Li Zhang1, Ke Shen1, Hashem A Taha1, Todd L Lowary1. 1. Alberta Glycomics Centre and Department of Chemistry , University of Alberta , 11227 Saskatchewan Drive , Edmonton , Alberta T6G 2G2 , Canada.
Abstract
A number of biologically relevant glycoconjugates possess 1,2- cis-furanosidic linkages, a class of glycosidic bond that remains challenging to introduce with high stereoselectivity. In this paper, we report an approach to one family of such linkages, α-xylofuranosides, via the use of thioglycoside donors possessing a conformationally restricting xylylene protecting group. The method was shown to provide the desired targets in good to excellent yield and stereoselectivity. Computational investigations support the proposal that the protecting group locks the electrophilic intermediate in these reactions into a conformation that leads to the high selectivity. The power of the methodology was demonstrated through the synthesis of a complex hexasaccharide motif from lipoarabinomannan, an immunomodulatory polysaccharide from mycobacteria.
A number of biologically relevant glycoconjugates possess 1,2- cis-furanosidic linkages, a class of glycosidic bond that remains challenging to introduce with high stereoselectivity. In this paper, we report an approach to one family of such linkages, α-xylofuranosides, via the use of thioglycoside donors possessing a conformationally restricting xylylene protecting group. The method was shown to provide the desired targets in good to excellent yield and stereoselectivity. Computational investigations support the proposal that the protecting group locks the electrophilic intermediate in these reactions into a conformation that leads to the high selectivity. The power of the methodology was demonstrated through the synthesis of a complex hexasaccharide motif from lipoarabinomannan, an immunomodulatory polysaccharide from mycobacteria.
Glycoconjugates containing
monosaccharide residues in the furanose
ring form are present in a number of pathogenic organisms, including Aspergillus fumigatus and Mycobacterium tuberculosis, the causative agents of invasive aspergillosis and tuberculosis,
respectively.[1−4] Generally present in the protective cell walls of the organisms
that produce them, these molecules play important roles in disease
progression and the interaction of the pathogen with the host immune
system. Access to structurally defined glycoconjugates incorporating
furanose rings is therefore of interest to develop probes for understanding
their biological roles and as a prelude to possible therapeutic, vaccine,
or diagnostic development.[1,2] Such species are most
conveniently obtained by chemical synthesis, and over the past several
years many methods for accessing furanose-containing glycans have
been developed.[5,6]As is the case for their
pyranoside counterparts, the stereocontrolled
synthesis of 1,2-cis furanosides (e.g., 1 and 2, respectively, Figure A) is more challenging than that of 1,2-trans furanosides. Participating protecting groups at O-2
cannot be employed, and the inherently flexible five-membered ring
often further complicates stereocontrol by providing two energetically
similar reaction pathways that produce mixtures of products.[5] Consequently, the majority of robust methods
for the stereocontrolled synthesis of 1,2-cis-furanosides
employ conformationally locking groups resulting in a single energetically
favored reaction pathway.[5,6] Such methods have previously
been applied to the synthesis of β-arabinofuranosides (3, Figure B),[7−9] β-fructofuranosides (4),[10] and α-galactofuranosides (5).[11−13]
Figure 1
(A) Schematic representation
of 1,2-cis-α-furanosides
(1) and 1,2-cis-β-furanosides
(2). (B) Examples of β-arabinofuranosides (3), β-fructofuranosides (4), and α-galactofuranosides
(5).
(A) Schematic representation
of 1,2-cis-α-furanosides
(1) and 1,2-cis-β-furanosides
(2). (B) Examples of β-arabinofuranosides (3), β-fructofuranosides (4), and α-galactofuranosides
(5).In this context, we have
previously reported the preparation of
β-arabinofuranosides using the 2,3-O-xylylene-protected
arabinofuranosyl thioglycoside 6 (Figure A).[14] This donor
gave high 1,2-cis-selectivity in reactions with a
range of acceptor molecules. The selectivity was proposed to arise
from the cyclic xylylene protecting group fixing the electrophilic
intermediate produced upon activation into a single conformationally
rigid species. This donor is a particularly attractive reagent when
there is a need to modify O-5 of the β-arabinofuranoside residue
subsequent to glycosylation.
Figure 2
(A) 2,3-O-Xylylene-protected
donor (6) for the stereoselective synthesis of β-arabinofuranosides
and 2,3-O-xylylene-protected donors (7 and 8) explored in this paper as reagents for synthesizing
α-xylofuranosides. (B) Central hypothesis of this work: that 7 and 8 will lead to an intermediate (e.g., 9) that preferentially reacts[15] with nucleophiles to produce α-glycosides. (C) 5-Thiomethyl-α-xylofuranoside
(α-MTX) residue (10) present in mycobacterial LAM.
(A) 2,3-O-Xylylene-protected
donor (6) for the stereoselective synthesis of β-arabinofuranosides
and 2,3-O-xylylene-protected donors (7 and 8) explored in this paper as reagents for synthesizing
α-xylofuranosides. (B) Central hypothesis of this work: that 7 and 8 will lead to an intermediate (e.g., 9) that preferentially reacts[15] with nucleophiles to produce α-glycosides. (C) 5-Thiomethyl-α-xylofuranoside
(α-MTX) residue (10) present in mycobacterial LAM.We describe here an extension
of that work by exploring 2,3-O-xylylene-protected
xylofuranosyl thioglycosides (7 and 8) as
agents for the stereocontrolled installation
of α-xylofuranoside bonds. We hypothesized that activation of 7 and 8 would lead to the formation of electrophilic
intermediate (e.g., an oxacarbenium ion or a related ion pair) restricted
in a conformation with C-3 below the plane formed by the other ring
atoms (the E3 conformation). Such a species would be expected
to react preferentially from the bottom face following the inside
attack model[15] of Woerpel, leading to the
α-glycoside (Figure B). Interest in synthesizing α-xylofuranosides arises,
in part, from the presence of 5-thiomethyl-α-xylofuranoside
(α-MTX) residues (e.g., 10, Figure C) in the polysaccharidelipoarabinomannan
(LAM), a key immunomodulatory polysaccharide in the cell wall of mycobacteria.[16] Among the organisms possessing this motif is M. tuberculosis. The preparation of α-MTX-containing
glycoconjugates is of interest to provide tools for studying the biological
role of this unusual monosaccharide.[17] We,[18,19] and others,[16,20,21] have previously prepared oligosaccharides containing α-MTX
residues, but the xylofuranosylation reactions often proceeded with
modest stereoselectivity, and frequently the mixtures were difficult
to purify.
Results
and Discussion
Preparation
of 2,3-O-Xylylene-Protected Xylofuranoside Thioglycoside
Donors
The key intermediate in the synthesis of 7 and 8 (Scheme ) is the previously reported trityl derivative 11, which can be conveniently prepared in four steps from d-xylose.[18,22] Reaction of 11 with α,α′-dibromo-o-xylene and NaH in DMF and subsequent removal of the trityl
group under acidic conditions gave alcohol 12 in 41%
yield over the two steps. The major byproduct was a dimer resulting
from two equivalents of the 11 reacting with the alkylating
agent. A p-methoxybenzyl (PMB) group was introduced
on O-5 to obtain building block 7 in 92% yield upon reaction
of 12 with sodium hydride and PMBCl. Alternatively, acetylation
of 12 with acetic anhydride and pyridine gave 8 in quantitative yield.
Scheme 1
Synthesis of 7 and 8
Optimization
of Glycosylation Conditions
With donors 7 and 8 in hand, we first explored their reaction with a model acceptor,
1,2:3,4-di-O-isopropylidene-α-d-galactopyranose
(13), as detailed in Table . The stereochemistry of the newly introduced
glycosidic linkage could be identified by the coupling constant between
H-1 and H-2 (3J) in the 1H
NMR spectrum of the product. For the 1,2-cis (α)
isomer, 3J is ∼5.0 Hz. In contrast, 3J is ∼2.5 Hz for the 1,2-trans (β) isomer.[14] Further support for
these assignments came from the 1H and 13C NMR
spectra of deprotected compounds, including a hexasaccharide target
synthesized (see below), using this methodology. Data obtained from
these deprotected compounds was consistent[23] with the assignments made using the trends outlined above.
Table 1
Optimization of Glycosylation Conditions Using 7 and 8
entry
donor
solvent
yieldb
α:βc
product
1
7
CH2Cl2
86%
1:3.2
14
2
7
CH2Cl2–CH3CN (1:3)
86%
1:1.3
14
3
7
CH3CH2CN
81%
1:2.3
14
4
7
Et2O
88%
2.3:1
14
5
7
toluene–dioxane (1:3)
82%
1.8:1
14
6
8
Et2O
66%
9.5:1
15
7
8d
Et2O
78%
9.5:1
15
8
8e
Et2O
73%
9.5:1
15
Unless
otherwise indicated, all
reactions employed 1.3 equiv of donor, 1.0 equiv of 13, 2.5 equiv of NIS, and 0.25 equiv of AgOTf.
Combined yield of α- and β-isomers.
Determined from the ratio of
H-1
resonances in the 1H NMR spectrum of the corresponding
isomer mixture.
1.7 equiv
of donor used.
2.5 equiv
of donor used.
Unless
otherwise indicated, all
reactions employed 1.3 equiv of donor, 1.0 equiv of 13, 2.5 equiv of NIS, and 0.25 equiv of AgOTf.Combined yield of α- and β-isomers.Determined from the ratio of
H-1
resonances in the 1H NMR spectrum of the corresponding
isomer mixture.1.7 equiv
of donor used.2.5 equiv
of donor used.The effect
of the solvent on the stereoselectivity of the reaction
was investigated first. When 7 (1.3 equiv) was reacted
with 13 (1.0 equiv) in the presence of N-iodosuccinimide (NIS) and silver triflate (AgOTf) in different solvents
(Table , entries 1–5),
disaccharide 14 was obtained with α:β selectivities
ranging from ∼1:3 to ∼2:1. The use of diethyl ether
(entry 4) or a 1:3 mixture of toluene and dioxane (entry 5) provided
the best α-selectivity. These results are consistent with previous
work indicating that that ethereal solvents tend to increase α-stereoselectivity
in glycosylations.[24] It should be noted,
however, that this effect in furanosides has not been widely studied.The effects of the O-5 protecting group in the donor were next
explored. When donor 7, which bears an electron-rich p-methoxybenzyl group at O-5, was reacted with 13 (entry 4), disaccharide 14 was generated in good yield
and with moderate α-selectivity (α:β = 2.3:1). The
use of 8, with an electron-withdrawing acetyl group at
O-5, gave the desired disaccharide 15 in good yield and
with very good α-selectivity (α:β = 9.5:1, entry
6). Interestingly, this trend is opposite to what was observed with
donor 6, which was used for the synthesis of β-arabinofuranosides.[14] With 6, electron-rich groups on
O-5 led to higher 1,2-cis-selectivity. We do not
understand why these differences are observed. Indeed, if the reaction
proceeds through an oxacarbenium ion intermediate (see below), a more
electron-rich donor should, in theory, be more selective.Finally,
glycosylations in diethyl ether with different equivalents
of 8 were examined (entries 7 and 8). When 1.7 equiv
of 8 was used, the desired disaccharide 15 was formed in the best yield (78%) compared to when either 1.3 or
2.5 equiv was used. The α-selectivity (α:β = 9.5:1)
was the same in all cases.In other optimization studies, the
reaction concentration was shown
to have no substantial effect on the yield and selectivity of glycosylations
using 8 as the donor (Table S1). In addition, the β-anomer of 8 was prepared
and shown to give glycoside products in yields and selectivity identical
to 8. The studies outlined above identified optimized
conditions for these glycosylations: reaction of 1.7 equiv of donor 8 and 1.0 equiv of a glycosyl acceptor in the presence of
2.5 equiv of NIS and 0.25 equiv of AgOTf in diethyl ether at room
temperature. Under these conditions, the desired glycoside was formed
in high yield and with very good α-selectivity.
Substrate
Scope of the Xylofuranosylation
Having optimized the reaction,
its use with a range of carbohydrate acceptors was investigated. Glycosylations
with mono- and disaccharides 16–23 proceeded smoothly, giving products 25–32 (Table , entries 2–9) in good to excellent yield (67–96%)
and α-stereoselectivity (α:β 7:1 to >20:1). The
results were not uniformly excellent, however. Glycosylation of 24 with 8 (entry 10) provided disaccharide 33 with essentially no stereoselectivity (α:β
1.2:1). In contrast, glycosylation of mannopyranosides 20 and 23, which contain a free C-4 hydroxyl group, afforded
disaccharides 29 and 32 in excellent yield
and α-stereoselectivity (α:β > 17.7:1). This
is significant, as the α-d-xylofuranosyl-(1→4)-α-d-mannopyranose linkage produced in these reactions is the same
core structure as that present in α-MTX-functionalized mycobacterial
LAM. It should be noted that another advantage of the xylylene protecting
group is that when mixtures are produced, typically the two products
are can be separated without difficulty. For example, the separation
of the two diastereomers of 33 could be easily achieved.
Table 2
Xylofuranosylation of Various Carbohydrate Acceptors
by 8a
Reaction conditions: 1.7 equiv of
donor, 1 equiv of acceptor, 2.5 equiv of NIS, and 0.25 equiv of AgOTf.
Combined yield of α-
and β-isomers.
Determined
from the ratio of H-1
resonances in the 1H NMR spectrum of the corresponding
isomer mixture.
Reaction conditions: 1.7 equiv of
donor, 1 equiv of acceptor, 2.5 equiv of NIS, and 0.25 equiv of AgOTf.Combined yield of α-
and β-isomers.Determined
from the ratio of H-1
resonances in the 1H NMR spectrum of the corresponding
isomer mixture.
Computational
Investigations of the Reaction
In an effort to understand
the selectivity obtained in the reaction, we carried out quantum mechanical
calculations to determine the energetics of the reaction coordinate.
First, the optimized geometry of the oxacarbenium ion that results
from activation of the thioglycoside was determined at the HF/6-311G(d,p)//B3LYP/6-31G(d)
level of theory. Although the true intermediate in these reactions
is likely not a “naked” oxacarbenium ion,[25] this intermediate was chosen to simplify the
calculations. In addition, a model with a 5-O-benzoyl
group (34, Figure ) was used to facilitate subsequent transition state calculations
to avoid undesired bond stretch/bend frequencies associated with the
methyl group of the acetateester. As we hypothesized at the outset
of this project (Figure B), the lowest energy conformation of 34 is the E3 envelope (Figure , middle). This is similar in structure to the lowest energy
conformation of a xylo-configured 2,3,5-tri-O-methylated oxacarbenium ion (35), calculated
earlier by Codée and co-workers.[26] Oxacarbenium ion 35 adopts a 4T3 twist conformation, with C-4 above and C-3 below the plane formed
by C-1, C-2, and the ring oxygen.
Figure 3
Structure of oxacarbenium ion 34 (left), its low-energy
conformation as determined by DFT calculations (middle), and a 2,3,5-tri-O-methyl-xylofuranosyl oxacarbenium (35) studied
earlier by Codée and co-workers (right).[26]
Structure of oxacarbenium ion 34 (left), its low-energy
conformation as determined by DFT calculations (middle), and a 2,3,5-tri-O-methyl-xylofuranosyl oxacarbenium (35) studied
earlier by Codée and co-workers (right).[26]From the oxacarbenium ion, two
pathways were examined: one where
the nucleophile (methanol was used in these calculations) can attack
from the top face resulting in 1,2-trans product
or from the bottom face resulting in a 1,2-cis product.
Both products were optimized, and their geometries are shown in Figure . To simplify the
calculations, we assumed that loss of the proton from the methanol
was fast and occurred following glycosylation (i.e., after C–O
bond formation). As such the optimized products are protonated glycosides.
The Supporting Information contains animations
showing the both approaches of the nucleophile to 34.
The 1,2-cis product adopts the same conformation
as the oxacarbenium ion (E3) whereas the 1,2-trans-product has a slightly different conformation (4T3). These are the same as the tri-O-methylated
derivatives studied previously by Codée and co-workers.[26] The relative energies that were observed indicated
that the 1,2-cis product is 6.3 kcal/mol lower in
energy than the 1,2-trans product.
Figure 4
Optimized geometries
of the 1,2-cis (left) and
1,2-trans (right) glycoside products formed upon
trapping of oxcarbenium ion 34 with methanol.
Optimized geometries
of the 1,2-cis (left) and
1,2-trans (right) glycoside products formed upon
trapping of oxcarbenium ion 34 with methanol.Using the optimized geometries of the two products,
we carried
out coordinate scans on the C1–O1 bond at the HF/6-31G(d) level
to obtain a picture of the energetics of the reaction coordinate.
As can be observed in Figure , the path toward the 1,2-cis product has
a lower energy barrier than that producing the 1,2-trans-product. These data suggest that nucleophilic trapping of 34 by a nucleophile from the bottom face of the ring is kinetically
favored. Each of the scan structures was used as input for transition
state calculations, which were done using HF/6-31G(d) level of theory.
Frequency analysis showed that one imaginary frequency was obtained
for each of the transition state structures. Geometries earlier in
the reaction coordinate showed imaginary frequencies of −147.9
and −155.2 cm–1 for the cis and trans products, respectively. These correspond
to OCH3 rotational vibrations. Higher energy structures
near the transition state showed imaginary frequencies of −186.9
and −233.9 cm–1 corresponding to the desired
C1–O1 bond stretching.
Figure 5
Coordinate scan of the C1–O1 bond from
1.5 to 2.4 Å
in the 1,2-trans (open circles) and 1,2-cis (filled triangles) methyl glycoside products arising from trapping
of 34 with methanol.
Coordinate scan of the C1–O1 bond from
1.5 to 2.4 Å
in the 1,2-trans (open circles) and 1,2-cis (filled triangles) methyl glycoside products arising from trapping
of 34 with methanol.Taken together, these computational results provide support
for
our hypothesis that electrophilic intermediates derived from 2,3-O-xylylene-protected-xylofuranoside glycosyl donors adopt
conformations that should, based on Woerpel’s inside attack
model,[15] lead to α-glycosides. Moreover,
the transition state calculations suggest that trapping of these electrophiles
by nucleophiles from the bottom face, producing α-glycosides,
is indeed energetically favored. Although a similar hypothesis was
proposed for reactions leading to β-arabinofuranosides from 6,[14] the calculations here represent
the first support for the origin of the stereoselectivity in glycosylation
reactions with xylylene-protected-furanoside donors. It should be
noted that remote participation by the O-5ester group could also
provide an explanation for the high α-selectivity. We think
this unlikely as the conformational rigidity imposed on the furanose
ring by the xylylene group would result in significant torsional strain
should remote participation occur. In addition, the use of other electron-withdrawing,
but not participating, groups at this position (see below) also give
high yields of the 1,2-cis-product.
Synthesis
of an α-MTX-Containing Hexasaccharide Fragment of LAM
The studies described above demonstrate the successful use of 2,3-O-xylylene-protected thioglycoside 8 for the
stereoselective synthesis of α-xylofuranosides and, in particular,
the α-d-xylofuranosyl-(1→4)-α-d-mannopyranose motif. We then turned our attention to applying the
methodology to a more complex biologically relevant molecule. We selected
as a target a hexasaccharide fragment of mycobacterial LAM (36, Scheme ). This molecule, and related α-MTX-containing compounds, has
recently been incorporated into a glycan array for use in mapping
the specificity of proteins that recognize mycobacterial glycans.[19]
Scheme 2
Retrosynthetic Analysis of 36
The previous synthesis of 36 did not employ 2,3-O-xylylene-protected
donors to introduce either the α-MTX
or the β-arabinofuranoside residues.[19] Therefore, we endeavored to use both 6 and a derivative
of 8 for their installation. To streamline the synthesis,
we explored the use of the 5-O-mesyl-2,3-O-xylylene-thioglycoside 37 in place of 8. We postulated that the electron-withdrawing nature of the
mesyl group would lead to high α-selectivity, and the resulting
product could then be subjected directly to a substitution reaction
to introduce the thiomethyl group. This approach would save two steps
(deacetylation and mesylation) postglycosylation compared to if 8 were used. We also planned to use thioglycoside 38 and the previously reported trisaccharide 39(19) in the synthesis of 36. The synthesis
of 37 and 38 is shown in Scheme and was achieved from 12 (above) and the known thioglycoside 40,[27] respectively.
Scheme 3
Synthesis of Thioglycosides 37 (A) and 38 (B)
Implementation of this strategy is shown in Scheme . The synthesis began with
the treatment
of trisaccharide 39 and thioglycoside 6 with
NIS and trimethylsilyl trifluoromethanesulfonate (TMSOTf) leading
to the formation of tetrasaccharide 42 in 80% yield.
The PMB group was then cleaved by treatment with TFA yielding a 91%
yield of alcohol 43. Glycosylation of 43 with 38, again promoted by NIS and TMSOTf, resulted
in an 82% yield of pentasaccharide 44, which, upon treatment
with hydrazine acetate, gave 45, the substrate for the
key xylofuranosylation reaction. We were pleased to discover that
the reaction of 45 with 37 did indeed give
the desired product, 46, in an excellent 87% yield. None
of the β-glycoside product was detected in the reaction. It
should be noted, however, that this reaction did require optimization.
In particular, donor 37 was not soluble in diethyl ether,
the solvent shown (Table ) to be optimal for the high degree of α-selectivity
in these reactions. Fortunately, the use of THF, in which 37 is soluble, provided the desired α-glycoside with high selectivity.
The high yield obtained in formation of 46 suggests that
the origin of the α-selectivity using 8 arises
from the electron-withdrawing nature of the O-5 substituent, not remote
participation. If participation were essential, the reaction of 45 with 37, which has a nonparticipating mesylate
at O-5, would be expected to give the product with lower 1,2-cis selectivity.
Scheme 4
Synthesis of Hexasaccharide 36
Having installed all of the
monosaccharide residues, the mesyl
group was displaced, and the benzoate esters cleaved, using an excess
of sodium thiomethoxide in acetonitrile at reflux. The conversion
of 46 into 47 proceeded in 74% yield. Finally,
the benzyl and xylylene protecting groups were cleaved using dissolving
metal reduction, leading to a 74% yield of hexasaccharide 36. The synthesis of the mycobacterial LAM fragment thus could be achieved
in seven steps and 26% overall yield from 39. This compares
favorably to our previous synthesis of this compound, also from 39,[19] which required 11 steps and
two difficult chromatographic purifications, leading to the formation
of the product in 12% overall yield.
Summary
In conclusion,
a novel stereoselective 1,2-cis α-xylofuranosylation
employing a conformationally constrained
2,3-O-xylylene-protected xylofuranosyldonor has
been developed. Optimization of the glycosylation conditions showed
that treatment of 1.7 equiv of p-tolyl 5-O-acetyl-1-thio-2,3-O-xylylene-α-d-xylofuranoside (8) with 1.0 equiv of various
acceptors, in the presence of 2.5 equiv of NIS and 0.25 equiv of AgOTf
in diethyl ether at room temperature, gives the product in high yield
and in high α-selectivity. Computational investigations suggest
that the reaction produces an electrophilic intermediate with a conformation
that favors the formation of the desired α-glycoside. In an
application of the method, we prepared a hexasaccharide fragment of
mycobacterial LAM in which two 1,2-cis-furanosyl
residues, a β-arabinofuranoside and an α-xylofuranoside,
were installed using 2,3-O-xylylene-protected thioglycoside
donors. This synthesis highlights the power of these reagents in the
assembly of glycans containing 1,2-cis-furanosides,
particularly those with modifications at O-5.
Experimental
Section
General
Methods
All reagents used were purchased from commercial
sources and were used without further purification unless noted. Solvents
used in reactions were purified by successive passage through columns
of alumina and copper under nitrogen. Unless stated otherwise, all
reactions were carried out under a positive pressure of argon. Reactions
were monitored by TLC on Silica Gel 60-F254 (0.25 mm),
and spots were visualized under UV light (254 nm) and/or stained by
charring with acidified anisaldehyde solution in ethanol. Column chromatography
was performed on Silica Gel 60 (40–60 μm) or C18 silica gel (35–70 μm, Toronto Research Chemicals). 1H NMR spectra were recorded at 500, 600, or 700 MHz, and chemical
shifts are referenced to CHCl3 (7.26, CDCl3)
or CH3OD (3.35, CD3OD). 13C NMR spectra
were recorded at 125 or 175 MHz, and 13C chemical shifts
are referenced to CDCl3 (77.06, CDCl3) or CD3OD (49.0, CD3OD). Assignments of NMR spectra were
made on the basis of two-dimensional experiments (1H–1H COSY, HSQC, and HMBC). ESI-MS spectra were recorded on samples
suspended in THF or CH3OH with added NaCl. Optical rotations
were measured at 22 ± 2 °C at the sodium D-line (589 nm)
and are in units of deg·mL(dm·g)−1.
General
Procedure for Glycosylation Reactions
To a mixture of donor
(1.7 mmol) and acceptor (1 mmol) in Et2O were added 4 Å
molecular sieves (200 mg) at rt. After stirring for 1 h, NIS (2.5
mmol) and AgOTf (0.25 mmol) were added to the mixture. The reaction
was monitored by TLC, and after stirring for 2 h at rt, the reaction
was quenched by the addition of Et3N. The solution was
diluted with CH2Cl2 and filtered through Celite.
The filtrate was then washed with satd aq. Na2S2O3 and brine. The organic layer was subsequently dried
with anhydrous Na2SO4, filtered, and concentrated,
and the residue was purified by flash silica gel column chromatography
(hexane–EtOAc) to give the product.
To a solution of 11(18) (2.26 g, 4.56 mmol) in DMF
(40 mL) were added NaH (400 mg, 10.04 mmol) and α,α′-dibromo-o-xylene (1.32 g, 5.02 mmol) slowly at 0 °C. After
stirring for at 2 h, the reaction was quenched by the addition of
satd aq. NH4Cl. Dilution of the mixture with CH2Cl2 provided a solution that was then washed with brine.
The organic layer was dried with anhydrous Na2SO4, filtered, and concentrated. The crude product was purified by flash
column chromatography (20:1 hexane–EtOAc) to give the crude
product. The obtained crude product was dissolved in CH2Cl2–methanol (7:3, 20 mL), and then p-TsOH (80 mg) was added. The mixture was stirred for 14 h and then
neutralized by the addition of Et3N. Then the mixture was
concentrated to syrup, which was purified by column chromatography
(6:1 hexane–EtOAc) to give 12 (990 mg, 41% over
two steps) as a colorless oil. Rf 0.3
(3:1, hexanes–EtOAc); [α]D = −78.3
(c = 0.9, CHCl3); 1H NMR (500
MHz, CDCl3) δH 7.43–7.32 (m, 6H,
ArH), 7.10–7.06 (m, 2H, ArH), 5.07 (d, 1H, J = 7.3 Hz, H-1), 4.93 (d, 1H, J = 12.7 Hz, OCH2Ph), 4.92 (d, 1H, J = 12.5
Hz, OCH2Ph), 4.82 (d, 1H, J = 12.7 Hz, OCH2Ph), 4.75 (d, 1H, J = 12.6 Hz, OCH2Ph), 4.26 (dd,
1H, J = 5.4, 7.9 Hz, H-3), 4.16 (ddd, 1H, J = 7.9, 5.1. 5.1 Hz, H-4), 4.04 (dd, 1H, J = 7.3, 5.4 Hz, H-2), 3.76–3.73 (m, 2H, H-5), 2.34 (app t,
1H, J = 7.1 Hz, OH), 2.31 (s, 3H ArCH3); 13C NMR (125 MHz, CDCl3) δC 137.8 (Ar), 136.3 (Ar), 135.2 (Ar), 132.4 (2C, Ar), 131.7
(Ar), 131.4 (Ar), 129.9 (Ar), 129.7 (Ar), 129.7 (2C, Ar), 129.6 (Ar),
89.0 (C-1), 85.3 (C-2), 81.4 (C-3), 80.6 (C-4), 69.8 (OCH2Ph), 68.90 (OCH2Ph), 61.8
(C-5), 21.1 (ArCH3); HRMS-ESI m/z [M + Na]+ calcd for C20H22NaO4S: 381.1131. Found: 381.1128.
To a solution of liquid NH3 (25 mL) at −78
°C was added sodium metal (50 mg) until a deep blue solution
was produced. A solution of 47 (37 mg, 0.02 mmol) in
THF (0.5 mL) was then added over a period of 1 min, making sure that
the deep blue color persisted. The reaction mixture was stirred at
−78 °C for 2 h, and CH3OH was then added until
the dark blue color disappeared and the solution became clear. The
solution was then warmed to rt by blowing air gently over the solution,
which also helped evaporate the NH3. When the reaction
mixture reached rt, CH3OH (8 mL) was added, and the pH
of the solution was brought to ∼8 by the careful addition of
Amberlite IR 120 H+ resin. The solution was filtered to
remove the resin, and the filtrate was concentrated. The residue was
purified by chromatography (6:2:1 CH3OH–NH4OH–H2O) on Iatrobeads 6RS-8060 to give 36 (14.5 mg, 74%) as a thick syrup that was later lyophilized from
water to give a fluffy solid. Rf = 0.43
(5:2:1 CH3OH–NH4OH–H2O); [α]D +72.6 (c = 0.2, CH3OH); 1H NMR (700 MHz, CD3OD) δH 5.37 (d, 1H, J = 4.2 Hz, H-1), 5.08 (d,
1H, J = 1.8 Hz, H-1), 5.03 (d, 1H, J = 4.8 Hz, H-1), 4.95 (d, 1H, J = 1.2 Hz, H-1),
4.84 (d, 1H, J = 1.2 Hz, H-1), 4.32 (dt, 1H, J = 7.2, 5.4 Hz), 4.15–4.10 (m, 2H), 4.09–3.77
(m, 26H), 3.72–3.64 (m, 7H), 3.42 (dt, 1H, J = 9.6, 6.5 Hz), 3.07–3.02 (m, 1H), 2.76–2.71 (m, 1H),
2.63–2.58 (m, 2H), 2.16 (s, 3H), 1.59–1.42 (m, 4H),
1.41–1.30 (m, 8H); 13C NMR (125 MHz, CD3OD) δC 109.7 (C-1), 109.5 (C-1), 107.4 (C-1), 104.8
(C-1), 102.4 (C-1), 101.7 (C-1), 89.5, 84.4, 84.0, 83.7, 83.5, 83.3,
82.3, 80.5, 79.2, 79.0, 78.9, 78.4, 77.2, 76.7, 76.1, 76.0, 73.5,
72.3, 72.2, 69.8, 68.9, 68.2, 68.2, 62.9, 62.4, 42.6 (CH2NH2), 34.5, 34.0, 30.7 (octyl CH2), 30.6 (octyl
CH2), 30.4 (octyl CH2), 28.0 (octyl CH2), 27.2 (octyl CH2), 16.3 (SCH3). HRMS-ESI m/z [M + H]+ calcd for C40H72NO25S: 998.4109. Found: 998.4134.
All calculations were performed in the Gaussian
09 program.[28] Geometry optimizations were
run for oxocarbenium ion 34, as well as the 1,2-cis product and 1,2-trans products of its
trapping with methanol at the HF/6-311G(d,p)[29−32]//B3LYP/6-31G(d)[33,34] level of theory. Using the optimized geometries of the two products,
a reaction coordinate scan was carried out on the C1–O1 bond
from 1.5 to 2.4 Å to obtain a picture of the energetics of the
reaction coordinate. Each of the scan structures was used as input
for transition state calculations using the HF/6-31G(d)[29,34] level of theory. Frequency analysis showed that one imaginary frequency
was obtained for each of the structures, as described above. All calculations
were done in the gas phase.
Authors: Erwin R van Rijssel; Pieter van Delft; Gerrit Lodder; Herman S Overkleeft; Gijsbert A van der Marel; Dmitri V Filippov; Jeroen D C Codée Journal: Angew Chem Int Ed Engl Date: 2014-07-31 Impact factor: 15.336
Authors: Shiva Kumar Angala; Juan Manuel Belardinelli; Emilie Huc-Claustre; William H Wheat; Mary Jackson Journal: Crit Rev Biochem Mol Biol Date: 2014-06-10 Impact factor: 8.250
Authors: Stefan van der Vorm; Thomas Hansen; Erwin R van Rijssel; Rolf Dekkers; Jerre M Madern; Herman S Overkleeft; Dmitri V Filippov; Gijsbert A van der Marel; Jeroen D C Codée Journal: Chemistry Date: 2019-04-17 Impact factor: 5.236
Authors: Matthias Krumb; Maximilian Jäger; Alice Voss; Loreen Immig; Karin Peters; Danuta Kowalczyk; Albrecht Bufe; Till Opatz; Otto Holst; Christian Vogel; Marcus Peters Journal: Chemistry Date: 2020-10-27 Impact factor: 5.020
Figure 1
Figure 2
Scheme 1
Figure 3
Figure 4
Figure 5
Scheme 2
Scheme 3
Scheme 4