Photochemical Surface Functionalization:
Synthesis, Nanochemistry and
Glycobiological Studies
Lingquan Deng
邓灵泉
Doctoral Thesis
Stockholm 2011
Akademisk avhandling som med tillstånd av Kungl Tekniska Högskolan i Stockholm
framlägges till offentlig granskning för avläggande av doktorsexamen i kemi med
inriktning mot organisk kemi, fredagen den 21 October, kl 13:00 i sal F3, KTH,
Lindstedtsvägen 26, Stockholm. Avhandlingen försvaras på engelska. Opponent är Dr.
Javier Rojo, Instituto de Investigaciones Químicas, CSIC, Seville, Spain.
ISBN 978-91-7501-072-4
ISSN 1654-1081
TRITA-CHE-Report 2011:47
© Lingquan Deng, 2011
E-Print, Stockholm
献给我的父亲、母亲和姐姐
To my family
Lingquan Deng, 2011: “Photochemical Surface Functionalization: Synthesis,
Nanochemistry and Glycobiological Studies”, Chemistry/Organic Chemistry,
KTH Chemical Science and Engineering, Royal Institute of Technology,
SE-100 44 Stockholm, Sweden.
Abstract
This thesis mainly deals with the development of photochemical approaches to
immobilize carbohydrates on surfaces for glycobiological studies. These
approaches have been incorporated into a number of state-of-the-art
nanobio-platforms, including carbohydrate microarrays, surface plasmon
resonance (SPR), quartz crystal microbalance (QCM), atomic force
microscopy (AFM), and glyconanomaterials. All the surfaces have displayed
good binding capabilities and selectivities after functionalization with
carbohydrates, and a range of important data have been obtained concerning
surface characteristics and carbohydrate-protein interactions, based on the
platforms established. Besides, a variety of non-carbohydrate and
carbohydrate-based molecules have been synthesized, during which process
the mutarotation of 1-glycosyl thiols and the stereocontrol in 1-S-glycosylation
reactions have been thoroughly studied.
Keywords: Carbohydrates; Glycobiology; Carbohydrate-protein interactions;
Perfluorophenylazide (PFPA); Photochemistry; Carbohydrate surface
immobilization; Carbohydrate microarrays; Surface plasmon resonance (SPR);
Quartz crystal microbalance (QCM); Atomic force microscopy (AFM);
Glyconanomaterials; Concanavalin A (Con A); Cyanovirin-N (CV-N);
1-Glycosyl thiols; Mutarotation; 1-S-glycosylation; Stereocontrol.
Abbreviations
Ac
AFM
aq.
Bn
BS-II
BSA
Bz
CBD
CD
Con A
CuAAC
CV-N
Cys
DCC
DCM
DIPEA
DLS
DMAP
DMF
DMPU
DMSO
DNA
EDAC
EF
EG
ELISA
Et
Gal
GBP
Glc
GlcNAc
GNP
HIV
HSQC
ISC
ITC
Man
Me
MS
MW
NHS
Acetyl
Atomic force microscopy
Aqueous
Benzyl
Bandeiraea simplicifolia II
Bovine serum albumin
Benzoyl
Carbohydrate binding domain
Circular dichroism
Concanavalin A
Cu-catalyzed azide alkyne cycloaddition
Cyanovirin-N
Cysteine
N,N'-Dicyclohexylcarbodiimide
Dichloromethane
N,N-Diisopropylethylamine
Dynamic light scattering
4-Dimethylaminopyridine
Dimethylformamide
N,N'-Dimethylpropyleneurea
Dimethyl sulfoxide
Deoxyribonucleic acid
1-Ethyl-3-(3-dimethylaminopropyl) carbodiimide
Enhancement factor
Ethylene glycol
Enzyme-linked immunosorbent assay
Ethyl
Galactose
Glycan-binding protein
Glucose
N-Acetylglucosamine
Glyconanoparticle
Human immunodeficiency virus
Heteronuclear single quantum correlation
Intersystem crossing
Isothermal titration calorimetry
Mannose
Methyl
Mass spectrometry
Molecular weight
N-Hydroxysuccinimide
NIS
NMR
NOESY
OEG
PAAm
PDT
PEG
PEO
PEOX
PFPA
PNA
PP
PS
PSA
QCM
QSAR
quant.
RCA-I
RNA
RT
SEM
SFM
SPM
SPR
TBA
TEA
TEG
TEM
TfOH
THF
TMSOTf
Ts
UV
UV-vis
WGA
w/w
N-Iodosuccinimide
Nuclear magnetic resonance
Nuclear Overhauser effect spectroscopy
Oligoethylene glycol
Polyacrylamide
Photodynamic therapy
Poly(ethylene glycol)
Poly(ethylene oxide)
Poly(2-ethyl-2-oxazoline)
Perfluorophenylazide
Peanut agglutinin
Polypropylene
Polystyrene
Pisum sativum
Quartz crystal microbalance
Quantitative structure-activity relationship
Quantitative
Ricinus communis Agglutinin I
Ribonucleic acid
Room temperature
Standard error of the mean
Scanning force microscopy
Scanning probe microscopy
Surface plasmon resonance
Tetrabutylammonium
Triethylamine
Triethylene glycol
Transmission electron microscopy
Trifluoromethanesulfonic acid
Tetrahydrofuran
Trimethylsilyl trifluoromethanesulfonate
4-Toluenesulfonyl
Ultraviolet
Ultraviolet-visible
Wheat-germ agglutinin
Weight to weight
List of Publications
This thesis is based on the following papers, referred to in the text by their
Roman numerals I-VIII:
I. pH-Dependent Mutarotation of 1-Thio-Aldoses in Water.
Unexpected Behavior of (2S)-D-Aldopyranoses
Rémi Caraballo*, Lingquan Deng*, Luis Amorim, Tore Brinck, and Olof
Ramström
*
These authors contributed equally to this work
J. Org. Chem. 2010, 75, 6115-6121.
II. Stereocontrolled 1-S-Glycosylation and Comparative Binding
Studies of Photoprobe-Thiosaccharide Conjugates with Their
O-Linked Analogs
Lingquan Deng, Xin Wang, Suji Uppalapati, Oscar Norberg, Hai Dong,
Adrien Joliton, Mingdi Yan, and Olof Ramström
Submitted for publication
III. Stereoselective Synthesis of Light-Activatable
Perfluorophenylazide-Conjugated Carbohydrates for Glycoarray
Fabrication and Evaluation of Structural Effects on Protein
Binding by SPR Imaging
Lingquan Deng, Oscar Norberg, Suji Uppalapati, Mingdi Yan, and Olof
Ramström
Org. Biomol. Chem. 2011, 9, 3188-3198.
IV. Perfluorophenyl Azide Immobilization Chemistry for Single
Molecule Force Spectroscopy of the Concanavalin A/Mannose
Interaction
Carolin Madwar, William Chu Kwan, Lingquan Deng, Olof Ramström,
Rolf Schmidt, Shan Zou, and Louis Cuccia
Langmuir 2010, 26, 16677-16680.
V. Photogenerated Carbohydrate Microarrays to Study
Carbohydrate-Protein Interactions Using Surface Plasmon
Resonance Imaging
Anuradha Tyagi, Xin Wang, Lingquan Deng, Olof Ramström, and
Mingdi Yan
Biosens. Bioelectron. 2010, 26, 344-350.
VI. Multivalent Glyconanoparticles with Enhanced Affinity to the
Anti-Viral Lectin Cyanovirin-N
Xin Wang*, Elena Matei*, Lingquan Deng*, Olof Ramström, Angela
Gronenborn, and Mingdi Yan
*
These authors contributed equally to this work
Chem. Comm. 2011, 47, 8620-8622.
VII. Photo-Click Immobilization of Carbohydrates on Polymeric
Surfaces - A Quick Method to Functionalize Surfaces for
Biomolecular Recognition Studies
Oscar Norberg, Lingquan Deng, Mingdi Yan, and Olof Ramström
Bioconjugate Chem. 2009, 20, 2364–2370.
VIII. Photo-Click Immobilization on Quartz Crystal Microbalance
Sensors for Selective Carbohydrate-Protein Interaction Analyses
Oscar Norberg, Lingquan Deng, Teodor Aastrup, Mingdi Yan, and Olof
Ramström
Anal. Chem. 2011, 83, 1000-1007.
Papers not included in this thesis:
IX. Control of the Ambident Reactivity of the Nitrite Ion
Hai Dong, Martin Rahm, Lingquan Deng, Tore Brinck, and Olof
Ramström
Submitted for publication
X. Synthesis of Glyconanomaterials via Photo-Initiated Coupling
Chemistry
Xin Wang, Oscar Norberg, Lingquan Deng, Olof Ramström and Mingdi
Yan
ACS Symp. Ser. 2011, in press.
XI. A Novel Carbohydrate-Anion Recognition System
Hai Dong, Lingquan Deng, and Olof Ramström
Preliminary manuscript
Table of Contents
Abstract
Abbreviations
List of publications
1. Introduction ..................................................................................... 1 1.1. 1.2. Carbohydrates and glycobiology – A general background..................... 1 Carbohydrate chemistry......................................................................... 2 1.2.1. Introduction .................................................................................................. 2 1.2.2. Mutarotation, conformation and anomeric effect ......................................... 3 1.2.3. Glycomimetics ............................................................................................. 5 1.3. Glycobiology .......................................................................................... 6 1.3.1. Biological roles of glycans ........................................................................... 6 1.3.2. Glycans in medicine and biotechnology ...................................................... 6 1.3.3. Glycan-mediated recognition ....................................................................... 7 1.3.4. Multivalency in carbohydrate-protein interactions ....................................... 7 1.4. 1.5. Methods and tools in glycobiological studies ......................................... 7 Carbohydrate immobilization ............................................................... 10 1.5.1. Survey of methods .................................................................................... 10 1.5.2. Photochemical immobilization methods .................................................... 13 1.5.3. Three photochemical carbohydrate immobilization approaches ............... 15 1.6. 2. Aims of this thesis ................................................................................ 17 Syntheses ..................................................................................... 19 2.1. Synthesis of non-carbohydrate molecules ........................................... 19 2.1.1. Synthesis of linker molecules .................................................................... 19 2.1.2. Synthesis of PFPA derivatives .................................................................. 21 2.2. Synthesis of non-PFPA derivatized glycosides .................................... 23 2.2.1. Synthesis of 1-thioglycosides .................................................................... 23 2.2.2. Synthesis of oligosaccharides ................................................................... 25 2.2.3. Synthesis of carbohydrate-azides ............................................................. 27 2.3. Synthesis of PFPA-conjugated carbohydrates .................................... 28 2.3.1. Synthesis of O-linked PFPA-carbohydrates .............................................. 28 2.3.2. Synthesis of S-linked PFPA-carbohydrates .............................................. 29 2.3.3. Synthesis of oligosaccharide-PFPA conjugates ........................................ 31 3. Mutarotation of 1-thioaldoses and stereocontrol in 1thioglycosylation (Papers I and II) ........................................................... 35 3.1. Mutarotation of 1-thioaldoses in aqueous media (Paper I) .................. 35 3.1.1. Introduction ................................................................................................ 35 3.1.2. pD-dependence of the mutarotation .......................................................... 35 3.1.3. Reversibility of the mutarotation ................................................................ 38 3.1.4. Proposed mechanisms of the mutarotation of 1-thioaldoses .................... 39 3.1.5. Conclusions ............................................................................................... 39 3.2. Stereocontrol in 1-thioglycosylation (Paper II) ..................................... 40 3.2.1. Introduction ................................................................................................ 40 3.2.2. Occurrence and identification of the anomerization .................................. 40 3.2.3. Controlling the stereochemistry in the synthesis ....................................... 41 3.2.4. Conclusions ............................................................................................... 44 4. Development of three carbohydrate immobilization approaches
and their applications (Papers II-VIII) ...................................................... 45 4.1. Photo-induced glycosurface fabrication approach I .................................. 46 4.1.2. SPR imaging analysis................................................................................ 47 4.1.3. Affinity analyses using SPR and ITC ......................................................... 49 4.1.4. AFM single molecular force analysis ......................................................... 51 4.1.5. Conclusions ............................................................................................... 55 4.2. Direct photoconjugation (Papers V and VI).......................................... 56 4.2.1. Photo-induced glycosurface fabrication approach II ................................. 56 4.2.2. Carbohydrate microarray studies .............................................................. 57 4.2.3. Glyconanoparticle platform for probing CV-N ............................................ 60 4.2.4. Conclusions ............................................................................................... 65 4.3. 5. Carbohydrate-conjugated photoprobes (Papers II-IV) ......................... 46 4.1.1. Photo-click carbohydrate immobilization (Papers VII and VIII) ............ 66 4.3.1. Photo-induced glycosurface fabrication approach III ................................ 66 4.3.2. Protein binding and surface specificity studies by QCM ........................... 67 4.3.3. Conclusions ............................................................................................... 71 Concluding remarks and outlooks ................................................ 73 Acknowledgements
Appendix
References
1. Introduction
1.1. Carbohydrates and glycobiology – A general background
Life is an extraordinarily complex phenomenon, and the understanding of
living things has been an everlasting quest in human history. In the 1950s to
1970s, the molecular biology revolution initiated the interrogation of living
systems at the molecular level. The central dogma in biology, that biological
information flows from DNA to RNA and then to proteins, has been
established since then. It is thus tempting to assume that the makeup and
functions of cells, tissues, organs, physiological systems and intact organisms
can be explained merely by these molecules. However, in fact, a few more
classes of molecules, including lipids and carbohydrates, are required to fulfill
the task and complete the puzzle.
Lipids and carbohydrates take part in some of the major posttranslational
modifications of proteins, and thus help to generate a vast variety of biological
complexes inherent in the development, growth and function of intact
organisms. Due to the modifications, these entities greatly diversify the effects
of the relatively small number of genes in a typical genome. Details and
functions of lipids are however not discussed in this thesis. Instead, it will
focus on carbohydrates and carbohydrate-containing biomolecules, which are
ubiquitous in living systems and play essential roles in various biological
processes, including modulating or mediating various fundamental processes
in cell-cell-, cell-matrix-, and cell-molecule interactions, or in the interactions
between different organisms, and acting as regulatory switches within the
nucleus and cytoplasm.1-3
The chemistry of carbohydrates has drawn considerable interest since the late
19th century. However, carbohydrates had long been considered mainly as an
energy source or as structural materials, and were believed to lack other
biological functions. Nevertheless, with the development of new technologies
for elucidating the structures and functions of carbohydrates, the word
“glycobiology” was first coined in the late 1980s,4 to recognize the new
frontier of molecular sciences, which combined the disciplines of carbohydrate
chemistry and biochemistry with a modern understanding of the cell biology of
carbohydrates, especially that of their conjugates with proteins and lipids.
Glycobiology can be defined as the branch of science concerned with the role
of sugars in biological systems.3 In the broadest sense, it is the study of the
structure, biosynthesis and biology of saccharides that are widely distributed in
nature. This research field has grown very rapidly in recent years. It involves
studies in the chemistry of carbohydrates, the enzymology of carbohydrate
1
formation and degradation, the recognition of carbohydrates by specific
proteins, the roles of carbohydrates in complex biological systems and their
analysis or manipulation by various techniques.3
1.2. Carbohydrate chemistry
1.2.1. Introduction
In nature, carbohydrates are produced in the process of photosynthesis, which
converts the light energy of the sun to chemical energy by combining carbon
dioxide and water to form carbohydrates and molecular oxygen (Scheme 1.1).
Scheme 1.1 Process of photosynthesis.
Structurally, the smallest carbohydrate would have three carbons, with an
aldehyde or keto group and two hydroxyl groups. Carbohydrates with three or
more carbons are successively called trioses, tetroses, pentoses, hexoses,
heptoses, octoses, and so on. Carbohydrates exist as aldoses, which are
polyhydroxyaldehydes, and ketoses, being polyhydroxyketones. Each aldose
and ketose further exists as two enantiomers, and the prefixes “D” and “L”
have been used to distinguish them. Thus a D- or L-carbohydrate is defined by
the fact that the hydroxyl group on the asymmetric carbon atom furthest from
the most oxidized carbon atom (aldehydo, keto, or carboxyl group) is
positioned right or left in the carbohydrate’s Fischer projection (Figure 1.1,
C5-OH), respectively.
Figure 1.1 Open-chain and ring forms of mannose.
A monosaccharide is the simplest form of carbohydrates that cannot be further
hydrolyzed. It can exist in either open-chain or ring form. Oligosaccharides are
linear or branched chains of monosaccharides attached to one another by
glycosidic linkages, and polysaccharides are typically referred to structures
composed of repeating oligosaccharide motifs.
A chiral center, termed the anomeric center, at C-1 for aldo carbohydrates or at
C-2 for keto carbohydrates is generated in the formation of the carbohydrate
ring. From this center, the monosaccharides are attached to other residues by
glycosidic linkage, and typically the residues are hydroxy-, amino-, or
2
thio-groups, or interlinked with the carbohydrate ring by a carbon atom,
forming O-, N-, S- or C-glycosides, respectively. When the attached residue is
a non-carbohydrate structure, it is usually termed as aglycone, and the
carbohydrate counterpart as glycone. Depending on the stereochemical
relationship between the hydroxyl groups attached to the anomeric carbon and
to the highest numbered chiral carbon in the carbohydrate ring, α or β linkages
are designated. More specifically, when the groups are cis (bound to the center
carbons in the same direction), the anomer is defined as α anomer, while if
they are trans (bound to the center carbons in opposite directions), the β
configuration is designated. The anomers are diastereomers.
It is important to realize that the two anomers confer very different structural
properties and biological functions upon saccharide sequences that are
otherwise identical in composition. A good illustration for this in nature is the
marked differences between starch and cellulose. Though both are
homopolymers of glucose, the former is mainly α1-4 linked and the latter β1-4
linked throughout the molecule. Even for monosaccharide anomers, they can
possess distinct biological activities.5 In this case, they are normally of similar
physicochemical properties which can hamper the separation of the two
stereoisomers.
1.2.2. Mutarotation, conformation and anomeric effect
The anomeric center of all reducing saccharides (those with the reducing
ability of the aldehyde or ketone in its terminal monosaccharide component)
can undergo an interconversion of isomers, accompanied by a change in
optical rotation. This process is known as mutarotation. It can be catalyzed by
dilute acid or base. In the process, the monosaccharide ring opens up and then
recloses to form a ring with the other anomeric configuration, or form other
isomers with different ring sizes (Scheme 1.2). Aldohexoses form
six-membered rings through a C-1—O—C-5 ring closure, and five-membered
rings via C-1—O—C-4 ring closure, producing pyranoses and furanoses,
respectively.
3
Scheme 1.2 Mutarotation of D-mannose.
Pyranoses mainly exist in a chair conformation, but can also exist in for
example half-chair, boat and skew conformations. The chair form is the most
stable and only the skew conformation has a comparable energy minimum,
although still about 20 kJ higher than the chair form. Major conformations for
furanoses are the envelope and twist forms. The chair conformation of
pyranoses further exists in two isomeric forms, as 1C4 and 4C1. The letter C
stands for “chair” and the numbers indicate the relative locations of the C-1
and C-4 carbon atoms, either above or below the reference plane of the chair,
which is made up by the other four atoms in the carbohydrate ring
(cf. Figure 1.1).
In a molecule with a chair conformation, the equatorially positioned
substituent should normally be energetically favored for steric reasons,
compared to its axial form. However, the anomerically bound groups
(aglycones) in carbohydrates do not completely follow this rule. Thus, the term
“anomeric effect” was coined to describe the increased preference for an
electronegative substituent, at the anomeric carbon in a carbohydrate, to be in
an axial rather than equatorial orientation.6,7 The intramolecular electrostatic
interactions of the two dipoles next to the anomeric center have been used to
explain the anomeric effect (Figure 1.2). Thus the anomeric configurations
arise as a result of the partial neutralization of the two dipoles. However, it is
noteworthy that a modern understanding of the anomeric effect includes
electrostatic effects, molecular orbital interactions, as well as solvent effects,
which all dictate the conformational preferences.
4
Figure 1.2 Dipole interactions and anomeric effect.
1.2.3. Glycomimetics
Due to the difficulties of purifying necessary carbohydrate ligands from
natural sources or synthesizing oligosaccharides that occur naturally, the
strategy of designing carbohydrate ligands as simplified carbohydrate
derivatives, which are often called glycomimetics, emerged and has been
constantly developed. Many advantages are associated with the synthesis of
glycomimetics, including easier access than the natural analogs, facile
modification of the structural properties, and that the glycomimetics can be
designed as non-biodegradable compounds, as well as low molecular weight
derivatives (drug candidates). It is not guaranteed that glycomimetics will
always work as ligands for the target receptors, but it is possible that they
display even higher receptor affinities than their naturally-occurring analogs.
When the structure of the receptor under investigation is known, it
dramatically helps the rational design of glycomimetics. If the receptor
structures, for example those of lectins, are unknown, then systematic
derivatization of the ligands is the way to go for identifying the essential
structural features for potent receptor binding. With the help of synthetic
glycomimetics, carbohydrate-protein interactions can be studied with regard to
their structural and functional details. Furthermore, in a therapeutic context,
carbohydrate-protein interactions can also be modulated or inhibited by
glycomimetics.
Two means of producing glycomimetics, among others, are worth mentioning.
The first concerns the strategies to block the anomeric reactivities of the
glycosidic linkage, which are based on the substitution of the hemiacetal
function with a more stable one. These enable the generation of glycomimetics
that are more resistant to glycosidases and to acidic cleavage, providing
promising potential therapeutics. To achieve such a goal, one of the two
hemiacetal oxygen atoms, or both, should be replaced. Figure 1.3 shows some
possible modifications around the anomeric center to generate such
glycomimetics.
5
Figure 1.3 Glycomimetics with modifications around the anomeric center.
The other means is the preparation of multivalent glycoconjugates which
mimic the polyantennary complex-type carbohydrates found on
naturally-occurring molecular constructs or biological surfaces. The design of
multivalent neoglycoconjugates with respect to the multivalency principle is of
fundamental importance in studying carbohydrate-protein interactions.8
Various scaffolds, including dendrimers, polymers, liposomes and different
nano-materials, have been developed to help achieving this goal.
1.3. Glycobiology
1.3.1. Biological roles of glycans
Glycan is a term that refers to any form of carbohydrates, from mono- to
polysaccharides, and either free or covalently linked to another molecule.3 Due
to the ubiquitous and complex nature of glycans, their biological roles are
substantially varied. Glycans thus play various subtle to crucial roles for the
development, growth, function and survival of an organism. Furthermore, the
diverse roles can be categorized into two parts: one is about the structural and
modulatory functions, which involve the glycans themselves or their
modulation toward the molecule to which they are attached; the other mainly
concerns specific recognition of glycans by glycan-binding proteins.
1.3.2. Glycans in medicine and biotechnology
Numerous biotherapeutic agents contain glycan components. These range from
natural products to rationally designed glycomimetics and recombinant
glycoconjugates, for example glycoproteins. The glycan components in these
agents play important roles for their biological activity and therapeutic
efficacy. Heparin, a sulfated glycosaminoglycan, and its derivatives are among
the most commonly used drugs all over the world.9,10 Glycoproteins are by
now one of the major products in the biotechnology industry, with sales in tens
of billions of dollars annually.11 Moreover, a number of human disease states
are characterized by changes in the glycosylation processes that could be of
potential diagnostic and/or therapeutic significance.3 As a result, carbohydrate
chemistry and glycobiology have been gaining increasing attention and interest
in modern biotechnology and pharmaceutical industries.
6
1.3.3. Glycan-mediated recognition
As mentioned above, glycans play two major classes of biological roles.
Besides their structural and modulatory functions, however, many of their
more specific biological roles are revealed via their recognition by
glycan-binding proteins (GBPs). Except for glycan-specific enzymes and
antibodies, the GBPs can be classified into two major groups: lectins and
glycosaminoglycan-binding proteins.3 Lectins are glycan-binding proteins that
are highly specific for their carbohydrate moieties. They bind to glycans
specifically by fitting them into shallow but relatively well-defined binding
pockets. In contrast, glycosaminoglycan-binding proteins interact with sulfated
glycosaminoglycans by cation-anion attractive forces. Though this type of
binding is based on ionic interactions, it can be specific as well. These groups
of proteins can be either endogenous to the organism that synthesizes their
cognate glycans or exogenous. In the former case, studies focus on the
development and function of a complex multicellular organism, while in the
latter, the interactions between microbial proteins that bind to specific glycans
on host cells are mostly studied.
1.3.4. Multivalency in carbohydrate-protein interactions
It has been generally recognized that single-site binding affinities in many
lectins are relatively low, usually with Kd values in the micromolar range. For
those lectins with low affinities, multivalent interactions between multiple
carbohydrate-binding domains (CBDs) and multiple glycans are usually
required to generate high-affinity binding, which is more relevant in vivo.
Thus, multivalent interactions can be defined as specific simultaneous
association of multiple ligands present on a molecular construct or biological
surface binding in a cooperative way to multiple receptors expressed on a
complementary entity.8
1.4. Methods and tools in glycobiological studies
The development of methods for obtaining various glycan ligands and tools for
studying carbohydrate-protein interactions has greatly facilitated modern
advances in glycobiology. Two major synthetic strategies have been developed
for efficient preparation of oligosaccharides and glycoconjugates: enzymatic
synthesis and chemical synthesis.12 Programmable one-pot synthesis and
automated solid-phase assembly of oligosaccharides are excellent examples in
this pursuit.13,14 Tools for studying carbohydrate-protein interactions have also
evolved rapidly in the past few decades. Techniques like X-ray
crystallography, UV-vis, fluorescence spectroscopy, circular dichroism (CD)
and NMR have greatly aided the insight into the molecular basis of structural
features and functional roles of various glycans. More recently, a number of
7
other techniques and tools have been incorporated in glycobiological studies,
such as microarrays, biosensors, nano-based technologies and bioinformatical
tools, to name just a few. More specifically, these techniques and tools include
isothermal titration calorimetry (ITC),15 mass spectrometry (MS),16
carbohydrate microarrays,17 surface plasmon resonance (SPR),18,19 quartz
crystal microbalance (QCM),20,21 scanning probe microscopy (SPM),22
glyconanotechnologies,23 computational simulations.24 Among these, two
major methods are especially advantageous: solution-based ITC and solidsupported binding techniques, the combination of which can result in
comprehensive elucidations of binding performances. ITC can directly provide
thermodynamic data of molecular interactions, thus reliably conveying
important binding information and supporting binding implications. On the
other hand, techniques and tools involving solid supports, such as microarrays,
biosensor instrumentations and nanoplatform-based techniques, have drawn
increasing attention and interest in glycobiology.25
ITC is a physical technique used for determining thermodynamic parameters
of interactions in solution. The binding of glycans to proteins can be measured
as a change in free energy by the ITC technique.15 It directly measures the
binding affinity (Ka), enthalpy changes (ΔH), and binding stoichiometry (n) of
the interactions between binding partners in solution, while the Gibbs energy
changes (ΔG) and the entropy changes (ΔS) can be derived from the
thermodynamic relationships: ΔG = -RTInKa = ΔH-TΔS, where R is the gas
constant and T is the absolute temperature. In operation, increasing amounts of
glycan are injected to a fixed concentration of protein solution in a
microcalorimetry cell by a syringe. The glycan solution is added at intervals
and the heat taken up or evolved from the binding is measured. The heat flow
spikes are then integrated with respect to time, providing the total heat
exchanged per injection. This heat effect relative to the molar ratio
[ligand]/[macromolecule] is then plotted. These data can be used to directly
define the thermodynamic parameters of binding and calculate the Kd of the
carbohydrate-protein interaction of interest.
Carbohydrate microarrays constitute a welcome development for elucidating
carbohydrate-protein interactions in a wide variety of contexts, from
cell-protein-, cell-cell- to host-pathogen interactions. It is an extension of both
enzyme-linked immunosorbent assay (ELISA)-type formats and modern DNA
and protein microarray technologies. The main advantages of microarray
analysis lie in its capability to screen a number of biological interactions in
parallel, its potential to present arrayed ligands that can mimic cell surface
display and generate high binding affinities of the interactions, as well as the
minimization of the amount of glycan probes needed for interaction studies.
On a single slide, up to hundreds or even thousands of spots can be printed at
8
the same time. This microarray technique has received much attention ever
since its introduction into the glycoscientific community, and it has
revolutionalized the study of carbohydrate-protein interactions.17,26-34
SPR is a technique for measuring the association and dissociation kinetics of
analytes with their complementary receptors. When either the analyte or the
receptor is immobilized on the SPR sensor chip and associated with the other
in a free form, a change in the refractive index of the layer in contact with a
gold film will occur. This change then leads to the transduction of an SPR
signal measured in resonance units (RU). The interactions are measured in real
time, and thus the association and dissociation data can be obtained. Compared
to carbohydrate microarrays, which usually rely on various protein labeling
formats in a step-wise batch mode, the SPR technique can evaluate biological
interactions on surfaces in a label-free manner.18,19,35-40 Further advantages of
SPR include the capability to measure low affinities from millimolar to
picomolar range, the requirement of small amount of samples and very fast
measurements.
QCM is known as a very sensitive mass-measuring technique in both gas
phase and liquid environment. It measures a mass per unit area by monitoring
the change in frequency of a quartz crystal resonator. The resonance frequency
is disturbed by addition or removal of a small mass on the surface of the
acoustic resonator. As a mass is deposited on the surface of the crystal, the
mass of the crystal increases. Consequently, the frequency of oscillation
decreases compared to the initial value. This frequency change can be
quantified and correlated to the mass change using the Sauerbrey equation.41 In
the gas phase, the technique has been widely used in determining thin-layer
thickness and for gas-adsorption studies.42,43 In liquid environment, affinities
of molecules to surface-bound binding partners can be effectively measured,
and interactions between biomolecules, carbohydrate-protein interactions, for
example, have been investigated.20,21,44,45
9
Atomic force microscopy (AFM), also known as scanning force microscopy
(SFM), is a type of scanning probe microscopy with very high resolution,
which can reach the order of fractions of a nanometer (over 1000 times better
than the optical diffraction limit). AFM has been recognized as one of the
foremost tools for measuring, imaging and manipulating matters at nanoscale.
The AFM uses a cantilever with a sharp tip at its far end for scanning the
specimen surface. During scanning, forces between the tip and the sample,
when they are brought in close proximity, would result in a deflection of the
cantilever according to Hooke’s law. Different forces can be measured in
different situations, including mechanical contact force, van der Waals’ forces,
chemical bonding, electrostatic forces, solvation forces, etc. Most commonly, a
laser spot is used to measure the deflection by reflecting from the top surface
of the cantilever into an array of photodiodes. In recent years, single molecule
force spectroscopy, based on AFM, has been extensively used to address,
detect and measure biomolecular interactions at the single molecule
level.22,46-49
Carbohydrate-functionalized nanomaterials, or glyconanomaterials, are a
variety of nanomaterials carrying carbohydrate ligands on their surfaces. These
functional nanomaterials have emerged as a very promising platform for
studies of carbohydrate-mediated recognitions, and for biomedical imaging,
diagnostics, therapeutics and drug delivery. Gold-, silver- and magnetic
nanoparticles, quantum dots, polymer nanoparticles, liposomes, silica
nanoparticles, carbon-based nanomaterials including fullerenes, carbon
nanotubes and graphene sheets, and virus-based nanoparticles, have been
demonstrated to be great core materials for the build-up of such nanoplatforms.
The platforms combine the unique properties of nanometer-scale objects with
the capability to present carbohydrate ligands in a multivalent manner, greatly
enhancing the binding affinities to their complementary partners and better
mimicking relevant biological recognition events.23,50-52 To achieve a good
performance of such glyconanomaterials, the proper presentations of
carbohydrate ligands on the surfaces, which are closely related to the coupling
chemistry, the type and length of the spacer and the ligand density, are critical.
1.5. Carbohydrate immobilization
1.5.1. Survey of methods
Partly because of the fact that solid-surfaces are better simulacra to cell
surfaces than solution systems, and in part due to their ability to provide
comprehensive hierarchical levels of molecular information,53-55
solid-supported techniques are playing increasingly important roles in studies
of various carbohydrate-protein interactions. As a result, the development of
10
methods for immobilizing carbohydrates on surfaces and resins has become an
important research topic.
Methods for carbohydrate immobilization fall into two general strategies: the
non-covalent and covalent immobilizations. For the former strategy, the
carbohydrate ligands are physisorbed to surfaces via non-covalent interactions,
including electrostatic- and van der Waals’ interactions, hydrophobic effects,
as well as hydrogen bonding. Non-conjugated polysaccharides,34
neoglycoproteins,56 neoglycolipids,26,57-59 carbohydrates with fluorous tags,60,61
ionic glycoconjugates,62,63 and biotinylated carbohydrates,64,65 have for
example been developed as glycan-bearing ligands in non-covalent
immobilization routes (Figure 1.4). Various surface-based carbohydratereceptor studies are subsequently feasible as long as sufficient stability of the
system can be preserved after incubation with the biological target and the
washing steps.
OH
OH
OH
O
O
O
O
O
O
SSSS S S
Au
c
b
S
S
S SS S
Au
OH
O
O
O
O
O
a
nitrocellulose
OH
OH
OH
O
O
n
O
non-covalent
immobilization
nitrocellulose
f luorinated
O
O O
e
f
O O
O
O
O
b
b
C8F17
f luorinated
OH OH
streptavidin
OH OH
d
streptavidin
Figure 1.4 Different non-covalent immobilization methods. (a) Physisorption of
non-conjugated polysaccharides on nitrocellulose coated surfaces. (b) Attachment of
neoglycoproteins. (c) Monolayer formation of neoglycolipids. (d) Immobilization of
fluorous tag derivatized carbohydrates on fluorinated surfaces. (e) Assembly of ionic
glycoconjugates
via
layer-by-layer
deposition.
(f)
Attachment
of
biotinylated
carbohydrates to streptavidin derivatized surfaces.
11
Covalent immobilization methods, in comparison, offer more stable functional
surfaces than the non-covalent alternatives. Thus the covalent strategy has
drawn greater attention and the majority of surface functionalization methods
are based on covalent bond formation. Figures 1.5.1 and 1.5.2 depict a range of
different chemical protocols for covalent immobilization of carbohydrates on
surfaces. These include:
1) Immobilization of functionalized glycans on surfaces: self-assembly of
carbohydrate-alkanethiols on gold surfaces,40,66-68 coupling between
maleimide-thiol,69,70 hydrazide-epoxide,71,72 amine-cyanuric chloride,73
amine-N-hydroxysuccinimide (NHS) ester,74-76 and aminooxy-aldehyde
functionalities,77 attachment via Diels-Alder,28 Staudinger,78 and copper
catalyzed 1,3-dipolar cycloaddition reactions,20,21,57,68 immobilization of
neoglycoproteins
to
aldehyde
slides,70,79
and
site-specific
photo-immobilization of carbohydrates on surfaces.17,18,80
Figure 1.5.1 Different covalent immobilization methods, part 1. (a) Self-assembly of
alkanethiol-functionalized carbohydrates on gold surfaces. (b) Immobilization of
carbohydrates using different coupling chemistries between two typical functionalities. (c)
Attachment
of
neoglycoproteins
to
aldehyde
surfaces.
(d)
Immobilization
of
photoprobe-derivatized carbohydrates on surfaces.
2) Direct immobilization of non-modified glycans on derivatized surfaces:
either through a two-step procedure, e.g., first reacting the free carbohydrates
with 2,6-diaminopyridine or N-methylaminooxy-containing bifunctional
12
linkers and then coupling the intermediates to NHS ester-coated surfaces;81,82
or via one-step routes, which are exemplified by immobilizing free
carbohydrates on aminooxy- or hydrazide-derivatized surfaces,83-85 and by
covalent attachment of unmodified glycans to surfaces derivatized with a layer
of photoreactive groups.19,23,86,87
photoprobe
photoprobe
Figure 1.5.2 Different covalent immobilization methods, part 2. (e) Attaching free glycans
via 2,6-diaminopyridine linker on NHS-ester surfaces. (f) Immobilizing free glycans via
N-methylaminooxy linker on NHS-ester surfaces. (g) Direct immobilization of free
carbohydrates on aminooxy- or hydrazide-derivatized surfaces. (h) Direct attachment of
unmodified glycans to surfaces derivatized with a layer of photoreactive groups.
1.5.2. Photochemical immobilization methods
Recently, photochemical immobilization methods to attach carbohydrates on
solid supports have shown great potential as a result of their simplicity,
robustness and versatility. Photoreactive functional groups like phenylazido
groups,17-23,44,88-93 benzoyl groups,87,94 and diazirine groups,80,86 which can be
converted under UV irradiation into nitrenes, ketyl radicals and carbenes,
13
respectively, have been used for such photochemical functionalization
purposes. Photochemical methods have proven to be advantageous from both
chemical and surface engineering points of view. From the chemical point of
view, the methods 1) provide a clean and easy means for carbohydrate
immobilization in which only photons are required as a reagent, 2) give robust
surfaces that ensure reproducible results and withstand storage, 3) are versatile
to be applicable to a wide range of target molecules, even those that are
normally not reactive to form covalent linkages, 4) are compatible with glycan
structures and aqueous media, and 5) are adaptable to both free and derivatized
carbohydrate structures. From the surface engineering point of view, the
photochemical functionalization methods 1) are versatile to be applicable to a
big variety of substrate structures, 2) are easily adaptable to many
state-of-the-art biotechnologies, and 3) are controllable by external means, for
example, patterning of such surfaces can be achieved by applying a
photomask.
Aryl azides are well-known chromophores that can insert into C-H, N-H bonds
and react with unsaturated carbon-carbon bonds.95 The photochemistry utilized
in our studies is based on a subgroup of aryl azides, the perfluorophenylazides
(PFPAs).89,96,97 The photochemistry of phenyl azide is complex (Scheme 1.3).
Under UV irradiation or heat, the phenyl azide decomposes and forms a singlet
phenyl nitrene, which is highly reactive and can quickly rearrange to a sevenmembered ketenimine (Scheme 1.3, pathway 1). The ketenimine can further
react with amines to give azepinamines or produce polymer tars in the absence
of a nucleophile. The singlet phenyl nitrene can also relax through intersystem
crossing (ISC) to generate the triplet phenylnitrene, which preferably happens
at low temperature and particularly in the presence of alcohols (pathway 2).
The lower energy triplet species can either take part in hydrogen abstraction to
produce aniline-type of structures or undergo bi-molecular reactions to yield
the corresponding azo-compound. However, these are the two reaction
pathways that are not useful for carbohydrate immobilization. To increase the
percentage of useful insertion or addition reactions as shown in the third
pathway (pathway 3), fluorine substituents have been introduced on the phenyl
ring, either in a perfluorinated manner or ortho to the azido group. This
modification can raise the energy barrier of the ring-expansion reaction and the
lifetime of the fluorinated singlet phenyl nitrene can thus be greatly increased,
so that the desired insertion or addition reaction yields are largely enhanced,
too.
14
Scheme 1.3 Simplified illustration of phenylazide photochemistry. Modified from
reference 97.
In fact, the PFPA chemistry has been applied to attaching organic molecules
on surfaces, by either thermal- or photo-activation, resulting in versatile
surface modifications for a wide range of applications.17-23,44,88-93,98-100
1.5.3. Three photochemical carbohydrate immobilization approaches
Three different approaches to photochemically immobilize carbohydrates on
surfaces can be envisaged: 1) a carbohydrate-conjugated photoprobe approach,
2) a direct-photoconjugation approach, and 3) a “hybrid” carbohydrate
immobilization approach. The first approach (referred to as approach I
throughout following text) is achieved by first derivatizing the carbohydrate
structures with a photoprobe, and then applying the entire conjugate molecule
for surface functionalization in a one-step procedure under UV irradiation
(Figure 1.6a). In our studies, PFPA was used as the photoprobe
(Papers II-IV).18,22,101 This approach ensures site-specific attachment of
carbohydrate ligands on surfaces, but it usually requires more labor-demanding
work on the synthesis of each carbohydrate-photoprobe conjugate.
15
Figure
1.6
The
three
approaches
developed
to
photochemically
immobilize
carbohydrates on surfaces. (a) The carbohydrate-conjugated photoprobe approach. (b)
The direct-photoconjugation approach. (c) The “hybrid” carbohydrate immobilization
approach. The letters X, Y, Z and W represent different functional groups.
The direct-photoconjugation approach (approach II) concerns first attaching a
bifunctional photoprobe on surfaces and then photochemically inserting the
photoprobe into the intended target carbohydrate structures (Figure 1.6b).
Bifunctional PFPA-thiols were chosen as model photoreactive molecules in
our studies (Papers V and VI).19,102 The advantage of this approach lies in its
ability to immobilize un-derivatized glycans directly onto surfaces, which can
greatly circumvent the difficulties in chemical derivatization of glycan
structures, facilitate the process of surface modification and make the method
easily applicable to immobilizing complex oligosaccharides and
polysaccharides.
The third approach is a hybrid type of carbohydrate immobilization (approach
III), based on a “photo-click” immobilization protocol (Figure 1.6c).
Specifically, it refers to the method of first attaching the bifunctional
PFPA-alkyne linker on surfaces and then coupling it to different
carbohydrate-azide molecules using the copper catalyzed 1,3-dipolar
cycloaddition “click” chemistry (Papers VII and VIII).20,21 This approach
enables site-specific attachment of carbohydrate ligands on surfaces for
subsequent biological interrogations, and considerably reduces the work load
in organic synthesis, since the simply modified carbohydrate-azide structures
are more easily obtainable compared to the entire carbohydrate-photoprobe
conjugate molecules. In addition, a large number of carbohydrate-azide
16
structures are readily accessible from commercial sources and among research
groups in the glycoscience community. Thus, this approach would enable the
direct use of a wide range of carbohydrate structures.
1.6. Aims of this thesis
The aims of the projects in this thesis are to:
1) Develop different approaches to photochemically functionalize surfaces
with carbohydrates for glycobiological studies. Three such approaches have
been envisaged and developed for the purpose. In approach I, the main task
was to further develop synthetic methodologies for obtaining target
carbohydrate-photoprobe structures in a simpler, yet more efficient way. In
approach II, the focus has been directed to engineering surfaces and interfaces
for direct photoconjugation of free carbohydrates. In approach III, the effort on
developing methods to easily and significantly expand the carbohydrate ligand
diversity was emphasized. A variety of carbohydrate structures, ranging from
monosaccharides to more complex structures, have been designed and
synthesized. Both photoprobe-derivatized and unmodified glycan structures,
from monosaccharides to trisaccharides, were targeted.
2) Design and develop different surface coatings and surface chemical
structures that are adapted to the photoligation process and can minimize
non-specific adsorption and enhance signal-to-noise ratio; applying the
carbohydrate-photoprobe structures to a variety of surface materials was also
considered. Concerning surface coatings, different polymeric materials were
mainly targeted and evaluated; while for surface chemical structures, the
effects of different structural features, including carbohydrate linkage type,
spacer length and spacer structure on subsequent protein binding have been
evaluated. Moreover, different glycan structures have been applied to various
types of surface materials, such as bare gold surfaces, polymer-coated gold
surfaces or glass slides, and carbon-based nanomaterials (data concerning the
last type of materials not shown in this thesis).
3) Combine the photochemical immobilization methods with modern nanoand biotechniques to establish a variety of state-of-the-art platforms, and apply
such platforms for carbohydrate-protein binding studies. These nano- and
biotechniques include the microarray-, SPR-, QCM-, AFM- and
nanomaterials-based technologies as mentioned above. Producing
functionalized surfaces with good specificity and sensitivity and trying to
obtain quantitative analysis of the interactions are basic goals in this part.
Moving from “proof-of-concept” studies to probing more biologically relevant
systems utilizing the established platforms is also within our research pursuit.
17
4) In addition to all the planned aims, solving those unexpected problems
identified during the course of research is another continuous effort for us. The
studies of mutarotation of 1-thioaldoses and stereocontrol in
1-thioglycosylation reactions, to some extent, belong to this type (Papers I and
II).101,103
18
2. Syntheses
In order to develop the photochemical carbohydrate immobilization
methodologies and investigate carbohydrate-protein interactions, a variety of
chemical structures need to be synthesized. Easy, efficient, and stereoselective
syntheses of target molecules are thus highly demanded. To meet the
objectives of this thesis, a series of non-carbohydrate structures, including
linker molecules and PFPA derivatives were first designed and synthesized.
These structures were used either in further synthesis of carbohydrate
derivatives or directly for surface modifications and functionalizations.
Secondly, a number of non-PFPA derivatized glycosides were targeted and
synthesized. This range of molecules include 1-glycosyl thiols,
oligosaccharides and carbohydrate-azides. The original purpose of
synthesizing 1-glycosyl thiols was to develop easier, yet more efficient
synthetic routes for obtaining desired PFPA-conjugated carbohydrate
structures. However, during the process, a conspicuous mutarotation
phenomenon of the 1-glycosyl thiols was observed and identified. As a result,
more such structures were synthesized and the mutarotation properties of these
molecules thoroughly studied. The oligosaccharides and carbohydrate-azide
molecules were used in the development of carbohydrate immobilization
approaches II and III (see pages 15 and 16 ) for studying protein bindings.
Thirdly, three groups of PFPA-conjugated carbohydrates were
stereoselectively synthesized, including O-linked PFPA-carbohydrates,
S-linked PFPA-carbohydrates and oligosaccharide-PFPA conjugates. The
stereoselective formation of the glycosidic linkage is one of the most
challenging aspects in glycoconjugate and oligosaccharide synthesis.104 This
remains a challenge for not only O-linked structures but also their S-linked
analogs. However, these challenges were addressed to ensure the desired
stereochemistry in different target structures. The obtained final structures
were then evaluated against the interrogating proteins.
2.1. Synthesis of non-carbohydrate molecules
2.1.1. Synthesis of linker molecules
A number of linker molecules with different lengths were synthesized as
spacers to link the carbohydrate and PFPA moieties in the PFPA-carbohydrate
structures. The linkers were based on the oligoethylene glycol (OEG) motif,
which has been widely adopted for various biotechnological and biomedical
applications.105 These were primarily chosen because of their water solubility,
stability and their ability to reduce non-specific binding of proteins. The
19
integration of ethylene glycol structures with defined length into other
synthetic molecules usually requires a selective funtionalization and
prolongation of symmetrical diols or diol derivatives. Thus, an iterative
method to produce different bifuntional and monodispersed OEGs with up to 9
units was developed (Scheme 2.1).18
For the shortest linker, triethylene glycol (TEG) derivative 4, the commercially
available chloride 1 was chosen as starting material. The TEG azide 2 was
prepared quantitatively from the chloride 1 by use of sodium azide in DMF at
90 °C overnight. Tosylation at the other end of azide 2 was achieved in good
yield according to an established method.106 SN2 substitution of the tosylate 3
was subsequently achieved by use of sodium iodide in DMF at 50 °C for
5 hours and the desired iodide 4 was obtained in a yield of 82%. To make
longer OEG derivatives, OEG diols 5 and 9 were first transformed into OEG
ditosylates 6 and 10, respectively. The ditosylates were subsequently
prolonged by nucleophilic substitution with the deprotonated azide 2 at one
end of the molecule to afford tosylates 7 and 11 in fair yields. The final iodides
8 and 12 were obtained following similar conditions as for linker 4.
Scheme 2.1 Synthesis of iodo-linkers 4, 8, and 12. a) NaN3, DMF, 90 °C, overnight
(quant.); b) TsCl, TEA, DMAP, THF, 0 °C-RT, 12 h (78%); c) NaI, DMF, 50 °C, 5 h (82%);
d) TsCl, TEA, DMAP, THF, 0 °C-RT, overnight (81%); e) 2, NaH, DMF, 1.5 h (59%); f) NaI,
DMF, 50 °C, 5 h (81%); g) TsCl, TEA, DMAP, DCM, 0 °C-RT, 2 h (84%); h) 2, NaH, DMF,
3 h (63%); i) NaI, acetone, reflux, 1 h (86%).
Several other linkers were synthesized to reduce non-specific protein bindings,
assist best binding capacity as diluter molecules (molecules that are used to
“dilute” the other major functional molecules in the mixture), or correct for the
non-specific protein bindings as control molecules in subseuqent surface
binding studies. These molecules were used for either the SPR or QCM
platform. For the preparation of photoreactive SPR surfaces, non-PFPAderivatized linkers (see one of the examples in Scheme 2.2) were synthesized
as diluter molecules to assist optimal interactions and reduce non-specific
binding.19 Again, the commercially available chloride 1 was chosen as starting
material. An iodo intermediate was generated by substitution of the chloro
functionality in compound 1 using sodium iodide, and thioester 13 was
20
subsequently obtained in the same pot using potassium thioacetate. Hydrolysis
of thioester 13 resulted in the desired thiol 14 and disulfide 15 in a combined
quantitative yield.
Scheme 2.2 Synthesis of thiol 14 and disulfide 15. a) i. NaI, DMF, 90 °C, 2 h; ii. KSAc,
DMF, 80 °C, 6 h (71%); b) K2CO3, MeOH, 3 h (quant.).
For the QCM platform, a control surface made from azidoethanol 21 (Scheme
2.3) was produced to correct for non-specific protein binding. This was
achieved by a three-step procedure: attachment of NHS-derivatized PFPA
(NHS-PFPA) ester on the surface, coupling of bifunctional linker 19, and
finally introduction of the hydroxyl linker 21 through click chemistry.21
Monofunctionalization of diol 5 produced alkyne derivatized linker 16,
followed by tosylation, azido substitution and reduction of the azide group at
the other end of the molecule. These yielded the alkyne-functionalized amine
linker 19. The azidoethanol 21 was obtained by substitution of the chloro
group of chloride 20 using sodium azide.
Scheme 2.3 Synthesis of alkyne-derivatized amine 19 and azide 21. a) KOH, propargyl
bromide, 60 °C, 3 h (36%); b) TsCl, KOH, 0 °C, 2 h (quant.); c) NaN3, TBAI, DMF, 45 °C,
overnight (67%); d) i. PPh3, THF, 30 °C, 19 h; ii. H2O, 30 °C, 25 h (70%); e) NaN3, TBABr,
110 °C, 18 h, 99%.
2.1.2. Synthesis of PFPA derivatives
Besides the non-PFPA derivatized linkers, several PFPA derivatives were also
designed and synthesized, either for further incorporation into the
PFPA-carbohydrate molecules or for separate use in different surface
chemistries. For example, NHS-PFPA ester 25 was produced from
commercially available pentafluorophenyl ester 22 (Scheme 2.4), via
substitution using sodium azide, hydrolysis, and esterification using
carbodiimide chemistry.107 The ester 25 was used for the development of both
carbohydrate immobilization approach I in synthesizing PFPA-carbohydrate
structures and approach III in stepwise surface derivatization.18,20
21
Scheme 2.4 Synthesis of NHS-PFPA ester 25. a) NaN3, Acetone/H2O (2:1), 90 °C, 2 h
(quant.); b) NaOH aq. (20% w/w):MeOH (1:18), 5 h (97%); c) NHS, EDAC, 35 °C, 25 h
(94%).
PFPA functionalized disulfide linker 27 (Scheme 2.5) was produced by
Steglich esterification,108 in which PFPA acid 24 and disulfide diol 26 were
reacted in the presence of DCC and catalytic amount of DMAP. The
compound was applied for the development of carbohydrate immobilization
approaches I and III in double surface photoligation.18,20,22
Scheme 2.5 Synthesis of PFPA disulfide 27. a) DMAP, DCC, DCM, 0 °C-RT, dark, 12 h,
40%.
A longer disulfide linker 29 (Scheme 2.6) was synthesized from starting thiol
28 by iodine promoted oxidation. The resulting disulfide 29 was further
reacted with PFPA acid 24 to produce structure 30 by the same method as for
its analog 27. Both disulfide 30 and its reduced thiol product 31 could be used
for the development of carbohydrate immobilization approach II in
goldnanoparticle surface functionalization.102
22
Scheme 2.6 Synthesis of PFPA-disulfide 30 and PFPA-thiol 31. a) I2, EtOH, 0.5 h
(quant.); b) DMAP, EDAC, DCM, 0 °C-RT, dark, 14 h (37%); c) Zn/HCl, EtOH/MeCN (1:1),
1 h (82%).
Alkyne-derivatized PFPA 32 (Scheme 2.7) was produced by reacting the
NHS-PFPA ester 25 and amine 19 in a straightforward way. This PFPA
derivative was used for further coupling with carbohydrate ligands on solidsupport for a stepwise functionalization.21
Scheme 2.7 Synthesis of alkyne-derivatized PFPA 32. a) MeCN, overnight (70%).
2.2. Synthesis of non-PFPA derivatized glycosides
2.2.1. Synthesis of 1-thioglycosides
The initial objective of producing 1-glycosyl thiols was to develop an efficient
synthetic route directly using unprotected glycosyl donors for linker
conjugation, circumventing the labor-intensive and time-consuming nature of
the protection-glycosylation-deprotection routes generally required for
synthesizing their O-linked analogs. However, in the process, mutarotation of
1-glycosyl thiols was observed and identified. Thus, a number of 1-glycosyl
thiols/thiolates were synthesized to expand the scope of the carbohydrate
analogs for a detailed study of the mutarotation process. Moreover, several of
these and a few other analogous structures were also used for the original
synthetic purpose.
23
Four common 1-glycosyl thiols 33-36 were thus synthesized starting from
either their corresponding commercially available free aldopyranoses 33a-36a
or peracetylated aldopyranoses 33b-36b (Scheme 2.8). After BF3·Et2O
promoted thioacetate substitution and Zemplén deacetylation,45 1-thio-β-Dglucopyranose 33, 1-thio-β-D-galactopyranose 34, 1-thio-α-D-mannopyranose
35 and 1-thio-α-D-altropyranose 36 were accessed in good overall yields.
Scheme 2.8 Synthesis of 1-glycosyl thiolates 33-36. a) I2, Ac2O, 0 °C-RT, 1 h (quant.); b)
BF3·Et2O, HSAc, DCM, 0 °C-RT, 24 h (76-89%); c) NaOMe, MeOH, 2 h (≤ 90%).
Disaccharide thiol derivative, 1-thio-β-D-lactose 40, was also obtained
following the same synthetic route (Scheme 2.9).
Scheme 2.9 Synthesis of 1-thio-β-D-lactose 40. a) I2, Ac2O, 0 °C-RT, 1 h (quant.); b)
BF3·Et2O, HSAc, DCM, 0 °C-RT, 24 h (78%); c) NaOMe, MeOH, 2 h (86%).
In synthesizing 1-thio-β-L-fucopyranose 44 (Scheme 2.10), the first two steps
were similar, but the final deprotection was achieved using lithium
hydroxide.109
O
HO
OH
OH
OH
41
O
a
AcO
OAc
42
OAc
OAc
O
b
AcO
OAc
43
SAc
OAc
O
c
HO
SH
OH
OH
44
Scheme 2.10 Synthesis of 1-thio-β-L-fucopyranose 44. a) I2, Ac2O, 0 °C-RT, 1 h (quant.);
+
b) BF3·Et2O, HSAc, DCM, 0 °C-RT, 24 h (84%); c) i. LiOH, MeOH, H2O, 4 h; ii. H
exchange resin (92%).
Only few reports on the efficient synthesis of 1-thio-α-L-fucopyranose 47
could be found in the literature. However, the α-thiofucose 47 was still
accessible through a straightforward strategy (Scheme 2.11), albeit in
relatively low overall yield. Thus, thioacetate 45 was formed prior to the
24
peracetylation, which led to the desired α-anomer 46 in 30% yield over two
steps.110 Subsequent deacetylation by lithium hydroxide provided the final
product 47.
Scheme 2.11 Synthesis of 1-thio-α-L-fucopyranose 47. a) HSAc, HCl, 0 °C-RT, 5 h; b)
+
Ac2O, pyridine, DMAP (30% over two steps); c) i. LiOH, MeOH, H2O, 4 h; ii. H exchange
resin (90%).
2.2.2. Synthesis of oligosaccharides
To expand the carbohydrate ligand variety used in our studies and extend the
study from “proof-of-concept” to more biologically relevant systems, several
α1-2-linked mannobiose and mannotriose structures were synthesized. These
structures are the outmost moieties from Man9 (Figure 2.1), an oligosaccharide
structure on the human immunodeficiency virus (HIV) envelop glycoprotein
gp120, which binds anti-HIV proteins like Cyanovirin-N (CV-N) at nanomolar
concentration.102,111,112
Figure 2.1 Structure of Man9 N-linked oligosaccharide. The α1-2-linked mannobiose and
mannotriose moieties are shown in the dashed boxes.
A central building block for synthesizing these mannose dimers and trimers is
the orthoester 51 (Scheme 2.12). The structure can be conveniently obtained
starting from commercially available free D-mannose 35a.113 After
peracetylation and bromination, orthoester 49 can be subsequently produced
from bromide 48 by lutidine promoted ring-closure. Subsequent deacetylation
and benzylation afforded desired building block 51 in good overall yield.
25
Scheme 2.12 Synthesis of orthoester 51. a) HBr in AcOH, DCM, 0 °C-RT, 12 h; b)
2,6-lutidine, MeOH/CHCl3 (1:1), 24 h (72% over two steps); c) K2CO3, MeOH, 2 h; d)
NaH, BnBr, DMF, 0 °C-RT, 24 h (87% over two steps).
Orthoester 51 was then transformed to either glycosyl acceptors 54 and 55 or
glycosyl donor 56, both via Lewis acid-promoted/catalyzed ring-opening
processes (Scheme 2.13).
Scheme 2.13 Synthesis of glycosyl donor 56 and glycosyl acceptors 54 and 55. a)
BnOH, BF3·Et2O, DCM, 0 °C-RT, 2 h (96% for 52), or TMSOTf, DCM, 0 °C, 1 h (quant. for
+
53); b) NaOMe, MeOH, H exchange resin, 1 h (97% for 54, 92% for 55); c) EtSH, HgBr2,
MeCN, 60 °C, 24h (72%).
The coupling of the respective glycosyl donor and acceptor was achieved
using NIS/TfOH-mediated chemistry (Scheme 2.14). The α-linkage in
dimannosides 57 and 58 was favored by neighboring group participation from
the acetyl group at C-2 in the glycosyl donor 56. Zemplén deacetylation and
palladium-catalyzed hydrogenolysis of the dimers 57 and 58 provided the
reducing mannobiose 61 and non-reducing mannobioside 62, respectively.
26
Scheme 2.14 Synthesis of mannose dimers 61 and 62. a) NIS, TfOH, DCM, -10 °C, 1 h
+
(78% for 57, 70% for 58); b) NaOMe, MeOH, H exchange resin, 3 h (92% for 59, quant.
for 60); c) H2, Pd/C, MeOH, 72 h (88% for 61, 80% for 62).
However, prior to debenzylation, mannose dimers 59 and 60 can also be
further linked to more carbohydrate rings (Scheme 2.15). When glycosyl donor
56 was used again with the NIS/TfOH-mediated coupling, mannose trimers 63
and 64 were successfully obtained. Similarly, two types of deprotections
subsequently afforded the reducing carbohydrate 67 and the non-reducing
structure 68.
Scheme 2.15 Synthesis of mannose timers 67 and 68. a) 56, NIS, TfOH, DCM, -10 °C, 1
+
h (79% for 63, 76% for 64); b) NaOMe, MeOH, H exchange resin, 3 h (77% for 65, 89%
for 66); c) H2, Pd/C, MeOH, 96 h (94% for 67, 85% for 68).
2.2.3. Synthesis of carbohydrate-azides
A range of carbohydrate-azide molecules were synthesized for the purpose of
27
testing the photo-click carbohydrate immobilization approach. All desired
structures were synthesized in few steps in high yields following previously
reported procedures. Mannose azide 70 and galactose azide 72 were produced
from the corresponding pentaacetates 35b and 34b (Scheme 2.16).26 Structure
76 was prepared from the peracetylated N-acetyl-β-D-glucosamine 73 through
the oxazoline intermediate 74, followed by acid-catalyzed ring-opening and
linker coupling,114,115 and finally Zemplén deprotection.
Scheme 2.16 Synthesis of carbohydrate-azides 70, 72 and 76. a) 21, BF3·Et2O, DCM,
0 °C-RT, 18 h (83%); b) NaOMe, MeOH, 2.5 h (95%); c) 21, BF3·Et2O, DCM, -40 °C-RT,
24 h (69%); d) NaOMe, MeOH, 2.5 h (97%); e) i. TMSOTf, 50 °C, 0.5 h, ii. TEA, 10 min
(98%); f) 21, H2SO4, DCM, 19 h (23%); g) NaOMe, MeOH, 2.5 h (92%).
2.3. Synthesis of PFPA-conjugated carbohydrates
2.3.1. Synthesis of O-linked PFPA-carbohydrates
One of the most straightforward ways to achieve photochemical
immobilization of carbohydrates would be to apply carbohydrate structures
derivatized with a PFPA tag on surfaces in a single step. Thus, the
aforementioned pentaacetates 33b-35b, 38 and azide linker 2 were utilized as
starting materials for the synthesis of such PFPA-carbohydrates
(Scheme 2.17). The corresponding carbohydrate-azides were produced with
the help of BF3·Et2O chemistry. Subsequent deacetylation and azide reduction
yielded the unprotected carbohydrate-amine intermediates, which were further
coupled to NHS-PFPA 25 to provide the desired PFPA-carbohydrate products.
Three monosaccharide derivatives, PFPA-glucose 77, PFPA-galactose 78,
PFPA-mannose 79, and one disaccharide derivative, PFPA-lactose 80, were
28
thus synthesized.
Scheme 2.17 Synthesis of PFPA-carbohydrates 77-80. a) BF3·Et2O, DCM, 0 °C-RT, 24 h
(55-63%); b) NaOMe, MeOH, 1 h (quant.); c) i. H2, Pd/C, MeOH, 1 h, ii. 25, DMF, dark,
2.5 h (58-67%).
In order to evaluate the effect of structural difference of the linkers on
subsequent protein binding, a linker bearing a triazole-ring was designed to
compare with the linear linker structures (Scheme 2.18). Thus, compound 81
was synthesized based on Cu-catalyzed azide-alkyne cycloaddition, producing
the desired triazole-ring in the linker. Subsequently, three steps of reactions
including azide substitution, reduction and amination afforded the final
PFPA-mannose 83 in good overall yield.
Scheme 2.18 Synthesis of PFPA-carbohydrate 83. a) CuI, DIPEA, MeCN, 48 h (76%); b)
i. NaN3, DMF, 65 °C, 20 h (95%), ii. NaOMe, MeOH, 5 h (80%); c) i. H2, Pd/C, MeOH, 1 h
(92%), ii. 25, MeCN, dark, overnight (69%).
2.3.2. Synthesis of S-linked PFPA-carbohydrates
Taking advantage of the fact that sulfur is less basic and more nucleophilic
29
than oxygen, glycosyl thiols are increasingly used in the synthesis of
thiooligosaccharides, thioglycopeptides and thioglycoproteins.116-118 The use of
glycosyl thiols was thus adopted in a straightforward, yet more efficient
synthetic route, compared to the ones for obtaining O-linked analogs, to
produce the S-linked PFPA-carbohydrate structures. Free glycosyl thiolates 33,
35 and 40 were used as donors, and after SN2 reactions with the iodo-linker 4,
the corresponding S-linked glycosides 84a-86a were conveniently obtained
(Scheme 2.19). Subsequent reduction and PFPA-coupling using the methods
mentioned above provided the final S-linked PFPA-carbohydrates 84-86, in
total yields up to around 70%. This synthetic pathway proved highly efficient,
requiring few steps and minimal purification, potentiating its use for
glycobiological studies. However, unexpected anomerization occurred, and
efforts to improve the anomeric purities were thus undertaken (see details in
the next chapter).
Scheme 2.19 Synthesis of S-linked PFPA-carbohydrates 84-86. a) DMF, 50 °C, 1 h
(82-98%); b) i. H2, Pd/C, MeOH, 1 h (quant.), ii. 25, DMF, dark, 2.5 h (77-86%).
To test the effects of different linker lengths on carbohydrate-protein
interactions at surfaces, two additional S-linked PFPA-mannose structures, 89
and 90, were synthesized with hexaethylene glycol- and nonaethylene
glycol-based linkers, respectively (Scheme 2.20).
30
Scheme 2.20 Synthesis of S-linked PFPA-carbohydrates 89 and 90. a) DMF, 50 °C, 1 h
(84% for 87 and 88); b) i. H2, Pd/C, MeOH, 1 h (quant.), ii. 25, DMF, dark, 2.5 h (81% for
89, 82% for 90).
2.3.3. Synthesis of oligosaccharide-PFPA conjugates
To expand the carbohydrate ligand diversity and investigate more biologically
relevant systems, α1-2-linked mannobiose- and mannotriose-PFPA molecules
96 and 102 were synthesized (Scheme 2.21 and 2.22). The azide linker 2 was
allowed to react with orthoester 51 directly to yield the carbohydrate-azide 91,
which was then deacetylated and further allowed to react with the glycosyl
donor 56, leading to azide derivatized dimannoside 93. Compound 93 was
deprotected in a similar manner as for its linker-free analogs (57 and 58),
described previously. The resulting amine 95 was finally allowed to react with
NHS-PFPA ester 25 through amidation, yielding the desired compound 96.
31
Scheme 2.21 Synthesis of PFPA-mannobiose 96. a) TMSOTf, DCM, 0 °C, 2.5 h (72%);
+
b) NaOMe, MeOH, H exchange resin, 1 h (quant.); c) 56, NIS, TfOH, DCM, -10 °C, 1 h
+
(63%); d) NaOMe, MeOH, H exchange resin, 2 h (93%); e) H2, Pd/C, MeOH, 4 d; f) 25,
DMF, dark, overnight.
A similar protocol to obtain the trimer derivative 102 from the dimer 94 was
utilized as for its linker-free analog 67 and 68 (Scheme 2.22). However, in the
case of 102, when the glycosyl donor 56 was allowed to react with acceptor
94, a mixture of α- and β-anomers was produced that turned out to be difficult
to purify. To circumvent this problem, a different glycosyl donor 98 was
synthesized and allowed to react with the acceptor 94.119 The desired α-product
99 was then obtained in 58% yield. Deprotection and amidation then afforded
the final mannotriose-PFPA structure 102.
32
Scheme 2.22 Synthesis of mannotriose 102. a) NaOMe, MeOH, 2 h; b) BzCl, pyridine,
DCM, 6 h (89% over two steps); c) 98, NIS, TfOH, DCM, -15 °C, 1 h (58%); d) NaOMe,
+
MeOH, H exchange resin, 2 h (92%); e) H2, Pd/C, MeOH, 5 d; f) 25, DMF, dark,
overnight.
33
34
3. Mutarotation of 1-thioaldoses and
stereocontrol in 1-thioglycosylation (Papers I
and II)
3.1. Mutarotation of 1-thioaldoses in aqueous media (Paper I)
3.1.1. Introduction
The mutarotation of natural reducing O-saccharides has been known for a long
time.120,121 The composition of the mutarotation products is likely governed by
a complex combination of factors including steric hindrance, stereoelectronicand solvent effects.122 As a result, even after over 160 years since its
discovery, the mutarotation of carbohydrates still remains a relatively
unpredictable/uncontrollable phenomenon.
Certain aldopyranose analogs, in which the endocyclic oxygen and/or the
glycosidic oxygen atom were/was replaced by a heteroatom (cf. Figure 1.3),
also undergo mutarotation.120,123,124 For the sulfur analogs, however, previous
studies on mutarotation have mainly been confined to structures with an
endocyclic sulfur at the C-5 position.125 Mutarotation of 1-thioaldoses, on the
other hand, have not been thoroughly investigated. In fact, free glycosyl
thiols/thiolates were generally considered relatively stable, partly explained by
the poor orbital overlap between the anomeric carbon and the sulfur atom,
which could possibly block the ring-opening and subsequent mutarotation. In
other words, it was relatively unclear under what conditions 1-thioaldoses
would effectively mutarotate. The importance of these molecules in synthesis
and in biological contexts, and the quest for fundamental understanding of
chemical processes led us to investigate and clarify whether 1-thioaldoses
undergo mutaroation. The pH-dependent mutarotation of 1-thioaldoses and
their equilibrium anomeric ratios in aqueous media were thus studied.
3.1.2. pD-dependence of the mutarotation
Stereochemically pure 1-thioaldoses, 1-thio-β-D-glucopyranose 33, 1-thio-β-Dgalactopyranose 34, 1-thio-α-D-mannopyranose 35, and 1-thio-α-Lfucopyranose 47, were initially synthesized and tested under acidic (pD 4),
neutral (pD 7) and basic (pD 9) conditions. The influences of the different
pD’s on the final anomeric compositions were thus demonstrated. The
mutarotation processes of the four 1-thioaldoses were conveniently followed
by 1H-NMR spectroscopy (Figure 3.1). Under acidic or neutral conditions,
mutarotation for all the thiocarbohydrates tested readily occurred and the
process was highly conspicuous.
35
Starting 1-thioaldoses
HO
HO
HO
HO
O
SNa
33
OH
OH
O
HO
SNa
34
OH
HO
HO
HO
SH
OH
O
35
O
SNa
OH
OH
OH
47
1
Figure 3.1 H-NMR kinetic studies of 1-thioaldose mutarotation. Fractions of β-anomers
formed from starting 1-thioaldoses 33, 34, 35 and 47 under acidic condition (pD 4).
More specifically, from Figure 3.1, it can be seen that gluco- (33), galacto(34), and fuco- (47) derivatives yielded anomeric mixtures with β-anomer
ratios of 74.1%, 78.1% and 74.3%, respectively, under acidic conditions
(pD 4). These values are slightly larger compared to the β-anomer ratios of
their corresponding hydroxy-aldoses under the same conditions, but still
following the same trend in favoring one of the anomers. In the mannose case,
for the hydroxy-mannose analog, 66.6% α-anomer was obtained at pD 4;
however, for the thio-mannose 35, the opposite anomer was preferentially
yielded (66.2% β-anomer in the final mixture) under the same experimental
conditions.
Under basic conditions, the 1-thioaldoses showed different mutarotation
behaviors and α-/β- compositions. The mutarotation of thiol analogs 33, 34
and 47 was almost completely restricted at pD 9 (> 95%, 94%, and < 5% of
β-anomer, respectively). To confirm these trends, 1-thio-β-L-fucopyranose 44
was synthesized, which was then tested and compared to its α-analog 47. As
can be seen from Figure 3.2, compound 44 barely mutarotated at pD 9, the
same as its α-analog 47. However, when 44 was subjected to acidic or neutral
conditions, a similar anomeric composition (80% of β-anomer) was formed as
for the α-anomer 47. Thus, the pD-dependence of the mutarotation for 1thioaldoses was clearly demonstrated.
36
Starting 1-thioaldoses
SH
O
OH
OH
SH
OH
O
OH
44
OH
OH
47
1
Figure 3.2 H-NMR kinetic studies of 1-thioaldose mutarotation. Fractions of β-anomers
formed from starting 1-thioaldoses 44 and 47 under basic condition (pD 9).
The 1-thio-α-D-mannopyranose 35 showed slightly different mutarotation
behavior (Figure 3.3). Nevertheless, under basic conditions, the compound still
proved to be more stable compared to under acidic or neutral conditions,
yielding only 24% of β-anomer in the final mixture at pD 9.
Starting 1-thioaldose
HO
HO
HO
OH
O
35
SNa
1
Figure 3.3 H-NMR kinetic studies of 1-thioaldose mutarotation. Fractions of β-anomers
formed from starting 1-thioaldose 35 under acidic, neutral and basic conditions (pD 4,
pD 7, pD 9).
These results pointed to the unusual importance of the axial hydroxyl group at
the C-2 position of the carbohydrate ring in determining the mutarotation
behavior. It was further hypothesized that when 1-thioaldoses pocessed a
37
(2S)-configuration, they would tend to display significant β-anomer preference
under acidic/neutral conditions, yet α-preference under basic conditions.
To test this hypothesis, 1-thio-α-D-altropyranose 36 was prepared, and the
mutarotation behaviors of this compound were tested under acidic, neutral and
basic conditions. The results proved consistent with the predictions. In this
case, complete conversion of the starting α-anomer to β-anomer occurred
within 5 minutes at pD 4 and pD 7, whereas no mutarotation was observed
within the same time span at pD 9.
3.1.3. Reversibility of the mutarotation
The mutarotation of reducing O-saccharides is known to be a reversible
process. Thus the property was also probed for two of the 1-thioaldoses, both
1-thio-α-D-mannopyranose 35 and 1-thio-α-D-altropyranose 36.
Starting 1-thioaldose
HO
HO
HO
OH
O
35
Figure 3.4
SNa
1
H-NMR kinetic studies of 1-thioaldose mutarotation. pD switches for
1-thiomannose 35.
In the thiomannose case, the anomeric mixture compositions changed from
favoring the β-anomer to the α-anomer in response to the pD switch from
acidic to basic conditions, or vice versa (Figure 3.4). These processes were
carried out over several cycles, and the mutarotation rates and anomeric
equilibrium ratios proved to be consistent. However, in the thioaltrose case, the
possibility of multiple pD switches was excluded due to a concomitant
conformational change of the structure 36 (Scheme 3.1). Nevertheless, the
reversibility of the thioaltrose analog was still demonstrated.
38
Scheme 3.1 Proposed 1-thio-altropyranose mutarotation processes.
3.1.4. Proposed mechanisms of the mutarotation of 1-thioaldoses
It was hypothesized that the mutarotation of 1-thioaldoses could follow two
mechanistic pathways (Scheme 3.2). One is likely to go through an
exo-ring-opening process similar to the 1-hydroxy- or 1-aminoaldoses.124,126
The other pathway would involve the formation of the oxocarbenium ion
intermediate (109). The generation of trace amount of thiofuranoses during the
kinetic experiments pointed to the acyclic intermediate formation during
mutarotation. On the other hand, the slow formation of aldoses (110) as well as
the characteristic smell of H2S indicated the involvement of the oxocarbenium
ion pathway.
Scheme 3.2 Proposed mutarotation mechanistic pathways.
3.1.5. Conclusions
The mutarotation of 1-thioaldopyranoses proved to occur readily in aqueous
media. A strong pD-dependence was demonstrated for the process; comparable
mutarotation rates were shown at lower and neutral pD, whereas at higher pD
the process proceeded slower or was almost blocked. The
(2S)-D-thioaldopyranoses tested showed an interesting tendency to
39
preferentially form the β-anomer under acidic or neutral conditions but
α-anomer under basic conditions. The reversibility of the mutarotation process
for these stuctures upon pD-switching was also demonstrated.
3.2. Stereocontrol in 1-thioglycosylation (Paper II)
3.2.1. Introduction
Thiosaccharides and S-linked glycoconjugates are highly valuable substrate
analogs for studies in glycobiology. Considerable effort has been made
towards
the
synthesis
and
biological
evaluations
of
such
glycomimetics.116-118,127-130 The use and importance of 1-thioaldoses for
1-thioglycosylation reactions were documented as early as in the 1960’s.131
Recently, this type of molecules have been further applied to numerous novel
systems including biohybrid polymer systems,132,133 dynamic combinatorial
chemistry,45,109,134,135 glycoclusters based on self-organizable scaffolds,8,136,137
glycoporphyrin-assisted
photodynamic
therapy
(PDT),8,138
and
52,139,140
nanotechnology.
Though the value of thioaldoses and thioglycosides has been increasingly
appreciated, anomerization during the 1-thioglycosylation processes,
especially when using unprotected 1-glycosyl thiols as donors, is rarely
reported in the literature. In fact, it has been generally regarded that glycosyl
thiolates retain their anomeric configuration in alkylation reactions.141
However, anomerization of glycosyl thioethers promoted by Lewis or protic
acid has been more widely realized and studied.142-144 In the following
paragraphs, the process of identifying mutarotation of glycosyl thiols and the
optimization of conditions to control the stereochemistry of the product in
1-thioglycosylation reactions will be discussed and presented.
3.2.2. Occurrence and identification of the anomerization
To synthesize the PFPA-derivatized, S-linked carbohydrates 84-86, 89 and 90,
a general route was initially designed, and the synthesis of mannose derivative
85 was first implemented (Scheme 3.3). Unprotected mannose thiolate 35 was
directly alkylated with a tosylated TEG linker 3 carrying an azide
functionality. Methanol was first used as the reaction solvent due to solubility
considerations. After reaction under ambient conditions, however, a mixture of
products was obtained after purification. Reducing (0 °C) or increasing
(40-60 °C) the temperature, did not prevent the mixture formation.
40
HO
HO
HO
HO
OH
O
+
TsO
HO
HO
a
O
O
N3
S
SNa
35
3
HO
b, c
HO
HO
OH
O
O
O
N3
85a
OH
O
O
S
O
O
F
F
N
H
F
85
N3
F
Scheme 3.3 Initial route for the synthesis of S-linked PFPA-mannose 85. a) MeOH, RT,
72 h, 52% (α/β = 5/1); b) H2, Pd/C, MeOH, 1.5 h, quant.; c) 25, DMF, dark, 2 h, 84%
(α/β = 5/1).
Although to some extent unexpected, a combination of NMR analyses
identified the formation of the β-anomer in the product mixture. This
observation was corroborated by a value of the homonuclear three-bond
coupling constant (J1,2) of 0.63 Hz,145 and the presence of a multiplet
resonating at δ 3.41 ppm, which was assigned to mannose H-5 (Figure 3.5).146
In addition, NOESY-NMR of the anomer displayed a strong nuclear
Overhauser effect between the H-1, H-3 and H-5 protons, clearly suggesting a
β configuration at the anomeric center. Finally, a 2D heteronuclear single
quantum correlation (HSQC) experiment indicated the stereochemistry of the
obtained structures. Thus, that anomerization was taking place under the
present conditions could be confirmed.
1
Figure 3.5 H-NMR spectrum of α- and β-anomers of compound 85a in D2O.
3.2.3. Controlling the stereochemistry in the synthesis
In order to obtain the desired final products with specific stereochemistry, a
number of factors were investigated including reaction time, temperature,
solvent, electrophile and presence of base. Evaluations of the first three
conditions in synthesizing mannose-azide 85a are summerized in Table 3.1.
41
Four
different
polar
aprotic
solvents,
DMSO,
DMF,
N,N’-dimethylpropyleneurea (DMPU) and MeCN, were tested. DMSO turned
out to be the best solvent in controlling the stereochemisry of the product; it
promoted the fastest coupling rates and yielded only desired α-product at all
different temperatures tested. Compared to DMSO, DMF gave relatively less
good results, however, in this case, under the conditions of 3 h of reaction at
50 °C, full conversion of the reaction and fairly high stereoselectivity
(α/β = 11/1) were achievable. Considering practical reasons, DMF was chosen
as reaction solvent, and further optimization of the reaction conditions were
conducted. DMPU and MeCN proved not suitable for the reaction.
Table 3.1 Time, temperature and solvents evaluated
a
Entry
Solvent
Temp.(°C)
Time (h) b
Product
(α/β ratio, yield)
1
DMSO
20
8
only α c
2
DMSO
50
0.5
only α c
3
DMSO
100
0.08
only α c
4
DMF
0
No reaction d
-
5
DMF
20
50
α/β: 7/1, 58% e
6
DMF
50
3
α/β: 11/1, 64% e
7
DMF
100
0.5
α/β: 4/1, 36% e
8
DMPU
0
24
side products
9
DMPU
20
16
side products
10
DMPU
50
1.5
side products
a
In all entries, 1.2 equiv. of glycosyl thiolate 35 and 1 equiv. of linker azide 3
were used. b Time to reach full conversion. c Determined by NMR, yields not
determined. d No reaction after 24 h, no mutarotation observed. e Isolated
yields, based on 50 mg of glycosyl thiolate scale reactions, α/β ratio
determined by NMR.
The effect of the electrophile on the stereocontrol of the reaction was
subsequently tested. Thus, the iodo linker 4 instead of the tosyl linker 3 was
used as the electrophile. Compared to the originally used tosyl moiety, the iodo
functionality was expected to result in an increase in reaction rate of the SN2
displacement, to trap the starting 1-α-mannose thiolate 35 and produce the
corresponding α-glycoside in high anomeric purity. This was supported by the
results: when the iodo linker was used with the thiolate at 50 °C in DMF, the
reaction reached completion within 1 hour, and only α-product was produced
with an isolated yield of 87% (Figure 3.6). In contrast, when the iodo-linker
was reacted with 1-mannose thiolate in MeOH at 50 °C, a mixture of anomeric
42
products was again formed.
1
Figure 3.6 H-NMR spectrum of stereo-controlled compound 85a in D2O.
From the presentation in the first part of this chapter, it is clear that
mutarotation of 1-glycosyl thiols is pH-dependent, where the process is
generally impaired under basic conditions. As a consequence, the effect of
added base was further tested. In this case, the tosyl electrophile 3, which
originally led to a mixture of anomeric products at 50 °C in DMF, was
subjected to glycosylation using 1-α-mannose thiolate 35. The results showed
that the added base, triethylamine (TEA), helped to increase the anomeric
purity at 50 °C in DMF. However, at higher temperature, 100 °C, the basic
environment was not able to efficiently block the mutarotation of the
1-thiomannose.
Moreover, the origin of the anomeric product mixture was clarified.
Anomerization of 1-thioglycosides has been proposed to be able to occur,
albeit under Lewis acidic conditions.147 This could as well lead to the final
anomeric mixture. However, experiments were carried out to exclude the
possibility of product anomerization in the present case. Thus, leaving the
glycoside product in the reaction vessel after reaction completion for
prolonged times did not result in increased anomerization. Likewise, reactions
using the purified α-mannose thioether 85a, in the presence of NaOTs (≥1
equivalent) in DMF at 50 °C for up to one week, yielded no anomerization
products. This supported the mechanistic origin to be mutarotation of the
1-glycosyl thiols, resulting in the formation of final anomeric product mixture.
Following the the optimized conditions for obtaining the anomerically pure
1-thioglycoside 85a, several other S-linked PFPA-carbohydrate conjugates
were synthesized, including monosaccharide derivative 84, disaccharide
derivative 86, as well as two analogs to PFPA-thiomannose 85, the
43
hexaethyleneglycol- and nonaethyleneglycol linker bearing molecules 89 and
90, respectively (cf. Scheme 2.19 and 2.20).
3.2.4. Conclusions
The occurrence and circumvention of anomerization during 1-S-glycosylation
reactions were demonstrated. Highly efficient and stereocontrolled synthesis of
a number of photoprobe-thiosaccharide conjugates was achieved. Mutarotation
of glycosyl thiols proved to be the origin of the anomeric mixtures formed.
Kinetic effects were used to circumvent anomerization. This highly efficient
and stereocontrolled protocol, in combination with a one-step method to
convert free O-saccharides to their corresponding S-analogs,127 will provide a
promising route to various S-linked glycoconjugates for glycobiological
studies.
44
4. Development of three carbohydrate
immobilization approaches and their
applications (Papers II-VIII)
Multivalent binding is prevalent in carbohydrate-protein interactions.148 It has
been well established that multivalency can offer numerous benefits regarding
enhanced receptor selectivity and affinity, and may induce particular
organization on the cell surface to control signal transduction pathways within
cells, which is not achievable with monovalent interactions.149 To mimic the
multivalent interactions, a number of solid-support based techniques have
emerged and since played increasingly important roles in glycobiological
studies. A key step and prerequisite to these studies, however, is the
fabrication of carbohydrate functionalized surfaces. Thus, the development of
methods to immobilize carbohydrates on surfaces has become an important
research topic. Meanwhile, it is also highly desirable to optimize both factors
involved in multivalency and the intrinsic properties of individual ligands to
obtain high specificity and increased affinity.
One of the ultimate applicable goals of studying carbohydrate-protein
interactions is to generate useful information for the development of diagnostic
tools and therapeutics. In this pursuit, a number of methods and tools have
been involved.8 To develop potent carbohydrate-based drugs, the following
steps can be undertaken: first, the presence of a carbohydrate binding protein,
assumingly on a bacterial or viral surface, needs to be confirmed. This can be
achieved by assessing the bacterial/viral genomes, proteomic analysis, and
searching for carbohydrate binding protein homology through bioinformatics.
Then, the isolated proteins can be labeled with fluorogenic probes and
subjected to carbohydrate microarrays to identify the best glycan lead.
Afterwards, the lead candidates can be validated using a range of other
techniques such as ELISA, ITC, SPR, QCM, AFM, and X-ray crystallography.
Next, a panel of simpler structures, compared to the original lead glycans, are
usually evaluated, mainly for the purpose of easing potential manufacturing.
Further lead optimizations can be achieved by classical quantitative
structure-activity relationship (QSAR) methods. The resulting glycomimetics
can be subsequently incorporated into multivalent scaffolds to obtain even
higher binding affinities.
Our efforts in developing robust and versatile carbohydrate immobilization
approaches using photochemistry will be presented in this chapter. Moreover,
the incorporation of these approaches in different nano- and biotechniques to
build useful and general platforms for studying carbohydrate-protein
45
interactions will be elaborated. Such techniques include carbohydrate
microarrays, ITC, SPR, QCM, AFM, and carbohydrate functionalized
goldnanoparticles. These techniques are among the most common tools that
have been extensively applied to the development of novel diagnostics and
therapeutics, as illustrated in the paragraph above. Optimization of the
platforms to achieve good selectivities and the example of the more
biologically relevant system will be emphasized.
4.1. Carbohydrate-conjugated photoprobes (Papers II-IV)
In this section, the photo-initiated glycosurface fabrication using
carbohydrate-conjugated photoprobes will be introduced. This carbohydrate
immobilization approach was subsequently incorporated in SPR and AFM
platforms to study carbohydrate-protein interactions. In the SPR case, the
effects of different structural features among the PFPA-carbohydrate ligands
on protein binding were evaluated. In the AFM study, single molecular forces
in the model mannose-Con A interaction system were measured, yielding
fundamental data like the unbinding/rupture forces and kinetics. Furthermore,
some of the solid supported SPR binding data were compared with the
solution-based ITC results. The ITC results proved consistent with the SPR
data, thus demonstrating the reliability of this solid-support platform for
studying protein binding.
4.1.1. Photo-induced glycosurface fabrication approach I
The first approach of photochemical fabrication of glycan-functionalized
surfaces was realized using carbohydrate-conjugated photoprobes by a double
surface photoligation methodology (Figure 4.1). Either aminated- or
gold-coated- glass substrates could be utilized by treating with NHS-PFPA
ester 25 or PFPA-disulfide 27, respectively. The resulting photoprobederivatized surfaces were subsequently covered with a polymer film followed
by UV irradiation. As a result, the polymer was covalently immobilized on the
surfaces by the fast and efficient C-H insertion reaction of the photogenerated
perfluorophenyl nitrene from PFPA. The matrix then served as the substrate
for the second round of photoligation to introduce the different
PFPA-derivatized carbohydrates. The resulting glycosurfaces were thus
prepared for subsequent protein binding studies.
46
O
O
N
O
O
O
O
O
S
F
F
F
F
F
F
F
F
N3
S
F
F
O
F
N3
F
N3
25
27
Figure 4.1 Immobilization of carbohydrates on surfaces using carbohydrate-conjugated
photoprobes by double surface photoligation.
In the fabrication process, a range of protein-resistant polymers, including
poly(2-ethyl-2-oxazoline) (PEOX), poly(ethylene glycol) (PEG) and
poly(ethylene oxide) (PEO), were evaluated. Of these, PEOX provided a more
stable performance in the SPR setup, and PEG was selected for the AFM
study.
4.1.2. SPR imaging analysis
To evaluate the effects of chemical and structural differences from individual
ligands on protein binding, five mannose-derivatized photoprobes, 79, 83, 85,
89 and 90, were applied on the SPR sensor surface (Figure 4.2). These
structures were varied with respect to anomeric attachment, S- and O-linked
carbohydrates, as well as linker structure and length. Structure 79 was
composed of a mannose head group linked by an O-glycosyl bond to a
TEG-based linker, which was further coupled to the PFPA moiety. In contrast,
structure 85 was linked through an S-glycosyl bond while all the other features
were the same as for compound 79. Compounds 85, 89 and 90 were all
S-linked glycoconjugates, however differing in linker lengths composed of
three to nine ethylene glycol (EG) units. Stucture 83 was designed to possess a
triazole ring in the linker part, with a linker length comparable to that of
47
compound 89. In addition, a galactose-derivatized PFPA structure (78) was
used as a reference compound in the study.
Figure 4.2 SPR responses of six different PFPA-carbohydrates (10 mM) toward Con A
(10 μM) interrogation. Each data point (± SEM) represents the average response of six
replicates for each compound. Δ%R in the Y-axis means the percentage of change in
reflectivity.
Con A,150 the plant lectin originally extracted from the jack bean, Canavalia
ensiformis, was used as a model lectin to probe the carbohydrate-protein
interactions. The binding data are shown in Figure 4.2. First of all, the results
clearly show that Con A selectively binds to its cognate mannose ligands
(compounds 79, 83, 85, 89 and 90), but not to the control compound
(galactose-derivative 78) on the sensor. This matches the established
specificity of the lectin, thus demonstrating good lectin selectivity and
minimum nonspecific binding for the photogenerated glycosurface.
The effects of structural variations among the mannose-derivatives on lectin
binding were also evidenced. As can be seen from the figure, the performances
of the O- and S-linked mannoside analogs, 79 and 85, respectively, upon
Con A interrogation were comparable. Similar binding capacities were
obtained, with that for the O-linked analog 79 being slightly higher. The result
correlated well with the literature reported solid-support ELISA comparison of
S- and O-linked glycoconjugates.151 The two molecules possessed only one
48
atom difference (the glycosyl S- and O-atom, respectively), thus providing a
minimally varied pair of ligands for assessing the reliability and sensitivity of
the photogenerated glycosurfaces. To confirm the result, the solid supported
SPR binding experiments were subjected to a range of different lectin
concentrations for further tests and were also cross-checked with the solution
based ITC data, which will be discussed below.
The effects of other structural variations, i.e., the different linker lengths and
linker structures, on protein binding were also addressed. Three different linker
lengths (3, 6 and 9 EG units) and two different linker structures (linear and
triazole-containing) were evaluated for protein binding on the same sensor.
Based on previous studies, linkers shorter than 3 EG units in the
PFPA-carbohydrate conjugates showed reduced binding efficiency. Only
linkers longer than the TEG moiety were targeted in the present study.
Structures 85, 89, and 90 possessed linear linkers with three, six and nine
ethylene glycol units, respectively. Compound 83 had an equivalent chain
length to 89 however containing a triazole-ring within the linker. As shown in
Figure 4.2, comparing compounds 85, 89 and 90, the lectin binding capacity
decreases as the spacer length increases over the tested range in the system
used. This observation is, however, to some extent expected, likely due to a
combination of potential entanglement and increased intramolecular
photoinsertion during UV irradiation. With respect to the linker structural
difference, the triazole-ring-bearing compound 83 shows slightly higher
binding capacity compared to its linear analog 89, when applied to lectin
interrogation. This agrees with previous studies, indicating that biomolecules
retain their biological activities after incorporation of the triazole rings.152-154
4.1.3. Affinity analyses using SPR and ITC
As mentioned above, further SPR experiments were carried out with varying
concentrations of interrogating Con A solution toward the O- and S-linked
mannose-PFPA conjugates. The carbohydrate concentration was thus fixed at
10 mM, and three different concentrations of Con A solution, 100 nM, 1 μM
and 10 μM, were employed for the lectin binding studies (Figure 4.3).
Consistent and reproducible results were obtained with good signal-to-noise
ratios at all concentrations tested, showing that the S-linked PFPA-mannose 85
bound the lectin slightly weaker than its O-linked analog 79.
49
Figure 4.3 SPR responses of O- and S-linked PFPA-carbohydrates 79 and 85,
respectively, toward Con A interrogation. Each data point (± SEM) represents the average
response of six replicates for each compound.
ITC analyses were subsequently conducted to support the results. ITC is a
well-established method which can provide accurate and quantitative
thermodynamic information of various binding events. Binding affinity data
can be conveniently derived from the measurement, so it was chosen for crosschecking the surface carbohydrate-protein interactions on the SPR sensor
chips. The resulting titration curves of O- and S-linked PFPA-mannose
towards Con A are displayed in Figure 4.4. The obtained thermodynamic data
for the binding are presented in Table 4.1. Compared to the binding affinity of
methyl α-D-mannopyranoside towards Con A (Ka = 7.0 × 103 M-1),155 the Ka
values of both compounds 79 and 85 against Con A fall in the same range,
with that of O-linked conjugate 79 being slightly higher than that of the
S-linked 85. This result is in good agreement with the results obtained from the
solid-supported SPR measurements, thus demonstrating the reliability of the
comparison of lectin bindings between the S- and O-linked carbohydrate
conjugates on the SPR platform, built upon the photochemical carbohydrate
immobilization approach I.
50
Figure 4.4 Calorimetric titration of Con A (80 µM) with carbohydrate derivatives 79 (a)
and 85 (b) (5 mM) at 25 °C: raw data obtained from 20 automatic injections, each of 2 µL.
The integrated curves show experimental points (square dots) and the best fit (solid line).
Table 4.1 Thermodynamic binding parameters of S- and O-linked PFPA-conjugated
carbohydrates with lectins at 25 °C
a
Comp.
Lectin
79
85
Con A a
Con A a
Ka
M-1 × 10-3
6.50 ± 1.04
4.06 ± 0.65
-ΔG
5.2 ± 0.7
4.9 ± 0.4
-ΔH
kcal/mol
13.5 ± 1.0
12.8 ± 1.1
-TΔS
8.3 ± 1.3
7.9 ± 0.9
Average of six experiments.
4.1.4. AFM single molecular force analysis
In the AFM instrumentation setup, not only carbohydrate-functionalized
surfaces are needed, but also AFM tips modified with proteins are prerequisite.
Thus, AFM tips functionalized with Con A using a bifunctional PEG,
NHS-PEG-NHS, were produced (Figure 4.5). The PEG moiety was used to
suppress non-specific interactions. Meanwhile, excess amount of a
monofunctional PEG, NHS-PEG-OMe, was applied on the tip to limit the
number of proteins attached and increase the possibility of obtaining single
molecular interactions.156
51
O
O
O
O
O
O
PEG
N O
PEG OMe
O N
O
O
NHS-PEG-NHS
N O
O
NHS-PEG-OMe
Figure 4.5 AFM tip functionalization. Tip functionalization with Con A via amide formation
between the NHS-ester and the terminal amino group or surface-exposed lysine residues
of Con A.
Mannose-PFPA conjugate 79 was used for probing the interactions, and it is
noteworthy that all the force measurements were carried out in pH<7
environment, where Con A would exist as a dimer rather than a tetramer,157 so
that measurement of multiple Con A/mannose interactions could be reduced.
The interaction force between single mannose/Con A pairs was obtained by
force-distance measurements, that is, after the AFM tip approaches the surface
and interactions between Con A and mannose are established, the tip is
subsequently retracted to measure the unbinding forces required to detach
Con A from mannose. In most cases, the recognition events could not be
observed since the probability of detecting single molecular interactions with
minimally functionalized AFM tips was very low. However, when it occurred,
force curves displaying rupture events could be obtained (Figure 4.6). The
shape of the nonlinear repture peak reflects the elastic properties of the
tethering PEG polymer, whereas the height of the rupture peak is determined
by the strength of the Con A/mannose interaction at a particular tip velocity.
The curves were fitted with a simplified extended freely-jointed chain model
(e-FJCPEG) to determine whether the unbinding events were resulted from the
specific binding partners.158
52
Figure 4.6 Typical retraction force curves for the unbinding of a Con A to a
PFPA-mannose-functionalized surface. The loading rate was 7120 pN/sec (tip velocity
400 nm/s, cantilever spring constant 17.8 pN/nm) with a relative trigger of 100 pN toward
the surface and a dwell time of 1 s. The fits to a simplified extended freely jointed chain
model (e-FJCPEG) are shown as solid lines. The Kuhn lengths determined from the fits
are: 868, 708, and 649 pm from top to bottom.
Control experiments were also performed to confirm the results. Several ways
of eliminating or blocking the Con A/mannose interactions were designed,
including the use of galactose-functionalized surfaces that do not specifically
bind Con A, using monofunctionalized PEG spacer on the AFM tip that do not
link to Con A, or adding free mannose to the buffer to saturate the Con A
binding sites in advance. These control experiments resulted in the absence of
specific interactions (<0.3%), thus reinforcing the reliability of the
aforementioned force measurements.
The random insertion points along the PEG polymer chains by the PFPA
photochemistry however led to the fact that the rupture distance can not be
used as a discriminating factor in analyzing the force curves. Nevertheless, the
PEG Kuhn lengths could be used to serve the analysis in this context.159 Thus,
curves with PEG Kuhn lengths between 600 and 900 pm at each tip velocity
were chosen for further analysis (Figure 4.7a). The distribution of unbinding
forces of the chosen curves was then analyzed with a Gaussian fit, yielding the
most probable unbinding force for the Con A/mannose interaction (Figure
4.7b). The value is typically reported with its corresponding loading rate,
which is the product of the tip velocity and the spring constant.46 In the present
study, for mannose surfaces on aminated glass, an average unbinding force of
70-80 pN for loading rates between 8000 and 40000 pN/s was obtained, while
for mannose surfaces on gold-coated glass, an unbinding force of 57 ± 20 pN
53
at 6960 pN/s was acquired. The results are in agreement with previous
reported unbinding forces in related mannose/Con A studies.
Figure 4.7 Histograms of single molecular force spectroscopy events for unbinding of
Con A from mannose at a loading rate of 8140 pN/sec (tip velocity 400 nm/s, cantilever
spring constant 20.35 pN/nm). (a) Distribution frequency of Kuhn lengths (n = 80); the
highlighted curves between 600 and 900 pm were selected for further analysis. (b)
Distribution frequency of unbinding forces (n = 39); a Gaussian fit was used to determine
the most probable unbinding force to be 69.5 ± 2.7 pN.
Furthermore, measuring the rupture forces of biomolecular interactions at
variable probe loading velocities (dynamic force spectoscopy) allowed
exploring the kinetics of the mannose/Con A interaction.160,161 In our
experiments, the loading rates were varied from 8140 to 40700 pN/s to probe
the binding kinetics. A single linear plot in the Con A/mannose dynamic force
spectroscopy could be obtained (Figure 4.8). From the plot, the dissociation
rate constant (koff) and the distance between the binding complex and the
transition state (xβ) were determined to be 0.16 s-1 and 5.3 Å, respectively.
Both the rupture forces and the kinetics obtained in our study are in good
agreement with previously reported results.162,163
54
Figure 4.8 Dynamic force spectroscopy of Con A/mannose unbinding at loading rates
between 8140 and 40700 pN/sec.
4.1.5. Conclusions
A robust photo-initiated carbohydrate immobilization approach has been
developed. Photoprobe-conjugated carbohydrates were designed and used for
the development. This approach was successfully incorporated into different
platforms to study carbohydrate-protein interactions.
From the initial SPR results, it was demonstrated that all the synthesized
PFPA-carbohydrate conjugates showed good selectivity for the lectin, however
varying in sensitivity. Especially, the O- and S-linked glycosides showed
comparable performances, indicating that the replacement of the glycosidic
oxygen for a sulfur atom resulted in no significant loss of binding. The
intermediate linker length proved to be optimal, and the linker structural
differences in terms of the linear and the triazole-ring features induced small
influence on the lectin binding in the surface system used. Furthermore, the
results showed that the photogenerated glycoarray system was capable of
distinguishing small variations in the glycoconjugate structures and conveying
their effects on binding in a convenient manner.
Further SPR analysis consistently confirmed the initial SPR results at different
protein concentrations, and ITC tests were also conducted to support the
findings. Both methods revealed that the S- and O-linked glycosides showed
comparable lectin affinities, thus demonstrating the reliability of using the
solid-support SPR platform, based on this photochemical immobilization
approach, to study carbohydrate-protein interactions. The combination of the
two methods, solid-phase based SPR and solution-phase based ITC, also
demonstrates potential in probing protein bindings in a more comprehensive
way.
The AFM platform provided fundamental binding data concerning single
molecular rupture forces and kinetics, which are difficult to obtain by other
55
means.
4.2. Direct photoconjugation (Papers V and VI)
In this part, the direct photoconjugation approach to immobilize carbohydrates
on surfaces will be introduced. As discussed in several earlier chapters, current
preparation approaches to carbohydrate-functionalized surfaces are generally
based on derivatized carbohydrate structures. However, the direct conjugation
of underivatized carbohydrates to surfaces remains a challenging task. In this
area, available methods mainly target the reducing end of the glycan chains,
but these methods usually induce ring opening of the terminal saccharide unit,
which can affect the binding properties of smaller saccharide structures.
Another often associated disadvantage with these methods is the requirement
of relatively harsh conditions. In constrast, the PFPA photochemistry provides
a facile and efficient conjugation method to immobilize underivatized
carbohydrates on surfaces.19,102,164
The direct photoconjugation approach was applied to generate carbohydrate
arrays on flat sensor substrates, and SPR was used to read out the binding
events in real time and at high throughput. To our delight, the carbohydrate
ligands immobilized on surfaces in this manner well retained their biological
activities, and the ligand presentation and surface composition had strong
influences on subsequent binding. These aspects will be discussed below in
details.
As already mentioned, multivalent display of carbohydrate ligands has proven
to be of high potential in addressing challenges in probing various
carbohydrate-involved recognition events. Scaffolds coupled with high density
of glycans have thus been developed to mimic the glycan clustering effects on
cell surfaces, enhancing the affinity to carbohydrate-binding entities.54,165,166 In
this context, glyconanoparticles (GNPs) became our desired target to exercise
the direct photoconjugation approach. The anti-HIV lectin, CV-N, and its
mutants, were probed in terms of binding sites and affinities. The idea of
preparing “multivalent glycomimetics” by grafting simple glycans on a
multivalent scaffold, instead of synthesizing complex glycan structures to
obtain high affinities, was tested.
4.2.1. Photo-induced glycosurface fabrication approach II
A key step of the direct photoconjugation approach is the introduction of
PFPA photoprobes to surfaces. Towards this end, the PFPA moiety can be
derivatized with different terminal groups, including thiols/disulfides,
phosphates or silanes, and these structures can subsequently be used to
functionalize different surfaces based on gold-, iron oxide- and silica
56
substrates, respectively. Following the PFPA-modification of surfaces, a
solution of underivatized carbohydrates is applied to the surfaces and the
attachment of underivatized carbohydrates can be achieved by fast UV
irradiation.
In the carbohydrate array case, gold-coated optical glass slides were first
treated with PFPA-derivatized thiols, and solutions of each carbohydrate were
subsequently printed onto the SPR sensors using a robotic printer (Figure
4.9a). Surfaces with different PFPA photoprobes, diluter molecules and ligand
densities were tested, and various underivatized carbohydrates including
mono- (Man, Glc, Gal), di- (Manα1-2Man) and trisaccharides
(Manα1-3[Manα1-6]Man) were immobilized. In the glyconanoparticle case,
uniform, ~22 nm gold nanoparticles were synthesized and in situ
functionalized with a PFPA-disulfide (Figure 4.9b), following a previously
established
procedure.91
Di(Manα1-2Man)
and
trisaccharides
(Manα1-2Manα1-2Man) were then photochemically conjugated to the
PFPA-functionalized gold nanoparticle surfaces by C-H insertion reactions as
described earlier.
Figure 4.9 The direct photoconjugation of free carbohydrates on: (a) flat SPR gold
surfaces, and (b) gold nanoparticle surfaces.
4.2.2. Carbohydrate microarray studies
Initially, the quest for an optimal PFPA photoprobe was pursued. Thus, a
number of thiol-functionalized PFPAs with varying lengths of spacers,
PFPA-C2-SH, PFPA-C6-SH, PFPA-C11-SH and PFPA-MUTEG (Figure 4.10),
were used to derivatize the SPR sensor surfaces. After exposure to each of
57
these thiols in ethanol (4 mM) for 3 h, the surfaces were printed with free
mannose (Man), glucose (Glc) and galactose (Gal) solutions as quadruplets.
Subsequent UV irradiation and washing provided the functionalized SPR
surfaces for protein interrogation. Con A was used again as a model protein,
and the results showed that the SPR responses increased with the length of the
spacer (Figure 4.10). PFPA-MUTEG gave the highest signal and was therefore
chosen as the photo-coupling probe in subsequent studies.
F
N3
F
F
O
F
F
N3
F
F
O
F
F
O
F
F
F
O
O
O
F
N3
F
4
N3
F
F
O
O
SH
9
O
SH
SH
SH
PFPA-C 2-SH PFPA-C 6-SH PFPA-C 11 -SH PFPA-MUTEG
Figure 4.10 SPR responses of Con A with Man, Glc and Gal immobilized on SPR
sensors functionalized with four different PFPA-thiol linkers. Each data point was the
average of three SPR sensors with four replicates for each ligand. Error bars represent
the standard derivation of 12 data sets.
Subsequently, a few diluter molecules were evaluated with respect to their
ability to suppress non-specific interactions. Mercapto-1-hexanol
(HS(CH2)6OH), 1-decanethiol (HS(CH2)9CH3), MUTEG and MDEG (Figure
4.11) were chosen and applied to the SPR sensor surfaces, respectively, to
form monolayers. The resulting surfaces were then exposed to Con A and the
SPR responses were recorded. From the results (Figure 4.11), it can be seen
that surfaces covered by HS(CH2)6OH or HS(CH2)9CH3 gave large
non-specific adsorption signals, while the responses from MUTEG and MDEG
surfaces showed to be much lower. The results are consistent with the general
finding that ethylene glycol units reduce non-specific protein adsorption.
MDEG proved to provide minimal background adsorption, and was thus
chosen as the diluter molecule to be mixed with PFPA-MUTEG in different
molar ratios to treat the SPR sensors.
58
HO
HO
4
HO
O
9
O
SH
HO(CH 2) 6SH
SH
SH
CH3(CH2)9 SH MUTEG
O
SH
MDEG
Figure 4.11 SPR responses of Con A with different thiol-linker derivatized and bare gold
surfaces. Each data point was the average of three SPR sensors with four replicates for
each ligand. Error bars represent the standard derivation of 12 data sets.
Various ratios of photoprobe molecule PFPA-MUTEG versus diluter molecule
MDEG, from neat PFPA-MUTEG to a mixture containing 90% MDEG
(Figure 4.12), were attached to the surfaces, and five underivatized
carbohydrates
including
Man,
Glc,
Gal,
Manα1-2Man
and
Manα1-3[Manα1-6]Man were subsequently printed and covalently attached on
the photoreactive sensor surfaces, in quadruplets in a 5 × 4 patterned array
(Figure 4.12). Con A binding results showed that the SPR responses generally
decreased with decreasing ratio of PFPA-MUTEG in the bifunctional linker
mixture, except when the surfaces were treated with the 90:10
[PFPA-MUTEG]/[MDEG] mixture. The results also revealed the rank of
binding affinities of the applied carbohydrate ligands against Con A. It can be
seen from the figure that trisaccharide Manα1-3[Manα1-6]Man displayed the
highest affinity, followed by disaccharide Manα1-2Man. The binding of Glc
was weaker than that of Man, while Gal gave minimal binding. This order of
affinities and selectivities agreed with that of the corresponding free ligands in
solution,167 thus demonstrating that the surface-attached ligands well retained
their recognition abilities on the platform used.
59
HO
HO
HO
HO
HO
HO
OH
O
OH
O
HO
O
HO
Man3
OH
O
HO
HO
O
HO
OH
O
OH
O
HO
HO
HO
O
OH
Man2
HO
HO
OH
O
OH
Man
HO
HO
HO
OH
O
OH
Glc
OH
O
OH
HO
OH
OH
Gal
Figure 4.12 SPR responses of Con A (100 nM) with an array of carbohydrate surfaces.
(a) SPR sensors were treated with PFPA-MUTEG and MDEG at various mole ratios at
indicated. (b) 5 × 4 Pattern of carbohydrate ligands printed on the SPR surfaces. Each
data point in (a) was the average of three SPR sensors with four replicates for each
ligand as shown in (b). Error bars were omitted for clarity.
Additional studies were carried out to test the regenerability of the
glycosurfaces and probe the detection sensitivities of different carbohydratefunctionalized sensors. The surfaces were subjected to 15 cycles of urea
regeneration, where consistent binding results could be obtained, indicating the
stability and robustness of the glycan microarray surfaces. The sensitivity tests
showed that the oligosaccharide-functionalized surfaces had higher detection
sensitivities than the monosaccharide-derivatized surfaces.
4.2.3. Glyconanoparticle platform for probing CV-N
In this study, a more biologically relevant system, the mannose-CV-N
interaction, which could aid the development of potential anti-HIV virucides,
was probed. As already mentioned, CV-N is a potent anti-HIV protein, more
specifically, an 11 kDa cyanobacterial lectin that can inhibit a number of
viruses, including HIV, at concentrations as low as nanomoles. Thus, it
represents a potential HIV microbicide candidate. The anti-HIV activity of
CV-N is mediated by binding to the high-mannose structures on the envelope
glycoprotein gp120.111,168 Previous studies have already established that the
binding epitopes on N-linked oligomannoses for CV-N comprise α1-2-linked
mannose moieties on the glycan’s D1 and D3 arms (cf. Figure 2.1).112,169,170
However, the use of multivalent glyconanoparticles to probe this lectin has not
60
yet been reported.
Low-mannose structures were thus attached onto gold nanoparticles by the
direct photoconjugation approach to study the glycan binding specificities and
affinities for CV-N. In order to quantitatively probe the affinity enhancement
introduced by the GNPs, a fluorescence competition assay was developed to
determine the apparent dissociation constants of GNP binding to CV-N (KD).
The results obtained from this assay were compared with the Kd values of
monomeric glycan binding to CV-N measured by ITC. The information gained
from the study was envisaged to help the development of new diagnostic and
therapeutic tools directed at combating viral infections.
Two CV-N variants were used in this study: CVNQ50C and CVNMutDB.
CVNQ50C is essentially a more stable wild-type variant, comprising two
separate glycan binding sites, one on Domain A and the other on Domain
B.112,170 α1-2-Linked mannotriose (Man3) moieties preferentially bind to
Domain A while α1-2-linked mannobiose (Man2) structures show preference
for Domain B.170,171 The Cys substitution at position 50 was introduced to
allow for specific fluorescence labeling of the CV-N without interfering with
glycan bindings (referred to as Cy5-labeled CV-Ns throughout text below). In
contrast, the CVNMutDB has its Domain B completely eliminated, keeping only
the binding site on Domain A active.
After the carbohydrate functionalization of the gold nanoparticles, the surface
ligand density was determined by a colorimetric assay using anthrone/sulfuric
acid.91 The mannobiose and mannotriose functionalized gold nanoparticles
(GNP-M2 and GNP-M3) (Figure 4.13) showed to present 1516 ± 232
Man2- and 1037 ± 148 Man3-entities per particle, respectively.
61
Figure 4.13 Synthesis of the α1-2-linked mannobiose and mannotriose functionalized
gold nanoparticles (GNP-M2 and GNP-M3).
With these two sets of binding partners prepared, a series of interaction studies
were subsequently implemented. It can be expected that when CVNQ50C is used
to interact with the GNPs, crosslinking between the two partners will occur
leading to formation of complexes. However, no or less crosslinking should be
possible between the GNPs and CVNMutDB. Indeed, when GNP-M3 was treated
with CVNQ50C, a red shift from 529 nm to 542 nm in the surface plasmon
resonance band of gold nanoparticles could be observed (Figure 4.14b),
indicating particle size growth.172 Transmission electron microscopy (TEM)
tests confirmed this indication by showing the presence of clusters of
aggregated particles (Figure 4.14d). Similar tests were also conducted for
CVNMutDB, but the results revealed neither SPR shift nor particle aggregations
(Figure 4.14a,c). Dynamic light scattering (DLS) measurements of CVNQ50Cand CVNMutDB-treated GNP-M3 particles yielded average particle sizes of
38.3 ± 4.6 nm and 25.9 ± 3.5 nm, respectively (Figure 4.14e), further
confirming the above conclusions. These results are in good agreement with
previous structural studies on the CV-N mutants.173
62
Figure 4.14 UV-vis spectra of GNP-M3 before (solid line) and after (dotted line) treatment
with: (a) CVN
CVN
mutDB
CVN
Q50C
mutDB
, and (b) CVN
, and (d) CVN
Q50C
Q50C
. TEM micrographs of GNP-M3 treated with: (c)
. (e) DLS measurements of CVN
MutDB
- (middle) and
- (right) treated GNP-M3 particles.
Next, the binding affinities of the GNPs with the CV-N variants were
quantitatively evaluated using a fluorescence-based competition assay.23 Free
ligand competitors, Man2 and Man3, were incubated with a fixed
concentration of Cy5-labeled CVNQ50C, in the presence of varying
concentrations of GNP-M2 and GNP-M3, respectively. After incubation, the
solution was centrifuged to remove all GNP-CVN complexes, and the
fluorescence intensity of the resulting supernatant was measured. The
difference in fluorescence intensity of Cy5-CVNQ50C before and after
incubation corresponds to the amount of the bound (and removed) CVNQ50C.
Concentration response curves for GNP-M2 and GNP-M3 thus permit to
determine the IC50 values (Figure 4.15a). However, to determine the binding
constants of GNPs with CVNQ50C, it is as well necessary to obtain the Kd
values of the monomeric ligands, Man2 and Man3, with CVNQ50C
63
(Figure 4.15b,c). These values were conveniently determined by ITC and the
dissociation constants, Kd1 for Domain A and Kd2 for Domain B
(Figure 4.15d), were calculated based on a two-site binding model.
Figure 4.15 Fluorescence binding assay studies. (a) Concentration response curves of
GNP-M2 (right) and GNP-M3 (left). (b) Schematic representation of a binding scenario.
(c) Modified Cheng-Prusoff equation based on a competitive two-site binding model,
where [M] is the concentration of the free ligand for the glycan-binding sites on Domains
Q50C
, respectively. The data were fitted using the maximum bound
A and B of CVN
fractions, fBmax1 and fBmax2, corresponding to the two binding sites, and dissociation
constants KD1 and KD2 as adjustable parameters. (d) ITC measurement of bindings
between free Man2 (left) / Man3 (right) and CVN
Q50C
.
These data, together with the IC50 values, were then used to calculate the
64
apparent dissociation constants, KD1 and KD2, of GNPs for the two binding
sites on CVNQ50C, Domain A and B, respectively. The data are summarized in
Table 4.2. From the results, it can be seen that both GNP-M2 and GNP-M3
exhibit affinity enhancements of several orders of magnitude compared to their
free monomeric glycan ligands, when binding to CVNQ50C. Considering the
enhancement of affinity per ligand, an increase of up to several hundred times
is still present for the gold nanoparticle-bound glycans (Table 4.2,
enhancement factor (EF) values in parentheses). Interestingly, for both
GNP-M2 and GNP-M3, the affinity enhancement is more pronounced for their
better binding domain. That is, when Man2 exhibits a higher affinity for the
binding site on Domain B, its multivalent analog GNP-M2 induces the affinity
enhancement by 176 folds for Domain B versus 8.2 folds for Domain A. In
contrast, for GNP-M3, the opposite was observed that the EF is higher for the
Domain A site (309) than for Domain B (3.6).
Table 4.2 Affinities for Man2/3 (Kd) and GNP-M2/3 (KD) binding to CVN
Q50C
. Numbers in
parentheses correspond to EF (= Kdi/( KDi × number of ligands on GNP))
Ligand
Kd1 or KD1 (Domain A)
Kd2 or KD2 (Domain B)
Man2
GNP-M2
Man3
GNP-M3
700 ± 50 µM
56.4 ± 7 nM (8.2)
3.4 ± 0.2 µM
0.011 ± 0.007 nM (309)
64 ± 4 µM
0.24 ± 0.1 nM (176)
43 ± 2 µM
11.8 ± 2.3 nM (3.6)
4.2.4. Conclusions
A new approach, the PFPA-based direct photoconjugation approach, to
immobilize carbohydrates on surfaces has been developed. This new approach
was applied to fabricate carbohydrate arrays that could be analyzed directly by
SPR. Using this approach, underivatized carbohydrates were covalently
attached to the SPR sensors, leading to robust and stable surfaces that could be
used repeatedly. Equally importantly, the surface bound glycan ligands
retained their affinities and selectivities toward their binding proteins.
Interaction studies also showed that the affinity was highly dependent on the
ligand density.
Furthermore, two mannose structures, Man2 and Man3, were successfully
grafted onto gold nanoparticles using the same approach. The resulting GNPs
were used to probe two CV-N variants, the single binding-site protein
CVNMutDB and the two binding-site protein CVNQ50C. The bindings occurred in
a predictable manner, where crosslinked complexes and aggregates were not
observed for CVNMutDB but were seen for CVNQ50C. Moreover, the GNPs
exhibited significantly higher affinity towards the CV-N lectins, compared to
65
the free glycan ligands. Substantial affinity enhancements per carbohydrate
ligand attached were also recorded. These data showed that gold nanoparticles
indeed serve as efficient multivalent scaffolds to enhance the apparent
affinities and to provide glycan clustering effects. The enhanced affinities
compare well with those of natural or synthetic high-mannose structures as
well as other types of synthetic multivalent ligands. Therefore, the idea of
grafting simple glycans on gold nanoparticle scaffolds to prepare multivalent
glycomimetics, instead of synthesizing complex glycan structures, to obtain
high affinities, was also demonstrated. These approaches and ideas hold strong
promise in the development of new and effective diagnostic tools and
therapeutics.
4.3. Photo-click carbohydrate immobilization (Papers VII and
VIII)
In this section, the design of carbohydrate immobilization approach III will be
described. This approach is a “hybrid” of PFPA photo-immobilization and
Cu-catalyzed azide alkyne cycloaddition (CuAAC, one of the best known
“click” reactions) methods. The photo-click approach combines the advantages
of approaches I and II described above. It introduces site-specific carbohydrate
attachment, reduces the synthetic burden, as well as extends this approach to a
large range of readily accessible carbohydrate-azide structures. The photoclick concept was designed to be realized in two experimental procedures,
which were further compared with immobilization approach I of using the
entire PFPA-carbohydrate entity. Moreover, one of the photo-click procedures
was elaborated with various polymer coatings, multiple carbohydrate
structures and interrogated by a number of different lectins.
The immobilization approach was tested on the QCM platform. Comparisons
of bindings on the constructed surfaces by the three different experimental
procedures, and additional evaluations of different surface performances using
one of the photo-click procedures will be discussed.
4.3.1. Photo-induced glycosurface fabrication approach III
Initially, two procedures of the photo-click approach were designed to
immobilize carbohydrates on the QCM chip surfaces. They were further
compared to approach I. All of the three procedures should theoretically result
in comparable final surfaces (Figure 4.16). The earlier described
double-surface photoligation method was used here to generate the polymer
surfaces and subsequent glycosurfaces, except for certain polystyrene (PS)
surfaces which were obtained from a commercial source. More specifically,
one of the photo-click procedures consisted of two steps (Figure 4.16,
procedure B): firstly, alkyne-derivatized PFPA 32 was attached to the surface
66
by UV irradiation. The alkyne surface was then coupled to the
carbohydrate-azide 70 using the CuAAC chemistry. The other procedure,
however, comprised three steps (Procedure C), and the alkyne surface in this
case was produced in two separate steps: first, the polymeric surface was
derivatized by the NHS-PFPA ester 25. The resulting surface was
subsequently immersed in a solution of alkyne-derivatized amine linker 19.
Amidation reactions on the surface afforded the same alkyne-functionalized
surface as in procedure B, which was then subjected to the carbohydrate-azide
70. In comparison, carbohydrate immobilization approach I (Procedure A) was
also utilized, in which the mannose-derivatized photoprobe 83 was
immobilized on the surface in a single photoligation step under UV light.
HO
HO
HO
OH
O
HO
OH
O
HO
O
70
OH
O
N3
OH
72
N3
HO
HO
O
HO
O
NHAc
76
N3
N3
21
Figure 4.16 Two photo-click procedures (B and C), in comparison with photochemical
approach I (procedure A), to immobilize carbohydrates on surfaces.
Subsequently, procedure C was utilized to further probe the capability of this
photo-click approach in generating a general platform to study carbohydrateprotein interactions. Thus, a number of additional polymers, including
polyacrylamide (PAAm), PEG, PEOX, and polypropylene (PP), were used to
coat the QCM sensor surfaces in the first photoligation step (cf. Figure 4.1).
Three different carbohydrate-azide structures, 70, 72 and 76, were used in the
third step of procedure C (cf. Figure 4.16). The polymeric surfaces were also
functionalized with 2-azidoethanol 21 to produce control surfaces.
4.3.2. Protein binding and surface specificity studies by QCM
The biorecognition properties of the surfaces were evaluated using the QCM
flow-through instrumentation, with repeated injections of lectins specific for
the carbohydrate ligands used in the studies. The QCM technique allows for
67
real-time monitoring of the association and dissociation of the lectins and
provides an easy way to assess the stability of the immobilization over several
injections.
In the first study, which mainly concerned the comparison of binding
performances based on the three immobilization procedures (cf. Figure 4.16),
lectins Con A and PNA (primarily specific for β-D-galactopyranose ligands)
were used to interrogate the mannose functionalized surfaces. The interactions
were evaluated by measuring the decrease in frequency upon lectin injection.
Figure 4.17 shows the typical binding curves for the two lectins used in the
study. As can be seen, when the α-D-mannopyranose-specific lectin Con A
was applied, the frequency of the oscillating manno-surface dropped
immediately upon injection as a consequence of the association of the lectin to
the surface. As the injection came to the end, dissociation took place where the
lectin was washed away from the surface. In comparison, the lectin PNA
showed no binding to the surface, in agreement with its known specificity.
Figure 4.17 Typical binding curves of Con A and PNA toward the mannose-functionalized
surfaces.
Comparison of the binding performances for all three experimental procedures
are presented in Figure 4.18. As shown, procedure A produced the highest and
most consistent binding for Con A, while at the same time presenting
negligible nonspecific binding of PNA to the surfaces. The observation is
understandable since procedure A is the simplest route with only one
photoligation step and is also very likely to produce the most glycan-dense
surfaces. Procedures B and C showed high and specific binding of Con A
toward the surfaces, although with different binding capacities. These results
confirm the generality of the different procedures to produce surfaces of
similar quality.
68
Figure 4.18 Binding results for the three different procedures (normalized for the
molecular weight of the lectins).
In the second study, procedure C was chosen and a few more polymers,
carbohydrates and lectins were used for the versatility evaluation of the
photo-click approach and the QCM platform. The carbohydrates and lectins
were chosen based on their binding specificities and previous experiences of
surface binding analyses. Con A and PSA are complementary lectins for α-Dmannose residues, whereas WGA and BS-II bind selectively to N-acetyl-β-Dglucosamine and RCA-I to β-D-galactose. To minimize nonspecific binding,
BSA was used to block all non-covered surface areas by several injections in
the beginning of the experiment. Several studies have shown that blocking
with BSA in the initial phase of the experiment lasts throughout the whole
experiment, thus obviating the need for BSA in the running buffer.21,45
The dissociation and association of the lectins to the surfaces could again be
clearly observed, with small levels of nonspecific binding found for some of
the lectins. These nonspecific interactions are hypothesized to be due to, e.g.,
surface structure and composition, blocking coverage by BSA, and pI values
of the lectins, etc. However, these can be corrected for by a control surface,
using the non-carbohydrate linker molecule 2-azidoethanol (21). Thus, a
corrected average binding value was obtained for each lectin on each surface
with good reproducibility over time (Figure 4.19). As shown, high binding
specificities were found for all lectins evaluated, and the expected trends were
followed. Con A and PSA displayed binding to the mannose surfaces, WGA
and BS-II to the N-acetylglucosamine surfaces, and RCA-I to the galactose
surfaces. The difference in binding capacities between the different lectins
mainly resulted from parameters like the lectin molecular weight and the
affinity for the carbohydrate ligands used in the study.
69
Figure 4.19 Corrected lectin binding to PEG-coated surfaces. Each data represents an
average of binding values from two separate injections of protein, subtracted with the
averaged binding from the 2-azidoethanol control surface.
The lectin binding profiles to all glycan-functionalized QCM sensor surfaces
are depicted in Figure 4.20. From the protein point of view, the Con A
interactions proved to be the most consistent in the study, with more than
40 Hz corrected binding to all mannose-functionalized surfaces and with
relatively insignificant nonspecific binding. PSA followed the same binding
pattern as for Con A, though with lower capacity for the selective mannose
interactions. The other lectins showed certain variations in binding among the
different polymeric surfaces. For example, relatively low bindings of WGA
and BS-II to the GlcNAc/PAAm surface can be seen while high bindings to
GlcNAc/PEG are observed. RCA-I also showed different capacities of binding
to the galactose-functionalized surface, although to a lower extent compared
with WGA and BS-II binding to the GlcNAc-surfaces.
Figure 4.20 Corrected binding values of lectins to the functionalized surfaces: (a) PEG.
(b) PEOX. (c) PAAm. (d) PS. (e) PP.
70
The different binding features can also be viewed from the polymeric surface
point of view. Comparison of the differences in sensitivity showed that the PS
surfaces in general displayed negative bindings to the mismatched
carbohydrate-functionalized surfaces, for example, RCA-I to Man and PSA to
GlcNAc. Whereas the PP surfaces displayed a resulting positive binding, e.g.,
PSA to GlcNAc and Con A to Gal. The hydrophilic polymer surfaces (PAAm,
PEG, and PEOX) generally showed a more coherent binding than the
hydrophobic polymer surfaces (PS and PP), demonstrated by the more
coherent corrected lower binding values to the mismatched surfaces.
4.3.3. Conclusions
The photo-click approach has been successfully developed to immobilize
carbohydrates on surfaces in a fast, straightforward and site-specific way. The
approach makes use of the highly efficient PFPA photoligation, in
combination with the CuAAC chemistry. QCM instrumentation was used as
the platform to incorporate the immobilization approach and to evaluate
subsequent protein binding performances of the glycan-surfaces. High
specificities and capacities were recorded for all surfaces, and all the surfaces
can also be applied with a variety of polymeric substrates.
In the first study on this platform, three different experimental procedures were
adapted for the glycosurface production, all resulting in functional surfaces
with good quality, and the two photo-click procedures can be easily expanded
with a larger range of carbohydrate-azide structures, as well as other
azide-containing molecules.
In the second study on the platform, the versatility of the approach and the
platform was further probed by diversifying the polymeric coating materials,
the carbohydrate ligands and the complementary interrogating lectins. The
predicted protein binding selectivities were observed for all the glyco-surfaces
developed, thus demonstrating the potential of this technology. PEG-coated
sensors proved to give very good binding performances; other formats also
showed high quality. The photochemically fabricated carbohydrate surfaces
are inherently robust, requiring only low-cost starting materials and ambient
conditions. In general, these features point to a convergent fabrication
approach that constitutes a versatile platform for sensing and other types of
glycobiological studies.
71
72
5. Concluding remarks and outlooks
This thesis has mainly been directed to developing new, straightforward and
robust photochemical approaches to immobilize carbohydrates on surfaces for
glycobiological studies. Three different approaches, namely the
carbohydrate-conjugated photoprobe approach, the direct photoconjugation
approach and the “hybrid” carbohydrate immobilization approach, have been
designed and successfully applied into a wide range of nanobio-platforms.
These platforms comprise some of the most common tools used in modern
glycobiological field, including carbohydrate microarrays, biosensors like SPR
and QCM, AFM and glyconanomaterials. The demonstrated compatibility with
a large number of applications, the robust and versatile coupling chemistry,
and the controllable and adaptable surface features involved in the
development, have endowed great potential for the photochemical
carbohydrate immobilization approaches presented.
All the glycan-functionalized surfaces have displayed good selectivities toward
protein binding. Moreover, the surfaces have been used to evaluate the effects
of structural variations in surface chemical components on subsequent binding
(SPR platform), to compare with solution-phase based ITC measurement of
carbohydrate-protein interactions, and to probe fundamental binding
parameters like the single molecular rupture forces and kinetics (AFM
platform). The surfaces can also be adapted to fabricating carbohydrate
microarrays using common present-day arrayers, enabling evaluation of
multiple glycan-protein bindings in parallel. Carbohydrate functionalized
surfaces based on gold nanoparticles have also been produced, leading to
quantitative analysis of their interactions toward the anti-HIV lectin CV-N.
The robustness of the surfaces has been demonstrated by the reproducible
binding results on the surfaces and their ability to withstand storage. The
versatility of the surfaces was demonstrated by their compatibility with a wide
range of polymeric coatings, glycan ligand structures and interrogating
proteins (QCM platform).
To make the above studies possible, synthetic methods have been developed
and carbohydrate structures synthesized. For example, a highly efficient and
stereocontrolled protocol, using unprotected thiocarbohydrates, to obtain
various S-linked glycoconjugates was developed. Furthermore, a long-standing
question, the mutarotation of 1-thioaldopyranoses, has been proven to occur
readily in aqueous media. A strong pH-dependence was demonstrated for the
process.
Looking forward, a series of research projects can be further pursued based on
the current thesis:
73
In order to more efficiently use the carbohydrate microarray platform for
applications in glycobiology, projects on extending the arrays with an
appropriate number of carbohydrate structures can be carried out. The
carbohydrate structures may be obtained either from synthetic means based on
our own experience and effort, or from commercial sources, consortia and
other research groups in glycobiology and glycochemistry.
Evaluation of additional methods is desired to expand the photochemical
carbohydrate immobilization approaches. For example, a site-specific
immobilization method to attach free glycans can be invisaged by using
hydroxylamine- or hydrazide-terminated PFPA-linker molecules. In this
context, studies to determine and elucidate the point of attachment of the free
carbohydrate rings by the direct photoconjugation approach (approach II) have
been initiated.
More detailed information about the glycan surfaces is valuable, which can be
obtained by various surface characterization tools and methods. For example,
to determine the immobilization efficiency of various PFPA-carbohydrates
with different linker lengths, respective PFPA-linkers derivatized by certain
fluorophores can be synthesized and used for the purpose.
Quantitative analyses of carbohydrate-protein interactions using our SPR,
QCM and microarray platforms are also highly desirable. Some attempts in
this direction are currently ongoing.
Moreover, the production of higher-density carbohydrate arrays is another
desired research aim, the combination of microarray technologies with
nanomaterials scaffolds will aid us in achieving such goals. For the
applications of the established glycan surfaces, current studies have been
mainly focused on different carbohydrate-binding protein analyses, however,
much more can be explored. These include for example high-throughput
ligand screening by the microarray platform and feasibility study of bacterial
or toxin analysis.
Last but not least, preparing carbohydrate-functionalized polymer
nanoparticles by use of the established photochemical approaches for various
diagnostic and therapeutic applications, such as targeted drug delivery and
release, is also of great interest, considering a number of unique properties of
polymeric materials and their compatibility in biological systems.
74
Acknowledgements
Herein I would like to take the opportunity to express my deepest gratitude and
heartfelt thanks to all the people who have guided, supported and helped me in
various ways during the past four years.
First and foremost, I sincerely thank my supervisor, Professor Olof Ramström,
for accepting me as a PhD student in the group, for guiding me all the way
along, for offering me various support and opportunities, for all kinds of
advices, for strong trust in me from day one and for your forever open office
door – amazing availability and patience whenever I want to discuss with you.
Olof, it has truly been a fantastic experience and great pleasure working with
you!
A very sincere thanks also goes to Professor Mingdi Yan, thank you for your
recommendation in the first place, for your everlasting guidance and support,
and for the fruitful and pleasant collaborations in the last four years.
I also feel grateful to:
Professor Tore Brinck, Professor Louis Cuccia and Professor Angela
Gronenborn for nice collaborations.
My other co-authors: Professor Hai Dong, Oscar Norberg, Dr. Rémi Caraballo,
Dr. Luis Amorim, Xin Wang, Suji Uppalapati, Dr. Anuradha Tyagi, Carolin
Madwar, William Chu Kwan, Dr. Rolf Schmidt, Dr. Shan Zou, Dr. Teodor
Aastrup, Dr. Elena Matei, Dr. Martin Rahm, Adrien Joliton, Dr. Delphine
Gallois-Montbrun and Dr. Manishkumar Shimpi for pleasant team work in
different projects.
Professor Torbjörn Norin for taking time and effort to preview my thesis and
provide constructive criticisms and valuable comments. Dr. Rémi Caraballo,
Dr. Morakot Sakulsombat and all the current Ramström group members for
helping with proof-reading and comment on the thesis.
My colleagues and friends in the Ramström group. A special thanks to Hai for
helping me with all sorts of matters when I just started in the group and arrived
in the country. Other former members including Zhichao, Rikard, Gunnar,
Pornrapee (Jom), Rémi, Marcus and Morakot, current members Oscar, Yan,
Lei, Manish, Niranjan, Juan, Na, Yang, Sheng, Fredrik, Javier, and my three
Erasmus students Adrien, Katharina and Philipp during the years – thank you
all for being good friends, for creating pleasant time in the lab and for
spreading a nice atmosphere throughout the group.
75
Professor Christina Moberg, Professor Torbjörn Norin, Professor Licheng Sun,
Professor Peter Somfai, Professor Anna-Karin Borg Karlson and Dr. Zoltán
Szabó for interesting graduate courses.
Dr. Ulla Jacobsson and Dr. Zoltán Szabó for support with the NMR
instrumentation and other matters.
Lena Skowron, Henry Challis and Ilona Mozsi for all kinds of help.
All former and present coworkers in the Chemistry/Organic Chemistry
Department for a friendly and pleasant working atmosphere.
The Swedish Research Council, the Royal Institute of Technology and the
China Scholarship Council for financial support.
The Aulin-Erdtman foundation and Knut och Alice Wallenbergs Stiftelse for
travel grants.
I would also like to warmly thank some of my non-research-related friends:
Lin Cao, Xiaoling Cao, Hairong Deng, Junmei Duan, Jing Fu, Hailong Li,
Minyue Li, Xiaogai Li, Han-Hsuan Lin, Zhonghai Lu, Zhanyu Ma, Yuxin Pei,
Jian Qin, Zhongwei Si, Yongbo Tang, Tianling Wei, Juan Wu, Sihong Wu,
Ailing Xu, Chunlin Xu, Ting Yang, Yongliang Yang, Peiqun Ye, Liang Zhang,
Qian Zhang, Qinglin Zhang, Chelsea, Mandy, Markus, Victor…. Thank you
guys for making my days in Stockholm this colorful and memorable.
I thank all the good friends and teachers in China for everlasting care and
support as well.
A special thanks to Aunt Zhilan, for so much care, help and support in the past
four years. You have made my life here really much easier and more
comfortable. Thank Xinxin for being cute and nice.
Elaine, thank you for the many years of love, and for being patient, considerate
and supportive.
My parents and sister in China, my gratitude becomes nonverbal when I think
of you. Thank you, thank you and thank you!
最衷心的感谢父母这么多年来对我最无私的爱、培育、鼓励和支持！也
谢谢姐姐一直以来的关心和爱。
76
Appendix
The following is a description of my contribution to publications I to VIII, as
requested by KTH.
Paper I: I contributed to the formation of the research problems (major
contribution to the initiation of the project), and shared the experimental work
and writing.
Paper II: I contributed to the formation of the research problems, performed
the majority of the experimental work and wrote the manuscript.
Paper III: I contributed to the formation of the research problems, performed
the majority of the experimental work and wrote the manuscript.
Paper IV: I contributed to the formation of the research problems, and
performed part of the experimental work.
Paper V: I contributed to the formation of the research problems, and
performed part of the experimental work.
Paper VI: I contributed to the formation of the research problems, and shared
the experimental work and writing.
Paper VII: I contributed to the formation of the research problems, and
performed part of the experimental work.
Paper VIII: I contributed to the formation of the research problems, and
performed part of the experimental work.
77
78
References
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
(10)
(11)
(12)
(13)
(14)
(15)
(16)
(17)
(18)
(19)
(20)
(21)
(22)
(23)
(24)
Varki, A. Glycobiology 1993, 3, 97-130.
Dwek, R. A. Chem. Rev. 1996, 96, 683-720.
Varki, A.; Cummings, R. D.; Esko, J. D.; Freeze, H. H.; Stanley, P.;
Bertozzi, C. R.; Hart, G. W.; Etzler, M. E. Essentials of Glycobiology,
Cold Spring Harbor (NY): Cold Spring Harbor Laboratory Press,
2009.
Rademacher, T. W.; Parekh, R. B.; Dwek, R. A. Annu. Rev. Biochem.
1988, 57, 785-838.
Kalyuzhin, O.; Zemlyakov, A.; Kalina, N.; Mulik, E.; Kuzovlev, F.;
Makarova, O. Bull. Exp. Biol. Med. 2008, 145, 623-625.
Edward, J. T. Chem. Ind. 1955, 1102-1104.
Lemieux, R. U. Pure Appl. Chem. 1971, 25, 527-548.
Chabre, Y. M.; Roy, R. Adv. Carbohydr. Chem. Biochem. 2010, 63,
165-393.
Coombe, D. R.; Kett, W. C. Cell. Mol. Life Sci. 2005, 62, 410-424.
Lever, R.; Page, C. P. Nat. Rev. Drug Discov. 2002, 1, 140-148.
The Future of Biotech. The 2010 Guide to Emerging Markets and
Technology, BioWorld Today Publishers, 2010.
Bertozzi, C. R.; Kiessling, L. L. Science 2001, 291, 2357-2364.
Sears, P.; Wong, C.-H. Science 2001, 291, 2344-2350.
Seeberger, P. H.; Werz, D. B. Nature 2007, 446, 1046-1051.
Dam, T. K.; Brewer, C. F. Chem. Rev. 2002, 102, 387-430.
Mischnick, P.; Niedner, W.; Adden, R. Macromol. Symp. 2005, 223,
67-77.
Pei, Z.; Yu, H.; Theurer, M.; Waldén, A.; Nilsson, P.; Yan, M.;
Ramström, O. ChemBioChem 2007, 8, 166-168.
Deng, L.; Norberg, O.; Uppalapati, S.; Yan, M.; Ramström, O. Org.
Biomol. Chem. 2011, 9, 3188-3198.
Tyagi, A.; Wang, X.; Deng, L.; Ramström, O.; Yan, M. Biosens.
Bioelectron. 2010, 26, 344-350.
Norberg, O.; Deng, L.; Aastrup, T.; Yan, M.; Ramström, O. Anal.
Chem. 2011, 83, 1000-1007.
Norberg, O.; Deng, L.; Yan, M.; Ramström, O. Bioconjugate Chem.
2009, 20, 2364-2370.
Madwar, C.; Chu Kwan, W.; Deng, L.; Ramström, O.; Schmidt, R.;
Zou, S.; Cuccia, L. A. Langmuir 2010, 26, 16677-16680.
Wang, X.; Ramström, O.; Yan, M. Anal. Chem. 2010, 82, 9082-9089.
Jansson, J. L. M.; Maliniak, A.; Widmalm, G. ACS Symp. Ser. 2006,
930, 20-39.
79
(25)
(26)
(27)
(28)
(29)
(30)
(31)
(32)
(33)
(34)
(35)
(36)
(37)
(38)
(39)
(40)
(41)
(42)
(43)
(44)
(45)
(46)
(47)
80
Larsen, K.; Thygesen, M. B.; Guillaumie, F.; Willats, W. G. T.;
Jensen, K. J. Carbohydr. Res. 2006, 341, 1209-1234.
Fazio, F.; Bryan, M. C.; Blixt, O.; Paulson, J. C.; Wong, C.-H. J. Am.
Chem. Soc. 2002, 124, 14397-14402.
Feizi, T.; Fazio, F.; Chai, W.; Wong, C.-H. Curr. Opin. Struct. Biol.
2003, 13, 637-645.
Houseman, B. T.; Mrksich, M. Chem. Biol. 2002, 9, 443-454.
Laurent, N.; Voglmeir, J.; Flitsch, S. L. Chem. Commun. 2008,
4400-4412.
Liu, Y.; Feizi, T. In: Glycoscience. Fraser-Reid, B., Tatsuta K., Thiem
J. (Eds), Springer-Verlag Berlin Heidelberg, 2008, p. 2121.
Love, K. R.; Seeberger, P. H. Angew. Chem. Int. Ed. 2002, 41,
3583-3586.
Park, S.; Lee, M.-R.; Shin, I. Chem. Commun. 2008, 4389-4399.
Song, E.-H.; Pohl, N. L. B. Curr. Opin. Chem. Biol. 2009, 13,
626-632.
Wang, D.; Liu, S.; Trummer, B. J.; Deng, C.; Wang, A. Nat.
Biotechnol. 2002, 20, 275-281.
D'Agata, R.; Grasso, G.; Iacono, G.; Spoto, G.; Vecchio, G. Org.
Biomol. Chem. 2006, 4, 610-612.
Karamanska, R.; Clarke, J.; Blixt, O.; MacRae, J.; Zhang, J.; Crocker,
P.; Laurent, N.; Wright, A.; Flitsch, S.; Russell, D.; Field, R.
Glycoconjugate J. 2008, 25, 69-74.
Linman, M. J.; Yu, H.; Chen, X.; Cheng, Q. ACS Appl. Mater.
Interfaces 2009, 1, 1755-1762.
Liu, W.; Chen, Y.; Yan, M. Analyst 2008, 133, 1268-1273.
Mercey, E.; Sadir, R.; Maillart, E.; Roget, A.; Baleux, F.;
Lortat-Jacob, H.; Livache, T. Anal. Chem. 2008, 80, 3476-3482.
Smith, E. A.; Thomas, W. D.; Kiessling, L. L.; Corn, R. M. J. Am.
Chem. Soc. 2003, 125, 6140-6148.
Sauerbrey, G. Z. Phys. 1959, 155, 206-222.
Grate, J. W.; Martin, S. J.; White, R. M. Anal. Chem. 1993, 65,
940A-948A.
Guilbaut, G. C. In: Methods and Phenomena. Their Applications in
Science and Technology, Lu C., Czanderna, A. W. (Eds), Elsevier,
New York, 1984, p. 251.
Pei, Y.; Yu, H.; Pei, Z.; Theurer, M.; Ammer, C.; Andre, S.; Gabius,
H.-J.; Yan, M.; Ramström, O. Anal. Chem. 2007, 79, 6897-6902.
Pei, Z.; Larsson, R.; Aastrup, T.; Anderson, H.; Lehn, J.-M.;
Ramström, O. Biosens. Bioelectron. 2006, 22, 42-48.
Bizzarri, A. R.; Cannistraro, S. Chem. Soc. Rev. 2010, 39, 734-749.
Neuman, K. C.; Nagy, A. Nat. Methods 2008, 5, 491-505.
(48)
(49)
(50)
(51)
(52)
(53)
(54)
(55)
(56)
(57)
(58)
(59)
(60)
(61)
(62)
(63)
(64)
(65)
(66)
(67)
(68)
(69)
(70)
(71)
(72)
Handbook of Molecular Force Spectroscopy. Noy, A. (Ed.), Springer:
New York, 2008.
Zlatanova, J.; Lindsay, S. M.; Leuba, S. H. Prog. Biophys. Mol. Biol.
2000, 74, 37-61.
de la Fuente, J. M.; Barrientos, A. G.; Rojas, T. C.; Rojo, J.; Cañada,
J.; Fernández, A.; Penadés, S. Angew. Chem. Int. Ed. 2001, 40,
2257-2261.
El-Boubbou, K.; Huang, X. Curr. Med. Chem. 2011, 18, 2060-2078.
Marradi, M.; Martín-Lomas, M.; Penadés, S. Adv. Carbohydr. Chem.
Biochem. 2010, 64, 211-290.
Cohen, M.; Varki, A. OMICS 2010, 14, 455-464.
Lee, Y. C.; Lee, R. T. Acc. Chem. Res. 1995, 28, 321-327.
Lundquist, J. J.; Toone, E. J. Chem. Rev. 2002, 102, 555-578.
Willats, W. G. T.; Rasmussen, S. E.; Kristensen, T.; Mikkelsen, J. D.;
Knox, J. P. Proteomics 2002, 2, 1666-1671.
Bryan, M. C.; Fazio, F.; Lee, H.-K.; Huang, C.-Y.; Chang, A.; Best,
M. D.; Calarese, D. A.; Blixt, O.; Paulson, J. C.; Burton, D.; Wilson, I.
A.; Wong, C.-H. J. Am. Chem. Soc. 2004, 126, 8640-8641.
Feizi, T.; Stoll, M. S.; Yuen, C.-T.; Chai, W.; Lawson, A. M. Meth.
Enzymol. 1994, 230, 484-519.
Mann, D. A.; Kanai, M.; Maly, D. J.; Kiessling, L. L. J. Am. Chem.
Soc. 1998, 120, 10575-10582.
Ko, K.-S.; Jaipuri, F. A.; Pohl, N. L. J. Am. Chem. Soc. 2005, 127,
13162-13163.
Pohl, N. L. Angew. Chem. Int. Ed. 2008, 47, 3868-3870.
Uzawa, H.; Ito, H.; Izumi, M.; Tokuhisa, H.; Taguchi, K.; Minoura, N.
Tetrahedron 2005, 61, 5895-5905.
Uzawa, H.; Ito, H.; Neri, P.; Mori, H.; Nishida, Y. ChemBioChem
2007, 8, 2117-2124.
Linman, M. J.; Taylor, J. D.; Yu, H.; Chen, X.; Cheng, Q. Anal. Chem.
2008, 80, 4007-4013.
Shao, M.-C. Anal. Biochem. 1992, 205, 77-82.
Houseman, B.; Mrksich, M. Top. Curr. Chem. 2002, 218, 1-44.
Qian, X.; Metallo, S. J.; Choi, I. S.; Wu, H.; Liang, M. N.; Whitesides,
G. M. Anal. Chem. 2002, 74, 1805-1810.
Zhang, Y.; Luo, S.; Tang, Y.; Yu, L.; Hou, K.-Y.; Cheng, J.-P.; Zeng,
X.; Wang, P. G. Anal. Chem. 2006, 78, 2001-2008.
Park, S.; Shin, I. Angew. Chem. Int. Ed. 2002, 41, 3180-3182.
Ratner, D. M.; Adams, E. W.; Su, J.; O'Keefe, B. R.; Mrksich, M.;
Seeberger, P. H. ChemBioChem 2004, 5, 379-383.
Lee, M.-R.; Shin, I. Angew. Chem. Int. Ed. 2005, 44, 2881-2884.
Park, S.; Lee, M.-R.; Shin, I. Nat. Protocols 2007, 2, 2747-2758.
81
(73)
(74)
(75)
(76)
(77)
(78)
(79)
(80)
(81)
(82)
(83)
(84)
(85)
(86)
(87)
(88)
(89)
(90)
(91)
(92)
(93)
82
Schwarz, M.; Spector, L.; Gargir, A.; Shtevi, A.; Gortler, M.; Altstock,
R. T.; Dukler, A. A.; Dotan, N. Glycobiology 2003, 13, 749-754.
Blixt, O.; Head, S.; Mondala, T.; Scanlan, C.; Huflejt, M. E.; Alvarez,
R.; Bryan, M. C.; Fazio, F.; Calarese, D.; Stevens, J.; Razi, N.;
Stevens, D. J.; Skehel, J. J.; van Die, I.; Burton, D. R.; Wilson, I. A.;
Cummings, R.; Bovin, N.; Wong, C.-H.; Paulson, J. C. Proc. Nat.
Acad. Sci. 2004, 101, 17033-17038.
Disney, M. D.; Seeberger, P. H. Chem. Biol. 2004, 11, 1701-1707.
Liang, P.-H.; Wang, S.-K.; Wong, C.-H. J. Am. Chem. Soc. 2007, 129,
11177-11184.
de Paz, J. L.; Spillmann, D.; Seeberger, P. H. Chem. Commun. 2006,
3116-3118.
Köhn, M.; Wacker, R.; Peters, C.; Schröder, H.; Soulère, L.;
Breinbauer, R.; Niemeyer, C. M.; Waldmann, H. Angew. Chem. Int.
Ed. 2003, 42, 5830-5834.
Adams, E. W.; Ratner, D. M.; Bokesch, H. R.; McMahon, J. B.;
O'Keefe, B. R.; Seeberger, P. H. Chem. Biol. 2004, 11, 875-881.
Chevolot, Y.; Bucher, O.; Léonard, D.; Mathieu, H. J.; Sigrist, H.
Bioconjugate Chem. 1999, 10, 169-175.
Bohorov, O.; Andersson-Sand, H.; Hoffmann, J.; Blixt, O.
Glycobiology 2006, 16, 21C-27C.
Xia, B.; Kawar, Z. S.; Ju, T.; Alvarez, R. A.; Sachdev, G. P.;
Cummings, R. D. Nat. Methods 2005, 2, 845-850.
Lee, M.-R.; Shin, I. Org. Lett. 2005, 7, 4269-4272.
Zhi, Z.-L.; Powell, A. K.; Turnbull, J. E. Anal. Chem. 2006, 78,
4786-4793.
Zhou, X.; Zhou, J. Biosens. Bioelectron. 2006, 21, 1451-1458.
Angeloni, S.; Ridet, J. L.; Kusy, N.; Gao, H.; Crevoisier, F.;
Guinchard, S.; Kochhar, S.; Sigrist, H.; Sprenger, N. Glycobiology
2005, 15, 31-41.
Carroll, G. T.; Wang, D.; Turro, N. J.; Koberstein, J. T. Langmuir
2006, 22, 2899-2905.
Liu, L.-H.; Dietsch, H.; Schurtenberger, P.; Yan, M. Bioconjugate
Chem. 2009, 20, 1349-1355.
Liu, L.-H.; Yan, M. Acc. Chem. Res. 2010, 43, 1434-1443.
Wang, H.; Zhang, Y.; Yuan, X.; Chen, Y.; Yan, M. Bioconjugate
Chem. 2011, 22, 26-32.
Wang, X.; Ramström, O.; Yan, M. J. Mater. Chem. 2009, 19,
8944-8949.
Wang, X.; Ramström, O.; Yan, M. Chem. Commun. 2011, 47,
4261-4263.
Wang, X.; Ramström, O.; Yan, M. Adv. Mater. 2010, 22, 1946-1953.
(94)
(95)
(96)
(97)
(98)
(99)
(100)
(101)
(102)
(103)
(104)
(105)
(106)
(107)
(108)
(109)
(110)
(111)
(112)
(113)
(114)
(115)
(116)
(117)
Carroll, G.; Wang, D.; Turro, N.; Koberstein, J. Glycoconjugate J.
2008, 25, 5-10.
Bayley, H.; Scroven, J. V. In: Azides and Nitrenes. Scriven, E. F.
(Ed.), 1984, p. 434.
Platz, M. S. Acc. Chem. Res. 1995, 28, 487-492.
Yan, M. Chem. Eur. J. 2007, 13, 4138-4144.
Liu, L.; Yan, M. Angew. Chem. Int. Ed. 2006, 45, 6207-6210.
Liu, L.-H.; Lerner, M. M.; Yan, M. Nano Lett. 2010, 10, 3754-3756.
Yan, M.; Cai, S. X.; Wybourne, M. N.; Keana, J. F. W. J. Am. Chem.
Soc. 1993, 115, 814-816.
Deng, L.; Wang, X.; Uppalapati, S.; Norberg, O.; Dong, H.; Joliton,
A.; Yan, M.; Ramström, O. Submitted 2011.
Wang, X.; Matei, E.; Deng, L.; Ramstrom, O.; Gronenborn, A. M.;
Yan, M. Chem. Commun. 2011, 47, 8620-8622.
Caraballo, R.; Deng, L.; Amorim, L.; Brinck, T.; Ramström, O. J.
Org. Chem. 2010, 75, 6115-6121.
Demchenko, A. V. In: Handbook of Chemical Glycosylation:
Advances in Stereoselectivity and Therapeutic Relevance.
Demchenko, A. V. (Ed.), 2008, pp 1-27.
Harris, J. M. In: Poly(ethylene glycol) Chemistry: Biotechnical and
Biomedical Applications; Harris, J. M. (Ed.), 1992, pp 1-14.
Lee, Y.; Koo, H.; Jin, G.-W.; Mo, H.; Cho, M. Y.; Park, J.-Y.; Choi, J.
S.; Park, J. S. Biomacromolecules 2005, 6, 24-26.
Keana, J. F. W.; Cai, S. X. J. Org. Chem. 1990, 55, 3640-3647.
Neises, B.; Steglich, W. Angew. Chem. Int.Ed. 1978, 17, 522-524.
André, S.; Pei, Z.; Siebert, H.-C.; Ramström, O.; Gabius, H.-J.
Bioorg. Med. Chem. 2006, 14, 6314-6326.
Hashimoto, H.; Shimada, K.; Horito, S. Tetrahedron: Asymmetry
1994, 5, 2351-2366.
Bewley, C. A.; Gustafson, K. R.; Boyd, M. R.; Covell, D. G.; Bax, A.;
Clore, G. M.; Gronenborn, A. M. Nat. Struct. Mol. Biol. 1998, 5,
571-578.
Bewley, C. A.; Otero-Quintero, S. J. Am. Chem. Soc. 2001, 123,
3892-3902.
Beignet, J.; Tiernan, J.; Woo, C. H.; Kariuki, B. M.; Cox, L. R. J. Org.
Chem. 2004, 69, 6341-6356.
Eklind, K.; Gustafsson, R.; Tidén, A. K.; Norberg, T.; Åberg, P. M. J.
Carbohydr. Chem. 1996, 15, 1161-1178.
Nakabayashi, S.; Warren, C. D.; Jeanloz, R. W. Carbohydr. Res. 1986,
150, c7-c10.
Driguez, H. Top. Curr. Chem. 1997, 187, 85-116.
Gamblin, D. P.; Scanlan, E. M.; Davis, B. G. Chem. Rev. 2009, 109,
131-163.
83
(118)
(119)
(120)
(121)
(122)
(123)
(124)
(125)
(126)
(127)
(128)
(129)
(130)
(131)
(132)
(133)
(134)
(135)
(136)
(137)
(138)
(139)
(140)
84
Pachamuthu, K.; Schmidt, R. R. Chem. Rev. 2006, 106, 160-187.
Zhang, Y.-M.; Mallet, J.-M.; Sinay, P. Carbohydr. Res. 1992, 236,
73-88.
Isbell, H. S.; Pigman, W.; Melville L. Wolfrom, R. S. T.; Derek, H.
Adv. Carbohydr. Chem. Biochem. 1970, 24, 13-65.
Pigman, W.; Isbell, H. S.; Melville, L. W.; Tipson, R. S. Adv.
Carbohydr. Chem. Biochem. 1968, 23, 11-57.
Juaristi, E.; Cuevas, G. Tetrahedron 1992, 48, 5019-5087.
Davis, B. G.; Fairbanks, A. J. Carbohydrate Chemistry, Oxford
University Press, Oxford, 2002.
Sinnott, M. L. Carbohydrate Chemistry and Biochemistry: Structure
and Mechanism, Royal Society of Chemistry, 2007.
Grimshaw, C. E.; Whistler, R. L.; Cleland, W. W. J. Am. Chem. Soc.
1979, 101, 1521-1532.
Smiataczowa, K.; Kosmalski, J.; Nowacki, A.; Czaja, M.; Warnke, Z.
Carbohydr. Res. 2004, 339, 1439-1445.
Bernardes, G. J. L.; Gamblin, D. P.; Davis, B. G. Angew. Chem. Int.
Ed. 2006, 45, 4007-4011.
Bundle, D. R.; Rich, J. R.; Jacques, S.; Yu, H. N.; Nitz, M.; Ling,
C.-C. Angew. Chem. Int. Ed. 2005, 44, 7725-7729.
Marcaurelle, L. A.; Bertozzi, C. R. Chem. Eur. J. 1999, 5, 1384-1390.
Thayer, D. A.; Yu, H. N.; Galan, M. C.; Wong, C.-H. Angew. Chem.
Int. Ed. 2005, 44, 4596-4599.
Horton, D.; Hutson, D. H.; Melville, L. W.; Tipson, R. S. Adv.
Carbohydr. Chem. 1963, 18, 123-199.
Chen, Y.; Chen, G.; Stenzel, M. H. Macromolecules 2010, 43,
8109-8114.
van Dongen, S. F. M.; de Hoog, H.-P. M.; Peters, R. J. R. W.; Nallani,
M.; Nolte, R. J. M.; van Hest, J. C. M. Chem. Rev. 2009, 109,
6212-6274.
Caraballo, R.; Dong, H.; Ribeiro, J. P.; Jiménez-Barbero, J.;
Ramström, O. Angew. Chem. Int. Ed. 2010, 49, 589-593.
Caraballo, R.; Sakulsombat, M.; Ramström, O. Chem. Commun.
2010, 46, 8469-8471.
Dondoni, A.; Marra, A. Chem. Rev. 2010, 110, 4949-4977.
Nelson, A.; Belitsky, J. M.; Vidal, S. b.; Joiner, C. S.; Baum, L. G.;
Stoddart, J. F. J. Am. Chem. Soc. 2004, 126, 11914-11922.
Hirohara, S.; Obata, M.; Alitomo, H.; Sharyo, K.; Ando, T.; Yano, S.;
Tanihara, M. Bioconjugate Chem. 2009, 20, 944-952.
Fernandez, J.; Mellet, C.; Defaye, J. J. Incl. Phenom. Macrocyclic
Chem. 2006, 56, 149-159.
Yu, M.; Yang, Y.; Han, R.; Zheng, Q.; Wang, L.; Hong, Y.; Li, Z.; Sha,
Y. Langmuir 2010, 26, 8534-8539.
(141)
(142)
(143)
(144)
(145)
(146)
(147)
(148)
(149)
(150)
(151)
(152)
(153)
(154)
(155)
(156)
(157)
(158)
(159)
(160)
(161)
(162)
(163)
(164)
(165)
(166)
Josse, S.; Le Gal, J.; Pipelier, M.; Cléophax, J.; Olesker, A.; Pradère,
J.-P.; Dubreuil, D. Tetrahedron Lett. 2002, 43, 237-239.
Boons, G.-J.; Stauch, T. Synlett 1996, 1996, 906,908.
Connolly, D. T.; Roseman, S.; Lee, Y. C. Carbohydr. Res. 1980, 87,
227-239.
Kartha, K. P. R.; Cura, P.; Aloui, M.; Readman, S. K.; Rutherford, T.
J.; Field, R. A. Tetrahedron: Asymmetry 2000, 11, 581-593.
Yu, H. N.; Ling, C.-C.; Bundle, D. R. J. Chem. Soc., Perkin Trans. 1
2001, 832-837.
Crich, D.; Li, H. J. Org. Chem. 2000, 65, 801-805.
Encinas, L.; Chiara, J. L. Eur. J. Org. Chem. 2009, 2009, 2163-2173.
Kiessling, L. L.; Young, T.; Grubber, T. D.; Mortell, K. H. In:
Glycoscience. Fraser-Reid, B., Tatsuta K., Thiem J. (Eds),
Springer-Verlag Berlin Heidelberg, 2008, p. 2484.
Choi, S. K. Synthetic multivalent molecules: concepts and biomedical
applications, Wiley, New York, 2004.
Sumner, J. B. J. Biol. Chem. 1919, 37, 137-142.
Nilsson, U.; Johansson, R.; Magnusson, G. Chem. Eur. J. 1996, 2,
295-302.
Best, M. D. Biochemistry 2009, 48, 6571-6584.
Devaraj, N.; Collman, J. QSAR Comb. Sci. 2007, 26, 1253-1260.
Hanson, S.; Greenberg, W.; Wong, C. H. QSAR Comb. Sci. 2007, 26,
1243-1252.
Schwarz, F. P.; Puri, K. D.; Bhat, R. G.; Surolia, A. J. Biol. Chem.
1993, 268, 7668-7677.
Hinterdorfer, P.; Dufrene, Y. F. Nat. Methods 2006, 3, 347-355.
Ratto, T. V.; Langry, K. C.; Rudd, R. E.; Balhorn, R. L.; Allen, M. J.;
McElfresh, M. W. Biophys. J. 2004, 86, 2430-2437.
Guo, S.; Ray, C.; Kirkpatrick, A.; Lad, N.; Akhremitchev, B. B.
Biophys. J. 2008, 95, 3964-3976.
Oesterhelt, F.; Rief, M.; Gaub, H. E. New J. Phys. 1999, 1, 6.1 - 6.11.
Bell, G. I. Science 1978, 200, 618-627.
Evans, E.; Ritchie, K. Biophys. J. 1997, 72, 1541-1555.
Krieg, M.; Helenius, J.; Heisenberg, C.-P.; Muller, D. J. Angew. Chem.
Int. Ed. 2008, 47, 9775-9777.
Ratto, T. V.; Rudd, R. E.; Langry, K. C.; Balhorn, R. L.; McElfresh,
M. W. Langmuir 2006, 22, 1749-1757.
Wang, X.; Norberg, O.; Deng, L.; Ramström, O.; Yan, M. ACS Symp.
Ser. 2011, in press.
Huskens, J. Curr. Opin. Chem. Biol. 2006, 10, 537-543.
Mammen, M.; Choi, S. K.; Whitesides, G. M. Angew. Chem. Int. Ed.
1998, 37, 2754-2794.
85
(167)
(168)
(169)
(170)
(171)
(172)
(173)
86
Mandal, D. K.; Kishore, N.; Brewer, C. F. Biochemistry 1994, 33,
1149-1156.
O'Keefe, B. R.; Shenoy, S. R.; Xie, D.; Zhang, W.; Muschik, J. M.;
Currens, M. J.; Chaiken, I.; Boyd, M. R. Mol. Pharmacol. 2000, 58,
982-992.
Barrientos, L. G.; Gronenborn, A. M. Mini-Rev. Med. Chem. 2005, 5,
21-31.
Shenoy, S. R.; Barrientos, L. G.; Ratner, D. M.; O'Keefe, B. R.;
Seeberger, P. H.; Gronenborn, A. M.; Boyd, M. R. Chem. Biol. 2002,
9, 1109-1118.
Bewley, C. A.; Kiyonaka, S.; Hamachi, I. J. Mol. Biol. 2002, 322,
881-889.
Daniel, M.-C.; Astruc, D. Chem. Rev. 2004, 104, 293-346.
Matei, E.; Furey, W.; Gronenborn, A. M. Structure 2008, 16,
1183-1194.