Synthesis and investigation of L-fuco- and D-glucurono-azafagomine

Article abstract of DOI:10.1039/b200884j

The new azasugars (3S,4R,5S)-4,5-dihydroxy-3-methylhexahydropyridazine (3, azafucofagomine) and (3S,4R,5R)4,5-dihydroxyhexahydropyridazine-3-carboxylic acid (4, azaglucuronofagomine) were synthesised. Azafucofagomine (3) was made from D-ribose in ten step

Full text of DOI:10.1039/b200884j

View Article Online / Journal Homepage / Table of Contents for this issue
Synthesis and investigation of L-fuco- and
D-glucurono-azafagomine
Henrik H. Jensen, Astrid Jensen, Rita G. Hazell and Mikael Bols*
1
Department of Chemistry, University of Aarhus, Langelandsgade 140, DK-8000 Aarhus,
Denmark. E-mail: mb@chem.au.dk; Fax: ϩ4586196199; Tel: ϩ4589423963
Received (in Cambridge, UK) 23rd January 2002, Accepted 21st March 2002
First published as an Advance Article on the web 12th April 2002
The new azasugars (3S,4R,5S)-4,5-dihydroxy-3-methylhexahydropyridazine (3, azafucofagomine) and (3S,4R,5R)-
4,5-dihydroxyhexahydropyridazine-3-carboxylic acid (4, azaglucuronofagomine) were synthesised. Azafucofagomine
(3) was made from -ribose in ten steps in a synthesis that involved partial 2,3-protection, deoxygenation of the
5-OH, reductive amination with tert-butyl carbazate, mesylation, cyclisation and deprotection. Compound 4 was
made from -xylose in 12 steps in a related way starting with 2,3,5-protection, reductive amination with tert-butyl
carbazate, mesylation and cyclisation. The key step in this synthesis is selective debenzylation of a primary benzyl
ether with acetyl bromide to produce a partially benzylated hexahydropyridazine that was oxidised to the acid and
deprotected. The 3-, 4- and 6-deoxy analogues of azafagomine [1, (3R,4R,5R)-4,5-dihydroxy-3-hydroxymethylhexa-
hydropyridazine] were also made. Compounds 3 and 4 were shown to be potent α-fucosidase and β-glucuronidase
inhibitors, respectively.
by synthesising the azafagomine analogues of -fucose and
-glucuronic acid, 3 and 4 (Fig. 2).
Introduction
Glycosidases and related enzymes are crucial in many biological
processes. Potent and selective inhibitors of these enzymes are
important, because they can be used to interfere with such
processes. Thus glycosidase inhibitors may be potential agents
against diabetes,1 cancer, AIDS,2 hepatitis,3 Gaucher’s disease3
and influenza.4 A particularly effective way of procuring
enzyme inhibitors is to design transition state analogues. Aza-
sugar inhibitors,5 in particular, are subject to interest in this
Fig. 2
respect because of their ability to mimic charge in a glycosidase
transition state. Positive charge can develop on the ring oxygen
as depicted in the transition state B, or on the anomeric carbon
as in A (Fig. 1). Which of the two transition states A and B is the
Two different routes to optically active 1-azafagomines have
been established through the synthesis of 1. This compound
can either be prepared from a carbohydrate11 or be made by
a hetero Diels–Alder reaction sequence linked to a lipase
catalysed resolution of a racemic intermediate.12 The former
general strategy is probably to be preferred, because it is
efficient and provides the highest possible optical purity, but
can only be applied to the synthesis of stereoisomers where the
pentose starting material is reasonably priced. While 1 can be
prepared from relatively inexpensive -xylose, synthesis of 2 by
this method would require expensive -ribose. For this reason 2
has been prepared by the chemoenzymatic route.10 Compound
3, on the other hand, is basically the mirror image of 2 but lacks
the 3Ј-hydroxy group. This means that it might be possible to
prepare 3 from inexpensive -ribose. Compound 4 has the
same stereochemistry as 1, which means that it could poten-
tially be made from -xylose as well. We here report the syn-
thesis 3 and 4, an investigation of their glycosidase inhibition,
Fig. 1
and an evaluation of the importance of the individual hydroxy
groups in 1.
more important appears to depend on the enzyme studied, but
most enzyme transition states are likely to have a component of
Results and discussion
each.6–8
1-Azafagomine (1) is a compound that is able to mimic
-Ribose contains the correctly configured hydroxy groups
present in the 4- and 5-positions of the target compound 3.
Similarly to the synthesis of 1 from -xylose the 4-hydroxy
group of -ribose must be replaced by hydrazine with inversion.
However, the 5-hydroxy group of -ribose also needs to be
removed. It was conceived that both objectives could con-
veniently be carried out by using the isopropylidene derivative 5
both transition states A and B and as such appears to be
an ideal mimic of the charge in the transition state. Com-
pound 1 is indeed a very good inhibitor of many glucosidases.9
The galacto-analogue 2 has also been prepared, and this
compound is a good galactosidase inhibitor.10 In the present
study we wished to complete the investigation of azafagomines
1190
J. Chem. Soc., Perkin Trans. 1, 2002, 1190–1198
This journal is © The Royal Society of Chemistry 2002
DOI: 10.1039/b200884j

View Article Online
Scheme 1
as a starting material. Acetonide 5 was obtained in one step
from -ribose by treatment with acetone and acid. If the 5-OH
of 5 could be deoxygenated to 7 without affecting the 1-OH, the
route would be paved for obtaining 3, similarly to the synthesis
of 1 from -xylose. In fact 7 is a known compound that has
been obtained from -ribono-1,4-lactone, which avoids poten-
tial chemoselectivity problems.13 However, synthesis from 5
should be more efficient.
Selective tosylation of 5 has previously been reported.14 We
did the reaction under slightly different conditions obtaining
monotosylate 6 in 73% yield (Scheme 1). Treatment of 6 with
NaI in dimethoxyethane gave the 5-iodo derivative, which was
hydrogenated at 1 atm in the presence Pd–C to give 7 in 74%
yield. Reductive amination of 7 using tert-butyl carbazate,
NaCNBH3 and acetic acid gave the hydrazine 8 in 97%
yield. Reaction of 8 with acetic anhydride in methanol gave the
N-acetyl compound that, without purification, was subjected
to mesylation using mesyl chloride and pyridine. This gave
O-mesylate 9 in 81% yield. Then, treatment of 9 with TMSOTf–
lutidine, a reagent selective for removal of Boc groups in the
presence of acetonide groups, gave a crude monohydrazide that
was subjected to cyclisation by being refluxed for 24 h in CHCl3
in the presence K2CO3. This gave 10 in 91% yield. Compound
10 gave a poorly resolved NMR spectrum, but as it was crystal-
line an X-ray crystal structure was used to confirm the structure
(Fig. 3). The structure shows the compound in the anticipated
chair conformation. Finally deprotection of 10 was carried out
by hydrolysis with refluxing 6 M hydrochloric acid giving 3 in
quantitative yield.
Fig. 3 X-Ray structure of compound 10.
Σσs where σs is the substituent contribution. In the present case
the pKa of N1 is predicted to be 6.3 and pKa of N2 is predicted
to be 5.9. The overall pKa becomes 6.4 from the equation pKa =
log(1/KaN1 ϩ 1/KaN2),15 and is thus reasonably well predicted by
this method.
The target compound 4 has the same stereochemistry as 1
and might therefore be made from -xylose. It was conceived
that the intermediate 11, which is made from -xylose in
8 steps,11 might be a potential starting material for 4 provided
a selective debenzylation of the primary benzyl ether could
be carried out (Scheme 2). In order to achieve this, 11 was
treated with neat acetyl bromide for 7 h at 25 ЊC. The hydrazine
was quickly acetylated, followed by a slower acetylation–
debenzylation reaction of the primary OH-group, affording a
triacetyl compound. The crude product was O-deacetylated
using NaOMe in MeOH to give diacetyl compound 12 in 73%
yield. A 7% yield of the fully benzylated diacetyl compound 13
The pKa of 3 was measured as 6.3 by titration. This value was
compared to the predicted value obtained by a recently pub-
lished empirical method as a confirmation of the latter.15
According to this method each of the nitrogens in a hexahydro-
pyridazine have their pKa calculated using the formula 7.3 Ϫ
J. Chem. Soc., Perkin Trans. 1, 2002, 1190–1198
1191

View Article Online
Fig. 5
the three deoxygenated analogues 16–18 in optically pure form
(Fig. 5). The 5-deoxy analogue 16 was obtained by a modifi-
cation of the chemoenzymatic synthesis of 1 (Scheme 3).12 In
this synthesis the optically active epoxide 19 was an inter-
mediate which, upon acidic hydrolysis, gave selective attack of
water at C-4. It was found that treatment of 19 with HI in acetic
acid gave the same selectivity. After acetylation the iodide 20
was obtained in 73% yield. This iodide was subjected to radical
reduction using Bu3SnH–AIBN to give the 4-deoxy derivative
21 in 65% yield. Finally deacetylation with NaOMe–MeOH
followed by hydrazinolysis with neat hydrazine at 100 ЊC for
24 h gave 16 in 57% yield.
The 4-deoxy analogue 17 was made by a modification of the
synthesis of 2.10 In this synthesis the iodides 22 and 23 were
intermediates (Scheme 4). This inseparable 1 : 3 mixture of
Scheme 2
was also obtained. Then oxidation of 12 using Jones reagent
gave a 66% yield of the acid 14. Finally hydrogenolysis at 1 atm
using Pd–C catalyst followed by acidic hydrolysis with 6 M
hydrochloric acid at 100 ЊC gave the target 4 in 74% yield.
The conformational behaviour of 4 is unusual. Compound 4
1
was observed to be predominantly in the C4 conformation
when the nitrogen was protonated regardless of whether the
carboxylate was protonated or not. The unprotonated com-
pound was, however, in the expected 4C1 conformation (Fig. 4).
Scheme 3
Fig. 4
This was seen by the unusually small J3,4 and J4,5 at neutral and
low pH. The compound had an estimated conformer ratio in
4
1
water of 9 : 1 between C1 and C4 at pH 11 and 1 : 4 at pH 1.
The known very similar piperidine 15 is however predominantly
in the 4C1 conformation as the hydrochloride.16
This behaviour can be explained by stereoelectronic sub-
1
stituent effects, which causes the C4 conformer to be more
4
1
basic than the C1 conformer.15 This makes the C4 conformer
favoured in acidic solution. The reason why only 4 and not also
15 flips predominantly to the 1C4 conformation must be because
1
the steric hindrance between axial substituents in 15 in a C4
conformation is larger than in 4. In particular, the 1,3-steric
interactions between the additional CH2 group and the 4-OH
in 15 must make the 1C4 conformation less favourable for
this compound than it is for 4 where the CH2 group has been
replaced by an NH-group.
Another aspect required to complete the investigation of 1
was the evaluation of the effect of the biological activity of the
different OH groups in 1. It was therefore necessary to obtain
Scheme 4
1192
J. Chem. Soc., Perkin Trans. 1, 2002, 1190–1198

View Article Online
Table 1 Ki values in µM at pH 6.8, 25 ЊC (—, not tested)
iodides was obtained from the cis diastereomer of epoxide 19
by reaction with HI followed by acetylation similar to the trans-
formation 19 into 20, and both isomers were converted into 2.
However, the major isomer 23 should be source for 17. There-
fore the mixture of 22 and 23 was reduced with Bu3SnH–AIBN
to give a mixture of de-iodo compounds 24 and 25. This mix-
ture was separated by chromatography to give 46% of 25 and
21% of 24. Both compounds were deprotected with NaOMe–
MeOH followed by hydrazine hydrate at 100 ЊC, which afforded
17 in 65% yield and 26 in 63% yield.
3
32
33
α-Fucosidase
(human placenta)
α-Fucosidase
0.63
0.81
6.4a
4b
—
0.0013c
(bovine kidney)
a Value from ref. 17. b Value from ref. 18. c Value from ref. 19.
The 3Ј-deoxy analogue 18 was made by a modification of the
synthesis of 4, since the intermediate 12 was ideal for this
purpose (Scheme 5). Compound 12 was chlorinated using
itors is their lack of a 2-hydroxy group. The results also suggest
that the hydrazine moiety is a better transition state mimic than
an amine in either position. It would be interesting to compare
the inhibition of 3 with the 2-deoxy analogue of 33 to confirm
this, but that compound has not been reported. This suggests
that a 2-hydroxy analogue of 3 would be a better α-fucosidase
inhibitor than 33 or the 2-hydroxy analogue of 32. Such a com-
pound would, however, probably be very unstable and form a
hydrazone.
The Ki value of 4 for the inhibition of β-glucuronidase is
shown in Table 220–23 together with a series of glucuronidase
inhibitors from the literature. Compound 4 is a weaker inhib-
itor than the corresponding isofagomine 15, but stronger than
the deoxynojrimycin 34. The order of inhibition isofagomine >
azafagomine > deoxynojirimycin has also been found for
inhibition of other β-glycosidases.8,11 The difference in potency
between the azafagomine and isofagomine has, however, never
been seen as large as in this case. One effect that could con-
tribute to the decreased inhibition of 4 compared to the iso-
fagomine 15 is the conformational behaviour discussed above.
Compound 4 probably binds in protonated form, but proton-
Scheme 5
1
ated 4 is predominantly in the undesirable C4 conformation,
methanesulfonyl chloride in pyridine at 80 ЊC in the presence of
CsCl. This procedure gave the 3Ј-chloro derivative 27 in 65%
yield. This compound was reduced with Bu3SnH–AIBN to give
a 3Ј-deoxy compound 28 in 66% yield. Finally, hydrogen-
olysis at 1 atm in the presence of HCl and Pd–C followed by
acidic hydrolysis using 6 M hydrochloric acid at 100 ЊC gave
67% of 18.
which decreases the concentration of desired 4C1 conformation
and thereby the observed inhibition.
The evaluation of the deoxy analogues 16–18 resulted in
some interesting observations (Table 3). It is clear from the data
that removal of hydroxy groups from 1 always reduces its
potency. However, the effect of removal of a hydroxy group is
less severe when (a) it is in the 6-position (sugar numbering) and
(b) the enzyme is almond β-glucosidase. This shows that all the
hydroxy groups of 1 contribute to binding, and together with
the observations above show that the best azasugar inhibitor
would be one with all the hydroxy groups of the substrate
intact. It is remarkable that the primary hydroxy is by far the
least important of the hydroxy groups. Removal of this OH
reduces binding by a factor of only 10–15 for three glycosidases.
However removal of the entire hydroxymethyl group, as in ( )-
29, reduces inhibition of the same three enzymes a 1000-fold, so
the presence of the methyl group is far more important than the
OH group. Compound 29 is a more conformationally flexible
molecule than 18, but this cannot explain the tremendous loss
of activity. It is likely that the methyl group in 18 displaces a
molecule of water from the active site which is important for the
overall binding process. This is confirmed by compound 4 also
being a stronger glucosidase inhibitor than 29 (Table 2). ( )-29
is also a surprisingly poor β-xylosidase inhibitor (Table 3).
The importance of the 3- and 4-hydroxy groups (carbo-
hydrate numbering) can be confirmed by comparing the Ki
values of ( )-30 and ( )-31 with those of 37 (Table 4). Both ( )-
30 and ( )-31 are extremely poor inhibitors of the glycosidases
investigated, with the exception of the inhibition of almond β-
glucosidase by ( )-31. The 4-deoxy analogue ( )-31 is 50 times
less potent than 37, which is very similar to the difference
in inhibition between azafagomine (1) and its 4-deoxy analogue
17. This confirms the relatively low importance of the 4-
hydroxy group for the inhibition of almond β-glucosidase,
The target compounds 3, 4, 16–18 were tested for inhibition
of a series of glycosidases. The known15 compound ( )-29 was
included in the evaluation since it was of interest as a nor-
hydroxymethyl analogue of 1. The two known15 deoxyiso-
fagomine analogues ( )-30 and ( )-31 were also investigated
as being useful to compare with 16 and 17. The unintended
product 26 was also investigated, but a racemic sample of 26
was used due to insufficient optically active material. ( )-26
was made in the same way, but starting with ( )-22 and ( )-23.10
Inhibition of the enzymes was measured at 25 ЊC and pH 6.8 in
a sodium phosphate buffer, with the exception of β-glucuron-
idase which was measured at 37 ЊC and pH 4.6. Ki values were
obtained by determination of Km with and without the presence
of an inhibitor and plotting the results in a Hanes plot. In all
cases competitive inhibition (or no inhibition) was observed.
Compound 3 was found to be a submicromolar inhibitor of
two α-fucosidases (Table 117–19). The compound was 5–10 times
more potent than the corresponding isofagomine 3216,17 yet
250–500 times weaker than the deoxynojirimycin analogue 33.5
This shows that (a) the price in inhibition potency of not having
a 2-OH is high and (b) a nitrogen in place of endocyclic oxygen
increases inhibition somewhat. This is similar to what has
previously been seen for galactosidase inhibitors.10 For α-
galactosidases the inhibition order is 1-deoxynojirimycin >
azafagomine > isofagomine. For α-glucosidase inhibition, on
the other hand, the order is azafagomine > 1-deoxynojiri-
mycin > isofagomine.8 The results suggest that an important
drawback of 3 and other azafagomines as α-glycosidase inhib-
J. Chem. Soc., Perkin Trans. 1, 2002, 1190–1198
1193

View Article Online
Table 2 Ki values in µM at pH 6.8, 25 ЊC (NI, no inhibition; —, not tested)
4
34
15
35
36
β-Glucuronidase (bovine liver)
β-Glucosidase (almond)
α-Glucosidase (yeast)
1
7
160
<560a
NI
0.079b
NI
NI
0.065c
0.039d
98c
NI
1200e
—
>560a
a Value from ref. 20. b Value from ref. 16. c Value from ref. 21. d Value from ref. 22. e Value from ref. 23.
Table 3 Ki values in µM at pH 6.8, 25 ЊC (—, not tested)
1
16
30
>3000
1300
—
17
18
29
β-Glucosidase (almond)
α-Glucosidase (yeast)
Isomaltase (yeast)
β-Galactosidase (A. oryzae)
α-Galactosidase (coffee bean)
β-Xylosidase (A. niger)
0.33a
6.9a
9
3
92
4
—
—
—
540
3600
690
—
—
1690
>3000
>3000
650
>3000
—
0.27a
702b, c
934b, c
—
—
—
a Value from ref. 11. b Value from ref. 9. c Obtained on racemic inhibitor.
Table 4 Ki values in µM at pH 6.8, 25 ЊC (—, not tested)
2
( )-26
37
( )-30
( )-31
β-Glucosidase (almond)
α-Glucosidase (yeast)
Isomaltase (yeast)
β-Galactosidase (A. oryzae)
α-Galactosidase (coffee bean)
0.13a
100
—
25
300
>3000
0.11b
86b
7.2b
—
—
2200
>3000
>3000
—
5.6
>3000
>3000
1100
570a
—
0.04a
0.28a
—
—
a Value from ref. 10. b Value from ref. 8.
which is also reflected by both 1 and 2 inhibiting this enzyme
strongly. The only discrepancy between the isofagomine ana-
logues and the azafagomine analogues is the observation that
16 is a moderately good β-glucosidase inhibitor while 30 is
extremely poor. The 3-deoxy analogue (carbohydrate number-
ing) of 2, 26, is a 400–10 000 times weaker inhibitor than 2,
which is another example of the importance of this OH-group
(Table 4).
In conclusion it has been found that azafagomines are good
inhibitors of both α- and β-glucosidases. All the hydroxy
groups are important for binding. In fact the general lack of a
2-hydroxy group (carbohydrate numbering) in an azafagomine
appears to be an important deficiency in the inhibitor in its
binding to α-glycosidases. A 2-hydroxyazafagomine, if suf-
ficiently stable, might indeed be a very strong α-glycosidase
inhibitor since 2-hydroxyisofagomine (noeuromycin) is much
more potent than isofagomine against α-glycosidases and since
azafagomines generally are more potent than isofagomines
against these enzymes.
oration was carried out on a rotatory evaporator with the
temperature kept below 40 ЊC. Glassware used for water-free
reactions was dried for a minimum of 2 hours at 130 ЊC before
use. Columns were packed with silica gel 60 (230–400 mesh) as
the stationary phase. TLC-plates (Merck, 60, F254) were visual-
ized by spraying with cerium sulfate (1%) and molybdic acid
(1.5%) in 10% H2SO4 and heating till coloured spots appeared.
1H NMR, 13C NMR and COSY were carried out on a Varian
Gemini 200 instrument. For samples in water, the water-signal
(δ 4.7) was used as the reference. Mass spectra were run on a
Micromass LC-TOF instrument. Optical rotations are given in
units of 10Ϫ1 deg cm2 gϪ1
.
2,3-Di-O-isopropylidene-5-O-tosyl-D-ribofuranose (6)
Primary alcohol 5 (1.69 g, 8.84 mmol) was dissolved in freshly
dried pyridine (10 ml). Dimethylaminopyridine (50 mg) was
added together with toluene-p-sulfonyl chloride (2.10 g,
11 mmol). The tosyl chloride was added in three portions,
1 hour apart, and the reaction was left to stir for 24 hours at
room temperature. After this time, water and dichloromethane
(50 ml of each) were added. The phases were separated and the
aqueous phase was extracted with dichloromethane (5 × 50 ml).
The combined organic extracts were dried (MgSO4) and the
solvents were removed. The residue underwent column chrom-
atography using EtOAc–pentane (3 : 7) as eluent. This yielded
2.00 g (73%) of the desired product 6. NMR spectra were
identical to those previously reported.14
Experimental
General
Solvents were distilled under anhydrous conditions. All
reagents were used as purchased without further purification.
Pyridine was dried over potassium hydroxide before use. Evap-
1194
J. Chem. Soc., Perkin Trans. 1, 2002, 1190–1198

View Article Online
2,3-Di-O-isopropylidene-5-deoxy-D-ribofuranose (7)
4.30–4.55 (m, 2H, H1a, H2), 4.05 (t, 1H, J 6.2 Hz, H3), 3.07 (s,
3H, SO2CH3), 2.8–3.0 (br s, 1H, H1b), 2.05 [s, 3H, NC(O)CH3],
1.50 (d, 3H, H5), 1.44 [s, 12H, C(CH3)3, CH3], 1.31 (s, 3H,
CH3). Cross-peaks corresponding to J-coupling between the
signals at 4.30–4.55 and 2.8–3.0 were observed in the COSY
spectrum; δC(C6D6) 173.7 [NC(O)CH3], 154.7 [NC(O)OtBu],
109.3 [OC(CH3)2O], 81.4, 78.4, 76.7, 75.7 [C2, C3, C4,
OC(CH3)3], 47.6 (C1), 38.9 (SO2CH3), 28.2 [OC(CH3)3], 27.9,
25.7 [OC(CH3)2O], 20.5, 18.8 [C5, C(O)CH3]. HRMS(ES):
Calcd. for C16H30N2O8 ϩ Na: 433.1621, found 433.1619.
Tosyl compound 6 (2.0 g, 6.49 mmol) was dissolved in freshly
dried dimethoxyethane (50 ml). Sodium iodide (1.947 g,
13 mmol) was added and the reaction mixture was refluxed for
2 hours. Dichloromethane and water (50 ml of each) were
added, whereafter the aqueous phase was extracted with
dichloromethane (3 × 50 ml). The combined organic extracts
were dried (MgSO4) and the solvents were removed. The resi-
due was dissolved in ethanol (28 ml) and triethylamine (2.1 ml)
and Pd–C (10%, 700 mg) were added. Hydrogen pressure was
applied (1 atm, room temperature) and the mixture was stirred
for 2 days. After this period of time an additional portion of
catalyst was added (150 mg) and the mixture was stirred over-
night. The reaction mixture was then filtered through a bed of
Celite® and concentrated to a residue that underwent chrom-
atography through a short column of silica gel using EtOAc–
pentane as eluent (1 : 3). This gave a yield of 832 mg (74%) of
the reduced compound 7. NMR spectra were identical to those
previously reported.13
(3S,4R,5S)-1-Acetyl-4,5-isopropylidenedioxy-3-methylhexa-
hydropyridazine (10)
Mesylate 9 (652 mg, 1.59 mmol) was dissolved in freshly dis-
tilled CH2Cl2 (30 ml) and 2,6-lutidine (1.48 ml, 12.7 mmol) and
TMSOTf (1.44 ml, 7.95 mmol) were added at 0 ЊC. The reaction
was allowed to reach room temperature over 1 hour and stirred
additionally for 6 hours at this temperature. A solution of
aqueous Na2CO3 (10%, 20 ml) was added and the aqueous
phase extracted with CHCl3 (15 × 20 ml). The volume of
organic solvent was reduced to ca. 250 ml, whereafter
anhydrous K2CO3 (3.5 g) was added and the mixture refluxed
for 24 hours. The mixture was filtered and the residue concen-
trated to a pale oil. This underwent filtration through silica
gel (eluent: first AcOEt–pentane 1 : 1, then AcOEt), which
gave 310 mg (91%) of the cyclised product 10 that appeared as
colourless crystals. Rf(AcOEt) = 0.39; mp(uncorr.) 143–144 ЊC;
[α]2D5 Ϫ21.0 (c 1, CHCl3); δH(CDCl3) 1.35, 1.50 [s, 6H,
C(OCH3)2], 1.80 (d, 3H, J 6.6 Hz, H3Ј), 2.14 [s, 3H,
NC(O)CH3], 2.75 (dd, 1H, J5,6ax 8.0 Hz, J6ax,6eq 13.6 Hz, H6ax),
2.98–3.11 (m, 1H, H3), 3.36 (d, 1H, J 13.0 Hz, NH ), 3.96 (dd,
1H, J3,4 2.0 Hz, J4,5 5.2 Hz, H4), 4.23 (m, 1H, H5), 4.43 (dd, 1H,
H6eq); δC(CDCl3) 172.8 [NC(O)CH3], 109.1 [OC(CH3)2O],
73.3, 70.3, 53.6, 42.6 (C3, C4, C5, C6), 28.3, 26.3 [OC(CH3)2O],
20.9, 15.3 [C3Ј, NC(O)CH3]; HRMS(ES): Calcd. for C10H18-
N2O3 ϩ Na: 237.1215, found 237.1237.
1-(2-tert-Butyloxycarbonylhydrazino)-1,5-dideoxy-2,3-di-
O-isopropylidene-D-ribitol (8)
Hemiacetal 7 (680 mg, 3.91 mmol) was dissolved in methanol
(17 ml) and tert-butyl carbazate (1.033 g, 7.82 mmol) was added
together with sodium cyanoborohydride (982 mg, 15.6 mmol).
Acetic acid was added until pH ∼ 5 was reached. The reaction
mixture was stirred at room temperature for 42 hours where-
after a saturated aqueous solution of NaHCO3 (10 ml) was
added. The methanol was then removed under reduced pressure
whereafter the aqueous phase was extracted with dichloro-
methane (3 × 20 ml). The organic extracts were dried (MgSO4),
the solvent removed, and the residue filtered through a short
column of silica gel using EtOAc–pentane (1 : 1) as eluent
(Rf = 0.38). This resulted in 1.096 g (97%) of the Boc-protected
hydrazine 8, which appeared as a colourless oil. [α]2D5 2.75 (c 2,
CHCl3); δH(CDCl3) 7.15 (br s, 1H, HNBoc), 4.70 (br s, 2H, OH,
NHNHBoc), 4.26 (quintet, 1H, J 4.8 Hz, H4), 3.76–3.90 (m,
2H, H2, H3), 3.10 (dd, 1H, J1a,2 9.2 Hz, J1a,1b 12.4 Hz, H1a),
3.00 (dd, 1H, J1b,2 4.2 Hz, H1b), 1.41 [s, 9H, C(CH3)3], 1.25–
(3S,4R,5S)-4,5-Dihydroxy-3-methylhexahydropyridazine (3)
Acetohydrazide 10 (304 mg, 1.42 mmol) was dissolved in
hydrochloric acid (6 M, 25 ml) and heated to 100 ЊC overnight
in a sealed flask. The solvent was removed, and the residue
underwent ion-exchange chromatography (Amberlite IR-120,
Hϩ). The product was released from the resin with 2.5%
NH4OH. This gave a quantitative yield (185 mg) of the desired
hexahydropyridazine 3, which appeared as a colourless powder.
Mp 155 ЊC (decomp.); [α]2D5 Ϫ28.0 (c 1, H2O); δH(D2O) 3.65–3.70
(m, 2H, H4, H5), 2.55–2.8 (m, 3H, H3, H6ax, H6eq), 0.98 (d,
3H, J 6.8 Hz, H3Ј); δC(D2O) 68.8, 67.3 (C4, C5), 54.4, 45.2 (C3,
C6), 14.1 (C3Ј); HRMS(ES): Calcd. for C5H12N2O2 ϩ Na:
155.0796, found 155.0793.
t
1.34 [m, 9H, OC(CH3)2O, H5]; δC(CDCl3) 156.9 (NCO2 Bu),
108.3 [C(OCH3)2], 82.4, 80.9, 74.6, 65.2, 51.8 [C1, C2, C3, C4,
OC(CH3)3], 28.4 [OC(CH3)3], 28.2, 25.6 [OC(CH3)2O], 20.5
(C5). HRMS(ES): Calcd. for C13H26N2O5 ϩ Na: 313.1739,
found 313.1737.
1-(1-Acetyl-2-tert-butoxycarbonylhydrazino)-1,5-dideoxy-2,3-
di-O-isopropylidene-4-O-methylsulfonyl-D-ribitol (9)
Secondary alcohol 8 (1.066 g, 3.68 mmol) was dissolved in
methanol (65 ml) and acetic anhydride (6.5 ml) was added. The
mixture was stirred at room temperature for 2 hours whereafter
a saturated aqueous solution of NaHCO3 (40 ml) was added.
The mixture was stirred for 15 min, whereafter the methanol
was removed under reduced pressure. The aqueous phase was
extracted with CH2Cl2 (5 × 40 ml) and the organic extract
washed with brine (40 ml) before being dried (MgSO4). The
residue after evaporation was dissolved in pyridine (6.6 ml) and
methanesulfonyl chloride (0.43 ml, 5.51 mmol) was added in
two portions (1 equiv. then 0.5 equiv.) separated by an 18 hour
interval. The reaction was finished according to TLC (AcOEt–
pentane 1 : 1) after 24 hours. Water (20 ml) and CH2Cl2 (20 ml)
were added, the phases separated and the aqueous phase was
extracted with CH2Cl2 (5 × 20 ml). The combined organic
extracts were dried (MgSO4) and concentrated. The remaining
oil underwent column chromatography on silica gel (eluent:
first AcOEt–pentane 1 : 2, then 1 : 1). This resulted in 1.225 g
(81%) of the desired compound 9, which appeared as a white
foam. Rf(AcOEt–pentane 1 : 1) = 0.29. [α]2D5 Ϫ15.2 (c 2, CHCl3);
δH(CDCl3) 6.9–7.1 (br s, 1H, NH ), 4.79 (qv, 1H, J 6.2 Hz, H4),
(3R,4R,5R)-1,2-Diacetyl-4,5-dibenzyloxy-3-hydroxymethylhexa-
hydropyridazine (12)
To monoacetyl compound 11 (521 mg, 1.15 mmol) freshly
distilled acetyl bromide (8 ml) was added at 0 ЊC under an
atmosphere of nitrogen. The ice bath was removed and the
mixture was stirred for 7 hours at room temperature. It was then
diluted with chloroform (10 ml) and carefully poured into an
ice-bath cooled flask containing chloroform (10 ml) and satur-
ated aqueous Na2CO3 solution (20 ml). The solution was
further neutralised by adding saturated NaHCO3 solution. The
aqueous phase was then extracted three times with chloroform
and the combined organic phases were dried over anhydrous
MgSO4. The drying agent was removed by filtration, and the
organic solvent was removed under reduced pressure. The
remaining oil was immediately deacetylated by being stirred in
methanol (10 ml) containing a catalytic amount of sodium
methoxide. After 1 hour a small lump of dry ice was added, and
J. Chem. Soc., Perkin Trans. 1, 2002, 1190–1198
1195

View Article Online
the methanol was removed by evaporation under reduced pres-
sure. The remaining oil underwent flash chromatography in
AcOEt–pentane 1 : 1, resulting in 40 mg (7%) of 13 (Rf = 0.5)
and 340 mg (73%) of 12 (Rf = 0.15), which both appeared
as colourless oils. δH(CDCl3) 12, 2.00–2.35 (m, 6H), 3.08 (d,
1H), 3.34–4.05 (m, 6H), 4.32–4.95 (m, 4.5H), 5.12 (t, 0.25H,
J 6.0 Hz), 5.26 (dd, 0.25H, J 4.0 Hz, J 11.4 Hz), 7.20–7.40 (m,
10H); 13, 1.98–2.20 (m, 6H), 3.08 (d, 1H, J 13.2 Hz), 3.40–3.94
(m, 4H), 4.26–4.89 (m, 7.25H), 5.35 (dd, 0.75H, J 4.4 Hz,
J 9.4 Hz), 7.05–7.40 (m, 15H); MS(ES) 13, Calcd. for
C30H34N2O5 ϩ Na: 525.2, found: 525.0; HRMS(ES) 12, Calcd.
for C23H28N2O5 ϩ Na: 435.1898, found: 435.1899.
quenched with water (5 ml). After another hour the reaction
mixture was extracted with AcOEt (3 × 15 ml) and the com-
bined organic phases washed with saturated solutions of
NaHCO3 and Na2SO3 (10 ml). After being dried over MgSO4,
filtered and evaporated, the residue underwent column chroma-
tography (AcOEt–pentane 1 : 1), which resulted in 210 mg
(73%) of iodide 20, which appeared as a colourless oil.
Rf(AcOEt–pentane 1 : 1) = 0.26. [α]2D5 Ϫ31.0 (c 1, CHCl3);
δH(CDCl3) 5.30 (t, 1H, t, J2,3;3,4 6.0 Hz, H3), 4.69 (dd, 1H,
J2,2Јa 4.4 Hz, J2Јa,2Јb 12.0 Hz, H2Јa), 4.56 (dd, 1H, J2,2Јb 5.4 Hz,
H2Јb), 4.1–4.2 (m, 1H, H2 or H4), 4.0–4.1 (m, 1H, H4 or H2),
4.01 (dd, 1H, J4,5a 4.0 Hz, J5a,5b 12.6 Hz, H5a), 3.79 (dd, 1H,
J4,5b 6.0 Hz, H5b), 3.02 (s, 3H, NCH3), 2.08 (s, 3H, CH3),
2.02 (s, 3H, CH3); δC(CDCl3) 170.6, 169.1 [C(O)CH3], 154.0,
153.1 [NC(O)N], 69.9 (C3), 60.4 (C2Ј), 58.8, 50.2 (C2, C5),
25.5 (NCH3), 20.9 [double intensity, C(O)CH3], 16.9 (C4);
HRMS(ES) Calcd. for C12H16N3O6I ϩ Na: 447.9983, found:
447.9984.
(3S,4R,5R)-1,2-Diacetyl-4,5-dibenzyloxyhexahydropyridazine-
3-carboxylic acid (14)
To a solution of 12 (191 mg, 0.46 mmol) in acetone (8 ml) at
0 ЊC, Jones reagent (ca. 0.3 ml; Jones reagent: 2.67 g of CrO3
was added to 2.3 ml of concentrated sulfuric acid and then
diluted to 10 ml) was added in three portions over 30 min. All
starting material had disappeared after 2 hours (TLC monitor-
ing). i-PrOH (1 ml) and water (10 ml) were then added, and the
acetone was removed under reduced pressure. The remaining
water was extracted five times with AcOEt, and the combined
organic phases were dried over MgSO4, filtered and concen-
trated. The remaining oil underwent flash chromatography in
AcOEt–pentane 1 : 1 containing 1% HCO2H (Rf = 0.45) result-
ing in 130 mg (66%) of 14. The compound appeared as a col-
ourless powder, mp 180–183 ЊC (uncorr.); δH(CDCl3) 1.97–2.24
(m, 6H), 3.48–3.51 (m, 1H), 3.70–3.72 (m, 1H), 3.99–4.83 (m,
6H), 5.70 (t, 1H, J 3.4 Hz), 7.00–7.40 (m, 10H); HRMS(ES)
Calcd. for C23H26N2O6 ϩ Na: 449.1689, found: 449.1689.
(؊)-(2R,3S)-3-Acetoxy-2-acetoxymethyl-8-methyl-1,6,8-
triazabicyclo[4.3.0]nonane-7,9-dione (21)
Iodide 20 (210 mg, 0.49 mmol) was dissolved in benzene
(2.5 ml) and n-Bu3SnH (0.288 mg, 0.99 mmol) and AIBN
(0.4 mg) were added. The temperature was raised to 80 ЊC, and
the mixture was stirred for 4.5 hours at this temperature. The
solvent was then removed under reduced pressure, and the resi-
due loaded directly onto a column of silica gel and eluted (first
CHCl3, then AcOEt–pentane 1 : 1) to give 95 mg (65%) of the
reduced product 21. Rf(AcOEt–pentane 1 : 1) 0.19. [α]2D5 Ϫ24.8
(CHCl3, c 1); δH(CDCl3) 5.12 (q, 1H, J 2.8 Hz, H3), 4.44 (m,
1H, H2), 4.32 (dd, 1H, J2,2Јa 6.0 Hz, J2Јa,2Јb 11.2 Hz, H2Јa), 4.16
(dd, 1H, J2,2Јb 7.4 Hz, H2Јb), 3.8–4.0 (m, 1H, H5a), 3.28 (dt, 1H,
J4a,5b 4.8 Hz, J4b,5b;5a,5b 11.8 Hz, H5b), 3.06 (s, 3H, NCH3), 2.00–
2.15 (m, 2H, H4a, H4b), 2.07 (s, 3H, CH3), 2.02 (s, 3H, CH3);
δC(CDCl3) 170.6, 169.9 [C(O)CH3], 154.9, 153.1 [NC(O)N],
64.7 (C3), 60.0 (C2Ј), 54.8, 40.0 (C2, C5), 25.3, 24.7 (NCH3,
C4), 21.1, 20.8 [C(O)CH3]; HRMS(ES): calcd. for C12H17O6N3
ϩ Na: 322.1015, found: 322.1020.
(3S,4R,5R)-4,5-Dihydroxyhexahydropyridazine-3-carboxylic
acid (4)
Carboxylic acid 14 (105 mg, 0.25 mmol) was dissolved in
methanol (12 ml) and 10% Pd–C (50 mg) was added. Hydrogen
pressure (1 atm) was applied and 2 drops of concentrated
hydrochloric acid were added. The mixture was stirred at room
temperature for 2 hours after which the catalyst was removed
by filtration through Celite®. The methanol was removed by
evaporation and to the remaining oil hydrochloric acid (6 M,
20 ml) was added. The flask was sealed and heated to 100 ЊC for
24 hours. The aqueous acid was then removed by evaporation
and the residue loaded onto a column of ion-exchange resin
(Amberlite IR-120, Hϩ), which was then carefully washed. The
compound was released with 5% NH4OH and subjected to
column chromatography eluting with 7 : 2 : 1 i-PrOH–water–
conc. NH4OH (Rf = 0.3). This gave 30 mg (74%) of 4. [α]2D5
Ϫ159.2 (c 0.18, H2O); δH(D2O, neutral) 2.93 (dd, 1H, J5,6ax
7.0 Hz, J6ax,6eq 13.2 Hz, H6ax), 3.42 (dd, 1H, J5,6eq 3.8 Hz,
H6eq), 3.44 (d, 1H, J3,4 6.6 Hz, H3), 3.78 (dt, 1H, H5), 3.87 (t,
1H, H4); δH(D2O, acidified with DCl to pH ∼ 1) 3.19 (dd, 1H,
J5,6ax 4.4 Hz, J6ax,6eq 13.2 Hz, H6ax), 3.60 (dd, 1H, J5,6eq 2.4 Hz,
H6eq), 3.88 (d, 1H, J3,4 4.0 Hz, H3), 4.00 (dt, 1H, H5), 4.18 (t,
1H, H4); δH(D2O, basified with Na2CO3 to pH ∼ 10) 2.56 (dd,
1H, J5,6ax 10.0 Hz, J6ax,6eq 13.0 Hz, H6ax), 3.08 (d, 1H, J3,4 9.2
Hz, H3), 3.16 (dd, 1H, J5,6eq 5.0 Hz, H6eq), 3.47 (t, 1H, H4),
3.56 (dt, 1H, H5); δC(D2O, neutral) 174.2 (COOH), 70.3, 67.3,
63.3 (C3, C4, C5), 48.7 (C6). HRMS(ES) Calcd. for C5H10NO4
ϩ Na: 185.0538, found 185.0539.
(؉)-(3R,4S)-4-Hydroxy-3-hydroxymethylhexahydropyridazine
(16)
Diacetate 21 (92 mg, 0.31 mmol) was deacetylated in methanol
(4 ml) containing a catalytic amount of sodium methoxide at
room temperature. After reaction completion (20 min) the
methanol was removed and N2H5OH (5 ml) was added. The
mixture was refluxed for 24 h. The solvent was then removed
and the residue underwent ion exchange (Amberlite IR-120,
Hϩ). The product was released with 5% NH4OH. Concentra-
tion followed by flash chromatography in EtOH–conc. NH4OH
9 : 1 (Rf = 0.44) produced 23 mg (57%) of 16. [α]2D5 24.7
(c 1, H2O); δH(D2O) 3.82 (dd, 1H, J3,3Јa 3.0 Hz, J3Јa,3Јb 12.0 Hz,
H3Јa), 3.61 (dd, 1H, J3,3Јb 6.3 Hz, H3Јb), 3.54 (ddd, 1H,
J4,5eq 5.0 Hz, J3,4 9.5 Hz, J4,5ax 11.0 Hz, H4), 3.10 (ddd, 1H,
J5eq,6eq 2.7 Hz, J5ax,6eq 5.0 Hz, J6ax,6eq 13.1 Hz, H6eq), 2.77 (dt,
1H, J5eq,6ax 3.0 Hz, H6ax), 2.61 (ddd, 1H, H3), 2.05 (tdd, 1H,
J5ax,5eq 13.1 Hz, H5eq), 1.50 (ddt, 1H, H5ax); δC(D2O) 66.7, 64.5
(C3Ј, C4), 60.6 (C3), 46.3 (C6), 33.4 (C5).
(؊)-(2S,4S)-4-Acetoxy-2-acetoxymethyl-8-methyl-1,6,8-
triazabicyclo[4.3.0]nonane-7,9-dione (25) and (؉)-(2R,3R)-3-
acetoxy-2-acetoxymethyl-8-methyl-1,6,8-triazabicyclo[4.3.0]-
nonane-7,9-dione (24)
(؊)-(2R,3R,4R)-3-Acetoxy-2-acetoxymethyl-4-iodo-8-methyl-
1,6,8-triazabicyclo[4.3.0]nonane-7,9-dione (20)
A mixture of iodides 22 and 23 (506 mg, 1.2 mmol) was dis-
solved in benzene (5 ml) and n-Bu3SnH and AIBN were added
according to the procedure for synthesis of compound 3. Work-
up and purification were done as described earlier. This yielded
162 mg (46%) of 25 [Rf(AcOEt–pentane 1 : 1) = 0.12] and 75 mg
(21%) of 24 [Rf(AcOEt–pentane 1 : 1) = 0.16]. 25: [α]2D5 = Ϫ21.3
Epoxide 19 (144 mg, 0.68 mmol) was dissolved in acetic acid
(2 ml) and a 57% aqueous solution of HI was added (173 mg,
1.35 mmol). The reaction mixture was allowed to stir at room
temperature for 1 h. Acetic acid anhydride (5 ml) was added
and allowed to react for 3 hours before the reaction was
1196
J. Chem. Soc., Perkin Trans. 1, 2002, 1190–1198

View Article Online
(c 1, CHCl3); δH(CDCl3) 5.13 (qv, 1H, J 3.2 Hz, H4), 4.20–4.52
(m, 3H, H2, H2Јa, H2Јb), 4.01 (dd, 1H, J5a,5b 13.0 Hz, H5a) 3.26
(dd, 1H, H5b), 3.04 (s, 3H, NCH3), 2.06 [s, 3H, C(O)CH3], 2.00
[s, 3H, C(O)CH3], 2.00–2.10 (m, 2H, H3a, H3b); δC(CDCl3):
170.9, 169.9 [C(O)CH3], 155.3, 153.4 [NC(O)N], 64.5, 62.6
(C4, C2Ј), 50.5, 47.9 (C2, C5), 28.8 (C3), 25.4 (NCH3), 21.4,
20.9 [C(O)CH3]. HRMS(ES): Calcd. for C12H17O6N3 ϩ Na:
322.1015, found: 322.1009. 24: [α]2D5 5.5 (CHCl3, c 1);
δH(CDCl3): 5.10 (td, 1H, J 5.0 Hz, J 10.2 Hz, H3), 4.41–4.62
(m, 3H, H2, H2Јa, H2Јb), 3.96 (td, 1H, J4a,5a;4b,5a 4.0 Hz,
J5a,5b 12.2 Hz, H5a), 3.23 (dt, 1H, J4a,5b 4.0 Hz, H5b), 3.05 (s,
3H, NCH3), 1.99–2.16 (m, 2H, H4a, H4b), 2.11 [s, 3H,
C(O)CH3], 2.00 [s, 3H, C(O)CH3]; δC(CDCl3) 170.8, 169.7
[C(O)CH3], 154.8, 153.1 [NC(O)N], 67.7, 59.1 (C2Ј, C3),
53.6, 42.5 (C2, C5), 25.4 (double intensity, C4, NCH3), 21.1,
20.8 [C(O)CH3]. HRMS(ES): Calcd. for C12H17O6N3 ϩ Na:
322.1015, found: 322.1014.
(3R,4R,5R)-4,5-Dihydroxy-3-methylhexahydropyridazine (18)
Diacetate 28 (36 mg, 0.09 mmol) was dissolved in methanol
(5 ml) and 10% Pd–C (50 mg) was added. Hydrogen pressure
(1 atm) was applied and 2 drops of concentrated hydrochloric
acid were added. The mixture was stirred at room temperature
for 18 hours after which the catalyst was removed by filtration
through Celite®. The methanol was removed by evaporation
and to the remaining oil hydrochloric acid (6 M, 6 ml) was
added. The flask was sealed and heated to 100 ЊC for 24 hours.
The aqueous acid was then removed by evaporation and the
residue subjected to column chromatography in 99 : 1 EtOH–
conc. NH4OH. This gave 8 mg (67%) of 18. [α]2D5 Ϫ12 (H2O,
c 0.3); δH(D2O) 3.65 (ddd, 1H, H5, J45 8.8 Hz, J56eq 4.6 Hz,
J56ax 10.4 Hz, H5), 3.36 (m, 1H, J6ax6eq 13.2 Hz, H6eq), 3.24
(t, 1H, J34 9.6 Hz, H4), 2.91 (dq, 1H, J3,3Ј 6.6 Hz, H3), 2.78 (dd,
1H, H6ax), 1.22 (d, 3H, CH3); δC(D2O) 76.8, 70.7 (C4, C5), 57.2
(C3), 51.4 (C6), 14.2 (C3Ј).
(؊)-(3S,5S)-5-Hydroxy-3-hydroxymethylhexahydropyridazine
(17)
X-Ray crystallography†
The crystal structure of 10, (3S,4R,5S)-1-Acetyl-4,5-isopropyl-
idenedioxy-3-methyl-hexahydropyridazine, C10H18N2O3, M =
214.27, was solved using data collected at 120 K from a colour-
less plate on a SIEMENS SMART CCD diffractometer. The
crystals are monoclinic, space group P21, with unit cell: a =
6.436(1) Å, b = 10.547(2) Å, c = 8.763(2) Å, β = 105.899(3)Њ, V =
572.1(2) Å3, Z = 2, µ = 0.092, Rint = 0.078, for 6253 measured
reflections of which 3168 were independent. Direct methods
were applied24 for the structure solution, and the structure
Diacetate 25 (88 mg, 0.29 mmol) was deprotected as described
for the synthesis of 16. This yielded 25 mg (65%) of the desired
product 17. Rf(EtOH–conc. NH4OH 9 : 1) = 0.4. [α]2D5 Ϫ5.0 (c 1,
H2O); δH(D2O) 1.15 (q, 1H, J 11.8 Hz, H4ax), 1.96–2.10 (dm,
1H, H4eq), 2.38 (dd, 1H, J5,6ax 10.8 Hz, J6ax,6ax 12.3 Hz, H6ax),
2.78–2.95 (m, 1H, H3), 3.10 (dd, 1H, J5,6eq 4.8 Hz, H6eq),
3.46 (dd, 1H, J3,3Јa 6.2 Hz, J3Јa,3Јb 11.7 Hz, H3Јa), 3.55 (dd, 1H,
J3,3Јb 4.8 Hz, H3Јb), 3.76 (m, 1H, H5); δC(D2O) 34.6 (C4),
51.8 (C6), 57.2 (C3), 62.8, 65.5 (C3Ј, C5). HRMS(ES) Calcd.
for C5H12O2N2 ϩ H: 133.0977, found: 133.0977.
refined by least-squares methods to a final R = 0.039, Rw
=
0.046, GoF = 1.37 for 2811 reflections with I > 3σ(I ) and 208
parameters. The structure is held together in chains by hydro-
gen bonds from N2 to the keto oxygen, O15, in a molecule
related by a screw axis.
(؊)-(3R,4R)-4-Hydroxy-3-hydroxymethylhexahydropyridazine
(26)
Diacetate 24 (75 mg, 0.25 mmol) was deprotected as described
for the synthesis of 16. This yielded 21 mg (63%) of the desired
product 26. Rf(EtOH–CHCl3–conc. NH4OH 7 : 2 : 1) = 0.26.
[α]2D5 Ϫ10.0 (c 0.25, H2O); δH(D2O) 3.99 (q, 1H, J 1.8 Hz, H4),
3.56 (d, 2H, J3,3Ј 7.0 Hz, H3Јa, H3Јb), 3.00–3.10 (m, 1H, H6ax),
2.94 (dt, 1H, H3), 2.81 (td, 1H, J5ax,6eq;5eq,6eq 4.0 Hz, J6ax,6eq 13.2
Hz, H6eq), 1.77–1.86 (m, 2H, H5ax, H5eq); δC(D2O) 63.4 (C4),
61.0 (double intensity, C3, C3Ј), 41.6 (C6), 30.8 (C5).
Enzyme inhibition
The enzyme assays were carried out as described previously.9
All assays were performed at pH 6.8 and 25 ЊC except the
β-glucuronidase assay which was performed at pH 4.6 and
37 ЊC. The inhibition constants (Ki) were obtained from the
formula Ki = [I]/(KMЈ/KMϪ1), where KMЈ and KM are Michaelis–
Menten constants with and without inhibitor present. KMЈ and
KM were obtained from a Hanes plot, which also was used to
ensure that inhibition was competitive. The following KM values
(without inhibitor) were obtained using 4-nitrophenyl glyco-
sides as substrates and the above conditions: β-glucuronidase
(bovine liver), 1 mM; β-glucosidase (almonds), 3.8 mM; α-gluco-
sidase (baker’s yeast), 0.25 mM; isomaltase (yeast), 1.3 mM;
β-galactosidase (Aspergillus oryzae), 1.4 mM; α-galactosidase
(green coffee beans), 0.6 mM; α-fucosidase (human placenta),
0.23 mM; α-fucosidase (bovine kidney), 0.24 mM.
(3S,4R,5R)-1,2-Diacetyl-4,5-dibenzyloxy-3-chloromethylhexa-
hydropyridazine (27)
Primary alcohol 12 (53 mg, 0.13 mmol) was dissolved in
pyridine (0.8 ml). The mixture was cooled to 0 ЊC and meth-
anesulfonyl chloride was added (45 µl, 4.5 equiv.). The mesyl
chloride was added in 3 portions with 1 hour intervals, and the
mixture was left to stir for 4 hours at 80 ЊC with a pinch of CsCl
(approx. 10 mg). After this time the solvents were removed, and
the residue underwent column chromatography using EtOAc–
pentane (1 : 3) as eluent. This yielded 36 mg (65%) of the
desired product 27. The NMR spectrum of 27 was complex due
to extensive rotamer formation. HRMS(ES): Calcd. for C23H27-
O4N2Cl ϩ Na: 453.1557, found: 453.1558.
† CCDC reference number 178401. See http://www.rsc.org/suppdata/
p1/b2/b200884j for crystallographic files in .cif or other electronic
format.
References
(3R,4R,5R)-1,2-Diacetyl-4,5-dibenzyloxy-3-methylhexahydro-
pyridazine (28)
1 L. J. Scott and C. M. Spencer, Drugs, 2000, 59, 521.
2 G. S. Jacob, P. Scudder, T. D. Butters, I. Jones and D. C. Tiemeier,
in Natural Products as Antiviral Agents, eds. C. K. Chu and
H. G. Cutler, Plenum Press, New York, 1992, pp. 137–151.
3 J. Alper, Science, 2001, 291, 2338.
4 W. G. Laver, N. Bischofberger and R. G. Webster, Sci. Am., 1999,
January, 280(1), 78.
5 A. E. Stütz, Iminosugars as Glycosidase Inhibitors: Nojirimycin and
Beyond, Wiley–VCH, Weinheim, 1999.
6 T. D. Heightman and A. T. Vasella, Angew. Chem., Int. Ed., 1999, 38,
750.
Chloride 27 (36 mg, 0.084 mmol) was dissolved in toluene
(2.5 ml) and n-Bu3SnH (42.5 µl, 2.2 equiv.) and AIBN (0.4 mg)
were added. The temperature was raised to 110 ЊC and the mix-
ture was stirred for 1.5 hours at this temperature. The solvent
was then removed under reduced pressure and the residue sub-
jected to column chromatography on silica gel in AcOEt–
pentane 1 : 7 to give 22 mg (66%) of the reduced product 28.
The NMR spectrum of 28 was complex due to extensive
rotamer formation. HRMS(ES): Calcd. for C23H28O4N2 ϩ Na:
419.1947, found: 419.1947.
7 D. L. Zechel and S. G. Withers, Acc. Chem. Res., 2000, 33, 11.
8 M. Bols, Acc. Chem. Res., 1998, 31, 1.
J. Chem. Soc., Perkin Trans. 1, 2002, 1190–1198
1197

View Article Online
9 M. Bols, R. Hazell and I. Thomsen, Chem. Eur. J., 1997, 3, 940.
10 H. H. Jensen and M. Bols, J. Chem. Soc., Perkin Trans. 1, 2001, 905.
11 B. V. Ernholt, I. B. Thomsen, A. Lohse, I. Plesner, K. B. Jensen,
R. G. Hazell, X. Liang, A. Jacobsen and M. Bols, Chem. Eur. J.,
2000, 6, 278.
12 X. Liang and M. Bols, J. Org. Chem., 1999, 64, 8485.
13 G. M. Escandar and L. F. Sala, Org. Prep. Proced. Int., 1990, 22,
623.
14 D. H. R. Barton, J. Camara, X. Cheng, S. D. Géro, J. Cs. Jaszberenyi
and B. Quiclet-Sire, Tetrahedron, 1992, 48, 9261.
15 H. H. Jensen, L. Lyngbye, A. Jensen and M. Bols, Chem. Eur. J.,
2002, 8, 1218.
18 X. Liang, B. O. Petersen, J. Duus and M. Bols, J. Chem. Soc., Perkin
Trans. 1, 2001, 2764.
19 G. Legler, A. E. Stütz and H. Immich, Carbohydr. Res., 1995, 272,
17.
20 T. Tsuruoka, H. Fukuyasu, M. Ishi, T. Usui, S. Shibahara and
S. Inouye, J. Antibiot., 1996, 49, 155.
21 Y. Nishimura, E. Shitara, H. Adachi, M. Toyoshima, M. Nakajima,
Y. Okami and T. Takeuchi, J. Org. Chem., 2000, 65, 2.
22 T. Niwa, T. Tsuruoka, S. Inouye, Y. Naito, T. Koeda and T. Niida,
J. Biochem., 1972, 72, 207.
23 G. Legler and F. Witassek, Hoppe-Seylers Z. Physiol. Chem., 1974,
355, 617.
16 Y. Ichikawa, Y. Igarashi, M. Ichikawa and Y. Suhura, J. Am. Chem.
Soc., 1998, 120, 3007.
17 A. Hansen, T. M. Tagmose and M. Bols, Tetrahedron, 1997, 53, 697.
24 A. Altomare, G. Cascarano, C. Giacovazzo, A. Guagliardi, A. G. G.
Moliterni, M. C. Burla, G. Polidori, M. Camalli and R. Spagna,
SIR97, University of Bari, Italy, 1997.
1198
J. Chem. Soc., Perkin Trans. 1, 2002, 1190–1198