Synthesis of Rovafovir Etalafenamide (Part III): Evolution of the Synthetic Process to the Phosphonamidate Fragment

Article abstract of DOI:10.1021/acs.oprd.0c00428

Phosphonamidate 1 is a key fragment in the assembly of rovafovir etalafenamide, a novel nucleotide reverse transcriptase inhibitor under development at Gilead Sciences for the treatment of HIV infection. An early manufacturing route, relying on simulated moving bed (SMB) chromatography for the separation of phosphorus diastereomers, was executed on scale to produce multiple batches of 1. However, developing alternative synthetic conditions became desirable in consideration of the high production cost, long lead time, and high process mass intensity (PMI) associated with SMB. Several strategies to improve these factors are described herein, including epimerization and recycling of the undesired (R)-phosphorus diastereomer, design of stereoselective approaches to establish the desired (S)-configuration at phosphorus, and identification of conditions or derivatives to allow for selective crystallization. Ultimately, a second-generation route to 1 was developed and demonstrated on scale. The new route achieves the separation of phosphorus diastereomers by means of selective crystallization, does not require SMB, and offers lower PMI, cost, and lead time.

Full text of DOI:10.1021/acs.oprd.0c00428

pubs.acs.org/OPRD
Article
Synthesis of Rovafovir Etalafenamide (Part III): Evolution of the
Synthetic Process to the Phosphonamidate Fragment
Andrea Ambrosi,* Dustin A. Bringley, Selcuk Calimsiz, Jeffrey A. O. Garber, Huy Huynh, Sankar Mohan,
Keshab Sarma, Jinyu Shen, Jonah Curl, Bernard Kwong, Olga Lapina, Edmund Leung, Lennie Lin,
Andrew Martins, Teague McGinitie, Jaspal Phull, Ben Roberts, Mary Rosario, Bing Shi, Eric A. Standley,
Li Wang, Xueqing Wang, and Guojun Yu
Cite This: Org. Process Res. Dev. 2021, 25, 1247−1262
Read Online
sı
*
Metrics & More
Article Recommendations
Supporting Information
ACCESS
ABSTRACT: Phosphonamidate 1 is a key fragment in the assembly of rovafovir etalafenamide, a novel nucleotide reverse
transcriptase inhibitor under development at Gilead Sciences for the treatment of HIV infection. An early manufacturing route,
relying on simulated moving bed (SMB) chromatography for the separation of phosphorus diastereomers, was executed on scale to
produce multiple batches of 1. However, developing alternative synthetic conditions became desirable in consideration of the high
production cost, long lead time, and high process mass intensity (PMI) associated with SMB. Several strategies to improve these
factors are described herein, including epimerization and recycling of the undesired (R)-phosphorus diastereomer, design of
stereoselective approaches to establish the desired (S)-configuration at phosphorus, and identification of conditions or derivatives to
allow for selective crystallization. Ultimately, a second-generation route to 1 was developed and demonstrated on scale. The new
route achieves the separation of phosphorus diastereomers by means of selective crystallization, does not require SMB, and offers
lower PMI, cost, and lead time.
KEYWORDS: phosphonamidate, phosphorus, stereochemistry, SMB, crystallization
relationship studies demonstrated that the phosphonamidate
prodrug rovafovir etalafenamide (GS-9131, 6) possesses
enhanced oral bioavailability and in vitro activity relative to its
precursor 5 and other prodrug moieties.3 The favorable
pharmacokinetic profile of rovafovir etalafenamide, along with
its unique activity against NRTI-resistant HIV-1 strains,
prompted its clinical development at Gilead Sciences.
1. INTRODUCTION
Phosphonamidate prodrugs have been part of the medicinal
chemistry toolbox since the early 1990s, when McGuigan and
co-workers pioneered the ProTide (pronucelotide) strategy to
improve the bioavailability and potency relative to the parent
phosphonic acids.1 Within this context, the prodrug tenofovir
alafenamide fumarate (TAF, 4, Figure 1), a nucleotide reverse
Our retrosynthetic approach to 6 (Scheme 1) involves
disconnection to the phosphonamidate fragment 1 and the
Scheme 1. Retrosynthetic Approach to 6
Figure 1. Phosphonamidate prodrugs.
Received: September 25, 2020
Published: May 5, 2021
transcriptase inhibitor (NRTI) approved for the treatment of
HIV and HBV infections, represents an example of successful
phosphonamidate prodrug implementation,2 achieving im-
proved properties compared to the phosphonic acid backbone
tenofovir (2) and the phosphonate diester prodrug tenofovir
disoproxyl fumarate (TDF, 3). Similarly, structure−activity
https://doi.org/10.1021/acs.oprd.0c00428
Org. Process Res. Dev. 2021, 25, 1247−1262
© 2021 American Chemical Society
1247

Organic Process Research & Development
pubs.acs.org/OPRD
Article
Scheme 2. First-Generation Route to Phosphonamidate 1
Scheme 3. Epimer Recycling Strategy
adenosine core 7. A preceding paper details the assembly of 6
from 1 and 7.4 Herein, we describe the evolution of the synthetic
process to phosphonamidate 1 from its inception, requiring
chromatographic separation of phosphorus diastereomers, to a
more efficient, second-generation variant relying on selective
crystallization.
needed for solvent exchange operations from step 2 [dichloro-
methane (DCM)/toluene to ethanol] to step 4 (ethanol to
DCM). Second, the BOMCl alkylation step (step 1) turned out
to be sensitive to the quality of the electrophile and often
required additional charges to reach full conversion. Third, the
intermediates involved in the displacement reaction (step 2)
were found to be highly susceptible to hydrolytic degradation,
thus requiring strictly anhydrous conditions to prevent the
formation of impurities.
Overall, these factors limited the scalability of the process and
revealed the need for alternative conditions to reduce cost and
lead time. Our research initially focused on developing a
recycling strategy to recover the unwanted isomer 13 from SMB
separation, hence improving the overall yield and reducing the
cost of the current process. Concurrently, in an effort to remove
SMB altogether to reduce the lead time, stereoselective synthesis
and selective crystallization were investigated as alternatives.
These endeavors culminated in the design of a more efficient
SMB-free second-generation route that overcomes the pre-
viously identified shortcomings.
The original synthetic route to 1 is shown in Scheme 2.5
Silylation of diphenylphosphite with N,O-bis(trimethylsilyl)-
trifluoroacetamide (BSTFA), followed by alkylation with
benzyloxymethyl chloride (BOMCl) delivered intermediate 8,
which was monohydrolyzed and isolated as the sodium salt 9 in a
fully telescoped sequence (77−88% overall yield). Conversion
of 9 to the phosphonochloridate 10 and displacement by L-
alanine ethyl ester (11) afforded a 45:55 mixture of phosphorus
epimers 12 and 13 favoring the undesired (RP)-isomer (61−
87% yield). The diastereomers were separated by simulated
moving bed (SMB) chromatography on chiral stationary phase
(CSP) to furnish the desired (SP)-diastereomer 12 (99.5:0.5 dr,
38−47% yield) as an ethanolic solution. Finally, hydrogenolysis
of the benzyl-protecting group and crystallization from methyl
tert-butyl ether (MTBE) delivered 1 (78−86% yield) as a single
diastereomer in >99% area-normalized (AN) purity.
This sequence was executed on manufacturing scale to
produce multiple batches of 1 in 22% overall yield (four steps).
However, as the demand for material increased, several issues
associated with the synthetic route became apparent. First,
reliance on SMB for the separation of diastereomers (step 3) was
identified as a major contributor to high cost, process mass
intensity (PMI),6 and long lead time in the production of 1. It
was estimated that SMB accounts for 36% of the production
time and 20% of the total PMI, considering all solvents required
for the chromatographic separation itself and solvents/water
2. RESULTS AND DISCUSSION
2.1. Recycling Strategy Applied to the Current
Process. The undesired (RP)-isomer 13 is obtained by SMB
separation of a nearly 1:1 mixture of phosphorus epimers as a
40−60 wt % ethanolic solution (raffinate, Scheme 3). In
previous manufacturing campaigns, this material was discarded,
thereby contributing to low overall yield and high waste burden.
To increase the efficiency of the synthetic route, the recycling of
13 via P-epimerization was examined.
1248
https://doi.org/10.1021/acs.oprd.0c00428
Org. Process Res. Dev. 2021, 25, 1247−1262

Organic Process Research & Development
pubs.acs.org/OPRD
Article
Table 1. Optimization of Epimerization Reaction Conditions
a
conditions
results (% AN)
entry
DBU (mol %)
PhOH (mol %)
SMB raffinate
solvent
drying agent
12
13
14 + PhOH
1
2
3
4
5
6
7
8
0.8
21.1
28.2
27.0
1.2
27.4
29.0
19.7
30.7
31.2
11.9
32.2
32.7
65.9
21.4
28.3
27.1
58.5
29.4
31.6
43.7
33.0
31.2
54.1
34.9
32.6
27.0
52.8
39.9
41.9
36.1
40.4
36.4
32.0
32.0
33.1
28.4
29.1
29.0
20
20
20
20
20
20
5
10
20
5
10
10
10
10
10
10
5
DCM
DCM
DCM
DCM
THF
toluene
DCM
DCM
DCM
DCM
DCM
DCM
4 Å MS powder
4 Å MS beads
BSA
4 Å MS powder
4 Å MS powder
4 Å MS powder
4 Å MS powder
4 Å MS powder
4 Å MS powder
4 Å MS powder
4 Å MS powder
9
5
5
10
11
12
13
10
20
a
Determined by UPLC−MS.
Previous studies with compound S1 had identified the
washed with water (3 × 2 V) to remove ethanol completely. The
resulting DCM solution was treated with 4 Å powder molecular
sieves (0.5 S) and subjected to the epimerization conditions.
The reaction provided 12/13 in 96% yield as a 47:53 mixture of
diastereomers after work-up.
combination of catalytic amounts of 1,8-diazabicyclo[5.4.0]-
undec-7-ene (DBU) and phenol in the presence of a drying
agent as suitable conditions for epimerization at phosphorus
(see Supporting Information). These conditions were used as
the starting point to further optimize the epimerization of 13.
The raffinate was received as a solution in ethanol from SMB
separation, with 13 accounting for 65.9% AN of the mixture
composition and in 98.8:1.2 dr (Table 1, entry 1). The raffinate
also contained a mixture of phenol and 147 [inseparable by
ultraperformance liquid chromatography (UPLC) analysis]
amounting to 27.0% AN. Given the incompatibility of 13 with
a mixture of alcohol and base (which would lead to the
displacement of phenol by the alcohol), solvent exchange from
ethanol to a suitable reaction solvent was necessary prior to the
initiation of the epimerization reaction. Screening of drying
agents and solvents (entries 2−7) identified DCM and 4 Å
powder molecular sieves as a suitable combination to reach a 1:1
mixture of isomers with little loss to hydrolysis (entry 3).
Omission of drying agents resulted in a higher loss to hydrolysis
(entry 2). After further optimization of these conditions, it was
determined that 20 mol % DBU allows to reach equilibrium as a
1:1 mixture within 20 h (entries 11−13). Additionally, the
amount of phenol has little impact on the reaction outcome
(entries 3, 10, and 13), as some phenol (typically 5−10 mol %) is
already present in the SMB raffinate.
Two limiting mechanisms can be envisioned for P-
epimerization of phosphonamidates (Figure 2):8 (i) a
Figure 2. Epimerization mechanisms.
dissociative, elimination−addition mechanism by deprotona-
tion at nitrogen and ejection of phenolate, proceeding through
metaphosphonimidate i, and (ii) an associative, SN2-type
mechanism by direct nucleophilic attack at phosphorus of
phenolate (formed from phenol and base) or nucleophilic base,
proceeding through the pentacoordinate intermediate or
transition state ii.9 Experimental evidence gathered over the
course of these studies suggests that the former hypothesis is
more plausible (Table 2). In addition to DBU, epimerization
also proceeds in the presence of guanidine bases (pKa BH+ > 16,
entries 1−3) but not with tertiary amines or other weaker bases
(pKa BH+ < 15, entries 5−7). The pKa value of the conjugate acid,
rather than its nucleophilicity, seems to be the determining
factor. Moreover, epimerization does not proceed with the N-
benzyl derivative 15 (entry 4), where deprotonation at nitrogen
is not possible. These observations support an elimination−
A 50 g scale experiment was carried out to assess the scalability
of the optimized conditions (Scheme 4). The ethanolic SMB
raffinate containing 13 was first diluted with DCM (2 V) and
Scheme 4. Optimized Reaction Conditions for P-
Epimerization
1249
https://doi.org/10.1021/acs.oprd.0c00428
Org. Process Res. Dev. 2021, 25, 1247−1262

Organic Process Research & Development
pubs.acs.org/OPRD
Article
Previously, the investigation of the stereoselective synthesis of
4 had revealed the possibility of establishing the desired
configuration at phosphorus via crystallization-induced asym-
metric transformation (CIAT)20 of a configurationally labile
phosphonochloridate intermediate. As 1 possesses the same
stereochemical arrangement at the phosphorus atom as 4, we
envisioned that, by installing a benzimidazole-derived chiral
auxiliary (16, Scheme 5) as a surrogate of the chiral
adeninopropanol chain in 4, a similar CIAT process could be
accessed during chlorination to provide diastereomerically
enriched 18. The benzimidazole moiety was chosen to
potentially impart crystallinity to the phosphonochloridate
intermediate 17. Moreover, the presence of a phenyl group
would aid the cleavage of the chiral auxiliary by hydrogenolysis
in the final step.
Table 2. Impact of Base Selection on the Epimerization
Reaction
Substrate 16 was synthesized in six steps from (R)-
phenyloxirane and benzimidazole (see Supporting Informa-
tion). The intended CIAT mechanism was investigated by
reacting 16 with thionyl chloride in various solvents [MeCN,
iPrOAc, MTBE, cyclopentyl methyl ether (CPME), 2-
methyltetrahydrofuran (MeTHF), toluene, trifluorotoluene,
and chlorobenzene] but proved unsuccessful. In most cases,
the phosphonochloridate 17 did not crystallize from the reaction
mixture or, when it did [in tetrahydrofuran (THF) and
MeTHF], no significant diastereomeric enrichment was
observed upon displacement with alanine.
Despite the unsuccessful attempts at obtaining enrichment via
CIAT, we envisioned that the benzimidazole-derived chiral
auxiliary could be leveraged in a DYKAT mechanism. Based on a
precedent of phosphonamidate synthesis via DYKAT,17a
reaction conditions were developed to access material enriched
in up to 65:35 dr from phosphonic dichloride 20 by sequential
treatment with 11 and phenol (Scheme 6). However, cleavage of
the chiral auxiliary by hydrogenation revealed that the selectivity
was in favor of the undesired (RP)-diastereomer 19. To exploit
such stereochemical preference to our advantage, an alternative
strategy was devised. Dichloride 20 was treated sequentially with
11 and an electron-poor phenol that could act as a leaving group
(Scheme 7). As expected, mixtures enriched in (RP)-
diastereomer were obtained. Among the activated phenols that
we tested (nitrophenols and pentafluorophenol), 3-nitrophenol
gave the highest dr (65:35). The resulting phosphonamidate 21
was then treated with sodium phenoxide in DCM, affording the
desired (SP)-diastereomer 18 in 79:21 dr.
a
b
c
d
Reference 10a. Reference 10b. Reference 10c. pKa values for
e
isothioureas are in the 7.3−10.3 range (ref 11). Substrate 15 was
used.
association mechanism through metaphosphonimidate i, with
the consequent loss of stereochemical information.
2.2. Stereoselective Synthesis. The development of
stereoselective methods to access phosphoric/phosphonic acid
derivatives as single isomers has been largely stimulated by the
well-established influence of the phosphorus configuration on
biological activity.3a,12,13 The existing stereoselective methods
rely on a chiral auxiliary,14 enzymatic kinetic resolution,15
dynamic kinetic asymmetric transformation (DYKAT)16 of a
configurationally labile intermediate,13,17 organocatalysis,18 or
stereospecific displacement on a stereodefined substrate.19 In an
effort to avoid physical methods of separation (either
chromatography or crystallization), we investigated the
possibility of synthesizing 1 stereoselectively by means of chiral
auxiliary and/or DYKAT strategies.
With a diastereoselective route to the desired (SP)-
diastereomer in hand, our attention turned to the final
Scheme 5. Use of Benzimidazole-Derived Chiral Auxiliary in a CIAT Process
1250
https://doi.org/10.1021/acs.oprd.0c00428
Org. Process Res. Dev. 2021, 25, 1247−1262

Organic Process Research & Development
pubs.acs.org/OPRD
Article
Scheme 6. Use of Benzimidazole-Derived Chiral Auxiliary in a DYKAT Process
phosphonic dichloride 22 and 11 provided the corresponding
phosphonamidates in modest diastereomeric enrichment in
favor of the undesired (RP)-isomer (40:60 dr, entry 1).
However, switching to N-benzyl Ala-OEt (24) as the
nucleophile improved the dr to 74:26, favoring the (SP)-isomer
(entry 2). Interestingly, the diastereoselectivity is affected by the
choice of the base for the individual additions. The highest (SP)-
selectivity was obtained with a mixture of N-methylimidozole
(NMI) and triethylamine (TEA). When NMI alone was used in
step 1, selectivity reversed to favor the (RP)-isomer (34:66 dr;
not listed). A more pronounced effect was also observed with 4-
(dimethylamino) pyridine (DMAP) [20:80 dr, favoring (RP),
entry 3]. It is possible that nucleophilic, unhindered bases
change the mechanism of substitution such that nucleophilic
addition of the base to the phosphonamidyl chloride occurs
prior to the addition of phenol. Unfortunately, these conditions
were ineffective with the displaceable 4-nitrophenol nucleophile,
providing little dr enhancement (41:59 dr, entry 4). In an effort
to improve diastereoselectivity and/or obtain crystalline
products whose dr could be further enhanced by crystallization,
additional substitution patterns were investigated. Gratifyingly,
it was found that the TsO-phosphonic dichloride 23, combined
with 24, yields excellent levels of selectivity (96:4 dr), yet
favoring the undesired (RP)-diastereomer (entries 6−7). In
contrast, reaction of 23 with unprotected alanine 11 resulted in
lower selectivity (59:41 dr, entry 5). As previously accomplished
Scheme 7. Use of Benzimidazole-Derived Chiral Auxiliary in
a DYKAT Process with 3-Nitrophenol
deprotection step. Contrary to our expectations, removal of the
benzylic chiral auxiliary by hydrogenolysis was problematic. Full
conversion of 18 to 1 could only be achieved with a high loading
of a mixture of Pd/C and Pd(OH)2/C (76 mol % each), under 4
bar of H2 at 60 °C over 6 days. Due to these challenges, chiral
auxiliary routes were not further investigated.
Encouraged by the diastereoselectivity trends observed with
the chiral auxiliary-bearing substrate 20, we further evaluated
DYKAT approaches using a variety of O- and/or N-substituted
derivatives (Table 3). Initial attempts using O-benzyl
Table 3. DYKAT Sequence Using N- and O-Derivatives
b
entry
R1
R2
step 1 conditions
Ar
step 2 conditions
(SP)/(RP) dr
1
2
3
4
5
6
7
Bn
H
NMI (2.2 equiv) MeCN, −5 to 22 °C
Ph
Ph
Ph
NMI (1.1 equiv) MeCN, −5 to 22 °C
NMI (1.1 equiv) DCM, −5 to 22 °C
NMI (1.1 equiv) DCM, 0 to 22 °C
NMI (1.15 equiv) DCM, 0 to 22 °C
NMI (1.0 equiv) DCM, 0 to 22 °C
DCM, 0 to 22 °C
40:60
74:26
20:80
41:59
59:41
4:96
a
Bn
Bn
Bn
H
Bn
Bn
TEA (3.0 equiv) NMI (1.0 equiv) DCM, −5 °C
DMAP (3.0 equiv) NMI (1.0 equiv) DCM, 0 to 22 °C
DMAP (3.0 equiv) NMI (1.0 equiv) DCM, 0 to 22 °C
NMI (3.5 equiv) DCM/MeCN, 0 °C
NMI (3.5 equiv) DCM/MeCN, 0 °C
NMI (3.5 equiv) DCM/MeCN, 0 °C
a
a
4-(NO2)-C6H4
Ph
Ph
Ts
3-(NO2)-C6H4
DCM, 0 to 22 °C
4:96
a
b
The free base of the alanine derivative was used. Determined by UPLC or 31P NMR.
1251
https://doi.org/10.1021/acs.oprd.0c00428
Org. Process Res. Dev. 2021, 25, 1247−1262

Organic Process Research & Development
pubs.acs.org/OPRD
Article
with substrate 20, the high (RP)-selectivity could be turned to
our advantage by carrying out the displacement with 3-
nitrophenol (yielding 25), followed by treatment with sodium
phenolate (Scheme 8). Through this sequence, the desired (SP)-
isomer 26 was obtained in 85:15 dr.
nucleophiles,22 typically requiring hydride bases23 or Ag2O24
to undergo alkylation with alkyl halides. However, such
conditions resulted in the rapid decomposition of 1, and
alternative benzylation conditions (benzylic bromide and
organic/inorganic base) were equally unsuccessful. Nonethe-
less, acid-catalyzed benzylation using trichloroacetimidates
allowed for the preparation of several analogues (see Supporting
Information, Table S3), four of which were found to be
crystalline solids and chosen for further development (Table 4,
Scheme 8. DYKAT Sequence Using Phosphonic Dichloride
23 and 3-Nitrophenol
Table 4. Crystalline Analogues of 12 and 13
Unfortunately, challenges were encountered in the conversion
of 26 to the unprotected phosphonamidate 1 by the removal of
the protecting groups. Tosyl cleavage (CeCl3·H2O/NaI) or
conversion to the O-benzyl derivative (TBAI/BnOH) could not
be accomplished without decomposition of the relatively
unstable phosphonamidate moiety. The N-benzyl group also
proved difficult to be removed by hydrogenolysis. With the
phosphonamidate derivative 15 (obtained through the con-
ditions in Table 3, entry 2), complete hydrogenolytic cleavage of
the N-benzyl group could only be achieved with a high loading of
Pd/C and Pd(OH)2/C (40 mol % each), under 4 bar of H2 at 60
°C over 2 days. Given the unsuccessful conversion of 26 to 1 and
the poor step/atom economy for other derivatives under
analysis, we chose to deprioritize research into diastereoselective
approaches.
2.3. Selective Crystallization. As the investigation of
stereoselective phosphonamidate synthesis was unable to deliver
a scalable, satisfactory alternative to SMB, our attention shifted
to the selective crystallization of the desired diastereomer. Two
opportunities exist in the current synthetic route for selective
crystallization: (i) crystallization of the benzylated intermediate
12 (or an alternative, O-protected precursor), and (ii)
crystallization of 1. The former option is preferable due to the
beneficial impact on PMI of positioning a low-yielding step
earlier in the synthesis.21
2.3.1. Synthesis and Selective Crystallization of Benzylated
Analogues. As 12 is a viscous liquid, research efforts focused on
the identification of crystalline, benzyl-substituted analogues of
12 to be exploited in a selective crystallization process, or, more
ambitiously, in a CIAT process. In order to be a practical
alternative to SMB, the intended analogue must satisfy the
following requirements: (i) be a crystalline, high-melting solid;
(ii) undergo selective crystallization (or CIAT) to provide
materials with high diastereomeric enrichment; (iii) contain a
functionalized benzylic protecting group easily removable by
hydrogenation; and (iv) generate a hydrogenation byproduct
that is easily removed from 1.
Benzylation of 1 with benzylic halides was identified as a quick
strategy to provide access to a variety of analogues of 12 to be
evaluated for their crystallinity and capacity of crystallizing
selectively from the corresponding (RP)-epimers. Hydroxy-
methyl phosphonates and related compounds are poor
entries 1−4). Three corresponding undesired phosphorus
epimers (entries 5−7) were also synthesized to compare their
crystalline properties and initiate studies of crystallization from
mixtures of diastereomers.
2.3.1.1. 9-Fluorenyl and 4-Phenylbenzyl Analogues. In
preparation for crystallization studies, we sought to determine
the thermodynamic preference for either diastereomer under
equilibrating conditions. Epimerization of 27a with DBU/
phenol was examined in several solvents (Table 5). A consistent
ca. 50:50 dr was obtained after 4 days (entries 1−5).
Epimerization of 28a in DCM led to the same 50:50 dr (entry
6), thus confirming the lack of a thermodynamic preference for
either diastereomer.
Subsequently, crystallization experiments were carried out to
determine the possibility of enriching 27a from a mixture of P-
epimers (Table 6). Mixtures of 27a and 28a with a starting 74:26
dr were dissolved in a mixture of solvent/antisolvent and
subjected to cooling crystallization (5 °C). Analysis of the
crystallized solids showed improvement of the dr from 74:26 up
to 98:2 (entries 1−4). However, when crystallization was
attempted starting from a 51:49 mixture, no enrichment was
observed (entries 5−8). The recovered solids were obtained as
nearly 50:50 mixtures of diastereomers and were identified by
XRPD as a new solid form (mp = 88−90 °C). Extensive solvent
screening conducted on the new polymorph did not deliver any
enrichment. These observations suggest the existence of the two
diastereomers as a quasi-racemate20b that prevents selective
crystallization of 27a from a 1:1 mixture.
1252
https://doi.org/10.1021/acs.oprd.0c00428
Org. Process Res. Dev. 2021, 25, 1247−1262

Organic Process Research & Development
pubs.acs.org/OPRD
Article
Table 5. Epimerization of 27a
a
entry
substrate
solvent
dr (27a/28a)
1
2
3
4
5
6
27a
DCM
THF
toluene
EtOAc
CH3CN
DCM
50:50
47:53
49:51
48:52
48:52
50:50
28a
a
Determined by UPLC.
Table 6. Selective Crystallization of 27a
Table 8. Selective Crystallization of 27c
a
a
a
a
entry
solvent
initial dr (27a/28a)
final dr (27a/28a)
entry
solvent
initial dr (27c/28c)
final dr (27c/28c)
1
2
3
4
5
6
7
8
THF/heptane 1:1
EtOAc/heptane 3:1
toluene/heptane 4:1
iPrOH/heptane 9:1
THF/heptane 2:3
toluene/heptane 4:1
MeTHF/heptane 1:1
MTBE/heptane 4:1
74:26
98:2
95:5
98:2
98:2
55:45
47:53
49:51
53:47
1
2
3
4
5
6
EtOAc/heptane 2:1
iPrOH
THF/heptane 1:1
EtOAc/heptane 3:2
iPrOH/heptane 7:1
iPrOAc/heptane 7:2
80:20
99:1
99:1
99:1
94:6
91:9
92:8
50:50
51:49
a
Determined by 31P NMR.
a
Determined by UPLC.
the crystallized solids showed that dr had improved to 99:1 when
starting from an 80:20 mixture (entries 1−3) and to ≥91:9 when
starting from a 50:50 mixture (entries 4−6).
Selective crystallization studies on the 4-phenylbenzyl
analogue 27b were executed in a similar fashion to 27a and
led to the same conclusions. Subjection of 27a to standard
epimerization conditions resulted in 50:50 dr, indicating the
absence of a thermodynamic preference for either diaster-
eoisomer. Moreover, selective crystallization trials from the dr =
49:51 input material resulted in no enrichment and detection of
a new crystalline form by XRPD (mp = 74 °C).
2.3.1.2. 9-Anthracenylmethyl Analogue. As with 27a and
27b, the 9-anthracenylmethyl analogue 27c was first subjected
to epimerization conditions (Table 7). While the reaction in
DCM returned the usual ca. 50:50 dr (entry 1), other solvents
resulted in diastereomeric mixtures slightly favoring 27c (entries
2−4).
The significant dr enrichment obtained during crystallization,
along with the slight thermodynamic preference for the desired
isomer 27c in solution under equilibrating conditions, led us to
believe that a CIAT process could be developed (Table 9).
Thus, 51:49 27c/28c mixtures were subjected to standard
epimerization conditions in several solvents in which 27c has
low solubility (<20 mg/g).25 The suspensions were allowed to
equilibrate for 1−4 days at the indicated temperature, the
mixtures were then filtered, and the solids and mother liquors
were analyzed for dr and % recovery. The high diastereomeric
enrichment achieved (≥95:5, entries 1, 2, 5) and the >50%
recovery suggest that a CIAT mechanism is indeed operative.
Moreover, when the conditions of entry 6 were reproduced and
the solvent was allowed to slowly evaporate over 3 days, the
initial suspension fully converted to a crystalline solid of 95:5 dr.
Unfortunately, high-yielding, nonevaporative reaction condi-
tions could not be developed without experiencing a dr erosion
(entries 4, 6, 7).
Selective crystallization studies were next carried out,
beginning from 80:20 and 50:50 mixtures of diastereomers
(obtained by chromatographic purification of epimerized
reaction mixtures). The solids were dissolved in a mixture of
solvent/antisolvent and cooled to −3 °C (Table 8). Analysis of
Table 7. Epimerization of 27c
a
entry
solvent
dr (27c/28c)
1
2
3
4
DCM
EtOAc
iPrOAc
51:49
54:46
58:42
iPrOAc/MTBE 7:3
68:32
a
Determined by 31P NMR.
1253
https://doi.org/10.1021/acs.oprd.0c00428
Org. Process Res. Dev. 2021, 25, 1247−1262

Organic Process Research & Development
pubs.acs.org/OPRD
Article
Table 9. Optimization of CIAT Conditions
a
b
b
entry
solvent (amount)
iPrOAc (10 V)
iPrOAc (5 V)
iPrOAc (7 V)
CPME (7 V)
MTBE (15 V)
T (°C)
solid recovery (%)
solid dr (27c/28c)
filtrate dr (27c/28c)
1
2
4
5
6
7
21
21
−3
21
21
21
46
57
62
50
64
65
97:3
95:5
92:8
95:5
88:12
93:7
54:45
75:25
39:61
52:48
25:75
76:24
iPrOAc (4 V), MTBE(2 V)
a
b
Based on the total 27c/28c input. Determined by 31P NMR.
Scheme 9. Hydrogenolysis of 27c
Concurrently, it was determined that the 9-anthracenylmethyl
group cannot be removed easily under hydrogenolytic
conditions due to the competing reduction of the 9,10-ring,
resulting in >40% of compound 29 under most of the catalytic
conditions screened (Scheme 9; see Supporting Information for
additional details).
2.3.1.3. Benzyloxycarbonyl Analogue. Investigation of the
selective crystallization of 27d revealed that, unlike analogues
27a and 27b, a 1:1 mixture of diastereomers does not give rise to
a new solid form, as determined by PXRD analysis. Moreover,
solubility studies showed a striking difference between the two
diastereomers in several solvents (Figure 3), with 27d being
typically more soluble than 28d.
Table 10. Enrichment of 27d by Selective Crystallization of
28d
initial dr
final dr
a
ab
,
entry
solvent
toluene
toluene
MTBE/toluene
2:1
T (°C)
(27d/28d)
(27d/28d)
1
2
3
−8
−20
−10
62:38
59:41
59:41
75:25
78:22
82:18
4
5
MTBE/toluene
2:1
MTBE/toluene
1:1
−29
55:45
53:47
86:14
84:16
−26
a
b
Determined by 31P NMR. In the mother liquor.
contrast with that of 27a and 27b, where no dr improvement
upon crystallization was observed from ca. 50:50 mixtures.
However, further experiments with toluene/MTBE solvent
mixtures failed to provide crystallization conditions that could
improve the dr above 85:15. Crystallization at this stage was
therefore disregarded as potentially inefficient.
2.3.2. Selective Crystallization of 1. 1 and its (RP)-epimer
(30) are both solid compounds. While 1 exhibits high
crystallinity and a sharp melting point at 73−74 °C, 30 is a
waxy solid melting at 38−40 °C. To develop a selective
crystallization of 1, its solubility was measured in various
solvents at 22 °C (Figure 4). 1 displays low solubility (<40 mg/
g) in aromatic hydrocarbons (toluene, p-xylene, and ethyl
benzene), butyl esters (butyl acetate, isobutyl acetate, and tert-
butyl acetate), and ethers (MTBE, CPME, tert-amyl methyl
ether, diisopropyl ether, and dibutyl ether). A complementary
solubility study using 30 was also performed with solvents in
which 1 is poorly soluble. The solubility difference was found to
be oftentimes remarkable, with 30 being approximately 20 times
more soluble than 1 in ethers and 50 times more soluble in
aromatic hydrocarbons.
Figure 3. Solubility of 27d and 28d.
Selective crystallization experiments were executed to
leverage the solubility difference and obtain diastereomerically
enriched 27d in the mother liquor. Mixtures of 27d/28d with
varying dr (53:47 to 62:38, favoring 27d) were subjected to
cooling crystallization in toluene and MTBE/toluene mixtures
(Table 10). The dr value in the mother liquors was found to have
improved up to 86:14 (entry 4). This behavior is in stark
With these data in hand, we felt confident that a cooling
crystallization would allow us to isolate 1 in high yield and dr
from a mixture of diastereomers. Among the solvents that
1254
https://doi.org/10.1021/acs.oprd.0c00428
Org. Process Res. Dev. 2021, 25, 1247−1262

Organic Process Research & Development
pubs.acs.org/OPRD
Article
Figure 4. Solubility of 1 and 30 at 22 °C.
displayed high solubility difference, toluene, CPME, and MTBE
were chosen for further investigation based on commercial
availability, cost, and toxicity considerations.
reached 5 °C. After aging at −23 °C for 17 h, 1 was isolated by
filtration in 32% yield and 96:4 dr.
2.4. Second-Generation Synthesis. With reliable selective
crystallization conditions in hand, a fully updated sequence to 1
was developed (Scheme 11). It was envisioned that moving to
Selective crystallization studies were performed by dissolving
a 50:50 mixture of 1 and 30 in DCM and then exchanging the
solvent to MTBE, CPME, or toluene (15 V). The slurries were
cooled to −24 °C over 2 h and then aged for an additional 2 h. 1
was isolated by filtration and dried, and its purity/dr was
determined by 31P NMR (Table 11). Among the solvents tested,
MTBE proved to be superior as it afforded 1 as a single
diastereomer in high purity (entry 3).
Scheme 11. Proposed Second-Generation Synthesis
Table 11. Selective Crystallization of 1
yield
a
purity
(% AN)
solid dr
filtrate dr
b
b
b
entry solvent
(%)
(1/30)
(1/30)
1
2
3
toluene
CPME
MTBE
70
64
68
94.0
95.2
100
94:6
97:3
100:0
76:24
77:23
83:17
a
b
Based on the maximum theoretical 50% yield. Determined by 31P
NMR.
The MTBE selective crystallization conditions were then
the benzyloxycarbonyl (Cbz) protecting group would eliminate
the problematic use of BOMCl, while still accommodating the
hydrogenolytic protecting group removal at the end of the
synthesis. In addition, the Cbz group should be much easier to
re-install than the benzyl group on 3026 to support a recycling
strategy via the DBU/PhOH conditions.27
The synthesis begins with a five-step, telescoped sequence
starting from diphenyl phosphite (Scheme 12). Notably,
commercial sources of diphenyl phosphite contain variable
amounts of phenol (4−20 mol %) and triphenyl phosphite (2−4
mol %), so any reaction conditions must be able to
accommodate these impurities. Silylation of diphenyl phosphite
using BSTFA, followed by Abramov reaction with paraformal-
dehyde at 90 °C provided the silyl ether 32 (Table 12, entry 1).
Whereas silyl phosphites react readily with simple aliphatic
aldehydes at room temperature,28 reactions of phosphonates
(other than methyl) with paraformaldehyde typically require
applied to a crude hydrogenolysis reaction mixture (Scheme
10). After hydrogenation of a mixture of 12 and 13 (44:56 dr),
the DCM organic phase was filtered through a pad of Celite to
remove the spent catalyst, followed by a water wash and solvent
exchange to MTBE (5 V). The solution was cooled to −23 °C
over 5 h and seeded with 1 (0.015 S) when the temperature
Scheme 10. Selective Crystallization of 1 from the
Hydrogenolysis Reaction Mixture
1255
https://doi.org/10.1021/acs.oprd.0c00428
Org. Process Res. Dev. 2021, 25, 1247−1262

Organic Process Research & Development
pubs.acs.org/OPRD
Article
Scheme 12. Telescoped Abramov/Cbz Protection/Hydrolysis Sequence
Table 12. Evaluation of Conditions for Silylation and Abramov Reaction of Diphenyl Phosphite
a
a
entry
silylation conditions
BSTFA (1.0 equiv) neat, 35 °C
BSTFA (1.0 equiv) neat, 35 °C
BSA (1.0 equiv) neat, 50 °C
TMSCl (1.0 equiv) Et3N (1.0 equiv) toluene (8 V), rt
Abramov conditions
conversion to 31 (%)
conversion to 32 (%)
1
2
3
4
(CH2O)n (1.5 equiv) toluene (2 V), 90 °C
trioxane (0.5 equiv) toluene (2 V), 90 °C
(CH2O)n (1.5 equiv) toluene (2 V), 90 °C
(CH2O)n (1.0 equiv), 100 °C
99.7
>99.9
93.5
89.6
>99.9
0
26.4
3.8
a
Determined by 31P NMR.
more forcing conditions.29 Trioxane was investigated as an
alternative source of formaldehyde, but it failed completely to
react in the Abramov reaction (entry 2). In addition, while more
common silylating reagents such as N,O-bis(trimethylsilyl)-
acetamide (BSA) or TMSCl are effective for the formation of 31,
the Abramov reaction proceeds poorly when these reagents are
used (entries 3 and 4).
Desilylation to the hydroxymethylphosphonate 33 was
initially accomplished by methanol treatment and removal of
the volatile methoxytrimethylsilane under vacuum. However,
hydrolysis to release phenol occurred to a varying extent and led
to reduced yields. More reproducible conditions were later
identified under biphasic conditions, where the presence of
potassium chloride helped to minimize hydrolysis.
Attention then turned to Cbz protection. Acylation
conditions involving stoichiometric triethylamine and catalytic
DMAP led to a complex reaction mixture (Table 13, entry 1).
Subsequent stability studies indicated that 33 readily decom-
poses when treated with alkyl amines and is only moderately
stable to DMAP. Instead, pyridine was identified as a suitable
base to carry out the protection reaction. However, reaction
screening with varying amounts of pyridine and CbzCl revealed
that the reaction tends to stall with a maximum of ca. 60%
conversion (entries 2 and 3). NMR analysis showed that this is
due to in situ decomposition of CbzCl to benzyl chloride, which
can be suppressed by moving to a more nonpolar solvent and
utilizing slow addition of CbzCl (entries 4 and 5). Alternatively,
the Cbz protection could be completed under Schotten−
Baumann conditions using the more benign potassium
bicarbonate as the base (entry 6). Indeed, these conditions
were found to streamline nicely with the aqueous desilylation
Table 13. Evaluation of Conditions for Cbz Protection of 33
a
entry
1
conditions
34 (% AN)
CbzCl (1.5 equiv), DMAP (0.1 equiv), Et3N
(1.5 equiv), DCM (10 V), 0 °C to rt
0 (decomp.)
2
3
4
5
6
7
CbzCl (1.0 equiv), pyridine (1.0 equiv), DCM (10 V), 58.8
rt
CbzCl (1.2 equiv), pyridine (2.0 equiv), DCM (10 V), 54.1
rt
pyridine (2.0 equiv), toluene (15 V), 0 °C, CbzCl
87.5
96.0
97.3
(1.1 equiv, charged over 20 min)
pyridine (2.0 equiv), toluene (15 V), 0 °C, CbzCl
(1.1 equiv, charged over 60 min)
CbzCl (1.0 equiv), DMAP (2 mol %), 10 wt % aq.
KHCO3 (1.3 equiv), toluene (5 V), rt
CbzCl (1.3 equiv), DMAP (4 mol %), toluene (5 V), 97.7 (91%
rt, 20 wt % aq. KHCO3, (1.5 equiv, charged over
40 min)
assay yield)
a
Determined by 31P NMR.
conditions and could be completed from the same biphasic
mixture. The best results were obtained with slow addition of
aqueous potassium bicarbonate, which helps to minimize the
hydrolytic degradation of 33 (entry 7).
Finally, the solution of 34 is solvent-exchanged into 2-
butanone and subjected to monohydrolysis conditions using
aqueous sodium hydroxide. A substoichiometric amount of
hydroxide (0.85 equiv) was found to be optimal to limit bis-
hydrolysis. After reaction completion, azeotropically distilling
the mixture at a constant volume with replacement of 2-
1256
https://doi.org/10.1021/acs.oprd.0c00428
Org. Process Res. Dev. 2021, 25, 1247−1262

Organic Process Research & Development
pubs.acs.org/OPRD
Article
butanone until the water content is below 1.5 wt % allows the
sodium salt 35 to crystallize directly from the reaction mixture.
The process is capable of purging phenol and other impurities
that were generated throughout the processing and typically
provides 35 in 60−70% yield (over 5 steps) and >95% purity.
The phosphonamidate moiety is installed subsequently from
35 through a two-step telescoped process (Scheme 13). In the
displacement providing 45:55 dr in favor of the undesired (RP)-
diastereomer 13. The explanation for the switch in diaster-
eoselectivity based on the protecting group is not clear at this
time.
In the final step, the DCM solution of 27d/28d is subjected to
hydrogenolysis using 10% Pd/C under 3.5 bar of hydrogen. In
comparison to the benzyl group deprotection in the previous
route, the Cbz deprotection requires half the catalyst loading
(0.05 vs 0.1 S, or 2 vs 4 mol %), proceeds in a much shorter
reaction time (6 h vs ca. 3 days), and affords a cleaner reaction
profile. The selective crystallization process discussed in Section
2.3.2 was successfully applied for the isolation of 1 in 40−45%
yield and >98:2 dr.
Scheme 13. Telescoped Chlorination/Displacement
Sequence and Final Hydrogenolysis Step
To demonstrate the scalability of the route, the sequence was
performed at 3 kg diphenylphosphite input scale (Scheme 14).
The initial five-step telescoped sequence resulted in 3.06 kg of
35 (78% yield corrected for the purity of diphenyl phosphite,
95.7% AN purity). 500 g of 35 was carried through the
chlorination/displacement sequence, obtaining the phosphona-
midate intermediates 27d/28d as a DCM solution in 85% yield
(56:44 dr, 98.2% AN purity). Finally, hydrogenolysis and
selective crystallization afforded 164 g of 1 as a white crystalline
solid in 46% yield and 99.4:0.6 dr (99.5% AN purity).
3. CONCLUSIONS
first step, Vilsmeier conditions are used to access the
corresponding phosphonochloridate 36, which is treated in
the second step with 11 and DIPEA to obtain a mixture of
phosphonamidate diastereomers 27d/28d. In the chlorination
step, a slight excess of oxalyl chloride is required to achieve full
conversion and avoid the unproductive formation of phosphonyl
anhydrides 37. However, the excess oxalyl chloride is
detrimental in the subsequent displacement step with 11 as it
leads to the formation of a bis-oxalyl alanine impurity 38. To
overcome this challenge, a distillation step was introduced after
chlorination to remove excess oxalyl chloride. As expected, the
purity profile and yield of the telescoped sequence displayed
variability depending on the water content of 11 and DIPEA. To
increase the robustness of the process, it became necessary to
predry solutions of 11 and DIPEA in DCM with TMSCl prior to
their addition to a DCM solution of 36. With the optimized
chlorination/displacement conditions, the diastereomeric mix-
ture of 27d/28d is obtained as a DCM solution in 85−97% assay
yield over two steps. Interestingly, the coupling reaction reliably
proceeds with ≥55:45 dr in favor of the desired (SP)-
diastereomer 27d. In contrast, in the previous manufacturing
route, the O-Bn-protected-phosphonochloridate 10 undergoes
In summary, stereoselective synthesis, selective crystallization,
and epimer recycling were extensively evaluated to access 1 in an
expeditious and cost-effective manner. Ultimately, the switch
from SMB to selective crystallization and the introduction of the
Cbz protecting group enabled access to an improved synthetic
route (Table 14). Although both first- and second-generation
routes entail a sequence of eight steps with two solid-state
isolations, the second-generation route proceeds with a higher
overall yield (36% increase) and lower PMI (16% reduction).
Contrary to our expectations, analysis of PMI metrics revealed
that solvent usage remains unaltered in the second-generation
route (approximately 200 kg of solvents per kg of 1, Figure 5).
The solvent-saving opportunity offered by the removal of SMB
from the sequence appears to be offset by the increased solvent
demand of step 1. However, the usage of raw materials and water
is significantly reduced on a kg/kg basis due to the combined
effect of several factors: high yields for all steps, favorable dr in
step 2, and lack of solvent exchange operations involving water.
Thanks to these improvements, cost and lead time of 1 are
estimated to decrease by 50% on a manufacturing scale.
Scheme 14. Second-Generation Route to Phosphonamidate 1
1257
https://doi.org/10.1021/acs.oprd.0c00428
Org. Process Res. Dev. 2021, 25, 1247−1262

Organic Process Research & Development
pubs.acs.org/OPRD
Article
Table 14. Comparison of First- and Second-Generation Routes
attribute
first-generation route
8 (4 manufacturing steps, 2 isolations)
SMB
22
second-generation route
synthetic steps
diastereomer resolution
overall yield
8 (3 manufacturing steps, 2 isolations)
selective crystallization
30
total
328
275
reactants/reagents
solvents
water
45
198
85
16
201
58
PMI
lead time
5 months
2.5 months
Figure 5. Comparison of PMI breakdown for first- and second-generation routes.
charged with diphenyl phosphite (3.02 kg, 12.9 mol, 1.0 equiv)
and warmed to an internal temperature of 35 °C. BSTFA (3.29
kg, 12.8 mol, 1.0 equiv) was added dropwise over 45 min while
maintaining the temperature of the batch at 35 °C, and the
mixture was aged for 30 min and then sampled for conversion to
31. 31P NMR (162 MHz, CDCl3): δ 124.2 ppm (92.9% AN).
Paraformaldehyde (0.57 kg, 18.9 mol, 1.5 equiv) and toluene
(6.5 L, 2.2 V) were added, and the batch was ramped linearly to
90 °C over a period of 2 h. The mixture was aged for an
additional 2 h and then sampled for conversion to 32. 31P NMR
(162 MHz, CDCl3): δ 16.3 ppm (93.8% AN). The batch was
cooled to 20 °C and diluted with aq. potassium chloride (20 wt
%, 9.0 L, 3.0 V) and toluene (10.6 L, 3.5 V). After stirring for 15
h, the organic layer was sampled for conversion to 33. 31P NMR
(162 MHz, CDCl3): δ 18.0 ppm (93.6% AN). 4-Dimethylami-
nopyridine (60.5 g, 0.03 equiv) and benzyl chloroformate (3.09
kg, 18.1 mol, 1.4 equiv) were added, followed by a slow addition
of aq. potassium bicarbonate over 30−60 min (20 wt %, 10.2 kg,
1.6 equiv). The mixture was aged for an additional 60 min and
then sampled for conversion to 34. 31P NMR (162 MHz,
CDCl3): δ 11.1 ppm (96.6% AN). The phases were allowed to
settle, the aqueous layer was separated, and the organic layer was
washed with aq. sodium chloride (5 wt %, 6.0 L, 2.0 V). The
resulting organic layer was concentrated in vacuo with a
maximum jacket temperature of 50 °C to a final volume of ca.
6 L (2.0 V); then, 2-butanone (12.0 L, 4.0 V) was added, and the
batch was re-distilled to ca. 6 L. 2-Butanone (18.0 L, 6.0 V) was
added, and the mixture was sampled to ensure that the toluene
content was below 10 wt % (relative to 2-butanone) and then
cooled to an internal temperature of 0 °C. Aq. sodium hydroxide
(20 wt %, 2.09 kg, 0.8 equiv) was charged dropwise over ca. 60
min, and the mixture was stirred for 30 min and then sampled for
conversion to 35. 31P NMR (162 MHz, CDCl3): δ 0.9 ppm
(95.7% AN). Aq. HCl (18.5 wt %) was charged until the solution
pH was ca. 7 (required 0.16 kg, 0.06 equiv), and the batch was
warmed to an internal temperature of 22 °C. 2-Butanone (39.2
L, 13.0 V) was charged followed by 35 seed crystals (15.0 g,
4. EXPERIMENTAL SECTION
4.1. General Experimental. All substrates and reagents
were commercially available and used as received. Diphenyl
phosphite was obtained from Millipore-Sigma with approx-
imately 86% w/w purity, as determined by 1H NMR using 1,3,5-
trimethoxybenzene as an internal reference. NMR spectroscopy
was performed on a Bruker Ascend 400 (400 MHz) system,
using CDCl3, DMSO-d6, or D2O as the solvent. For NMR
analysis, 200 μL of the reaction solution or 20 mg solid was
added to 0.75 mL of the deuterated solvent. UPLC analysis was
performed using a Waters Acquity UPLC system as follows: (i)
Chiral purity of 27d/28d: Waters Acquity BEH Phenyl column
(1.7 μm, 2.1 × 100 mm) with 0.4 mL/min flow rate and 30 °C
column temperature. The analytical method utilized two mobile
phases: mobile phase A consisted of 0.1% (v/v) formic acid in
H2O and mobile phase B consisted of MeOH. The gradient
elution method was completed as follows: 1.0 min 35% of
mobile phase B; 6.0 min 55% mobile phase B; 16.0 min 75%
mobile phase B; 18.0 min 90% mobile phase B; 20.0 min 90%
mobile phase B; and 20.1 min 35% mobile phase B. The eluting
material was detected using a UV detector set at 207 nm. (ii)
Purity of 1: Waters Acquity UPLC HSS T3 column (1.8 μm, 2.1
× 100 mm) with 0.5 mL/min flow rate and 25 °C column
temperature. The analytical method utilized two mobile phases:
mobile phase A consisted of 0.1% (v/v) formic acid in H2O and
mobile phase B consisted of MeCN. The gradient elution
method was completed as follows: 2.0 min 3% of mobile phase
B; 15.0 min 90% mobile phase B; 17.0 min 90% mobile phase B;
17.1 min 3% mobile phase B; and 22.0 min 3% mobile phase B.
The eluting material was detected using a UV detector set at 260
nm. (iii) Chiral purity of 1: CHIRALPAK IC-3 column (3 μm,
4.6 × 150 mm) with 0.4 mL/min flow rate of isocratic elution of
7:3 TFA (0.1% v/v) in water/iPrOH at 40 °C column
temperature over 30 min. The eluting material was detected
using a UV detector set at 260 nm.
4.2. Sodium Phenyl[(benzyloxycarbonyloxy)methyl]-
phosphonate (35). A 16 L, jacketed glass-lined reactor was
1258
https://doi.org/10.1021/acs.oprd.0c00428
Org. Process Res. Dev. 2021, 25, 1247−1262

Organic Process Research & Development
pubs.acs.org/OPRD
Article
0.003 equiv), and the slurry was aged for 15 min. With a
maximum jacket temperature of 55 °C, the batch was distilled
under vacuum while maintaining a constant volume via addition
of 2-butanone, until the Kf of the mother liquor was NMT 1.5%
(required 32.8 L of fresh 2-butanone, 10.9 V). The batch was
then cooled to 22 °C over 1 h, filtered, washed with 2-butanone
(15.0 L, 5.0 V), and dried under vacuum at 45 °C for 20 h to
28d was obtained as a 15.9 wt % stock solution in DCM [98.16%
AN purity by UPLC, 85.4% assay yield, 55:45 (SP)/(RP)]. For
characterization purposes, 27d and 28d were obtained by
chromatographic purification of a portion of the DCM solution.
Data for 27d: 1H NMR (400 MHz, CDCl3): δ 7.40−7.35 (m,
5H), 7.32−7.27 (m, 2H), 7.20−7.13 (m, 3H), 5.22−5.16 (m,
2H), 4.57 (dd, J = 14.2, 9.0 Hz, 1H), 4.47 (dd, J = 14.2, 7.1 Hz,
1H), 4.20−4.02 (m, 3H), 3.66 (t, J = 10.6 Hz, 1H), 1.30 (d, J =
7.1 Hz, 3H), 1.19 (t, J = 7.1 Hz, 3H). 13C NMR (101 MHz,
CDCl3): δ 173.4 (d, J = 5.6 Hz), 154.7 (d, J = 10.4 Hz), 149.8 (d,
J = 8.5 Hz), 134.7, 129.8, 128.8, 128.6, 128.5, 125.2 (d, J = 0.9
Hz), 120.5 (d, J = 4.6 Hz), 70.5, 62.2 (d, J = 156.8 Hz), 61.5,
49.7, 21.3 (d, J = 4.1 Hz), 14.0. 31P NMR (162 MHz, CDCl3): δ
17.9 ppm. HRMS [M + H]+: calcd for C20H25O7NP, 422.1363;
found, 422.1360. Data for 28d: 1H NMR (400 MHz, CDCl3): δ
7.41−7.34 (m, 5H), 7.31−7.27 (m, 2H), 7.19−7.13 (m, 3H),
5.23−5.16 (m, 2H), 4.54 (d, J = 8.0 Hz, 2H), 4.18−4.05 (m,
1H), 4.12 (q, J = 7.2 Hz, 1H), 4.11 (q, J = 7.2 Hz, 1H), 3.60 (t, J
= 10.5 Hz, 1H), 1.31 (d, J = 7.1 Hz, 3H), 1.22 (t, J = 7.1 Hz, 3H).
13C NMR (101 MHz, CDCl3): δ 173.6 (d, J = 5.8 Hz), 154.7 (d,
J = 10.1 Hz), 149.7 (d, J = 8.6 Hz), 134.7, 129.7, 128.7, 128.6,
128.5, 125.1 (d, J = 1.3 Hz), 120.6 (d, J = 4.7 Hz), 70.4, 61.8 (d, J
= 155.3 Hz), 61.5, 49.6 (d, J = 1.8 Hz), 21.3 (d, J = 4.2 Hz), 14.0.
31P NMR (162 MHz, CDCl3): δ 18.9 ppm. HRMS [M + H]+:
calcd for C20H25O7NP, 422.1363; found, 422.1361.
provide 3.06 kg of 35 as a white solid (95.7% AN purity by 31
P
NMR, 78% yield corrected for purity of diphenyl phosphite). 1H
NMR (400 MHz, DMSO-d6): δ 7.41−7.29 (m, 5H), 7.25−7.17
(m, 2H), 7.12−7.04 (m, 2H), 7.03−6.95 (m, 1H), 5.09 (s, 2H),
4.12 (d, J = 8.3 Hz, 2H) ppm. 13C NMR (101 MHz, D2O): δ
155.2 (d, J = 9.6 Hz), 151.4 (d, J = 7.7 Hz), 134.8, 129.6 (d, J =
1.1 Hz), 128.5, 128.4, 128.2, 124.2 (d, J = 1.1 Hz), 120.8 (d, J =
3.7 Hz), 70.1, 61.7 (d, J = 160.3 Hz) ppm. 31P NMR (162 MHz,
DMSO-d6): δ 6.3 ppm. HRMS [M + H]+: calcd for C15H16O6P,
323.0684; found, 323.0685; [M − H]−: calcd for C15H14O6P,
321.0528; found, 321.0533.
4.3. Ethyl (S)-[(Benzyloxycarbonyloxy)methyl]-
(phenoxy)phosphoryl-L-alaninate (27d) and Ethyl (R)-
[(Benzyloxycarbonyloxy)methyl](phenoxy)phosphoryl-
L-alaninate (28d). A three-neck, 2 L round-bottom flask was
charged with oxalyl chloride (0.22 kg, 1.70 mol, 1.2 equiv) and
toluene (0.58 L, 1.2 V). A 15 L jacketed glass-lined reactor was
charged with 35 (0.50 kg, 1.46 mol, 1.0 equiv), DMF (5.36 g, 7.3
mmol, 0.05 equiv), and toluene (5.2 L, 10.4 V). DCM (0.38 L,
0.8 V) was charged via a spray head. The reactor contents were
adjusted to 15 °C, and the oxalyl chloride/toluene solution was
added slowly over 2 h while maintaining the temperature of the
batch at 25 °C. The resulting slurry was aged at 22 °C for 20 h
and then sampled for conversion to the monochloridate 36. 31P
NMR (162 MHz, CDCl3): δ 26.2 ppm (100% AN). The reactor
contents were distilled under vacuum with a maximum jacket
temperature of 40 °C to a final volume of ca. 1.5 L (3.0 V). The
temperature was adjusted to 22 °C, and DCM (0.7 L, 1.4 V) was
added via a spray head. A three-neck, 3 L round-bottom flask was
charged with 11 (0.25 kg, 1.59 mol, 1.1 equiv) and DCM (2.7 L,
5.4 V). After aging for 15 min at room temperature,
chlorotrimethylsilane (39 g, 0.36 mol, 0.25 equiv) was added,
and the solution was aged for 20 h. A three-neck, 3 L round-
bottom flask was charged with DIPEA (0.47 kg, 3.64 mol, 2.5
equiv), DCM (0.38 L, 0.8 V), chlorotrimethylsilane (16 g, 0.15
mol, 0.1 equiv), and toluene (0.29 L, 0.6 V) and aged for 2 h at
room temperature. The solution of 11 in DCM was charged to
the reactor containing the monochloridate 36 over 30 min, and
the mixture was aged for 1 h at 22 °C. The reactor contents were
adjusted to 5 °C, and the DIPEA/DCM/toluene solution was
added slowly over 2 h while maintaining the temperature of the
batch at 10 °C. After aging for 1 h at 5 °C, the mixture was
sampled for conversion to 27d/28d. UPLC: 95.52% AN, 55:45
(SP)/(RP). Water (2.5 L, 5.0 V) and MTBE (3.4 L, 6.8 V) were
charged while keeping the internal temperature at 25 °C. The
reactor contents were adjusted to 22 °C and aged for 15 min.
The phases were allowed to settle, the aqueous layer was
separated, and the organic layer was washed sequentially with aq.
K2CO3 (0.25 M, 2.5 kg, 5.0 S), twice with aq. HCl (0.1 M, 2.5 kg,
5.0 S), and water (2.5 kg, 5.0 S). The resulting organic layer was
concentrated under vacuum with a maximum jacket temper-
ature of 40 °C to a final volume of ca. 0.8 L (1.6 V). DCM (1.1 L,
2.2 V) was added via a spray head. The reactor contents were
concentrated under vacuum until the distillation stalled, and a
dark mobile oil was obtained. DCM (2.3 L, 4.6 V) was added,
and the reactor contents were adjusted to 22 °C. 3.30 kg of 27d/
4.4. Ethyl (S)-(Hydroxymethyl) (phenoxy)phosphoryl-
L-alaninate (1) and Ethyl (R)-(Hydroxymethyl)
(phenoxy)phosphoryl-L-alaninate (30). A 1 L Parr reactor
was charged with water-wet 10% Pd/C (26.15 g, 0.02 equiv) and
a stock solution of 27d/28d in DCM (3.30 kg solution, 0.52 kg
27d/28d, 1.23 mol). After degassing, the reactor contents were
adjusted to 9 °C and hydrogenated at 3.5 bar. After 5.5 h, the
mixture was sampled for conversion to 1/30. 31P NMR (162
MHz, CDCl3): δ 26.1 ppm (43.8% AN, 30), 24.8 ppm (55.0%
AN, 1). The mixture was filtered through a Buchner funnel
coated with Celite (0.15 kg). The reactor and Celite pad were
rinsed forward with DCM (2.9 L, 5.5 V). The filtrates were
combined and transferred into a 5 L glass-lined reactor. The
solution was concentrated under vacuum with a maximum
jacket temperature of 30 °C to a final volume of ca. 0.8 L (1.5 V).
The reactor contents were co-distilled with MTBE (2.7 L, 5.2 V)
to a final volume of ca. 0.8 L (1.5 V). MTBE (2.7 L, 5.2 V) was
charged to the reactor, and the mixture was sampled for the
1
determination of toluene content by H NMR. Co-distillation
with MTBE was repeated until the toluene content with respect
to MTBE was 2.5 mol % or less. The reactor contents were
adjusted to 30 °C and ramp-cooled to −18 °C over 5 h. The
mixture was seeded with 1 (7.8 g, 0.02 equiv) when the internal
temperature reached 10 °C. The mixture was aged at −18 °C for
14 h, filtered, and the cake was rinsed with MTBE (0.2 L, 0.4 V)
precooled at −18 °C . The wet cake was dried under vacuum at
20 °C for 5 h to provide 0.164 kg of 1 as a white solid [99.52%
AN purity by UPLC, 46.1% yield, 99.4:0.6 (SP)/(RP)]. 1H NMR
(400 MHz, CDCl3): δ 7.31 (dd, J = 8.6, 7.2 Hz, 2H), 7.22 (dt, J =
8.7, 1.3 Hz, 2H), 7.18−7.11 (m, 1H), 4.15 (q, J = 7.1 Hz, 2H),
4.09−4.01 (m, 1H), 4.01 (dd, J = 14.8, 5.5 Hz, 1H), 3.96 (dd, J =
14.8, 5.2 Hz, 1H), 3.86 (br s, 1H), 3.40 (br s, 1H), 1.34 (d, J =
7.0 Hz, 3H), 1.24 (t, J = 7.2 Hz, 3H). 13C NMR (101 MHz,
CDCl3): δ 174.2 (d, J = 4.7 Hz), 150.2 (d, J = 9.2 Hz), 129.6,
124.8, 120.5 (d, J = 4.6 Hz), 61.5, 58.2 (d, J = 149.8 Hz), 49.6,
20.9 (d, J = 4.5 Hz), 14.0. 31P NMR (162 MHz, CDCl3): δ 24.8
ppm. HRMS [M + H]+: calcd for C12H19O5NP, 288.0995;
found, 288.0993. For characterization purposes, 30 was
1259
https://doi.org/10.1021/acs.oprd.0c00428
Org. Process Res. Dev. 2021, 25, 1247−1262

Organic Process Research & Development
pubs.acs.org/OPRD
Article
obtained as a white solid after chromatographic purification of
the mother liquor. 1H NMR (400 MHz, CDCl3): δ 7.36−7.27
(m, 2H), 7.22 (dt, J = 8.6, 1.3 Hz, 2H), 7.20−7.11 (m, 1H),
4.20−4.11 (m, 1H), 4.17 (q, J = 7.1 Hz, 2H), 4.00 (dd, J = 14.9,
6.9 Hz, 1H), 3.93 (dd, J = 14.9, 2.9 Hz, 1H), 3.48 (br s, 1H), 3.17
(br s, 1H), 1.37 (d, J = 7.2 Hz, 3H), 1.26 (t, J = 7.2 Hz, 3H). 13C
NMR (101 MHz, CDCl3): δ 174.9 (d, J = 4.3 Hz), 150.0 (d, J =
9.2 Hz), 129.5, 124.8 (d, J = 1.2 Hz), 120.7 (d, J = 4.4 Hz), 61.6,
58.0 (d, J = 147.6 Hz), 49.5 (d, J = 1.2 Hz), 20.8 (d, J = 4.8 Hz),
13.9. 31P NMR (162 MHz, CDCl3): δ 26.1 ppm. HRMS [M +
H]+: calcd for C12H19O5NP, 288.0995; found, 288.0995.
Bing Shi − Process Chemistry, Gilead Sciences Inc., Foster City,
California 94404, United States
Eric A. Standley − Process Chemistry, Gilead Sciences Inc.,
Foster City, California 94404, United States
Li Wang − Process Development, Gilead Alberta ULC,
Edmonton, Alberta T6S 1A1, Canada
Xueqing Wang − Process Development, Gilead Alberta ULC,
Edmonton, Alberta T6S 1A1, Canada
Guojun Yu − Process Development, Gilead Alberta ULC,
Edmonton, Alberta T6S 1A1, Canada
Complete contact information is available at:
https://pubs.acs.org/10.1021/acs.oprd.0c00428
ASSOCIATED CONTENT
■
Notes
sı
* Supporting Information
The authors declare the following competing financial
interest(s): The authors are employees of Gilead Sciences,
Inc., Foster City, California, USA, and Gilead Alberta ULC,
Edmonton, Alberta, Canada.
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/acs.oprd.0c00428.
General procedures, experimental details, synthesis of
starting materials, and NMR data (PDF)
ACKNOWLEDGMENTS
■
The authors are grateful to Amy Cagulada, Nisha Shah, Elan
Shao, and Mark Smith for performing reaction safety assess-
ments, Alban Pereira and Alejandro Corona for assistance with
obtaining characterization data, and Chiajen Lai for helpful
discussion on selective crystallization.
AUTHOR INFORMATION
■
Corresponding Author
Andrea Ambrosi − Process Chemistry, Gilead Sciences Inc.,
Foster City, California 94404, United States; orcid.org/
0000-0002-2889-5913; Email: andrea.ambrosi@
gilead.com
REFERENCES
■
(1) (a) Pertusati, F.; Serpi, M.; McGuigan, C. Medicinal Chemistry of
Nucleoside Phosphonate Prodrugs for Antiviral Therapy. Antiviral
Chem. Chemother. 2012, 22, 181−203. (b) Pradere, U.; Garnier-
Amblard, E. C.; Coats, S. J.; Amblard, F.; Schinazi, R. F. Synthesis of
Nucleoside Phosphate and Phosphonate Prodrugs. Chem. Rev. 2014,
114, 9154−9218. (c) Slusarczyk, M.; Serpi, M.; Pertusati, F.
Phosphoramidates and phosphonamidates (ProTides) with antiviral
activity. Antiviral Chem. Chemother. 2018, 26, 204020661877524.
(d) Mehellou, Y.; Rattan, H. S.; Balzarini, J. The ProTide Prodrug
Technology: From the Concept to the Clinic. J. Med. Chem. 2018, 61,
2211−2226.
Authors
Dustin A. Bringley − Process Chemistry, Gilead Sciences Inc.,
Foster City, California 94404, United States
Selcuk Calimsiz − Process Development, Gilead Alberta ULC,
Edmonton, Alberta T6S 1A1, Canada
Jeffrey A. O. Garber − Process Development, Gilead Alberta
ULC, Edmonton, Alberta T6S 1A1, Canada
Huy Huynh − Process Development, Gilead Alberta ULC,
Edmonton, Alberta T6S 1A1, Canada
Sankar Mohan − Process Development, Gilead Alberta ULC,
Edmonton, Alberta T6S 1A1, Canada
Keshab Sarma − Process Chemistry, Gilead Sciences Inc., Foster
City, California 94404, United States
Jinyu Shen − Process Development, Gilead Alberta ULC,
Edmonton, Alberta T6S 1A1, Canada
Jonah Curl − Process Development, Gilead Alberta ULC,
Edmonton, Alberta T6S 1A1, Canada
Bernard Kwong − Process Development, Gilead Alberta ULC,
Edmonton, Alberta T6S 1A1, Canada
Olga Lapina − Process Chemistry, Gilead Sciences Inc., Foster
City, California 94404, United States
Edmund Leung − Process Development, Gilead Alberta ULC,
Edmonton, Alberta T6S 1A1, Canada
Lennie Lin − Process Development, Gilead Alberta ULC,
Edmonton, Alberta T6S 1A1, Canada
Andrew Martins − Process Development, Gilead Alberta ULC,
Edmonton, Alberta T6S 1A1, Canada
Teague McGinitie − Analytical Chemistry, Gilead Alberta
ULC, Edmonton, Alberta T6S 1A1, Canada
Jaspal Phull − Process Development, Gilead Alberta ULC,
Edmonton, Alberta T6S 1A1, Canada
Ben Roberts − Process Chemistry, Gilead Sciences Inc., Foster
City, California 94404, United States
Mary Rosario − Process Development, Gilead Alberta ULC,
Edmonton, Alberta T6S 1A1, Canada
(2) (a) Ballatore, C.; McGuigan, C.; De Clercq, E.; Balzarini, J.
Synthesis and evaluation of novel amidate prodrugs of PMEA and
PMPA. Bioorg. Med. Chem. Lett. 2001, 11, 1053−1056. (b) Chapman,
H.; Kernan, M.; Prisbe, E.; Rohloff, J.; Sparacino, M.; Terhorst, T.; Yu,
R. Practical synthesis, separation, and stereochemical assignment of the
PMPA pro-drug GS-7340. Nucleosides, Nucleotides Nucleic Acids 2001,
́
20, 621−628. (c) Pertusati, F.; Hinsinger, K.; Flynn, A. S.; Powell, N.;
Tristram, A.; Balzarini, J.; McGuigan, C. PMPA and PMEA prodrugs
for the treatment of HIV infections and human papillomavirus (HPV)
associated neoplasia and cancer. Eur. J. Med. Chem. 2014, 78, 259−268.
(d) Ray, A. S.; Fordyce, M. W.; Hitchcock, M. J. M. Tenofovir
alafenamide: A novel prodrug of tenofovir for the treatment of Human
Immunodeficiency Virus. Antiviral Res. 2016, 125, 63−70.
(3) (a) Cihlar, T.; Ray, A. S.; Boojamra, C. G.; Zhang, L.; Hui, H.;
Laflamme, G.; Vela, J. E.; Grant, D.; Chen, J.; Myrick, F.; White, K. L.;
Gao, Y.; Lin, K.-Y.; Douglas, J. L.; Parkin, N. T.; Carey, A.; Pakdaman,
R.; Mackman, R. L. Design and Profiling of GS-9148, a Novel
Nucleotide Analog Active against Nucleoside-Resistant Variants of
Human Immunodeficiency Virus Type 1, and Its Orally Bioavailable
Phosphonoamidate Prodrug, GS-9131. Antimicrob. Agents Chemother.
2008, 52, 655−665. (b) Mackman, R. L.; Ray, A. S.; Hui, H. C.; Zhang,
L.; Birkus, G.; Boojamra, C. G.; Desai, M. C.; Douglas, J. L.; Gao, Y.;
Grant, D.; Laflamme, G.; Lin, K.-Y.; Markevitch, D. Y.; Mishra, R.;
McDermott, M.; Pakdaman, R.; Petrakovsky, O. V.; Vela, J. E.; Cihlar,
T. Discovery of GS-9131: Design, synthesis and optimization of
amidate prodrugs of the novel nucleoside phosphonate HIV reverse
transcriptase (RT) inhibitor GS-9148. Bioorg. Med. Chem. 2010, 18,
3606−3617.
1260
https://doi.org/10.1021/acs.oprd.0c00428
Org. Process Res. Dev. 2021, 25, 1247−1262

Organic Process Research & Development
pubs.acs.org/OPRD
Article
(4) Standley, E. A.; Bringley, D. A.; Calimsiz, S.; Ng, J. D.; Sarma, K.;
Shen, J.; Siler, D. A.; Ambrosi, A.; Chang, W.-T. T.; Chiu, A.; Davy, J.
A.; Doxsee, I. J.; Esanu, M. M.; Garber, J. A. O.; Kim, Y.; Kwong, B.;
Lapina, O.; Leung, E.; Lin, L.; Martins, A.; Phoenix, J.; Phull, J.; Roberts,
B. J.; Shi, B.; St-Jean, O.; Wang, X.; Wang, L.; Wright, N.; Yu, G.
Synthesis of Rovafovir Etalafenamide (Part I): Active Pharmaceutical
Ingredient Process Development, Scale-Up, and Impurity Control
Strategy. Org. Process Res. Dev. 2021.
(5) Yu, R. H. C.; Brown, B. H.; Polniaszek, R. P.; Graetz, B. R.; Sujino,
K.; Tran, D. D.-P.; Triman, A. S.; Kent, K. M.; Pfeiffer, S. Processes and
intermediates for preparing anti-HIV agents. WO 2012159047 A1,
2012.
Bansal, S.; Espiritu, C.; Keilman, M.; Lam, A. M.; Steuer, H. M. M.; Niu,
C.; Otto, M. J.; Furman, P. A. Discovery of a β-d-2′-Deoxy-2′-α-fluoro-
2′-β-C-methyluridine Nucleotide Prodrug (PSI-7977) for the Treat-
ment of Hepatitis C Virus. J. Med. Chem. 2010, 53, 7202−7218.
(13) Pertusati, F.; McGuigan, C. Diastereoselective synthesis of P-
chirogenic phosphoramidate prodrugs of nucleoside analogues
(ProTides) via copper catalysed reaction. Chem. Commun. 2015, 51,
8070−8073.
(14) Roman, C. A.; Balzarini, J.; Meier, C. Diastereoselective Synthesis
of Aryloxy Phosphoramidate Prodrugs of 3′-Deoxy-2′,3′-didehydro-
thymidine Monophosphate. J. Med. Chem. 2010, 53, 7675−7681.
(15) Xiang, D. F.; Bigley, A. N.; Desormeaux, E.; Narindoshvili, T.;
Raushel, F. M. Enzyme-Catalyzed Kinetic Resolution of Chiral
Precursors to Antiviral Prodrugs. Biochemistry 2019, 58, 3204−3211.
(16) Steinreiber, J.; Faber, K.; Griengl, H. De-racemization of
Enantiomers versus De-epimerization of DiastereomersClassifica-
tion of Dynamic Kinetic Asymmetric Transformations (DYKAT).
Chem.Eur. J. 2008, 14, 8060−8072.
(6) (a) Curzons, A. D.; Mortimer, D. N.; Constable, D. J. C.;
Cunningham, V. L. So you think your process is green, how do you
know?Using principles of sustainability to determine what is green−a
corporate perspective. Green Chem. 2001, 3, 1−6. (b) Jimenez-
Gonzalez, C.; Ponder, C. S.; Broxterman, Q. B.; Manley, J. B. Using the
Right Green Yardstick: Why Process Mass Intensity Is Used in the
Pharmaceutical Industry To Drive More Sustainable Processes. Org.
Process Res. Dev. 2011, 15, 912−917.
(17) (a) Cavalier, J.-F.; Fotiadu, F.; Verger, R.; Buono, G. New Highly
Diastereoselective Synthesis of Phosphoramidates. A Route to Chiral
Methyl p-Nitrophenyl Alkylphosphonates. Synlett 1998, 73−75.
(b) Tran, K.; Beutner, G. L.; Schmidt, M.; Janey, J.; Chen, K.; Rosso,
V.; Eastgate, M. D. Development of a Diastereoselective Phosphor-
ylation of a Complex Nucleoside via Dynamic Kinetic Resolution. J.
Org. Chem. 2015, 80, 4994−5003. (c) Wang, L.; Cao, S.; Du, Z.; Wu,
Q.; Bian, Z.; Kang, C.; Gao, L.; Zhang, J. 4-Dimethylaminopyridine-
catalyzed dynamic kinetic resolution in asymmetric synthesis of P-
chirogenic 1,3,2-oxazaphospholidine-2-oxides. RSC Adv. 2016, 6,
89665−89670. (d) Cini, E.; Barreca, G.; Carcone, L.; Manetti, F.;
Rasparini, M.; Taddei, M. Stereoselective Synthesis of Sofosbuvir
through Nucleoside Phosphorylation Controlled by Kinetic Reso-
lution. Eur. J. Org. Chem. 2018, 2622−2628.
(7) The hydrolysis of phosphonamidates 12/13 to yield 14 is an
extremely facile reaction in the presence of trace amounts of base and
water.
(8) (a) Cox, J. R.; Ramsay, O. B. Mechanisms of Nucleophilic
Substitution in Phosphate Esters. Chem. Rev. 1964, 64, 317−352.
(b) Harger, M. J. P. An elimination-addition mechanism for some
phosphonamidic chloride-amine reactions. Tetrahedron Lett. 1981, 22,
4741−4742. (c) Harger, M. J. P. Evidence for elimination−addition
mechanisms in the reactions of N-t-butyl-P-alkylphosphonamidic
chlorides with t-butylamine and isopropylamine. J. Chem. Soc., Perkin
Trans. 1 1983, 2127−2131. (d) Freeman, S.; Harger, M. J. P.
Associative and dissociative mechanisms for the reactions of N-t-butyl-
P-phenylphosphonamidic chloride with isopropylamine and t-butyl-
amine: competitive, kinetic, and stereochemical studies. J. Chem. Soc.,
Perkin Trans. 2 1988, 81−90.
(18) (a) Liu, S.; Zhang, Z.; Xie, F.; Butt, N. A.; Sun, L.; Zhang, W. First
catalytic enantioselective synthesis of P-stereogenic phosphoramides
via kinetic resolution promoted by a chiral bicyclic imidazole
nucleophilic catalyst. Tetrahedron: Asymmetry 2012, 23, 329−332.
(b) Wang, L.; Du, Z.; Wu, Q.; Jin, R.; Bian, Z.; Kang, C.; Guo, H.; Ma,
X.; Gao, L. Organocatalytic Enantioselective Synthesis of P-Stereogenic
Chiral Oxazaphospholidines. Eur. J. Org. Chem. 2016, 2024−2028.
(c) DiRocco, D. A.; Ji, Y.; Sherer, E. C.; Klapars, A.; Reibarkh, M.;
Dropinski, J.; Mathew, R.; Maligres, P.; Hyde, A. M.; Limanto, J.;
Brunskill, A.; Ruck, R. T.; Campeau, L.-C.; Davies, I. W. A
multifunctional catalyst that stereoselectively assembles prodrugs.
Science 2017, 356, 426−430.
(9) van Bochove, M. A.; Swart, M.; Bickelhaupt, F. M. Nucleophilic
Substitution at Phosphorus (SN2@P): Disappearance and Reappear-
ance of Reaction Barriers. J. Am. Chem. Soc. 2006, 128, 10738−10744.
(19) Ross, B. S.; Ganapati Reddy, P.; Zhang, H.-R.; Rachakonda, S.;
Sofia, M. J. Synthesis of Diastereomerically Pure Nucleotide
Phosphoramidates. J. Org. Chem. 2011, 76, 8311−8319.
(10) (a) Kaljurand, I.; Kutt, A.; Sooväli, L.; Rodima, T.; Mäemets, V.;
̈
(20) (a) Anderson, N. G. Developing Processes for Crystallization-
Induced Asymmetric Transformation. Org. Process Res. Dev. 2005, 9,
800−813. (b) Brands, K. M. J.; Davies, A. J. Crystallization-Induced
Diastereomer Transformations. Chem. Rev. 2006, 106, 2711−2733.
(21) Anderson, N. G. Route Selection. In Practical Process Research &
Development; Anderson, N. G., Ed.; Academic Press: San Diego, 2000;
pp 27−52.
(22) Gelat, F.; Lacomme, C.; Berger, O.; Gavara, L.; Montchamp, J.-L.
Synthesis of (phosphonomethyl)phosphinate pyrophosphate ana-
logues via the phospha-Claisen condensation. Org. Biomol. Chem.
2015, 13, 825−833.
(23) (a) Zhou, X.; Reilly, J. E.; Loerch, K. A.; Hohl, R. J.; Wiemer, D.
F. Synthesis of isoprenoid bisphosphonate ethers through C−P bond
formations: Potential inhibitors of geranylgeranyl diphosphate
synthase. Beilstein J. Org. Chem. 2014, 10, 1645−1650. (b) Berger,
O.; Montchamp, J.-L. General synthesis of P-stereogenic compounds:
the menthyl phosphinate approach. Org. Biomol. Chem. 2016, 14,
7552−7562. (c) Mary, F.; Arrachart, G.; Leydier, A.; Pellet-Rostaing, S.
Synthesis of organophosphorus ligands with a central oxygen atom and
their applications in solvent extraction. Tetrahedron 2019, 75, 3968−
3976.
Leito, I.; Koppel, I. A. Extension of the Self-Consistent Spectrophoto-
metric Basicity Scale in Acetonitrile to a Full Span of 28 pKa Units:
Unification of Different Basicity Scales. J. Org. Chem. 2005, 70, 1019−
̀
1028. (b) Garrido, G.; Koort, E.; Rafols, C.; Bosch, E.; Rodima, T.;
Leito, I.; Rosés, M. Acid−Base Equilibria in Nonpolar Media. Absolute
pKa Scale of Bases in Tetrahydrofuran. J. Org. Chem. 2006, 71, 9062−
9067. (c) Room, E.-I.; Kutt, A.; Kaljurand, I.; Koppel, I.; Leito, I.;
̃ ̃
̈
Koppel, I. A.; Mishima, M.; Goto, K.; Miyahara, Y. Brønsted Basicities
of Diamines in the Gas Phase, Acetonitrile, and Tetrahydrofuran.
Chem.Eur. J. 2007, 13, 7631−7643.
̌
̌
(11) Hanusek, J.; Hejtmánková, L.; Sterba, V.; Sedlák, M. Influence of
substitution on kinetics and mechanism of ring transformation of
substituted S-[1-phenylpyrrolidin-2-on-3-yl]isothiuronium salts. Org.
Biomol. Chem. 2004, 2, 1756−1763.
(12) (a) Lee, W. A.; He, G.-X.; Eisenberg, E.; Cihlar, T.; Swaminathan,
S.; Mulato, A.; Cundy, K. C. Selective Intracellular Activation of a Novel
Prodrug of the Human Immunodeficiency Virus Reverse Transcriptase
Inhibitor Tenofovir Leads to Preferential Distribution and Accumu-
lation in Lymphatic Tissue. Antimicrob. Agents Chemother. 2005, 49,
1898−1906. (b) Sofia, M. J.; Bao, D.; Chang, W.; Du, J.; Nagarathnam,
D.; Rachakonda, S.; Reddy, P. G.; Ross, B. S.; Wang, P.; Zhang, H.-R.;
1261
https://doi.org/10.1021/acs.oprd.0c00428
Org. Process Res. Dev. 2021, 25, 1247−1262

Organic Process Research & Development
pubs.acs.org/OPRD
Article
(24) Wei, Y.; Qiu, G.; Lei, B.; Qin, L.; Chu, H.; Lu, Y.; Zhu, G.; Gao,
Q.; Huang, Q.; Qian, G.; Liao, P.; Luo, X.; Zhang, X.; Zhang, C.; Li, Y.;
Zheng, S.; Yu, Y.; Tang, P.; Ni, J.; Yan, P.; Zhou, Y.; Li, P.; Huang, X.;
Gong, A.; Liu, J. Oral Delivery of Propofol with Methoxymethylphos-
phonic Acid as the Delivery Vehicle. J. Med. Chem. 2017, 60, 8580−
8590.
(25) Solubility of 27m in MTBE: 3 mg/g; CPME: 10 mg/g; iPrOAc:
16 mg/g.
(26) While re-benzylation of 30 would require use of benzyl
trichloroacetimidate (Section 2.1), acylation with a carbamate
protecting group is expected to be quite facile under mild conditions.
See, for instance, Yan, B.; Spilling, C. D. Stereospecific Pd(O)-
Catalyzed Malonate Additions to Allylic Hydroxy Phosphonate
Derivatives: A Formal Synthesis of (−)-Enterolactone. J. Org. Chem.
2004, 69, 2859−2862.
(27) Recycling must involve reprotection of 30. Attempts at direct
epimerization of 30 resulted in its rapid decomposition when exposed
to DBU.
(28) Engel, R. Phosphorus Addition at sp2 Carbon. Org. React. 1988,
36, 175−248.
(29) (a) Kiyokawa, K.; Suzuki, I.; Yasuda, M.; Baba, A. Synthesis of
Cyclopropane-Containing Phosphorus Compounds by Radical Cou-
pling of Butenylindium with Iodo Phosphorus Compounds. Eur. J. Org.
Chem. 2011, 2163−2171. (b) Padilha, G.; Kaufman, T. S.; Silveira, C.
C. Wittig−Horner mediated synthesis of 4-vinyl sulfide derivatives of
pyrazoles. Tetrahedron Lett. 2016, 57, 3349−3353.
1262
https://doi.org/10.1021/acs.oprd.0c00428
Org. Process Res. Dev. 2021, 25, 1247−1262