An improved and robust scale-up process aided with identification and control of critical process impurities in darunavir ethanolate

Article abstract of DOI:10.1007/s11164-019-03948-4

A robust and safe industrial process, including five isolations and drying steps for widely prescribed anti-HIV (protease inhibitor) drug darunavir ethanolate 2, has been developed. A salient feature of this process is the development of procedures enabling the efficient synthesis of multi-kilogram quantity of darunavir ethanolate, and process demonstrations through plant scale preparation are offered where darunavir molecule has been prepared with overall > 70% chemical yield and > 99.8% purity without involving any purification procedure(s), with all possible process impurities below than the desired limit (not more than 0.08%) were isolated, synthesized and characterized. The developed process is entirely robust, very efficient and demonstrated up to kilograms scale.

Full text of DOI:10.1007/s11164-019-03948-4

Research on Chemical Intermediates
https://doi.org/10.1007/s11164-019-03948-4
An improved and robust scale‑up process aided
with identifcation and control of critical process impurities
in darunavir ethanolate
Sathyanarayana Girigani1,2 · Harnam Singh1 · Sankar Rao Kola1 ·
Vijayalaxmi Dayanand Yelmeli1 · Venu Gopal Regula1 · Sakshi Shah1 ·
Neelu Jain2 · Pramod Kumar1
Received: 7 March 2019 / Accepted: 3 August 2019
© Springer Nature B.V. 2019
Abstract
A robust and safe industrial process, including fve isolations and drying steps for
widely prescribed anti-HIV (protease inhibitor) drug darunavir ethanolate 2, has
been developed. A salient feature of this process is the development of procedures
enabling the efcient synthesis of multi-kilogram quantity of darunavir ethanolate,
and process demonstrations through plant scale preparation are ofered where daru-
navir molecule has been prepared with overall >70% chemical yield and >99.8%
purity without involving any purifcation procedure(s), with all possible process
impurities below than the desired limit (not more than 0.08%) were isolated, synthe-
sized and characterized. The developed process is entirely robust, very efcient and
demonstrated up to kilograms scale.
Keywords Darunavir · Safe industrial process · Reduction · Deprotection ·
Coupling · Critical process impurities
Electronic supplementary material The online version of this article (https://doi.org/10.1007/s1116
4-019-03948-4) contains supplementary material, which is available to authorized users.
*
*
Harnam Singh
harnamsingh@microlabs.in
Pramod Kumar
pramodkumar@microlabs.in
Sathyanarayana Girigani
satyanarayana@microlabs.in
1
2
Chemical Research and Development, API R&D Centre, Micro Labs Ltd., Bommasandra-Jigani
Link Road, KIADB Industrial Area, Bangalore, Karnataka 560105, India
Sri Satya Sai University of Technology and Medical Sciences, Sehore, Madhya Pradesh 466001,
India
Vol.:(0123456789)
1 3

S. Girigani et al.
Fig. 1 A second-generation non-peptidic HIV protease inhibitors and raw materials
Introduction
As HIV is unrecoverable and it seems that after cancer, this is the greatest threat to
life, millions of AIDS patients from diferent age group are expecting from scientifc
community for better treatment.1 To fulfll this gap, chemists and biologists contin-
ued eforts have introduced several classes of potent drugs like highly active antiret-
roviral therapy (HAART) using more than three drugs ranging from HIV reverse
transcriptase to protease inhibitors brought about this remarkable breakthrough [1].
Among this efective blend, HIV protease inhibitors are so essential for HAART that
the Food and Drug Administration (FDA) has approved diferent protease inhibitors
and clinically used for the treatment of HIV/AIDS such as amprenavir, indinavir,
saquinavir, nelfnavir and ritonavir.2,3
Recently, FDA has approved few more HIV protease inhibitors including tiprana-
vir (1) and darunavir ethanolate (DAR 2) focused on drug-resistant HIV. The devel-
opment of tipranavir and darunavir ethanolate, a second-generation non-peptidic
HIV protease inhibitors [2], with marked improved resistance profles, has opened
a new perspective on the treatment of antiretroviral therapy (ART) [3] experienced
HIV patients with poor viral load control (Fig. 1).
Whereas, DAR 2 (brand name Prezista, formerly known as TMC114) which is in
second-generation protease inhibitor class has been fully recommended by Ofce of
AIDS Research Advisory Council (OARAC) for the treatment of HIV in adults and
children 6 years of age and older [4]. It was approved by the FDA on June 23, 2006
to overcome the problems with early protease inhibitor like severe side efects and
drug toxicities [5], require a high therapeutic dose, expensive to aford and show a
disturbing susceptibility to drug-resistant mutations; a second-generation protease
1
http://www.cancer.org/acs/groups/cid/documents/webcontent/002295-pdf.pdf.
2
https://aidsinfo.nih.gov/education-materials/fact-sheets/21/58/fda-approved-hiv-medicines.
3
http://emedicine.medscape.com/article/1533218-overview#a2.
1 3

An improved and robust scale-up process aided with…
inhibitor DAR 2 is discovered. Furthermore, DAR 2 is the frst drug which is more
efcacious and cost-efective than other protease inhibitors.
A strong demand for DAR 2 led to the development of several synthetic schemes
for its synthesis [6–13] involving the transformation of tert-Butyl[(S)-1-((R)-oxiran-
2-yl)-2-phenylethyl]carbamate 3 as starting material in most of the synthetic proto-
cols to DAR 2, albeit involving multistep procedures using reagents such as palla-
dium, sodium azide and trifuoroacetic acid (TFA) which are known to be potentially
hazardous at the commercial scales besides being environmentally degrading along
with having process drawbacks such as tedious workup, excessive efuent genera-
tion, low isolated yields, arduous isolation/purifcation procedures, formation of
stagnant impurities/by-products that are stif to eliminate even after multiple puri-
fcation. Search for a practically simpler cost-efective method led Tibotec Pharma-
ceuticals Limited [14] to design a protocol by involving existing 3 as the key starting
material, employing simple addition, substitution, reduction, deprotection and cou-
pling strategies as key steps in the construction of the darunavir skeleton.
The scale-up process for the this protocol has been achieved by converting 3 into
4-amino-N-((2R,3R)-3-amino-2-hydroxy-4-phenylbutyl)-N-isobutylbenzene sulfon-
amide 9 in multistep procedures by involving the use of reagents such as palladium
metal and concentrated hydrochloric acid, and thus, it was made to participate in a
coupling with 4, and providing DAR 2 in two further steps. Although this protocol
provides ready access to 2 via isolated four to fve intermediates, it is too long in
addition to have some environmental and toxicological drawbacks. The major dis-
advantages of this route are the use of potentially hazardous reagents in addition to
generation of large amounts of efuents besides having low isolated yields and high
production cost.
A perusal of these synthetic processes disclosed in the existing studies [6–14]
made it clear that the only one synthetic route [14] could result into the synthesis of
DAR 2 successfully at commercial level till now. Most of these synthetic protocols
have made their academic value instead of commercial aspects besides involving use
of fancy reagents, more efuent generation along with low isolated yield. The estab-
lished commercial viable protocol has also some major disadvantages such as use
of potentially hazardous reagents [6], generation of large amounts of efuents along
with the formation of process/degradation impurities [7], which could not be washed
out in purifcations at mid or at the end of the process, struggling to aford DAR 2
with stringent quality profle (any unspecifed impurity: not more than 0.08% w/w)
as API [15].
Therefore, evaluation of these reported synthetic strategies for DAR 2 prepara-
tion besides understanding of their advantages/disadvantages in consideration of
commercial aspect such as process robustness and low cost driven us to work on
the synthetic strategy disclosed in US Patent 6,248,775 B1 [14] and 7,700,645 B2
[16]. Though route is an improved one, still more improvements are required to get
the right quality of the API consistently. Thus, process was studied thoroughly dur-
ing the laboratory development stage and introduced some major modifcations to
encounter the shortcomings of the listed process [15, 19] for making DAR 2 pro-
cess plant friendly and commercially viable. We envisioned that an efcient route
with optimized process to the robust level for DAR 2 without involving hazardous
1 3

S. Girigani et al.
reagents should make economic sense besides giving a new practical industrial
method within the acceptable environmental and toxicological concerns [17]. We
further justifed that the new method would be more appealing if it can employ the
same commercially available starting material; 3 as listed in most of the synthetic
routes. In this contribution, we report the design and technical aspects of our new
improved process, which enables the commercial production of 2 from 3.
Experimental
Materials
All materials were purchased from commercial suppliers. Unless specifed oth-
erwise, all reagents and solvents were used as supplied by manufacturers. Melting
points were determined by open air capillary with Buchi M-565 and are uncor-
rected. IR spectra of samples were recorded on Shimadzu IR Afnity-I FT-IR
1
spectrophotometer. H NMR spectra (400 MHz) and 13C NMR spectra (100 MHz)
were recorded in CDCl3, DMSO-d6, on Bruker 400 MHz NMR (ASCEND, 5 mm
PABBO) instrument, and mass spectra were determined on Velos Pro ion trap mass
spectrophotometer (Thermo Scientifc). TLC was done on Merck TLC silica gel
60F254 plates using mobile phase MeOH/DCM (8:2 and 9:1).
Synthetic procedure
Preparation of tert‑butyl((1S,2R)‑1‑benzyl‑2‑hydroxy‑3‑[isobutyl[(‑4‑nitrophenyl)
sulfonyl]‑amino]propyl]carbamate (7)
A mixture of isobutyl amine (1000 mL, 9.981 mol), IPA (200 mL, 1.0 volume) and
200 g (0.759 mol) of 3 was stirred for 12–14 h at 20–30 °C. After completion of
reaction (TLC), volatiles were removed under vacuum. Water (1000 mL) was added
to the residue followed by stirring for 2 h, and subsequent fltration resulted to aford
intermediate 6, after drying at 50–60 °C in air oven over a period of 10–12 h. Dried
solid was then mixed with DCM (740 mL), TEA (115 mL, 0.835 mol) at 15–25 °C
followed by slow addition of 4-nitrobenzenesulfonyl chloride solution (168 g,
0.759 mol, in 300 mL of DCM) at 15–25 °C. Resulted reaction mixture was stirred
for 1–2 h at 15–25 °C. After reaction completion (TLC), volatiles were evapo-
rated under vacuum completely. Water (1500 mL) and sodium hydroxide solution
(200 mL; 0.700 mol) was added to distilled residue followed by stirring for 2 h. The
resulting aqueous slurry was fltered, washed with water and dried in air oven at
55–60 °C over a period of 10–12 h to yield the intermediate 7. (367 g, 93%): Rf
1
(3–6=0.50 and 6–7=0.80); H NMR (400 MHz, DMSO-d6) δ 0.81–0.85 (m, 6H,
–2CH3), 1.24 (s, 9H, –3CH3), 1.94–1.99 (m, 1H, –CH), 2.91–2.96 (m, 2H, –CH2),
3.04–3.17 and 3.42–3.50 (m, 5H, –2CH2, –CH(OH)), 4.93 (d, J=6.4 Hz, 1H, –CH),
6.64 (d, J=8.8 Hz, 1H, –NH), 7.24–7.11 (m, 5H, arom-H), 8.04 (d, J=8.4 Hz, 2H,
arom-H), 8.36 (d, J=8.4 Hz, 2H, arom-H); 13C NMR (100 MHz, DMSO-d6) δ 19.5,
1 3

An improved and robust scale-up process aided with…
19.7, 25.7, 28.0, 35.3, 51.1, 55.0, 71.1, 77.4, 124.2, 125.5, 127.7, 128.4, 129.0,
139.3, 145.4, 149.3, 155.1; IR (KBr, cm−1) 3562, 3363, 2960, 1697; MS (ES+) m/z:
522.23 (MH+).
Preparation
of 4‑Amino‑N‑((2R,3S)‑3‑amino‑2‑hydroxy‑4‑phenylbutyl)‑N‑isobutylbenzene
sulfonamide (9)
Mixture of 7 (300 g, 0.575 mol), wet Raney nickel (60 g, 20% w/w) and metha-
nol (1500 mL) was taken in stainless steel autoclave at 20–25 °C under hydrogen
gas pressure (5.0 kg). Resulted reaction slurry was stirred for 3–4 h. After reaction
completion (TLC), the reaction mixture was fltered through hyfo bed and washed
with methanol (300 mL). Filtrate was evaporated under vacuum followed by degas-
sing for 30 min at 40–45 °C. Ethyl acetate (600 mL) and aqueous HCl (254 mL,
2.299 mol; 30–35%) was added to the distilled residue and stirred at 20–25 °C for
over a period of 4–5 h. After reaction completion (TLC), reaction mass was fltered
and wet solid was suck dried for 30–40 min. A mixture of suck dried solid and water
(2400 mL) was treated with sodium hydroxide solution (1500 mL, 1.725 mol) at
10–20 °C over a period of 2–3 h. The resulting heterogeneous slurry was fltered,
washed with water and dried in air oven for 8–10 h at 45–50 °C to aford intermedi-
ate 9. (192 g, 85.3%): Rf (7–8=0.32 and 8–9=0.21); 1H NMR (400 MHz, CDCl3) δ
0.92 (dd, J=8 Hz and 12 Hz, 6H, –2CH3), 1.85–1.90 (m, 1H, –CH), 2.44–2.52 (m,
2H, –CH2), 2.78–2.85 (m, 1H, –CH2), 2.98–3.02 (m, 2H, –2CH2), 3.09–3.19 (m;
2H, –CH2), 3.24–3.32 (m, 1H, –CH(NH2)), 3.55 (br-s, 1H, –OH), 3.70–3.75 (m, 1H,
–CH), 4.17(s, 2H, arom-NH2), 6.68 (d, J=4.0 Hz, 2H, -arom-H), 7.20–7.34 (m, 5H,
arom-H), 7.60 (d, J=8.0 Hz, 2H, arom-H); 13C NMR (100 MHz, CDCl3) δ 19.9,
20.2, 27.2, 39.4, 52.7, 55.7, 58.7, 73.3, 114.1, 126.4, 126.5, 128.6, 129.3, 129.5,
138.9, 150.6; IR (KBr, cm−1) 3452, 3348, 3240, 2951, 1649, 1598, 1321, 1147; MS
(ES+) m/z: 392.20 (MH+).
Preparation of darunavir (10)
Mixture of 9 (100 g, 0.254 mol) and DMSO (300 mL) was cooled to 15 °C followed
by addition of 5 as solution (72.7 g; 0.266 mol in 300 mL of DMSO) at 15–30 °C.
DMSO reaction mass was thus stirred for 1–2 h. After reaction completion (TLC),
ammonium hydroxide (2.1 g, 0.16 mol) was added to reaction mass followed by
water addition and subsequent stirring for 1.0 h. Aqueous slurry was thus fltered,
washed with water and treated wet solid with IPA to aford, after drying for 10–12 h
at 50–55 °C under vacuum 10 having HPLC Purity >99.80% with dimer-2 impurity
13 and oxazolidinone impurity 14 to the level of not more than 0.08% (for char-
acterization data of these, see Supporting Information): (132 g, 94%), Rf=0.85;
1
mp 72–74 °C; H NMR (400 MHz, DMSO-d6) δ 0.78 (d, J=6.8 Hz, 3H, –CH3),
0.84 (d, J=6.4 Hz, 3H, –CH3), 1.03 (d, 3H, IPA-CH3), 1.26 (m, 1H, –CH2), 1.40
(m, 1H, –CH2), 1.93 (m, 1H, –CH), 2.46–2.49 (m, 2H, –CH2), 2.52–2.68 (m, 2H,
–CH2), 2.75–2.77 (m, 1H, –CH2), 2.93–3.04 (m, 2H, –CH), 3.26–3.28 (m, 1H,
1 3

S. Girigani et al.
–CH), 3.56–3.62 (m, 4H, –CH2), 3.63–3.72 (m, 1H, –CH–OH), 4.30 (d, 1H, IPA),
4.83–4.84 (m, 1H, –CH–O), 4.94 (d, J=6.0 Hz, 1H, –CH), 5.50 (d, J=5.2 Hz, 1H,
–CH)), 5.94 (s, 2H, arom-NH2), 6.59 (d, J=8.8 Hz, 2H, arom-H), 7.12–7.13 (m,
1H, –CONH), 7.19–7.23 (m, 5H, arom-H), 7.37 (d, J=8.4 Hz, 2H, arom-H); 13C
NMR (100 MHz, DMSO-d6) δ 20.0, 25.4, 25.5, 26.3, 35.1, 45.0, 52.7, 55.8, 57.2,
62.0, 68.8, 70.3, 72.3, 72.7, 108.7, 112.6, 123.6, 125.6, 127.8, 128.9, 129.1, 139.4,
152.6, 155.1; IR (KBr, cm−1) 3419, 3344, 2962, 2904, 1705; MS (ES+) m/z: 548.27
(MH+).
Preparation of darunavir ethanolate (2)
Mixture of 10 (120 g, 0.219 mol) and ethanol (700 mL) was heated to 75–80 °C to
make it clear solution. Ethanolic clear solution thus obtained was cooled gradually to
0–5 °C and stirred for 1–2 h followed by fltration and washing of fltered solid with
chilled ethanol, dried for 10–12 h at 50–55 °C under vacuum to aford the targeted
darunavir ethanolate API (2) having HPLC Purity>99.80% with dime-2 impurity
13 and oxazolidinone impurity 19 to the level of not more than 0.08% (for character-
ization data of these, see Supporting Information): (117 g, 90%), mp 99–101 °C; 1H
NMR (400 MHz, DMSO-d6) δ 0.78 (d, J=6.8 Hz, 3H, –CH3), 0.84 (d, J=6.4 Hz,
3H, –CH3), 1.05 (t, J=6.8, 3H, Ethanol-CH3), 1.26 (m, 1H, –CH2), 1.42 (m, 1H,
–CH2), 1.83 (m, 1H, –CH2), 2.46–2.65 (m, 4H, –CH2 and –CH) 2.93–3.01 and 2.91
(m, 2H, –CH2), 3.26–3.45 (m, 3H, Ethanol-CH2, and –CH2), 3.56–3.63 (m, 4H,
–2CH2), 3.71–3.372 (m, 1H, –CH), 3.84–3.88 (m, 1H, –CH), 4.31 (t, J=4.8 Hz, 1H,
Ethanol-OH), 4.84 (d, J=8.0 Hz, 1H, –CH), 4.94 (d, J=6.4 Hz, 1H, –CHOH), 5.51
(d, J=5.2 Hz, 1H, –CH), 5.94 (s, 2H, arom-NH2), 6.60 (d, J=8.8 Hz, 2H, arom-H),
7.12–7.23 (m, 6H, –NH and arom-H), 7.38 (d, 2H, arom-H); 13C NMR (100 MHz,
CDCl3) δ 18.4, 20.0, 25.4, 26.3, 35.1, 45.0, 52.7, 55.8, 56.0, 57.3, 68.8, 70.3, 72.3,
72.8, 108.8, 112.6, 123.7, 125.6, 127.8, 128.9, 129.2, 139.4, 152.6, 155.1; IR (KBr,
cm−1) 3466, 3437, 3360, 2964, 1707, 1597, 1313; MS (ES+) m/z: 592.18 (MH−).
tert‑Butyl[(1S,2R)‑1‑benzyl‑2‑hydroxy‑3‑(isobutylamino)
propyl]‑N‑[(1S,2R)‑1‑benzyl‑2‑hydroxypropyl]carbamate (Dimer‑1 impurity; 11)
1H NMR (400 MHz, DMSO-d6) δ 0.87 (dd, J=6.4 Hz and 17.2 Hz, 3H, –CH3), 1.25
(s, 18H, –C(CH3)3), 1.68–1.76 (m, 1H, –CH), 2.18–2.20 (m, 2H, –CH2), 2.50–2.57
(m, 4H, –CH2), 3.03 (d, 2H, –CH2), 3.49 (br-s, 4H, –CH2 and –CH), 4.68 (s, 2H,
–NH), 6.63 (d, 2H, –NH), 7.12–7.25 (m, 10H, arom-H); 13C NMR (100 MHz,
DMSO-d6) δ 20.7, 20.9, 25.8, 28.1, 36.1, 55.0, 58.7, 63.9, 70.3, 77.2, 125.5, 127.8,
129.1, 139.6, 155.2; IR (KBr, cm−1): 3356, 1683, 1529; MS (ES+) m/z: 600.65
(MH+).
tert‑Butyl‑N‑[(2S,3R)‑4‑[(4‑(N‑hydroxylamino)phenyl)sulfonyl‑(2‑methylpropyl)
amino]‑3‑hydroxy‑1‑phenylbutan‑2‑yl‑]carbamate (N‑hydroxyl impurity; 12)
1H NMR (400 MHz, DMSO-d6) δ 0.82 (dd, J=7.2 Hz and 3.6 Hz, 6H, –CH3), 1.40
(s, 9H, –C(CH3)3), 1.91–2.08 (m, 1H, –CH), 2.67–2.77 (m, 2H, –CH2), 2.91–3.00
1 3

An improved and robust scale-up process aided with…
(m, 2H, –CH2), 3.30 (br-s, 1H, –CH2), 3.48–3.52 (m, 1H, –CH2), 3.61 (d, J=5.6 Hz,
1H, CHOH), 4.94 (d, J=6.4 Hz, 1H, –CH), 6.68 (d, J=9.2 Hz, 1H, –NH), 6.87 (d,
J=8.8 Hz, 2H, arom-H), 7.14 (d, J=6.8 Hz, 1H, arom-H), 7.18–7.24 (m, 4H, -arom-
H), 7.55 (d, J=8.8 Hz, 2H, arom-H), 8.66 (s, 1H, NHOH), 8.97 (s, 1H, –NHOH);
13C NMR (100 MHz, DMSO-d6) δ 19.9, 20.0, 26.2, 28.1, 35.2, 52.4, 55.1, 56.7,
72.4, 77.3, 111.3, 125.5, 127.6, 127.8, 128.3, 129.1, 139.6, 155.2; IR (KBr, cm−1):
3363, 1693, 1598, 1332; MS (ES+) m/z: 508.09 (MH+).
N‑[(1S,2R)‑3‑[[(4‑Amino(3R,3aS,6aR)‑hexahydrofuro(2,3‑b)‑furan‑3‑yl‑oxycarbonyl]
(phenyl)sulfonyl](2‑methylpropyl)amino]‑2‑hydroxy‑1‑(phenylmethyl)propyl]
carbamic acid(3R,3aS,6aR)‑hexahydrofuro[2,3‑b]‑furan‑3‑yl ester (dimer‑2 impurity;
13)
1H NMR (400 MHz, DMSO-d6) δ 0.82 (d, J=8.0 Hz, 6H, –C(CH3)3), 1.20 (m,
1H, –CH2), 1.40 (m, 1H, –CH2), 1.80 (m, 1H, –CH2), 1.94–1.97 (m, 2H, –CH2 and
–CH), 2.74–2.79 (m, 3H, –CH2), 2.99–3.06 (m, 3H, –CH2 and –CH), 3.27 (m, 1H,
–CH), 3.56–3.61 (m, 4H, –CH2 and –CH), 3.69–3.73 (m, 2H, –CH2), 3.85–3.87
(m, 3H, –CH2 and –CH), 3.96–4.00 (m, 1H, –CH2), 4.82–4.84 (m, 1H, –CH2),
5.00 (d, J=8.0 Hz, 1H, –OH), 5.16 (dd, J=8 Hz, 1H, –CH), 5.50 (d, J=5.6 Hz,
1H, –CH), 5.63 (d, J=5.2 Hz, 1H, –CH), 7.12–7.14 (m, 1H, –NH), 7.19–7.25 (m,
5H, arom-H), 7.66 (dd, J=8 Hz and 24 Hz, 4H, arom-H), 10.22 (s, 1H, –NH); 13C
NMR (100 MHz, DMSO-d6) δ 25.5, 26.7, 35.1, 44.6, 44.9, 52.2, 55.8, 56.6, 68.6,
70.0, 70.2, 72.3, 73.3, 108,6, 108.7, 117.7, 125.6, 127.7, 128.2, 129.0, 132.2, 139.3,
142.6, 152.6, 155.0; IR (KBr, cm−1): 3342, 2962, 2875, 1718, 1595, 1315; MS
(ES+) m/z: 704.40 (MH+).
4‑Amino‑N‑(2‑methylpropyl)‑N‑[[(4S,5R)‑2‑oxo‑4‑(phenylmethyl)‑5‑oxazolidinyl]
methyl] benzenesulfonamide (oxazolidinone impurity; 14)
1H NMR (400 MHz, CDCl3) δ 0.91 (d, J=8.0 Hz, 6H, –C(CH3)3), 1.89–193 (m,
1H, –CH), 2.60–2.65 (m, 1H, –CH2), 2.84–2.97 (m, 2H, –CH2), 3.06–3.17 (m,
1H, –CH2), 3.25–3.34 (m, 1H, –CH2), 3.74–3.85 (m, 1H, –CH2), 4.06–4.10 (m,
1H, –O–CH), 4.78–4.87 (m, 1H, –NH–CH), 5.12 (s, 1H, –NH), 7.05–7.07 (m, 2H,
arom-H), 7.14–7.18 (m, 2H, arom-H), 7.21–7.24 (m, 1H, arom-H), 7.36–7.37 (m,
2H, arom-H), 7.65 (d, 2H, –arom-H); 13C NMR (100 MHz, CDCl3) δ 19.9, 20.0,
26.7, 35.9, 48.0, 55.9, 57.0, 78.8, 114.8, 127.3, 128.0, 128.8, 129.1, 129.3, 136.1,
149.6, 157.8; IR (KBr, cm−1): 3367, 2960, 2929, 1753, 1317; MS (ES+) m/z: 418.51
(MH+).
Results and discussion
It was apparent that, in order to achieve an efcient synthesis, the best option
would be to fnd a suitable reaction conditions that can functionalize epoxide ring
of the 3 in such a way so as to install an isobutyl amine moiety in the desired
position at room temperature, instead of heating to refux which is supposed to
1 3

S. Girigani et al.
Scheme 1 Improved and robust process16 for the preparation of DAR (2) starting from 3. Reagents
and conditions: (a) isobutyl amine/IPA, 20–30 °C; (b) DCM, TEA, 4-nitrobenzenesulfonyl chloride,
15–30 °C; (c) MeOH, Raney Ni, H2 gas, 20–30 °C; (d) aqueous HCl, EtOAc, 20–30 °C; (e) water, NaOH
solution, 20–30 °C; (f) DMSO, 2,5-dioxopyrrolidin-1-yl((3R,3aS,6aR)-hexahydrofuro[2,3-b]furan-3-yl)
carbonate, 15–30 °C; (g) EtOH, 75–80 °C
result into dimer-1 impurity; 11 formation (which is not disclosed till now and
made it know while we developed the process), followed by substitution of sec-
ondary amine proton of 6 with 4-nitrobenzenesulfonyl chloride and successive
reduction of nitro group to amine without use of palladium metal besides hav-
ing the process control for formation of N-hydroxyl impurity; 12 below than the
desired limit (Scheme 1).
Next, synthetic sequence in the formation of DAR 2 is deprotection of N-Boc
group which would be more appealing if performed in mild acidic condition at room
temperature (RT) rather than refuxing it in strong acidic solution followed by cou-
pling of 4 to aliphatic amine of compound 9, without formation of critical process
impurities such as dimer-2 impurity; 13 and oxazolidinone impurity; 14, which were
not exterminating by any of the purifcation methods, if formed, leading to the DAR
2 after treating it with ethanol (Fig. 2).
As described in Scheme 1, our synthesis of 2 begins with the addition of the
isobutyl amine moiety to epoxide ring of 3 with excess of isobutyl amine reagent
(13.15 mol. equiv.) along with IPA (1.0 volume) for 12–14 h at 20–30 °C. The solu-
bility of starting material is slight poor in isobutyl amine, thus warranting the use of
1.0 T of IPA which increases the solubility of starting material. It was found that, at
this concentration, a clear solution results as the reaction approaches toward comple-
tion followed by slow precipitation of desired compound 6. TLC examination of this
clear solution indicates a clean reaction with partial formation of 6. A time study of
this reaction carried out shows that the reaction is incomplete below 10 h, whereas
at a time range of 12–14 h the reaction is completed when temperature is 20–30 °C.
After the completion of the reaction (TLC), the volatiles are removed under vacuum
followed by chasing of left over solvents traces with water. The resulting viscous
material thus obtained is stirred with water at 20–30 °C and afording in separation
of solid 6 as crude, which on treatment with IPA at 40–50 °C results to yield 6 in
80–85%. However, detailed study was performed to have a suitable solvent selection
1 3

An improved and robust scale-up process aided with…
O
NH
NHOH
O
HO
N
O
O
O
O
S
O
HN
O
N
H
N
OH
OH
N-Hydroxyl impurity (12)
Process impuritiy at stage 6 to 7
Dimer-1 impurity (11)
Process impurity at stage 3 to 6
O
H
O
O2
O
S
O
N
H
and
HN
N
N
H
OH
O
SO2
O
H
NH
O
O
O
O
H2N
Oxazolidinone impurity (14)
Critical process Impurities at stage 9 to 10
H
Dimer-2 Impurity (13)
Fig. 2 Structures of process impurities formed in stage 3–6, 6–7 and 9–10 in the process
Table 1 Solvent selection for the formation of 6
S. No. Solvent
Isobutyl
Yield (%) Dimer-1
Remarks
amine
impurity
(%)
(mol. equiv.)
1.
2.
3.
1-Propanol
12–14
12–14
12–14
94.00
80.00
–
~10–15
~13–18
~10–15
Impurity formation and incomplete
reaction
Toluene
Acetone
Impurity formation and incomplete
reaction
Impurity formation and incomplete
reaction
4.
5.
DMSO
MEK
12–14
12–14
78.00
–
~2–4
~10–15
Incomplete reaction
Impurity formation and incomplete
reaction
6.
DMF
12–14
91
~10–15
Impurity formation and incomplete
reaction
7.
8.
IPA
DMA
12–14
12–14
94.67
88
~1.09
~10–15
–
Impurity formation and incomplete
reaction
9.
Isobutyl amine11 24–26
93
1.3
–
1 3

S. Girigani et al.
Table 2 Screening of solvent for the formation of 7
S. No.
Solvent
4-Nitrobenzenesulfonyl Base
chloride (mol. equiv.)
Yield (%)
Remarks
1.
2.
3.
4.
5.
6.
7.
8.
Acetone11
Acetone
EtOAc
Toluene
DCM
1.1
1.1
1.1
1.1
1.1
1.1
1.1
1.1
NaHCO3
55–60
80–85
80–85
70–75
93–96
90–95
85–90
55–60
Incomplete reaction
Thick reaction mass
Thick reaction mass
Thick reaction mass
Clear reaction mass
Thick reaction mass
Free fowing reaction mass
Clear reaction mass
TEA
TEA
TEA
TEA
TEA
TEA
TEA
MEK
MIBK18
DMF
along with quantity of isobutyl amine as a reagent for the formation of 6. The results
are summarized in Table 1.
After having the selection study for desired solvent, dimer-1 impurity; 11 forma-
tion (~10–15%) was observed on TLC in most of the solvents, while the formation
of 6 was executed. The best results could be obtained when IPA was used as solvent
in combination with isobutyl amine at proposed reaction conditions or when reac-
tion was performed only in isobutyl amine at 20–30 °C for around 12–14 h resulting
to aford >80% yield of compound 6.
Thereafter, considering our next strategy for substitution of secondary amine pro-
ton of 6 with 4-nitrobenzenesulfonyl chloride, we attempted a reaction of 6 with
listed procedures such as using inorganic base (NaHCO3, Na2CO3) and acetone/or
DCM [13] as solvent to obtain an intermediate 7. However, a mixture of 6, 7 and
4-nitrobenzenesulfonic acid was obtained due to degradation of 4-nitrobenzenesul-
fonyl chloride in aqueous basic reaction medium. Reaction of 6 with 4-nitrobenze-
nesulfonyl chloride in acetone and organic base like TEA was the next to be tried,
and it resulted in isolation of the expected 7 in 80–85% yield. Although the yield of
7 in this case was satisfactory, the reproducibility of yield as well as quality was not
up to the mark. Therefore, we tried this substitution reaction in diferent solvents by
using preferable organic base (TEA) to have a better reaction profle with desired
yield and quality (Table 2).
Surprisingly, in most of the solvents along with TEA as base, reaction was either
resulting to form additional by-products or resulting to have a thick reaction mass
which was ofering constraints in completion of reaction conversion as well as in fl-
tration. The best results could be achieved when reaction for the formation of 7 was
executed in DCM/or MIBK as a reaction solvent; however, DCM was preferred over
MIBK due to having low boiling point which could be easily evaporated after reac-
tion completion for making isolation of 7 as solid with ease in 90% of isolated yield.
After having the detailed study for the formation of 6 and 7 separately, it was
considered that if both stages could be performed in one pot without isolation of
6 would be real assets in the way for manufacturing the robust process for DAR 2.
Therefore, after a few trials, we were able to perform these two chemical transfor-
mations such as isobutyl amine addition and substitution of aminic proton of 6 in the
1 3

An improved and robust scale-up process aided with…
Table 3 Screening of reducing catalyst for the formation of 8
S. No.
Reducing reagents
8 (%)
7 (%)
Remarks
1.
2.
3.
4.
5.
6.
Zn/NH4Cl
Zn/HCOONH4
Zn/NH2NH2
Fe/CH3COOH
Pd/C
~85
~80
~50
~70
98.91
98.94
~7–10
~10–15
~40–50
~20
ND
ND
Incomplete reaction and impurities formation
Incomplete reaction and impurities formation
Incomplete reaction and impurities formation
Incomplete reaction and impurities formation
–
–
Raney Ni
synthetic sequence for preparation of compound 7 in tandem without confrming and
analyzing the intermediate 6. Thus, volatiles are removed under vacuum from the
reaction mixture after the reaction completion of 6 followed by fltration and drying
of fltered solid, without any purifcation and analysis, is taken up in DCM and TEA
and treated with 4-nitrobenzenesulfonyl chloride solution in 2–3 volumes of DCM
for 1–2 h at 15–30 °C to aford, after fltration, 7 in 88–93% overall yield from 3.
Combining these two steps ofers signifcant advantages, as both steps necessitate
the use of a single reactor at the plant level, thus lowering the reaction time and util-
ity cost besides saving the costs involved in purifcation and analysis of 6. Moreover,
the overall yield from 3–7 is about 15–20% higher by a one-pot procedure.
The formation of compounds 8 from 7 was achieved by treating it with Raney
nickel in methanol and stirring the resulting mixture under hydrogen gas pressure
for 3–4 h at 20–40 °C. As the reaction proceeds, the temperature keeps on increas-
ing and is controlled at 35–40 °C till the end of the reaction. However, several con-
ventional reducing agents were tried, which are reported for reduction of nitro to
amine in literature such as Zn/NH4Cl [18], Zn/HCOONH4 [19], Zn/NH2NH2 [20],
Fe/CH3COOH [21] Pd/C [13, 22, 23] and Raney nickel [24, 25] during laboratory
development for achieving the formation of 8 without/or minimum formation of by-
products along with complete conversion of 7–8 (Table 3).
As noticeable, reduction of nitro to amine was either incomplete or form unde-
sired by-products when reagent(s), other than Pd/C and Raney nickel, were tried
for the formation of 8. Best results for reducing nitro to amine could be obtained,
while reaction was performed either with wet Pd/C (10%) or Raney nickel along
with methanol. Considering very low cost of Raney nickel over Pd/C was selected as
reagent of choice for the formation of 8. As a result, the metallic catalyst is removed
by fltration after reaction completion and is washed with methanol. Excess metha-
nol is evaporated, and the resulting thick oily mass is dissolved in ethyl acetate at
20–30 °C for next chemical transformation. Being a thick oily nature of 8, isolation
of this intermediate was avoided and preferred in situ conversion of 8–9 without fur-
ther any detailed study. After a few tryouts, we were able to perform these two steps
such as reduction and the N-Boc deprotection without confrming and analyzing the
intermediate 8.
Initial investigation for confrming N-Boc deprotection, after having a thick oily
residue of 8, was initiated with known conventional reagent like TFA [26, 27] in
1 3

S. Girigani et al.
Table 4 Screening of acid for the formation of 9
S. No.
Acid
Solvent
8 (%)
9 (%)
~80
Yield (%)
70
Remarks
1.
TFA
DCM
~7–10
Incomplete reaction
and impurities
formation
2.
Conc. HCl
EtOAc
ND
99.88
88
–
DCM as solvent led us with incomplete reaction besides formation of several by-
products which were hard to wipe out in the next stage(s). Understanding this, we
attempted N-Boc deprotection with easily available commercial reagent, i.e., aque-
ous hydrochloric acid solution (30–35%) [28] along with ethyl acetate as reaction
solvent (Table 4).
Results were quite impressive when 8 was treated with aqueous HCl solution for
the formation of 9 with ethyl acetate in place of TFA/DCM reagent/solvent system.
Based on quality, experimental output, process and handling ease, aqueous HCl as
deprotecting reagent was selected for preparation of 9 and resulted in 85–90% yield
by a one-pot procedure. Combining these two steps ofers signifcant advantages, as
both steps necessitate the use of a single reactor at the plant level, thus lowering the
reaction time and utility cost besides saving the costs involved in isolation of 8.
Final step of DAR 2 preparation involves the condensation of key pharmaceu-
tically active candidate (3R,3aS,6aR)-hexahydrofuro[2,3-b]furan-3-ol (4) with 9 to
structure darunavir (10) skeleton, was achieved by treating 9 with 5, derivative of
4, in diferent solvents, in presence/or absence of base [29, 30]. However, during
process development, we encountered the formation of very critical process impuri-
ties such as dimer-2 impurity; 13 [32] and oxazolidinone impurity; 14 (which is not
disclosed in reported routes till now and made it know, while we developed the pro-
cess), which were tightly entangled to the DAR 2 and were very hard to remove/or
minimize/or control even after multiple purifcation/crystallization procedures. This
was understood that this might be due to the strong behavior of darunavir API skel-
eton to form solvate with every single solvent and reported in the studies also [19,
31, 32]. Because of strong afnity of darunavir to form solvate with almost all sol-
vents, impurities could also be crystallized out as a result of having almost similar
structure, from reaction solution, along with darunavir without solubilizing/or wash-
ing out in mother liquor, while purifcations/crystallizations were tried. Therefore,
it was very signifcant and mandatory to control these impurities formation, below
than the desired limit, during reaction conversion rather than going for unpromising
purifcations. Before to come across for its cause, it was necessary to identify and
confrm these impurities, with the aid of LCMS, NMR, mass, etc. (see supporting
information).
In next attempt, a number of experiments were performed by varying all process
parameters such as molar ratio of 5, reaction time, temperature and reaction media
(DCM, DMF, DMSO, THF and acetone) for the preparation of 10 from 9 (Table 5).
To our delight, desired compound 10 could be obtained in >99.80% purity
by reacting 9 with compound 5 (1.05 mol. equiv.) in DMSO at 15–30 °C having
1 3

An improved and robust scale-up process aided with…
Table 5 Screening of solvent for
S. No. Solvent 9 (%) 10 (%) Dimer-2
Oxazolidinone
impurity; 13 impurity; 14 (%)
(%)
the formation of 10
1.
2.
3.
4.
5.
DCM
0.10 99.58 0.06
0.20
0.06
1.16
0.04
0.10
DMSO16 0.01 99.76 0.02
DMF
THF
Acetone
27.78 66.37 2.26
57.10 37.63 0.80
2.96 94.79 1.37
impurities formation below than desired limit (NMT 0.08%). It was also understood
that under these reaction conditions if temperature raises beyond the range of 30 °C,
have directly impact on the formation of impurities (13 and 14) and temperature less
than 15 °C causes freezing of reaction mass, which again suppressed the reaction
conversion rate along with increased reaction time, further resulting to have a possi-
bility of impurities formation above than controlled limits. Therefore, condensation
of 9 with 5 is achieved by executing reaction in DMSO at 15–30 °C and at the end of
the reaction at 15–30 °C, aqueous ammonia solution (0.24 mol. equiv.) is added to
the reaction mass for decomposition of left over traces of compound 5, which could
further react with 10 if remains as such, followed by slow addition of water to reac-
tion mass for solid precipitation, stirred and fltered to aford 10 in 93–97% yield.
The optimized synthesis is proven economically rewarding in terms of increased
yield, reduced timings as well as workup hassles along with low level of known/
unknown impurities below than the controlled limit. The newly developed protocol
for DAR 2 has been found to be a robust process as it has achieved reproducible
results on commercial scale. As per studies, DAR 2 in dose form is used as in the
form of ethanolate solvate. Therefore, no further attempts were made to optimize the
ethanolate solvate formation and implemented the same prior art reported process
[6, 12, 13] for the formation of DAR 2 at the end could be aforded solid DAR 2
with desirable yield of 85–90% and purity >99.80% (any unspecifed impurity=not
more than 0.08% w/w) under considerations of ICH guidelines [33],4 requirements
and its daily in-take dose. It is worth mentioning that in fnal process step, purifca-
tion/re-crystallization and such operations were avoided to make this scale-up proto-
col most efcient, inexpensive, productive and superior in comparison with existing
processes/patents.
Conclusion
Starting from the conventional and commercial available epoxide starting material
3, the newly developed production process for DAR 2 has an overall yield of around
70% and involves fve isolation and drying steps, namely 6, 7, 9, 10 and the target
4
http://www.fda.gov/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/ucm065005.htm.
1 3

S. Girigani et al.
product 2. The new process ofers distinctive advantages over earlier reported pro-
cedures in terms of less efuent generation, avoiding the use of highly expensive
reagent like palladium besides having the robust process for controlling the process/
degradation by-products below than the desired limit (not more than 0.08%) with-
out any purifcation procedures as well as signifcant cost reduction on commercial
scale. The structural evolution of critical impurities (11–14) led to their identifca-
tion and origin followed by instituting a suitable control strategy that could result
to highly pure DAR 2 in a consistent way. To the best of our knowledge, we have
used the industrially important Raney nickel for nitro reduction and DMSO as sol-
vent in condensation reaction for the frst time in the synthesis of darunavir. The
by-products of this process, namely isobutylene, N-hydroxysuccinimide, NaCl and
Et3N·HCl, are water soluble as well as non-toxic. This process is amenable for the
large-scale production of DAR 2.
Acknowledgements The authors greatly appreciate fnancial support for this work from Micro Labs Ltd.,
API Division Centre, ML-27, Bangalore. We thank our group colleagues for their appreciated contribu-
tion. The Analytical Department of Micro Labs Limited acknowledged for providing support for analyti-
cal and spectral data.
Compliance with ethical standards
Confict of interest The authors declared that they have no confict of interest.
References
1. J.E. Arts, J.D. Hazuda, Cold Spring Harb Perspect Med 2, a007161 (2012)
2. M. Mahdi, Z. Szojka, J.A. Mótyán, J. Tozsér, Viruses 7, 6152 (2015)
3. N. Kumarasamy, A. Patel, S. Pujari, Indian J. Med. Res. 134, 787 (2011)
4. Centers for Disease Control and Prevention (2012, November 30). Vital signs: HIV infection, test-
ing, and risk behaviors among youths - United States. Morbidity and Mortality Weekly Report
(MMWR) 61(47), pp. 971–976
5. M.R. Ghante, R.S. Shelar, S.D. Sawant, M.M. Kadam, Int. J. Pharm. Pharm. Sci. 6, 240 (2014)
6. M.L. Vazluez, R.A. Mueller, J.J. Talley, D.P. Getman, G.A. DeCrescenzo, J.N. Freskos, R.M.
Heintz, D.E. Berternshaw, U.S. Patent RE 42,889 E, 2011
7. A.N. Kobe, T.S. Akashi, H.M. Takasago, K.I. Kakogawa, N.Y. Matsubara, U.S. Patent 5,929,284,
1999
8. N.M.F. Goyvaerts, P.T.B.P. Wigerinck, H.B. Zinser, B.M. Ebert, U.S. Patent 7,772,411 B2, 2010
9. A.K. Ghosh; G.M. Bilcer, T. Devasamudram, U.S. Patent 7,897,635 B2, 2011
10. P.T.B.P. Wigerinck, D.L.N.G. Surleraux, W.G. Verschueren, H.A. DeKock, W.A.A. Aelterman, U.S.
Patent Appl. 2011/0160468 A1, 2011
11. M. Mizhiritskii, E. Marom, U.S. Patent Appl. 2012/0296101 A1, 2012
12. K.R. Babu, V.K. Rao, C.N. Raju, Der Pharma Chem. 3, 389 (2011)
13. A.K. Ghosh, S. Leshchenko, M. Noetzel, J. Org. Chem. 69, 7822 (2004)
14. M.L. Vazluez, R.A. Mueller, J.J. Talley, D.P. Getman, G.A. DeCrescenzo, J.N. Freskos, R.M.
Heintz, D.E. Berternshaw, U.S. Patent 6,248,775 B1, 2001
15. Q.P.J.L. Mario, K.B.R. Romanie, V.R. Jan, L.C.S. Maria, K.J.H.M. Hero, L.F.A. Marie, Patent EP
1732931 B1, 2016
16. H.W.P. Vermeersch, D.J.C.T. Thone, D.V. Marie-Louise, U.S. Patent 7,700,645 B2, 2010
17. H. Singh, T.G. Varadaraju, S. Shah, S.R. Kola, P.K. Sahu, V.G. Regula, V.R. Rayavarpu, P. Kumar,
S.K. Dubey, Indian Patent 2451/CHE/2014 A, 2016
18. S.M. Kelly, B.H. Lipshutz, Org. Lett. 16, 98 (2014)
1 3

An improved and robust scale-up process aided with…
19. D.C. Gowda, B. Mahesh, S. Gowda, Indian J. Chem. 40B, 55 (2001)
20. D.C. Gowda, S. Gowda, Indian J. Chem. 42B, 180 (2003)
21. H. Chen, C.N. Nilsen, A. Choudhury, K.L. Sorgi, Arkivoc 14, 1 (2008)
22. J.A. Mangravite, J. Chem. Educ. 60, 439 (1983)
23. D.C. Gowda, S. Gowda, Indian J. Chem. 39B, 70 (2000)
24. H.D. Burge, D.J. Collins, Ind. Eng. Chem. Prod. Dev. 19, 389 (1980)
25. I. Pogorelic, M. Filipan- Litvic, S. Merkas, G. Ljubic, I. Cepanec, M. Litvic, J. Mol. Catal. A:
Chem. 274, 202 (2007)
26. M. Srinivasan, A. Yurek-George, A. Ganesan, Mol. Divers. 9, 291 (2005)
27. A. Isidro- Llobet, M. Alvarez, F. Albericio, Chem. Rev. 109, 2455 (2009)
28. D.S. Cofey, M.N. Hawk, S.W. Pedersen, S.J. Ghera, P.G. Marler, P.N. Dodson, M.L. Lytle, Org.
Process Res. Dev. 8, 945 (2004)
29. M. Mizhiritskii, E. Marom, U.S. Patent Appl. 20120296101 A1, 2012
30. V. Ahire, S. Sasne, A. Deshmukh, K. Kumbhar, A. Bhatnagar, D. Verma, R. Vyas, G. Singh, N.
Bhise, U.S. Patent Appl. 20120237770 A1, 2012
31. P. Kar, R. Lipowsky, V. Knecht, J. Phys. Chem. B. 117, 5793 (2013)
32. E. Marom, U.S. Patent Appl. 20120035142 A1, 2012
33. M. Chen, Family Health International Biostatistics Workshop (New Delhi, India, 2007)
Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published
maps and institutional afliations.
1 3