Aqueous methanolysis of an α-d-N-acetylneuraminyl pyridinium zwitterion: Solvolysis occurs with no intramolecular participation of the anomeric carboxylate group

Article abstract of DOI:10.1002/poc.776

The synthesis of 3,4-dihydro-2H-pyrano[3,2-c]pyridinium α-D-N-acetylneuraminoate (4), the subsequent rate constants for aqueous methanolysis and the associated reaction products are reported. When compared with the analogous 2-deoxyglucopyranosyl 4′-bromoisoquinolinium salt (11), the rate of solvolysis of 4 displays a reduced sensitivity towards the ionizing power of the solvent. Solvolysis product analysis showed that substitution occurs predominantly with inversion of configuration. Hence it can be concluded that the methanolysis reactions of 4 proceed via dissociative transition states with no intramolecular nucleophilic participation by the anomeric carboxylate group. Copyright

Full text of DOI:10.1002/poc.776

JOURNAL OF PHYSICAL ORGANIC CHEMISTRY
J. Phys. Org. Chem. 2004; 17: 478–482
Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/poc.776
Aqueous methanolysis of an a-D-N-acetylneuraminyl
pyridinium zwitterion: solvolysis occurs with no
intramolecular participation of the anomeric
carboxylate groupy
Tara L. Knoll and Andrew J. Bennet*
Department of Chemistry, Simon Fraser University, 8888 University Drive, Burnaby, British Columbia V5A 1S6, Canada
Received 24 August 2003; revised 29 October 2003; accepted 29 October 2003
ABSTRACT: The synthesis of 3,4-dihydro-2H-pyrano[3,2-c]pyridinium ꢀ-D-N-acetylneuraminoate (4), the subse-
quent rate constants for aqueous methanolysis and the associated reaction products are reported. When compared with
the analogous 2-deoxyglucopyranosyl 40-bromoisoquinolinium salt (11), the rate of solvolysis of 4 displays a reduced
sensitivity towards the ionizing power of the solvent. Solvolysis product analysis showed that substitution occurs
predominantly with inversion of configuration. Hence it can be concluded that the methanolysis reactions of 4 proceed
via dissociative transition states with no intramolecular nucleophilic participation by the anomeric carboxylate group.
Copyright # 2004 John Wiley & Sons, Ltd.
KEYWORDS: nucleophilic substitution; hydrolysis; methanolysis; carbohydrates; oxacarbenium ion; sialoside; zwitterion
INTRODUCTION
equilibrated.7 In accord with these results, the lifetime in
aqueous solution of the glucopyranosylium ion (1) has
been estimated to be around (1.0–2.5) ꢀ 10ꢁ12 s,8,9 which
is too short to allow solvent equilibration.10 Based on this
estimated lifetime for the glucopyranosylium ion, one
would expect that many reactions of glucopyranosides
should occur via concerted mechanisms. However, only
reactions in which the aglycon departs as an anion have
been shown to occur via ANDN mechanisms.11–13 In
comparison with glucopyranosides, sialosides (N-acetyl-
neuraminides) are a distinct subfamily of carbohydrates
because of the carboxylate group that is attached to the
anomeric centre. It has been proposed that this pendant
carboxylate group stabilizes the oxacarbenium ion (2) via
an electrostatic interaction.14 As a result, it is expected
that reactions at the tertiary anomeric centre in N-
acetylneuraminides, as reported by Chou et al. for the
spontaneous hydrolysis of pyridinium N-acetylneurami-
nyl zwitterions (3a),15 should occur via DN þAN mechan-
isms. However, it is also possible that intramolecular
nucleophilic participation could become a viable path-
way during the reactions of N-acetylneuraminides when
departure of an anionic leaving group is not catalyzed.
This alternative pathway has been proposed to occur
during the spontaneous hydrolysis of 3b.16 It should be
noted that because of the rapid mutarotation of N-
acetylneuraminic acid17 it is impossible to probe the
stereochemical outcome for hydrolysis reactions of these
carbohydrates. Hence in the present study, the syn-
thesis and aqueous methanolysis of 3,4-dihydro-2H-
pyrano[3,2-c]pyridinium ꢀ-D-N-acetylneuraminoate (4)
Over the years, innumerable researchers have probed
nucleophilic substitution mechanisms for phosphoryl,
acyl and acetal transfer reactions.1 One of the most
prominent research groups in this area of biological
physical organic chemistry is that of Jencks. Indeed,
several reaction mechanism probes in current use origi-
nated from this group. For example, diffusion-limited
trapping, usually by azide ion, of a carbenium ion
intermediate is used to provide a ‘clock’ from which
the intermediate’s lifetime can be estimated.2–4 For the
specific case of acetal hydrolysis, Young and Jencks used
this methodology to predict that the archetypal methoxy-
methyl oxacarbenium ion (CH3OCHþ2 ) has no existence
in water and, as a consequence, the reactions of methox-
ymethyl acetals occur via concerted ANDN mechanisms.2
This prediction has subsequently been shown to be
correct.5,6
In 1980, Sinnott and Jencks’ extensive study on the
solvolyses of several glucopyranosyl derivatives led to
the conclusion that these reactions were dissociative in
nature but that the cationic intermediate was not solvent
*Correspondence to: A. J. Bennet, Department of Chemistry, Simon
Fraser University, 8888 University Drive, Burnaby, British Columbia
V5A 1S6, Canada.
E-mail: bennet@sfu.ca
Contract/grant sponsor: Natural Sciences and Engineering Research
Council of Canada; Contract/grant number: 121348–02.
Contract/grant sponsor: Simon Fraser University.
ySelected paper part of a special issue entitled ‘Biological Applications
of Physical Organic Chemistry dedicated to Prof. William P. Jencks’.
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2004; 17: 478–482

AQUEOUS METHANOLYSIS OF CARBOHYDRATE-BASED ZWITTERIONS
479
was undertaken in order to investigate the stereochemical
outcome for the reactions of pyridinium N-acetylneura-
minyl zwitterions.
J3e,4 ¼ 7.0 Hz, H-3e), 4.04 (dd, 1 H, J9a,9b ¼ 12.5 Hz,
J9a,8 ¼ 5.8 Hz, H-9a), 4.21–4.29 (m, 2 H, H-5, H-9b),
4.49 (t, 2 H, J2 ,3 ¼ 5.2 Hz, H-20), 5.00 (dd,
1 H, J6,5 ¼ 11.4 Hz, J6,7 ¼ 3.1 Hz, H-6), 5.42 (dd, 1 H,
0
0
J4,3e ¼ 7.0 Hz, J4,3a ¼ 2.6 Hz, H-4), 5.49–5.55 (m, 2 H,
0
0
0
H-8, H-7), 7.11 (d, 1 H, J7 ,8 ¼ 7.4 Hz, H-8 ), 8.62 (d, 1 H,
H-70), 8.83 (s, 1 H, H-50). 13C NMR (400 MHz, D2O) ꢁ
20.1, 20.8, 20.9, 21.5, 22.0, 23.1, 29.7, 37.9, 48.7, 63.1,
67.0, 67.1, 69.4, 70.2, 73.5, 94.1, 114.0, 121.8, 139.2,
142.0, 165.7, 167.4, 169.9, 170.0, 170.8, 171.1, 171.6.
N-[(5-Acetamido-3,5-dideoxy-D-glycero-ꢀ-D-galacto-non-
2-ulopyranosyl)onate]-30,40-dihydro-20H-pyrano[30,20-c]-
pyridinium (4). Deprotection of 7 by use of the method
described by Chou et al.,15 followed by reversed-phase
HPLC purification [mobile phase: 5% aqueous MeOH
containing 1% (v/v) HOAc on a C-18 column) gave 4 as a
1
white solid in 25% yield. H NMR (400 MHz, D2O), ꢁ
1.89 (t, 1 H, J3a,3e þ J3a,4 ¼ 22.9 Hz, H-3a), 2.07 (m, 2 H,
EXPERIMENTAL
0
H-30), 2.90 (t, 1 H, J3 ,4 ¼ 6.4 Hz, H-4 ), 3.23 (dd, 1 H,
0
0
J3e,3a ¼ 12.2 Hz, J3e,4 ¼ 3.9 Hz, H-3e) 3.62–3.67 (m, 2 H,
Methanol was dried by distillation from its magnesium
alkoxide salt. Deionized water was further purified by use
of a Milli-Q ultra-pure water system. NMR spectra were
acquired at operating frequencies of 400 and 100 MHz for
1H and 13C NMR, respectively, using either CDCl3 or
D2O as the solvent and internal reference. Coupling
constants (J) are reported in hertz. The two anomeric
methyl N-acetylneuraminides (8a and 8b) were pur-
chased from Toronto Research Chemicals (TRC), and
N-acetylneuraminic acid (9) from Rose Chemicals.
H-7, H-9a), 3.84–4.04 (m, 5 H, H-4, H-5, H-6, H-8, H-
0
0
0
9b), 4.51 (t, 2 H, J2 ,3 ¼ 5.1 Hz, H-2 ), 7.24 (d, 1 H,
0
0
0
0
0
J7 ,8 ¼ 7.4 Hz, H-8 ), 8.57 (dd, 1 H, J5 ,7 ¼ 2.0 Hz, H-
70), 8.66 (d, 1 H, H-50). 13C NMR (400 MHz, D2O), ꢁ
22.2, 24.2, 24.7, 43.1, 53.8, 65.4, 70.6, 70.7, 72.6, 73.7,
77.2, 96.5, 117.5, 126.1, 140.7, 142.4, 171.4, 172.3,
177.7. HRMS (ESI): calcd for C19H27N2O9 [M þ Hþ],
427.1717; found, 427.1717.
Syntheses
N-[Methyl (5-acetamido-4,7,8,9-tetra-O-acetyl-3,5-di-
deoxy-D-glycero-ꢀ-D-galacto-non-2-ulopyranosyl)onate]-
30,40-dihydro-20H-pyrano[30,20-c]pyridinium tetrafluoro-
borate (7). The fully protected N-acetylneuraminyl
chloride 6 (1.00 g, 1.96 mmol)18 was added to a suspen-
Kinetics
The solvolysis reactions of 4 were conducted at 65 ꢂC and
were monitored by following the decrease in absorbance
at 266 nm using a Cary 3E UV–visible spectrophotometer
equipped with the Cary six-cell Peltier constant-tempera-
ture accessory. Reactions were initiated by the injection
of a stock solution of 4 (10 ml, 3 mM) into a methanol–
water mixture (1.00 ml) containing N-methylmorpholine
(3 equiv.). Rate constants were calculated by non-linear
least-squares regression of the absorbance versus time
data to a standard first-order rate equation.
˚
sion of dried 4 A molecular sieves (3.0 g) in anhydrous
THF (40 ml). This mixture was stirred under a nitrogen
atmosphere for 10 min, following which 3,4-dihydro-2H-
pyrano[3,2-c]pyridine (5)19 (6.63 g, 49.0 mmol) was
added and the reaction mixture was cooled using an
ice–water bath in the dark. Subsequently, silver tetra-
fluoroborate (0.05 g, 2.56 mmol) was added to the stirred
solution, which was then allowed to warm to room
temperature, and the reaction mixture was kept in the
dark at this temperature for 60 h. The crude product,
which was obtained after a standard work-up, was pur-
ified by flash chromatography [silica gel, MeOH–CH2Cl2
(1:10)] to give a white solid (0.56 g, 44%): RF 0.24
Determination of pKa
1
[MeOH–CH2Cl2 (1:7)]. H NMR (400 MHz, CDCl3), ꢁ
1.89 (m, 4 H, CH3, H-3a), 2.01 (s, 3 H, CH3), 2.05 (s, 3 H,
CH3), 2.12–2.17 (m, 5 H, CH3, H-30), 2.27 (s, 3 H, CH3),
3.00 (m, 2 H, H-40), 3.48 (dd, 1 H, J3a,3e ¼ 14.0 Hz,
The pKa of the conjugate acid of 3,4-dihydro-2H-
pyrano[3,2-c]pyridine was calculated from absorbance
versus pH data measured in buffered solutions (10 mM) of
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2004; 17: 478–482

480
T. L. KNOLL AND A. J. BENNET
the pyridine (1 ꢀ 10ꢁ4 M) between pH 5.5 and 9.0 (no
Table 1. Observed rate constants for the aqueous metha-
nolysis of 4 at 65.0 ꢂCa
ionic strength adjustment).
[MeOH] (%, v/v)
105 ꢀ kobs (sꢁ1
)
100
80
60
50
40
20
0
10.8 ꢄ 0.3
4.58 ꢄ 0.29
2.85 ꢄ 0.02
2.18 ꢄ 0.01
1.64 ꢄ 0.01
1.10 ꢄ 0.04
0.868 ꢄ 0.028
Product studies
The solvolytic product studies were performed by heating
at 65 ꢂC in a sealed vial that contained a
for 7 ꢀ t
½hyd
solution of 4 (1.05–1.12 mg) and N-methylmorpholine
(3 equiv.) in a binary solvent mixture (10 ml). After
cooling, the reaction vials were opened and the solvent
was removed under reduced pressure. After drying under
vacuum (ꢃ0.01 mm Hg) for 48 h, the resulting solid
residues were dissolved in D2O (0.5 ml) and their
1H NMR spectra were acquired. The reaction products
formed were identified by comparing the newly acquired
NMR spectra with those of two methyl sialosides (8a and
8b), sialic acid (9) and the glycal 10.
a
Mean value of three runs; quoted error ¼ ꢂnꢁ1
.
the slope of this plot is 0.71, it is clear that a reduction in
the solvent’s polarity produces a smaller rate-accelerating
effect for the solvolysis of 4 relative to the reported effect
for 11.20
Product studies on the solvolysis reactions of 4 were
performed in the presence of the sterically hindered base
N-methylmorpholine to ensure that the acid-labile products
(8a and 8b) were stable to the solvolytic conditions. Listed
in Table 2 are the observed products for the reactions of 4
in both methanol and 50% (v/v) aqueous methanol.
RESULTS
After synthesis of 3,4-dihydro-2H-pyrano[3,2-c]pyridine
(5),19 the pKa of its conjugate acid was calculated as
7.18 ꢄ 0.02 by fitting the measured absorbance (245 nm)
versus pH data to a standard titration equation. The
synthetic route devised by Chou et al.15 was used to
couple the basic pyridine (5) to a fully protected N-
acetylneuraminyl chloride (6), resulting in 7, which, after
deprotection, gave 4 (Scheme 1; see Experimental section
for full details).
DISCUSSION
The central question addressed in this work is whether
zwitterions such as 4 react with (pathway a; Scheme 2) or
without (pathway b; Scheme 2) anchimeric assistance by
the anomeric carboxylate group.
Listed in Table 1 are the observed first-order rate
constants for the aqueous methanolysis of 4 at 65 ꢂC.
In order to facilitate a direct comparison with
the published data for the solvolytic reaction of 11, the
methanolysis reactions of 4 were performed without the
addition of ionic strength adjusting salts.20 Figure 1
shows a plot of log(kobs)4 versus log(kobs)11. Given that
It is known that positively charged glycosyl pyridinium
salts hydrolyse via transition states that have significant
C—N bond cleavage.21–23 As a result, the expected
effects of solvent polarity changes on the two possible
pathways shown in Scheme 2 are different. Pathway a is
analogous to a ‘type III’ SN2 reaction where charge at
the transition state is greatly reduced in comparison with
Scheme 1
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2004; 17: 478–482

AQUEOUS METHANOLYSIS OF CARBOHYDRATE-BASED ZWITTERIONS
481
Therefore, a reaction proceeding via pathway a should
increase in rate in a lower polarity solvent since H-
bonding to the carboxylate group should be weaker at
the transition state relative to that at the ground state. In
contrast, a reaction proceeding via pathway b will have
the positive charge that originates on the pyridinium ring
being delocalized at the transition state and, as a result,
only a modest increase in reaction rate is expected. For
the pathway b scenerio, H-bonding to the carboxylate
group should be similar at the transition and ground
states. Therefore, based on the observation that the
solvolytic rate constants for 4 display a smaller sensitivity
to changes in solvent polarity than do the corresponding
reactions of 11, a compound that lacks the appended
carboxylate group, it can be concluded that the carbox-
ylate group in 4 is not directly involved as a nucleophile
during these solvolysis reactions.
Figure 1. Plot of log(kobs)4 versus log(kobs 11 for the aqu-
)
eous methanolyses of 4 and 11 at 65 ꢂC. The kinetic data for
11 were taken from Ref. 20. The line shown is the best linear
least-squares fit through the data points
The observed solvolysis products for the reactions of 4
are also consistent with reactions proceeding via pathway
b (Scheme 2). Specifically, in 100% methanol the only
substitution product observed is the inverted methyl ꢃ-D-
N-acetylneuraminide (8b), not 8a, the product expected
for reaction via pathway a. In contrast, it has been
reported that the aqueous methanolysis of 12 in 20%
(v/v) MeOH–H2O, at pH 5.0 and 6.0, yields an approxi-
mately 1:1 mixture of methyl ꢀ- and ꢃ-D-N-acetylneur-
aminides (8a and 8b).14,25 Clearly, the reactions of these
two different N-acetylneuraminides do not proceed via a
common, solvent-equilibrated, intermediate. It is likely
that at least some of the reaction products formed from 12
are made at the solvent-separated ion pair stage: reactions
of glycosides in which the leaving group is anionic, such
as ꢀ-glucopyranosyl fluoride,11,12 are more prone to react
via an ANDN (SN1) mechanism than are the correspond-
ing reactions of analogous neutral leaving group, yielding
substrates such as ꢀ-glucopyranosyl pyridinium
salts.21,22 The nucleophilic selectivity for the reaction
of 4 in 50% (v/v) MeOH–H2O (kMeOH/kHOH ¼ 1.58),
calculated according to Eqn (1),26 is larger than the
Table 2. Observed products formed during the reactions of
4 in aqueous methanol at 65 ꢂCa,b
[MeOH]
(%, v/v) (10) (%) (9) (%)
Glycal
GlyOH ꢀ-GlyOMe ꢃ-GlyOMe
(8a) (%)
(8b) (%)
100
10
—
NDc
NDc
90
35
50d
14
50e
a
b
c
All solvents contained 3 mol equiv. of N-methylmorpholine.
Percentages do not necessarily add up to 100 because of rounding.
Not detected; it was estimated that <5% methyl ꢀ-D-N-acetylneurami-
nide (8a) was formed.
[H2O]/[MeOH] ¼ 2.25 in this solvent mixture.
d
e
Integral for ꢃ-D-sialic acid corrected assuming that 7% of the total sialic
acid present was in the ꢀ-form.
the ground state, and thus a reduction in solvent po-
larity should lead to a very large increase in reaction
rate.1 The Taft ꢀ value of 1.17 for water24 suggests that
water is a better H-bond donor than methanol (ꢀ ¼ 0.93).
Scheme 2
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2004; 17: 478–482

482
T. L. KNOLL AND A. J. BENNET
corresponding value reported for the reactions of 11 in
the same solvent (kMeOH/kHOH ¼ 1.07).20
REFERENCES
1. March J. In Advanced Organic Chemistry: Reaction, Mechan-
isms, and Structure (4th edn). Wiley–Interscience: New York,
1992; 293–369.
2. Young PR, Jencks WP. J. Am. Chem. Soc. 1977; 99: 8238–
8248.
3. Richard JP, Jencks WP. J. Am. Chem. Soc. 1984; 106: 1373–
1383.
4. Richard JP, Rothenberg ME, Jencks WP. J. Am. Chem. Soc. 1984;
106: 1361–1372.
5. Craze GA, Kirby AJ, Osborne R. J. Chem. Soc., Perkin Trans. 2
1978; 357–368.
6. Knier BL, Jencks WP. J. Am. Chem. Soc. 1980; 102: 6789–6798.
7. Sinnott ML, Jencks WP. J. Am. Chem. Soc. 1980; 102: 2026–
2032.
8. Amyes TL, Jencks WP. J. Am. Chem. Soc. 1989; 111: 7888–7900.
9. Huang X, Surry C, Hiebert T, Bennet AJ. J. Am. Chem. Soc. 1995;
117: 10614–10621.
10. Richard JP, Tsuji Y. J. Am. Chem. Soc. 2000; 122: 3963–3964.
11. Banait NS, Jencks WP. J. Am. Chem. Soc. 1991; 113: 7951–7958.
12. Zhang Y, Bommuswamy J, Sinnott ML. J. Am. Chem. Soc. 1994;
116: 7557–7563.
13. Bennet AJ, Kitos TE. J. Chem. Soc., Perkin Trans. 2 2002; 1207–
1222.
14. Horenstein BA, Bruner M. J. Am. Chem. Soc. 1998; 120: 1357–
1362.
15. Chou DTH, Watson JN, Scholte AA, Borgford TJ, Bennet AJ.
J. Am. Chem. Soc. 2000; 122: 8357–8364.
16. Ashwell M, Guo X, Sinnott ML. J. Am. Chem. Soc. 1992; 114:
10158–10166.
17. Friebolin H, Supp M, Brossmer R, Keilich G, Ziegler D. Angew.
Chem., Int. Ed. Engl. 1980; 19: 208–209.
kMeOH ½Gly ꢁ OMeꢅ½HOHꢅ
¼
ð1Þ
kHOH
½Gly ꢁ OHꢅ½MeOHꢅ
This greater selectivity of the N-acetylneuranimyl
carbenium ion (2) relative to that of the 2-deoxygluco-
pyranosylium ion (1b) for the more nucleophilic metha-
nol is consistent with cation 2 having a longer lifetime
than cation 1b in solution. The reported lifetimes for the
cations 1b9 and 214 in water are 1.0 ꢀ 10ꢁ11 and
ꢆ 3 ꢀ 10ꢁ11 s, respectively.
As a final point, glycal 10 formation probably occurs
with proton-abstraction assistance provided by the leav-
ing group pyridine at the point of the ion–molecule
complex. As such, this proposed a mechanism is similar
to that put forward by Thibblin and Saeki for the
solvolysis reactions of 1-(1-methyl-1-phenylethyl)pyridi-
nium cations.27
18. Roy R, Laferriere C. Can. J. Chem. 1990; 68: 2045–2054.
19. Sliwa H, Cordonnier G. J. Heterocycl. Chem. 1975; 12: 809–810.
20. Zhu J, Bennet AJ. J. Org. Chem. 2000; 65: 4423–4430.
21. Hosie L, Marshall PJ, Sinnott ML. J. Chem. Soc., Perkin Trans. 2
1984; 1121–1131.
22. Huang X, Tanaka KSE, Bennet AJ. J. Am. Chem. Soc. 1997; 119:
11147–11154.
23. Zhu J, Bennet AJ. J. Am. Chem. Soc. 1998; 120: 3887–3893.
24. Kamlet MJ, Abboud JLM, Abraham MH, Taft RW. J. Org. Chem.
1983; 50: 2877–2887.
Acknowledgements
25. Horenstein BA, Bruner M. J. Am. Chem. Soc. 1996; 118:
10371–10379.
26. Ta-Shma R, Rappoport Z. Adv. Phys. Org. Chem. 1992; 27:
239–291.
27. Thibblin A, Saeki Y. J. Org. Chem. 1997; 62: 1079–1082.
The authors gratefully acknowledge the Natural Sciences
and Engineering Research Council of Canada and Simon
Fraser University for financial support of this work.
Copyright # 2004 John Wiley & Sons, Ltd.
J. Phys. Org. Chem. 2004; 17: 478–482