Continuous Flow C-Glycosylation via Metal-Halogen Exchange: Process Understanding and Improvements toward Efficient Manufacturing of Remdesivir

Article abstract of DOI:10.1021/acs.oprd.0c00370

As remdesivir is the first approved treatment for COVID-19 (SARS-CoV-2), its production is likely to be of vital importance in the near future. Continuous flow processing has been demonstrated as a key technology in the manufacturing of high-volume active pharmaceutical ingredients and is considered for use in this synthetic sequence. In particular, the challenging C-glycosylation of a pyrrolotriazinamine via metal-halogen exchange was identified as a transformation with significant potential benefit, as exemplified by calorimetric analysis of each reaction step. Multiple simplifications of this process were attempted in batch but in general were found to be unfruitful. The five-feed process was then transferred to a flow setup, where specific conditions were found to circumvent solid formation and permit stable processing. Detailed optimization of stoichiometries provided an improvement upon batch conditions with a total residence time of <1 min.

Full text of DOI:10.1021/acs.oprd.0c00370

This is an open access article published under a Creative Commons Attribution (CC-BY)
License, which permits unrestricted use, distribution and reproduction in any medium,
provided the author and source are cited.
pubs.acs.org/OPRD
Communication
Continuous Flow C‑Glycosylation via Metal−Halogen Exchange:
Process Understanding and Improvements toward Efficient
Manufacturing of Remdesivir
Timo von Keutz, Jason D. Williams,* and C. Oliver Kappe*
Cite This: https://dx.doi.org/10.1021/acs.oprd.0c00370
Read Online
sı
*
Metrics & More
Article Recommendations
Supporting Information
ACCESS
ABSTRACT: As remdesivir is the first approved treatment for COVID-19 (SARS-CoV-2), its production is likely to be of vital
importance in the near future. Continuous flow processing has been demonstrated as a key technology in the manufacturing of high-
volume active pharmaceutical ingredients and is considered for use in this synthetic sequence. In particular, the challenging C-
glycosylation of a pyrrolotriazinamine via metal−halogen exchange was identified as a transformation with significant potential
benefit, as exemplified by calorimetric analysis of each reaction step. Multiple simplifications of this process were attempted in batch
but in general were found to be unfruitful. The five-feed process was then transferred to a flow setup, where specific conditions were
found to circumvent solid formation and permit stable processing. Detailed optimization of stoichiometries provided an
improvement upon batch conditions with a total residence time of <1 min.
KEYWORDS: flow chemistry, organometallics, magnesium−halogen exchange, active pharmaceutical ingredient synthesis,
exothermic chemistry
reaction with 2,3,5-tri-O-benzyl-D-ribonolactone (2).3 This
INTRODUCTION
■
step has already been subjected to multiple different synthetic
The rapid spread of the SARS-CoV-2 virus (COVID-19) in the
first half of 2020 has caused mass disruption worldwide and an
enormous number of cases and deaths, which are unprece-
iterations to improve its performance,10 including lithium−
bromide exchange methods, numerous different protecting
dented in recent history.1 Accordingly, treatments for the
groups (most notably, some success in initial routes using 1,2-
bis(dimethylsilyl)ethane),11 and a range of additives to
disease have been sought mostly as repurposed drugs because
they have already passed the initial human safety milestones.2
improve the selectivity. Despite these efforts, few details have
been published, and discussions of the key findings and the
finalized route are generally not included. No patent for a
manufacturing route exists, but on the basis of evidence in the
published literature, the manufacturing route likely includes
this C-glycosylation transformation.6a Alternative synthetic
routes have been suggested, such as the use of artificial-
intelligence-based synthesis planning software;12 however, to
our knowledge no laboratory data on these routes have been
disseminated.
This C-glycosylation reaction of 1 with 2 is expected to
access substantial benefits by flow processing. The excellent
heat- and mass transfer characteristics instilled by continuous
processing make this an ideal method of carrying out
organometallic chemistry.13,14 The long addition times
required for exothermic reactions in batch can be circum-
vented, avoiding the need to hold unstable intermediates for
prolonged time periods. As a result, these organometallic
Of these potential treatments, remdesivir (4), a broad-
spectrum antiviral that was originally developed to treat
Ebola,3 has received significant attention, becoming the first
treatment to be approved by the U.S. Food and Drug
Administration (FDA) and, more recently, receiving approval
in Europe.4 As a result, the demand for this compound will
likely be tremendous, requiring an efficient manufacturing
route, with the accompanying technology and supply chain in
place.5 Many published studies have already begun to improve
or redefine the manufacturing route.6
As highlighted in a recent article, synthetic chemistry has a
key role to play in managing this pandemic by providing
efficient routes to key active pharmaceutical ingredients
(APIs).7 In particular, the global disruption of manufacturing
supply chains has encouraged many countries to refocus their
efforts to become less reliant on importing key APIs and their
intermediates. Implementation of continuous flow technology
is anticipated to have significant benefits for API manufacture
in the coming years,8 specifically, increased throughput and
added flexibility, which will allow for an agile response in times
of dramatically increased demand such as a global pandemic.9
The published synthesis of 4 (Scheme 1a) begins with a C-
glycosylation of iodinated pyrrolotriazinamine 1 performed by
transient N-TMS protection using an organometallic (trace-
less) base followed by magnesium−halogen exchange and
Special Issue: Flow Chemistry Enabling Efficient
Synthesis
Received: August 15, 2020
https://dx.doi.org/10.1021/acs.oprd.0c00370
Org. Process Res. Dev. XXXX, XXX, XXX−XXX
© XXXX American Chemical Society
A

Organic Process Research & Development
pubs.acs.org/OPRD
Communication
a
Scheme 1. Reported C-Glycosylation Process in the Synthesis of Remdesivir
a
(a) The key C-glycosylation requires protection of the free NH2 group in pyrrolotriazinamine 1 followed by metal−halogen exchange and
nucleophilic addition to ribonolactone 2 to afford glycosylated pyrrolotriazinamine 3. A higher yield of 69% has recently been reported for this step
performed on larger scale with some process modifications.6a (b) A stepwise analysis of the C-glycosylation is presented, with the calorimetrically
measured exotherm for each reagent addition shown in red, to estimate the total processing time required on a large scale.
reactions can often be performed at a higher temperature in
flow compared with batch. Consequently, there can be
significant savings in the time and energy consumed in
running these processes, alongside the potential for improved
impurity profiles. In order to harness these benefits, we
endeavored to explore the transformation in detail and
undertake the transfer from batch to continuous flow
processing.
structure 1c, depicted as the MgCl species for simplicity) is
the nucleophilic component for C-glycosylation. A significant
exotherm of 174 kJ/mol was observed, demonstrating the
exceptionally fast and energetic nature of this particular
exchange reaction.
Finally, the nascent Grignard 1c is added to a solution of
ribonolactone 2. The resulting exotherm of 68 kJ/mol is
moderate in magnitude but occurs quickly. Again, this implies
that the manufacturing process will be limited by significantly
extended addition periods rather than the actual times required
for the individual chemical transformations.
Protection and Magnesium−Halogen Exchange:
Attempted Process Modifications. In the analysis of this
route, multiple alternative reagents or procedures were
identified. Because of the lack of previously published details
around this step, it is difficult to discern which of these have
been previously attempted and the corresponding results.
First, it was proposed that a bis(N-benzyl)-protected starting
material would obviate the requirement for in situ TMS
protection (Scheme 2a). This would also be cleaved in the
later O-benzyl deprotection step,3 meaning that no additional
process step would be required. When this N,N-dibenzylation
reaction was carried out, two products were observed: the
expected product 5a was formed alongside the undesired
regioisomeric product 5b (see the Supporting Information for
characterization of the two products) in yields of 43% and
49%, respectively.
When the literature glycosylation conditions were employed
using the expected regioisomer 5a (without TMSCl or
PhMgCl), a reasonable quantity of the expected adduct was
observed (62% by HPLC area). The main side product
observed was the protonated (desiodo) species, due to the
acidity of ribonolactone 2. The other regioisomeric starting
material 5b showed a lower yield of the desired product (41%
by HPLC area), presumably because of the increased steric
hindrance around the nucleophilic center. This approach could
RESULTS AND DISCUSSION
■
Initial Process Assessment and Calorimetry Data.
Prior to initiation of synthetic efforts, a study was carried out in
which the unit operations in this process3b were examined.
Here the exotherms for the individual reagent additions were
measured in separate experiments in order to estimate the
required addition periods on the manufacturing scale and gain
an appreciation of the potential time savings (Scheme 1b).
First, trimethylsilyl chloride (TMSCl) is added to a solution
of iodinated pyrrolotriazinamine 1. This results in a mild
exotherm of 22 kJ/mol. However, more importantly when flow
processing is considered, a significant quantity of solid is
formed. This solid was found to be the HCl salt of the
protected heterocycle (tentative structural representation 1a;
see the Supporting Information for details), resulting from the
HCl released from the successive N-silylations.
Next, 2 equiv of a first Grignard reagent is added, resulting
in a moderate exotherm of 122 kJ/mol. This reagent is
assumed to function as a traceless base to deprotonate the
formed HCl salt or, if only 1 equiv of TMSCl is used, to
deprotonate the NH. The recorded exotherm period was
significantly prolonged, which is proposed to be due to the
heterogeneous nature of the suspension of 1a, meaning that
mass transfer limits the rate of this reaction.
To N-protected species 1b is added a second Grignard
reagent, iPrMgCl·LiCl, to perform a magnesium−iodide
exchange. The resulting aryl Grignard species (proposed
B
https://dx.doi.org/10.1021/acs.oprd.0c00370
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Communication
place of the heavier and more expensive iodide 1. However, the
desired Mg−Br exchange was not observed in the presence of
iPrMgCl·LiCl within 1 h at 0 °C or room temperature. When
the reaction mixture was warmed to 60 °C, a small quantity
(20% yield) of the desired exchange product was formed in 1
h, rendering this process too slow for use in flow processing
even at elevated temperature (Scheme 2c).
Scheme 2. Attempted Modifications of the C-Glycosylation
Process
a
Finally, a combination of Grignard reagent (for N
protection) and organolithium (for Li−Br exchange) was
examined in order to identify a suitable alternative to the
current conditions (Scheme 2d). In the case of nBuLi,
complete exchange was achieved only at −80 °C. Reactions
at −40 and 0 °C showed only 46% and 18% Mg−Br exchange,
respectively. This was attributed to reaction with the N-TMS
groups at higher temperatures. Furthermore, these results
appeared to be irreproducible and likely were strongly
dependent on the addition period, as reported elsewhere.6b
Alternative organolithium reagents (MeLi and MeLi·LiBr)
showed no Li−Br exchange. Since the attempted modifications
showed no clear advantage over the originally reported
conditions,3 the original choice of reagents (TMSCl, PhMgCl,
and iPrMgCl) were maintained for the transition to flow. It
should be noted that because of safety concerns surrounding
benzene formation, an alternative Grignard reagent (e.g.,
tolylmagnesium chloride) would be preferable for larger-scale
development work.
Nucleophilic Addition: Impurities and Analysis. After
the formation of the heteroaryl Grignard reagent, its addition
to 2,3,5-tri-O-benzyl-D-ribonolactone 2 furnishes the desired
glycosylated compound 3. As mentioned in previous reports,
this step has several key challenges for reaction with an
organometallic reagent (Scheme 3). Primarily, the α-proton of
this lactone is relatively acidic, favoring deprotonation by the
Grignard reagent rather than nucleophilic attack. The products
of this pathway are simply the starting pyrrolotriazinamine 7
and the enolate form of lactone 2a, which returns the starting
material 2 upon acidic quench.
Additional complications arising from the proposed ketone
intermediate 3a were also observed. Since this intermediate is
more electrophilic than the starting material 2, it is prone to
undergo reaction with the heteroaryl Grignard or any
remaining Grignard from the previous step. This results in
the two major impurities 8 and 9, respectively. To control
these impurities, it may be advantageous to maintain an excess
of lactone 2.
a
(a) N-Benzyl protection resulted in two isomers, 5a and 5b.
Glycosylation of 5a provided potentially promising results (see the
Supporting Information). (b) The use of alkyl Grignard reagents with
heteroaryl iodide resulted in a mixture of deprotonation and Mg−I
exchange. (c) Unsuccessful Mg−Br exchange provided only minor
arylmagnesium formation even at elevated temperatures and
lengthened reaction times. (d) Combination of Grignard and
organolithium reagents provided irreproducible results or no
exchange.
It is noteworthy that this ketone intermediate 3a has not
been previously proposed but is likely to be present under
nonacidic conditions. In the 13C NMR spectrum of product 3
in DMSO-d6, a signal at 189.0 ppm was observed, clearly
indicating the presence of a ketone. In contrast, the NMR
spectrum in CDCl3 showed no peaks above 160 ppm, implying
the presence of the furanose form 3 (see the Supporting
Information for details).
Flow Process Optimization. During the initial reaction
analysis, a key focus was placed upon the prevention of solid
formation, which would render this process incompatible with
straightforward flow processing.15 Thus, it was identified that
the HCl salt (proposed structure 1a, which arises from mixing
the heterocycle 1 with TMSCl) could be avoided by premixing
TMSCl with PhMgCl prior to addition of heterocycle 1. It is
assumed that this instantly quenches any formed HCl, so the
basic pyrrolotriazinamine nitrogen atoms remain free. The
reaction appeared to be unaffected by the time allowed for
represent a useful method for future work, yet optimization of
the N-benzyl protection step would be required for it to be
synthetically useful.
In the published procedure,3 PhMgCl is used as a base to
facilitate the initial deprotonation/N-TMS protection. In view
of the relatively high molecular weight of this Grignard reagent
and its toxic byproduct (benzene), alternatives were examined.
Ideally, it would be possible to simply use 3 equiv of the same
Grignard reagent to perform the deprotonation and Mg−I
exchange, simplifying the process (Scheme 2b). However, in
our hands, the Mg−I exchange appeared to be rapid when an
alkyl Grignard (e.g., EtMgCl) was used, even in the presence of
two acidic NH groups. The formed heteroaryl Grignard 1c was
observed to quench itself, presumably through deprotonation
of an NH in another substrate molecule.
Considering the rapid Mg−I exchange, it was envisaged that
this reaction could be performed using heteroaryl bromide 6 in
C
https://dx.doi.org/10.1021/acs.oprd.0c00370
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Communication
Scheme 3. Representation of the Reaction Pathway, Including Side Product Formation by Onward Reaction of the Product
(Top Arrow) or Simple Deprotonation of Lactone 2 (Bottom Arrow)
premixing of these reagents, and no solid formation was
observed at any point.
Scheme 4. Schematic Layout of “Flow Setup A” Used in
This Study
a
By implementing the described change in addition order, it
was possible to begin transferring the process to continuous
flow. To achieve this, a FlowPlate16,17 microreactor (Ehrfeld)
was used (process plate with “TG” geometry, internal volume
0.4 mL, channel width 0.2 mm in the mixing structure or 0.5
mm in the residence channel; see the Supporting Information
for further details) with the aim of providing optimal mass- and
heat transfer in a manner that can be maintained for larger-
scale processing.17 The five required reaction streams were
mixed as depicted in Scheme 4. The PhMgCl and TMSCl are
mixed in a Y-piece prior to entering the FlowPlate reactor
because of the low exothermicity and residence time sensitivity
of this step. This stream is then mixed with heteroaryl iodide 1,
followed by addition of iPrMgCl and then lactone electrophile
2.
In order to maintain the process intensity, heteroaryl iodide
1 was used as a 0.4 M solution, which is close to its solubility
limit in THF (∼0.5 M). Similarly, a concentrated solution of
TMSCl (2 M) was used, and both Grignard reagents remained
undiluted (both 2 M). Initial experiments demonstrated that
temperatures below 0 °C could not be used with those
concentrations because of the instant formation of solids.
During operation at temperatures of 40 °C or above, pressure
fluctuation was observed as a result of partial blockages, which
resulted in inconsistent reagent delivery and unrepresentative
results. Furthermore, it was determined that a significant
deficiency of PhMgCl versus TMSCl (1 equiv of PhMgCl and
2 equiv TMSCl) also resulted in solid formation, presumably
due to formation of the HCl salt.
a
The displayed flow rates correspond to the conditions used in Table
1, entry 2.
With these stoichiometries, experiments were carried out to
determine the required residence time for Mg−I exchange.
Although the LiCl salt is proposed to provide a higher rate of
exchange,18 iPrMgCl itself was found to provide full exchange
in a residence time of 2.6 s. This short reaction time enables a
high reaction throughput even in a small reactor volume.
Furthermore, it appeared as though a short residence time of
∼9 s (slightly variable, based on equivalents of reagents used)
was sufficient for complete reaction of lactone 2, since
increasing this final residence time unit showed no decrease
in quenched heterocycle 7 (see Supporting Information).
The detailed flow optimization began by comparing results
versus a benchmark batch reaction carried out as previously
described (Table 1, entry 1),3a which achieved a 40% yield of
D
https://dx.doi.org/10.1021/acs.oprd.0c00370
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Communication
Table 1. Summary of Results from Flow Optimization Experiments Using the Conditions Outlined in Scheme 4
a
yields [%]
entry
equiv of TMSCl
equiv of PhMgCl
equiv of iPrMgCl
equiv of lactone 2
1
7
desired product 3
8
9
b
1
2
2
2
2
2
2
2
2
1.5
1.5
1.2
1.5
2
1.8
1.5
1.8
1
1
−
1
0.1
0.1
0.4
0.6
0.4
0.3
0.6
0.2
3.0
9.1
70.3
8.2
39.6
−
15.4
−
7.1
2.7
4.7
14.0
−
c
2
1.3
1.3
1.3
1.3
1.3
1.3
1.3
1.3
1.3
1.3
1.4
3
4
5
6
7
12.6
21.8
32.8
30.4
28.7
33.2
39.8
42.9
27.9
46.5
43.2
23.9
10.5
14.3
13.9
7.8
13.7
5.6
2.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
18.4
25.9
22.9
25.6
18.7
14.1
12.7
7.9
2
1.5
1.5
2
2
2
4.3
3.8
d
8
12.0
8.4
10.2
12.6
12.7
e
9
e
10
10.3
40.1
3.5
e
11
2
2
e
12
10.1
7.1
a
b
Determined by HPLC area % at 254 nm. The experiment was performed in batch as a benchmark. In this experiment, iPrMgCl·LiCl was used.
c
d
No lactone 2 was added in this experiment to obtain an idea of the performance of the first two reaction steps. A batch quench was used to
e
maintain an excess of 2. Flow setup B was used, providing additional residence time for the N-protection step.
the desired product 3 by HPLC analysis (corresponding to the
reported isolated product yield of 42%), alongside a 9% yield
of protonated heterocycle 7. Side products 8 and 9 were
observed in 15% and 14% yield, respectively. It was found that
in flow a slightly higher quantity of iPrMgCl (1.3 equiv) was
required to achieve complete Mg−I exchange.
The first flow experiments were carried out using a 1.6 mL
sample loop to inject heterocycle 1. An experiment was carried
without lactone 2 (i.e., simply forming Grignard 1c, followed
by a quench) to determine the performance of these first two
steps (entry 2). Although a relatively low purity of the
quenched product 7 was observed, the majority of the side
products did not absorb at 310 nm, implying that the
heterocycle was not incorporated (see the Supporting
Information for the corresponding chromatograms). Com-
pared with batch experiments, the first flow experiment
including lactone 2 showed a very high quantity of side
product 9 (entry 3), although this was found to improve
significantly when the number of equivalents of lactone 2 was
increased (entry 4).
In an attempt to decrease the quantity of side product 8, the
stoichiometries of PhMgCl and TMSCl were modulated
(entries 5−7). Although the quantities of side products 8
and 9 decreased, a significantly higher quantity of the
protonated product 7 was observed instead. This signifies
problems in the initial N protection. The most favorable of
these results was observed when 2 equiv of TMSCl was
maintained, but with 1.5 equiv of PhMgCl (entry 5).
It was envisaged that maintaining an excess of lactone 2
throughout the course of reaction may decrease side product 9.
To achieve this, an experiment was attempted in which the
Grignard (formed in flow) was added to a batch of lactone 2
(i.e., a “batch quench”). The resulting yield was not a
significant improvement on the previous setup (entry 8).
Another consideration was that the N-protection step may
be far slower than Mg−I exchange and nucleophilic attack.
Accordingly, the TMSCl/PhMgCl stream was combined with
heterocycle 1 prior to entering the FlowPlate and was provided
an extra 2 mL residence time unit (∼38 s; see “flow setup B” in
the Supporting Information for details). Gratifyingly, this
provided a significant improvement upon the previous results
(entry 9). The stoichiometry of PhMgCl was reduced to limit
the yield of side product 9 (entries 10 and 11), but some
heteroaryl iodide 1 remained since some iPrMgCl was now
being consumed by the HCl byproduct of N protection. A
slight increase in the number of iPrMgCl equivalents provided
almost complete Mg−I exchange with a lower quantity of 9.
The resulting yield of the desired product 3 was optimal under
these conditions (47%; entry 12). These conditions were
found to be stable for operation over 17 min (3.8 g, 14.7 mmol
of starting material processed) with no pressure fluctuations.
Because of the poor commercial availability of pyrrolotriazin-
amine 7, a longer scale-out run was, unfortunately, not possible
in this setting.
Although the development of this continuous flow
procedure has not resulted in a significant yield improvement
versus the original published procedure (47% vs 40% HPLC
area in our hands),3a there is a marked improvement in
processability. The batch procedure uses three different
temperature regimes (20, 0, and −20 °C) and exotherm-
limited addition times, leading to a reaction time of several
hours (which would likely be much extended on the
manufacturing scale). In contrast, the developed flow
procedure uses only a single temperature zone and enables
complete reaction in a residence time of <1 min, allowing for
E
https://dx.doi.org/10.1021/acs.oprd.0c00370
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Communication
Author
high throughput upon scale-up. In order to facilitate this
straightforward scale-up, the chemistry was performed in a
commercial flow chemistry platform with a defined scale-up
strategy.17
Timo von Keutz − Center for Continuous Flow Synthesis and
Processing (CC FLOW), Research Center Pharmaceutical
Engineering GmbH (RCPE), 8010 Graz, Austria; Institute of
Chemistry, University of Graz, 8010 Graz, Austria
CONCLUSION
Complete contact information is available at:
https://pubs.acs.org/10.1021/acs.oprd.0c00370
■
The use of continuous flow processing for organometallic
reactions has been proven as an efficient manufacturing
technique and may be advantageous for use in the C-
glycosylation step in the synthesis of remdesivir. The potential
advantage was illustrated by a series of calorimetric analyses
that demonstrated significant exotherms arising from four
sequential reagent addition steps. It was expected that the
improved heat transfer afforded by flow processing would
allow substantially faster processing of the reaction mixtures.
Upon initial assessment, numerous potential process
improvements could be applied in order to streamline this
lengthy transformation. Accordingly, a number of these were
attempted, and their related drawbacks have been discussed in
some detail. Additionally, the main side reactions encountered
in this reaction and their associated routes of formation have
been analyzed.
Finally, the reaction was adapted to operate under
continuous flow conditions. Reaction parameters permitting
stable operation were identified, which proved to be a
significant challenge because of the numerous possible sites
of protonation and potentially insoluble metal salts. It was
found that the Mg−I exchange was exceptionally fast in this
case, which allowed a residence time of just 2.6 s for this step,
with a total residence time of just 56 s for the overall three-step
sequence. Further optimization of the reagent stoichiometries
suppressed the identified impurities arising from onward
reaction of the intermediate, maximizing the quantity of C-
glycosylated pyrrolotriazine 3. This corresponds to a
throughput of 51.8 mmol/h for starting material 1, with a
proven scale-up pathway in this reactor using smart-
dimensioning concepts.17 It is envisaged that these results
will prove to be valuable to future remdesivir manufacturing
efforts toward the goal of effective continuous processing.
Funding
The CC FLOW Project (Austrian Research Promotion Agency
FFG 862766) is funded through the Austrian COMET
Program by the Austrian Federal Ministry of Transport,
Innovation and Technology (BMVIT), the Austrian Federal
Ministry for Digital and Economic Affairs (BMDW), and the
State of Styria (Styrian Funding Agency SFG).
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
́
The authors thank Rene Lebl for additional analytical work.
ABBREVIATIONS
API, active pharmaceutical ingredient; TMS, trimethylsilyl
■
REFERENCES
■
(1) (a) Liu, C.; Zhou, Q.; Li, Y.; Garner, L. V.; Watkins, S. P.;
Carter, L. J.; Smoot, J.; Gregg, A. C.; Daniels, A. D.; Jervey, S.; Albaiu,
D. Research and Development on Therapeutic Agents and Vaccines
for COVID-19 and Related Human Coronavirus Diseases. ACS Cent.
Sci. 2020, 6, 315−331. (b) Zhou, P.; Yang, X. L.; Wang, X. G.; Hu, B.;
Zhang, L.; Zhang, W.; Si, H. R.; Zhu, Y.; Li, B.; Huang, C. L.; et al. A
Pneumonia Outbreak Associated with a New Coronavirus of Probable
Bat Origin. Nature 2020, 579, 270−273.
(2) (a) Maxmen, A. More than 80 Clinical Trials Launch to Test
Coronavirus Treatments. Nature 2020, 578, 347−348. (b) De Savi,
C.; Hughes, D. L.; Kvaerno, L. Quest for a COVID-19 Cure by
Repurposing Small-Molecule Drugs: Mechanism of Action, Clinical
Development, Synthesis at Scale, and Outlook for Supply. Org. Process
Res. Dev. 2020, 24, 940−976.
(3) (a) Warren, T. K.; Jordan, R.; Lo, M. K.; Ray, A. S.; Mackman,
R. L.; Soloveva, V.; Siegel, D.; Perron, M.; Bannister, R.; Hui, H. C.;
et al. Therapeutic Efficacy of the Small Molecule GS-5734 against
Ebola Virus in Rhesus Monkeys. Nature 2016, 531, 381−385.
(b) Siegel, D.; Hui, H. C.; Doerffler, E.; Clarke, M. O.; Chun, K.;
Zhang, L.; Neville, S.; Carra, E.; Lew, W.; Ross, B.; et al. Discovery
and Synthesis of a Phosphoramidate Prodrug of a Pyrrolo[2,1-
f ][Triazin-4-Amino] Adenine C-Nucleoside (GS-5734) for the
Treatment of Ebola and Emerging Viruses. J. Med. Chem. 2017, 60,
1648−1661.
ASSOCIATED CONTENT
■
sı
* Supporting Information
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/acs.oprd.0c00370.
Further details of the reaction setup, experimental
results, and characterization data (PDF)
(4) (a) First COVID-19 treatment recommended for EU authorisation,
European Medicines Agency, June 25, 2020. https://www.ema.
europa.eu/en/news/first-covid-19-treatment-recommended-eu-
authorisation (accessed 08-11-2020). (b) European Commission secures
EU access to Remdesivir for treatment of COVID-19. European
Comission, July 29, 2020. https://ec.europa.eu/commission/
presscorner/detail/en/ip_20_1416 (accessed 08-11-2020).
(5) (a) On the remdesivir manufacturing process: “The process is
both resource- and time-intensive, with some individual manufactur-
ing steps taking weeks to complete.” See: Working to Supply
Remdesivir for COVID-19. Gilead,June 24, 2020. https://www.gilead.
com/purpose/advancing-global-health/covid-19/working-to-supply-
remdesivir-for-covid-19 (accessed 08-14-2020). (b) Jarvis, L. M.
Scaling up remdesivir amid the coronavirus crisis. Chem. Eng. News,
April 20, 2020. https://cen.acs.org/biological-chemistry/infectious-
disease/Scaling-remdesivir-amid-coronavirus-crisis/98/web/2020/04
(accessed 08-14-2020).
AUTHOR INFORMATION
■
Corresponding Authors
Jason D. Williams − Center for Continuous Flow Synthesis and
Processing (CC FLOW), Research Center Pharmaceutical
Engineering GmbH (RCPE), 8010 Graz, Austria; Institute of
Chemistry, University of Graz, 8010 Graz, Austria;
orcid.org/0000-0001-5449-5094; Email: jason.williams@
rcpe.at
C. Oliver Kappe − Center for Continuous Flow Synthesis and
Processing (CC FLOW), Research Center Pharmaceutical
Engineering GmbH (RCPE), 8010 Graz, Austria; Institute of
Chemistry, University of Graz, 8010 Graz, Austria;
orcid.org/0000-0003-2983-6007; Email: oliver.kappe@
uni-graz.at
F
https://dx.doi.org/10.1021/acs.oprd.0c00370
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
pubs.acs.org/OPRD
Communication
(6) During the preparation of this communication, three studies
concerning the manufacturing route to remdesivir appeared. See:
(a) Vieira, T.; Stevens, A. C.; Chtchemelinine, A.; Gao, D.; Badalov,
P.; Heumann, L. Development of a Large-Scale Cyanation Process
Using Continuous Flow Chemistry En Route to the Synthesis of
Remdesivir. Org. Process Res. Dev. 2020, DOI: 10.1021/ac-
s.oprd.0c00172. (b) Xue, F.; Zhou, X.; Zhou, R.; Zhou, X.; Xiao,
D.; Zhong, W.; Gu, E.; Guo, X.; Xiang, J.; Wang, K.; Yang, L.; Qin, Y.
Improvement of the C-Glycosylation Step for the Synthesis of
Remdesivir. Org. Process Res. Dev. 2020, 24, 1772. (c) Paymode, D. J.;
Cardoso, F. S. P.; Agrawal, T.; Tomlin, J. W.; Cook, D. W.; Burns, J.
M.; Stringham, R. W.; Sieber, J. D.; Gupton, B. F.; Snead, D. R.
Expanding Access to Remdesivir via an Improved Pyrrolotriazine
Synthesis: Supply Centered Synthesis. Org. Lett. 2020, DOI: 10.1021/
acs.orglett.0c02848.
Aktoudianakis, E.; Cho, A.; Kim, C. U. Practical Synthesis of 1′-
Substituted Tubercidin C-Nucleoside Analogs. Tetrahedron Lett.
2012, 53, 484−486.
(12) For examples of computer-aided synthetic route design toward
́
remdesivir, see: (a) Szymkuc, S.; Gajewska, E. P.; Molga, K.; Wołos,
A.; Roszak, R.; Beker, W.; Moskal, M.; Dittwald, P.; Grzybowski, B. A.
Computer-Generated “Synthetic Contingency” Plans at Times of
Logistics and Supply Problems: Scenarios for Hydroxychloroquine
and Remdesivir. Chem. Sci. 2020, 11, 6736−6744. (b) Lin, Y.; Zhang,
Z.; Mahjour, B.; Wang, D.; Zhang, R.; Shim, E.; McGrath, A.; Shen,
Y.; Brugger, N.; Turnbull, R.; Jasty, S.; Trice, S.; Cernak, T.
Reinforcing the Supply Chain of COVID-19 Therapeutics with
Expert-Coded Retrosynthetic Software. ChemRxiv 2020,
DOI: 10.26434/chemrxiv.12765410.v1.
(13) For reviews of organometallic chemistry in flow, see:
(a) Degennaro, L.; Carlucci, C.; De Angelis, S.; Luisi, R. Flow
Technology for Organometallic-Mediated Synthesis. J. Flow Chem.
2016, 6, 136−166. (b) Colella, M.; Nagaki, A.; Luisi, R. Flow
Technology for the Genesis and Use of (Highly) Reactive
Organometallic Reagents. Chem. - Eur. J. 2020, 26, 19−32.
(c) Yoshida, J.; Takahashi, Y.; Nagaki, A. Flash Chemistry: Flow
Chemistry That Cannot Be Done in Batch. Chem. Commun. 2013, 49,
9896−9904.
(14) For selected examples of industrially relevant organometallic
chemistry in flow, see: (a) Thaisrivongs, D. A.; Naber, J. R.; Rogus, N.
J.; Spencer, G. Development of an Organometallic Flow Chemistry
Reaction at Pilot-Plant Scale for the Manufacture of Verubecestat.
Org. Process Res. Dev. 2018, 22, 403−408. (b) Hafner, A.; Mancino,
V.; Meisenbach, M.; Schenkel, B.; Sedelmeier, J. Dichloromethyl-
lithium: Synthesis and Application in Continuous Flow Mode. Org.
Lett. 2017, 19, 786−789. (c) Hafner, A.; Meisenbach, M.; Sedelmeier,
J. Flow Chemistry on Multigram Scale: Continuous Synthesis of
Boronic Acids within 1 S. Org. Lett. 2016, 18, 3630−3633. (d) Hafner,
A.; Filipponi, P.; Piccioni, L.; Meisenbach, M.; Schenkel, B.;
Venturoni, F.; Sedelmeier, J. A Simple Scale-up Strategy for
Organolithium Chemistry in Flow Mode: From Feasibility to
Kilogram Quantities. Org. Process Res. Dev. 2016, 20, 1833−1837.
(e) Stueckler, C.; Hermsen, P.; Ritzen, B.; Vasiloiu, M.; Poechlauer,
P.; Steinhofer, S.; Pelz, A.; Zinganell, C.; Felfer, U.; Boyer, S.;
Goldbach, M.; De Vries, A.; Pabst, T.; Winkler, G.; Lavopa, V.;
Hecker, S.; Schuster, C. Development of a Continuous Flow Process
for a Matteson Reaction: From Lab Scale to Full-Scale Production of
a Pharmaceutical Intermediate. Org. Process Res. Dev. 2019, 23, 1069−
1077.
(15) For a review of oscillatory flow reactors, including their
applications in solid handling, see: Bianchi, P.; Williams, J. D.; Kappe,
C. O. Oscillatory Flow Reactors for Synthetic Chemistry Applications.
J. Flow Chem. 2020, 10, 475−490.
(16) For selected examples of organometallic chemistry carried out
in the FlowPlate reactor, see: (a) Sagmeister, P.; Williams, J. D.;
Hone, C. A.; Kappe, C. O. Laboratory of the Future: A Modular Flow
Platform with Multiple Integrated PAT Tools for Multistep Reactions.
React. Chem. Eng. 2019, 4, 1571−1578. (b) von Keutz, T.; Cantillo,
D.; Kappe, C. O. Continuous Flow Synthesis of Terminal Epoxides
from Ketones Using in Situ Generated Bromomethyl Lithium. Org.
Lett. 2019, 21, 10094−10098.
(7) Hardy, M. A.; Wright, B. A.; Bachman, J. L.; Boit, T. B.; Haley,
H. M. S.; Knapp, R. R.; Lusi, R. F.; Okada, T.; Tona, V.; Garg, N. K.;
Sarpong, R. Treating a Global Health Crisis with a Dose of Synthetic
Chemistry. ACS Cent. Sci. 2020, 6, 1017−1030.
(8) For reviews of the advantages of continuous flow processing for
API manufacture, see: (a) Gutmann, B.; Cantillo, D.; Kappe, C. O.
Continuous-Flow Technology - A Tool for the Safe Manufacturing of
Active Pharmaceutical Ingredients. Angew. Chem., Int. Ed. 2015, 54,
6688−6728. (b) Porta, R.; Benaglia, M.; Puglisi, A. Flow Chemistry:
Recent Developments in the Synthesis of Pharmaceutical Products.
Org. Process Res. Dev. 2016, 20, 2−25. (c) Baumann, M.; Baxendale, I.
R. The Synthesis of Active Pharmaceutical Ingredients (APIs) Using
Continuous Flow Chemistry. Beilstein J. Org. Chem. 2015, 11, 1194−
1219. (d) McWilliams, J. C.; Allian, A. D.; Opalka, S. M.; May, S. A.;
Journet, M.; Braden, T. M. The Evolving State of Continuous
Processing in Pharmaceutical API Manufacturing: A Survey of
Pharmaceutical Companies and Contract Manufacturing Organiza-
tions. Org. Process Res. Dev. 2018, 22, 1143−1166. (e) Berton, M.; de
Souza, J. M.; Abdiaj, I.; McQuade, T. D.; Snead, D. R. Scaling
continuous API synthesis from milligram to kilogram: extending the
enabling benefits of micro to the plant. J. Flow Chem. 2020, 10, 73−
92.
(9) (a) Lee, S. L.; O’Connor, T. F.; Yang, X.; Cruz, C. N.;
Chatterjee, S.; Madurawe, R. D.; Moore, C. M. V.; Yu, L. X.;
Woodcock, J. Modernizing Pharmaceutical Manufacturing: From
Batch to Continuous Production. J. Pharm. Innov. 2015, 10, 191−199.
(b) de Souza, R. O. M. A.; Watts, P. Flow Processing as a Tool for
API Production in Developing Economies. J. Flow Chem. 2017, 7,
146−150. (c) Malet-Sanz, L.; Susanne, F. Continuous Flow Synthesis.
a Pharma Perspective. J. Med. Chem. 2012, 55, 4062−4098.
(10) Numerous patents reporting different variations of this
synthetic step have been published, but with relatively little detail.
See: (a) Mackman, R. L.; Parrish, J. P.; Ray, A. S.; Theodore, D. A.
Methods and Compounds for Treating Paramyxoviridae Virus
Infections. US 2011045102, 2011. (b) Butler, T.; Cho, A.; Kim, C.
U.; Saunders, O. L.; Zhang, L. 1′-Substituted Carba-Nucleoside
Analogs for Antiviral Treatment. US 2009041447, 2009. (c) Mack-
man, R. L.; Parrish, J. P.; Ray, A.; Theodore, D. A. Methods and
Compounds for Treating Paramyxoviridae Virus Infections. WO
2012/012776 A1, 2012. (d) Chun, B. K.; Clarke, M. O. H.; Doerffler,
E.; Hui, H. C.; Jordon, R.; Mackman, R. L.; Parrish, J. P.; Ray, A. S.;
Siegel, D. Methods for Treating Filoviridae Virus Infections. WO
2016/069826 A1, 2016. (e) Clarke, M. O. H.; Jordan, R.; Mackman,
R. L.; Ray, A.; Siegel, D. Methods for Treating Filoviridae Virus
Infections. WO 2017/184668 A1, 2017. (f) Chun, B. K.; Clarke, M.
O. H.; Doerffler, E.; Hui, H. C.; Jordon, R.; Mackman, R. L.; Parrish,
J. P.; Ray, A. S.; Siegel, D. Methods for Treating Filoviridae Virus
Infections. US 2019/0275063 A1, 2019. Also see in particular refs 6a
and 6b.
(17) (a) Plouffe, P.; Macchi, A.; Roberge, D. M. From Batch to
Continuous Chemical Synthesis-a Toolbox Approach. Org. Process
Res. Dev. 2014, 18, 1286−1294. (b) Mielke, E.; Plouffe, P.; Mongeon,
S. S.; Aellig, C.; Filliger, S.; Macchi, A.; Roberge, D. M. Micro-Reactor
Mixing Unit Interspacing for Fast Liquid−Liquid Reactions Leading
to a Generalized Scale-up Methodology. Chem. Eng. J. 2018, 352,
682−694.
(18) Krasovskiy, A.; Knochel, P. A LiCl-Mediated Br/Mg Exchange
Reaction for the Preparation of Functionalized Aryl- and Hetero-
arylmagnesium Compounds from Organic Bromides. Angew. Chem.,
Int. Ed. 2004, 43, 3333−3336.
(11) (a) Cho, A.; Saunders, O. L.; Butler, T.; Zhang, L.; Xu, J.; Vela,
J. E.; Feng, J. Y.; Ray, A. S.; Kim, C. U. Synthesis and Antiviral
Activity of a Series of 1′-Substituted 4-Aza-7,9-Dideazaadenosine C-
Nucleosides. Bioorg. Med. Chem. Lett. 2012, 22, 2705−2707.
(b) Metobo, S. E.; Xu, J.; Saunders, O. L.; Butler, T.;
G
https://dx.doi.org/10.1021/acs.oprd.0c00370
Org. Process Res. Dev. XXXX, XXX, XXX−XXX