O-Me ratio.
With successful proof of concept on the in situ conversion
of O-Me 37 to N-Me 26, we used high-throughput reaction
screening to optimize the reaction conditions. Magnesium
hydroxide [Mg(OH)2] was selected as the base for optimization
due to ready availability and very low cost. Trimethylsulfoxo-
nium iodide [Me3S(O)I] was selected as the methylation
reagent15 due to relative cost, safety profile, and ease of handling
when compared with methyl iodide. Both DMSO and DMF
were suitable as reaction solvents; however, NMP was chosen
for intrinsic safety reasons.
The optimized methylation conditions [2 equiv of Mg(OH)2,
2 equiv of Me3S(O)I, NMP, 100 °C, 6 h) provided >99%
conversion and 92% isolated yield of 26 after in situ crystal-
lization, filtration, and drying. Addition of at least 1 equiv of
water was essential for complete conversion of the O-Me to
the N-Me product. Under these reaction conditions MeI is
released at the reaction temperature resulting in an initial 4:1
mixture of 26:37. In situ, iodide-promoted demethylation of 37
followed by remethylation recycles the undesired O-methyl
isomer to 26 in a single-pot reaction. The reaction was generally
complete after 5 h at 100 °C.16
During the final coupling reaction between free amine 27
and oxadiazole 11, unselective acylation of the phenolic OH
occurs to give the bis-acylated compound 31. Hydrolysis of 31
using MeNH2 or KOH furnishes the coupled phenol 1. Attempts
to overcome the unselective acylation issue using alternative
coupling reagents were unsuccessful. Thus, in order to reduce
the need for 2.2 equiv of expensive oxadiazole 11 to obtain
high conversion/yield, protection of the OH group in amine 27
was employed. Selection of the ‘right’ protecting group was
critical. The group needed to be easily and economically
installed in essentially quantitative yield and provide a stable,
subsequent chemical steps. Evaluation of each step of the
existing route with respect to yield, productivity, efficiency,
waste generation (PMI),14 and solvent use was carried out, and
a number of potential areas for improvement were identified.
Considerable work on the methylation reaction was carried
out in the first-generation route development. Apart from the
moderate yield (70%), the reaction suffered from lower than
ideal productivity and very high solvent use, resulting in the
production of a large amount of waste. In fact the PMI for this
single step was over 100. Since, based on our prior work,
complete kinetic regioselectivity was unlikely to be achieved,
our attention shifted to the demethylation of 23 followed by a
N-methylation recycle, which would theoretically permit an
eventual 100% overall yield (Scheme 6). Unfortunately, at-
tempted demethylation of 23 under a variety of conditions
resulted in extensive decomposition, due in part to the lability
of the methyl ester functionality. In order to stabilize the
intermediate to the relatively harsh demethylation conditions,
we decided to look at reversing the step order and to prepare
the amide prior to methylation.
Amide 36 was prepared from hydroxypyrimidinone 7 using
1.2 equiv of 4-FBA in the presence of one equiv of triethylamine
in methanol at reflux for 6 h. Addition of acetic acid and water
to induce crystallization provided amide 36 in 99% isolated yield
after filtration and drying (Scheme 11).
Methylation of amide 36 under the previously optimized
conditions [MeI, Mg(OMe)2, DMSO] used for ester 7 provided
similar N-Me to O-Me regioselectivity (78:22). With the
increased stability of amide 36 vs ester 7, higher temperatures
and longer reaction times were investigated. Heating the mixture
for 4 h at 65 °C gave an 80:20 selectivity for the desired
N-methylated product 26. Remarkably, extended heating (20
h) provided a 99:1 selectivity for 26 (Scheme 12). In order to
explore the O-Me to N-Me recycle strategy further, design of
experiments (DOE) and extensive use of high-throughput
(15) Malik, S.; Nadir, U. K.; Pandey, P. S. Synth. Commun. 2008, 38, 3074.
(16) The scope, generality, and mechanism of the O-Me to N-Me
rearrangement is under investigation and will be reported separately
in due course.
(14) PMI ) product mass intensity [total raw material input (kg)/quantity
of bulk API (kg)]. Auge´, J. Green Chem. 2008, 10, 225.
78
•
Vol. 15, No. 1, 2011 / Organic Process Research & Development

Scheme 12. Methylation of 36 and conversion of O-Me to N-Me in situ using MgI2
Scheme 13. Preparation of free-amine pivalate ester 33
isolable intermediate that could be stored and used directly in
the coupling reaction. Ideally the intermediate would be
crystalline and nonhygroscopic, obviating the tedious and
expensive drying step of amine 27 in the existing process. Last
the protecting group would ideally be cleaved under the same
conditions (or milder) as used in the current process. The
oxadiazole side chain is unstable under basic conditions and
care is needed to avoid decomposition. Consequently, a number
of acyl protecting groups were evaluated including acetate,
propionate, benzoate, and pivalate. A combination of positive
attributes led us to select the pivalate ester derivative 33 for
further development.17
Pivalate ester 33 was prepared in excellent yield (99%) and
purity using a highly optimized through-process (Scheme 13).
Cbz-amide 26 was acylated with pivaloyl chloride in the
presence of triethyl amine and a catalytic amount (0.01 mol
%) of DMAP in ethyl acetate. Without DMAP, the reaction
was sluggish and typically stalled at 90-95% conversion.
Attempted hydrogenation of 32 using the previously employed
conditions (Pd/C, MsOH, MeOH) was complicated due to the
insolubility of the MsOH salt of the amine 33 in the reaction
medium, making catalyst filtration impossible. After screening
several different acids, glycolic acid was identified as a suitable
counterion with excellent solubility properties for the resulting
salt. Glycolic acid proved to be an essential additive allowing
complete conversion to the amine without formation of the des-
fluoro impurity 34 and maintaining solubility of the free amine
in the reaction mixture.18 In practice, after pivalate ester
formation, the crude reaction mixture was washed with water
to remove the NEt3 ·HCl. Methanol, glycolic acid, and Pd/C
were added, and the mixture hydrogenated at 5 psi of hydrogen
for 2-3 h, at 20-25 °C. Upon complete conversion, the catalyst
was filtered, and the pH of solution was adjusted to pH 9 with
triethylamine to crystallize the pivalate ester, 33, as the free
base. The isolated solid was a high-melting, crystalline solid
that proved stable to storage in the solid state and gratifyingly
was nonhygroscopic between 0 and 95% relative humidity and
could be dried easily using conventional drying oven technology.
With crystalline, dry, nonhygroscopic 33 in-hand, reoptimi-
zation of the final amide coupling was carried out (Scheme 14).
In the original manufacturing route a mixture of solvents was
required. THF was used to azeotropically dry the amine 27,
acetonitrile was used to prepare the acyl chloride 11 and IPA/
water used to crystallize the free phenol 1. The volume
productivity was low, while solvent waste generation was high
and the multisolvent waste mixture produced was difficult to
recycle. In order to simplify the process and increase productiv-
ity, the coupling and isolation were optimized to a single solvent.
Thus, acyl chloride preparation was carried out in acetonitrile
at 0-5 °C. A slurry of 33 and NMM in acetonitrile was added
to 11 at -10 °C. On completion of the reaction aqueous
potassium hydroxide was added to convert 35 into crude 1.
Acetic acid was then added followed by water to crystallize
the free phenol 1 in 97% overall isolated yield from 33. When
compared to the original process to couple 11 with 27, the new
process is less complex from a process standpoint, uses
(17) Humphrey, G. R.; Miller, R. A.; Maligres, P. E.; Weissman, S. Process
for preparing N-substituted hydroxypyrimidinone carboxamides. PCT
Int. Appl. WO/2009/088729, 2009.
(18) Des-F 34 results in formation of des-F 1, an impurity not rejected by
crystallization.
Vol. 15, No. 1, 2011 / Organic Process Research & Development
•
79

Scheme 14. Optimized synthesis of free phenol 1
approximately 80% less solvent, and is about 3 times more
productive. The new process also uses only 1.15 equiv of
expensive acid chloride 11 (vs 2.2 equiv), does not employ any
solvent switches or time-consuming concentrations, and pro-
vides API 1 in almost quantitative yield.
volume with MTBE to remove residual ammonia to provide
35b,19 (∼50 wt % in MTBE, 73.3 assay kg) in 99% yield.
(Cyano-dimethyl-methyl)carbamic Acid Benzyl Ester (4).
A mixture of aminonitrile 3 (36.8 assay kg, 437 mol) and
MTBE (307 kg) was cooled to 10 °C and benzylchloroformate
(86.1 kg, 505 mol) added over 1 h at 10-25 °C. Di-
isopropylethylamine (73.1 kg, 566 mol) was added over 1 h at
10-25 °C and the batch aged at 25 °C for 1 h. Water (83.8
kg) was added, the phases were agitated for 30 min, and the
lower aqueous layer was separated. The organic phase was
washed with water (83.9 kg) and concentrated under reduced
pressure and solvent switched to heptane at a volume of about
550 L. The resultant slurry was filtered and washed with heptane
(300 kg). The cake was dried to afford 45b,20 (85.6 kg, > 99.9
wt %, > 99.9 LCAP%) in 90% yield, mp 101.0-101.7 °C.
[1-(N-Hydroxycarbamimidoyl)-1-methyl-ethyl]carbam-
ic Acid Benzyl Ester (5). A mixture of Cbz-aminonitrile 4 (109
kg, 499 mol) and IPA (220 kg) was warmed to 58 °C and 50%
(w/w) aqueous hydroxylamine (35.9 kg, 544 mol) added over
25 min. The solution was aged at 60 °C for 5 h and cooled to
-2 °C, and heptane (198 kg) was added over 30 min. The
resultant slurry was filtered and washed with heptane (300 kg).
The cake was dried at 40 °C under reduced pressure to afford
55 (114.1 kg, >99.9 wt %, 99.8 LCAP%) in 91% yield, mp
160.0-160.5 °C.
Conclusion
The first-generation manufacturing process was rapidly and
successfully scaled up to provide multiton quantities of bulk
drug with consistently high purity. This route satisfied API
requirements from Phase II to initial commercial manufacturing
batches. The yield of the key hydroxypyrimidinone 7 was
increased from 15% in the initial medicinal chemistry route to
44%.
Improvements in the conversion of hydroxypyrimidinone 7
to raltegravir 1 resulted in an increase in overall yield from
only 20% in the original medicinal chemistry synthesis to 51%
in the first-generation manufacturing route and finally to 84%
for the second-generation manufacturing route. The described
process was successfully demonstrated at the multihundred
kilogram scale. In addition to reducing the overall cost to
produce raltegravir, the improvement in overall yield, coupled
with a 3- to 5-fold increase in productivity for most steps,
resulted in an overall reduction of organic and aqueous waste
generation by 65%.
2-(1-Benzyloxycarbonylamino-1-methyl-ethyl)-5,6-dihy-
droxy-pyrimidine-4-carboxylic Acid Methyl Ester (7). A
slurry of amidoxime 5 (63.1 kg, 251 mol) and methanol (180
kg) was cooled to 15 °C. Dimethyl acetylenedicarboxylate (38.8
kg, 273 mol) was added over 30 min, maintaining the batch
temperature between 15 and 25 °C. The resultant solution was
aged at 25 °C for 2 h to obtain >99% conversion and gave
Michael adducts 6Z/6E (ratio ∼65:35 by 1H NMR).
Experimental Section
General. NMR data were obtained using a Bruker AM400
or AM500 spectrometer. HPLC assays were carried out using
a C-18 reversed-phase column eluted with 0.1% H3PO4 (aq)
and acetonitrile. All isolated yields reflect correction for purity
based on HPLC assays. All reagents and solvents were used as
received without further purification.
2-Amino-2-methyl-propionitrile (3). A solution of acetone
cyanohydrin 2 (75.0 kg, 882 mol) and MTBE (10 kg, from a
line flush) was cooled to 10 °C, and anhydrous ammonia (22.6
kg, 1329 mol) was charged over 2 h, maintaining the reaction
temperature between 10-25 °C and the pressure below 30 psi.
Once complete conversion was obtained (>99% as determined
by GC assay), the vessel was vented to a thermal oxidizer and
purged with nitrogen for 15 min to remove residual ammonia.
MTBE (73 kg) was added and the batch flushed at constant
For characterization purposes, a small sample was purified
by flash chromatography (silica gel, EtOAc/hexanes) to give
(19) (a) Taillades, J.; Commeyras, A. Tetrahedron 1974, 30, 3407–3414.
(b) Rousset, A.; Lasperas, M.; Taillades, J.; Commeyras, A. Tetra-
hedron 1980, 36, 2649–2661.
(20) (a) Demko, Z. P.; Sharpless, K. B. Org. Lett. 2002, 4, 2525–2527. (b)
Crescenzi, B.; Gardelli, C.; Muraglia, E.; Nizi, E.; Orvieto, F.; Pace,
P.; Pescatore, G.; Petrocchi, A.; Poma, M.; Rowley, M.; Scarpelli,
R.; Summa, V. Preparation of N-substituted hydroxypyrimidinone
carboxamide inhibitors of HIV integrase. PCT Int. Appl. WO/2003/
035077, 2003.
80
•
Vol. 15, No. 1, 2011 / Organic Process Research & Development

6Z and 6E. 6Z as colorless needles, mp 76.5-77.0 °C. 1H NMR
(CDCl3, 400 MHz) δ: 7.38-7.26 (m, 5 H), 5.75 (s, 1 H), 5.66
(br s, 3 H), 5.08 (s, 2 H), 3.82 (s, 3 H), 3.71 (s, 3 H), 1.54 (s,
6 H). 13C NMR (CDCl3, 100 MHz) δ: 165.5, 163.2, 161.2,
155.6, 154.9, 136.0, 128.6 (2 C), 128.3, 128.1 (2 C), 102.3,
66.9, 54.3, 52.6, 51.5, 25.9 (2 C). HRMS (ESI) calculated for
C18H23N3O7 (M + Na)+ 416.1434, found 416.1438. 6E as
EtOAc) to give O-methylated byproduct 23 as colorless needles,
mp 103.0-103.5 °C. 1H NMR (CDCl3, 400 MHz) δ: 10.45 (s,
1 H), 7.37-7.26 (m, 5 H), 6.56 (br s, 1 H), 5.10 (s, 2 H), 4.06
(s, 3 H), 4.00 (s, 3 H), 1.74 (s, 6 H). 13C NMR (CDCl3, 125
MHz) δ: 169.5, 161.4, 161.2, 155.2, 143.6, 137.3, 131.1, 128.4
(2 C), 128.1 (2 C), 128.0, 66.1, 57.2, 54.7, 53.3, 26.9 (2 C).
HRMS (ESI) calculated for C18H21N3O6 (M + Na)+ 398.1328,
found 398.1321.
1
colorless needles, mp 65.0-66.1 °C. H NMR (CDCl3, 400
MHz) δ: 7.38-7.26 (m, 5 H), 5.83 (s, 1 H), 5.40 (br s, 2 H),
5.08 (s, 1 H), 5.07 (s, 2 H), 3.88 (s, 3 H), 3.69 (s, 3 H), 1.56 (s,
6 H). 13C NMR (CDCl3, 100 MHz) δ: 166.9, 163.4, 161.3,
160.6, 155.6, 135.8, 128.6 (2 C), 128.3 (2 C), 128.1, 94.9, 67.0,
54.5, 52.8, 51.4, 26.0 (2 C). HRMS (ESI) calculated for
C18H23N3O7 (M + Na)+ 416.1434, found 416.1430.
{1-[4-(4-Fluoro-benzylcarbamoyl)-5-hydroxy-1-methyl-6-
oxo-1,6-dihydro-pyrimidin-2-yl]-1-methyl-ethyl}carbamic Acid
Benzyl Ester (26). To a slurry of N-methylpyrimidinone 22
(1.40 kg, 3.73 mol) in EtOH (14 L) at 4 °C was slowly added
4-fluorobenzylamine (1.05 kg, 8.14 mol) over 15 min. The
slurry becomes very thick, and vigorous stirring is required.
The reaction mixture was warmed to 72 °C over 2 h and kept
at this temperature for 2 h (>99.5% conversion).
The solution was concentrated to about 160 L under reduced
pressure and solvent switched, by feed and bleed at constant
volume, to xylenes. The final batch temperature was maintained
below 70 °C during the solvent switch. The mixture was heated
to 125 °C, aged for 2 h, warmed to 135 °C for 5 h to complete
the reaction and then cooled to 60 °C. Methanol (45 kg) was
added and the resultant slurry aged at 35 °C for 1 h. MTBE
(145 kg) was added over 1 h at 20-25 °C. The slurry was
cooled to -10 °C over 2 h and aged overnight. The slurry was
filtered, the cake washed with 9:1 of MTBE:methanol (140 kg)
and dried at 40 °C under reduced pressure to afford 7 (49.6 kg,
> 99.5 wt %, 98.4 LCAP%) in 54% overall yield, mp
186.3-187.0 °C. 1H NMR (CDCl3, 400 MHz) δ: 12.23 (br s,
1 H), 10.75 (br s, 1 H), 7.24 (m, 5 H), 6.10 (br s, 1 H), 4.98 (s,
2 H), 3.98 (s, 3 H) 1.69 (s, 6 H). 13C NMR (CDCl3, 100 MHz)
δ: 169.9, 159.4, 155.1, 153.5, 149.4, 136.2, 128.3 (2 C), 128.0,
127.9 (2 C), 126.1, 66.7, 55.6, 53.2, 26.4 (2 C). HRMS (ESI)
calculated for C17H19N3O6 (M + H)+ 362.1352, found 362.1352.
2-(1-Benzyloxycarbonylamino-1-methyl-ethyl)-5-hydroxy-
1-methyl-6-oxo-1,6-dihydro-pyrimidine-4-carboxylic Acid
Methyl Ester (22). To a mixture of hydroxypyrimidinone 7
(2.00 kg, 5.53 mol) and DMSO (16 L) was added a solution of
8 wt % of Mg(OMe)2 in MeOH (13.4 L, 11.1 mol, 2.0 equiv).
Excess MeOH was then evaporated under vacuum (30 mmHg)
at 40 °C. After cooling to 20 °C, MeI (1.38 L, 22.2 mol) was
added dropwise, and the mixture was stirred at 20-25 °C for
2 h, then at 60 °C for 5 h and cooled to 20 °C.
2 M HCl (20 L) was added followed by 5 wt % sodium
bisulfite (2 L). Water (40 L) was added over 40 min, and the
slurry was stirred for 40 min and the resultant slurry cooled to
5 °C. The crystalline product was collected by filtration, washed
with water (20 L) and then with 9:1 of MTBE/MeOH (30 L)
to remove the O-Me byproduct. The product was dried
overnight at rt under vacuum with N2 sweep to give 22 (1.49
kg) in 70% yield, as an off-white crystalline solid, mp
198.5-199.0 °C. 1H NMR (CDCl3, 400 MHz) δ: 10.35 (br s,
1 H), 7.38-7.26 (m, 5 H), 5.41 (br s, 1 H), 5.01 (s, 2 H), 3.96
(s, 3 H), 3.64 (s, 3 H), 1.70 (s, 6 H). 13C NMR (CDCl3, 100
MHz) δ: 169.7, 159.5, 154.3, 151.0, 147.9, 136.0, 128.5 (2 C),
128.2 (2 C), 128.1, 123.5, 66.9, 60.3, 53.0, 32.9, 27.8 (2 C).
HRMS (ESI) calculated for C18H21N3O6 (M + H)+ 376.1509,
found 376.1507.
The reaction mixture was cooled to 60 °C, and AcOH (0.55
L) was added over 30 min. Water (6.7 L) was added over 30
min to afford a slurry. After 30 min at 60 °C, the remaining
7.3 L water was added over 30 min, and the reaction mixture
was allowed to cool to ambient temperature. The crystalline
solid was collected by filtration, slurry washed with 1:1 of water/
EtOH (2 × 4 L), and dried on the filter pot under vacuum with
N2 sweep to constant weight to afford Cbz-amide 26 (1.60 kg,
99 LCAP%) in 90% yield, as white crystalline solid, mp
182.0-182.8 °C. (NMR spectra were acquired at 0 °C: the
1
major carbamate rotamer is reported.) H NMR (600 MHz,
CDCl3) δ 11.95 (br s, 1 H), 7.83 (t, J ) 6.0 Hz, 1H), 7.36-7.28
(m, 7 H), 7.07-7.03 (m, 2 H), 5.42 (s, 1 H), 5.00 (s, 2 H), 4.57
(d, J ) 6.0 Hz, 2 H), 3.65 (s, 3 H), 1.67 (s, 6 H); 13C NMR
(150 MHz, CDCl3) δ 168.5, 162.4 (d, JCF ) 246 Hz), 159.8,
154.5, 151.1, 146.7, 135.9, 133.1 (d, JCF ) 3 Hz), 129.6 (d,
JCF ) 8 Hz), 128.8, 128.7, 128.4, 124.4, 115.9 (d, JCF ) 21
Hz), 67.2, 57.3, 42.5, 33.1, 28.1. HRMS (ESI) calculated for
C24H25FN4O5 (M + H)+ 469.1887, found 469.1882.
2-(1-Amino-1-methyl-ethyl)-5-hydroxy-1-methyl-6-oxo-
1,6-dihydro-pyrimidine-4-carboxylic Acid 4-Fluoro-benzyl-
amide (27). A solution of amide 26 (1.60 kg, 3.41 mol),
methanesulfonic acid (0.20 kg, 3.41 mol, 1.0 equiv), and 5%
Pd/C (0.16 kg, 10 wt %) in methanol (12.6 L) at 12 °C was
hydrogenated at 15 psi for 2-3 h (>99.5% conversion). The
reaction mixture was filtered through Solka Floc (0.72 kg), then
washed with 4.5:1 of methanol/water (24 L). The combined
filtrate was concentrated to a total volume 15.4 L at 10-23 °C
and neutralized by 2.5 NaOH to pH ) 8 at 5-23 °C. The
resulting slurry was aged at 0-5 °C for 1 h. The crystalline
solid was collected by filtration, rinsed with cold water (5.0
L), and dried under vacuum with nitrogen sweep to afford
monohydrated free amine 27 (1.19 kg, 99.7 LCAP% purity,
100.2 wt % by HClO4 titration) in 99% yield, mp 188.0-189.3
°C. 1H NMR (400 MHz, DMSO-d6) δ: 10.86 (br s, 1 H), 7.35
(dd, J ) 8.6, 5.6 Hz, 2 H), 7.15 (bt, J ) 8.6 Hz, 2 H), 4.48 (d,
J ) 6.0 Hz, 2 H), 3.58 (s, 3 H), 1.63 (s, 6 H); 13C NMR (100
MHz, DMSO-d6) δ: 169.3, 164.2, 162.2 (d, JCF ) 243 Hz),
154.9, 153.2, 137.0 (d, JCF ) 2 Hz), 130.2 (d, JCF ) 8 Hz),
123.3, 116.1 (d, JCF ) 22), 57.4, 42.4, 33.9, 27.6 (2 C). HRMS
(ESI) calculated for C16H19FN4O3 (M + H)+ 335.1519, found
335.1517.
For characterization, a small sample of the mother liquors
was purified by flash chromatography (silica gel, hexane/
Vol. 15, No. 1, 2011 / Organic Process Research & Development
•
81

Potassium 5-Methyl-[1,3,4]oxadiazole-2-carboxylate (30).
Ethyl oxalyl chloride (4.01 kg, 29.4 mol) was slowly added to a
mixture of methyl tetrazole 28 (2.50 kg, 29.7 mol; [while 5-me-
thyl-tetrazole is not rated an explosive material (see: http://www.
chemicalbook.com/ChemicalProductProperty_EN_CB4290179.
htm), it is an irritant and highly flammable. It should be
handled only after all hazard data has been reviewed and
proper PPE obtained]) and triethylamine (3.03 kg, 30.0 mol)
in toluene (32 L) at 0 °C at such a rate that the temperature
stayed below 5 °C at all times. (Caution: 29 is a potentially
explosive intermediate that will quickly generate nitrogen
gas at <50 °C. Conduct this chemistry only after training
for handling explosive compounds, and conduct the heating
only behind a shield.) The slurry was stirred for 1 h at 0-5
°C. The triethylamine hydrogen chloride salt was filtered off.
The solid was washed with 27 L of cold toluene (5 °C). The
combined filtrates were kept at 0 °C at all times and were
slowly added to a hot solution of toluene (50 °C, 15 L) over
40-50 min. (Caution: N2 gas evolves. Use the same
precautions during heating as above.) Most of the intermedi-
ate 29 had been converted to the corresponding non-explosive
oxadiazole ester at this moment. The solution was continually
stirred at 70 °C for 1 h. After cooling to 20 °C, the toluene
solution was washed with 5 L of 10% brine, and solvent
was switched to ethanol. EtOH was added to adjust the final
volume to approximately 40 L.
150 mL. The resulting slurry was then cooled to 0-5 °C and
stirred for 1 h. The crystalline solid was collected by filtration
and washed with MeOH (70 mL) and water (3 × 100 mL)
The product was dried overnight under vacuum with nitrogen
sweep at room temperature. Raltegravir 1 (free phenol) was
obtained as a white crystalline solid (24.6 g, 99 wt %) in 88%
1
yield, mp 143.0-144.1 °C. H NMR (400 MHz, DMSO-d6)
δ: 12.20 (br s, 1 H), 9.86 (br s, 1 H), 9.10 (t, J ) 6.4 Hz, 1 H),
7.43 (dd, J ) 8.6, 5.7 Hz, 2 H), 7.19 (app. t, J ) 8.6 Hz, 2 H),
4.54 (d, J ) 6.4 Hz, 2 H), 3.52 (s, 3 H), 2.59 (s, 3 H), 1.77 (s,
6 H); 13C NMR (100 MHz, DMSO-d6) δ: 169.5, 166.9, 162.4
(d, JCF ) 243 Hz), 159.8, 159.2, 153.7, 152.9, 146.8, 135.9 (d,
JCF ) 3 Hz), 130.6 (d, JCF ) 8 Hz), 125.6, 116.2 (d, JCF ) 22),
58.7, 42.8, 34.1, 28.1 (2 C), 11.8. HRMS (ESI) calculated for
C20H21FN6O5 (M + H)+ 445.1636, found 445.1633.
{1-[4-(4-Fluoro-benzylcarbamoyl)-5,6-dihydroxy-pyrimi-
din-2-yl]-1-methyl-ethyl}-carbamic Acid Phenethyl Ester
(36). A slurry of pyrimidine diol 7 (50.0 g, 136 mmol) in
methanol (81 mL) was warmed to 55 °C and triethylamine (16.4
g, 163 mmol) added in one portion. The solution was warmed
to 65 °C and 4-fluorobenzylamine (20.4 g, 163 mmol) added
over 30-40 min at 65-68 °C. The mixture was aged at gentle
reflux for 7 h. The solution was cooled to 55 °C and acetic
acid (16.3 g, 271 mmol) added over 5 min. Water (17 mL)
was added and the reaction solution seeded with product (50
mg) at 60 °C. The resultant slurry was aged at 60 °C for 30
min and the remaining water (65 mL) added over 30 min at 60
°C and then cooled to 20 °C over 2 h. The slurry was filtered,
and the cake was washed with 1:1 of methanol/water (60 mL)
and dried in a nitrogen stream overnight to afford amide 36
(62.3 g, 96 wt %, KF ) 1.75% w/w) in 98% yield, mp
110.3-111.5 °C. 1H NMR (600 MHz, CDCl3) δ: 12.37 (br s,
2H), 7.95 (br t, J ) 6.0 Hz, 1 H), 7.35-7.31 (m, 2 H),
7.25-7.19 (br m, 5 H), 7.08-7.04 (m, 2 H), 6.49 (s, 1 H),
4.96 (s, 2 H), 4.58 (d, J ) 6.0 Hz, 2 H), 1.64 (s, 6 H); 13C
NMR (150 MHz, CDCl3) δ: 168.5, 162.6 (d, JCF ) 247 Hz),
160.0, 155.3, 154.1, 147.9, 136.5, 133.1 (d, JCF ) 3 Hz), 129.7
(d, JCF ) 8 Hz), 128.6, 128.1, 127.9, 127.1, 116.0 (d, JCF ) 22
Hz), 66.8, 55.5, 42.7, 26.7. HRMS (ESI) calculated for
C23H23FN4O5 (M + H)+ 455.1731, found 455.1737.
{1-[4-(4-Fluoro-benzylcarbamoyl)-5-hydroxy-1-methyl-6-
oxo-1,6-dihydro-pyrimidin-2-yl]-1-methyl-ethyl}-carbamic
acid phenethyl ester (26). A mixture of amide 36 (62.4 g, 95.4
wt %, 131 mmol), magnesium hydroxide (15.8 g, 271 mmol),
trimethylsulfoxonium iodide (59.4 g, 271 mmol) and water (1.46
mL) was warmed to 100 °C over 2 h and maintained at 100 °C
for 5 h. The mixture was cooled to 20 °C and methanol (108
mL) was added. 5 N HCl (54 mL) was added over 15 min
followed by seed crystals of product (50 mg). The mixture was
aged for 15 min and then a solution of 2.4 M aq sodium bisulfite
(3 mL) was added over 1-2 min. The mixture was aged for
1-2 h at 31-35 °C. Five N HCl (54 mL) was added over 1 h.
The slurry was gradually cooled to 10 °C over 1 h. The slurry
was filtered and the filter cake washed with 1:1 of methanol/
water (200 mL). The white granular, crystalline product was
dried to a constant weight under a stream of nitrogen to give
Cbz-amide 26 (42 g, 99 wt % purity) in 89% yield, mp
181.8-182.5 °C.
The solution was cooled to 10 °C and aqueous KOH (8.0
L) was added over 30 min. The resulting thick slurry was then
stirred for 40 min at room temperature while the oxadiazole K
salt crystallized. The solid was collected by filtration and washed
with EtOH (11 L) and MTBE (15 L). The crystalline solid was
dried overnight under vacuum at 20 °C with nitrogen sweep to
give oxadiazole K salt 30 (4.48 kg) in 91% yield, mp 258.3 °C
1
dec. H NMR (400 MHz, CD3OD/D2O ) 2:1) δ: 2.62 (S, 3
H); 13C NMR (100 MHz, CD3OD/D2O ) 2:1) δ: 165.8, 161.8,
158.3, 9.8. MS (FAB) m/z 129 (corresponding to free acid, M
+ H, 100%).
Raltegravir (1) from Amine (27). A slurry of oxadiazole
K salt 30 (22.8 g, 96.1 wt %, 132 mmol) in acetonitrile (200
mL) and DMF (0.3 mL) was cooled to 0-5 °C and oxalyl
chloride (16.7 g, 132 mmol) added keeping the internal
temperature <5 °C. The slurry was aged for 1 h at 0-5 °C.
A slurry of free amine monohydrate 27 (22.2 g, 63 mmol)
and THF (260 mL). was azeotropically dried at 45-55 °C to a
KF of <150 ppm. The resulting slurry (about 280 mL) was
cooled to 0-5 °C and NMM (33.4 g, 330 mmol) added. The
acetonitrile slurry of 11 was added slowly maintaining the
reaction temperature below 5 °C. The slurry was aged for 1 h.
The reaction mixture was quenched with 40% MeNH2 aqueous
(31 g, 396 mmol) and the yellow slurry stirred for 30 min at 5
°C. The mixture was acidified to pH 4.3 with 2 N HCl, followed
by addition of water (210 mL). EtOAc (200 mL) was added,
and the mixture was heated to 30 °C until all solids dissolved.
The organic layer was separated. The aqueous layer was
extracted with EtOAc (150 mL) at 30 °C. The combined organic
layers were washed with brine (200 mL) and water (100 mL).
The organics were concentrated to a volume of (100 mL) under
vacuum and solvent-switched to MeOH to a final volume of
82
•
Vol. 15, No. 1, 2011 / Organic Process Research & Development

2,2-Dimethyl-propionic Acid 2-(1-Amino-1-methyl-ethyl)-
4-(4-fluorobenzylcarbamoyl)-1-methyl-6-oxo-1,6-dihydro-
pyrimidin-5-yl Ester (33). To a slurry of Cbz-amide 26 (50.0
g, 98 wt %, 105 mmol) in ethyl acetate (150 mL) was added
triethylamine (13.7 g, 136 mmol) and DMAP (12.7 mg, 0.15
mmol). The slurry was cooled to 10 °C and pivaloyl chloride
(15.1 g, 126 mmol) added over about 30 min maintaining the
reaction temperature 10-15 °C. On complete addition, the
slurry was aged 15 min at 10-15 °C. Water (50 mL) was added
and the mixture warmed to rt. The lower aqueous layer was
cut. Methanol (100 mL) and 67 wt % glycolic acid (16.6 g,
125.5 mmol) were added to the organic layer and the crude
mixture hydrogenated at 5 psi and 20 °C using 5 wt % Pd/C
(1.15 g, 50% wet). On complete reaction, the catalyst was
removed by filtration through a Celite plug and the cake washed
with methanol (50 mL). The filtrates were combined, and
triethylamine (15.8 g, 136 mmol) was added in one portion at
rt. The mixture was seeded with amine product 33 (50 mg)
and aged for 1 h. Water (150 mL) was added over 1 h at rt.
The slurry was aged for 2 h and filtered. The cake was washed
with 1:1 of methanol/water (100 mL) and dried to afford 33
(42.9 g, 99 wt %) in 98% yield, as a white crystalline solid,
mp 210-211 °C. 1H NMR (600 MHz, CDCl3) δ 7.88 (br t, J
) 6.0 Hz, 1 H), 7.29-7.25 (m, 2 H), 7.04-7.00 (m, 2 H),
4.53 (d, J ) 6.0 Hz, 2 H), 4.01 (s, 3 H), 1.60 (br s, 2 H), 1.59
(s, 6 H), 1.43 (s, 9 H); 13C NMR (150 MHz, CDCl3) δ 175.9,
162.4 (d, JCF ) 246 Hz), 161.9, 160.5, 160.4, 138.8, 136.6,
133.9 (d, JCF ) 3.1 Hz), 129.4 (d, JCF ) 7.9 Hz), 115.8 (d, JCF
) 21.4 Hz), 56.5, 42.6, 39.4, 34.4, 31.3, 27.3. HRMS (ESI)
calculated for C21H27FN4O4 (M + H)+ 419.2095, found
419.2099.
the above oxadiazole acid chloride 11 at -10 °C over 45 min.
20% aq potassium hydroxide (389 g) was added and the reaction
mixture aged at 0-5 °C. HOAc (137 mL) was added over 5
min and the reaction mixture warmed to 15 °C. Water (800
mL) was added slowly. Seed 1 (2 wt %) was added to prevent
supersaturation. The slurry was aged for 1 h and filtered. The
cake was washed with 2.5:1 water/acetonitrile (300 mL), water
(300 mL) and dried to afford raltegravir 1 (104 g, >99.5 wt %)
in 97% yield, mp 143.5-144.5 °C.
Raltegravir Potassium 1. To a suspension of phenol 1 (200
g, 4.50 mol) in ethanol (270 mL) and water (270 mL) was added
aq KOH (45 wt %, 35.7 mL, 423 mmol) at 20 °C. After 30
min the cloudy solution was filtered through a 1µm filter,
seeded, at rt with potassium salt 1 (9.0 g). The slurry was aged
for 1 h and EtOH (2.9 L) added at 20 °C over 1 h. The slurry
was cooled to <5 °C and aged for 2 h. The batch was filtered,
washed with EtOH (500 mL), and dried to afford raltegravir
potassium 1 (194 g, > 99 wt %) as a white crystalline solid in
94% yield, mp 274.2-275.2 °C. 1H NMR (500 MHz, DMSO-
d6) δ: 11.65 (t, J ) 6.0 Hz, 1 H), 9.75 (s, 1 H), 7.36 (dd, J )
8.6, 5.7 Hz, 2 H), 7.14 (app. t, J ) 8.6 Hz, 2 H), 4.48 (d, J )
6.0 Hz, 2 H), 3.43 (s, 3 H), 2.58 (s, 3 H), 1.73 (s, 6 H); 13C
NMR (125 MHz, DMSO-d6) δ: 168.7, 167.0, 166.6, 162.1 (d,
JCF ) 243 Hz), 159.7, 158.3, 153.1, 139.6, 138.0 (d, JCF ) 3
Hz), 130.2 (d, JCF ) 8 Hz), 123.7, 116.0 (d, JCF ) 22), 58.4,
42.1, 33.3, 28.1 (2 C), 11.7.
Acknowledgment
We thank the following individuals from Merck Research
Laboratories for their support during this work: Richard G. Ball,
Nancy Tsou, Jennifer R. Foley, Spencer Dreher, David L.
Hughes, Tony Hudgens, Jaemoon Lee, Amar J. Mahajan,
Dermot O’Brien, Michael Palucki, Paul Phillips, Vanessa M.
Pruzinsky, Mary Stanik, Dietrich Steinhubel, and Timothy J.
Wright.
Raltegravir (1) from Amine (33). A mixture of oxadiazole
K-salt 30 (45.7 g, 0.27 mol), acetonitrile (250 mL), and DMF
(0.55 mL) was cooled to -5 °C and oxalyl chloride (33.4 g,
0.27 mol) added over 30 min. The slurry was aged at 0-5 °C
for 1 h, then cooled to -10 °C.
A solution of amine 33 (99.9 g, 0.24 mol) and acetonitrile
(150 mL) was cooled to -10 °C and N-methylmorpholine (28.9
g, 0.28 mol) added. The free amine/NMM slurry was added to
Received for review September 22, 2010.
OP100257R
Vol. 15, No. 1, 2011 / Organic Process Research & Development
•
83