Targeting a Mirabegron precursor by BH3-mediated continuous flow reduction process

Article abstract of DOI:10.1016/j.cattod.2017.09.058

A continuous-flow reduction of (R)-2-hydroxy-N-[2-(4-nitrophenyl)ethyl]-2-phenylacetamide, involved in the synthetic pathway of Mirabegron, has been developed. This study demonstrated the possibility to safely handling BH₃ complexes within microfluidic reactors using 2-MeTHF as greener alternative to traditional solvents, and without requiring any additive such as DMI. In addition, NMR and HPLC purity analysis revealed that the sole by-product of this process is the diamine 3, which wouldn't affect the following synthetic steps towards Mirabegron.

Full text of DOI:10.1016/j.cattod.2017.09.058

Accepted Manuscript
Title: Targeting a Mirabegron Precursor by BH3-Mediated
Continuous Flow Reduction Process
Authors: Sonia De Angelis, Claudia Carlucci, Modesto de
Candia, Gabriele Rebuzzini, Paolo Celestini, Massimiliano
Riscazzi, Renzo Luisi, Leonardo Degennaro
PII:
S0920-5861(17)30673-9
DOI:
Reference:
https://doi.org/10.1016/j.cattod.2017.09.058
CATTOD 11063
To appear in:
Catalysis Today
Received date:
Revised date:
Accepted date:
27-7-2017
18-9-2017
29-9-2017
Please cite this article as: Sonia De Angelis, Claudia Carlucci, Modesto de Candia,
Gabriele Rebuzzini, Paolo Celestini, Massimiliano Riscazzi, Renzo Luisi, Leonardo
Degennaro, Targeting a Mirabegron Precursor by BH3-Mediated Continuous Flow
Reduction Process, Catalysis Today https://doi.org/10.1016/j.cattod.2017.09.058
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

Targeting a Mirabegron Precursor by BH3-Mediated Continuous Flow
Reduction Process
Sonia De Angelis,a Claudia Carlucci,a Modesto de Candia,a Gabriele Rebuzzini,b Paolo Celestini,b
Massimiliano Riscazzib Renzo Luisi,*,a and Leonardo Degennaroa
a FLAME-Lab - Flow Chemistry and Microreactor Technology Laboratory; Department of Pharmacy — Drug Sciences,
University of Bari “A. Moro” Via E. Orabona 4, Bari 70125, Italy.
bCOSMA S.p.A, Via Colleoni 15/17, Ciserano (BG) 24040, Italy
E mail: renzo.luisi@uniba.it
Graphical abstract
CATTOD Highlight



Flow chemistry set up for pharmaceutical applications
Continuous flow reduction using BH3 solutions
Straightforward and green preparation of Mirabegron precursor
Abstract
A continuous-flow reduction of (R)-2-hydroxy-N-[2-(4-nitrophenyl)ethyl]-2-phenylacetamide, involved in
the synthetic pathway of Mirabegron, has been developed. This study demonstrated the possibility to safely
handling BH3 complexes within microfluidic reactors using 2-MeTHF as greener alternative to traditional
solvents, and without requiring any additive such as DMI. In addition, NMR and HPLC purity analysis
revealed that the sole by-product of this process is the diamine 3, which wouldn’t affect the following
synthetic steps towards Mirabegron.
1. Introduction
The phenylethanolamine core occurs in important molecules such as neurotransmitter noradrenaline and it is
also contained in many naturally occurring molecules such as several alkaloids and compounds of
pharmaceutical interest with different activities ranging from bronchodilation to antifungal activity (Fig. 1).
[1,2]
1

Among pharmaceutical molecules containing the phenylethanolamine scaffold, 2-(2-aminothiazol-4-yl)-N-
[4-(2-{[(2R)-2-hydroxy-2-phenylethyl]amino}ethyl)phenyl]acetamide, commonly known as Mirabegron®, is
a β-3 adrenergic agonist for the treatment of overactive bladder (OAB) with symptoms of urge urinary
incontinence and urinary frequency. The main synthetic pathways leading to Mirabegron, reported as
patents, are summarized in Scheme 1. Whichever pathway is followed for the production of Mirabegron, a
key building block is represented by amine 3 (Scheme 1). Amine 3 can be targeted by reduction of amide 1
to amine 2 using BH3·THF in THF and 1,3-dimethyl-2-imidazolidinone (DMI) (Scheme 1, path a). [3]
Alternatively, other reducing agents, such as BH3·DMS, BH3·THF, NaBH4, NaBH4/BF3Et2O, LiBH4,
LiAlH4, in suitable solvents (alcohols, ethers and polar aprotic mixtures of solvents) have been proposed. [4]
As different approach to amine 3, the nucleophilic ring-opening of toxic and harmful stireneoxide (Scheme
1, path b), provides derivative 5 which in turn can be converted to derivative 3 upon reductive conditions.
Nevertheless, from a safety and green perspective, either the use of borane as reducing agent, or the use of
toxic chemicals (i.e. stireneoxide) should be carefully evaluated when planning an industrial production. As
far as safety and toxicity are concerns, it is worth pointing out that flow chemistry is emerging as a
green technology for handling of toxic and potentially hazardous chemicals. [5-7] In fact, in recent
years the "flow chemistry" and the development of new microfluidic devices are playing an
important role, both in academic and industrial research, as viable alternative to batch processing,
offering in many cases sustainable synthetic routes. [8-12] The continuous flow production attracted
the interest of chemical and pharmaceutical industry, ensuring lower costs, and effectively tackling
out safety concerns. [13-15] The use of continuous flow processing allowed also exploring reaction
conditions not accessible using traditional batch chemistry. [16] The high surface area in a
microflow reactor allows for very fast heat transfer, and precise temperature control, decreasing the
risk for runaway reactions. Another important point is the higher efficiency in mixing, realized in
flow microreactor compared to batch, thus eliminating or considerably reducing concentration
gradients that could negatively affect a fast reaction.
Based on these premises, we report in this work the development of a continuous flow process for the
preparation of intermediate 2 (Scheme 1) by reduction of (R)-2-hydroxy-N-[2-(4-nitrophenyl)ethyl]-2-
phenylacetamide 1 using BH3·DMS as reducing agent in a greener solvent such as 2-MeTHF. [17]
2. Material and methods
2.1 General (Standard techniques)
2

Nuclear magnetic resonance spectra were recorded on 300 and 500 MHz spectrometers. Chemical shifts for
1H NMR spectra were recorded in parts per million from tetramethylsilane with the solvent resonance as the
internal standard (chloroform, d = 7.26 ppm).
2.2 General (HPLC analysis)
Each sample has been diluted to final concentration of 0.1 mM in HPLC grade MeOH, filtered, and injected
for analysis. Samples were analyzed on Waters HPLC 1525 binary pump system, integrated with Water 2487
UV-Vis dual wavelength detector and an autosampler. Chromatographic conditions: UV detection, 1 = 290
nm; 2 = 270 nm; Column: Kinetex 5 C18 100A, 150  4.6 mm (Phenomenex), equipped with C18 pre-
column; column temperature: 25 °C; injection volume: 10 μL; flow rate 0.5 mL/min. Mobile phase
composition: linear gradient in 10 min, from 90% 10 mM Ammonium formate (AF) (pH 5)/10% MeOH v/v
to 30% 10 mM AF (pH 5)/70% MeOH v/v; then reset to start conditions in 5 min. Total analysis time 20
min.
2.3 Reagents
(R)-2-hydroxy-N-(4-nitrophenethyl)-2-phenylacetamide, BH3·THF 1M in THF, BH3·DMS 2M in THF,
BH3·DMS 1M in 2-MeTHF were commercially available (Sigma Aldrich, Merck, TCI Europe, Alfa Aesar).
THF and 2-MeTHF solutions of BH3·THF and BH3·DMS complexes were prepared from commercial
solutions and their concentrations were determined by 1H NMR analysis using BF3∙Et2O as internal standard.
[18]
2.4 Flow set up
All experiments were planned by Vapourtec Flow Commander Software in order to set-up solution
flow rates and residence time of each experiment (Table 3). The following solutions were prepared
and loaded in two different 2 mL PTFE loops: solution A: solution of (R)-2-hydroxy-N-(4-
nitrophenethyl)-2-phenylacetamide 1 (0.2M in dry 2-MeTHF) and solution B: solution of
BH3·DMS in dry 2MeTHF (Table 3, 0.6M, entry 7, 1M, entry 2), obtained by diluting commercially
available 2M solutions. The solutions were pumped in a 10 mL PTFE coil reactor at a flow rate of
125 µL/min (entry 7) at 90 °C with a residence time of 40 minutes, and at a flow rate of 167 µL/min
(Table 3, entry 2) at 90 °C with a residence time of 30 minutes. The resulting solution was collected
under steady state conditions, after 40 minutes and after 30 minutes (Table 3, entry 7 and 2
respectively). The output was quenched with H2O and extracted with ethyl acetate.
3. Results and Discussion
3

Inspired by our previous work concerning the safe handling of BH3 in a microfluidic reactor, we decided to
investigate the reduction step of amide 1 to amine 2 directly in a flow system. [19-27]
The first set of reactions were conducted in a flow reactor consisting of a 4 mL coil reactor, a 5 mL coil loop
fed with HPLC pumps. Table 1 collects the output of this investigation.
Initially, the effect of temperature and residence time was evaluated (Table 1, entries 1-3). Modest yields
were observed with higher residence time using BH3·DMS as the reducing agent. Similar results were
observed using BH3·THF as reducing agent (Table 1, entry 4), whereas increasing the temperature up to 60
°C (Table 1, entry 5) a doubling of the yield was obtained, highlighting a key role of the temperature.
However, using BH3·DMS at 60 °C, furnished only 25% yield of the desired amine 2 (Table 1, entry 6).
Interestingly, reintroducing the solution obtained under the conditions of entry 6 into the loop, and reacted
with additional 5 equivalents of BH3·DMS resulted in 70% yield of 2 (Table 1, entry 7). This last result
suggested that longer residence times and higher loading of reducing agent would provide better yields.
Eventually, temperature could be considered as critical parameter. Nevertheless, the decomposition of the
reducing agent need to be taken into account working at higher temperatures.
3.1 Analysis of the reduction process under batch conditions
Building on our previous 11B NMR experiments [18] on the stability of ethereal solutions of BH3 complexes,
we planned to test the reduction of 1 to 2 using BH3·DMS and BH3·THF complexes in batch conditions, and
with different solvents (Table 2). Under batch conditions, full conversion was observed in THF after 90 and
60 minutes using BH3·THF and BH3·DMS respectively (Table 2, entries 1, 2). According to our NMR,
evidences suggesting a higher stability of BH3 complexes in 2-MeTHF, [18] a faster reaction was observed
in this solvent with both BH3·THF and BH3·DMS complexes (Table 2, entries 3,4). Nicely, the use of the
more stable BH3·DMS complex in 2-MeTHF provide full conversion in 15 min reaction time.
With these evidences in hand, the reduction process was reinvestigated using a different flow reactor, and
different reaction conditions. In particular, 2-MeTHF was employed as the solvent for two main reasons: a)
its higher boiling point; b) the higher stability of the BH3 complexes. In addition, it is worth pointing out that
2-MeTHF is a greener alternative to other ethereal solvents; it has a higher boiling point with respect to THF
(bp 2-MeTHF = 80 °C; bp THF = 67 °C), a lower solubility in water (14 g per 100 g at 23 °C), and it is
obtained by renewable feedstocks. [28,29] The new flow set-up used two 2 mL loops, a 10 mL PTFE coil
reactor, and a back pressure regulator all connected to a Vapourtec system (Fig. 2). This flow system allowed
using temperatures close or even higher than the boiling point for the reaction solvent.
4

By using this new flow set-up (Table 3, entries 1, 2), high conversions were observed with both reducing
agents (87% with BH3·THF, and 92% with BH3·DMS respectively) even if the reaction with BH3·DMS
proceeded to full conversion using 2-MeTHF as the solvent, 5 equiv of BH3 complex, at 90 °C and with 30
minutes residence time. Interestingly, under these conditions, the product of double reduction 3 was also
observed though in very low yields (Table 3, entries 1, 2). Moreover, being derivative 2 just the precursor of
diamine 3, its presence in the reaction mixture wouldn’t represent a problem for further processing,
according to Scheme 1. In order to further improve the process, after optimizing the reaction conditions in
terms of temperature, solvent and reducing agent, we tackled the problem of the lowest number of
equivalents of BH3·DMS required for full conversion (Table 3, entries 3-8).
As reported in Table 3, reducing the amount of BH3·DMS from 4 equiv. to 2 equiv., while maintaining the
same residence time, resulted in lower conversion and lower yield of amine 2 (Table 3, entries 2-6). The best
compromise was obtained using 3 equiv of BH3·DMS with 40 minutes of residence time (Table 3, entry 7).
Under these conditions, a 94% yield of amine 2 was obtained jointly to 4.6% of diamine 3, so reaching
98.6% of useful compounds for further processing. Increasing the residence time, and reducing the amount
of BH3·DMS resulted in lower yield and conversion (Table 3, entry 8).
3.2 NMR monitoring of the flow reduction
The reaction mixtures obtained under flow conditions reported in Table 3, were analyzed by 1H NMR in
order to evaluate the amount of compounds 1, 2 and 3 in such mixture (Fig. 3). By 1H NMR analysis, clear
signals for the aromatic protons of 1, 2 and 3 could be recognized allowing for a quantitative evaluation.
According to Table 3, entries 2 and 7 gave the best results in terms of conversions and yields.
Moreover, in the case of entry 7, where a lower amount of reducing agent was used, 1.4% of starting material
could be detected, we found that acidic work up allowed for complete removal of amide 1 and full recover of
reduction products 2 (96.6%) and 3 (3.4%) (see Supplementary material). [30]
It is worth mentioning that such conclusions were confirmed by a HPLC quantitative analysis (see
Supplementary material).
4. Conclusions
In conclusion, a continuous-flow reduction of (R)-2-hydroxy-N-[2-(4-nitrophenyl)ethyl]-2-phenylacetamide,
involved in the synthetic pathway of Mirabegron, has been developed. This study demonstrated the
possibility to safely handly BH3 complexes within microfluidic reactors using 2-MeTHF as greener
alternative to traditional solvents, and without requiring any additive such as DMI. In addition, NMR and
5

HPLC purity analysis revealed that the sole by-product of this process is the diamine 3, which wouldn’t
affect the following steps towards Mirabegron.
Acknowledgements
We thank the University of Bari and Laboratorio SISTEMA, (Code PONa300369) financed by Italian Miur.
6

References
[1] Patent WO2010147631 A1.
[2] Patent WO200842898 A2.
[3] Patent WO 2003037881 A1.
[4] Patent WO 2014132270 A2.
[5] B. Gutmann, D. Cantillo, C. O. Kappe, Angew. Chem. Int. Ed. 54 (2015) pp. 6688-6729.
[6] R. Singh, H.-J. Lee, A. Kumar Singh, D.-P. Kim, Korean J. Chem. Eng. 33 (2016) pp. 2253-2267.
[7] M. B. Plutschack, B. Pieber, K. Gilmore, P. H. Seeberger, Chem. Rev. (2017) DOI
10.1021/acs.chemrev.7b00183.
[8] W. Reschetilowski, Microreactors in Preparative Chemistry Ed., Wiley-VCH: Weinheim, 2013.
[9] T. Wirth, Microreactors in Organic Synthesis and Catalysis, 2nd ed., Ed., Wiley-VCH: Weinheim, 2013.
[10] V. Hessel, J. C. Schouten, A. Renken, Y. Wang, J.-i. Yoshida, Handbook of Micro Reactors, Eds.,
Wiley-VCH: Weinheim, 2009.
[11] S. V. Luis, E. Garcia-Verdugo, Chemical Reactions and Processes under Flow Conditions, Eds., RSC
Green Chemistry, Royal Society of Chemistry: Cambridge, 2010.
[12] D. Dallinger, C. O. Kappe, Current Opinion in Green and Sustainable Chemistry 7 (2017) pp. 6-12.
[13] L. Pellegatti, J. Sedelmeier, Org. Process Res. Dev. 19 (2015) pp. 551-557.
[14] P. Poechlauer, J. Colberg, E. Fisher, M. Jansen, M. D. Johnson, S. G. Koenig, M. Lawler, T. Laporte, J.
Manley, B. Martin, A. O’Kearney-McMullan, Org. Process Res. Dev. 17 (2013) pp. 1472-1478.
[15] S. G. Newman, K. F. Jensen, Green Chem. 15 (2013) pp. 1456-1472.
[16] K. F. Jensen, AIChE J., 63 (2017) pp. 858-869.
[17] A. D. Mamuye, S. Monticelli, L. Castoldi, W. Holzera, V. Pace, Green Chem. 17 (2015) pp. 4194-4197.
[18] De Angelis, S., Celestini, P., Rebuzzini, G., Degennaro, L., Luisi, R. Synthesis 49 (2017) pp. 1969-
1971.
[19] L. Degennaro, M. Zenzola, P. Trinchera, L. Carroccia, A. Giovine, G. Romanazzi, A. Falcicchio, R.
Luisi, Chem. Commun. 50 (2014) pp. 1698-1700.
[20] M. Zenzola, L. Degennaro, P. Trinchera, L. Carroccia, A. Giovine, G. Romanazzi, P. Mastrorilli, R.
Rizzi, L. Pisano, R. Luisi, Chem. – Eur. J. 20 (2014) pp. 12190-12200.
[21] L. Degennaro, D. Maggiulli, C. Carlucci, F. Fanelli, G. Romanazzi, R. Luisi, Chem. Commun. 52
(2016) pp. 9554-9557.
[22] S. De Angelis, M. De Renzo, C. Carlucci, L. Degennaro, R. Luisi, Org. Biomol. Chem. 14 (2016) pp.
4304-4311.
[23] L. Degennaro, C. Carlucci, S. De Angelis, R. Luisi, J. Flow Chem. 6 (2016) pp. 136-166.
[24] L. Carroccia, B. Musio, L. Degennaro, G. Romanazzi, R. Luisi, J. Flow. Chem. 3 (2013) pp. 29-33.
[25] A. Giovine, B. Musio, L. Degennaro, A. Falcicchio, A. Nagaki, J.-i. Yoshida, R. Luisi, Chem. Eur. J. 19
(2013) pp 1872-1876.
7

[26] L. Degennaro, F. Fanelli, A. Giovine, R. Luisi, Adv. Synth. Catal. 357 (2015) pp 21-24.
[27] L. Degennaro, A. Tota, S. De Angelis, M. Andresini, C. Cardellicchio, M. A. Capozzi, G. Romanazzi,
R. Luisi Eur. J. Org. Chem. (2017) doi: 10.1002/ejoc.201700850.
[28] D. Prat, O. Pardigon, H.-W. Flemming, S. Letestu, V. Ducandas, P. Isnard, E. Guntrum, T. Senac, S.
Ruisseau, P. Cruciani, P Hosek, Org. Process Res. Dev. 17 (2013) pp. 1517-1525.
[29] D. Prat, A. Wells, J. Hayler, H. Sneddon, C. R. McElroy, S. Abou-Shehadad, P. J. Dunne, Green Chem.
18 (2016) pp. 288-296.
[30] Attempts to induce, under batch conditions, complete reduction of amine 2 to diamine 3 failed even
using a large excess of BH3 DMS.
8

O
O
N
H
OH
OH
OH
H
H
H
HO
HO
N
HO
HO
N
NH2
O
H3CO
O
OCH3
Adrenaline
Neurotransmitter
Noradrenaline
Neurotransmitter
(+)-hydrastine
alkaloid
(-) ephedrine
alkaloid
O
N
HN
Cl
Cl
O
N
HO
N
HO
N
O
N
O
O
O
N
N
O
O
N
Cl
N
(2S,3S)-cyproconazole
cereal antifungal
(2S, 4R)-ketoconazole
human antifungal
(R)-bambuterol
bronchodilator
Fig. 1. Pharmaceutical molecules containing the phenylethanolamine core.
Flow commander
software
HPLC pumps
2 x 2ml PTFE loops
10 ml PTFE coil
Fig. 2. Complete flow set-up Vapourtec
9

Fig. 3. 1H-NMR monitoring of the flow reduction
Scheme 1 Synthetic pathways leading to Mirabegron
10

Table 1 Optimization study of the reduction of 1 with BH3 in a continuous flow microreactor.
Reducing agent
[mol/L]
Reducing agent Flow rate a
T
tR
Yield b
(%)
14
Entry
(equiv)
(µL/min)
66.67
66.67
33.33
66.67
66.67
66.67
66.67
(°C)
25
50
25
25
60
60
60
(min)
30
1
2
3
4
5
6
7
BH3·DMS [1M]
BH3·DMS [1M]
BH3·DMS [1M]
BH3·THF [1M]
BH3·THF [1M]
BH3·DMS [1M]
BH3·DMS [1M]
5
5
30
14
5
60
22
5
30
20
5
30
38
5
30
25
10 (2 runs)c
60 (2 runs)c 70
c
a Flow rates for solutions of 1 and BH3. b Determined by 1H-NMR using the internal standard. The
solution resulting from entry 6 was reintroduced into the loop and reacted with additional 5 equiv.
of BH3·DMS with a tR of additional 30 min.
Table 2 Batch study of the reduction of amide 1 to amine 2.
Entry
BH3 Complex
BH3·THF
T (min)
90
Solvent
THF
1
2
3
4
60
THF
BH3·DMS
BH3·THF
30
2-MeTHF
2-MeTHF
15
BH3·DMS
Table 3 Optimization study of the reduction of 1 (0.2 M) in 2-MeTHF and BH3DMS (2M in THF)
11

BH3
BH3
Flow rate
(µL/min)
167
tR
(min) (%)a
1
2
3
(%)a
Entry
Complex
BH3·THF
BH3·DMS
BH3·DMS
BH3·DMS
BH3·DMS
BH3·DMS
BH3·DMS
BH3·DMS
(equiv)
(%)a
86.6
91.7
96
1
2
3
4
5
6
7
8
5
30
30
30
30
30
30
40
50
10.4
0
3
5
167
8.3
2
4
167
2
3
167
6.3
23
4.7
1.4
10
89.2
77
4.5
0
2
167
2.5
3
167
94
1.3
4.6
6
125
94
2.5
100
84
a Flow rate of 1 and BH3 Complex b Determined by 1H-NMR using internal standard
12