Mono- and oligosaccharides are prevalent motifs in biologically active natural products
such as glycopeptides, glycoproteins, proteoglycans, and glycolipids.[1] Deoxy sugars also constitute key intermediates with widespread applications in medicinal
and pharmaceutical research.[2] Therefore, the development of efficient methods for the synthesis of these glycomolecules
in high stereoselectivity and with broad structural diversity is undoubtedly appealing
in chemistry, biology, and related fields.[2f]
[3] Remarkable progress has been made, especially in transition-metal-catalyzed and
organocatalyzed conventional glycosylations for the preparation of complex carbohydrates.[4] In this study, we focused on the iodoglycosylation and iodocarboxylation of glycals
by using anionic chiral Co(III) complexes as phase-transfer catalysts, affording the
corresponding 2-deoxy-2-iodoglycosides and 2-deoxy-2-iodoglycosyl carboxylates, which
are of high synthetic and biological importance.[5] Although iodoglycosylations and iodoacetoxylations through stoichiometric glycosylations
using glycals and alcohols or acetic acid in the presence of an electrophilic iodine
reagent (NIS) or an iodate reagent such as NH4I, NaI, I2, or TMSI(OAc)2 with an oxidant [for example, H2O2, CAN, Cu(OAc)2, PhI(OAc)2,[6] PPh3,
[6h] or TfOH[5d]] have been reported, the exploration of high-performance catalytic systems that
are, in principle, distinct from previous ones remains essential for the development
of mild and efficient stereoselective glycosylation methods. Furthermore, there are
few reports on the catalytic stereocontrol of such reactions with chiral catalysts,
due to the difficulty of asymmetric intermolecular halogenation[7]
[8] or glycosylation.[4]
[9]
The potential of the octahedral chiral-at-metal complexes, in which the metal center
does not serve as a catalytic center to activate substrate by coordination but merely
provides a rigid framework and an environment of centrochirality, is less well recognized.[10] We have recently developed Brønsted acids and sodium salts of anionic chiral Co(III)
complexes as efficient catalysts for the highly enantioselective bromoaminocyclization
of olefins and for the Povarov reaction of enol ethers with 2-azadienes.[11] More importantly, the chiral Co(III)-complex-templated Brønsted acids have been
proved to function as bifunctional phase-transfer catalysts to shuttle N-bromosuccinimide (NBS) to the reaction solution and to control stereoselectivity.[11c]
[d] We therefore speculated that such anionic chiral Co(III) complexes might also serve
as alternative chiral-anion-mediated catalysts[12] with an iodo-cation source for the iodoglycosylation[4]
[6a]
[13] of glycals 2
[14] with alcohols 3 or carboxylic acids 5 (Scheme [1]).
Scheme 1 Chiral Co(III)-complex-templated phase-transfer catalysis
Here, we present our preliminary studies on catalytic iodoglycosylation and iodocarboxylation
reactions for the preparation of 2-deoxy-2-iodoglycosides 4 and 2-deoxy-2-iodoglycosyl carboxylates 6 by using anionic chiral Co(III) complexes 1.
The diastereoselective iodoglycosylation of 3,4,6-tri-O-benzyl-d-glucal (2a) and benzyl alcohol (3a) with NIS at room temperature was initially tested, and the corresponding 2-deoxy-2-iodoglycoside
4aa was obtained in 69% yield with a 2:1 dr (α/β ratio) without a catalyst (Table [1], entry 1). As expected, the reaction occurred smoothly in the presence of various
Lewis acids (entries 2–9), delivering the product in good yields (up to 86%), albeit
with only up to 2.5:1 dr, even at a low temperature of –40 °C (entries 6 and 8).[5d] Several Brønsted acids, such as p-toluenesulfonic acid and the chiral phosphoric acids PA1 and PA2, were also tested, and the results suggested that the proton has no impact on the
diastereoselectivity (entries 10–12). When DMAP and PPh3 were employed as the catalysts, the α/β ratio of 4aa was only 3:1 (entries 13 and 14). A series of anionic chiral Co(III) complexes, either
as sodium salts or Brønsted acids, were then screened (entries 15–21), Λ-(S,S)-1c afforded the best diastereomeric ratio of 4:1 in the highest yield of 86% (entry
18). The metal-centered chirality in the chiral Co(III)-complex-templated Brønsted
acids had little effect on the stereochemical outcome of the iodoglycosylation (entries
16, 19, and 20). Several iodinating reagents, such as 1,3-diiodo-5,5-dimethylhydantoin
(DIH) and N-iodosaccharine (NISC), were then tested, and NIS was found to be the optimal iodine
source for this protocol (entries 18, 22, and 23). To our delight, changing the ratio
of 2a, 3a, and NIS ratio slightly improved the diastereoselectivity (entry 24). Moreover, a
screening of the reaction parameters, including the solvent, temperature, and additives,
suggested that the reaction in CH2Cl2 at room temperature gave a higher α/β ratio of 5:1 with 4 Å MS (entries 24–34).
Table 1 Optimization of the Conditions for Iodoglycosylationa
|
Entry
|
Catalyst
|
Solvent
|
I+ source
|
Yieldb (%)
|
drc (α/β)
|
1
|
–
|
CH2Cl2
|
NIS
|
69
|
2:1
|
2d
|
ZnBr2
|
CH2Cl2
|
NIS
|
76
|
1.5:1
|
3d
|
Sc(OTf)3
|
CH2Cl2
|
NIS
|
86
|
1:1
|
4d
|
Mg(OTf)2
|
CH2Cl2
|
NIS
|
85
|
1.6:1
|
5d
|
TMSOTf
|
CH2Cl2
|
NIS
|
78
|
1.5:1
|
6d,e
|
TMSOTf
|
CH2Cl2
|
NIS
|
55
|
2:1
|
7d
|
TfOH
|
CH2Cl2
|
NIS
|
59
|
1.8:1
|
8d,e
|
TfOH
|
CH2Cl2
|
NIS
|
45
|
2.5:1
|
9d
|
AgOTf
|
CH2Cl2
|
NIS
|
50
|
2.5:1
|
10d
|
TsOH·H2O
|
CH2Cl2
|
NIS
|
69
|
1.5:1
|
11
|
PA1
|
CH2Cl2
|
NIS
|
61
|
2:1
|
12
|
PA2
|
CH2Cl2
|
NIS
|
51
|
2:1
|
13d
|
DMAP
|
CH2Cl2
|
NIS
|
54
|
3:1
|
14d,e
|
PPh3
|
CH2Cl2
|
NIS
|
85
|
3:1
|
15
|
Λ-(S,S)-1a
|
CH2Cl2
|
NIS
|
82
|
3:1
|
16
|
Δ-(R,R)-1a
|
CH2Cl2
|
NIS
|
85
|
3.5:1
|
17
|
Λ-(S,S)-1b
|
CH2Cl2
|
NIS
|
56
|
3.8:1
|
18
|
Λ-(S,S)-1c
|
CH2Cl2
|
NIS
|
86
|
4:1
|
19
|
Δ-(R,R)-1c
|
CH2Cl2
|
NIS
|
78
|
3.5:1
|
20
|
Δ-(S,S)-1c
|
CH2Cl2
|
NIS
|
50
|
2:1
|
21
|
Λ-(S,S)-1d
|
CH2Cl2
|
NIS
|
79
|
3:1
|
22
|
Λ-(S,S)-1c
|
CH2Cl2
|
DIH
|
40
|
3:1
|
23
|
Λ-(S,S)-1c
|
CH2Cl2
|
NISC
|
12
|
3:1
|
24f
|
Λ-(S,S)-1c
|
CH2Cl2
|
NIS
|
82
|
5:1
|
25f
|
Λ-(S,S)-1c
|
CCl4
|
NIS
|
74
|
3.5:1
|
26f
|
Λ-(S,S)-1c
|
CHCl3
|
NIS
|
85
|
3.8:1
|
27f
|
Λ-(S,S)-1c
|
DCE
|
NIS
|
79
|
4.5:1
|
28f
|
Λ-(S,S)-1c
|
toluene
|
NIS
|
76
|
3:1
|
29f,g
|
Λ-(S,S)-1c
|
CH2Cl2
|
NIS
|
51
|
4:1
|
30f,h
|
Λ-(S,S)-1c
|
CH2Cl2
|
NIS
|
74
|
4.5:1
|
31f,i
|
Λ-(S,S)-1c
|
CH2Cl2
|
NIS
|
60
|
3:1
|
32f,j
|
Λ-(S,S)-1c
|
CH2Cl2
|
NIS
|
80
|
4:1
|
33f,k
|
Λ-(S,S)-1c
|
CH2Cl2
|
NIS
|
79
|
4:1
|
34f,l
|
Λ-(S,S)-1c
|
CH2Cl2
|
NIS
|
64
|
4:1
|
a Unless otherwise noted, the reaction was performed with glycal 2a (0.1 mmol), 3a (0.15 mmol), NIS (0.12 mmol), catalyst (0.01 mmol), 4 Å MS (100 mg) in the solvent
(1 mL) at r.t. under N2 in the absence of light for 24 h.
b Yield of isolated product 4aa.
c Determined by 1H NMR.
d Catalyst (20 mmol%) was used.
e The reaction was carried out at –40 °C.
f The 2a/3a/NIS ratio was 1:1.2:1.1.
g The reaction was carried out at 35 °C.
h The reaction was carried out at 0 °C.
i The reaction was carried out at –20 °C.
j 3 Å MS (100 mg) was used instead of 4 Å MS.
k 5 Å MS (100 mg) was used instead of 4 Å MS.
l No additive was used.
With the optimized reaction conditions in hand,[15] we next explored the scope of the iodoglycosylation with respect to the alcohol
3. As shown in Table [2], various substituents on the aryl moiety of the benzyl alcohols 3b–f were tolerated, affording the corresponding 2-deoxy-2-iodoglycosides 4 in up to 73% yield with up to 4.5:1 dr (Table [2], entries 1–5). The electronic nature of the substrates had no evident influence
on the reactivity. In addition, 1-naphthylmethanol was also well tolerated and provided
the product 4ag with 4:1 dr (entry 6). Nonbenzylic aliphatic alcohols 3h–n were also suitable substrates, giving the desired glycosides 4ah–an in high yields and up to 6.5:1 dr (entries 7–13). Sterically hindered alcohols 3j and 3k readily participated in the reaction with up to 6.5:1 dr (entries 9 and 10). The
Z-alkene in 3n was also well tolerated (4:1 dr; entry 13). More importantly, the iodoglycosylation
proceeded under mild conditions with monosaccharides such as the primary alcohol 3o and the secondary alcohol 3p, giving the corresponding disaccharides with up to 9:1 dr (entries 14 and 15).
The scope of the iodoglycosylation using Brønsted acid of anionic chiral Co(III)-complexes
1 with regard to the glycals 2 was also evaluated (Table [3]). Glycals with other hydroxy-protecting groups, such as 2b and 2c, were good substrates for the iodoglycosylation reaction, giving the corresponding
products in moderate to high yields and with up to 8:1 dr (entries 1–8). Moreover,
tri-O-benzyl-d-galactal (2d), the C4-epimer of 2a, afforded the corresponding products in moderate yields with similar diastereoselectivities
(up to 6.5:1 dr; entries 9 and 10).[16]
Table 2 Scope of the Alcohols 3 for the Iodoglycosylationa
|
Entry
|
Alcohol
|
Product
|
Yieldb (%)
|
drc (α/β)
|
drd (Lit.)
|
1
|
(3b)
|
4ab
|
72
|
4:1
|
–
|
2
|
(3c)
|
4ac
|
62
|
4.5:1
|
–
|
3
|
(3d)
|
4ad
|
73
|
4:1
|
–
|
4
|
(3e)
|
4ae
|
56
|
4:1
|
–
|
5
|
(3f)
|
4af
|
66
|
4.5:1
|
–
|
6
|
CyOH (3g)
|
4ag
|
75
|
4:1
|
–
|
7
|
Ph(CH2)2OH (3h)
|
4ah
|
84
|
4.5:1
|
–
|
8
|
(3i)
|
4ai
|
78
|
4.5:1
|
4.8:1[6h]
|
9
|
i-PrOH (3j)
|
4aj
|
73
|
5:1
|
2:1[6h]
|
10
|
t-BuOH (3k)
|
4ak
|
72
|
6.5:1
|
–
|
11
|
Me(CH2)7OH (3l)
|
4al
|
78
|
4:1
|
2.3:1[6h]
|
12
|
(3m)
|
4am
|
78
|
4:1
|
–
|
13
|
(3n)
|
4an
|
76
|
4:1
|
–
|
14e
|
(3o)
|
4ao
|
88
|
6.5:1
|
2.5:1[6h]
|
15e
|
(3p)
|
4ap
|
53
|
9:1
|
–
|
a Unless otherwise noted, the reaction was performed with glycal 2a (0.1 mmol), alcohol 3 (0.12 mmol), NIS (0.11 mmol), Λ-(S,S)-1c (0.01 mmol), and 4 Å MS (100 mg) in CH2Cl2 (1 mL) at r.t., under N2 in the absence of light for 24 h.
b Yield of isolated product.
c Determined by 1H NMR.
d Anomer ratios reported in the literature.
e The reaction was carried out on a 0.1 mmol scale for 72 h with a 2a/3/NIS ratio of 2:1:2.4.
Table 3 The Scope of Glycals for the Iodoglycosylationa
|
Entry
|
Glycal
|
Alcohol
|
Product
|
Yieldb (%)
|
drc (α/β)
|
drd (Lit.)
|
1
|
|
3a
|
4ba
|
80
|
8:1
|
5:1[6d]
9:1[6e]
|
2
|
3c
|
4bc
|
54
|
5:1
|
–
|
3
|
3f
|
4bf
|
55
|
4:1
|
6:1[6d]
11:1[6e]
|
4
|
3g
|
4bg
|
70
|
5:1
|
–
|
5
|
3h
|
4bh
|
63
|
4.5:1
|
–
|
6
|
|
3a
|
4ca
|
60
|
4:1
|
–
|
7
|
3c
|
4cc
|
66
|
5:1
|
–
|
8
|
3j
|
4cj
|
61
|
5:1
|
–
|
9
|
|
3c
|
4dc
|
47
|
4:1
|
–
|
10
|
3k
|
4dk
|
52
|
6.5:1
|
–
|
a Unless otherwise noted, the reaction was performed with glycal 2 (0.1 mmol), alcohol 3 (0.2 mmol), NIS (0.15 mmol), Λ-(S,S)-1c (0.01 mmol), and 4 Å MS (100 mg) in CH2Cl2 (1 mL) at r.t. under N2 in the absence of light for 48 h.
b Yield of isolated product.
c Determined by 1H NMR.
d Anomer ratios reported in the literature.
Encouraged by these results, we next studied the iodocarboxylation of glycal 2a with carboxylic acids and sodium salts 5 to give the corresponding 2-deoxy-2-iodoglycosyl carboxylates 6 (Table [4]).[17] The reaction of the sodium salts 5a and 5c and the acids 5b and 5d in the absence of the chiral-Co(III)-complex-templated Brønsted acid were first tested,
and products 6a and 6b were obtained with low diastereoselectivities (Table [4], entries 1-4). Interestingly, the presence of the Brønsted acid of a chiral Co(III)
complex Λ-1c resulted in enhanced diastereoselectivity (up to 6.5:1 dr; entries 5–8), but the
sodium salts of the acids still gave low yields (entries 5 and 6). It is suggested
that the presence of an acidic proton might promote the reaction and that the bulky
chiral Co(III) complexes influence the stereocontrol. The protocol also tolerated
other carboxylic acids, and the corresponding glycosyl carboxylates 6c–e were obtained with good diastereoselectivities (up to 7.5:1 dr; entries 9–11).
Table 4 The scope of the Iodocarboxylationa
|
Entry
|
Acid or Na Salt
|
Product
|
Catalyst
|
Yieldb (%)
|
drc
|
1
|
BzONa (5a)
|
6a
|
–
|
14
|
3.5:1
|
2
|
BzOH (5b)
|
6a
|
–
|
82
|
2:1
|
3
|
AcONa (5c)
|
6b
|
–
|
29
|
2.5:1
|
4
|
AcOH (5d)
|
6b
|
–
|
70
|
1.8:1
|
5
|
BzONa (5a)
|
6a
|
Λ-(S,S)-1c
|
28
|
3.8:1
|
6
|
AcONa (5c)
|
6b
|
Λ-(S,S)-1c
|
28
|
5.5:1
|
7
|
BzOH (5b)
|
6a
|
Λ-(S,S)-1c
|
74
|
6.5:1
|
8
|
AcOH (5d)
|
6b
|
Λ-(S,S)-1c
|
65
|
5.5:1
d
|
9
|
4-F3CC6H4CO2H (5e)
|
6c
|
Λ-(S,S)-1c
|
67
|
6:1
|
10
|
(5f)
|
6d
|
Λ-(S,S)-1c
|
62
|
6:1
|
11
|
2,6-Cl2C6H4CO2H (5g)
|
6e
|
Λ-(S,S)-1c
|
75
|
7.5:1
|
a Unless otherwise noted, the reaction was performed with glycal 2a (0.1 mmol), 5 (0.15 mmol), NIS (0.12 mmol), Λ-(S,S)-1c (0.01 mmol), and 4 Å MS (100 mg) in CH2Cl2 (1 mL) at r.t. under N2 in the absence of light for 12 h.
b Yield of isolated product.
c Determined by 1H NMR.
d The reported anomer ratio of 5d was 4:1.[6f]
Next, we performed a 1H NMR spectral analysis of mixtures of NIS with Brønsted acid Λ-(S,S)-1c or benzoic acid (5b) in CDCl3 at room temperature. The presence of succinimide in the former mixture clearly revealed
that Λ-(S,S)-1c reacted with NIS to produce a new complex,[11c]
[d] thereby leading to the formation of succinimide, while no evidence was found to
show that there is an interaction between benzoic acid (5b) and NIS in the latter mixture (Figure [1]).
Figure 1 1H NMR analysis of mixtures of NIS with Brønsted acid Λ-(S,S)-1c or benzoic acid (5b) in CDCl3.
On the basis of the interesting experimental results and our previous work,[11c]
[d] we propose a plausible mechanism for these reaction of a chiral-Co(III)-templated
Brønsted acid in combination with NIS. As shown in Scheme [2], the Brønsted acid 1 might undergo an exchange reaction with NIS to generate a reactive chiral iodinating
reagent 7, as suggested by the 1H NMR spectral analysis. The chiral ion-pair 7 might then undergo diastereoselective reaction with the glycal 2 to afford the 2-deoxy-2-iodoglycoside 4 or a 2-deoxy-2-iodoglycosyl carboxylate 6 through nucleophilic addition of the alcohol 3 or carboxylic acid 5, respectively, with regeneration of Brønsted acid 1. The stereochemical outcomes were not good enough to have been caused by the competition
of the rapid noncatalytic reaction of NIS.
Scheme 2 Proposed mechanism
In summary, we have developed a diastereoselective iodoglycosylation or iodocarboxylation
of glycals with NIS mediated by a chiral-Co(III)-complex-templated Brønsted acid.
The catalytic reaction proceeds under mild conditions, providing convenient access
to 2-deoxy-2-iodoglycosides or 2-deoxy-2-iodoglycosyl carboxylates in up to 88% yield
and with up to 9:1 dr. The anionic chiral Co(III) complexes also function as bifunctional
phase-transfer catalysts to shuttle the N-iodosuccinimide to the reaction solution. Further studies will focus on the development
of stereoselective halogenations catalyzed by Brønsted acids of anionic chiral Co(III)
complexes.