Six-Step Gram-Scale Synthesis of the Human Immunodeficiency Virus Integrase Inhibitor Dolutegravir Sodium

Article abstract of DOI:10.1021/acs.oprd.1c00139

A short and practical synthesis for preparing the active pharmaceutical ingredient dolutegravir sodium was developed. The convergent strategy starts from (R)-3-amino-1-butanol and establishes the BC ring system in a 76% isolated yield over four steps. Ring A was constructed by a one-pot 1,4-addition to diethyl-(2E/Z)-2-(ethoxymethylidene)-3-oxobutandioate and subsequent MgBr2·OEt2-mediated regioselective cyclization. Amide formation with 2,4-difluorobenzylamine was either performed from the free carboxylic acid or through aminolysis of the corresponding ethyl ester. Final salt formation afforded dolutegravir sodium in a 48-51% isolated yield (HPLC purity of 99.7-99.9%) over six linear steps.

Full text of DOI:10.1021/acs.oprd.1c00139

pubs.acs.org/OPRD
Article
Six-Step Gram-Scale Synthesis of the Human Immunodeficiency
Virus Integrase Inhibitor Dolutegravir Sodium
Jule-Philipp Dietz, Tobias Lucas, Jonathan Groß, Sebastian Seitel, Jan Brauer, Dorota Ferenc,
B. Frank Gupton, and Till Opatz*
Cite This: https://doi.org/10.1021/acs.oprd.1c00139
Read Online
sı
*
Metrics & More
Article Recommendations
Supporting Information
ACCESS
ABSTRACT: A short and practical synthesis for preparing the active pharmaceutical ingredient dolutegravir sodium was developed.
The convergent strategy starts from (R)-3-amino-1-butanol and establishes the BC ring system in a 76% isolated yield over four
steps. Ring A was constructed by a one-pot 1,4-addition to diethyl-(2E/Z)-2-(ethoxymethylidene)-3-oxobutandioate and subsequent
MgBr2·OEt2-mediated regioselective cyclization. Amide formation with 2,4-difluorobenzylamine was either performed from the free
carboxylic acid or through aminolysis of the corresponding ethyl ester. Final salt formation afforded dolutegravir sodium in a 48−
51% isolated yield (HPLC purity of 99.7−99.9%) over six linear steps.
KEYWORDS: dolutegravir sodium, active pharmaceutical ingredient, antivirals, integrase inhibitors, carbamoyl pyridones
attributed to the late-stage introduction of the expensive
amino alcohol 13.
INTRODUCTION
■
Infection with the human immunodeficiency virus (HIV) has
become controllable in recent years due to enormous progress
in development of highly active drugs, which are given for
antiretroviral therapy (ART).1 ART requires the adminis-
tration of at least three different antiviral drugs to suppress the
development of resistances.2 Differentiated by the target
enzyme, there are several classes of HIV-inhibiting drugs.
The class of integrase strand transfer inhibitors (INSTIs)
interferes with the HIV integrase enzyme and prevents it from
inserting viral DNA into the human genome. INSTIs have
been introduced in 2007 with the launch of raltegravir (1)
followed by elvitegravir (2, 2012), dolutegravir (3, 2013),
bictegravir (4, 2018), and cabotegravir (5, 2021)3 (Scheme
1).4
The last three compounds exhibit a high similarity in their
molecular structures, assigning them to the group of carbamoyl
pyridine INSTIs. Dolutegravir (3), usually administered orally
as its sodium salt, was recently recommended by the World
Health Organization for first-line treatment of HIV initiating
ART.5 As a consequence, the demand of this important
medication could further rise.
All synthetic approaches to 3 that have been published
before 2019 have been carefully reviewed.4,6 Since then, four
more routes were disclosed.7−10 Most of the published
syntheses follow a similar strategy, which is represented here
by the hitherto most efficient approach from Micro Labs
(2016)11 (Scheme 2). The highly functionalized pyridone 7
(ring A) is constructed first, which then undergoes cyclization
with (R)-3-amino-1-butanol (13) to construct rings B and C.
Deprotection of the usually protected enol and treatment with
sodium hydroxide furnish dolutegravir sodium (15). The
seven-step synthesis by Micro Labs afforded 15 in a 35%
overall yield. This retained synthesis concept could be
Since the discovery of dolutegravir and the emerging
demand of (R)-3-amino-1-butanol (13), more efficient
syntheses of this crucial building block have been developed
with the consequence of a decreasing market price.12 Thus, an
earlier introduction of 13 could add additional value by
bringing more diversification to the synthetic portfolio of
dolutegravir. Extending the scope of industrially applicable
synthetic routes should encourage generic manufacturing to
ensure global supply of this important drug.
Here, a new synthetic route should be investigated by taking
13 as a starting material (Scheme 3). By using commodity
chemicals, the ring system BC (16) should be constructed first.
Next, 16 should be reacted with readily accessible diethyl-(2E/
Z)-2-(ethoxymethylidene)-3-oxobutandioate (17) to install
ring A. Amide coupling with 2,4-difluorobenzylamine and
salt formation follow in the last step. All in all, an industrially
feasible synthesis route was intended to be developed.
RESULTS AND DISCUSSION
■
Synthesis of the Ring System BC. The synthesis of
amine 16 has already been described twice in the patent
literature. Lerner et al. reported a five-step synthesis starting
from ethyl bromoactetate (19) furnishing amine 16 in a 24%
overall yield.13 The crucial cyclization with amino alcohol 13
gave only a 55% yield under microwave conditions, and
Received: April 21, 2021
https://doi.org/10.1021/acs.oprd.1c00139
Org. Process Res. Dev. XXXX, XXX, XXX−XXX
© XXXX American Chemical Society
A

Organic Process Research & Development
pubs.acs.org/OPRD
Article
Scheme 1. Structures of Raltegravir (1), Elvitegravir (2), Dolutegravir (3), Bictegravir (4), and Cabotegravir (5)
Scheme 2. Synthesis of Dolutegravir by Micro Labs (2016)
Scheme 3. Retrosynthetic Strategy toward Dolutegravir (15)
chromatographic steps were required. During our synthetic
work, a patent from the Virginia Commonwealth University
was disclosed, which describes a similar but more efficient
approach toward 16.10 The methyl ester of intermediate 24
was prepared in three steps from methyl glycinate hydro-
chloride (23) in a 66% yield. The cyclization step was
performed in a toluene/methanol/acetic acid mixture and
achieved a 71% yield after column chromatography. Removing
the Cbz group at higher hydrogen pressure gave slightly better
results, and amine 16 was finally obtained in 44% over five
steps. The synthesis was even performed on a multigram scale
but required two chromatographic steps (Scheme 4).
A potentially shorter approach could be achieved by first N-
acylating 13 with chloroacetyl chloride (25) (Scheme 5).
Subsequent alkylation with aminoacetaldehyde dimethyl acetal
(8) would furnish the acyclic precursor 27. Acid-catalyzed
intramolecular transacetalization would in turn afford the
desired amine 16.
When the acylation of 13 was performed under standard
conditions (NEt3, DCM, 0 °C), a mixture of desired 26 and
N,O-bis-acylated compound 28 was obtained (Scheme 6).
B
https://doi.org/10.1021/acs.oprd.1c00139
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
Scheme 4. Known Syntheses of Amine 16
Scheme 5. Retrosynthetic Proposal toward Amine 16
Scheme 7. N-Acylation of 13 with Chloroacetyl Chloride
In the next step, 26 should be used for alkylating
aminoacetaldehyde dimethyl acetal (8) to form the secondary
amine 27. The reaction was performed either by heating in
toluene or by stirring at r.t. in DCM in the presence of sodium
iodide (Scheme 8). After aqueous workup, only 33−45% of
crude 27 could be isolated. As already reported for the
previous reaction, a significant quantity of 27 remained in the
aqueous phase, which could not be separated from excess
amine 8. Through LC−MS, one main impurity could be
identified as the tertiary amine 29. To circumvent its formation
and to solve the water solubility issue, another strategy was
chosen.
By converting primary amine 8 into a secondary and more
lipophilic amine first, subsequent alkylation would lead to a
tertiary and better extractable amine. Benzylation of 8 proved
to be a good option as it can be easily reverted by
hydrogenation. N-Benzyl-2,2-dimethoxyethylamine (30) was
prepared in two different ways (Scheme 9). When heating
chloroacetaldehyde-dimethylacetal (31) with an excess of
benzylamine (32) in toluene, 77% of 30 was isolated after
distillation. According to a procedure from Luu et al., 30 could
also be prepared through reductive amination from stochio-
metric amounts of benzaldehyde (33) and aminoacetaldehyde
dimethyl acetal (8).14 The latter method furnished 30 in a 97%
isolated yield.
Nevertheless, 28 could be easily saponified to 26 by just
adding an aqueous base to the crude reaction mixture. After
extractive workup, only a 40% isolated yield of 26 was
obtained. It turned out that large amounts of 26 had remained
in the aqueous phase. When water was removed in vacuo and
the salty residue was suspended in ethyl acetate, a further 40%
of 26 could be isolated.
Due to the high polarity of 26, the whole process was
adjusted by using continuous extraction. After complete
saponification of 28, which was followed by GC−MS, the
organic phase was separated and additionally extracted with
water. The combined aqueous phases were neutralized with an
acid, transferred into a Kutscher-Steudel apparatus, and
continuously extracted for 72 h with ethyl acetate. This
procedure enabled the isolation of 26 in a 91% isolated yield
(Scheme 7). It should be noted that 26 showed high purity
1
according to H NMR and no further purification step was
required. As an ambient-pressure distillation is involved in this
continuous extraction, apparatus improvements can likely
accelerate the procedure.
Scheme 6. N-Acylation of 3-(R)-Amino-1-butanol (13)
C
https://doi.org/10.1021/acs.oprd.1c00139
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
Scheme 8. N-Alkylation of Aminoacetaldehyde Dimethyl Acetal (8) with 26
Scheme 9. Synthesis of N-Benzyl-2,2-dimethoxyethylamine (30) by N-Alkylation or Reductive Amination
Using secondary amine 30 instead of primary amine 8 for
the alkylation gave much better results (Scheme 10). The
LC−MS. As the undesired diastereomer with opposite
configuration at the acetalic center was not isolated and
characterized, diastereoselectivity can only be assumed based
on the second largest HPLC peak to amount at least to 44:1
(for details, see the Supporting Information). When the
reaction was performed on an 11.5 g scale, crude 35 was
isolated in a 92% yield as a brown oil, which only contained
slight impurities. For the workup, the reaction mixture was
neutralized with sodium hydroxide and extracted with ethyl
acetate. The stereoconfiguration of 35 was verified by VCD
spectroscopy and comparison of the experimental data with
calculations using density functional theory (for more details,
see the Supporting Information). The enantiomeric excess was
determined by chiral HPLC. Racemic 35 was prepared as a
reference material in the same way starting from rac-3-amino-
1-butanol (rac-35).
Scheme 10. Synthesis of Tertiary Amine 34 from 26 and 30
by Heating in MeCN in the Presence of KI and K2CO3
reaction was performed in acetonitrile (MeCN) in the
presence of potassium iodide (KI) and potassium carbonate
(K2CO3). Complete conversion of 26 was observed after 24 h
at r.t. or after 3 h when heating to reflux. For workup, MeCN
was removed and an extraction from water/ethyl acetate
furnished 34 as a slightly brownish oil. Again, the reaction
proceeded very cleanly, and no further purification of the
product was necessary.
The benzyl group was subsequently removed by hydro-
genation (Scheme 11), and crude amine 16 was obtained in a
Scheme 11. Debenzylation of 35 by Hydrogenation
34 cyclized cleanly in aqueous hydrochloric acid to the
desired diastereomer of oxazinone 35. At least 6 N HCl was
necessary to achieve complete conversion after 48 h at r.t.
(Table 1). Only small amounts of byproducts were detected by
Table 1. Intramolecular Transacetalization of 34
96% yield (78% IY over four steps) as a brown-orange oil,
which solidified after a while to a beige and free-flowing solid.
1
According to LC−MS and H NMR, minor impurities could
be detected in the crude material (see the corresponding
spectra in the Supporting Information). Crude 16 was used for
the next step, and no further purification efforts were
investigated.
Construction of Ring A. Jee et al. also reported a synthesis
route to dolutegravir starting from the ring system BC (16)
(Scheme 12).10 Intermediate 39, which contains the crucial
tricarbonyl moiety, was installed in two steps in a 89% yield.
Ring A was constructed by regioselective cyclization using
magnesium bromide ethyl etherate (MgBr2·OEt2) and triethyl-
amine (65% yield of 40). Subsequent amide coupling afforded
the desired dolutegravir 3 in a 67% yield. The route depicts the
a
entry
solvent
conversion [%]
IY [%]
1
2
3
1 N HCl
3 N HCl
6 N HCl
4
34
100
-
-
92
b
a
b
Determined by LC−MS and UV absorption at 254 nm. 11.5 g
scale.
D
https://doi.org/10.1021/acs.oprd.1c00139
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
Scheme 12. Synthesis of Dolutegravir (3) by Jee et al.
first one reported where the ring system AB was constructed
first. The overall yield was only 17%, and five chromatographic
steps were required.
Enaminone 39 should be also accessible in a single step by
1,4-addition from the corresponding enol ether 17. According
to a procedure of Jones,15 17 was prepared from the
commercially available diethyl oxalacetate sodium salt (in
two steps) (Scheme 13). The salt 41 had to be acidified first to
Amine 16 readily underwent 1,4-addition to 43 in several
solvents (DCM, MeCN, EtOH, THF, and toluene). The
formation of only one geometrical isomer of 43 could be
detected by NMR and LC−MS (for more details, see the
Supporting Information). Regarding the regioselective cycliza-
tion, it turned out that using strong bases like KOtBu, NaOEt,
or NaH predominantly led to the formation of a product
mixture. Applying the conditions of Jee et al., using a
combination of magnesium bromide ethyl etherate (MgBr2·
OEt2) and triethylamine (NEt3) looked more promising, and
predominant conversion to ester 18 (81 area % (254 nm), 35
area % (315 nm)) could be observed (entry 1, Table 2).10
Nevertheless, there was still significant byproduct formation
according to the HPLC trace at 315 nm and 1H NMR.
According to LC−MS data (m/z = 351), it is assumed that the
main byproduct could be pyrrole 45 (Figure 1), which could
form in an undesired 5-exo cyclization with the keto group
adjacent to the ester group.
Scheme 13. Synthesis of Enol Ether 17 by Condensation of
42 with TEOF and Ac2O
obtain diethyl oxalacetate (42), which was then condensed
with triethyl orthoformate (TEOF) in the presence of acetic
anhydride (Ac2O). The reaction could be performed in a
distillation apparatus, and 17 was directly distilled out of the
reaction mixture resulting in a 72% yield over two steps.
Furthermore, the reported workup method, which consisted
of dissolving the crude reaction mixture in sat. NaHCO3
solution followed by extraction, led to formation of barely
soluble magnesium salts, which complicated the extraction.
Table 2. Screening Conditions for Building Up Ring A
a
area % 18
entry
base
t [h]
254 nm
315 nm
1
2
3
4
5
NEt3
DIPEA
2,6-lutidine
N,N-dimethylaniline
pyridine
2
2
81
91
90
48
67
91
94
35
44
37
9
22
71
84
24
24
2
65
65
b
6
pyridine
a
b
Determined by LC−MS and UV absorption. Three equivalents were used.
E
https://doi.org/10.1021/acs.oprd.1c00139
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
acid conditions showed to be appropriate also for ester 18 and
were thus optimized (for more details, see the Supporting
Information). After heating overnight in the presence of an
excess of amine and acetic acid (both 2.5 equiv), full
conversion to dolutegravir (3) was detected by LC−MS. It
turned out be more efficient when the aqueous workup was
omitted. After complete conversion of ester 18, all volatiles
were removed in vacuo, and the residue was dissolved in hot
EtOH and treated with sodium hydroxide (NaOH). After
filtration and washing, the filtered salt was heated again in
EtOH and hot-filtered to increase the purity. After drying,
dolutegravir sodium (DTG-Na, 15) was obtained in a 70%
isolated yield (HPLC purity of 99.7% (254 nm)) (Scheme 17).
The amide coupling of acid 19 with amine 10 has been
reported in the patent literature. By using the expensive
coupling reagent HATU (1-[bis(dimethylamino)methylene]-
1H-1,2,3-triazolo[4,5-b]pyridinium 3-oxide hexafluorophos-
phate), N-methylpyrrolidine, and DMF as the solvent, 15
was isolated in a 55% yield after purification by preparative
HPLC.18 The reaction has also been described for preparing
bictegravir (4), but 1,1-carbonyldiimidazole (CDI) was used
instead as a coupling reagent in this case.19 Following this
protocol, acid 19 was activated with CDI by stirring for 2 h in
dimethyl carbonate (DMC) at 80 °C. Amine 10 was added at
r.t., and clean conversion to 3 was detected by LC−MS after 2
h. Aqueous workup afforded crude dolutegravir (3), which was
converted to the sodium salt as mentioned above giving 15 in a
94% isolated yield showing an HPLC purity of 99.9% (254
nm) (Scheme 18).
Figure 1. Proposed structure of byproduct 45.
Additionally, ester 18 was partly saponified under these
conditions.
For optimization, bases other than NEt3 were investigated
first. When diisopropylethylamine (DIPEA) was used,
conversion slightly improved to 91 area % (254 nm) and 44
area % (315 nm) (entry 2, Table 2). Using 2,6-lutidine as a
base resulted in a similar result to NEt3 (entry 3, Table 2),
while N,N-dimethylaniline gave only low conversion (entry 4,
Table 2). With pyridine, the reaction proceeded much slower
but also cleaner (91 area % (254 nm) and 71 area % (315 nm)
conversion after 65 h) (entry 5, Table 2). Increasing the
equivalents of pyridine (3.0) additionally improved the
conversion (entry 6, Table 2).
For the workup, the reaction mixture was cooled, quenched
with 1 N HCl, and extracted with DCM. When the reaction
was performed on a 2 g scale, a 95% isolated yield of crude 18
was obtained. The orange-reddish solid showed purities of 94
area % (254 nm) and 87 area % (315 nm) (Scheme 14).
Crude 18 could be washed with ethyl acetate to remove
impurities, but as some material was lost in this step, the crude
material was used for the next step. As an alternative procedure
for purification, it turned out that saponification furnished acid
19 in a pure form. A simple one-pot procedure was developed
by adding aqueous 1 N NaOH to the DCM extract of ester 18.
After stirring the two-phase mixture overnight, 18 was
completely saponified to acid 19. Impurities stayed in the
organic phase, and acid 19 precipitated out of the aqueous
phase after acidification. Filtration and drying afterward
afforded 19 as a colorless solid in a 72% yield over three
steps (Scheme 15).
When only 1.7 equiv of CDI were used, incomplete
conversion was observed. As a consequence, the final product
contained more nonremovable traces of acid 19 or ester 18.
The reaction was once performed in a smaller scale (0.3 g)
with skipping the aqueous workup the same as reported for
aminolysis of ester 18. A similar isolated yield (91%) was
obtained, but 15 was slightly less pure (99.7%, 254 nm).
CONCLUSIONS
While investigating the regioselective cyclization, it was also
considered to replace MgBr2 OEt2 by cheaper and more widely
available MgCl2. From all tested conditions, only heating in
MeCN showed promising results (for more details, see the
Supporting Information). Conversion to ester 18 was usually
lower (85 area % (254 nm), 73 area % (315 nm)). As also
partly saponification to acid 19 was observed, it appeared more
reasonable to drive the reaction completely toward 18. After
removing MeCN, workup and saponification were performed
as mentioned above to furnish a 53% isolated yield of 19
(Scheme 16).
■
A practical synthesis route to dolutegravir sodium starting from
(R)-3-amino-1-butanol (13) was introduced (Scheme 19).
First, a new four-step and highly yielding synthesis to the ring
system BC containing amine 16 was developed. It is
noteworthy that the reaction sequence proceeded without
significant byproduct formation and no cost-intensive
purification steps had to be used. The regioselective cyclization
to ring A was carefully optimized and significantly improved.
Furthermore, efficient access to acid 19 was demonstrated.
Ester 18 was isolated in a higher yield admittedly but showed a
lower purity than acid 19. Both compounds could be
transformed to desired DTG-Na in similar overall yields.
Conversion of ester 18 with 2,4-difluorobenzylamine required
harsher conditions but less expensive reagents (toluene and
Amide Coupling. Aminolysis of an ethyl ester moiety with
2,4-difluorobenzylamine by heating in toluene in the presence
of acetic acid has already been reported in the literature for
other dolutegravir building blocks.11,16,17 The toluene/acetic
Scheme 14. Synthesis of Ester 18 by One-Pot 1,4-Addition of Amines 16 with 17 Followed by Regioselective Cyclization
F
https://doi.org/10.1021/acs.oprd.1c00139
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
Scheme 15. Synthesis of Acid 19 by One-Pot 1,4-Addition of Amines 16 to 43 Followed by MgBr2-Mediated Regioselective
Cyclization and Saponification
Scheme 16. Synthesis of Acid 19 by One-Pot 1,4-Addition of Amines 16 to 43 Followed by MgCl2-Mediated Regioselective
Cyclization and Saponification
Scheme 17. Synthesis of DTG-Na (15) by Aminolysis of Crude Ester 18 and Subsequent Salt Formation
Scheme 18. Synthesis of DTG-Na (15) by Amide Coupling of Acid 19 and Subsequent Salt Formation
acetic acid), while acid 19 showed a cleaner conversion under
milder conditions by using CDI as a coupling reagent.
Ultimately, both transformations represent attractive routes
and enable new synthetic access to DTG-Na.
downfield from tetramethylsilane (1H NMR and 13C NMR).
Deuterated solvents (CDCl3 and DMSO-d6) served as an
internal reference. The reported signal splittings were
abbreviated as follows: sb = broad singlet, s = singlet, d =
doublet, and t = triplet. Coupling constants J are reported in
Hz. ESI-MS spectra were recorded on a 1260-series Infinity II
HPLC system (Agilent Technologies) with a binary pump and
an integrated diode array detector coupled to an LC/MSD
Infinitylab LC/MSD (G6125B LC/MSD) mass spectrometer.
For high-resolution (HR) mass spectra, an Agilent 6545 Q-
TOF spectrometer and a suitable external calibrant were used.
Analytical HPLC was carried out with an Agilent 1260 Infinity
system equipped with a binary pump, a diode array detector,
and an LC/MSD InfinityLab LC/MSD (G6125B LC/MSD)
mass spectrometer. An Ascentis Express C18 column (2.7 μm,
2.1 mm × 30 mm, 40 °C) or an ACE C18 PFP column (3 μm,
4.6 mm × 150 mm, 40 °C) with gradient elution using
acetonitrile/water (+0.1% formic acid) and a flow rate of 1.0
EXPERIMENTAL SECTION
■
All employed chemicals were commercially available and used
without prior purification except 2,4-difluorobenzylamine,
which was distilled and stored over a nitrogen atmosphere.
Anhydrous solvents were taken from a solvent purification
system and under a nitrogen atmosphere. Oven-dried glass-
ware was dried in an oven at 150 °C overnight, assembled
while still hot, cooled to room temperature, and then purged
with nitrogen. NMR spectra were recorded on a Bruker
Avance-III HD instrument (1H NMR, 300 MHz; 13C NMR, 75
MHz) or a Bruker Avance-III HD instrument (1H NMR, 400
MHz; 13C NMR, 101 MHz; 19F NMR, 377 MHz) with a 5 mm
BBFO probe. The chemical shifts δ were expressed in ppm
G
https://doi.org/10.1021/acs.oprd.1c00139
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
Scheme 19. Here Reported Synthesis of DTG-Na (15)
mL/min was used. Chiral HPLC was performed on a 1260-
series Infinity II HPLC system (Agilent Technologies) in a
normal phase and isocratic mode with EtOH/n-hexane as the
mobile phase. A Daicel Chiralpak IF-3 column (3 μm, 4.6 mm
× 250 mm, 40 °C) was used for enantiomeric excess
determination. Gas chromatography was performed on an
Agilent 8890 gas chromatograph equipped with a 5977 GC/
MS detector. An Agilent Technologies HP 5MS UI column
(30 m × 0.25 mm × 0.25 μm) as a stationary phase with
helium as a carrier gas and a flow rate of 1.2 mL/min was used.
The following parameters were used: an inlet temperature of
250 °C, a transfer line temperature of 250 °C, an ion source
temperature of 230 °C, an MS-quadrupole temperature of 150
°C, and an initial oven temperature of 40 °C for 2 min with a
temperature ramp of 50 °C/min to 320 °C over 5.6 min
followed by 7.4 min holding. IR spectroscopy was conducted
on a Bruker Tensor 27 FTIR spectrometer using a diamond
ATR unit. Thin-layer chromatography was performed on
Merck F254 silica gel plates. Spots were visualized with UV light
(λ = 254 nm) or stained with appropriate reagents. Melting
̈
points are uncorrected and were taken by using a Kruss
KSP1N digital melting point apparatus. Optical rotations were
measured on a PerkinElmer 241 MC polarimeter.
4-Hydroxybutan-2-one Oxime, 44. According to a
modified procedure by Budidet et al.,20 a solution of 4-
hydroxybutan-2-one (95%, 5.0 mL, 55 mmol, 1.0 equiv) in
EtOH (60 mL) was cooled in an ice bath. Hydroxylamine
hydrochloride (4.6 g, 66.0 mmol, 1.2 equiv) was added, and
the pH was adjusted to 6 by slow addition of aq. sodium
hydroxide solution (40 wt %). The colorless suspension was
stirred for 4 h at r.t. (complete conversion detected by GC−
MS) before it was filtered. The solvent was removed in vacuo
at 40 °C, and the residue was suspended in EtOAc (50 mL).
After drying over Na2SO4, all volatiles were removed in vacuo
at 40 °C to obtain 44 as a mixture of anti/syn isomers (5.60 g,
54.3 mmol, 99%) as a colorless viscous oil. M (C4H9NO2) =
103.12 g/mol. Rf (SiO2) = 0.21 (EtOAc), stained with a
ninhydrin reagent. IR (ATR): ν = 3249, 2889, 1660, 1427,
1
1370, 1261, 1050 cm−1. H NMR, COSY (400 MHz, DMSO-
d6): δ = 10.26/10.17 (s, 1H, −NOH), 4.62−4.56/4.55−4.48
(m, 1H, −OH), 3.59−3.50 (m, 2H, H-4), 2.41/2.25 (t, 3J = 6.8
Hz, H-3), 1.78/1.73 (s, 3H, H-1) ppm. 13C NMR, HSQC,
HMBC (100 MHz, DMSO-d6): δ = 154.1/153.8 (C-2), 58.4/
57.3 (C-4), 38.8/32.1 (C-3), 20.2/13.5 (C-1) ppm. GC−MS:
m/z = 58.1 (100%). ESI-HRMS: calcd for [C4H9NO2 + H]+,
m/z = 104.0706; found, m/z = 104.0703.
rac-3-Aminobutan-1-ol, rac-13. According to a modified
procedure by Budidet et al.,20 a suspension of 4-hydroxy-2-
butanone oxime (44, 8.60 g, 83.4 mmol) and Raney nickel (10
wt %) in MeOH (70 mL) was hydrogenated in an autoclave
for 28 h (10 bar H2, 45 °C) (reaction control by TLC). The
suspension was suction-filtered over celite, and the celite cake
was washed several times with MeOH. All volatiles were
removed in vacuo at 40 °C in order to obtain rac-13 as a
colorless oil (6.91 g, 77.6 mmol, 93%), which was used for the
next step without further purification. M (C4H11NO) = 89.14
g/mol. Tb = 81−83 °C (22 mbar); lit. 95−97 °C (28 mbar).21
Rf (SiO2) = 0.19 (EtOAc:MeOH:NEt3 = 2:1:1), stained with
ninhydrin reagent. IR (ATR): ν = 3347, 3280, 3183, 2957,
1
2924, 2870, 1599, 1455, 1375, 1062 cm−1. H NMR, COSY
H
https://doi.org/10.1021/acs.oprd.1c00139
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
(400 MHz, CDCl3): δ = 3.84−3.72 (m, 2H, H-1), 3.17−3.04
(m, 1H, H-3), 2.69 (sB, 3H, −OH and −NH2), 1.66−1.57 (m,
1H, Ha-2), 1.53−1.42 (m, 1H, Hb-2), 1.13 (d, 3J = 6.4 Hz, 3H,
H-4) ppm. 13C NMR, HMBC, HSQC (101 MHz, CDCl3): δ =
62.4 (C-1), 48.0 (C-3), 39.5 (C-2), 25.8 (C-4) ppm. ESI-
HRMS: calcd for [M + H]+, m/z = 90.0913; found, m/z =
90.0913. The spectrometric data are consistent with literature
values.22
(R)-2-Chloro-N-(4-hydroxybutan-2-yl)acetamide, 26.
In an oven-dried Schlenk flask, 13 (4.00 g, 44.9 mmol, 1.0
equiv) was dissolved in DCM (80 mL). NEt3 (10.0 mL, 71.8
mmol, 1.6 equiv) was added, and the solution was cooled in an
ice bath. Chloroacetyl chloride (25, 5.3 mL, 67 mmol, 1.5
equiv) was added dropwise over 10 min, cooling was removed,
and the dark red-brown solution stirred for 1 h at r.t.
(complete consumption of 13 detected by TLC). While
cooling, first water (64 mL) and then 3 N NaOH (32 mL)
were added to the reaction mixture. The cooling bath was
removed, and the two-phasic mixture was stirred vigorously for
90 min at r.t. (complete saponification to 26 detected by GC−
MS). The organic phase was separated and extracted with
water (3 × 40 mL). The combined aqueous phases were
cooled and adjusted to pH = 7−8 using conc. HCl. The
solution was transferred into a Kutscher-Steudel apparatus and
extracted continuously with EtOAc for 72 h. The orange
organic phase was dried over Na2SO4 and filtered, and the
solvent was removed in vacuo at 40 °C. 26 was obtained as an
orange-brown viscous oil (6.73 g, 40.6 mmol, 91%) and used
(C11H17NO2) = 195.26 g/mol. Tb = 130−138 °C (13 mbar);
lit., 147−149 °C (18 mbar).23 Rf (SiO2) = 0.35 (EtOAc + 1%
NEt3). IR (ATR): ν = 2934, 2830, 1454, 1192, 1127, 1056
1
cm−1. H NMR, COSY (300 MHz, CDCl3): δ = 7.34−7.20
3
(m, 5H, Ar-H), 4.49 (t, J = 5.5 Hz, 1H, H-2), 3.81 (s, 2H,
3
−CH2Ar), 3.37 (s, 6H, 2× OCH3), 2.75 (d, J = 5.5 Hz, 2H,
H-1), 1.54 (sB, 1H, −NH−) ppm. 13C NMR, HMBC, HSQC
(75 MHz, CDCl3): δ = 140.3 (Ar-C-1), 128.5 (Ar-C-3 and Ar-
C-5), 128.3 (Ar-C-2 and Ar-C-6), 127.1 (Ar-C-4), 104.1 (C-
2), 54.1 (2× −OCH3), 54.0 (−CH2Ar), 50.7 (C-1) ppm. ESI-
MS: m/z = 196.1 (100%, [M + H]+). The spectroscopic data
are consistent with literature values.14
(R)-2-(Benzyl(2,2-dimethoxyethyl)amino)-N-(4-hy-
droxybutan-2-yl)acetamide, 34. To a solution of 30 (6.11
g, 36.9 mmol, 1 equiv) in MeCN (40 mL) were added K2CO3
(10.2 g, 73.8 mmol, 2.0 equiv) and KI (6.13 g, 36.9 mmol, 1.0
equiv) while stirring, followed by a solution of 26 (7.21 g, 36.9
mmol, 1.0 equiv) in MeCN (40 mL) and additional MeCN
(130 mL). The suspension was heated to reflux for 3 h (full
conversion detected by LC−MS, 254 nm). The solvent was
removed in vacuo at 40 °C, and the salt-like residue was
suspended in EtOAc (150 mL) and water (210 mL). The
mixture was transferred into a separatory funnel, and the
organic phase was separated. The aqueous phase was extracted
with EtOAc (2 × 100 mL), and the combined organic phases
were dried over Na2SO4. After removing all volatiles in vacuo
at 40 °C, 34 was obtained as a brown oil (11.63 g, 35.8 mmol,
19
589
97%). M (C17H28N2O4) = 324.42 g/mol. [α] = −29.6
for the next step without any further purification. M
(CHCl3, c = 10 mg/mL). Rf (SiO2) = 0.32 (EtOAc + 2%
NEt3). IR (ATR): ν = 3324, 2933, 2833, 1649, 1529, 1453,
20
589
(C6H12ClNO2) = 165.63 g/mol. [α] = −32.7 (CHCl3, c =
1
1120, 1063 cm−1. H NMR, COSY (300 MHz, CDCl3): δ =
10 mg/mL). Rf (SiO2) = 0.30 (EtOAc), stained with ninhydrin
reagent. IR (ATR): ν = 3280, 2936. 1651, 1543, 1056 cm−1.
1H NMR, COSY (300 MHz, CDCl3): δ = 6.68 (sB, 1H, -NH-),
4.27−4.12 (m, 1H, H-2′), 3.71−3.54 (m, 2H, H-4′), 3.01 (sB,
1H, -OH), 1.93−1.80 (m, 1H, Ha-3′), 1.53−1.41 (m, 1H, Hb-
7.59 (m, 1H, −NH−), 7.38−7.22 (m, 5H, Ar-H), 4.39 (t, 3J =
5.3 Hz, H-2″), 4.20−4.04 (m, 1H, H-2′), 3.92−3.80 (m, 1H,
−OH), 3.72 (s, 2H, PhCH2−), 3.59−3.46 (m, 1H, Ha-4′),
3.35 (s, 3H, −OCH34), 3.33 (s, 3H, −OCH3), 3.33−3.30 (m,
3
3
3′), 1.26 (d, J = 1.3 Hz, 3H, H-1′) ppm. 13C NMR, HMBC,
1H, Hb-4′), 3.20 (d, J = 1.8 Hz, 2H, H-2), 2.71 (dd, J = 5.3
Hz, 4J = 1.1 Hz, 2H, H-1″), 1.89−1.76 (m, 1H, Ha-3′), 1.33−
HSQC (75 MHz, CDCl3): δ = 166.6 (C-1), 58.9 (C-4′), 43.2
(C-2′), 42.6 (C-2), 39.5 (C-3′), 20.9 (C-1′) ppm. GC−MS:
m/z = 120.1 (100%). ESI-HRMS: calcd for [M + H]+, m/z =
166.0629; found, m/z = 166.0633.
1.24 (m, 1H, Hb-3′), 1.22 (d, 3J = 6.7 Hz, 3H, H-1′) ppm. 13
C
NMR, HMBC, HSQC (75 MHz, CDCl3): δ = 172.0 (CO),
138.0 (Ar-C), 129.0 (Ar-C), 128.7 (Ar-C), 127.8 (Ar-C),
102.9 (C-2″), 60.7 (PhCH2−), 58.6 (C-4′), 58.6 (C-2), 57.3
(C-1″), 54.1 (−OCH3), 53.9 (−OCH3), 41.4 (C-2′), 40.5 (C-
3′), 21.2 (C-1′) ppm. ESI-HRMS: calcd for [M + H]+, m/z =
325.2122; found, m/z = 325.2130.
N-Benzyl-2,2-dimethoxyethylamine, 30. For method 1,
benzylamine (32, 12.2 g, 113 mmol, 3.0 equiv) and
chloroacetaldehyde-dimethylacetal (31, 4.71 g, 38 mmol, 1.0
equiv) were dissolved in toluene (50 mL) and heated to reflux
for seven days. The suspension was cooled in an ice bath and
filtered, toluene was then removed in vacuo at 40 °C. The
residue was distilled under vacuum to afford 30 as a colorless
liquid (5.71 g, 29.2 mmol, 77%). For method 2, according to a
modified procedure from Luu et al.,14 to a solution of
aminoacetaldehyde dimethyl acetal (8, 7.91 g, 75.2 mmol) in
dry methanol (300 mL), prepared in an oven-dried Schlenk
flask under a nitrogen atmosphere, was added freshly distilled
benzaldehyde (33, 7.60 mL, 75.2 mmol, 1.0 equiv), and the
solution was stirred for 18 h at r.t. While cooling in an ice bath,
NaBH4 (4.30 g, 114 mmol, 1.5 equiv) was added portion-wise
over 3 min. The ice bath was removed, and the suspension was
stirred at r.t. for 18 h. The reaction was quenched by adding
sat. NaHCO3 solution (30 mL), and the mixture was extracted
with DCM (3 × 60 mL). The combined organic phases were
washed once with brine (150 mL), dried over Na2SO4, and
filtered. After removing all volatiles in vacuo at 40 °C, 30 was
obtained as a clear colorless liquid (14.19 g, 72.7 mmol, 97%)
and used for the next step without further purification. M
(4R,9aS)-8-Benzyl-4-methylhexahydro-2H,6H-
pyrazino[2,1-b][1,3]oxazin-6-one, 35. Acetal 34 (11.53 g,
35.5 mmol) was suspended in water (160 mL) and cooled in
an ice bath. Conc. HCl (160 mL) was added through a
dropping funnel over 15 min while stirring. The cooling bath
was removed, and the slightly yellow solution was stirred at r.t.
for 48 h (complete conversion detected by LC−MS at 254
nm). The solution was cooled again in an ice bath, and sodium
hydroxide pellets (75 g) were added in small portions over 2 h
followed by sodium bicarbonate powder (4 g) in order to
adjust the pH to 7−8. While approaching the desired pH, 35
started precipitating, resulting in a murky beige suspension.
EtOAc (300 mL) was added while stirring, and the mixture
was transferred into a separating funnel. The organic phase was
separated, and the aqueous phase was extracted with EtOAc (2
× 200 mL). The combined organic phases were dried over
Na2SO4 and filtered, and the solvent was removed in vacuo at
40 °C. Crude 35 was obtained as a thick orange-brown oil
(8.49, 32.6 mmol, 92%, ≥99.8% ee according to chiral HPLC).
I
https://doi.org/10.1021/acs.oprd.1c00139
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
Racemic 35 was synthesized analogously from rac-13. M
cooling down, distillation under high vacuum afforded 17 as a
yellow clear liquid (12.0 g, 49.2 mmol, 72%). M (C11H16O6) =
244.24 g/mol. Tb = 116−120 °C (0.45 mbar); lit., 155−160
°C (1.3 mbar).15 Rf (SiO2) = 0.15 and 0.81 (EtOAc). IR
19
(C15H20N2O2) = 260.34 g/mol. [α] = −46.1 (CHCl3, c = 10
589
mg/mL). Rf (SiO2) = 0.32 (EtOAc). IR (ATR): ν = 2969,
1
2860, 1653, 1455, 1328, 1197, 1095, 1068 cm−1. H NMR,
1
COSY (300 MHz, CDCl3): δ = 7.36−7.23 (m, 5H, Ar-H),
4.97−4.86 (m, 2H, H-4 and H-9a), 3.93−3.81 (m, 2H, H-2),
(ATR): ν = 2987, 2937, 1359, 1256, 1177, 1019 cm−1. H
NMR, COSY (300 MHz, CDCl3): δ = 7.90 (s, 1H, CH−),
7.88 (s, 1H, CH′−), 4.40−4.32 (m, 2H, CHOCH2−),
4.40−4.32 (m, 2H, CHOCH′2−), 4.32−4.27 (m, 2H, O
C-4-OCH2−), 4.32−4.27 (m, 2H, OC-4′-OCH2−), 4.27−
4.18 (m, 2H, OC-1-OCH2−), 4.27−4.18 (m, 2H, OC-1′-
OCH2−), 1.44 (t, 3J = 7.2 Hz, 3H, CHOCH2CH3), 1.43 (t,
2
2
3.57 (d, J = 13.1 Hz, 1H, −NCHaAr), 3.54 (d, J = 13.1 Hz,
2
4
1H, −NCHbAr), 3.24 (dd, J = 16.0 Hz, J = 1.9 Hz, 1H, Ha-
7), 3.02 (dd, 2J = 16.3 Hz, 4J = 0.8 Hz, 1H, Hb-7), 2.97 (ddd, 2J
= 12.0 Hz, 3J = 4.9 Hz, 4J = 1.9 Hz, 1H, Ha-9), 2.44 (ddd, 2J =
3
4
12.0 Hz, J = 6.7 Hz, J = 0.8 Hz, 1H, Hb-9), 2.19−2.04 (m,
3
3
1H, Ha-3), 1.43−1.34 (m, 1H, Hb-3), 1.27 (d, J = 7.1 Hz, 4-
3J = 7.2 Hz, 3HCHOCH2CH′3), 1.36 (t, J = 7.2 Hz, 3H,
CH3) ppm. 13C NMR, HSQC, HMBC (75 MHz, CDCl3): δ =
166.3 (C-6), 136.3 (Ar-H), 129.4 (Ar-H), 128.6 (Ar-H), 127.7
(Ar-H), 79.3 (C-9a), 62.8 (C-2), 61.6 (−NCH2Ar), 57.5 (C-
7), 54.3 (C-9), 41.7 (C-4), 29.8 (C-3), 15.9 (−CH3) ppm.
ESI-HRMS: calcd for [M + H]+, m/z = 261.1594; found, m/z
= 261.1598.
OC-4-OCH2CH3), 1.35 (t, 3J = 7.2 Hz, 3H, OC-4-
3
OCH2CH′3), 1.28 (t, J = 7.2 Hz, 3H, OC-1-OCH2CH3),
3
1.27 (t, J = 7.2 Hz, 3H, OC-1-OCH2CH′3) ppm. 13C
NMR, HSQC, HMBC (75 MHz, CDCl3): δ = 185.2 (C-3),
183.2 (C-3′), 170.1 (CH−), 170.0 (CH′−), 165.2 (C-1),
164.2 (C-4), 164.0 (C-4′), 163.5 (C-1′), 109.6 (C-2), 108.3
(C-2′), 74.7 (CH−O−CH2−), 74.6 (CH−O−C′H2−),
62.2 (OC-1-OCH2−), 62.0 (OC-1′-OCH2−), 61.1 (O
C-4-OCH2−), 61.0 (OC-4′-OCH2−), 15.4 (
CHOCH2CH3), 15.4 (CHOCH2C′H3), 14.3 (OC-1-
OCH2CH3), 14.2 (OC-1′-OCH2CH3), 14.1 (OC-4-
OCH2CH3), 14.1 (OC-4′-OCH2CH3) ppm. ESI-MS: m/z
= 217.1 (100%, [M-Et + H]+). Ethyl enol ether hydrolyzes
during the LC−MS run to the free enol. The spectrometric
data are consistent with literature values.15
(4R,9aS)-4-Methylhexahydro-2H,6H-pyrazino[2,1-b]-
[1,3]oxazin-6-one, 16. A solution of oxazinone 35 (8.44 g,
324 mmol) in MeOH (120 mL) was degassed for 10 min by
purging with nitrogen. Palladium (10% on carbon, 0.85 g) was
added and the mixture was purged with hydrogen three times.
The mixture stirred under hydrogen atmosphere for seven days
(complete conversion detected by LC−MS and UV detection
at 254 nm). After purging for 10 min with nitrogen, the
mixture was suction-filtered over celite and washed several
times with MeOH. All volatiles were removed in vacuo at 40
°C. Crude 16 was obtained as a viscous yellow-orange oil (5.31
g, 31.2 mmol, 96%), which solidified to a slight yellowish solid
after a while. 16 was used for the next step without further
Ethyl (4R,12aS)-7-Hydroxy-4-methyl-6,8-dioxo-
3,4,6,8,12,12a-hexahydro-2H-pyrido[1′,2′:4,5]pyrazino-
[2,1-b][1,3]oxazine-9-carboxylate, 18. Amine 16 (2.00 g,
11.8 mmol, 1.0 equiv) was added in a single portion to a
solution of enol ether 17 (2.87 g, 11.8 mmol, 1.0 equiv) in dry
DCM (80 mL), which was prepared in an oven-dried Schlenk
flask under a nitrogen atmosphere. The yellow-greenish
solution was stirred at r.t. for 24 h (full conversion of 17
detected by LC−MS and UV detection at 254/315 nm) before
it was cooled in an ice bath. MgBr2·OEt2 (3.79 g, 14.7 mmol,
1.25 equiv) was added all at once under a nitrogen reverse
flow, and the suspension was stirred for 10 min before dry
pyridine (2.8 mL, 35 mmol, 3.0 equiv) was dripped into the
yellowish suspension within 2 min. A clear orange-red solution
formed immediately, which was stirred at r.t. for three days
(full conversion of 43 detected by LC−MS and UV detection
at 254/315 nm). The suspension was cooled in an ice bath,
and 1 N HCl (60 mL) was added while stirring. The mixture
was transferred into a separatory funnel and vigorously shaken
before the organic phase was separated. The aqueous phase
was extracted with DCM (2 × 50 mL), combined organic
phases were dried over Na2SO4 and filtered, and the solvent
was removed in vacuo at 40 °C. Crude 18 was obtained as an
orange fluffy solid (3.61 g, 11.2 mmol, 95%, HPLC purities of
94% (254 nm) and 87% (315 nm)) and used for the next step
without further purification. To obtain a pure material, crude
18 was heated in EtOAc (0.1 g/2 mL) and cooled down to r.t.
first then to −24 °C (freezer). The solid was filtered off,
washed with ice-cold EtOAc, and dried in vacuo at 40 °C. M
(C15H18N2O6) = 322.32 g/mol. Tm = 90−96 °C (sintering to
an orange resin) and 104−106 °C (resin melts to a yellow-
greenish liquid). [α]25819= −35.6 (CHCl3, c = 10 mg/mL). Rf
(C18-SiO2) = 0.58 (EtOH:H2O = 1:1 + 10% HOAc). IR
(ATR): ν = 2979, 1725, 1632, 1452, 1282, 1262, 1181, 1090,
purification. M (C8H14N2O2) = 170.21 g/mol. Tm = 56−61
19
589
°C; 79−83 °C (rac.). [α] = −103.1 (CHCl3, c = 10 mg/
mL). Rf (SiO2) = 0.31 (EtOAc + 10% MeOH + 5% NEt3),
stained with ninhydrin reagent. IR (ATR): ν = 3307, 2968,
1
2865, 1643, 1451, 1324, 1193, 1080, 1064 cm−1. H NMR,
COSY (300 MHz, CDCl3): δ = 5.04−4.92 (m, 1H, H-4),
4.85−4.80 (m, 1H, H-9a), 4.00−3.85 (m, 1H, H-2), 3.48 (d, 2J
2
= 17.3 Hz, 1H, Ha-7), 3.38 (d, J = 17.3 Hz, 1H, Hb-7), 3.14
2
3
2
(dd, J = 13.6 Hz, J = 3.8 Hz 1H, Ha-9), 2.97 (dd, J = 13.6
3
Hz, J = 4.3 Hz 1H, Hb-9), 2.16−2.01 (m, 1H, (Ha-3), 1.78
3
(sB, 1H, −NH−), 1.41−1.32 (m, 1H, Hb-3), 1.26 (t, J = 7.1
Hz, 3H, −CH3) ppm. 13C NMR, HSQC, HMBC (75 MHz,
CDCl3): δ = 167.8 (C-6), 78.8 (C-9a), 63.0 (C-2), 50.3 (C-7),
48.1 (C-9), 42.2 (C-4), 29.9 (C-3), 15.9 (−CH3) ppm. ESI-
MS: m/z = 171.1 (100%, [M + H]+). The spectrometric data
are consistent with literature values.10
Diethyl-(2E/Z)-2-(ethoxymethylidene)-3-oxobutan-
dioate, 17. A diethyl oxalacetate sodium salt (41, 95%, 15.0 g,
67.8 mmol) was weighed into an Erlenmeyer flask and
suspended in EtOAc (90 mL). The suspension was cooled in
an ice bath, and 1 N HCl (86 mL) was added while stirring.
After all of the salt was dissolved, the biphasic murky mixture
was transferred into a separatory funnel. The organic phase was
separated, and the aqueous phase was extracted with EtOAc (2
× 45 mL). The combined organic phases were dried over
Na2SO4 and filtered, and the solvent was removed in vacuo at
30 °C. To the orange-brown oily residue (13.1 g) were added
triethyl orthoformate (20.7 mL, 122 mmol, 1.8 equiv) and
acetic anhydride (17.9 mL, 190 mmol, 2.8 equiv). The flask
was equipped with a distillation apparatus, and the solution
was heated for 1 h to 120 °C, for 1 h to 130 °C, and for 1 h to
140 °C, while a colorless clear liquid was distilled off. After
1
1048 cm−1. H NMR, COSY (300 MHz, CDCl3): δ = 12.33
(sB, 1H, −OH), 7.91 (s, 1H, H-10), 5.40−5.28 (m, 1H, H-
J
https://doi.org/10.1021/acs.oprd.1c00139
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
12a), 5.00−4.87 (m, 1H, H-4), 4.40−4.19 (m, 3H, Ha-12 and
−OCH2−), 4.12−3.91 (m 3H, H-2 and Hb-12), 2.29−2.11
(m, 1H, Ha-3), 1.60−1.49 (m, 1H, Hb-3), 1.43 (t, 3J = 7.1 Hz,
3H, 4-CH3), 1.33 (t, 3J = 7.0 Hz, 3H, −OCH2CH3) ppm. 13C
NMR, HMBC, HSQC (75 MHz, CDCl3): δ = 169.8 (C-8),
164.3 (−COOEt), 162.5 (C-6), 156.5 (C-7), 141.5 (C-10),
115.3 (C-6a), 114.8 (C-9), 76.4 (C-12a), 62.8 (C-2), 61.1
(−OCH2), 52.6 (C-12), 44.8 (C-4), 29.5 (C-3), 15.7 (4-CH3),
14.4 (−OCH2CH3) ppm. ESI-HRMS: calcd for [M + H]+, m/
z = 323.1238; found, m/z = 323.1227.
(10 mL). The aqueous phase was cooled in an ice bath, and
conc. HCl was slowly added for acidification to pH = 1−2. A
colorless solid precipitated, which was vacuum-filtered and
washed with water (4 × 2 mL). The solid was dried in air first
then at 70 °C under high vacuum to obtain 19 (0.46 g, 1.56
mmol, 53%). M (C13H14N2O6) = 294.26 g/mol. Tm = 248−
21
589
252 °C (decomposition). [α] = −120.9 (MeCN, c = 10 mg/
mL). Rf (C18-SiO2) = 0.65 (EtOH:H2O = 2:1 + 20% HOAc).
IR (ATR): ν = 1735, 1646, 1618, 1546, 1461, 1440, 1346,
1
1285, 1095, 1081 cm−1. H NMR, COSY (400 MHz, DMSO-
(4R)-7-Hydroxy-4-methyl-6,8-dioxo-3,4,6,8,12,12a-
hexahydro-2H-pyrido[1′,2′:4,5]pyrazino[2,1-b][1,3]-
oxazine-9-carboxylic Acid, 19. For method 1, a solution of
enol ether 17 (2.87 g, 11.8 mmol, 1.0 equiv) in dry DCM (80
mL) was prepared in an oven-dried Schlenk flask under a
nitrogen atmosphere. Amine 16 (2.00 g, 11.8 mmol, 1.0 equiv)
was added all at once, and the yellow-greenish solution was
stirred at r.t. for 24 h (full conversion of 17 detected by LC−
MS and UV absorption at 254/315 nm) and then cooled in an
ice bath. MgBr2·OEt2 (3.79 g, 14.7 mmol, 1.25 equiv) was
added all at once under a nitrogen reverse flow, and the
suspension was stirred for 10 min before dry pyridine (2.8 mL,
35 mmol, 3.0 equiv) was dripped into the yellowish suspension
within 2 min. A clear orange-red solution formed immediately,
which was stirred at r.t. for three days (full conversion of 43
detected by LC−MS and UV detection at 254 nm). The
solution suspension was cooled in an ice bath, and 1 N HCl
(60 mL) was added while stirring. The mixture was transferred
into a separating funnel and vigorously shaken, and the organic
phase was separated. The aqueous phase was extracted with
DCM (2 × 50 mL), and to the combined organic phases was
added 1 N NaOH (60 mL). The two-phase mixture was stirred
vigorously at r.t. for 24 h (complete saponification detected by
LC−MS and UV detection at 254 nm). The aqueous phase
was separated, extracted once with DCM (50 mL), and cooled
in an ice bath. The pH was adjusted to 1−2 by slow addition of
conc. HCl (5 mL) where a colorless suspension formed. The
solid was vacuum-filtered and washed several times with cold
water (5 × 3 mL). The colorless solid was dried in air first then
at 70 °C in fine vacuum to obtain 19 (2.49 g, 8.46 mmol,
72%). For method 2, a solution of enol ether 17 (0.72 g, 2.94
mmol, 1.0 equiv) in dry MeCN (25 mL) was prepared in an
oven-dried Schlenk flask under a nitrogen atmosphere. Amine
16 (0.5 g, 2.94 mmol, 1.0 equiv) was added in a single portion,
and the yellow-greenish solution was heated to reflux for 16 h
(full conversion of 17 detected by LC−MS and UV absorption
at 254/315 nm). The orange solution was cooled to r.t. before
anhydrous MgCl2 (0.35 g, 3.67 mmol, 1.25 equiv) was added.
The suspension was stirred for 10 min at r.t. before dry
pyridine (0.71 mL, 8.81 mmol, 3.0 equiv) was added. The
mixture was heated to reflux under a nitrogen atmosphere for
24 h (95% conversion of 43 as judged by LC−MS and UV
detection 254 nm). MeCN was removed in vacuo at 40 °C,
and to the brownish salty residue was added DCM (15 mL).
The suspension was cooled in an ice bath, and 1 N HCl (15
mL) was added while stirring vigorously. The mixture was
transferred into a separating funnel and shaken vigorously. The
organic phase was separated, and the aqueous phase was
extracted with DCM (2 × 10 mL). 1 N NaOH (15 mL) was
added to the combined organic phases, and the mixture was
stirred vigorously at r.t. for 24 h (complete saponification
detected by LC−MS and UV detection at 254 nm). The
aqueous phase was separated and extracted once with DCM
d6): δ = 15.39 (sB, 1H, −COOH), 12.77 (sB, 1H, −OH), 8.67
(s, 1H, H-10), 5.52−5.47 (m, 1H, H-12a), 4.84−4.75 (m, 1H,
2
3
H-4), 4.65 (dd, J = 13.9 Hz, J = 4.6 Hz, 1H, Ha-12), 4.43
(dd, 2J = 13.9 Hz, 3J = 5.9 Hz, 1H, Hb-12), 4.09−3.99 (m, 1H,
Ha-2), 3.95−3.87 (m, 1H, Hb-2), 2.09−1.97 (m, 1H, Ha-3),
3
1.60−1.53 (m, 1H, Hb-3), 1.34 (t, J = 7.0 Hz, 3H, −CH3)
ppm. 13C NMR, HMBC, HSQC (101 MHz, DMSO-d6): δ =
172.2 (C-8), 165.4 (−COOH), 161.8 (C-6), 153.6 (C-7),
141.1 (C-10), 118.7 (C-6a), 113.0 (C-9), 76.0 (C-12a), 62.0
(C-2), 51.5 (C-12), 44.9 (C-4), 29.1 (C-3), 15.2 (−CH3)
ppm. ESI-HRMS: calcd for [M + H]+, m/z = 295.0925; found,
m/z = 295.0929.
Sodium (4R,12aS)-9-((2,4-Difluorobenzyl)carbamoyl)-
4-methyl-6,8-dioxo-3,4,6,8,12,12a-hexahydro-2H-
pyrido[1′,2′:4,5]pyrazino[2,1-b][1,3]oxazin-7-olate (Do-
lutegravir Sodium), 15. For method 1, an oven-dried
Schlenk flask was charged with ester 18 (94%, 1.00 g, 2.92
mmol, 1.0 equiv), dry toluene (30 mL), acetic acid (0.42 mL,
7.29 mmol, 2.5 equiv), and 2,4-difluorobenzylamine (10, 0.87
mL, 7.29 mmol, 2.5 equiv) under a nitrogen atmosphere. The
mixture was heated to reflux for 22 h (full conversion of 18
detected by LC−MS and UV detection at 254/315 nm). All
volatiles were removed in vacuo at 40 °C, and the residue was
dissolved in EtOH (30 mL) by heating to reflux. NaOH (0.13
g, 3.21 mmol, 1.1 equiv) was added, and the solution was
heated again to reflux for 2 min during which a beige
suspension formed. The mixture was stirred to r.t., filtered, and
washed with EtOH (4 × 2 mL). The yellow filter cake was
dried in air overnight and transferred into a new flask. EtOH
(15 mL) was added, and the suspension was heated again to
reflux, hot-filtered, and washed with hot ethanol (4 × 2 mL).
The solid was dried in air overnight then for 2 h at 70 °C in
fine vacuum to obtain 15 as a faint yellow solid (0.90 g, 2.04
mmol, 70%, purity of 99.7% (254 nm)). For method 2, an
oven-dried flask was charged with acid 19 (1.00 g, 3.40 mmol,
1.0 equiv), carbonyl diimidazole (97%, 1.14 g, 6.80 mmol, 2.0
equiv), and dry dimethyl carbonate (30 mL) under a nitrogen
atmosphere. The suspension was heated to 80 °C for 2 h
during which a nearly clear orange solution formed. After
cooling to r.t., 2,4-difluorobenzylamine (10, 0.81 mL, 6.80
mmol, 2.0 equiv) was added dropwise within 2 min, and the
solution was stirred for 2 h at r.t. (full conversion of 19
detected by LC−MS and UV detection at 254/315 nm). The
solvent was removed in vacuo at 40 °C, and the residue was
redissolved in DCM (30 mL) and 1 N NaOH (30 mL). After
stirring for 18 h at r.t., the colorless suspension was transferred
into a separatory funnel. The organic phase was separated, and
the aqueous phase was extracted twice with DCM (30 mL).
The aqueous phase was cooled in an ice bath, acidified (pH =
1−2) with conc. HCl, and extracted with DCM (3 × 25 mL).
The combined organic phases were dried over Na2SO4, and all
volatiles were removed in vacuo at 40 °C. The colorless foamy
residue was dissolved by heating in EtOH (30 mL). NaOH
K
https://doi.org/10.1021/acs.oprd.1c00139
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
(0.15 g, 3.74 mmol, 1.1 equiv) was added to the hot solution,
and heating to reflux was continued for 2 min. The suspension
was stirred to r.t., filtered, and washed with EtOH (5 × 3 mL).
The solid was dried in air overnight then for 2 h at 70 °C
under high vacuum to obtain 15 as a faint yellow solid (1.41 g,
3.19 mmol, 94%, purity of 99.9% (254 nm)). M
(C20H18F2N3NaO5) = 441.37 g/mol. Tm = 314 °C
https://pubs.acs.org/10.1021/acs.oprd.1c00139
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported by the Bill and Melinda Gates
Foundation through the Medicines for All initiative. We thank
Dr. J. C. Liermann (Mainz) for NMR spectroscopy and Dr. C.
J. Kampf (Mainz) for mass spectrometry. Parts of this research
were conducted using the supercomputer MOGON and/or
advisory services offered by the Johannes Gutenberg University
Mainz (hpc.uni-mainz.de), which is a member of the AHRP
(Alliance for High Performance Computing in Rhineland
Palatinate, www.ahrp.info) and the Gauss Alliance e.V. The
authors gratefully acknowledge the computing time granted on
the supercomputer MOGON at the Johannes Gutenberg
University Mainz (hpc.uni-mainz.de).
(decomposition); lit., 296 °C.24 [α] = −46.4 (DMSO-d6, c
21
589
= 10 mg/mL). IR (ATR): ν = 2975, 2913, 1641, 1537, 1504,
1
1424, 1321, 1274, 1258, 1106, 1093 cm−1. H NMR, COSY
(400 MHz, DMSO-d6): δ = 10.69 (t, 3J = 6.0 Hz, 1H, −NH−),
7.89 (s, 1H, H-10), 7.39−7.27 (m, 1H, Ar-H-6), 7.25−7.14
(m, 1H, Ar-H-3), 7.06−6.93 (m, 1H, Ar-H-5), 5.22−5.10 (m,
1H, H-12a), 4.87−4.72 (m, 1H, H-4), 4.50 (d, 3J = 6.0 Hz, 2H,
−NHCH2Ar), 4.36−4.24 (m, 1H, Ha-11), 4.21−4.08 (m, 1H,
Hb-11), 4.04−3.87 (m, 1H, Ha-2), 3.86−3.73 (m, 1H, Hb-2),
1.96−1.76 (m, 1H, Ha-3), 1.45−1.29 (m, 1H, Hb-3), 1.23 (t, 3J
= 7.0 Hz, 3H, −CH3) ppm. 13C NMR, HMBC, HSQC (101
MHz, DMSO-d6): δ = 177.9 (C-8), 167.0 (C-7), 166.0
1
3
(−CONH−), 162.0 (dd, J = 247 Hz, J = 12.3 Hz, Ar-C2),
REFERENCES
■
1
3
161.2 (C-6), 158.8 (dd, J = 249 Hz, J = 12.3 Hz, Ar-C4),
(1) GLOBAL HIV AND AIDS STATISTICS; 2019, https://www.
avert.org/global-hiv-and-aids-statistics (accessed March 29, 2021).
(2) Cihlar, T.; Fordyce, M. Current status and prospects of HIV
treatment. Curr. Opin. Virol. 2016, 18, 50−56.
3
3
134.4 (C-10), 130.5 (dd, J = 9.2 Hz, J = 6.2 Hz, Ar-C6),
2
4
115.0 (C-9), 111.3 (dd, J = 20.9 Hz, J = 3.6 Hz, Ar-C5),
2
108.9 (C-6a), 103.7 (t, J = 25.7 Hz, Ar-C3), 75.6 (C-12a),
62.0 (C-2), 53.1 (C-11), 43.1 (C-4), 35.4 (3J = 3.7 Hz,
−NHCH2Ar), 29.2 (C-3), 15.3 (-CH3) ppm. 19F NMR (377
MHz, DMSO-d6): δ = −112.9, −115.2 ppm. ESI-MS: m/z =
418.2 (100%, [M−Na]−). The spectroscopic data are
consistent with literature values.11
(3) FDA Approves First Extended-Release, Injectable Drug Regimen for
Adults Living with HIV; 2021, https://www.fda.gov/news-events/
press-announcements/fda-approves-first-extended-release-injectable-
drug-regimen-adults-living-hiv (accessed March 21, 2021).
(4) Hughes, D. L. Review of Synthetic Routes and Final Forms of
Integrase Inhibitors Dolutegravir, Cabotegravir, and Bictegravir. Org.
Process Res. Dev. 2019, 23, 716−729.
(5) In 2020, our web migration project tackled over 180,000 pages of
content and over 200,000 publications. Much of our content has been
updated, made more dynamic and may no longer be found in the same
place; 2021, https://www.who.int/hiv/pub/arv/arv-update-2019-
policy/en/ (accessed March 21, 2021).
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/acs.oprd.1c00139.
■
sı
Optimization studies, chromatograms, VCD spectrosco-
py, computational details, and NMR spectra (PDF)
XYZ coordinates of benzyloxazinone 35 (ZIP)
(6) Topics in heterocyclic chemistry. Synthesis of heterocycles in
̌
contemporary medicinal chemistry; Casar, Z.; Barth, R., Eds.; 44,
Springer: Switzerland, 2016.
(7) Phull, M. S.; Shah, B.; Kadlag, M.; Rao, D. R.; Malhotra, G.
Continues flow process for the preparation of active pharmaceutical
ingredients - polycyclic carbamoyl pyridone derivatives and
intermediates thereof. Patent WO 2019/159199 A1, August 22, 2019.
(8) Yasukata, T.; Masui, M.; Ikarashi, F.; Okamoto, K.; Kurita, T.;
Nagai, M.; Sugata, Y.; Miyake, N.; Hara, S.; Adachi, Y.; et al. Practical
Synthetic Method for the Preparation of Pyrone Diesters: An Efficient
Synthetic Route for the Synthesis of Dolutegravir Sodium. Org. Process
Res. Dev. 2019, 23, 565−570.
(9) Aoyama, Y.; Hakogi, T.; Fukui, Y.; Yamada, D.; Ooyama, T.;
Nishino, Y.; Shinomoto, S.; Nagai, M.; Miyake, N.; Taoda, Y.; et al.
Practical and Scalable Synthetic Method for Preparation of
Dolutegravir Sodium: Improvement of a Synthetic Route for Large-
Scale Synthesis. Org. Process Res. Dev. 2019, 23, 558−564.
(10) Jee, J.; Gade, N.; Roper, T. Novel pyrrole and pyridone
derivatives and uses thereof. Patent WO 2020/081143 A1, April 23,
2020.
(11) Sankareswaran, S.; Mannam, M.; Chakka, V.; Mandapati, S. R.;
Kumar, P. Identification and Control of Critical Process Impurities:
An Improved Process for the Preparation of Dolutegravir Sodium.
Org. Process Res. Dev. 2016, 20, 1461−1468.
(12) https://medicines4all.vcu.edu/our-portfolio/bmgf-projects/r-
3-aminobutanol-batch-/ (accessed March 12, 2021).
(13) Lerner, C.; Krei, L.; Hilpert, H. Pyrimidone derivatives and
their use in treatment, amelioration or prevention of a viral disease.
Patent WO 2017/158151 A1, September 21, 2017.
AUTHOR INFORMATION
Corresponding Author
Till Opatz − Department of Chemistry, Johannes Gutenberg
University, 55128 Mainz, Germany; orcid.org/0000-
0002-3266-4050; Email: opatz@uni-mainz.de
■
Authors
Jule-Philipp Dietz − Department of Chemistry, Johannes
Gutenberg University, 55128 Mainz, Germany
Tobias Lucas − Department of Chemistry, Johannes Gutenberg
University, 55128 Mainz, Germany
Jonathan Groß − Department of Chemistry, Johannes
Gutenberg University, 55128 Mainz, Germany
Sebastian Seitel − Department of Chemistry, Johannes
Gutenberg University, 55128 Mainz, Germany
Jan Brauer − Department of Chemistry, Johannes Gutenberg
University, 55128 Mainz, Germany
Dorota Ferenc − Department of Chemistry, Johannes
Gutenberg University, 55128 Mainz, Germany
B. Frank Gupton − Department of Chemical and Life Sciences
Engineering, Virginia Commonwealth University, Richmond,
Virginia 23284, United States; orcid.org/0000-0002-
8165-1088
(14) Luu, Q. H.; Guerra, J. D.; Castaneda, C. M.; Martinez, M. A.;
̃
Saunders, J.; Garcia, B. A.; Gonzales, B. V.; Aidunuthula, A. R.; Mito,
Complete contact information is available at:
L
https://doi.org/10.1021/acs.oprd.1c00139
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Article
S. Ultrasound assisted one-pot synthesis of benzo-fused indole-4,9-
dinones from 1,4-naphthoquinone and α-aminoacetals. Tetrahedron
Lett. 2016, 57, 2253−2256.
(15) Jones, R. G. The Synthesis of Ethyl Ethoxymethyleneox-
alacetate and Related Compounds. J. Am. Chem. Soc. 1951, 73, 3684−
3686.
(16) Sumino, Y.; Masui, M.; Yamada, D.; Ikarashi, F.; Okamoto, K.
Process fpr preparing compound having hiv integrase inhibitory
activity. Patent US 2014/0011995 A1, January 09, 2014.
(17) Ziegler, R. E.; Desai, B. K.; Jee, J.-A.; Gupton, B. F.; Roper, T.
D.; Jamison, T. F. 7-Step Flow Synthesis of the HIV Integrase
Inhibitor Dolutegravir. Angew. Chem., Int. Ed. 2018, 57, 7181−7185.
(18) Sumino, Y.; Okamoto, K.; Masui, M.; Yamada, D.; Ikarashi, F.
Process for preparing compund having HIV integrase activity. Patent
EP 2602260 A1, June 12, 2013.
(19) Phull, M.; Rao, D.; Birari, D. A process for the preparation of
bictegravir and intermediate thereof. Patent WO 2018/229798 A1,
December 20, 2018.
(20) Budidet, S.; Dussa, N.; Kaki, G.; Yatcherla, S.; Sanapureddy, J.;
Danda, S.; Katuroju, S.; Meenakshisunderam, S. An improved
proecess for the preparation of dolutegravir. Patent WO 2014/
128545 A2, August 24, 2014.
(21) Juhász, M.; Lázár, L.; Fulöp, F. Substituent effects in the ring-
̈
chain tautomerism of 4-alkyl-2-aryl substituted oxazolidines and
tetrahydro-1,3-oxazines. J. Heterocycl. Chem. 2007, 44, 1465−1473.
(22) Gu, N.; Liu, M.; Wang, H.; Sun, S.; Zhou, Z.; Hu, W.; Yu, J.-T.;
Cheng, J. Iridium-catalyzed annulation between 1,2-diarylethanone
and 3-aminopropanol toward site-specific 2,3-diaryl pyridines.
Tetrahedron Lett. 2017, 58, 3398−3400.
(23) Kaye, I. A.; Minsky, I. New Compounds. Preparation of N-
Substituted Aminoacetals. J. Am. Chem. Soc. 1949, 71, 2272−2273.
(24) Hotter, A.; Thaler, A.; Lebar, A.; Jankovic, B.; Naversnik, K.;
Klancar, U.; Abramovic, Z. Novel Hydrates of Dolutegravir Sodium.
Patent WO 2016/016279 A1, February 16, 2016.
M
https://doi.org/10.1021/acs.oprd.1c00139
Org. Process Res. Dev. XXXX, XXX, XXX−XXX