Glycosylation with ulosonates under Mitsunobu conditions: Scope and limitations

Article abstract of DOI:10.1039/d0nj03044a

A systematic study was performed by using Mitsunobu conditions (diethyl azodicarboxylate, Ph3P or n-Bu3P in THF or CH3CN) for glycosylations with methyl 3,4,5,7-tetra-O-benzoyl-α-d-gluco-hept-2-ulopyranosonate. From a set of 47 O-, N-, S- and C-nucleophiles, phenols and N-hydroxy compounds with a pKa of 5-8, phthalimide, benzotriazole, 6-chloropurine, an oxazolidinedione and several tetrazoles with a pKa of 4-8, and thiophenol gave the corresponding products in moderate to very good yields, while C-nucleophiles were unreactive. Trihaloacetanilides underwent O-glycosylation to give O-glycosyl-N-aryl trihaloacetimidates which could not be made by conventional O-imidoylations. All reactions were highly stereoselective to produce the β(d) anomers only. With methyl (5-acetamido-4,7,8,9-tetra-O-acetyl-3,5-dideoxy-d-glycero-d-galacto-2-nonulopyranose)onate phenols and benzotriazole resulted in the expected products, but all other nucleophiles failed to react. While these transformations gave anomeric mixtures in a ratio close to 1?:?1 at room temperature, the α-selectivity increased to 92?:?8 at -30 °C. An o-nitrophenyl sialoside was converted to a spiro-benzoxazinone derivative by reduction of the nitro group and subsequent spontaneous ring closure.

Full text of DOI:10.1039/d0nj03044a

View Article Online
View Journal
NJC
New Journal of Chemistry
Accepted Manuscript
A journal for new directions in chemistry
This article can be cited before page numbers have been issued, to do this please use: N. Kánya, S. Kun,
G. Batta and L. Somsák, New J. Chem., 2020, DOI: 10.1039/D0NJ03044A.
Volume 42
This is an Accepted Manuscript, which has been through the
Number
6
21 March 2018
Pages 3963-4776
Royal Society of Chemistry peer review process and has been
accepted for publication.
NJC
New Journal of Chemistry
rsc.li/njc
A journal for new directions in chemistry
Accepted Manuscripts are published online shortly after acceptance,
before technical editing, formatting and proof reading. Using this free
service, authors can make their results available to the community, in
citable form, before we publish the edited article. We will replace this
Accepted Manuscript with the edited and formatted Advance Article as
soon as it is available.
You can find more information about Accepted Manuscripts in the
Information for Authors.
Please note that technical editing may introduce minor changes to the
text and/or graphics, which may alter content. The journal’s standard
Terms & Conditions and the Ethical guidelines still apply. In no event
shall the Royal Society of Chemistry be held responsible for any errors
or omissions in this Accepted Manuscript or any consequences arising
from the use of any information it contains.
ISSN 1144-2546
PAPER
Chechia Hu, Tzu-Jen Lin et al.
Yellowish and blue luminescent graphene oxide quantum dots
prepared via a microwave-assisted hydrothermal route using
H2O2 and KMnO4 as oxidizing agents
rsc.li/njc

Page 1 of 37
New Journal of Chemistry
1
2
3
4
View Article Online
DOI: 10.1039/D0NJ03044A
Glycosylation with Ulosonates under Mitsunobu Conditions:
5
6
Scope and Limitations
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Nándor Kánya, Sándor Kun, Gyula Batta, László Somsák
Department of Organic Chemistry, University of Debrecen,
PO Box 400, H-4002 Debrecen, Hungary
Abstract
A systematic study was performed by using Mitsunobu conditions (diethyl azodicarboxylate,
Ph3P or n-Bu3P in THF or CH3CN) for glycosylations with methyl 3,4,5,7-tetra-O-benzoyl-α-
D-gluco-hept-2-ulopyranosonate. From a set of 47 O-, N-, S- and C-nucleophiles, phenols and
N-hydroxy compounds with a pKa of 5-8, phthalimide, benzotriazole, 6-chloropurine, an
oxazolidinedione and several tetrazoles with a pKa of 4-8, and thiophenol gave the
corresponding products in moderate to very good yields, while C-nucleophiles were
unreactive. Trihaloacetanilides underwent O-glycosylation to give O-glycosyl-N-aryl
trihaloacetimidates which could not be made by conventional O-imidoylations. All reactions
were highly stereoselective to produce the β(D) anomers only. With methyl (5-acetamido-
4,7,8,9-tetra-O-acetyl-3,5-dideoxy-D-glycero-D-galacto-2-nonulopyranose)onate phenols and
benzotriazole resulted in the expected products, but all other nucleophiles failed to react.
While these transformations gave anomeric mixtures in a ratio close to 1 : 1 at room
temperature, the α-selectivity raised to 92 : 8 at -30 °C. An o-nitrophenyl sialoside was
converted to a spiro-benzoxazinone derivative by reduction of the nitro group and subsequent
spontaneous ring closure.
Keywords
Mitsunobu glycosylation; stereoselective reaction; O-glycosyl trihaloacetimidate; O-
glycoside; N-glycoside; S-glycoside.
 Corresponding author. Tel: +3652512900 ext 22348; e-mail: somsak.laszlo@science.unideb.hu.
1

New Journal of Chemistry
Page 2 of 37
1
2
3
4
View Article Online
DOI: 10.1039/D0NJ03044A
Introduction
5
6
7
8
9
The Mitsunobu reaction, during its more than half a century long history, has become a very
useful and versatile tool for exchanging a hydroxyl group to a nucleophile in the presence of a
trisubstituted phosphine and an azodicarboxylic acid diester.1-3 The nucleophilic partner is
most frequently applied in its acidic form (NuH) with a pKa ≤ 15, preferably below 11. This
method has become well-appreciated in carbohydrate chemistry, as well, and found
applications in substitutions of all types of hydroxyl groups on the sugar ring as amply
discussed in the above reviews. The replacement of the glycosidic hydroxyl group is of
utmost importance in the synthesis of various O-, S- and N-glycosides and the application of
the Mitsunobu protocol in this field has also been surveyed recently.4 The formation of C-
glycosyl compounds from sugar derived lactols by using this methodology, not covered in the
previous overview, was also studied in details.5 In general, the substrates of the Mitsunobu-
type transformations are primary and secondary alcohols. Tertiary alcohols are much less
reactive, therefore, only sporadic examples can be found in the literature for their use.2
Tertiary α-hydroxy esters gave only traces of N-alkylated hydrazinedicarboxylates under
Mitsunobu conditions.6, 7 The anomeric hydroxyl group of aldoses resembles secondary
alcohols and their transformations are well known.4, 5 On the other hand, the glycosidic OH of
ketoses is similar to tertiary alcohols, and it is not very surprising that very few reactions were
carried out with ketose derived lactols as substrates. Thus, transformations of O-benzoyl
protected fructofuranoses were reported in two papers. The formation of three products were
observed in a reaction of 1,3,4,6-tetra-O-benzoyl-D-fructofuranose with p-nitrophenol in the
presence of tri-n-butyl-phosphine (n-Bu3P) and diethyl azodicarboxylate (DEAD) in DMF.
The major component of the mixture showed the p-nitrophenyl group to be present (1H NMR,
IR) while 13C NMR suggested that it was an α-furanoside, however the reaction was not
investigated further.8 Later, facile transformation of the same substrate was reported in the
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2

Page 3 of 37
New Journal of Chemistry
1
2
3
4
View Article Online
DOI: 10.1039/D0NJ03044A
presence of Ph3P and diisopropyl azodicarboxylate (DIAD) in THF with MeOH (used as
5
6
solvent), i-PrOH, o- and p-nitrophenol, p-methoxybenzylamine.9
7
8
9
Another type of glycosyl donor with a tertiary glycosidic hydroxyl group was the neuraminic
acid derivative methyl (5-acetamido-4,7,8,9-tetra-O-acetyl-3,5-dideoxy-D-glycero-D-galacto-
2-nonulopyranose)onate which was reacted with a range of phenols in the presence of Ph3P
and DEAD in CH3CN to give mixtures of the expected aryl sialoside anomers (α/β ratios
50:50 to 74:26, not separated) in 50-90 % yields. In many of these reactions the formation of
the corresponding Neu5Ac-2-en type glycal was also observed in various ratios (from traces
to 80 % relative to the glycosides).10 In the synthesis of fluorogenic reagents for
neuraminidase titration a similar chemistry was applied to get difluoro-4-methylumbelliferyl
3-fluoro-3-deoxy sialosides (Ph3P, DIAD, THF, 19-56 % yields; formation of by-products
was not reported, maybe not even investigated).11
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Recently, we have applied the Mitsunobu methodology to get aryl glycosides from methyl
3,4,5,7-tetra-O-benzoyl-α-D-gluco-hept-2-ulopyranosonate (1, Table 1) and sterically
hindered nitrophenols and nitronaphthols (Ph3P, DEAD, THF, 76-87 % yields) with complete
inversion of the anomeric configuration (Table 1, entries 1, 2). This transformation proved
superior to the nucleophilic substitution reaction of methyl (3,4,5,7-tetra-O-benzoyl-α-D-
gluco-hept-2-ulopyranosyl bromide)onate by the mentioned phenols.12 In these reactions
formation of the corresponding elimination product methyl 2,6-anhydro-3,4,5,7-tetra-O-
benzoyl-D-arabino-hept-2-enonate13 (2) was observed by TLC and isolated in one case in
11% yield.14
In the present article our recent studies are showcased on the reactions of methyl 3,4,5,7-tetra-
O-benzoyl-α-D-gluco-hept-2-ulopyranosonate (1) and methyl (5-acetamido-4,7,8,9-tetra-O-
acetyl-3,5-dideoxy-D-glycero-D-galacto-2-nonulopyranose)onate (33) with further O- as well
as N-, S- and C-nucleophiles under Mitsunobu conditions.
3

New Journal of Chemistry
Page 4 of 37
1
2
3
4
5
6
7
View Article Online
DOI: 10.1039/D0NJ03044A
Results and Discussion
Experiments with 1 and O-nucleophiles
8
9
As a continuation of our former study12 on the transformation of 115 with some nitrophenol
types (3 and 4 to give 7 and 8, respectively, in Table 1), a variety of further O-nucleophiles
with pKa values ranging from 0.3 to 10.8 (5 and 6 in Table 1 and compounds shown in Fig. 1)
were reacted with 1 under the previously applied conditions (3 equiv. of DEAD, PPh3, and
NuH each, THF, rt). In each case the formation of the glycal-type side product 2 was detected
by TLC, however, it was never isolated.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
From the ten new O-nucleophiles only N-hydroxyphthalimide (5) and 1-hydroxybenzotriazole
(6) gave the corresponding products 9 and 10, respectively, as single anomers in acceptable
yields. Compounds 5 and 6 have similar acidities to those of 3 and 4. Compounds less acidic
than 3-6 such as benzaldoxime, hexafluoro-isopropanol, and acetohydroxamic acid proved
unreactive, while more acidic ones like some carboxylic acids and nitrophenols caused the
decomposition of 1 (Fig. 1). From these experiments it can be concluded that the ideal range
of pKa with O-nucleophiles is approximately between (5)6-8.
4

Page 5 of 37
New Journal of Chemistry
1
2
3
View Article Online
DOI: 10.1039/D0NJ03044A
Table 1. Successful transformations of 1 with O-nucleophiles.a
4
5
6
7
8
DEAD (3 eq.)
Ph3P (3 eq.)
OBz
O
OBz
O
OBz
O
BzO
BzO
BzO
BzO
BzO
BzO
NuH
(3 eq.)
COOMe
Nu
+
+
COOMe
THF, rt
BzO
BzO
MeOOC
OH
9
BzO
2
1
7-10
3-6
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Entry
1.
NuH
CH3
pKa of NuHb
Product
Yield (%)
7812
CH3
OBz
O
BzO
BzO
HO
O
7.8
BzO
MeOOC
O2N
CH3
O2N
CH3
3
7
OBz
O
BzO
BzO
HO
O2N
O
2.
3.
4.
6.1
6.1
7.4
8712
BzO
MeOOC
O2N
4
8
O
OBz
O
O
BzO
BzO
O
N OH
N
51
BzO
MeOOC
O
O
5
6
9
OBz
OH
O
BzO
BzO
N
O
42c
N
N
N
BzO
N
MeOOC
N
10
a Order of addition: 1. Ph3P, 2. NuH, 3. substrate 1, 4. DEAD
b pKa values are taken from a ScienceFinder search of the corresponding materials.
c Corrected for 80% conversion of 1.
5

New Journal of Chemistry
Page 6 of 37
1
2
3
4
5
6
7
8
9
View Article Online
DOI: 10.1039/D0NJ03044A
OH
O
F3C
CF3
N
OH
N
H
9.3
OH
9.7
no reaction
10.8
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
OBz
O
COOH
4.2
BzO
BzO
COOH
BzO
4-5
NO2
HO
HO
H2N
COOH
O2N
NO2
O2N
NO2
4.1
0.33
2.3
decomposition
Figure 1. O-Nucleophiles with pKa values which
failed in attempted transformations of 1 (pKa values
are taken from a ScienceFinder search of the
corresponding materials).
Experiments with 1 and N-nucleophiles
Before exploring the reaction of 1 with different N-nucleophiles, optimization experiments
were carried out with phthalimide (11). The first attempt was performed under the same
conditions as with the O-nucleophiles (Table 2, entry 1) to give the substitution product 12 in
30% yield. Changing the phosphine, the azodicarboxylate, the solvent or varying the
temperature did not give better yields for 12 (entries 2-8). In a last experiment the order of
addition of the reagents was modified. Until then 1, the phosphine and 11 were dissolved in
THF and the azodicarboxylate was added last. In the attempt in entry 9 DEAD was first added
to the phosphine in THF, followed by 11 and finally the substrate 1, however, no significant
difference was observed in the yield. Therefore, we have used the original reaction conditions
for all further experiments. The modest yield can be explained by the formation of glycal 2
(which was isolated in 21% yield from the crude mixture obtained in entry 1) and the
6

Page 7 of 37
New Journal of Chemistry
1
2
3
4
5
6
7
8
9
10
hydrazindicarboxylate formed in the reaction may also be a competitive nucleoDpOhI:il10e..11003V9iIe/Dwt0AiNrstiJc0le3O0n4l4inAe
noteworthy, that a reaction of the glycosyl bromide of 1 (Br instead of OH) with potassium
phthalimide gave only glycal 2.
Table 2. Optimization experiments with 1 and phthalimide 11.a
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
R3P (3 eq.)
O
O
OBz
O
OBz
O
(R'OCO)2N2 (3 eq.)
solvent
BzO
BzO
BzO
BzO
+
N
NH
COOMe
BzO
MeOOC
BzO
OH
O
O
(3 eq.)
11
1
12
Entry
1.
R
R’
Et
Et
iPr
Bn
Et
Et
Et
Et
Et
Solvent
THF
Temperature (°C)
Reaction time (h)
overnight
48
Yield of 12 (%)
Ph
Ph
Ph
Ph
rt
-5
30
2.
THF
28
3.
THF
rt
overnight
overnight
overnight
overnight
overnight
3
27
4.
THF
rt
26
5.
n-Bu
Ph
THF
rt
trace
6.
Toluene
MeCN
THF
rt
30
7.
Ph
rt
29
decomposition
33
8.
Ph
reflux
rt
9.b
Ph
THF
overnight
a Order of addition: 1. Ph3P, 2. NuH 11, 3. substrate 1, 4. DEAD
b Order of addition: 1. Ph3P, 2. DEAD, stirred at rt for 10 min, 3. NuH 11, 4. substrate 1
Altogether 20 N-nucleophiles in a pKa range of ~12-4 were subjected to the Mitsunobu
conditions with 1. Besides phthalimide 11 compounds 13-20 (Table 3) gave isolable products
21-28, respectively, in modest yields ranging from 30 to 42 %. Several other N-nucleophiles
shown in Fig. 2 were unreactive or gave only multicomponent mixtures without an isolable
main product. It is noteworthy that in each reaction a single anomer was formed (Table 3,
entries 1-9) and with bidentate nucleophiles only one regioisomer was obtained in most cases
(entries 3, 5, 7-9). For the 6-chloropurine derivative 22 (entry 3) glycosylation in the N-9
position, as depicted, was established on the basis of the 13C chemical shift of the purine C-5
7

New Journal of Chemistry
Page 8 of 37
1
2
3
4
View Article Online
DOI: 10.1039/D0NJ03044A
atom at 131.9 ppm following a literature study.16 The reaction with tetrazole 17 (entry 6)
5
6
7
8
resulted in two isomers 25a and 25b, due to the tautomerism of the tetrazole ring. Substitution
in the N-1 vs N-2 positions can be established by the 13C chemical shift difference of the
tetrazole C-5 carbons: N-1 substituted derivatives resonate around 141 ppm, while N-2
isomers at ~152 ppm.17-19 The measured values of 141.3 ppm for 25a and 152.7 ppm for 25b
clearly identify the two isomers as depicted in entry 6. In each reaction with 5-substituted
tetrazoles (entries 5, 7-9), the formation of the 2-glycosyl compounds was observed only and
no product was derived from the 1H-tautomers of the respective tetrazoles. This substitution
pattern followed from the tetrazole C-5 resonances appearing above 162 ppm as shown in
some previous examples.19 Similarly to the O-nucleophiles, compounds with pKa values above
8.5 did not react. The reactive compounds 11, 13-20 have pKa values from 8.3 to 4.0,
nevertheless, this is well overlapping with the pKa range of the failed compounds (Fig. 2). The
failure with the purine derivatives xanthine and uric acid (compare with 6-chloropurine 14 in
entry 3) and the sulfonamides might be due to steric hindrance of the nucleophilic centres.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 3. Successful reactions of 1 with N-nucleophiles.a
DEAD (3 eq.)
Ph3P (3 eq.)
OBz
O
OBz
O
BzO
BzO
BzO
BzO
NuH
(3 eq.)
COOMe
Nu
+
THF, rt
BzO
BzO
MeOOC
OH
12, 21-28
1
11, 13-20
Entry
1.
NuH
pKa of NuHb
Product
O
Yield (%)
O
OBz
O
BzO
BzO
NH
N
8.3
30
BzO
MeOOC
O
O
12
11
13
OBz
O
N
BzO
BzO
N
N
2.
8.2
32
N
N
H
N
BzO
MeOOC
21
8

Page 9 of 37
New Journal of Chemistry
1
2
3
4
5
6
7
8
9
View Article Online
DOI: 10.1039/D0NJ03044A
N
N
N
N
N
Cl
OBz
O
N
BzO
BzO
N
H
3.
7.5
31
N
BzO
MeOOC
Cl
14
22
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
O
H
N
OBz
O
O
O
BzO
BzO
N
O
4.
5.
6.1
6.0
33
32
30
O
BzO
MeOOC
O
15
23
NH2
NH2
OBz
O
N
N
BzO
BzO
N
NH
N
N
BzO
MeOOC
N
16
N
24
OBz
O
N
N
BzO
BzO
N
N
BzO
MeOOC
NH
N
25a
6.
4.9
N
17
N
OBz
N
N
O
BzO
BzO
N
33
42
N
BzO
MeOOC
25b
OBz
O
N
N
N
N
N
BzO
BzO
N
N
7.
8.
4.5
4.0
N
H
BzO
MeOOC
18
19
26
OBz
O
N
N
N
N
N
N
BzO
BzO
Br
40
N
N
H
BzO
MeOOC
Br
27
OBz
O
N
N
N
BzO
BzO
OBz
O
OBz
HN
N
N
BzO
BzO
N
O
MeOOC
BzO
9.
4.4
36
N
BzO
20
BzO
OBz
OBz
28
a Order of addition: 1. Ph3P, 2. NuH, 3. substrate 1, 4. DEAD
b pKa values are taken from a ScienceFinder search of the corresponding materials.
9

New Journal of Chemistry
Page 10 of 37
1
2
3
4
5
6
7
8
9
View Article Online
DOI: 10.1039/D0NJ03044A
OBz
OBz
O
O
O
H
H2N
N
O
H
BzO
BzO
BzO
BzO
F3C
N
COOMe
N
CF3
NH
N
N
BzO
BzO
H
O
O
11.1
9.5
12
11-12
O
O
O
O
O
H
N
H
Ph
HN
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
N
HN
HN
HN
NH
NH
O
N
H
N
O
N
O
N
H
O
N
H
H
O
O
9.4
8.6
8.5
7.6
5.6
no reaction
NHTs
NO2
7.4
NHTf
NO2
6.0
decomposition
Figure 2. N-Nucleophiles with pKa values which failed in attempted transformations of 1
(pKa values are taken from a ScienceFinder search of the corresponding materials).
Experiments with 1 and trihaloacetanilide derivatives: a special result
A further N-nucleophile reacted with 1 was 2-nitro-trifluoroacetanilide 29a (pKa = 7.6) with
the intent to get intermediate A for spirocyclization toward a spiro-benzopyrazinone in a
similar way as described previously12 (Scheme 1). To this end the compound obtained and
belived to have structure A was subjected to catalytic hydrogenation. Surprisingly, this
treatment resulted in 1, that would not have been possible had the N-glycosyl-
trifluoroacetanilide A been formed in the Mitsunobu step. A more careful look at the spectral
characteristics of the obtained compound revealed it to be 30a. This was based on the 13C
signal of 141.8 ppm (2JC-F = 38.0 Hz) assignable to an imidate type carbon similar to those
found in the widely applied glycosylating agents O-glycosyl-N-phenyl trifluoroacetimidates
(e. g. B and C, imidate chemical shifts: 143.1 and 142.8 ppm, respectively).20 The formation
10

Page 11 of 37
New Journal of Chemistry
1
2
3
4
View Article Online
DOI: 10.1039/D0NJ03044A
of 30a could be rationalized by the reaction of the iminol tautomeric form of 29a providing a
5
6
7
8
presumably more acidic nucleophile which is also sterically more accessible.
Trifluoroacetanilide 29b and trichloroacetanilide 29c were also reacted with 1. Unfortunately,
the products of these reactions, 30b and 30c, respectively, could not be purified completely,
nevertheless, the 13C NMR spectra contained resonances at 140.5 (2JC-F = 36.7 Hz) and 141.3
ppm, respectively, to show the formation of imidates. In order to check the possibility to form
the anomeric O-glycosyl-N-phenyl trifluoroacetimidate, 1 was reacted with the corresponding
imidoyl chloride obtained from 29a by treating it with thionyl chloride at reflux temperature
for one hour. However, no conversion of 1 was observed overnight at rt or reflux in the
presence of K2CO3 in chloroform or acetone. This lack of reactivity may be ascribed to the
steric hindrance of the axial OH in a tertiary position in 1, however, the Mitsunobu protocol
still allows to get the β-anomer 29a.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
O-Alkylation of amide type substrates due to the amide-iminol (NH-C=O→N=C-OH)
tautomerism is rather well known in Mitsunobu chemistry.2 Nevertheless, the known
examples refer to tautomerizations where the iminol becomes part of an aromatic ring giving
a phenolic nucleophile, or the O-substitution is intramolecular to result in ring formation.
Intermolecular non-phenolic versions are scarce, we could locate only one analogous reaction
in the literature for the transformation of an allylic alcohol type substrate into an O-allyl-N-
phenyl trichloroacetimidate by the corresponding trichloroacetanilide in 10 % yield.21
To the best of our knowledge, this is a new way in carbohydrate chemistry to obtain an
imidate type leaving group that may find applications with OH-groups other than the
glycosidic one.
11

New Journal of Chemistry
Page 12 of 37
1
2
3
4
5
6
7
8
9
View Article Online
DOI: 10.1039/D0NJ03044A
OBz
O
K2CO3 in acetone or CHCl3
BzO
BzO
at rt or reflux
COOMe
BzO
Cl
CF3
OH
obtained
by SOCl2
treatment
N
1
+
29a
of
NO2
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
O
R
HO
R
155.0 ppm
30b
in
NH
N
X
amide
iminol
OBz
X
O
BzO
BzO
COOMe
NO2
29
a
b
c
BzO
X = NO2, R = CF3
X = H, R = CF3
X = H, R = CCl3
O
N
CF3
DEAD, Ph3P
THF, rt
141.8 ppm
143.1 ppm
OBz
OBz
O
X
O
BzO
BzO
BzO
BzO
N
N
O
O
BzO
BzO
R
CF3
B
MeOOC
30a
(52 %)
OBz
O
30b,c
142.8 ppm
N
BzO
BzO
were not
isolated
in pure
state
O
CF3
OBz
O
H2/Pd/C
(78 %)
BzO
BzO
BzO
O
N
C
BzO
1
CF3
A
MeOOC
O2N
Scheme 1. Experiments with trihaloacetanilide derivatives.
Experiments with 1 and S-nucleophiles
Mitsunobu conditions can also be used to replace OH-groups by S-nucleophiles, nevertheless,
(hetero)aromatic thiols and thioacids suit better than aliphatic thiols due to acidity differences.2
Although thioglycosides of type 32 are more easily prepared from the glycosyl bromide of 1
(Br instead of OH) with thiolates,12 the reactions of 1 with thiophenol (31), ethanethiol, and 2-
mercapto-benzothiazole were investigated (Scheme 2). While 31 gave the expected 32 in
modest yield, the two other thiols showed no conversion over 3 hours and a slow formation of
glycal 2 was observed overnight. The unreactive nature of the two latter thiols is most probably
12

Page 13 of 37
New Journal of Chemistry
1
2
3
4
View Article Online
DOI: 10.1039/D0NJ03044A
due to their insufficient acidity (both are above the optimal pKa range established with O- and
5
6
7
8
N-nucleophiles), though 2-mercapto-benzothiazole can be used in other Mitsunobu
transformations.2
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
OBz
O
BzO
BzO
+
PhSH (3 eq.)
pKa = 6.6
COOMe
BzO
OH
31
1
DEAD (3 eq.)
Ph3P (3 eq.)
THF, rt
OBz
O
BzO
BzO
S
BzO
MeOOC
32
(39 %)
S
SH
SH
10.6
N
9.8
no reaction
Scheme 2. Reaction of 1 with thiols (pKa values
are taken from a ScienceFinder search of the
corresponding materials).
Experiments with 1 and C-nucleophiles
A detailed study was carried out for the C-glycosylation of O-perbenzylated glycopyranoses
(mainly 2,3,4,6-tetra-O-benzyl-D-glucopyranose) under Mitsunobu conditions.5 The
combination of DEAD and n-Bu3P in benzene was found as the most advantageous of the
many reaction conditions investigated. Sulfonylmethane type C-nucleophiles gave the
expected C-glycosyl derivatives, however, with enolizable substituents, having e. g. ester or
13

New Journal of Chemistry
Page 14 of 37
1
2
3
4
View Article Online
DOI: 10.1039/D0NJ03044A
ketone moieties on the central methylene group, the O-glycosylation became a competitive or
5
6
exclusive reaction pathway.
7
8
9
We have attempted the reaction of 1 and 10 C-nucleophiles of a wide range of pKa (Fig. 3) six
of which were also investigated in the above study.5 DEAD and n-Bu3P were tried in THF,
acetonitrile, acetone, toluene and benzene, but the desired bis-C,C-glycosyl compound was
not formed in either experiment, instead, the glycal-type side product 2 was detected in all
cases. In the presence of Ph3P no conversion was observed either. The lack of reactivity might
be explained by the very weak nucleophilicity of these CH-acidic molecules and the steric
hindrance brought about on encountering of the tertiary reaction centre and the secondary C-
nucleophilic reagent.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
O
O
EtOOC
COOEt
NC
SO2Ph
12.0
S
NC
11.8
C-glycosylation
O-glycosylation of the enol
(62 %)
(35 %)
14.0
NC
O
CN
O2N
COOEt
9.2
unreactive
SO2Ph
11.1
NO2
unreactive
11
S
O
O
O
O
O
NO2
N
N
N
N
O
O
O
O
O
7.3
unreactive
7.7
6.3
4.7
unreactive
Figure 3. Attempted C-nucleophiles that proved unreactive towards 1 in this work and their
pKa values (pKa values are taken from a ScienceFinder search of the corresponding materials).
The framed compounds were also used in a former C-glycosylation study of 2,3,4,6-tetra-O-
benzyl-D-glucopyranose with the indicated results.5
14

Page 15 of 37
New Journal of Chemistry
1
2
3
4
View Article Online
DOI: 10.1039/D0NJ03044A
Experiments with the sialic ester 33
5
6
7
8
9
After determining the scope and limitations of the Mitsunobu glycosylation by 1, we turned to
the neuraminic acid derivative 33 already examined as a reagent in glycosylation of phenols
(for comparison, three of those experiments with phenols 34-36 to give 38-40, respectively,
are shown in Table 4, entries 1-3).10
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
First we tried to extend the reaction to o-nitrophenol 37, however, under the applied
conditions (entry 5) no glycosylation took place, only the formation of glycal 43 was
observed, and the same result was obtained in THF (entry 6). However, on cooling the
reaction mixture to -30 °C and paying close attention to the dropwise addition of DEAD over
10 minutes, the expected 41 was formed in 76 % yield with a 92:8 α/β ratio (entry 7). This
high stereoselectivity was a significant improvement to the previously achieved ones
(compare entries 1-3). Glycal 43 was detected in traces only in this experiment. To check the
effect of the lower temperature, p-nitrophenol 36 (entry 4) was reacted under these conditions
to give the expected 40 in very good yield but with much higher α-selectivity than the
literature result in entry 3. Therefore, it can be concluded that lowering the temperature is
beneficial for the α-selectivity of these reactions.
Table 4. Experiments with the neuraminic acid derivative 33.a
46
47
48
49
50
51
52
53
54
55
56
OAc
OAc
OAc
DEAD (3 eq.)
Ph3P (3 eq.)
OAc
OAc
OAc
OH
COOMe
COOMe
Nu
COOMe
+ NuH
(3 eq.)
+
AcO
AcHN
AcO
AcHN
O
AcO
AcHN
O
O
Conditions
AcO
33
AcO
AcO
43
13 34 37
38 42
,
-
-
Conditions
Solvent/
pKa of
Yield (%)
57 Entry
NuH
Product
NuHb Temp. (°C)/
R. time (h)
(α:β)
58
59
60
15

New Journal of Chemistry
Page 16 of 37
1
2
3
4
5
6
7
8
9
View Article Online
DOI: 10.1039/D0NJ03044A
OAc
AcO
OAc
COOMe
O
OH
7610
O
AcHN
1.
2.
9.8
CH3CN/0/3
(59:41)
AcO
38:43 1:0.3c
34
35
38
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
OAc
OAc
COOMe
O
OH
5010
AcO
AcHN
O
10.4
CH3CN/0/3
CH3CN/0/3
(50:50)
AcO
H3CO
39:43 1:0.8c
OCH3
39
OAc
OAc
COOMe
O
OH
9010
(55:45)
40:43 1:trace
AcO
AcHN
O
3.
4.
7.2
AcO
O2N
36
NO2
40
CH3CN/-30/8
DEAD added
dropwise
78
(92:8)
40:43 1:trace
40
over 10 min
OH
5.
6.
7.1
CH3CN/0/3
43
43
45
NO2
37
THF/rt/
overnight
detected, but
not isolated
OAc
OAc
COOMe
CH3CN/-30/8
DEAD added
dropwise
76
(92:8)
40:43 1:trace
NO2
AcO
AcHN
O
O
7.
8.
AcO
over 10 min
41
OAc
OAc
COOMe
N
N
CH3CN/-30/8
DEAD added
dropwise
N
AcO
AcHN
O
N
N
68
(~70:30)
N
8.2
AcO
H
over 10 min
13
42
a Order of addition: 1. Ph3P, 2. NuH, 3. substrate 33, 4. DEAD
b pKa values are taken from a ScienceFinder search of the corresponding materials.
57 c The corresponding N-glycosyl-N,N’-hydrazinedicarboxylate was also observed in a 0.1 ratio relative to 38 or 39.
58
59
60
16

Page 17 of 37
New Journal of Chemistry
1
2
3
4
5
6
7
8
9
View Article Online
Using the above improved conditions, the transformation of 33 was attempted by DoOthI: e10r.1O039-,/DN0N-J03044A
, S- and C-nucleophiles (Fig. 4) which proved to be reactive in reactions with 1, but
surprisingly, only benzotriazole 13 was successful to produce 42 (Table 4, entry 7). In all
other experiments, formation of glycal 43 was faster than the glycosylation.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
O
OH
N
N
N
N
N OH
O
O
5
6
N
NH
N
N
H
Cl
O
11
19
14
H
N
OBz
O
HN
N
N
N
BzO
BzO
Br
N
N
N
BzO
20
SH
NC
SO2Ph
Figure 4. Unreactive nucleophiles studied with the
neuraminic acid derivative 33.
Catalytic reduction of the nitro group in 41α led to a 68 % yield of the spiro-benzoxazinone
44 (Scheme 3) formed by an instantaneous attack of the amino group on the ester in a similar
ring closure described earlier with the corresponding o-nitrophenyl glycoside of 1.12
17

New Journal of Chemistry
Page 18 of 37
1
2
3
4
5
6
7
8
9
View Article Online
DOI: 10.1039/D0NJ03044A
OAc
OAc
COOMe
O
AcO
AcHN
O
AcO
41
NO2

10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
EtOAc, rt, 2h
H2, Pd/C
OAc
H
N
OAc
O
AcO
AcHN
O
O
AcO
44
(68%)
Scheme 3. Reductive ring closure of the O-(o-
nitrophenyl) sialoside 41α.
Structure elucidations
Based on the generally accepted mechanism of the Mitsunobu substitutions (see also
mechanistic considerations below) an inversion of the anomeric centre in 1 could be expected,
nevertheless, we sought for further corroborating proofs of the structures. It is conspicious
that all of the new compounds 9, 10, 12, 23-30, 32 derived from 1 show an optical rotation
([α]D) in the –17 - –62 range (Table 5) suggesting that they are of similar structure (1 and the
corresponding anhydro-aldonic acid ester A with equatorial COOMe groups are slightly
dextrorotatory: +14 and +52, respectively). The formerly obtained O-glycosides, such as F,
also have [α]D in the –5 - –69 range.12 Another comparison can be made with aldopyranosides
having the same aglycons as the present Mitsunobu products. Such comparisons are possible
with the pairs H-I, J-9, K-10, L-30, M-22, and N-32 where the compounds without the ester
group are available and sufficiently characterized in the literature. In all but one cases the
attachment of the COOMe moiety to the α-anomeric position results in a laevorotatory change
(the exception is the pair of benzotriazole derivatives M-22 where the direct attachment of the
extended aromatic system to the anomeric centre may be responsible for the anomaly).
18

Page 19 of 37
New Journal of Chemistry
1
2
3
4
View Article Online
DOI: 10.1039/D0NJ03044A
Further considerations can be made on the basis of changes in the sugar proton chemical shifts
5
6
caused by the introduction of the COOMe group. Previous experiences with similar
7
8
4
compounds of the C1/5C2 (D) conformation showed that substitution of the anomeric
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
hydrogen in the α-position by atoms/groups of high electronegativity, such as Br, OH or
COOMe, brought about upfield shifts especially for the hydrogens on the same side of the
ring, i. e. H-3/H-4 and H-5/H-6 (compare chemical shift differences Δδ in Table 5 for ring
protons of D-1, D-E, F-G pairs). Such differences are also remarkable for the pairs K-10, M-
22, N-32. Due to conformational change in 9 and 30 towards 2,5B (or a related skew boat)
3
indicated by the relatively small JH-3,H-4 coupling constants of 4-6 Hz, such comparisons are
not possible for the pairs J-9 and L-30. This change in the conformation has been observed in
some of the O-glycosidic compounds (besides 9 and 30 also in I and other similar structures
reported previously12) and may be ascribed to a release of steric strain due to the axial
COOMe and a gain of anomeric stabilization by an nO-6→σ*C-2-O-2 delocalization in the
indicated boat like conformation. This phenomenon has not been investigated in more depth.
To further confirm the anomeric configuration, we selected four of our products 10, 12, 23
and 25b and measured their 1D and/or 2D offset-compensated and zero-quantum filtered
ROESY-NMR spectra.22 In each case, weak, but observable peaks were detected (Figure 5
and Table 5) between H-4, H-6 and the protons of the methyl group of the ester at the
anomeric position, indicating that the methoxycarbonyl group was in the α-position.
3
In all four cases (10, 12, 23, 25b), the JH-3,CO heteronuclear coupling constant between the
methoxycarbonyl carbon atom and H-3 was also measured by using our HSQMBC method23
which turned out to be in the range of 4.0-6.8 Hz (Table 5). Though the scatter of the coupling
constant is significant (probably due to the very different nature of Nu units attached to C-2),
all these couplings accord with dihedral angles excluding the 30-130 degree range with a cis
25
5
3
configuration.24,
Since the ring conformations must be C2 according to big JH-3,H-4
19

New Journal of Chemistry
Page 20 of 37
1
2
3
4
3
View Article Online
DOI: 10.1039/D0NJ03044A
homonuclear couplings and, therefore, H-3 must be in an axial position, hence the JH-3,CO
5
6
7
8
heteronuclear couplings are relevant to determine the anomeric configuration. Taking into
consideration that similar ulosonates, e. g. 1 and E (Table 5) have a 3JH-3,CO coupling constant
of 2.1 and 2.4 Hz,15 respectively, and in other cases with an axial carbonyl group26, 27 the
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
3
value is JH-3,CO ≥ 4 Hz, it can be concluded that in these compounds the Nu unit is in the β-
position.
A)
OBz
OBz
O
6
O
O
BzO
BzO
BzO
BzO
4
Nu
O
H
H
BzO
O
BzO
ROE
H3C
H
H
Nu
H3C
O
No ROE
B)
COOMe
H-4
H-5
H-6
H-3
H-7,7'
Figure 5. A) Possibility for ROE effects between H-4, H-6 and the methyl ester in the
Mitsunobu products. B) Exemplary ROESY spectrum of 12 (blue, with irradiation at the OMe
frequency; effects below 0.05 % are disregarded in Table 5) in comparison with the basic 1H
NMR spectrum (red).
For the assignment of the anomeric configuration of the compounds derived from sialic ester
33
the
method
described
in
the
literature
was
used.10
20

Page 21 of 37
New Journal of Chemistry
1
2
3
4
5
Table 5. Selected physical and spectroscopic data of the new compounds and related products
6
δ [ppm], (3J [Hz])
7
8
9
Compound
[α]D
Aldopyranosides:
Ulopyranosides:
H-2
H-3
5.71
H-3
H-4
5.96
H-4
H-5
5.69
H-5
H-6
4.18
COOMe
3.68
OBz
O
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
+14
BzO
BzO
COOMe
BzO
D15
BzO
BzO
BzO
H
+52
5.90
0.19
6.20
0.24
5.83
0.14
4.72
0.54
3.86
3.77
O
COOMe
Δδ (1-D)
Δδ (E-D)
B3zO
OH
3JH-3,CO = 2.1 Hz
115
BzO
BzO
BzO
H
+129
5.84
0.13
6.17
0.21
5.85
0.16
4.77
0.49
O
COOMe
B3zO
Br
3JH-3,CO = 2.4 Hz
E15
+0.929
+4230
5.48
5.91
5.70
4.24
OBz
O
BzO
BzO
N3
-0.5431
BzO
F (for NMR data28)
BzO
BzO
BzO
H
+41
5.70
0.22
6.36
0.45
5.81
0.11
4.51
0.27
3.90
O
N3
Δδ (G-F)
B3zO
COOMe
3JH-3,CO = 4.1 Hz
G15
21

New Journal of Chemistry
Page 22 of 37
1
2
3
4
5
6
7
8
9
OBz
O
+66
-12
-21
No data available.
NO2
NO2
BzO
BzO
O
O
BzO
H32
OBz
O
6.00
(6.2)
6.05
6.11
5.06
3.71
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
BzO
BzO
(6.2, 8.2)
(9.3, 8.5)
(9.4, 6.2, 3.1)
BzO
MeOOC
I12
OBz
5.97
5.86
5.81
4.23
O
O
BzO
BzO
O
N
BzO
O
J33
O
-23
6.13
(4.2)
5.97
6.64
4.89
3.89
(4.2, 7.8)
(10.7, 7.8)
(10.7, 4.3, 4.3)
BzO
BzO
O N
O
OBz
O
COOMe
BzO
9
O
N
O
O
BzO
BzO
BzO
COOMe
O
OBz
probable ring conformation of 9
38
39
40
41
42
43
44
45
46
22

Page 23 of 37
New Journal of Chemistry
1
2
3
4
5
6
7
8
9
OBz
O
+8
5.97
6.05
5.8
4.15
BzO
O
BzO
N
BzO
N
N
K34
-34
6.31
(8.4)
0.34
6.14
(8.3, 8.3)
0.09
5.89
8.5, 9.9
0.09
4.96
(9.9, 3.2, 3.2)
0.81
3.99
BzO
BzO
BzO
H
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
O
O
N
B3zO
MeOOC
3JH-3,CO = 4.0 Hz
Δδ (10-K)
N
N
1D ROE %
0.13
0.21
not observeda
0.26
10
OBz
O
+60
-48
~5.9-5.8
~5.9-5.8
~5.9-5.8
~4.5-4.2
BzO
BzO
N
O
BzO
CF3
L20
CF3
6.02
(5.0)
5.90
(6.7, 5.0)
6.07
(9.4, 6.8)
5.18
(9.4, 3.8, 3.8)
3.78
BzO
BzO
O
O
N
OBz
BzO
MeOOC
O2N
30a
CF3
N
5.97
(4.6)
5.87
(4.6, 7.5)
6.04
(7.4, 8.6)
5.14
(8.2, 4.5, 3.9)
BzO
BzO
O
O
OBz
BzO
MeOOC
30b
(1H NMR data for the partially
purified compound)
23

New Journal of Chemistry
Page 24 of 37
1
2
3
4
X
5
6
N
7
8
9
10
11
12
13
CF3
O
BzO
BzO
BzO
COOMe
O
OBz
probable ring conformation of
30a (X = NO ) and 30b (X = H)
14
2
O
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
-17
6.88
(10.2)
6.28
(10.0, 9.7)
5.93
(9.7, 10.0)
not observeda
4.74
(10.1, 4.6, 2.8)
3.98
BzO
BzO
BzO
H
O
N
B3zO
MeOOC
1D ROE %
0.05
0.24
0.4
O
3JH-3,CO = 4.1 Hz
12
-74
-19
5.99
6.26
6.19
4.77
OBz
O
BzO
BzO
N
N
N
N
BzO
M35
6.74
(8.4)
0.75
6.37
(8.3, 7.7)
0.11
5.91
(9.8, 7.7)
-0.28
4.98
(9.7, 4.5, 2.8)
0.21
3.87
OBz
O
BzO
BzO
Δδ (21-M)
N
N
BzO
MeOOC
21
24

Page 25 of 37
New Journal of Chemistry
1
2
3
4
5
6
7
8
9
10
11
N
6.40
(8.8)
-30
-22
6.52
(8.4, 8.4)
5.92
(8.8, 9.2)
4.82
(multiplet)
3.86
3.92
N
N
Cl
OBz
O
BzO
N
BzO
BzO
MeOOC
22
O
6.54
(10.1)
6.36
(9.8, 9.8)
5.89
(9.5, 9.5)
4.51
(multiplet)
BzO
H
O
B3zO
BzO
N
12
13
14
O
BzO
1D ROE %
0.08
0.25
not observeda
0.47
MeOOC
O
3JH-3,CO = 6.8 Hz
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
23
NH2
-25
6.79
(8.1)
6.04
(8.1, 7.2)
5.85
(9.7, 7.1)
5.09
(9.6, 4.9, 3.0)
3.91
OBz
O
N
BzO
BzO
N
N
N
BzO
MeOOC
24
OBz
O
-27
-29
6.30
(8.8)
6.44
(8.6, 8.6)
5.87
(8.4, 9.4)
4.95
(9.8, 4.0, 2.5)
3.91
3.93
N
N
BzO
BzO
N
N
BzO
MeOOC
25a
6.86
(8.6)
not observeda
6.19
(8.1, 8.1)
5.91
(8.2, 9.3)
not observeda
5.19
(9.3, 4.5, 2.9)
BzO
BzO
BzO
H
O
B3zO
N
N
N
N
2D ROESY %
0.35
0.84
MeOOC
3JH-3,CO = 5.0 Hz
25b
OBz
-30
7.03
(8.1)
6.26
(8.0, 7.2)
6.01
(9.6, 7.2)
5.31
(9.8, 5.2, 3.4)
3.97
N
N
O
BzO
N
BzO
N
BzO
MeOOC
25

New Journal of Chemistry
Page 26 of 37
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
26
OBz
O
-24
-62
6.90
(8.0)
6.15
(8.1, 7.5)
5.91
(9.6, 7.2)
5.22
(9.7, 5.0, 3.0)
3.91
3.61
N
N
N
BzO
BzO
N
BzO
MeOOC
Br
27
OBz
O
6.77
(8.0)
6.00
(multiplet)
5.77
(multiplet)
5.18
(multiplet)
N
N
N
BzO
BzO
OBz
N
BzO
MeOOC
O
BzO
OBz
OBz
28
OBz
+94
+14
5.50
5.92
5.61
4.20
O
BzO
BzO
SPh
SPh
BzO
N36
OBz
O
5.70
(9.6)
6.03
(9.6, 9.5)
5.55
(9.4, 10.0)
4.81
(10.0, 4.4, 2.5)
3.82
BzO
BzO
BzO
MeOOC
Δδ (31-N)
0.2
0.11
-0.06
0.61
31
a Not visible in the spectrum or the intensity is <0.05 %.
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
26

Page 27 of 37
New Journal of Chemistry
1
2
3
4
View Article Online
DOI: 10.1039/D0NJ03044A
Mechanistic considerations
5
6
7
8
9
Usually, the Mitsunobu reaction takes place with an inversion of the stereogenic reaction
centre via an SN2-like mechanism.2 However, when a glycosidic OH-group is involved, a
more complex picture can be drawn taking into account the possible anomerization, as well as
SN1-like pathways with the formation of glycosylium ion type intermediates and anchimeric
assistance of the C-2 substituent as explained in details in a recent review.4.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The ulosonate type substrates in the present study have a further COOMe substituent at the
anomeric centre which may exert additional effects in terms of cation destabilization,
preference of conjugation, and steric hindrance, all of which can influence the outcome of a
particular reaction. As widely accepted, the azodicarboxylate I, the phosphine (R’3P) and the
conjugate acid of the nucleophile (HNu) may form the phosphonium intermediate III via
betaine II. Reaction of III with the ulosonate V may give the glycosyloxyphosphonium
intermediate IV along with hydrazinedicarboxylate VI. Intermediate IV may be stabilized by
the neighbouring ester group. Cation IV can be transformed in various ways that may be
influenced by the effects of the COOMe as well as the 3-benzoyloxy substituent (if present).
The most straightforward way can be the backside attack of Nu– to give the anomerically
inverted substitution product VIII. A parallel pathway can be the attack of the H-3 proton by
Nu– (or even by R’3P especially if it is the more basic n-Bu3P) to give the glycal type by-
product IX in a stereoelectronically favourable bimolecular elimination. Glycal IX was
observed in each reaction since its formation is promoted by the conjugating COOMe
substituent as well as by the sterically crowded reaction centre. Also, the sterically bulkier
nucleophiles may preponderantly attack at H-3, and thus, only glycal formation was observed
with C-nucleophiles. Intermediate IV may also undergo a R’3PO loss to give glycosylium ion
VII, nevertheless, this step may not be favoured due to the cation destabilizing effect of the
COOMe group. In this respect, there may be a difference between the glycosylium ions VIIa
27

New Journal of Chemistry
Page 28 of 37
1
2
3
4
View Article Online
DOI: 10.1039/D0NJ03044A
and VIIb derived from 1 and 33, respectively. While VIIb „suffers” from the electron
5
6
7
8
withdrawing effect of the ester substituent only, in VIIa the additional negative inductive
effect of the 3-OBz substituent may contribute to the destabilization. Though the participating
3-OBz can stabilize the neighbouring cation once formed, this effect can’t assist the cleavage
of R’3PO since the two groups are in a 1,2-cis relative configuration. Therefore, this
monomolecular pathway is probably highly unfavourable with 1. On the other hand, the
anchimeric effect of the 3-OBz group would also facilitate the formation of VIII from VIIa.
These considerations are supported by the experience that no anomerization was observed in
the reactions of 1. In the absence of a C-3 participating group in VIIb the nucleophile can
attack from both sides as reflected by the formation of VIII and X with modest or no
selectivity in room temperature reactions of 33.10 Nevertheless, lowering the temperature
prefers the α-side attack with less steric hindrance for the approaching nucleophile to give
products of type VIII almost exclusively. Cation VII can also loose the H-3 proton to yield
glycal IX.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
28

Page 29 of 37
New Journal of Chemistry
1
2
3
4
5
6
7
8
9
View Article Online
DOOI: 10.1039/D0NJ03044A
H
O
O
N
OR
N
OR
N
OR
H Nu
E2
Nu
+
+
RO
N
RO
N
RO
N
O
O
O
R'3P
PR'3
I
II
III
R'3P
E2
SN2
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
H
OBz
O
OMe
O
O
O
BzO
BzO
COOMe
OR
C
O
COOMe
OH

COOMe
PR'3
3
BzO
1
O
HO
IV
O
V
PR'3
O
H
N
O
RO
N
H
3
COOMe
COOMe
O PR'3
BzO
O
SN1/E1
SN2
VI
VIIa
E2
OAc
OAc
H
O
OH
O
Nu
Nu
COOMe
AcO
AcHN
O
SN1
COOMe
COOMe
AcO
33
VII
E1
VIII
Nu
-H
SN1
COOMe
H
O
O
O
3
COOMe
Nu
H
VIIb
X
IX
Scheme 4. Possible mechanistic pathways of the Mitsunobu reaction with ulosonates.
Conclusion
A wide range of O-, N-, S-, and C-nucleophiles have been investigated as reaction partners in
Mitsunobu glycosylations with ulosonic acid methyl esters. Substrates with a tertiary OH
group are very scarcely applied under these conditions due to the steric hindrance of the
alcoholic functional group. The series of experiments allowed to establish the scope and
limitations of this transformation with methyl 3,4,5,7-tetra-O-benzoyl-α-D-gluco-hept-2-
ulopyranosonate (1). Thus, among O-nucleophiles phenols and N-hydroxy compounds with a
pKa value in the range of 5-8 proved suitable. Less acidic oximes, alcohols and hydroxamic
29