Synthesis of an influenza neuraminidase inhibitor intermediate via a highly diastereoselective coupling reaction.

Article abstract of DOI:10.1021/ol017268v

[reaction: see text]. A highly diastereoselective coupling reaction between TBSOP (3) and trityl sulfenimine 4 was developed which provided influenza neuraminidase inhibitor intermediate 7 in 80% yield and >99% de after crystallization. The reaction was shown to be reversible with the high diastereoselectivity resulting from a favorable H-bonding interaction in the major diastereomer.

Full text of DOI:10.1021/ol017268v

ORGANIC
LETTERS
2002
Vol. 4, No. 9
1427-1430
Synthesis of an Influenza
Neuraminidase Inhibitor Intermediate via
a Highly Diastereoselective Coupling
Reaction
,†
David M. Barnes,† Maureen A. McLaughlin,* Tetsuro Oie,‡
Michael W. Rasmussen,† Kent D. Stewart,‡ and Steven J. Wittenberger†
Global Pharmaceutical R&D, Process Research & DeVelopment, Abbott Laboratories,
1401 Sheridan Road, D-R450, Bldg. R8, North Chicago, Illinois 60064-4000, and
Department of Structural Biology, Abbott Laboratories, 100 Abbott Park Road,
Abbott Park, Illinois 60064-6100
maureen.mclaughlin@abbott.com
Received December 19, 2001 (Revised Manuscript Received March 21, 2002)
ABSTRACT
A highly diastereoselective coupling reaction between TBSOP (3) and trityl sulfenimine 4 was developed which provided influenza neuraminidase
inhibitor intermediate 7 in 80% yield and >99% de after crystallization. The reaction was shown to be reversible with the high diastereoselectivity
resulting from a favorable H-bonding interaction in the major diastereomer.
The influenza virus infects between 25 and 75 million people
annually in the United States and results in estimated health
care and lost productivity costs of over 10 billion dollars
per year.1 In recent years, significant advances in the
understanding of the virus and its replication process have
resulted in the identification of the influenza neuraminidase
enzyme as a promising target for pharmaceutical intervention.
Inhibition of neuraminidase results in the inability of the
newly produced virus to be released from the infected cell,
disrupting viral replication.2
compound (Scheme 1),4b it was immediately recognized that
development of a highly selective coupling reaction (3 + 4)
and the isolation of a single diastereomer of intermediate 2
from the reaction mixture would be imperative to the success
of this route.4c Therefore, initial efforts focused on under-
standing the factors which affect the selectivity of the
(3) Maring, C.; McDaniel, K.; Krueger, A.; Zhau, C.; Sun, M.; Madigan,
D.; DeGoey, D.; Chen, H.-J.; Yeing, M. C.; Flosi, W.; Grampovnik, D.;
Kati, W.; Klein, L.; Stewart, K.; Stoll, V.; Saldivar, S.; Montgomery, D.;
Carrick, R.; Steffy, K.; Kempf, D.; Molla, A.; Kohlbrenner, W.; Kennedy,
A.; Herrin, T.; Xu, Y.; Laver, W. G. Presented at the 14th International
Conference on Antiviral Research, AntiViral Res. 2001, 50, A76; Abstract
129.
(4) (a) For X-ray crystallographic data for compounds 7 and 11 see ref
4b, supplementary data. (b) DeGoey, D. A.; Chen, H.-J.; Flosi, W. J.;
Grampovnik, D. J.; Yeung, C. M.; Klein, L. L.; Kempf, D. J. J. Org. Chem.
2002, ASAP article. (c) A manuscript describing the process research and
initial scale-up is in preparation. For a presentation of this synthesis, see:
Bhagavatula, L.; Chen, H.-J.; DeGoey, D.; DeMattei, J.; Flosi, W. J.;
Grampovnik, D.; Gupta, A.; Hill, D.; Klein, L.; Kurukulasuriya, R.; Manna,
S.; McLaughlin, M.; Nichols, P.; Oie, T.; Premchandran, R.; Rasmussen,
M.; Stewart, K.; Tian, Z.; Wittenberger, S. J. Abstract Prepared for
Presentation at ChiraSource2001, Philadelphia, PA, Nov 2001.
Abbott’s influenza neuraminidase inhibitor A-315675 (1)
is a potent inhibitor of influenza viral replication.3 Upon
undertaking process optimization of the current route to this
† Global Pharmaceutical R&D, Process Research & Development.
‡ Department of Structural Biology.
(1) Service, R. F. Science 1997, 275, 756.
(2) (a) Varghese, J. N. Drug DeV. Res. 1999, 46, 176 and references
therein. (b) Palese, P.; Jobita, K.; Ueda, M.; Compans, R. W. Virology 1974,
61, 397. (c) Liu, C.; Eichelberger, M. C.; Compans, R. W.; Air, G. M. J.
Virol. 1995, 69, 1099.
10.1021/ol017268v CCC: $22.00 © 2002 American Chemical Society
Published on Web 04/03/2002

Scheme 1
Scheme 2a
o
a (a) BF3‚OEt2 (2 equiv), Et2O, -78 C, 4-5:1 (ratio 7:8); (b)
TMSOTf (2 equiv), THF, -78 oC, 5:1; (c) TfOH (0.8 equiv), THF/
o
heptane, -40 C, 18:1.
reaction, which would also facilitate product isolation through
crystallization.
conversion to acetate 114a,b (Figure 1). A screen of various
Lewis acids (BF3‚OEt2, ZnCl2, Ti(OiPr)4, Yb(OTf)3, Cu-
(OTf)2, TiCl4, TMSOTf, TBSOTf, TIPSOTf, Et2AlCl) in-
dicated this ratio could be slightly increased to 5-6:1 by
using TMSOTf in Et2O or THF.
Furan-, pyrrole-, and thiophene-based siloxydienes have
been shown to undergo highly selective aldol-type reactions
with a variety of chiral substrates to yield an array of highly
functionalized molecules in high enantiomeric and diastereo-
meric purity.5 In our case, coupling of N-tert-butoxycarbonyl-
2-(tert-butyldimethylsiloxy)pyrrole6,7 (3, TBSOP, 1.5 equiv)
with S-trityl sulfenimine 48 using BF3‚OEt2 (2 equiv) in THF
at -78 °C resulted in a 4-5:1 ratio of C4-C5 erythro
diastereomers 7 and 8 (Scheme 2).4b Notably, C5-C6 threo
diastereomers 9 and 10 were not observed. The structure of
major isomer 79 was identified by X-ray crystallographic
analysis, and minor isomer 8,10 isolated by chromatography
as an oil, was also characterized by X-ray analysis after
Figure 1. Derivative of minor diastereomer 8, structurally char-
acterized by X-ray crystallographic analysis.
(5) (a) Rassu, G.; Zanardi, F.; Battistini, L.; Casiraghi, G. Synlett 1999,
1333. (b) Rassu, G.; Auzzas, L.; Pinna, L.; Battistini, L.; Zanardi, F.;
Marzocchi, L.; Acquotti, D.; Casiraghi, G. J. Org. Chem. 2000, 65, 6307.
(c) Rassu, G.; Auzzas, L.; Pinna, L.; Zanardi, F.; Battastini, L.; Casiraghi,
G. Org. Lett. 1999, 1, 1213.
While optimizing the reaction stoichiometry and solvent
composition using TMSOTf, two important observations
were made. First, the diastereomer ratio increased slightly
as the reaction proceeded, possibly indicating a reversible
reaction. To further investigate this observation, the reaction
was carried out with TMSOTf (1.5 equiv) either at -78 °C
for extended reaction times (18 h) or at higher temperature
(-40 °C). In the reaction carried out at -78 °C, the
diastereomeric ratio increased from 4:1 (7 to 8) at 3 h to
11:1 after stirring for 18 h (HPLC assay yield ) 83%).
Similarly, in the -40 °C reaction the ratio increased from
9:1 at 1 h to a 19:1 ratio after 6 h (HPLC assay yield )
88%). These results strongly suggest that the reaction is
reversible and equilibrates to a thermodynamic mixture.
(6) (a) Bocchi, V.; Chierici, L.; Gardini, G. P.; Mondelli, R. Tetrahedron
1970, 26, 4073. (b) Casiraghi, G.; Rassu, G.; Spanu, P.; Pinna, L. J. Org.
Chem. 1992, 57, 3760.
(7) The literature procedure for preparation of TBSOP was unsuitable
for large scale preparation so an alternate synthesis was developed. The
results will be published shortly.
(8) For preparation and use of sulfenimine, sulfinimine, and sulfonimines,
see: (a) Davis, F. A.; Zhou, P.; Chen, B.-C. Chem. Soc. ReV. 1998, 27, 13.
(b) Branchaud, B. P. J. Org. Chem. 1983, 48, 3531. (c) Davis, F. A.; Slegeir,
W. A. R.; Evans, S.; Schwartrz, A.; Goff, D. L.; Palmer, R. J. Org. Chem.
1973, 38, 2809.
(9) tert-Butyl (2R)-2-((1R,2S)-2-methoxy-2-methyl-1-(((tritylsulfenyl)
amino)pentyl)-5-oxo-2,5-dihydro-1H-pyrrole-1-carboxylate (7): 1H NMR
(400 MHz, CDCl3) δ 7.32 (dd, J ) 2.0, 6.1 Hz, 1H), 7.29-7.19 (m, 15 H),
6.02 (dd, J ) 1.4, 6.1 Hz, 1H), 4.83 (m, 1H), 3.86 (dd, J ) 3.1, 11.5 Hz,
1H), 3.05 (s, 3H), 2.62 (d, J ) 11.2 Hz, 1H), 1.62-0.98 (m, 4H), 1.36 (s,
9H), 0.92 (t, J ) 6.66 Hz, 3H), 0.43 (s, 3H); 13C NMR (100 MHz, CDCl3)
δ 168.2, 150.5, 148.2, 144.8, 130.1, 127.6 (2C), 126.6, 82.4, 79.2, 71.1,
66.9, 64.6, 48.9, 38.7, 27.9, 19.2, 17.1, 14.9. HRMS m/z [M + H] calcd
for C35H43N2O4S, 587.2944; found, 587.2953.
(10) tert-Butyl (2S)-2-((1S,2S)-2-methoxy-2-methyl-1-(((tritylsulfenyl)-
amino)pentyl)-5-oxo-2,5-dihydro-1H-pyrrole-1-carboxylate (8): 1H NMR
(400 MHz, CDCl3) δ 7.28-7.18 (m, 16H), 6.02 (dd, J ) 1.7, 6.1 Hz, 1H),
4.90 (m, 1H), 3.89 (dd, J ) 2.0, 11.5 Hz, 1H), 3.06 (s, 3H), 2.42 (d, J )
11.5 Hz, 1H), 1.54 (s, 9H), 1.28-0.85 (m, 4H), 1.16 (s, 3H), 0.57 (t, J )
7.1 Hz, 3H); 13C NMR (100 MHz, CDCl3) δ168.6, 150.1, 148.2, 144.3,
129.9, 128.1, 127.7, 126.5, 82.6, 79.1, 71.8, 67.1, 65.3, 49.2, 37.3, 28.2,
20.9, 17.5, 14.7.
1428
Org. Lett., Vol. 4, No. 9, 2002

To confirm this hypothesis, the two diastereomers were
separated by chromatography and subjected to the reaction
conditions (7 or 8, 0.5-2 equiv of TBSOP, 1-2 equiv of
TMSOTF in THF/heptane at -40 °C). The major diastere-
omer 7 was converted to a 93:7 mixture of 7 to 8 whereas
minor diastereomer 8 was converted to a 74:26 mixture of
diastereomers. Complete conversion to the thermodynamic
ratio was not achieved since both reactions were complicated
by the formation of a significant amount of the product
resulting from loss of Boc group from 7. However, these
results clearly illustrate the reversibility of the reaction.11
The second observation was that although the reaction
proceeded smoothly to completion on small (50-100 mg)
scale, on a larger scale significant amounts (30-50%) of
imine starting material were recovered. Attempts to push the
reaction to completion through longer reaction times, higher
temperatures, or excess Lewis acid were unsuccessful and
resulted in product decomposition and low yields. This
dependence on reaction scale suggested that the larger
relative amount of water present in the smaller scale reaction,
perhaps leading to hydrolysis of TMSOTf, might have
facilitated the conversion. The ensuing experiment demon-
strated that the reaction was more effectively carried out with
triflic acid. Subsequent investigation of the reaction param-
eters showed that the ratio could be reliably improved to
18:1, with an HPLC assay yield of 95% of desired isomer
7, by performing the reaction with triflic acid (0.8 equiv) in
THF/heptane at -40 °C.12 The reaction typically shows
complete conversion to a 4:1 ratio of diastereomers within
15 min, and then equilibration to 18:1 over approximately 2
h.
>99% ee (ee ) 89% for imine 4).12a,b Since it had not yet
been demonstrated whether optical purity could be raised in
subsequent purifications, isolation of 7 with high ee was key
to the success of this route. However, while investigating
lots of imine 4 with lower ee (80%) it was discovered that
these lots provided product of equal or lower ee than the
incoming imine. Further studies showed the presence of a
racemic crystal form of slightly lower solubility, identified
by X-ray powder pattern, which prevented any enhancement
of ee at this stage of the synthesis. Therefore an alternate
synthesis that provided imine of >95% ee was developed
so that coupled product 7 could be obtained in >95% ee.4c
To investigate the source of the high selectivity achieved
in the coupling reaction, two imine analogues, 12 and 14
(Scheme 3), were prepared. When subjected to the reaction
Scheme 3
conditions (0.9 equiv of triflic acid, THF, -40 °C), imine
12 cleanly provided adduct (()-13 as a single diastereomer
in 70% yield. Conversely, when the reaction was carried out
on imine 14, the reaction mixture consisted of primarily
starting material after 2 h. Closer examination revealed that
the pyrrole enantiomers (()-15 were unstable under the
reaction conditions and rapidly reverted to starting material.
However, by carrying out the reaction with BF3‚OEt2 at -78
°C for 30 min, pyrrole (()-15 was isolated as a single
diastereomer. The threo diastereomer was not detected by
Initial attempts at isolation of 7 in high enantiomeric and
isomeric purity were very promising. After washing the crude
reaction mixture with 0.5 M NaHCO3 followed by brine, a
solvent switch to heptane (<1% THF by GC) and crystal-
lization afforded 7 in 80-85% isolated yield, >99% de and
(11) The formation of imine 4 was observed by HPLC during the course
of the equilibrations, further implicating the reversibility of the reaction
rather than another mechanism of interconversion.
(12) (a) Sample experimental for 7: To a solution of imine (4, 30.4 g,
75.3 mmol) and TBSOP (3, 31.3 g, 105 mmol, 1.4 equiv) in heptane/THF
(1:4, 520 mL) at -40 °C was added triflic acid (5.33 mL, 60.2 mmol, 0.8
equiv) dropwise at a rate to maintain T ) -35 °C. The reaction was
maintained at -40 °C and monitored by HPLC. After 1.5 h, HPLC showed
8 peak area % imine remaining (diastereomer ratio ) 12:1). Additional
triflic acid (0.15 equiv) was added and the reaction stirred for 2.5 h at which
time HPLC showed 4 peak area % imine remaining (diastereomer ratio )
19:1). The reaction was poured into 0.5 M NaHCO3 (500 mL) stirred for
10 min; the layers were separated. The organics were washed with brine
(200 mL) and dried over Na2SO4. Yield ) 96% (HPLC assay yield of THF/
heptane solution). HPLC conditions: Zorbax SB-C8 4.6 × 250 mm 5-µm
column; 1.5 mL/min flow rate; mobile phase 50:50 water (0.1% H3PO4):
acetonitrile gradient over 15 min to 10:90, hold 5 min; column temperature
35 °C; UV detection at 230 nm. Retention times: TBSOP (3) 14.2 min;
imine 4 15.5 min; major diastereomer 7 16.0 min; minor diastereomer 8
15.2 min. The solvent was switched to heptane (<1 peak area % THF by
GC) by distillation under low vacuum (T ) 30-40 °C), maintaining solvent
level at 400 mL until distillation was complete. The product begins to
crystallize out during distillation. The suspension was distilled to a volume
of 240 mL and allowed to cool to room temperature and stir overnight.
The solid was filtered and washed with heptane (2 × 30 mL). The solid
was dried in vacuo at 40 °C to give 4 (37.6 g, 85.2%, de>99%): mp 150-
152 °C. See footnote 10 for NMR data. (b) Chiral HPLC conditions for
major diastereomer 7: Chiralpak AD 4.6 × 250 mm column; 1 mL/min
flow rate; isocratic mobile phase 97:3 hexane:isopropanol; column tem-
perature ambient; UV detection at 210 nm. Retention times: 4.6, 5.2 min.
1
LC-MS or H NMR in either reaction. Furthermore, the
decreased stability of (()-15 relative to (()-13 and 7
suggests the presence of the side chain methoxy group has
a significant impact on the outcome of the reaction.
To further elucidate the source of the unexpectedly high
selectivity of the reaction of TBSOP (3) with imine 4, the
reaction was examined by molecular modeling of the four
possible diastereomeric ammonium salts corresponding to
7-10. Energy minimization using a Hartree-Fock 6-31G*
basis set showed that all low energy conformations contained
a hydrogen bond between the trityl thioamine group and the
methoxy oxygen. The distance between the side chain N
(δ+) and the ring pyrrolidine nitrogen (δ-) as well as the
size of the alkyl group proximal to sulfur (methyl or propyl)
also impacted the relative energies of the four diastereomers.
Thus, isomers 9 and 10 which show a trans (N-N distance
) 3.8 Å) relationship about the C5-C6 bond are significantly
higher in energy (∆E >5 kcal/mol) than 7 and 8 which show
a gauche (N-N distance ) 2.8 Å) conformation, due to a
Org. Lett., Vol. 4, No. 9, 2002
1429

favorable electrostatic interaction between the two charged
atoms. Furthermore, the five-membered ring formed by the
hydrogen-bonding interaction required a steric interaction
between the sulfur atom and the C7 substituent, favoring
isomer 7 (methyl) over isomer 8 (propyl). These calculations
predict a ∆E of 0.5 kcal/mol between 7 and 8 which at -40
°C translates to a thermodynamic distribution of 70:30
between isomers 7 and 8 with ,1% of 9 and 10 (∆E >5
kcal/mol). This prediction is in qualitative agreement with
the experimentally observed ratio of 95:5 (isomers 9 and 10
not observed).
developed. The coupling of TBSOP and S-trityl sulfenimine
with triflic acid followed by crystallization afforded 7 in 80-
85% yield and >99% de and has been demonstrated in the
preparation of >3 kg of 7. The reaction has been shown to
be reversible with the high selectivity arising from a favorable
H-bonding interaction in the major diastereomer.
Acknowledgment. The authors thank the Structural
Chemistry department at Abbott Laboratories for spectro-
scopic data and Rebekah Dobrin for preparation of adduct
15.
In conclusion, a highly diastereoselective route to prepare
a key intermediate in the synthesis of A-315675 has been
OL017268V
1430
Org. Lett., Vol. 4, No. 9, 2002