Reactivity–Stereoselectivity Mapping for the Assembly of Mycobacterium marinum Lipooligosaccharides

Article abstract of DOI:10.1002/anie.202010280

The assembly of complex bacterial glycans presenting rare structural motifs and cis-glycosidic linkages is significantly obstructed by the lack of knowledge of the reactivity of the constituting building blocks and the stereoselectivity of the reactions i

Full text of DOI:10.1002/anie.202010280

A Journal of the Gesellschaft Deutscher Chemiker
Angewandte
Chemie
International Edition
www.angewandte.org
Accepted Article
Title: Reactivity-Stereoselectivity Mapping for the Assembly of
Mycobacterium Marinum Lipooligosaccharides
Authors: Thomas Hansen, Tim P. Ofman, Joey G. C. Vlaming, Ivan
A. Gagarinov, Jessey van Beek, Tessa A. Goté, Jacoba
M. Tichem, Gijs Ruijgrok, Herman S. Overkleeft, Dmitri V.
Filippov, Gijsbert A. van der Marel, and Jeroen D. C. Codée
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Angew. Chem. Int. Ed. 10.1002/anie.202010280
Link to VoR: https://doi.org/10.1002/anie.202010280

10.1002/anie.202010280
Angewandte Chemie International Edition
Reactivity-Stereoselectivity Mapping for the Assembly of
Mycobacterium Marinum Lipooligosaccharides
Thomas Hansen,‡ Tim P. Ofman,‡ Joey G. C. Vlaming,‡ Ivan A. Gagarinov, Jessey van Beek,
Tessa A. Goté, Jacoba M. Tichem, Gijs Ruijgrok, Herman S. Overkleeft, Dmitri V. Filippov,
Gijsbert A. van der Marel, Jeroen D. C. Codée*
‡ TH, TO and JV contributed equally
Leiden University, Leiden Institute of Chemistry, Einsteinweg 55, 2333 CC Leiden, The Netherlands
* Correspondence to: jcodee@chem.leidenuniv.nl
ABSTRACT
The assembly of complex bacterial glycans presenting rare structural motifs and cis-glycosidic linkages
is significantly obstructed by the lack of knowledge of the reactivity of the constituting building blocks
and the stereoselectivity of the reactions in which they partake. We here report a strategy to map the
reactivity of carbohydrate building blocks and apply it to understand the reactivity of the bacterial
sugars, caryophyllose, a rare C12-monosaccharide, containing a characteristic tetrasubstituted
stereocenter. We mapped reactivity-stereoselectivity relationships for caryophyllose donor and
acceptor glycosides, by a systematic series of glycosylations in combination with the detection and
characterization of different reactive intermediates using experimental and computational techniques.
The insights garnered from these studies enabled the rational design of building blocks with the required
properties to assemble mycobacterial lipooligosaccharide fragments of M. marinum.
INTRODUCTION
The bacterial glycan repertoire is equally vast and diverse.[1–5] As opposed to the mammalian
carbohydrate biosynthesis machinery that employs a limited set of 9 monosaccharides[6] to build
oligosaccharides and glycoconjugates, the bacterial biomachinery can introduce a wide variety of
substitution patterns.[1–5] Bacterial monosaccharides can feature diversely substituted amino groups,
deoxy centers, carbonyl groups, and tetrasubstituted tertiary carbon atoms at various positions on the
carbohydrate ring. Tertiary-C sugars can be found in various natural products, having attractive
biological properties.[7–10] Often these tertiary C-atoms are substituted with a small alkyl group,
commonly a methyl substituent, but more complex architectures in which functionalized alkyl chains
are attached can be found as well. For example, the tertiary C-sugar caryophyllose (Car, see Figure 1A)
is found in mycobacterial lipooligosaccharides (LOSs).[11–13] This unique structure bears a hydroxylated
C6-chain at the tetrasubstituted tertiary C4-atom.
1
This article is protected by copyright. All rights reserved.

10.1002/anie.202010280
Angewandte Chemie International Edition
OH
OH
OH
A
O
O
OH OH
OH
OH
OH
OH
OH
OH
caryophyllose (Car)
yersiniose A (YerA)
B
LOS-IV
O
LOS-III
LOS-II
R4
Zc*
OH
N
R3
O
LOS-I
NH
O
HO
R2
HO
HO
O
HO
HO
HO
HO
IX
HO
O
HO
O
HO
HO
O
O
HO
O
O
O
O
VIII
R2
HO
OH
III
O
OH
IV
HO
II
O
O
HO
VII
O
O
OH
R1: 2,4-dimethyl-branched
fatty acid chain
R2: H or OH
R3: H or COOH
R4: H or OCH3
O
HO
OH
OH
O
O
HO
R1
MeO
OH
O
OH
OH
VI
O
V
O
I
R1
O
O
R1
C
OTBS
O
O
LOS-IV fragment 1
O
O
Cl
N
2
4
4
3
OH
CO2Bn
BnO
MeO
BnO
OTBS ONap O
OMe
MeO
from
D
-glucose
from
D
-glucose
HO
O
5
6
pyrrolidone
2
N3
N
O
HOOC
2 building blocks
O
O
3 building blocks
cis-glycosylation
(DFT-guided)
BnO
SPh
HN
peptide
O
OBn
coupling
O
O
OH
HO
HO
O
4-azido-fucose donor
3
O
NapO
O
HO
OH
O
cis-glycosylation
additive-based
glycosylation
HO
BnO
SPh
O
HO
O
SmI2-mediated C-C
formation
O
HO
OH
O
HO
caryophyllose donor
4
O
OH
HO
Figure 1. Lipooligosaccharides from M. marinum and the target fragment with a retrosynthetic analysis. (A) Tertiary C-
sugar caryophyllose (Car) found in mycobacterial lipooligosaccharides and the related yersiniose A (YerA). (B) LOS-IV from
M. marinum. with numbering introduced by Rombouts et al.11-13 (C) Retrosynthetic analysis for LOS-IV fragment 1.
The mycobacterial LOSs are major constituents of the thick and waxy cell wall of mycobacteria.[11–
16] Being at the host-pathogen interface, they play an important role in the interaction with the immune
system. Because it is exceedingly laborious to purify these lipophilic compounds from the bacterial cell
wall, it has proven difficult to establish the precise role of these glycolipids in shaping an immune
response. Mycobacterium marinum is a waterborne pathogen that is most closely related to
Mycobacterium tuberculosis, and causes tuberculosis-like infections. As such it is often used as a surrogate
to study host-pathogen interactions involved in M.tb infections. M. marinum produces four LOS
structures (LOS-I–IV; Figure 1B), which all share an acylated trehalose core functionalized with
species-specific glycans. The LOS-II, LOS-III, and LOS-IV structures of M. marinum contain several
unusual carbohydrate monosaccharides, including the tertiary C-sugar caryophyllose as well as an N-
acylated 4-amino-4-deoxy-D-fucose (NAcFuc).[11–13,17–19] The complex carbohydrates of the higher
LOS-structures thus seem to play an important role in immune evasion although the exact mode of
action of these remains ill-understood.
2
This article is protected by copyright. All rights reserved.

10.1002/anie.202010280
Angewandte Chemie International Edition
The compelling bioactivity, intriguing structural features, and the fact that well-defined pure LOS
structures cannot be obtained from natural sources in sufficient amounts for biological studies
motivated us to develop synthetic chemistry to attain these complex structures. Although great progress
has been made in oligosaccharide synthesis, the assembly of bacterial glycans presenting rare structural
modifications and challenging cis-glycosidic linkages still presents a major obstacle as the reactivity of
the required building blocks is not well understood.[20–37] We here report an approach to map the
reactivity-stereoselectivity relationships for the tertiary C-sugar caryophyllose and its truncated
counterpart yersiniose A (YerA; Figure 1A). This has allowed us to effectively construct the Car-Car-
NAcFuc LOS-IV fragment 1 (Figure 1C), and related shorter fragments, equipped with an alkene spacer
for future conjugation purposes. The approach taken to understand the reactivity and stereoselectivity
of these rare and challenging bacterial monosaccharides hinges on the detection and characterization of
different reactive intermediates using experimental and computational techniques. These combined
studies have enabled the rational design of building blocks with the desired reactivity and selectivity to
assemble the spacer equipped LOS-IV fragment 1 with complete stereoselectivity. Our study discloses
the intrinsic reactivity of tertiary-C Car donors, which can act as a prototype for related tertiary-C sugars
that can fuel biological research.
RESULTS AND DISCUSSION
The Car-Car-NAcFuc carbohydrate 1 was assembled from the three key monomeric building blocks,
pyrrolidone 2, 4-amino-4-deoxy-D-fucose 3 and caryophyllose 4 (Figure 1C). The design of the latter
building block was based on reactivity studies, as outlined below. Pyrrolidone 2 can be synthesized
based on the work of the Lowary group from D-serine and the 4-azido-fucose donor 3 can be made from
D-glucose by deoxygenation of C6 and an inversion of the C4 position, following established
procedures.[38] Car donor 4 can be synthesized from building blocks 5 and 6 by a SmI2-mediated C-C
bond formation, as originally described by Prandi and co-workers.[39,40]
Our first goal was the generation of sufficient amounts of the Car donor glycosides, required to
map the reactivity of these building blocks and build the target fragment 1. To this end acid chloride 5
and 2,6-dideoxy-4-keto-glucose 6 were assembled. The synthesis of 5 is depicted in Scheme 1A and
started from methyl-α-D-glucopyranose. Epoxide 7 was readily prepared in three steps (>150 gram
scale). Regioselective opening of the epoxide with LiAlH4 afforded digitoxose-configured 8 in good
yield (78%, >120 gram scale). Installation of the temporary 2-methylnaphthyl protecting group resulted
in fully protected 9. The 4,6-O-benzylidene protecting group was removed using a catalytic amount of
I2 to yield diol 10 (99%), and the primary alcohol was converted into iodide 11 using
triphenylphosphine, iodide and imidazole. Radical reduction using NaBCNH3 and AIBN yielded the
partially protected D-digitoxose 12 (74%, >100 gram scale).[41] Hydrolysis of D-digitoxose 12 with 25%
v:v aq. AcOH, followed by the treatment with ethanethiol and concentrated HCl afforded the linear
diethyl dithioacetal 13 (67% over two steps, 50 gram scale). Subsequently, both hydroxyl functions of
the dithioacetal were protected with a TBS group using TBSOTf and pyridine to yield the fully
protected 14 (62%).
3
This article is protected by copyright. All rights reserved.

10.1002/anie.202010280
Angewandte Chemie International Edition
1. i. I2, PhCH(OMe)
2, Δ
A
ii. MsCl, pyr
iii. KOH, Δ
HO
HO
HO
Ph
O
Ph
O
3. NapBr
NaH
Ph
O
O
O
O
2. LiAlH4
O
O
O
O
HO
55% over
3 steps
quant.
78%
O
OMe
OMe
OH
OMe
ONap OMe
9
7
8
4. I2
99%
OH
7. i. AcOH
5. I2, TPP,
Imidazole, Δ
6. NaBCNH3
AIBN, Δ
I
HO
HO
ii. EtSH, HCl
O
O
O
SEt
HO
HO
66%
67%
over 2 steps
74%
OMe
OMe
OH
ONap SEt
13
OMe
ONap
12
ONap
11
ONap
10
[steps 1-6,
>100 gram scale]
8. TBSOTf
pyridine
62%
OTBS
OTBS
OTBS
OTBS
9. NaHCO3
I2
OH
Cl
SEt
10. KMnO4
11. (COCl)2
OTBS ONap O
16
OTBS ONap O
81%
75%
OTBS ONap O
15
OTBS ONap SEt
14
5
(used crude)
[steps 7-10,
>30 gram scale]
1. TBI
TPP, Δ
2. AIBN
Bu3SnH, Δ
3. i. O=SnBu3
ii. BnBr, Δ
Br
HO
O
O
B
HO
HO
O
O
HO
HO
HO
HO
60%
31%
HO
HO
18
99%
BnO
OMe
Br
17
OMe
OMe
[steps 1-3,
>10 gram scale]
OMe
19
O
O
4. DMP
90%
O
6. i. Zn(BH4)2
ii. 6 M HCl
iii. CDI
5. SmI2, Δ
OH
6
OTBS
O
O
7. Ac2O, H2SO4
12
8
5
4
NapO
5
2 O
O
11
9
7
10
3
1
BnO
94%
82%
OMe
58%
BnO
20
OTBS ONap O
BnO
6
over 3 steps
[steps 6-8,
>5 gram scale]
O
OMe
[gram
scale]
OMe
O
21
O
O
O
yersiniose A donor
caryophyllose donor
O
O
O
O
BnO
BnO
O
BnO
O
NapO
O
O
BnO
R
BnO
23
SPh
BnO
SPh
BnO
ONap OBn
24
R = OAc 22
R = SPh
BnO
25
SPh
OBn
8.
O
4
O
61%
O
Scheme 1. Synthesis of 4, 5, 6, 23, 24, and 25. (A) Synthesis of building block 5. Reagents and conditions: (1) i. benzaldehyde
dimethyl acetal, I2, CH3CN; ii. MsCl, pyridine; iii. KOH, THF/MeOH (55% over three steps); (2) LiAlH4, Et2O (78%); (3)
NapBr, NaH, DMF (quant.); (4) I2, MeOH (99%); (5) imidazole, triphenylphosphine, I2, toluene, 75 °C (66%); (6)
NaBCNH3, AIBN, t-BuOH, 80 °C (74%); (7) i. aq. 25% AcOH, reflux; ii. EtSH, aq. 37% HCl (67% over 2 steps); (8) TBSOTf,
pyridine, DCM (62%); (9) I2, NaHCO3, acetone, H2O (81%); (10) KMnO4 aq., NaH2PO4 aq., t-BuOH (75%); (11) (COCl)2,
pyridine. (B) Synthesis of donor 4 and 23. Reagents and conditions: (1) tribromoimidazole, triphenylphosphine, toluene, reflux
(60%); (2) AIBN, Bu3SnH, toluene, reflux (99%); (3) i. tributyltin oxide, toluene, reflux; ii. benzyl bromide, reflux (31%);
(4) DMP, DCM (91%); (5) SmI2, 5, THF, 50 °C, 15 min (82%); (6) i. Zn(BH4)2, THF; ii. 6 M HCl, MeOH; iii. CDI, DCM
(58% over three steps); (7) Ac2O, H2SO4, 1 min (94%); (8) thiophenol, BF3·OEt2, DCM (61%). Pyr = pyridine, NapBr = 2-
(bromomethyl)naphthalene, TBI = 2,4,5-tribromoimidazole, TBSOTf = t-butyldimethylsilyl trifluoromethanesulfonate, TPP
= triphenylphosphine, AIBN = azobisisobutyronitrile, DMP = Dess-Martin periodinane, CDI = carbonyldiimidazole.
Treating dithioacetal 14 with I2 and NaHCO3 in acetone/water delivered the corresponding aldehyde
15 in 81% yield (30 gram scale). Oxidation with potassium permanganate in t-BuOH/water of aldehyde
15 furnished the protected acid 16 (75%), which could be converted to building block 5 with pyridine
and oxalyl chloride.
As depicted in Scheme 1B building block 6 was also synthesized from methyl-α-D-glucopyranoside,
starting with the regioselective bromination of the C3- and C6-position using tribromoimidazole in
good yield (60%, >30 gram scale). Removal of the bromides using tributyltin hydride and AIBN
afforded the required dideoxy glucoside 18 in excellent yield (99%, 15 gram scale). The reaction of 18
with tributyltin oxide, followed by benzyl bromide provided the benzylated glucoside 19 in 31% yield.
4
This article is protected by copyright. All rights reserved.

10.1002/anie.202010280
Angewandte Chemie International Edition
Oxidation of the C4-alcohol in 19 with Dess-Martin periodinane then afforded key building block 6
(90%).
To build the tertiary C-sugar having the required Car-configuration, a SmI2-promoted C-C bond
coupling was employed using acyl chloride 5 and ketone 6 (Scheme 1B). The best yield for this cross-
coupling was obtained by premixing both coupling partners and quickly adding them, by canula, to a
warm (50 °C) solution of SmI2 in THF under completely inert atmosphere. This procedure reliably
delivered ketone 20 with the required stereochemistry at C4 in 82% yield (gram scale). A chelation
controlled reduction of ketone 20 with Zn(BH4)2 in THF then afforded the free alcohol. After removal
of the silyl protection groups and protection of the two vicinal diols using carbonyldiimidazole
caryophyllose 21 was obtained in 58% (over three steps, 15 gram scale). Proof for the stereochemistry
of the C7 position was obtained by NOESY NMR experiments (See SI). The anomeric methoxy group
of caryophyllose 21 was then converted to an acetyl group using H2SO4 in acetic anhydride. The
anomeric acetate 22 was formed in excellent yield (94%, >5 gram scale) and subsequently transformed
into the key caryophyllose thioglycoside 4 under the aegis of thiophenol and BF3·OEt2. Following a
highly similar route, the per-O-Bn caryophyllose thioglycoside 24 was constructed (See SI). We also
assembled yersiniose A (YerA) donors 23 and 25, to be used as model donors to map the reactivity-
selectivity of these type of donors (See SI).
With all donors in hand, we set out to study the glycosylation properties of the building blocks
under pre-activation conditions (Figure 2A). To do so we first investigated the possible reactive
intermediates that can play a role during the glycosylation of these donors. Covalent species, such as
anomeric triflates are formed, which can undergo a SN2-like substitution or serve as a reservoir for more
reactive oxocarbenium ion type species that partake in substitution reactions with more SN1-character.
We started our investigation with the detection of the formation of reactive covalent species by the use
of variable-T NMR.[42] We first tested the per-O-benzyl donor 24. To this end a mixture of 24 and Ph2SO
(1.3 eq.) in CD2Cl2 was treated with Tf2O (1.3 eq.) at −80 °C (Figure 2B).[43] Directly after the addition,
NMR data (1H, HSQC, COSY) were recorded, to reveal the generation of a single new species. The
signals of the anomeric H and C atoms appeared at δ 4.7 ppm and δ 90.4 ppm for 1H and 13C respectively,
which is significantly upfield from signals corresponding to an anomeric triflate or oxosulfonium triflate
1
species (generally found at H: δ ~5-6.5 ppm and 13C: δ ~105-110 ppm).[42,44–47] Warming the sample
did not lead to any degradation of the initially formed product, and therefore it could be isolated. NMR
analysis (1H, 13C, HSQC, COSY, NOESY and HMBC) identified the formed species to be bicyclic
compound 26. Similarly, upon activation of the structurally simpler yersiniose donor 25, a
corresponding bicycle was formed. These bicycles are formed by nucleophilic attack of the oxygen atom
in the C7 benzyl ether on the activated C1 position.
5
This article is protected by copyright. All rights reserved.

10.1002/anie.202010280
Angewandte Chemie International Edition
A
O
O
O
Ph2SO/Tf2O
Nu
PO
O
PO
PO
PO
X
Nu
SPh
α / β
1,2-cis / 1,2-trans
B
glycosyl covalent reactive intermediates (SN2-like pathways)
undesired bicycle
BnO
7
BnO
O
O
BnO
NapO
1
1
BnO
O
BnO
SPh
BnO
ONap OBn
7
RO
24
26
Ph2SO, TTBP, -80 °C
then Tf2O
OBn
O
major products (>80%) at -80 o
after 5 min
C
O
pre-activation
conditions
O
O
O
O
O
O
O
NapO
O
NapO
O
Ph
S
BnO
SPh
BnO
Ph
4
27
O
O
OTf
major product even at -60 o
C
O
after 1 h, but slowly coverts to α-species
O
C
glycosyl cation reactive intermediates (SN1-like pathways)
undesired bicycle
formation
kcal mol-1
(II)
(I)
<0.5
1.0
2.0
3.0
4.0
5.0
6.0
7.0
8.0
>8.5
Figure 2. Mapping the relevant reactive intermediates by a combined experimental and computational approach. (A)
The reaction mechanism continuum operational during glycosylation reactions. Glycosylation reactions are best considered as
taking place at a continuum between two formal extremes of the mechanisms, including the SN1 and SN2 mechanism. (B) Upon
activation with Ph2SO/Tf2O of donor 24, the undesired fused bicycle 26 was formed. This side reaction makes these per-O-
benzylated caryophyllose donors unsuitable for efficient glycosylation reactions. (C) Conformational energy landscape (CEL)
maps of selected pyranosyl oxocarbenium ions in which the found local minima are indicated with their respective energy. All
energies are as computed at PCM(CH2Cl2)-B3LYP/6-311G(d,p) at T=213.15 K and expressed as solution-phase Gibbs free
energy.
Cyclization reactions on activated glycosyl donors have been reported before (for example from a C6-
OBn to form a 1,6-anhydrosugar)[48], but the rate with which the caryophyllose/yersiniose cyclization
takes place is striking. Apparently, the architecture in these systems is intrinsically geared for this
intramolecular nucleophilic cyclization. To prevent this cyclization, the C7-OH was tethered to the C4-
hydroxyl by the use of a carbonate protection group. Activation of the thus obtained donor 4, using the
conditions described above, resulted in the formation of several species, amongst which the anomeric
b-oxosulfonium triflate 27 species (1H: δ 5.8 ppm; 13C: δ 107.6 ppm) as the dominant reactive
6
This article is protected by copyright. All rights reserved.

10.1002/anie.202010280
Angewandte Chemie International Edition
intermediate ( 80% based on 1H-NMR). To support that this is indeed the oxosulfonium triflate, more
Ph2SO (+1.7 eq.) was added after the activation, which led to the increase of the oxosulfonium signals
and the disappearance of the signals corresponding to the anomeric triflate. Upon slow warming of the
mixture, this species gradually converted into the anomeric a-triflate and a-oxosulfonium triflate
species (See SI for all variable-T NMR results).
To study the reactive intermediates on the other side of the reaction mechanism continuum, we
studied the caryophyllose and yersiniose oxocarbenium ions by the use of DFT computations. We
recently developed a DFT protocol to compute the relative energy of all possible glycopyranosyl
oxocarbenium ion conformers, filling the complete conformational space these ions can occupy
generating conformational energy landscape (CEL) maps.[49–51] Based on these CEL maps, a prediction
can be made on the stereochemical outcome of reactions involving these ions. Figure 2C shows the CEL
maps of the two YerA oxocarbenium ions (these were selected as the substituted C6-chain of
caryophyllose would demand a significant increase in computing cost). The lowest energy structures
are shown next to the CEL maps with their corresponding energy (with the lowest energy depicted in
black/purple). The CEL map of oxocarbenium ion I (Figure 2C, left) shows that this species
preferentially takes up a 4H3 conformation. A second local minimum was found on the other side of the
CEL map, revealing the B1,4 conformer to be only slightly higher in energy (DGCH = 0.5 kcal/mol).
Cl
2
2
This latter conformer explains the rapid formation of the bicycles found upon activation of donors 24
and 25 as the C7 ether is perfectly positioned to attack the C1 position in this cation. The CEL map of
oxocarbenium ion II (Figure 2C, right) reveals a single minimal energy conformer. This 4H3 conformer
is preferentially attacked from the diastereotopic face that leads to a chair-like transition state, and thus
based on this analysis this cation is predicted to serve as a 1,2-cis-selective glycosylating species.
We next probed the donors 4 and 23 for their stereoselectivity in glycosylation reactions (Table 1).
To this end, we performed a matrix of glycosylation reactions with a set of model alcohol nucleophiles
of gradually decreasing nucleophilicity.[52,53] The trends observed relate to changes from an SN2-type
substitution reaction of the covalent intermediate for the most nucleophilic alcohols (EtOH and MFE),
to reactions involving more oxocarbenium character (for the poorest nucleophiles; TFE, HFIP and
TES-d). The outcome of the glycosylation reactions for both the caryophyllose and yersiniose donor
show clear trends with changing nucleophilicity of the used acceptors. The caryophyllose donor 4 and
yersiniose donor 23 behave very similarly and with decreasing nucleophilicity the 1,2-cis-selectivity
increases for both systems. Even with strong nucleophiles, somewhat more of the 1,2-cis-product is
formed, which may be explained by the direct displacement of the b-oxosulfonium triflate 27 species.
The increasing 1,2-cis selectivity can be accounted for by an increase of SN1 character in the
glycosylation reaction, as the weaker nucleophiles require a more electrophilic glycosylating agent. The
4
CEL maps revealed the H3 oxocarbenium ion conformers to be most stable and a stereoselective
addition to these ions can explain the formation of the a-products.
Table 1. Experimentally found stereoselectivities for model glycosylation reactions with ethanol, 2-fluoroethanol, 2,2-
difluoroethanol, 2,2,2-trifluoroethanol, 1,1,1,3,3,3-hexafluoro-2-propanol, triethylsilane-d, 1-buten-4-ol, 28 and 29. The
1
stereoselectivity of the reaction is expressed as 1,2-cis:1,2-trans and based on the H-NMR spectroscopy. Experimental
7
This article is protected by copyright. All rights reserved.

10.1002/anie.202010280
Angewandte Chemie International Edition
conditions: pre-activation based glycosylation conditions; Ph2SO (1.3 eq.), TTBP (2.5 eq.), DCM (0.05 M), then Tf2O (1.3
eq.), then nucleophile (2 eq.), –80 °C to –60 °C.
O
O
O
O
O
O
O
O
NapO
O
BnO
23
SPh
BnO
SPh
4
O
O
67:33
(94%)
50:50
(60%)
HO
83:17
(100%)
66:34
(76%)
F
HO
F
87:13
(63%)
80:20
(100%)
HO
F
>98:2
(76%)
>98:2
(77%)
HO
CF3
CF3
CF3
>98:2
(16%)
>98:2
(28%)
HO
donor
hydrolysis
>98:2
(54%)
D
Si
63:37
(97%)
59:41
(86%)
HO
77:23
(50%)
61:39
(63%)
>98:2
(54%)
>98:2
(74%)
To test nucleophiles relevant for the assembly of LOS IV-fragment 1, three acceptors (1-buten-
4-ol, 28 and 29) were probed. Acceptor 28 and 29 represent truncated versions of the caryophyllose
acceptor, and 1-buten-4-ol will be used to serve as a conjugation-ready linker moiety. Acceptor 28 is
protected with benzyl groups, known to be electronically neutral, while 29 is protected with an electron-
withdrawing carbonate group. The difference in reactivity between these two acceptors is mirrored in
the stereoselectivity of the glycosylation reactions with donors 4 and 23, with the more nucleophilic
dibenzylated alcohol 28 providing an a/b-mixture, while the less nucleophilic alcohol 29, exclusively
formed the 1,2-cis-products. These results indicate the need for an electron withdrawing protecting
8
This article is protected by copyright. All rights reserved.

10.1002/anie.202010280
Angewandte Chemie International Edition
group on the caryophyllose building block, when employed as an acceptor. The cyclic carbonate
spanning hydroxyl groups at C10 and C11 in the synthesized caryophyllose building blocks thus serves
this purpose.
After having established the glycosylation properties of the donors, we undertook the
construction of the target Car-Car-NAcFuc LOS-IV fragment 1 from building blocks 2, 3 and 4 (Figure
3A). Because of the high reactivity of 3-butene-1-ol, we sought to modify the reactivity of the reactive
intermediates formed upon activation of the donor glycoside. To this end we turned to the use of an
additive mediated glycosylation strategy. Various strategies have recently been developed to use
exogenous nucleophiles to generate reactive intermediates of which the reactivity can be tuned to match
the reactivity of the nucleophile that is to be glycosylated. Based on the work of Mukaiyama and co-
workers[54–57], and others[58], we have introduced triphenylphosphine oxide (TPPO)[59] to modulate the
reactivity of anomeric iodides, in which the anomeric α-iodide serves as a reservoir for the more reactive
β-iodide (or β-phosphoniumiodide), which is the actual glycosylating species (See SI for the complete
reactivity-selectivity mapping study with additives). Thus, caryophyllose 4 was pre-activated in the
usual manner, after which a mixture of tetrabutylammonium iodide and TPPO was added, together with
3-butene-1-ol. This led to the generation of spacer-equipped caryophyllose 30 in 60% yield and
excellent stereoselectivity (>98:2; cis:trans). Subsequent HCl-mediated deprotection of the 2-
methylnaphthyl protection group yielded caryophyllose acceptor 31 (61%).[60] Coupling of this
acceptor with donor 4 using pre-activation conditions afforded disaccharide 32 in 50% yield and with
complete 1,2-cis selectivity, in line with the results obtained with the model acceptor. Deprotection of
the 2-methylnaphthyl protection group of Car-Car 32 required more acid compared to the deprotection
of 30, because of the presence of more Lewis basic entities in the substrate, but furnished acceptor 33
in a similar yield (60%). Coupling of acceptor 33 to 4-azidofucose donor 3 under pre-activation
conditions provided 32 in 43% yield with the exclusive formation of the 1,2-cis product (See SI for the
complete reactivity-selectivity mapping study performed with this donor). A Staudinger reduction was
used to generated the free amine. Surprisingly, this transformation proceeded very sluggishly (reduction
of the 4-azido fucose monosaccharide proceeded readily with TPP in 79% yield, see SI) even with the
more reactive trimethyl phosphine. The crude product was directly coupled to the pyrrolidone 2, to
yield the completely protected Car-Car-NAcFuc LOS-IV fragment 35. Deprotection was done by
saponification of the carbonate protection groups and the benzyl ester on the pyrrolidone, followed by
debenzylation under Birch condition, to successfully yield the target structure 1 in 40% yield over the
two deprotection steps.
The structure and purity of compound 1, were confirmed by NMR and HRMS analysis. It was
observed that 1 exists as a mixture of atropisomers, in line with the behavior of related pyrrolidone-4-
aminofucose monosaccharides, prepared by Lowary and co-workers.[38,61] Figure 3B compares the 1H-
13C HSQC NMR spectra of the synthetic LOS-IV fragment 1 with the natural product, isolated by
Rombouts et al.[12] The blue signals originate from the synthesized compound, the red signals are from
the natural product, and all residues from the natural product that are absent in the synthetic fragment
are grey. From the overlay it is apparent that the spectra closely match, indicating that the assembled
fragment resembles the natural product very well.
9
This article is protected by copyright. All rights reserved.

10.1002/anie.202010280
Angewandte Chemie International Edition
1. Ph2SO, TTBP
O
A
ethyl maleimide
O
3. Ph2SO, TTBP,
then Tf2O
O
O
O
4
then Tf2O, -80 °C → -60 °C
then TBAI, TPPO
O
O
O
-80 °C → -60 °C
then 31
O
O
O
RO
HO
O
O
O
O
O
BnO
NapO
O
O
RO
60%
50%
BnO
4
SPh
BnO
O
BnO
[gram
scale]
O
O
O
O
2. HCl/HFIP
TES
4. HCl/HFIP
TES
O
O
R = ONap 30
R = OH 31
R = ONap 32
R = OH 33
O
O
O
MeO
O
60%
BnO
O
O
61%
O
5. Ph2SO, TTBP,
then Tf2O
3
N
BnO2C
HN
43%
-80 °C → -60 °C
then 33
O
6. i. PMe3
ii. HATU, TEA
O
N3
O
O
O
O
O
O
2
BnO
O
O
O
O
BnO
O
O
BnO
O
O
15%
over 2 steps
O
BnO
O
O
BnO
O
O
O
O
BnO
BnO
34
O
O
O
O
BnO
O
O
O
35
MeO
O
O
HO
O
O
O
O
O
Zc*
N
OH
O
HOOC
HN
smaller LOS-IV
fragments
40%
over 2 steps
7. i. LiOH
ii. Na, NH3
HO
O
O
HO
OH
MeO
O
O
OH
VIII
HO
HO
O
HO
O
OH
HO
OH
IX
HO
O
HO
OH
LOS-IV fragment 37
N
O
HO
HOOC
HN
O
4
LOS-IV fragment 36
O
HO
VII
OH
O
HO
2
3
O
HO
LOS-IV fragment 1
HO
HO
O
2
4
4
3
B
Figure 3. (A) Assembly of LOS-IV fragment 1. Reagents and conditions: (1) Ph2SO, TTBP, N-ethyl maleimide, then Tf2O,
then TBAI, TPPO, then 3-buten-1-ol, –80 °C to 40 °C (60%); (2) HCl/HFIP, TES, DCM (61%); (3) Ph2SO, TTBP, DCM,
then Tf2O, then 31, –80 °C to –60 °C (50%); (4) HCl/HFIP, TES, DCM (60%); (5) Ph2SO, TTBP, DCM, then Tf2O, then
33, –80 °C to –60 °C (43%); (6) i. trimethylphosphine, THF ii. 2, TEA, HATU, CH3CN (15% over 2 steps); (7) i. LiOH,
H2O, THF ii. Na, NH3, t-BuOH, THF (40% over 2 steps). Tf2O = trifluoromethanesulfonic anhydride, TTBP = 2,4,6-tri-t-
butylpyrimidine, TBAI = t-butylammonium iodide, TPPO = triphenylphosphine oxide, HFIP = hexafluorisopropanol, TES =
triethylsilane,
HATU
=
1-[bis(dimethylamino)methylene]-1H-1,2,3-triazolo[4,5-b]pyridinium
3-oxide
hexafluorophosphate, TEA = triethylamine. (B) 1H-13C HSQC NMR overlay of the acidic OS-IV fraction isolated by
Rombouts et al. (red = residues of the natural product present in the synthesized fragment 1, and grey = residues absent of the
natural product in the synthesized fragment 1), and synthesized compound 1 (blue). I to IX correspond to the nine
monosaccharides of the OS-IV. In the overlay most signals overlap. Only signals close to the linker on VII are slightly off,
because this area is different from the natural compound, which is linked to a xylose.
10
This article is protected by copyright. All rights reserved.

10.1002/anie.202010280
Angewandte Chemie International Edition
CONCLUSION
In conclusion, we have reported a systematic evaluation of the reactivity of tertiary-C sugar building
blocks, caryophyllose and yersiniose. An approach, consisting of a systematic series of glycosylation
reactions in combination with the detection and characterization of different reactive intermediates
using variable-T NMR and conformational energy landscape computations, were used to assess
reactivity-stereoselectivity relationships. We found for these 4-C-branched sugars that ether
functionalities in the appended side-chain readily attack the activated anomeric center of the
caryophyllose and yersiniose donors, leading to unproductive glycosylation reactions. This behavior has
been explained using the conformational preference of oxocarbenium ion intermediates that can form.
Prevention of this nucleophilic attack is a prerequisite to generate effective donor glycosides and could
be achieved by tethering of the C4 side-chain. We found that tethered Car and YerA donors can
efficiently form the desired 1,2-cis linkages, as long as weak nucleophiles are employed in the
glycosylation. In order to achieve 1,2-cis selectivity, the reactivity of the Car-acceptors was tuned using
electron-withdrawing protecting groups. The rationally designed building blocks enabled the first
effective and stereoselective assembly of a Car-Car-NAcFuc LOS-IV fragment, and related shorter
fragments. The approach taken here can be used to uncover the reactivity of rare bacterial saccharides.
The insight gathered will be a solid base to inform future syntheses of bacterial oligosaccharides and
glycoconjugates to fuel immunological- and biological research.
SUPPORTING INFORMATION
Experimental and computational procedures, reactivity-selectivity mapping of 4-azidofucose donor,
additive controlled model glycosylations, CEL maps of all studied ions, variable-temperature NMR,
NMR spectra of all new and selected compounds.
ACKNOWLEDGEMENTS
Quantum chemical calculations were performed at the SurfSARA HPC center in Amsterdam; access
was provided via NWO-Rekentijd grant 17569 (to T.H. and J.D.C.C.). This work was supported by an
ERC-CoG (726072) and NWO VICI (VI.C.182.020) grant awarded to J.D.C.C. We kindly
acknowledge Mark Somers for technical support.
AUTHOR CONTRIBUTION
T.H., G.A.M., J.D.C.C. designed the experiments. T.H., T.P.O., J.G.C.V., I.A.G., J.B., T.A.G., J.M.T.,
G.R., carried out the experimental work. T.H. performed all computational work. T.H., T.P.O.,
J.G.C.V., I.A.G., J.B., T.A.G., J.M.T., G.R., H.S.O., D.V.F., G.A.M., J.D.C.C. were involved in scientific
discussions and critically reviewed the manuscript.
COMPETING INTEREST
The authors declare no competing interests.
11
This article is protected by copyright. All rights reserved.

10.1002/anie.202010280
Angewandte Chemie International Edition
REFERENCES
[1] H. L. P. Tytgat, S. Lebeer, Microbiol. Mol. Biol. Rev. 2014, 78, 372–417.
[2] S. Herget, P. V. Toukach, R. Ranzinger, W. E. Hull, Y. A. Knirel, C.-W. von der Lieth, BMC
Structural Biology 2008, 8, 35.
[3] S. I. Elshahawi, K. A. Shaaban, M. K. Kharel, J. S. Thorson, Chem. Soc. Rev. 2015, 44, 7591–7697.
[4] C. J. Thibodeaux, C. E. Melançon, H. Liu, Angewandte Chemie International Edition 2008, 47,
9814–9859.
[5] C. J. Thibodeaux, C. E. Melançon, H. Liu, Nature 2007, 446, 1008–1016.
[6] D. B. Werz, R. Ranzinger, S. Herget, A. Adibekian, C.-W. von der Lieth, P. H. Seeberger, ACS
Chem. Biol. 2007, 2, 685–691.
[7] J. Zeng, G. Sun, W. Yao, Y. Zhu, R. Wang, L. Cai, K. Liu, Q. Zhang, X.-W. Liu, Q. Wan, Angew.
Chem. Int. Ed. Engl. 2017, 56, 5227–5231.
[8] S. Bera, B. Chatterjee, D. Mondal, RSC Adv. 2016, 6, 77212–77242.
[9] S. Wang, J. Sun, Q. Zhang, X. Cao, Y. Zhao, G. Tang, B. Yu, Angewandte Chemie International
Edition 2018, 57, 2884–2888.
[10] K. C. Nicolaou, J. Renaud, P. G. Nantermet, E. A. Couladouros, R. K. Guy, W. Wrasidlo, J. Am.
Chem. Soc. 1995, 117, 2409–2420.
[11] Y. Rombouts, A. Burguière, E. Maes, B. Coddeville, E. Elass, Y. Guérardel, L. Kremer, J Biol
Chem 2009, 284, 20975–20988.
[12] Y. Rombouts, E. Elass, C. Biot, E. Maes, B. Coddeville, A. Burguière, C. Tokarski, E. Buisine, X.
Trivelli, L. Kremer, Y. Guérardel, J. Am. Chem. Soc. 2010, 132, 16073–16084.
[13] Y. Rombouts, L. Alibaud, S. Carrère-Kremer, E. Maes, C. Tokarski, E. Elass, L. Kremer, Y.
Guérardel, J. Biol. Chem. 2011, 286, 33678–33688.
[14] T. L. Lowary, Acc. Chem. Res. 2016, 49, 1379–1388.
[15] B. Bai, C. Chu, T. L. Lowary, Isr. J. Chem. 2015, 55, 360–372.
[16] A. Nobre, S. Alarico, A. Maranha, V. Mendes, N. Empadinhas, Microbiology, 2014, 160, 1547–
1570.
[17] A. Burguière, P. G. Hitchen, L. G. Dover, L. Kremer, M. Ridell, D. C. Alexander, J. Liu, H. R.
Morris, D. E. Minnikin, A. Dell, G. S. Besra, J. Biol. Chem. 2005, 280, 42124–42133.
[18] H. Ren, L. G. Dover, S. T. Islam, D. C. Alexander, J. M. Chen, G. S. Besra, J. Liu, Molecular
Microbiology 2007, 63, 1345–1359.
[19] D. Sarkar, M. Sidhu, A. Singh, J. Chen, D. A. Lammas, A. M. van der Sar, G. S. Besra, A. Bhatt, J.
Bacteriol. 2011, 193, 2336–2340.
[20] S. S. Nigudkar, A. V. Demchenko, Chem. Sci. 2015, 6, 2687–2704.
[21] A. V. Demchenko, in Handbook of Chemical Glycosylation (Ed.: A.V. Demchenko), Wiley-VCH
Verlag GmbH & Co. KGaA, 2008, pp. 1–27.
[22] P. O. Adero, H. Amarasekara, P. Wen, L. Bohé, D. Crich, Chem. Rev. 2018, 118, 8242–8284.
[23] H. S. Hahm, M. Hurevich, P. H. Seeberger, Nature Communications 2016, 7, 1–8.
[24] R. A. Mensink, T. J. Boltje, Chemistry – A European Journal 2017, 23, 17637–17653.
[25] W.-L. Leng, H. Yao, J.-X. He, X.-W. Liu, Acc. Chem. Res. 2018, 51, 628–639.
[26] C.-C. Wang, J.-C. Lee, S.-Y. Luo, S. S. Kulkarni, Y.-W. Huang, C.-C. Lee, K.-L. Chang, S.-C.
Hung, Nature 2007, 446, 896–899.
[27] X. Zhu, R. R. Schmidt, Angewandte Chemie International Edition 2009, 48, 1900–1934.
[28] S. van der Vorm, T. Hansen, J. M. A. van Hengst, H. S. Overkleeft, G. A. van der Marel, J. D.
C. Codée, Chemical Society Reviews 2019, 48, 4688–4706.
[29] T. Hansen, H. Elferink, J. M. A. van Hengst, K. J. Houthuijs, W. A. Remmerswaal, A. Kromm, G.
Berden, S. van der Vorm, A. M. Rijs, H. S. Overkleeft, D. V. Filippov, F. P. J. T. Rutjes, G. A. van
der Marel, J. Martens, J. Oomens, J. D. C. Codée, T. J. Boltje, Nature Communications 2020, 11,
2664.
[30] H. Liu, T. Hansen, S.-Y. Zhou, G.-E. Wen, X.-X. Liu, Q.-J. Zhang, J. D. C. Codée, R. R. Schmidt,
J.-S. Sun, Org. Lett. 2019, 21, 8713–8717.
12
This article is protected by copyright. All rights reserved.

10.1002/anie.202010280
Angewandte Chemie International Edition
[31] B. Hagen, S. van der Vorm, T. Hansen, G. A. van der Marel, J. D. C. Codée, in Selective
Glycosylations: Synthetic Methods and Catalysts (Ed.: C.S. Bennett), Wiley-VCH Verlag GmbH
& Co. KGaA, 2017, pp. 1–28.
[32] L. Bohé, D. Crich, Carbohydrate Research 2015, 403, 48–59.
[33] Q. Zhang, A. Gimeno, D. Santana, Z. Wang, Y. Valdés-Balbin, L. M. Rodríguez-Noda, T.
Hansen, L. Kong, M. Shen, H. S. Overkleeft, V. Vérez-Bencomo, G. A. van der Marel, J.
Jiménez-Barbero, F. Chiodo, J. D. C. Codée, ACS Cent. Sci. 2019, 5, 1407–1416.
[34] D. Crich, Acc. Chem. Res. 2010, 43, 1144–1153.
[35] L. K. Mydock, A. V. Demchenko, Org. Biomol. Chem. 2010, 8, 497–510.
[36] P. H. Seeberger, D. B. Werz, Nature 2007, 446, 1046–1051.
[37] M. Marianski, E. Mucha, K. Greis, S. Moon, A. Pardo, C. Kirschbaum, D. A. Thomas, G. Meijer,
G. von Helden, K. Gilmore, P. H. Seeberger, K. Pagel, Angewandte Chemie International Edition
2020, 59, 6166–6171.
[38] L. Wang, M. Dong, T. L. Lowary, J. Org. Chem. 2015, 80, 2767–2780.
[39] J. Prandi, Carbohydrate Research 2001, 332, 241–247.
[40] J. Prandi, G. Couturier, Tetrahedron Letters 2000, 41, 49–52.
[41] J. Guiard, Y. Rahali, J.-P. Praly, European Journal of Organic Chemistry 2014, 2014, 4461–4466.
[42] T. G. Frihed, M. Bols, C. M. Pedersen, Chem. Rev. 2015, 115, 4963–5013.
[43] J. D. C. Codée, R. E. J. N. Litjens, R. den Heeten, H. S. Overkleeft, J. H. van Boom, G. A. van
der Marel, Org. Lett. 2003, 5, 1519–1522.
[44] B. Hagen, S. Ali, H. S. Overkleeft, G. A. van der Marel, J. D. C. Codée, J. Org. Chem. 2017, 82,
848–868.
[45] S. van der Vorm, H. S. Overkleeft, G. A. van der Marel, J. D. C. Codée, J. Org. Chem. 2017, 82,
4793–4811.
[46] D. Crich, W. Cai, J. Org. Chem. 1999, 64, 4926–4930.
[47] D. Crich, S. Sun, Journal of the American Chemical Society 1997, 119, 11217–11223.
[48] H. M. Christensen, S. Oscarson, H. H. Jensen, Carbohydrate Research 2015, 408, 51–95.
[49] T. Hansen, L. Lebedel, W. A. Remmerswaal, S. van der Vorm, D. P. A. Wander, M. Somers, H.
S. Overkleeft, D. V. Filippov, J. Désiré, A. Mingot, Y. Bleriot, G. A. van der Marel, S.
Thibaudeau, J. D. C. Codée, ACS Cent. Sci. 2019, 5, 781–788.
[50] S. van der Vorm, T. Hansen, E. R. van Rijssel, R. Dekkers, J. M. Madern, H. S. Overkleeft, D. V.
Filippov, G. A. van der Marel, J. D. C. Codée, Chemistry – A European Journal 2019, 25, 7149–
7157.
[51] J. M. Madern, T. Hansen, E. R. van Rijssel, H. A. V. Kistemaker, S. van der Vorm, H. S.
Overkleeft, G. A. van der Marel, D. V. Filippov, J. D. C. Codée, J. Org. Chem. 2019, 84, 1218–
1227.
[52] S. van der Vorm, T. Hansen, H. S. Overkleeft, G. A. van der Marel, J. D. C. Codée, Chem. Sci.
2017, 8, 1867–1875.
[53] M. G. Beaver, K. A. Woerpel, J. Org. Chem. 2010, 75, 1107–1118.
[54] T. Mukaiyama, Y. Kobashi, T. Shintou, Chem. Lett. 2003, 32, 900–901.
[55] T. Mukaiyama, Y. Kobashi, Chem. Lett. 2003, 33, 10–11.
[56] Y. Kobashi, T. Mukaiyama, BCSJ 2005, 78, 910–916.
[57] Y. Kobashi, T. Mukaiyama, Chem. Lett. 2004, 33, 874–875.
[58] A.-H. A. Chu, S. H. Nguyen, J. A. Sisel, A. Minciunescu, C. S. Bennett, Org. Lett. 2013, 15,
2566–2569.
[59] L. Wang, H. S. Overkleeft, G. A. van der Marel, J. D. C. Codée, J. Am. Chem. Soc. 2018, 140,
4632–4638.
[60] A. G. Volbeda, H. A. V. Kistemaker, H. S. Overkleeft, G. A. van der Marel, D. V. Filippov, J. D.
C. Codée, J. Org. Chem. 2015, 80, 8796–8806.
[61] S. H. Reisberg, Y. Gao, A. S. Walker, E. J. N. Helfrich, J. Clardy, P. S. Baran, Science 2020, 367,
458–463.
13
This article is protected by copyright. All rights reserved.

10.1002/anie.202010280
Angewandte Chemie International Edition
TOC
O
variable-T NMR
OTf
O
Ph
S
Ph
assembly
LOS-IV
reactivity?
selectivity?
fragments
CEL maps
O
OH
OH
0 °
OH
OH
90 °
O
O
OH OH
OH
OH
ar op
OH
yllo
OH
OH
OH
HO
O
c
c
a
r
y
y
o
h
p
h
y
sell
(
ose
Car)
270 °
180 °
= 1,2-cis bond
quaternary
stereocenter
=
glycosylation reactions
F
OH
O
LG
F
OH
F
Sweet dreams are made of this! A joint experimental and computational approach was used to assess
reactivity-stereoselectivity relationships of the bacterial sugar caryophyllose, rare C12-
a
monosaccharide, containing a characteristic tetrasubstituted stereocenter. The gained fundamental
insights, guided the rational design of building blocks for the effective and stereoselective assembly of
mycobacterial lipooligosaccharide fragments.
ORCID
Thomas Hansen: 0000-0002-6291-1569
Tim P. Ofman: 0000-0002-1687-2191
Joey G. C. Vlaming: 0000-0002-6493-5783
Ivan A. Gagarinov: 0000-0003-1742-3870
Herman S. Overkleeft: 0000-0001-6992-4760
Dmitri V. Filippov: 0000-0002-6978-7425
Gijsbert A. van der Marel: 0000-0001-8917-3458
Jeroen Codée: 0000-0003-3531-2138
14
This article is protected by copyright. All rights reserved.