Efficient method for the preparation of peracetylated Neu5Ac2en by flash vacuum pyrolysis

Article abstract of DOI:10.1021/jo900224t

Peracetylated Neu5Ac2en methyl ester, an intermediate in the synthesis of the influenza neuraminidase inhibitor Relenza, has been synthesized in high yields from peracetylated Neu5Ac methyl ester by flash vacuum pyrolysis. Mechanistic evidence including deuterium labeled studies and DFT (B3LYP) calculations suggest this transformation proceeds via an intramolecular syn-elimination.

Full text of DOI:10.1021/jo900224t

Efficient Method for the Preparation of
Peracetylated Neu5Ac2en by Flash Vacuum
Pyrolysis
Evan J. Horn*,† and Jacquelyn Gervay-Hague
Department of Chemistry, UniVersity of California - DaVis,
One Shields AVenue, DaVis, California 95616
horne@uci.edu
ReceiVed February 2, 2009
FIGURE 1. Structures of Neu5Ac and related derivatives, 1, 3, and
Relenza.
TMSOTf catalyzed elimination of the peracetylated methyl ester
of Neu5Ac (2). More recent advances in the preparation of 1
include oxidation6 or elimination7 of peracetylated Neu5Ac thiogly-
cosides, and ꢀ-elimination of the peracetylated glycosyl chloride
methyl ester in refluxing Na2PO4. However, these recent syntheses
rely on activating Neu5Ac at C-2, and require several steps. To
date, the most efficient synthesis of peracetylated Neu5Ac2en has
been the method presented by C.-H. Wong and co-workers8
whereby the peracetylated methyl ester of Neu5Ac is treated with
PPh3HBr in acetonitrile, yielding 1 in 96%.
Despite the presence of these established procedures, we believed
the synthesis of 1 could be easily achieved by a thermally induced
ꢀ-elimination of the C-2 acetate of peracetylated Neu5Ac. We
hypothesized that under thermal conditions, the acetate alone could
be a sufficiently good leaving group to undergo elimination. Initial
attempts at developing this chemistry produced mixed results:
heating the peracetylated acid in pyridine yielded the desired
eliminated product, but in low yields and produced mainly
decarboxylation and dimer formation.9 Efforts to reduce these side
products by first protecting the acid as a methyl ester failed, as the
methyl ester was not observed to undergo elimination or decar-
boxylation in solution.
While thermal elimination of the C-2 acetate in peracetylated
Neu5Ac proved to be unsuccessful in solution, we questioned
whether this transformation could be possible in the gas phase under
conditions of flash vacuum pyrolysis (FVP). FVP involves the
sublimation of a compound at reduced pressure, followed by
quickly heating the vapor to pyrolytic temperatures and then
immediately cooling the vapor to cryogenic levels.10 Under these
conditions, the compound to be pyrolyzed is subjected to the high
temperatures of the column for a very short period of time (usually
on the order of 10-2 s) and therefore the occurrence of side reactions
and degradation is minimized. FVP has seen a number of synthetic
Peracetylated Neu5Ac2en methyl ester, an intermediate in the
synthesis of the influenza neuraminidase inhibitor Relenza, has
been synthesized in high yields from peracetylated Neu5Ac
methyl ester by flash vacuum pyrolysis. Mechanistic evidence
including deuterium labeled studies and DFT (B3LYP) calcula-
tions suggest this transformation proceeds via an intramolecular
syn-elimination.
The sialic acid Neu5Ac, found naturally as a cell surface
carbohydrate on mammalian cells, is involved in a number of vital
biological processes such as cellular recognition and adhesion, and
viral receptor recognition.1,2 For these reasons, derivatives of
Neu5Ac have been of general interest to researchers, as biological
probes and cell surface enzyme inhibitors. The commercial
influenza neuraminidase inhibitor Relenza (4-deoxy-4-guanidino-
Neu5Ac2en) is one such derivative. Developed by von Itzstein and
co-workers and put on the market jointly by Biota Ltd. and Glaxo-
Smith-Klein this drug is one of 2 neuraminidase inhibitors presently
on the market for treatment of the symptoms associated with
influenza. Currently Relenza is being stockpiled around the world
in preparation for any potential outbreak of influenza.
The 2-3 unsaturation present in Relenza is a common structural
motif found in a number of sialic acid derivatives possessing
neuraminidase binding activity. It is also a useful synthetic handle
for the preparation of sialic acid derivatives functionalized at C-3.3
For these reasons several methods have been developed for the
preparation of the glycal, Neu5Ac2en (1) (see Figure 1). These
include the most commonly used ꢀ-elimination of peracetylated
Neu5Ac glycosyl chloride, catalyzed by Et3N4 or DBU,5 and
(6) Kononov, L. O.; Komarova, B. S.; Nifantiev, N. E. Russ. Chem. Bull.
† Present address: Department of Chemistry, 1102 Natural Sciences II,
University of California - Irvine, Irvine, CA 92697.
(1) von Itzstein, M.; Thompson, R. J. Curr. Med. Chem. 1997, 4, 185–210.
(2) Angata, T.; Varki, A. Chem. ReV. 2002, 102, 439–469.
(3) Sun, X.; Kanie, Y.; Guo, C.; Kanie, O.; Suzuki, Y.; Wong, C.-H. Eur. J.
Org. Chem. 2000, 2643–2653.
(4) Meindl, P.; Tuppy, H. Monatsh. Chem. 1969, 100, 1295–1306.
(5) Okamoto, K.; Kondo, T.; Goto, T. Bull. Chem. Soc. Jpn. 1987, 60, 631–
636.
2002, 51, 698–702.
(7) Ikeda, K.; Konishi, K.; Sano, K.; Tanaka, K. Chem. Pharm. Bull. 2000,
48, 163–165.
(8) Burkart, M. D.; Vincent, S. P.; Wong, C.-H. Chem. Commun. 1999, 1525–
1526.
(9) Horn, E. J.; Saludes, J. P.; Gervay-Hague, J. Carbohdr. Res. 2008, 343,
936–940.
(10) Duffy, E. F.; Foot, J. S.; McNab, H.; Milligan, A. A. Org. Biomol. Chem.
2004, 2, 2677–2683.
10.1021/jo900224t CCC: $40.75
Published on Web 05/04/2009
2009 American Chemical Society
J. Org. Chem. 2009, 74, 4357–4359 4357

FIGURE 3. Relative energy diagram of intramolecular syn-elimination
of model tetrahydropyran system. Numbers in bold represent energies
(B3LYP/6-31 g(d)) in kcal/mol relative to the first structure.
simply decomposed. Interestingly, at temperatures above 420 °C
we began to see formation of the 4,5-fused-dihydro-oxazole 3, a
precursor used to prepare Neu5Ac derivatives functionalized at the
C-4 position.13
FIGURE 2. 1H NMR spectra (δ 6.6-2.4) of 2 and crude FVP products
pyrolyzed at varying temperatures. Characteristic signals a, b, and c
correspond to H-3eq of 2, H-3 of 1, and H-3 of 3 respectively. All spectra
were recorded at 400 MHz in CDCl3 and referenced to residual CHCl3 at
δ 7.26. See Supporting Information for detailed experimental procedures.
We found 410 °C to be the optimal temperature for the
elimination with our apparatus, yielding nearly quantitative product
without the need for workup or purification. Further optimization
in terms of mass recovery and overall yield was examined by
varying the carrier solvent, and substrate concentration. We found
that the total % conversion from 2 f 1 (as analyzed by crude 1H
NMR) to be mostly independent of carrier solvent, and we were
able to successfully carry out the pyrolysis using benzene, acetone,
methanol, acetonitrile, and dichloromethane with similar product
purity recovered. However, in terms of overall mass recovery,
dichloromethane produced the highest yields. We attribute this to
its low boiling point and heat of vaporization which, out of all the
solvents screened, cooled the column the least, allowing for precise
control of the temperature and high product recovery. However,
we believe this observation to be a consequence of our crude
apparatus, and feel that an industrially designed device could carry
out this method using any number of solvents.
With our conditions fully optimized, we have been able to
achieve repeatable yields of 93%.14 We have used this method
to prepare upward of 1 g of the eliminated product 1 in high
purity without trouble.
The clean conversion of this reaction suggested an inherent
reactivity of the C-2 acetate, and the fact that this reaction takes
place in the vapor phase, without the aid of a solvent or base
catalyst suggested to us that the transformation proceeds via an
intramolecular syn-elimination (Scheme 1).15
SCHEME 1. Proposed Intramolecular Syn-Elimination
applications,11 including the syn-elimination of acetates,12 however
these have mostly been limited to aromatic and low molecular
weigh substrates. To our knowledge there have been no reported
synthetic applications of FVP in carbohydrate systems.
We constructed a homemade flash vacuum pyrolysis apparatus
(Figure S1, see Supporting Information for details) that would allow
us to explore the possibility of acetate pyrolysis with sialic acid.
Our first attempts at the gas phase elimination of peracetylated
Neu5Ac were unsuccessful; the required temperature to volatilize
the substrate resulted only in decomposition. However, initial trials
using the methyl ester 2 showed small amounts of product
formation by crude 1H NMR analysis [confirmed via olefinic C-3
proton of 1 at δ 6.01 (d, J ) 3.3 Hz)]. With this promising lead,
we set out to optimize the reaction by varying packing material,
column length, temperature, pressure, and flow solvent. After
numerous trials we settled on an apparatus of column dimensions
21 cm × 1.8 mm (i.d.) packed with 3 mm diameter pyrex beads.
Using these parameters, we further optimized the pyrolysis by
In an effort to determine the feasibility of our proposed
mechanism under the conditions of flash vacuum pyrolysis we
employed density functional theory calculations on a model
tetrahydropyran system (Figure 3). By visual inspection of the
optimized geometry (B3LYP/6-31g(d))16 of the model substrate it
is clear that the constrained ring geometry places the C-2 acetate
oxygen in close proximity to the C-3 equatorial proton (2.53 Å)
but out of reach of the C-3 axial proton. Transition state calculations
revealed the desired transition state shown in Figure 3, with a ∆Gq
of 29.26 kcal/mol. Correlating this energy barrier with the tem-
1
screening the reaction at a range of temperatures. The H NMR
spectra of crude FVP product taken at varying temperatures is
displayed in Figure 2. Note that as the column temperature was
raised, the conversion of 2 to 1 increased to a point of near
quantitative transformation (410 °C), yet as the temperature
increased further, side product formation was observed and
eventually the temperature reached a point where the compound
(13) von Itzstein, M.; Jin, B.; Wu, W.-Y.; Chandler, M. Carbohydr. Res.
1993, 244, 181–185.
(14) This yield reflects the weight of pure recovered product.
(15) Wertz, D. H.; Allinger, N. L. J. Org. Chem. 1977, 42, 698–703.
(11) McNab, H. Contemp. Org. Synth. 1996, 3, 373–396.
(12) DePuy, C. H.; King, R. W. Chem. ReV. 1960, 60, 431–457.
4358 J. Org. Chem. Vol. 74, No. 11, 2009

2
SCHEME 2. Synthesis of D Labeled Substrate and
intermediates. Even so, this pathway could be considered more or
less a stepwise version of the syn-elimination.
Resulting Mechanistic Studya
In conclusion, we have developed a reagent-free method for the
preparation of Neu5Ac2en by flash vacuum pyrolysis. We believe
the method described herein to be the most efficient preparation
of peracetylated Neu5Ac2en reported to date; in terms of yield,
solvent and reagent cost, time, and waste. Furthermore, we feel
that this chemistry could easily be scaled to an industrial level as
a flow-like process representing a relatively green synthesis, as it
delivers the desired product in high yields without need for reagents
or purification. Given the current increased demand for Relenza
as a prevention for potential influenza pandemics, this method has
potential application toward the commercial production of the drug.
Experimental section
General Procedure for Flash Vacuum Pyrolysis. The apparatus
(shown in Figure S1, Supporting Information), consisting of a 21 ×
1.8 cm (i.d.) quartz column packed with 3 mm diameter pyrex beads
was allowed to heat to the desired temperature in open atmosphere.
The apparatus was then opened to a belt driven benchtop high-vac,
with pressure monitored by digital manometer. A continuous flow of
dry argon was introduced via a 21 gauge needle connected to an argon
manifold equipped at one end to a bubble trap. The apparatus was
flushed with argon for 20 min before an injection was made. The
sample to be pyrolyzed was dissolved in the degassed carrier solvent
at concentrations ranging from 5 to 200 mg per mL. The dissolved
sample was then injected dropwise into the injection flask where it
was quickly pulled into the hot column, pyrolyzed, and condensed in
the collection flask cooled by liquid N2. After all the sample solution
had been injected, the apparatus was allowed to sit for 2 min before
the vacuum was cut off and the collection flask removed. The pressure
during the pyrolysis was kept between 60 and 200 mBarr.
a Ratio determined by 1H NMR integration and HRMS of purified
products.
perature of the reaction (410 °C) we can make a rough estimation
of the t1/2 of the reaction from eq 1 and 2, which we have calculated
to be 1.42 × 10-3 s.
KBT
k )
e-∆G/RT
(1)
(2)
h
t1/2 ) -ln(0.5)/k
Considering the short time compounds spend in the hot zone
during FVP, we find this calculated value to be in agreement
with our experimental observations, and these computational
results support our proposed mechanism. Transition states of
other potential mechanisms were explored including a stepwise
E1 elimination, in which the intermediate cation would be
stabilized by oxonium formation. In this mechanism, deproto-
nation could likely occur at either the axial or equatorial C-3
proton. However, attempts at finding a reasonable transition state
for this mechanism failed, resulting only in transition state
structures that closely resembled our initial concerted mechanism.
4,7,8,9-Penta-O-acetyl-2-deoxy-2,3-dehydro-N-acetylneuramin-
ic Acid Methyl Ester (1). 2 (200 mg, 0.38 mmol) was dissolved in
2 mL of degassed dichloromethane. The solution was injected dropwise
into the injection flask of the FVP apparatus, with a column temperature
of 420 °C and a pressure of 65-90 mBar. After all of the solution
had been injected an additional 0.5 mL of degassed dichloromethane
was injected to wash the injection line. The apparatus was then allowed
to sit undisturbed for 2 min before being pressurized with argon to
atmospheric pressure. The collection flask was then removed and
allowed to warm to rt the collected solution was concentrated to yield
To further investigate the mechanism of this transformation, we
prepared the C-3 deuterated substrate 4 (Scheme 2). The base
catalyzed H f D exchange at C-3 of Neu5Ac proceeds via a ring-
opening/tautomerization sequence, and shows higher selectivity for
exchange at the axial position,17 presumably due to the extended
chirality of the carbohydrate motif. After determining ratio of
1
1 as yellow oil (170 mg, 95%). H and 13C NMR spectra were
consistent with literature.5,19
Acknowledgment. We gratefully acknowledge Professor Peter
Dervan for helpful discussions leading to this project, as well as
Michael Lowedwyk and Professor Dean Tantillo for assistance with
GAUSSIAN, and discussions in computational chemistry. Funding
for this work was provided by National Science Foundation Grant
CHE-0518010. NMR spectrometer funding was provided by NIH
Grant RR1973.
1
deuterated isomers of 4 (4 possible, 3 found) by H NMR and
HRMS, the isotopically labeled substrate was subjected to FVP
conditions, and the recovered product purified by HPLC. Analysis
of the purified mixture of 5 and 1 revealed a ratio of 20:1
respectively.18 This corresponds to nearly complete selective
elimination of the C-3 equatorial proton, and is in agreement with
our proposed syn-elimination mechanism. However, we feel this
evidence does not necessarily disprove the stepwise E1 mechanism,
and we have not ruled out tight ion pairing between the charged
Note Added after ASAP Publication. Due to a production
error Figure 3 was incorrect; the correct version was published
May 6, 2009, then again on May 8, 2009.
Supporting Information Available: Experimental procedures
including a detailed description of our FVP apparatus, computa-
tional data, and representative H NMR spectra of crude FVP
product. This material is available free of charge via the Internet
at http://pubs.acs.org.
(16) All calculations were performed using the GAUSSIAN03W suite of
programs. Geometries and energies reported herein are from B3LYP/6-31g(d)
optimizations, characterized by frequency analysis. Reported energies include
zero point energy corrections, unscaled. Free energies were calculated at 410
°C, the optimized FVP temperature. See supporting information for more details
including coordinates for all calculated structures and full GAUSSIAN 03
references.
1
JO900224T
(17) Schmidt, H.; Friebolin, H. J. Carbohydr. Chem. 1983, 2, 405–413.
1
(18) Ratio of 5:1 determined by H NMR integration ratio of C-3 proton at
(19) Kulikova, N. Y.; Shpirt, A. M.; Kononov, L. O. Synthesis 2006, 4113–
4114.
6.01 ppm relative to other peaks in the spectrum.
J. Org. Chem. Vol. 74, No. 11, 2009 4359