1
Glycosylated Green Fluorescent Protein for Carbohydrate
Binding Protein Analysis
A thesis submitted to The University of Manchester for the degree of
Doctor of Philosophy
in the Faculty of Engineering and Physical Sciences
2013
Andrew James Martin
School of Chemistry
2
CONTENTS
LIST OF TABLES AND FIGURES ................................................................................. 7
ABBREVIATIONS ........................................................................................................ 18
EXPLANATORY NOTES ............................................................................................. 21
ABSTRACT .................................................................................................................... 22
DECLARATION ............................................................................................................ 23
COPYRIGHT STATEMENT ......................................................................................... 24
ACKNOWLEDGEMENTS ............................................................................................ 25
Chapter 1: Introduction ................................................................................................... 26
1.1 Protein Glycosylation ............................................................................................ 27
1.1.1 Types of Protein Glycosylation ...................................................................... 28
1.1.2 N-Linked Glycosylation ................................................................................. 29
1.1.3 O-Linked Protein Glycosylation .................................................................... 32
1.1.4 Carbohydrate Mediated Signalling ................................................................ 36
1.2 Medical Applications of Carbohydrates ............................................................... 36
1.2.1 Carbohydrate Based Antibiotics .................................................................... 37
1.2.2 Carbohydrate Based Vaccines........................................................................ 38
1.2.3 Carbohydrates for Cell Specific Drug Delivery ............................................. 38
1.3 Carbohydrate Binding Protein Analysis (Considerations) .................................... 41
1.3.1 Polyvalency .................................................................................................... 41
1.3.2 Heterogeneity ................................................................................................. 43
1.3.3 Synthetic Glycoconjugate Scaffolds .............................................................. 44
1.4 Artificial Glycoproteins ........................................................................................ 46
1.4.1 Synthetic Strategies for Neoglycoproteins ..................................................... 47
1.4.2 Chemical Glycosylation ................................................................................. 50
3
1.4.3 Glycosylating Cysteines ................................................................................. 53
1.5 Green Fluorescent Protein (GFP) .......................................................................... 55
1.5.1 GFPuv ............................................................................................................ 58
1.5.2 Glycosylated GFP .......................................................................................... 59
1.6 Project Aims .......................................................................................................... 60
Chapter 2: The Generation, Expression and Purification of GFPuv Mutants ................. 62
2.1 Generating GFPuv Cysteine Mutants.................................................................... 63
2.1.1 Addition of Hexahistidine Tag to GFPuv ...................................................... 64
2.2 Site Directed Mutagenesis..................................................................................... 65
2.2.1 Inverse PCR ................................................................................................... 65
2.2.2 The Quickchange Method .............................................................................. 67
2.3 Generation of Cysteine Mutants by DNA Shuffling ............................................. 68
2.3.1 General Considerations .................................................................................. 68
2.3.2 Design of Polycysteine Mutants for DNA Shuffling ..................................... 69
2.3.3 DNA Shuffle of Shuffle 1 and sGFPuv_C48A .............................................. 70
2.4 Expression of GFPuv ............................................................................................ 73
2.4.1 Optimisation of Protein Expression ............................................................... 73
2.4.2 Expression of Mutants ................................................................................... 76
2.5 Purification of GFPuv ........................................................................................... 77
2.5.1 Gradient Immobilised Metal Affinity Chromatography (IMAC) .................. 77
2.5.2 Anion Exchange Column ............................................................................... 77
2.5.3 Size Exclusion Chromatography .................................................................... 78
2.5.4 Stepwise IMAC .............................................................................................. 78
2.6 Summary ............................................................................................................... 80
Chapter 3: Synthesis of Aminoethyl Glycosides ............................................................ 81
4
3.1 General Considerations ......................................................................................... 82
3.2 Synthesis of Monosaccharides .............................................................................. 84
3.3 Synthesis of Aminoethyl Trimannoside (41) ....................................................... 86
3.4 Synthesis of Tetramannosides (48 and 49) and Pentamannoside (50) .................. 93
3.5 Activation of Glycosides for Glycosylation of Cysteines ..................................... 95
3.6 Summary ............................................................................................................... 96
Chapter 4: The Glycosylation of GFPuv Mutants .......................................................... 98
4.1 General Considerations ......................................................................................... 99
4.2 Analysis of Protein samples by ESI-MS ............................................................... 99
4.3 Cysteine Reactivity Screen ................................................................................. 104
4.4 Final Glycosylation Procedure ............................................................................ 106
4.5 Analysis of Protein Samples by LCMS .............................................................. 107
4.6 Production of Neoglycoprotein Library .............................................................. 109
4.7 Glycosylation of Lysines..................................................................................... 111
4.8 Summary ............................................................................................................. 113
Chapter 5: The Enzymatic Modification of Glycosides ................................................ 115
5.1 Glycotransferase Screening on Trimannoside (41) ............................................. 116
5.1.1 Screening of Mannosides Against Glycotransferases .................................. 116
5.1.2 Screening of Mannosides Against Yeast Microsomal Extracts ................... 123
5.2 Modification of Lactosylated GFPuv Using Tran-sialidase ............................... 124
5.3 Summary ............................................................................................................. 128
Chapter 6: Lectin Binding Assays................................................................................. 129
6.1 General Considerations ....................................................................................... 130
6.2 Fluorescence Based Plate Assay ......................................................................... 130
6.2.2 Lectins Chosen for Initial Screens ............................................................... 131
5
6.2.3 Plate Assay Results ...................................................................................... 133
6.2.4 Fluorescence Based Assay Summary........................................................... 138
6.3 Isothermal Titration Calorimetry (ITC) .............................................................. 140
6.3.1 Titration of Me-α-Man Against ConA ......................................................... 142
6.3.2 Titration of GFPuv Against ConA ............................................................... 144
6.3.3 ITC Summary ............................................................................................... 146
6.4 Summary ............................................................................................................. 147
Chapter 7: Conclusions and Future Experiments .......................................................... 148
7.1 Conclusions ......................................................................................................... 149
7.2 Future Work ........................................................................................................ 151
Chapter 8: Experimental Details ................................................................................... 153
8.1 Experimental Details for Chapter 2..................................................................... 154
8.1.1 General Methods .......................................................................................... 154
8.1.2 Production of GFPuv Mutant Library .......................................................... 158
8.1.3 Protein Expression and Purification ............................................................. 166
8.2 Experimental Details for Chapter 3..................................................................... 169
8.2.1 General Procedure 1: Peracetylation with Acetic Anhydride and Pyridine177
............................................................................................................................... 169
8.2.2 General Procedure 2: Deacetylation with Sodium Methoxide177
................. 170
8.2.3 General Procedure 3: Hydrogenolysis of N-Cbz-protecting Groups177
....... 170
8.2.4 Synthesis of Aminoethyl Mannoside (27)177
............................................... 170
8.2.5 Synthesis of Aminoethyl Glucoside (32)177
................................................. 174
8.2.6 Synthesis of Aminoethyl Galactoside (33)177
.............................................. 176
8.2.7 Synthesis of Aminoethyl N-Acetyl glucosamine (34)177
............................. 179
8.2.8 Activation of Glycosides for Glycosylation of Cysteines130
........................ 182
6
8.2.9 Synthesis of Polymannosides178
................................................................... 182
8.3 Experimental Details for Chapter 4..................................................................... 192
8.3.1 Glycosylation of GFPuv Mutants ................................................................ 192
8.3.2 MS Analysis of Proteins .............................................................................. 193
8.4 Experimental Details for Chapter 5..................................................................... 194
8.4.1 Enzymatic Screening of Mannosides ........................................................... 194
8.4.2 Transialidase (TcTs) Reactions .................................................................... 196
8.5 Experimental Details for Chapter 6..................................................................... 196
8.5.1 Lectin 96-Well Plate Assay .......................................................................... 197
8.5.2 ITC Measurements ....................................................................................... 197
REFERENCES .............................................................................................................. 198
APENDICIES ............................................................................................................... 208
Appendix 1: DNA Sequences of GFPuv_WT and GFPuv_C48A_I229C ................ 208
Appendix 2: The DNA Sequences of sGFPuv_C48A and Shuffle 1 ........................ 209
Appendix 3: Screen Capture of a Typical Stepwise IMAC GFPuv_WT Purification
................................................................................................................................... 210
Appendix 4: The Fluorescence Spectra of GFPuv Mutants ...................................... 211
Appendix 5: Screen Capture of a Typical Polymannoside Purification ................... 212
Appendix 6: HSQC-TOCSY of Trimannoside (43).................................................. 213
Appendix 7: HMBC of Trimannoside (43) ............................................................... 214
7
LIST OF TABLES AND FIGURES
Figure 1.1 The structures of asparagine N-linked to GlcNAc (1) and serine O-linked to
GalNAc(2). ...................................................................................................................... 28
Figure 1.2 Proposed mechanism for amide activation towards glycans.15
..................... 29
Figure 1.3 N-linked glycosylation pathway of a correctly folded protein. Solid lines
represent the reactions common to all N-linked glycoproteins. Dashed lines show one of
the many possible routes through the Golgi, producing a complex N-linked glycan.12
. 31
Figure 1.4 Examples of N-linked glycan types: high mannose (3), complex (4) and
hybrid (5). ........................................................................................................................ 32
Figure 1.5 Examples of the initial steps in mucin type O-linked glycan synthesis. The
Tn-antigen (6) can be converted to sialyl Tn-antigen (7) by the enzyme ST6GalNAc.
Alternatively the Tn-antigen can be converted to mucin core 1 (8) or core 3 (9)
structures by enzymes C1GalT and C3GnT respectively. .............................................. 34
Figure 1.6 The 8 core structures of mucin type, O-linked glycans. ............................... 35
Figure 1.7 Structures (10-12) of O-linked glycans identified in mouse colon tissue.32
. 35
Figure 1.8 The structure of Urdamycin A (13). Carbohydrate components are shown in
red. ................................................................................................................................... 37
Figure 1.9 The structure of an anti ovarian cancer vaccine. Carbohydrate component is
shown in red. MUC5AC corresponds to a peptide linker. KLH = Keyhole limpet
hemocyanin, a commonly used immunogenic protein.46
................................................ 38
Figure 1.10 An example of lectin-directed enzyme-activated therapy (LEAPT). In this
case the enzyme is rhamnosidase and the drug released is 5-flurouracil (15).56
............. 40
Figure 1.11 The structure of mannose-6-phosphate (16). .............................................. 40
Figure 1.12 Examples of the polyvalent presentation of lectins and carbohydrates. A)
Cell-cell/cell-surface interaction. B) The separation of glycoproteins by lectin affinity
chromatography............................................................................................................... 42
8
Figure 1.13 Some of the most widely used classes of glycoconjugate scaffold. A)
Linear glycopolymers. B) Glycodendrimers. C) Glycosylated nanoparticles. D)
Neoglycoproteins. Green circles represent glycans. ....................................................... 45
Figure 1.14 Chemical structures of glucoside (17) and galactoside (18). Hydroxyl group
attached to C4 shown in red. ........................................................................................... 46
Figure 1.15 The native chemical ligation (NCL) of two synthetic peptides to form a
natural peptide bond. ....................................................................................................... 48
Figure 1.16 General scheme for the native chemical ligation of peptides to form natural
peptide linkages that do not contain cysteine. ................................................................. 49
Figure 1.17 Examples of reagents which can be used to modify lysine side chains and
their products. .................................................................................................................. 51
Figure 1.18 The bioorthogonal reaction of a ketones with aminooxy compounds or
hydrazine compound to form stable oxime or hydrazone linkages respectively. ........... 52
Figure 1.19 Examples of “click” chemistry cycloaddition reactions between azides and
alkenes or alkynes. .......................................................................................................... 52
Figure 1.20 Examples of unnatural amino acids suitable for “click” chemistry reactions;
azidohomoalanine (19), p-Azido-L-phenylalanine (20), homopropargylglycine (21),
homoallylglycine (22) and two examples of UAAs utilising strain promoted reaction
technology (23 and 24).................................................................................................... 53
Figure 1.21 Summary of commonly used methods of glycosylating cysteines.125
........ 54
Figure 1.22 The formation of GFP’s fluorophore. Fluorophore shown in red............... 55
Figure 1.23 Ribbon diagram of GFP showing the β sheets in yellow, α helix in red,
fluorophore in blue and loops in green. .......................................................................... 56
Figure 1.24 Ribbon diagram of the crystal structure of GFPuv with three amino acids
(E6, C48 and I229) residues highlighted (in yellow). ..................................................... 58
Figure 2.1 Binding of a hexahistidine tagged protein to an immobilised metal affinity
column. ............................................................................................................................ 63
Figure 2.2 Schematic representation of the cloning of GFPuv into a pET-30a vector. . 64
9
Figure 2.3 The additional amino acid sequence introduced on to the N-terminus of
GFPuv. Hexahistidine tag is shown in red and the GFPuv section is shown in green. .. 65
Figure 2.4 Schematic representation of the inverse PCR method for the C48A mutation.
......................................................................................................................................... 65
Figure 2.5 Comparison of the products of the inverse PCR (Inv 1 and Inv 2) with the
GFPuv sequence around the C48A mutation site. Repeat units of the forward primer
used in the PCR are highlighted in green. ....................................................................... 66
Figure 2.6 Schematic representation of the Quickchange site directed mutagenesis
method for the C48A mutation. ...................................................................................... 67
Figure 2.7 Schematic representation of how homologous genes can be combined in a
primerless PCR to create new chimeric genes for screening. ......................................... 68
Figure 2.8 Amino acid sequence of the sGFPuv_C48A gene purchased for DNA
shuffling experiments. Amino acids highlighted in red correspond to the residues
exchanged for cysteine in the Shuffle 1 gene.................................................................. 70
Figure 2.9 Example expression plate containing transformants from DNA shuffle
products. Highlighted section on the left is enlarged on the right. ................................. 70
Figure 2.10 Summary of the active, GFPuv mutants discovered by the DNA shuffling.
Each column corresponds to a new gene. White sections correspond to segments of
sGFP_C48A and the green sections correspond to segments of Shuffle 1. Naturally
occurring C48 has been highlighted yellow. The mutant names correspond to their
library designation. .......................................................................................................... 71
Figure 2.11 Summary of the 12 inactive GFPuv mutants screened. Each column
corresponds to a new gene. White sections correspond to segments of sGFPuv_C48A
and the green sections correspond to segments of Shuffle 1. Naturally occurring C48
has been highlighted yellow. The mutant names correspond to their library designation.
......................................................................................................................................... 73
Figure 2.12 SDS-PAGE gels of the samples taken from cultures expressing
GFPuv_WT at 22, 30 at 37°C. From left to right the lanes correspond to; protein ladder,
0 h, 1 h, 2 h, 3 h, 4 h, 5 h, 6 h, 7 h, 8 h and 24 h after induction..................................... 74
10
Figure 2.13 Western blots of the samples taken from cultures expressing GFP_WT at
22, 30 at 37°C. From left to right the lanes correspond to 0 h, 1 h, 2 h, 3 h, 4 h, 5 h, 6 h,
7 h, 8 h and 24h after induction....................................................................................... 75
Figure 2.14 The mass of GFPuv_WT purified relative to total cell mass and total
soluble protein produced. Data relating to 400 mL cultures grown at 37°C until an OD
of 0.7 was reached followed by protein expression at different temperatures. ............... 75
Figure 2.15 A) SDS gel of selected fractions collected after IMAC column. B) The 96-
deep well plate in which the samples were collected. C) The 96-deep well plate in which
the samples were collected illuminated by blue light and viewed through a light filter. 77
Figure 2.16 A) SDS-PAGE gel of fluorescent fractions collected from anion exchange
chromatography of GFPuv_WT. B) SDS-PAGE gel of fluorescent fractions collected
from size exclusion chromatography of GFPuv_WT. .................................................... 78
Figure 2.17 SDS of all mutants purified by stepwise IMAC purification. From left to
right the lanes contain 5 μg of mutants B10, C5, D1, C5, D1, D5, F1, F11, G1, G3, S6,
E6C and I229C. ............................................................................................................... 79
Figure 2.18 Deconvoluted MS of GFPuv_E6C_I229C from 10000-70000 Da after
stepwise IMAC. .............................................................................................................. 80
Figure 3.1 The use of aminoethyl mannoside (27) in carbohydrate arrays and in the
synthesis of glycopeptides. (a) The reaction of amino ethyl mannoside with an activated
array surface. (b) The conversion of amino ethyl mannoside in to an α-halo carbonyl
compound capable of reacting with thiols. (c) The reaction of the activated mannoside
(28) with a cysteine (25) containing peptide. .................................................................. 82
Figure 3.2 Synthesis of aminoethyl mannoside (27). (a) Ac2O in pyridine. (b) BnNH2 in
THF. (c) Cl3CCN, K2CO3 in Dichloromethane (DCM). (d) N-Cbz-aminoethanol,
TMSOTf in DCM. (e) NaOMe in MeOH. (f) Pd/C, H2 in MeOH. (g) N-Cbz-
aminoethanol, BF3.Et2O in DCM. ................................................................................... 84
Figure 3.3 Structures of aminoethyl glucose (32), aminoethyl galactose (33) and
aminoethyl GlcNAc (34). ................................................................................................ 84
11
Figure 3.4 Summary of glycosylation reactions performed on peracetylated
monosaccharides (30, 35-37). ......................................................................................... 85
Figure 3.5 Retrosynthetic analysis of trimannoside (41), demonstrating how it can be
synthesised from aminoethyl mannoside (42) and a mannosyl donor. ........................... 86
Figure 3.6 Structure of mannosyl acceptor (42) with carbon numbering labelled in red.
......................................................................................................................................... 87
Figure 3.7 The reaction of sodium periodate with a mixture of trimannosides (43-45) 88
Figure 3.8 The synthesis of acetobromo mannose (46) from peracetyl mannose (30). . 89
Figure 3.9 The reaction of aminoethyl mannoside (42) with acetobromo mannose (46)
to form a mixture of mannosides including the trimannoside (43). ................................ 89
Figure 3.10 Structure of trimannoside (43) with carbohydrate carbon atoms labelled. . 90
Figure 3.11 A section of a non-decoupled HSQC of trimannoside (43) showing the
coupling of the anomeric protons with their respective carbons..................................... 91
Figure 3.12 Section of a HMBC spectrum showing the coupling of C3 to H1’ in
trimannoside (43). ........................................................................................................... 91
Figure 3.13 The deacylation of trimannoside (43) to trimannoside (47) using sodium
methoxide in methanol. ................................................................................................... 92
Figure 3.14 The hydrogenation of trimannoside (47) to trimannoside (48) using a Pd/C
catalyst in water............................................................................................................... 92
Figure 3.15 Structures of tetramannosides (48 and 49) and pentamannoside (50). ....... 93
Figure 3.16 Conversion of aminoethyl mannoside (27) in to mannosyl iodoacetamide
(28) via reaction with iodoacetic anhydride in sodium bicarbonate buffer. ................... 95
Figure 3.17 Structure of aminoethyl lactose (51) donated by Dr R. Sardzik (The
University of Manchester)............................................................................................... 95
Figure 3.18 structures of glycosyl iodoacetamides produced; mannosyl iodoacetamide
(28), glucosyl iodoacetamide (52), galactosyl iodoacetamide (53), glucosamine
iodoacetamide (54) and lactosyl iodoacetamide (55)...................................................... 96
12
Table 3.1 Summary of HRMS data of glycosyl iodoacetamides produced; mannosyl
iodoacetamide (28), glucosyl iodoacetamide (52), galactosyl iodoacetamide (53),
glucosamine iodoacetamide (54) and lactosyl iodoacetamide (55). ............................... 96
Figure 4.1 The chemical glycosylation of a peptide with an iodoacetamide, under
conditions originally reported.181
.................................................................................... 99
Figure 4.2 A) The acquired spectrum of horse heart myoglobin (HHM), containing the
multiply charged protein peaks of the “charge envelope.” B) The deconvoluted
spectrum of HHM produced by MassLynx 4.0. ............................................................ 100
Table 4.1 Calculated and measure mass values for GFPuv mutants and horse heart
myoglobin (HHM). ....................................................................................................... 101
Figure 4.3 A) The ESI mass spectrum of GFPuv_E6C before to treatment with TCEP.
B) The ESI mass spectrum of the same sample of GFPuv_E6C after treatment with
TCEP. ............................................................................................................................ 103
Figure 4.4 The reaction of mutant GFPuv_E6C with iodoacetamide in ammonium
carbonate buffer. ........................................................................................................... 104
Figure 4.5 Deconvoluted mass spectra (31700-32100 Da) of samples taken from the
reaction of 0.1 mM GFPuv_E6C_I229C with 1 mM iodoacetamide (56). A) Mass
spectrum after 0 hours. B) Mass spectrum after 1 hour. C) Mass spectrum after 5 hours.
D) Mass spectrum after 24 hours. ................................................................................. 105
Table 4.2 Summary of the results of the reaction of four GFP mutants (0.1 mM) with
iodoacetamide (56) (1 mM) over 24 hours. In each case the mass corresponds to the
only significant peaks present in the mass spectra. ....................................................... 106
Figure 4.6 The finalised procedure for glycosylation of all GFPuv mutants. .............. 107
Figure 4.7 UV (205 nm) trace of a typical LCMS run of a GFPuv mutant. ................ 108
Table 4.3 Calculated and measure mass values for GFPuv mutants and HHM using an
Agilent 1100, HPLC system coupled to an Agilent 1100 LC/MSD SL quadrupole mass
spectrometer. ................................................................................................................. 109
Figure 4.8 A) Measured mass spectrum of GFPuv_C5. B) The deconvoluted molecular
ion peak of GFPuv_C5. B) Measured mass spectrum of GFPuv_C5 after reaction with
13
mannosyl iodoacetamide (28). D) The deconvoluted molecular ion peak of GFPuv_ C5
after reaction with mannosyl iodoacetamide (28). ........................................................ 110
Figure 4.9 The structure of 3,3 -dithiobis(sulfosuccinimidylpropionate) DTSSP.
Produced by ................................................................................................................... 111
thermo scientific as a reversible protein crosslinker. .................................................... 111
Figure 4.10 The reaction of GFPuv with DTSSP followed by the reduction of the
disulfides within the crosslinkers by TCEP to give thiol modified lysines. ................. 112
Figure 4.11 MALDI spectra of GFPuv_WT (blue), GFPuv_CL (GFPuv_WT after
treatment with DTSSP followed by reduction with TCEP) (red) and GFP_CL_Man10
(GFPuv_WT derivatised with approximately 10 mannosides) (green). ....................... 113
Figure 5.1 Structures of aminoethyl mannosides (27 and 41) used for glycotransferase
screening. ...................................................................................................................... 116
Figure 5.2 The natural action of N-acetylglucosaminyltransferase (GnT-I) on the N-
glycan core structure Man5GlcNAc2 (58)...................................................................... 117
Figure 5.3 The natural action of protein-O-mannose N-acetylglucosaminyltransferase I
(POMGnT-I) on an O-linked glycopeptides. ................................................................ 117
Figure 5.4 Known substrates for GnT-I (donated by Dr. S. Gluchowska, Trinity College
Dublin). Mannotriose-di-(N-acetyl-D-glucosamine) (60) and 3α,6α-mannotriose (61)
(commercially available from Sigma). .......................................................................... 118
Figure 5.5 MALDI-TOF spectrum of a 1:4 (linker:spacer) SAM on gold. A = mass
peak corresponding to a spacer-spacer homodimer. B = mass peak corresponding to a
spacer-linker heterodimer.............................................................................................. 119
Figure 5.6 The reaction of trimannoside (41) with an activated SAM on a gold plate to
form a carbohydrate array. ............................................................................................ 120
Figure 5.7 MALDI-TOF MS spectrum of trimannoside (41) carbohydrate array. A =
mass peak corresponding to a spacer-spacer homodimer. B = mass peak corresponding
to a spacer-linker heterodimer. C = peak corresponding to a heterodimer covalently
bound to trimannoside (41). .......................................................................................... 121
14
Figure 5.8 MALDI-TOF MS spectrum of mannoside (27) carbohydrate array. A = mass
peak corresponding to a spacer-spacer homodimer. B = mass peak corresponding to a
spacer-linker heterodimer. D = peak corresponding to a heterodimer covalently bound
to mannoside (27). ......................................................................................................... 122
Figure 5.9 A) MALDI-TOF spectrum of trimannoside (41) attached to a SAM on gold.
B) MALDI-TOF spectrum of trimannoside (41) attached to a SAM on gold after
treatment with GnT-I and UDP-GlcNAc. R= SAM spacer-linker hetrodimer. ........... 123
Figure 5.10 Structure of aminoethyl sialyllactose (63). ............................................... 125
Figure 5.11 The reaction of immobilised lactose with fetuin in the presence of trans-
sialidase (TcTs) enzyme to produce immobilised sialyllactose. ................................... 125
Figure 5.12 A) Deconvoluted mass sprectum of GFPuv_I229C. B) Deconvoluted mass
spectrum of GFPuv_I229C_Lac. .................................................................................. 126
Figure 5.13 Mass spectra of samples taken from the reaction of trans-sialidase (TcTs)
with GFPuv_I229C_Lac in the presence of fetuin. A) Reaction after 30 minutes. B)
Reaction after 1 hour. C) Reaction after 2.5 hours. ...................................................... 127
Figure 6.1 Schematic diagram of the lectin plate assay. The protein avidin (red) is
covalently bound to the surface of a 96-well plate, which enables the capture of
biotinylated lectins (blue). Glycosylated GFPuv mutants can then interact with the
immobilised lectins. ...................................................................................................... 131
Figure 6.2 The results of screening the interactions of unglycosylated, mannosylated
and galactosylated GFPuv mutants against streptavidin coated 96-well plates. Samples
are grouped according to their number of glycosylation sites. ..................................... 133
Figure 6.3 The results of screening the interactions of unglycosylated, mannosylated
and galactosylated GFPuv mutants against ConA coated 96-well plates. Samples are
grouped according to their number of glycosylation sites. ........................................... 135
Figure 6.4 The results of screening the interactions of unglycosylated, mannosylated
and galactosylated GFPuv mutants against GNL coated 96-well plates. Samples are
grouped according to their number of glycosylation sites. ........................................... 136
15
Figure 6.5 The results of screening the interactions of unglycosylated, mannosylated
and galactosylated GFPuv mutants against jacalin coated 96-well plates. Samples are
grouped according to their number of glycosylation sites. ........................................... 137
Figure 6.6 Ideal ITC plots. A) The raw data obtained from an ideal series of injections.
B) An ideal plot of the molar energy changes for a bivalent (N = 2) interaction. ........ 141
Figure 6.7 Equations for calculating Gibb’s free energy. R = The molar gas constant
8.314 J mol-1 K-1, T = The temperature in Kelvin. ..................................................... 141
Figure 6.8 Calorimetric data obtained from titration of native ConA (32 μM) with Me-
α-Man (5 mM). A) Raw data from 30 injections of 1 μL each of Me-α-Man into ConA.
B) Integrated curve showing the line of best fit. ........................................................... 143
Figure 6.9 Calorimetric data obtained from titration of native ConA (22 μM) with
GFPuv_C5_Man4 (240 μM). A) Raw data from 30 injections of 1 μL each of
GFPuv_C5_Man4 into ConA. B) Integrated curve showing the line of best fit. .......... 145
Figure 6.10 Raw calorimetric data from 30 injections of 1 μL each of: A) ITC buffer
into ITC buffer. B) GFPuv_C5 (240 μM) into ConA (22 μM). .................................... 146
Table 8.1 The PCR program used for in vitro DNA amplification. ............................. 157
Table 8.2 The PCR program used for site directed mutagenesis. ................................ 159
Table 8.3 The PCR program used for inverse PCR site directed mutagenesis. ........... 161
Table 8.4 The three genes (Shuffle 1-3) designed for DNA shuffle cysteine screen of
GFPuv. Numbers correspond to the amino acids to be substituted for cysteine in each
gene. .............................................................................................................................. 162
Table 8.5 The PCR program used for the DNA shuffle. .............................................. 164
Table 8.6 The PCR program used for the amplification of the DNA shuffle products.
....................................................................................................................................... 165
Figure 8.1 The reaction of mannose (29) with acetic anhydride to form peracetyl
mannose (30). ................................................................................................................ 170
Figure 8.2 The reaction of peracetylated mannose (30) with benzyl N-(2-hydroxyethyl)-
carbamate to form mannoside (31). .............................................................................. 171
16
Figure 8.3 The deprotection of mannoside (31) with NaOH and MeOH to give
mannoside (27). ............................................................................................................. 172
Figure 8.4 The hydrogenation of mannoside (42) using a Pd/C catalyst and hydrogen
gas to give mannoside (27). .......................................................................................... 173
Figure 8.5 The reaction of Peracetyl β-D-glucopyranose (36) and N-Cbz-ethanolamine
to produce glucoside (39). ............................................................................................. 174
Figure 8.6 The deprotection of glucoside (39) with NaOH in MeOH to produce
glucoside (64). ............................................................................................................... 175
Figure 8.7 The hydrogenation of glucoside (39) using a Pd/C catalyst and hydrogen gas
to produce glucoside (32). ............................................................................................. 175
Figure 8.8 The reaction of peracetyl β-D-galactose (37) with N-Cbz-ethanolamine to
produce galactoside (40). .............................................................................................. 176
Figure 8.9 The deprotection of galactoside (40) using NaOH in MeOH to form
galactoside (65). ............................................................................................................ 177
Figure 8.10 The hydrogenation of galactoside (65) using a Pd/C catalyst and hydrogen
gas to produce galactoside (33). .................................................................................... 178
Figure 8.11 The reaction of β-D-Glucosamine pentaacetate (35) with of N-Cbz-
ethanolamine to produce glucoside (38). ...................................................................... 179
Figure 8.12 The deprotection of glucoside (38) using NaOH and MeOH to produce
glucoside (66). ............................................................................................................... 180
Figure 8.13 The hydrogenation of glucoside (66) using a Pd/C catalyst and hydrogen
gas to produce glucoside (34). ...................................................................................... 181
Figure 8.14 Activation of aminoethyl mannoside (27), for reaction with cysteines, via
reaction with iodoacetic anhydride to produce glycosyl iodoacetamide (28). .............. 182
Figure 8.15 The reaction of peracetyl mannose (30) with HBr to produce acetobromo
mannoside (46). ............................................................................................................. 182
Figure 8.16 The reaction of acetobromo mannoside (46) and mannoside (42) to produce
trimannoside (43). ......................................................................................................... 183
17
Figure 8.17 The numbering scheme used for the assignment of NMR spectra of
trimannoside (43). ......................................................................................................... 185
Table 8.7 1H and
13C chemical shifts of the atoms found in the carbohydrate constituent
of trimannoside (43). Additional signals are listed below a long with the remaining
characterisation undertaken. .......................................................................................... 186
Figure 8.18 A section of a multiplicity edited HSQC of trimannoside (43). ............... 186
Figure 8.19 The deprotection of trimannoside (43) with NaOH and MeOH to produce
trimannoside (47). ......................................................................................................... 187
Figure 8.20 The hydrogenation of trimannoside (47) using a Pd/C catalyst and
hydrogen gas to produce trimannoside (41). ................................................................. 188
Figure 8.21 The structures of polymannoside side products 48, 48 and 50. ................ 189
18
ABBREVIATIONS
2-NBDG 2-(N-(7-Nitrobenz-2-oxa-1,3-diazol-4-yl)Amino)-2-deoxyglucose
ASGPR Asialoglycoprotein receptor
Aq Aqueous
BCA Bicinchoninic acid
C1GalT GalNAc 3-beta-galactosyltransferase
C3CnT GalNAc 3-beta-galactosyltransferase
ConA Concanavalin A
DCM Dichloromethane
DMSO Dimethyl sulfoxide
DTSSP 3,3 -dithiobis (sulfosuccinimidylpropionate)
DTT Dithiothreitol
EDC N-ethyl-N’-(dimethylaminopropyl)-carbodiimide
ER Endoplasmic reticulum
ERAD Endoplasmic-reticulum-associated protein degradation
ESI Electrospray ionisation
d.p. Decimal place
Fuc Fucose
Gal Galactose
GalNAc N-Acetylgalactosamine
GFP Green fluorescent protein
GFPuv Green fluorescent protein (ultraviolet)
19
Glc Glucose
GlcNAc N-Acetylglucosamine
GNL Galanthus nivalis lectin
GnT-1 N-Acetylglucosaminyltransferase I (GnT-I)
GPI Glycosyl phosphatidyl inositol
HHM Horse heart myoglobin
HIV Human immunodeficiency virus
HMPT Hexamethylphosphorous triamide
HPLC High performance liquid chromatography
IMAC Immobilised metal affinity chromatography
IPTG Isopropyl β-D-1-thiogalactopyranoside
ITC Isothermal titration calorimetry
KLH Keyhole limpet hemocyanin
Lac Lactose
LCMS Liquid chromatography–mass spectrometry
LEAPT Lectin-directed enzyme-activated therapy
Lit. Literature value
MALDI-TOF Matrix-assisted laser desorption/ionisation-time of flight
Man Mannose
Mol eq Mole equivalents
MS Mass spectrometry
MSH O-mesitylenesulfonylhydroxylamine
20
Neu5Ac N-acetylneuraminic acid
NCL Native chemical ligation
NHS N-hydroxysuccinimide
ON Over night
PCR Polymerase chain reaction
PBS Phosphate buffered saline
PEG Polyethylene glycol
POMGnT-1 Protein-O-mannose N-acetylglucosaminyltransferase I
r.t. Room temperature
SAM Self assembled monolayer
Sat. Saturated
SDS-PAGE Sodium dodecyl sulfate polyacrylamide gel electrophoresis
Siglecs Sialic acid-binding immunoglobulin-type lectins
ST6GalNAc α-2-6-sialyltransferase
TCEP Tris(2-carboxyethyl)phosphine
TcTs Transialidase
TOF Time of flight
UAA Unnatural amino acids
UDP Uridine diphosphate
UV Ultra violet
WT Wild type
21
EXPLANATORY NOTES
Mutant List
WT = GFPuv with N-terminal hexahistidine tag (NB all following mutants were derived
from this construct)
E6C = E6C
C48A = C48A
I229C = I229C
E6C_C48A = E6C, C48A
E6C_I229C = E6C, I229C
E6C_C48A_I229C = E6C, C48A, I229C
B10 = C48A, S202C, N212C, I229C
C5 = S30C, T38C, T43C, K52C
D1 = C48A, I229C
D4 = C48A, I229C
D5 = T38C, T43C
F1 = C48A, S202C
F11 = K52C
G1 = C48A, L221C, I229C
G3 = C48A, N105C, I188C
S6 = L15C, T38C, T43C, C48A, K52C
C5+2 = E6C, S30C, T38C, T43C, K52C, I229C
CL = WT modified with the cross linker DTSSP
22
ABSTRACT
The interactions of glycoconjugates with carbohydrate binding proteins are responsible
for a wide range of recognition events in vivo; including immune response, cell
adhesion and signal transduction. Glycoconjugates have already found many medicinal
uses as therapeutic and diagnostic agents, but their full potential is yet to be realised.
Access to a variety of homogeneously glycosylated glycoproteins is essential for the
study of these important carbohydrate binding events. This requires the chemical
synthesis and attachment of biologically relevant glycans to unglycosylated protein
scaffolds in a site selective manner.
Here we describe the use of a range of glycosyl iodoacetamides to glycosylate proteins
selectively via their cysteine residues. We have chosen the green fluorescent protein
mutant GFPuv for use as a protein scaffold due its known tolerance of two cysteine
mutations (E6C and I229C) and the previous successful derivatisation of these cysteines
with iodoacetamides.1 The inherent fluorescence of GFPuv also makes it a useful
candidate for fluorescence based binding assays or cell labelling studies.
16 active, mutants of GFPuv were created using a mixture of site directed mutagenesis
and DNA shuffling (including one mutant containing six reactive cysteine residues).
This was achieved by producing random combinations of two synthetic variants of
GFPuv, one of which contained 33 surface cysteines. 94 bacterial colonies expressing
active GFPuv were then sequenced and the new chimeric genes analysed.
Four monosaccharides and one trisaccharide (N-glycan core mimic) suitable for the
chemical glycosylation via cysteines were synthesised and successfully used to create a
selection of homogeneous neoglycoproteins. These neoglycoproteins were
demonstrated to interact differently with different lectins (including ConA, GNL and
Jacalin) in a qualitative fluorescence based assay. Interactions were shown to vary with
glycan structure, position of glycosylation sites and the number of glycosylation sites.
23
DECLARATION
I hereby declare that no portion of the work referred to in the thesis has been submitted
in support of an application for another degree or qualification of this or any other
university or other institute of learning.
24
COPYRIGHT STATEMENT
i. The author of this thesis (including any appendices and/or schedules to this
thesis) owns certain copyright or related rights in it (the “Copyright”) and s/he
has given The University of Manchester certain rights to use such Copyright,
including for administrative purposes.
ii. Copies of this thesis, either in full or in extracts and whether in hard or
electronic copy, may be made only in accordance with the Copyright, Designs
and Patents Act 1988 (as amended) and regulations issued under it or, where
appropriate, in accordance with licensing agreements which the University has
from time to time. This page must form part of any such copies made.
iii. The ownership of certain Copyright, patents, designs, trade marks and other
intellectual property (the “Intellectual Property”) and any reproductions of
copyright works in the thesis, for example graphs and tables (“Reproductions”),
which may be described in this thesis, may not be owned by the author and may
be owned by third parties. Such Intellectual Property and Reproductions cannot
and must not be made available for use without the prior written permission of
the owner(s) of the relevant Intellectual Property and/or Reproductions.
iv. Further information on the conditions under which disclosure, publication and
commercialisation of this thesis, the Copyright and any Intellectual Property
and/or Reproductions described in it may take place is available in the
University IP Policy (see http://documents.manchester.ac.uk/
DocuInfo.aspx?DocID=487), in any relevant Thesis restriction declarations
deposited in the University Library, The University Library’s regulations (see
http://www.manchester.ac.uk/library/aboutus/regulations) and in The
University’s policy on Presentation of Theses
25
ACKNOWLEDGEMENTS
Firstly I would like to thank my PhD supervisor Prof. Sabine L. Flitsch for her support,
guidance and inspiration over the past four years. I would also like to thank all members
of the Turner and Flitsch research groups for their help and advice in the lab. Special
thanks to Dr Robert Sardzik for his invaluable assistance with the chemical aspects of
this project and to Prof Josef Vogelmiere for his crash course in molecular biology.
Many thanks also to D. Adri Botes, Oxyrane and the EPSRC funding council for
making this project possible.
I am very grateful for the technical expertise and help I have received from Reynard
Spiess (HRMS and Protein ESI-MS) and Matthew Cliff (NMR).
I would also like to thank my family for their unwavering support and encouragement
over the last four years. As well as providing a much needed sense of perspective
throughout.
Finally many thanks to my wonderful wife Dr. Hannah Reed for her patience, love and
support throughout the last several years. It is difficult to imagine how much more of a
challenge it would have been without her.
26
Chapter 1: Introduction
CHAPTER 1
27
1.1 Protein Glycosylation
Glycosylation has been known to play a vital role in protein function for decades, but
for some time the reasons for these carbohydrate derivatisations was the subject of
debate. Some theories suggested protein glycosylation function was predominantly for
structural or stability enhancement, while others focused more on immunological and
cell signalling roles. In 1965 Edwin H. Eylar hypothesised that glycosylation played a
role in protein trafficking and that the primary structure of the protein would determine
the glycosylation sites.2 This hypothesis, like the many other suggestions for the
function of protein glycosylation is now known to be true. The remarkable versatility of
the carbohydrates and their biological roles was aptly summarised in the title of a
review by Ajit Varki in 1993; “All of the theories are correct”.3 This review is currently
the most citied article from Glycobiology and concludes that it is unlikely that a theory
for predicting any given glycan’s function, purely from its structure is possible; because
it is likely to have multiple functions and that these may alter according to its biological
environment.
Glycosylation is now known to play a vast number of roles, from altering
physiochemical properties of proteins such as conformation, solubility and stability, to
more sophisticated interactions such as protein trafficking, cell signalling and immune
response.3,4
A range of extremely specialised functions of glycoconjugates have also be
discovered including heavily glycosylated proteins, which prevent the nucleation of ice
in some arctic fish. This biological antifreeze allows them to survive in waters as cold
as -2°C without cellular damage.5 A layer of glycans is also crucial in the protection of
lysosomal membrane proteins such as LAMP-1 and LAMP-2, from degradation by the
hydrolytic enzymes they contain.6
Examples such as these demonstrate the extreme
versatility of glycoconjugation in nature and there have been several reviews in this area
over the years.4,7,8
An increasing amount of research is being dedicated to carbohydrate
synthesis and carbohydrate-binding protein interactions and subsequently the
importance of the field of glycobiology continues to grow.
CHAPTER 1
28
1.1.1 Types of Protein Glycosylation
In nature, carbohydrates (or glycans) can be attached to proteins in a number of ways.
Random glycation can occur if the concentrations of reducing sugars become too high,
however this is only significant in certain disease states, such as diabetes.9 Glycans are
usually attached to proteins enzymatically, which can occur in one of three ways; N-
linked, O-linked or attached to the C-terminus. This C-terminus linkage is used to attach
a glycosyl phosphatidyl inositol (GPI) anchor to proteins. These glycolipids anchor their
proteins to cell membranes. However, as GPI anchors are not linked to proteins via
glycosidic bonds these are generally not considered to be a true glycosylation.6,10
N-linked proteins are linked to their glycans through the nitrogen of the amide side
chain of asparagine. This linkage is always made to N-acetylglucosamine (GlcNAc)
which is then linked to the rest of the glycan. The structure of GlcNAc N-linked to an
asparagine (1) is shown in figure 1.1. O-linked proteins are linked to their glycans
through the oxygen atom of a hydroxyl group of serine, threonine, hydroxylysine or
hydroxyproline. This linkage can be made with a range of different monosaccharides
but is often to N-acetylgalactosamine (GalNAc). The structure of serine O-linked to
GalNAc (2) is also shown in figure 1.1.6 About 90% of known glycoproteins carry N-
linked glycans, many of these proteins also feature some O-linked glycans and only
10% of known glycoproteins carry purely O-linked glycans.11
Large glycans can extend
over 3 nm from their protein’s surface and effectively act as separate domains, whilst
others work together with their proteins to interact with other biomolecules.12
Figure 1.1 The structures of asparagine N-linked to GlcNAc (1) and serine O-linked to GalNAc(2).
CHAPTER 1
29
1.1.2 N-Linked Glycosylation
Despite the incredible diversity in N-linked glycan structures, they all share some
common features due to their synthetic origin. Initially a 14-saccharide core, containing
three glucose (Glc), nine mannose (Man) and two GlcNAc monomers is synthesised as
a precursor attached to the endoplasmic reticulum (ER) membrane. As proteins are
formed in the lumen of the ER an oligosaccharyltransferase attaches this core structure
to any Asn-X-Ser/Thr (X = any amino acid except Pro) sequon it recognises.13,14
As
well as recognising this relatively short sequon the active site of this enzyme contains
several negatively charged amino acid side chains. These side chains are believed to
hold the amide group of asparagine in such a position that the lone pair of the nitrogen
can no longer conjugate with the neighbouring carbonyl group (figure 1.2).15
This
would allow it to perform a nucleophilic attack on GlcNAc, which would not otherwise
be possible.
Figure 1.2 Proposed mechanism for amide activation towards glycans.15
This family of enzymes, unlike many other oligosaccharyltransferases, has a very broad
specificity towards different polypeptide substrates so these core N-linked glycans are
attached to the large majority of proteins as they synthesised. However this N-
glycosidic bond formation is only 90% efficient, which leads to some heterogeneity in
the population of glycosylated products or glycoforms of the same protein.16
CHAPTER 1
30
After the initial addition of the 14-saccharide core structure two terminal Glc units are
sequentially removed by ER glucosidases I and II respectively as shown in figure 1.3.
The resulting mono glucosylated glycan is recognised by the protein chaperones,
calnexin and calreticulin, which aid protein folding. Glucosidase II then removes the
one remaining Glc unit leaving a high Man glycan, which is recognised by ER
glucosyltransferase. This enzyme is able to recognise proteins folded incorrectly or
incompletely and will reattach a Glc unit if this is the case, preventing it from leaving
the calnexin-calreticulin cycle.14,17
Proteins that fail to fold correctly are removed at this point and recycled via the
Endoplasmic-reticulum-associated protein degradation (ERAD) patway.18
ER-
mannosidase I removes the α(1-2) linked Man from the α(1-3) branch of core N-linked
glycans to leave a 10-saccharide (eight Man and two GlcNAc monomers) structure
shown in figure 1.3. This enzyme acts on most glycans before leaving the ER but is also
thought to be part of the signal for ERAD. Due to its relatively low rate of activity it
could limit the amount of times a protein can participate in the calnexin-calreticulin
cycle before it is recycled. To ensure only unwanted peptides are digested they are first
removed to the cystol to be broken down by 26S proteasomes.14,19
CHAPTER 1
31
Figure 1.3 N-linked glycosylation pathway of a correctly folded protein. Solid lines represent the
reactions common to all N-linked glycoproteins. Dashed lines show one of the many possible routes
through the Golgi, producing a complex N-linked glycan.12
Correctly folded proteins are then transported to the Golgi apparatus where more Man
trimming occurs along with many other modifications that lead to the variety of
different N-linked glycan structures which are present throughout the rest of the cell.
These structures can include a large variety of different monosaccharides including:
galactose (Gal), fucose (Fuc), and N-acetylneuraminic acid (Neu5Ac). Glycoproteins
can spend up to 15 minutes travelling through the whole stack of Golgi towards the cell
surface.12
Unlike the ER, all of the Golgi’s enzymes are membrane bound which
enables a range of different glycosylation pathways to occur simultaneously.20
Mature
N-linked glycans are divided into three different groups; high mannose, complex and
hybrid. Examples of these three classes of N-linked glycans are shown in figure 1.4
(structures 3-5 respectively).6
CHAPTER 1
32
N-linked glycosylation plays a crucial role in the folding process of large proteins,
allowing structures that would not otherwise be stable to form.21
The direct effect of N-
linked glycosylation is to favour certain conformations in the nearby peptide chain
which often induces the formation of a β-turn.22
A third of all N-linked glycosylation
sites occur on β-turns, which suggests that this is an important function.23,24
The polar
nature of glycans solubilises proteins, which helps to prevent aggregation of non folded
proteins and also orientates unfolded segments towards the rest of the protein.12
Figure 1.4 Examples of N-linked glycan types: high mannose (3), complex (4) and hybrid (5).
1.1.3 O-Linked Protein Glycosylation
O-linked glycosylation often occurs in the Golgi, when proteins are already folded. The
biosynthesis of O-linked glycans is quite different from N-linked glycans and less well
understood. Often glycosylation sites are clustered together in contrast to N-linked sites,
which are usually dispersed. Also unlike N-glycosylation there is no universal amino
acid sequon for O-glycosylations making it very difficult to predict where they will
occur.6,25
Statistical studies have come up with some general rules for when O-linked
glycosylation may occur:26
1. O-linked glycosylation sites are tissue specific due to different isoforms of
glycotransferases being present.
CHAPTER 1
33
2. Only exposed residues can be glycosylated by these enzymes because it is a post
folding event.
3. Threonines are more likely to be glycosylated than serines.
4. Regions rich in proline and valine are more likely to be glycosylated.
Many other observations have been made; for example tryptophan is never found
adjacent to an O-linked glycosylation site and proline is often found at positions –1
and +3 in relation to O-linked glycosylation sites. All of these observations have been
used to develop computer programs that are making the prediction of O-linked
glycosylation sites increasingly accurate.26
Some tissue specific sequences have been
identified and the enzymes involved in O-linked glycosylations are generally more
protein specific than with N-linked glycosylations. For example O-linked Fuc in the
epidermal growth factor domains occurs at a serine or threonine in the sequence -Cys-
X-X-Gly-Gly-Thr/Ser-Cys- and O-linked Gal in collagen occurs on a hydroxylysine in
the sequence –Gly-Xaa-Hyl-Gly.27,28
Unlike N-linked glycans which all begin as a 14-saccharide core structure, O-linked
glycans are built sequentially on their proteins, one monomer at a time. The most
common type of O-linked glycosylation in higher eukaryotes are known as mucin type
glycans, which were originally found on mucin proteins but can also occur on other
proteins. Mucin type glycans are characterised by beginning with a GalNAc linked to a
serine or threonine. In other non-mucin type O-linked glycosylations the
monosaccharide linking the glycan to its protein can also be Fuc, Man, Glc, GlcNAc,
xylose or Gal.25,28
In the biosynthesis of mucin type glycans GalNAc is attached to the protein by a family
of enzymes known as polypeptide N-acetylgalactosamine transferases, which use
uridine diphosphate (UDP)–GalNAc as its source of monosaccharide.29
This GalNAc
residue attached to a peptide (6) is known as the Tn-antigen and is the starting point for
CHAPTER 1
34
all mucin type glycans (figure 1.5). The Tn-antigen can be sialylated by GalNAc α-2-6-
sialyltransferase (ST6GalNAc), to produce sialyl Tn-antigen (7). Alternatively the Tn-
antigen could be elongated by GalNAc 3-beta-galactosyltransferase (C1GalT) or
GalNAc β-1,3-N-acetylglucosaminyltransferase (C3GnT) to produce the core 1 (8) or
core 3 (9) mucin structures respectively (figure 1.5).
Figure 1.5 Examples of the initial steps in mucin type O-linked glycan synthesis. The Tn-antigen (6) can
be converted to sialyl Tn-antigen (7) by the enzyme ST6GalNAc. Alternatively the Tn-antigen can be
converted to mucin core 1 (8) or core 3 (9) structures by enzymes C1GalT and C3GnT respectively.
There are eight core mucin structures that have been identified and these are shown in
figure 1.6.30,31
These core structures can then further elaborated to produce a huge
variety of mucin type glycans. Three examples (10-12) of complex mucin type glycans
detected in colon tissue form mice are shown in figure 1.7.32
CHAPTER 1
35
Figure 1.6 The 8 core structures of mucin type, O-linked glycans.
Figure 1.7 Structures (10-12) of O-linked glycans identified in mouse colon tissue.32
Originally mucin glycans were thought to have non-specific roles in protecting the
gastrointestinal and respiratory tracts, maintaining viscoelasticity, hydrodynamic
protease resistance and pH buffering. Now they are known to be involved in many other
CHAPTER 1
36
important functions including; recognition, protein trafficking and modulation of
protein function.6,33
They also act as part of the innate immune system by binding the
lectins of microorganisms, which helps to protect areas such as the gastrointestinal and
respiratory tracts.34,35
1.1.4 Carbohydrate Mediated Signalling
Perhaps the most impressive function of carbohydrates in nature is their ability to act as
recognition domains that can be modified without altering the structure of the cell
component to which they are attached. This can modulate interactions with other
biomolecules, providing an effective and compact communication mechanism. For
example if a mannose attached to a newly synthesised glycoprotein is phosphorylated
by N-acetylglucosamine-1-phosphotransferase the glycoprotein will be transported to a
lysosome, unlike the majority of glycoproteins, which will pass through the Golgi
apparatus.12,36
Carbohydrate recognition plays a role in almost all bodily recognition
functions including immune response, fertilisation and blood group determination.37,38
The potential benefits and applications possible from understanding carbohydrate
recognition interactions are enormous. Thousands of glycan structures are known, but
most of these have features in common and a relatively small number of
monosaccharides are used in their assembly. In many cases such as the calnexin-
calreticulin cycle or the asialoglycoprotein receptor (ASGPR) in hepatocytes, the
difference of only one monosaccharide can determine the fate of a glycoprotein. This
begs the question: exactly how much of a complex glycan is necessary for recognition
and how much is due to its biosynthetic origins? A greater understanding of these
recognition processes could benefit several aspects of biotechnology.
1.2 Medical Applications of Carbohydrates
Proteins constitute a large proportion of candidates for new therapeutic and diagnostic
agents; for example hormones, enzymes, clotting factors, growth factors and
CHAPTER 1
37
monoclonal antibodies. Derivatisation of these therapeutic and analytical agents can
often improve or modulate activity just as with natural proteins. Carbohydrates are also
increasingly found in small molecule therapeutic agents such as antibiotics. The more
that can be understood about natural carbohydrate interactions, the wider the range of
potential benefits that can be achieved.
1.2.1 Carbohydrate Based Antibiotics
Glycan modification of antibiotics such as vancomycin is seen as one way in which
medicine can combat the rapidly increasing antibiotic resistance of some
microorganisms. Many antibiotics contain carbohydrate components and altering this
aspect has been demonstrated to be effective in circumventing resistance in some
cases.39
Naturally occurring antibiotics such as Urdamycin A (13) (figure 1.8) often
contain carbohydrate components, which are essential for their activity. Structures like
these have provided inspiration for many antibacterial and antitumour agents.40,41
Figure 1.8 The structure of Urdamycin A (13). Carbohydrate components are shown in red.
Another approach in the development of new antibiotics, is the use of synthetic
carbohydrate binding molecules that can disrupt pathogens by interacting with their
glycans.42
While these agents do not themselves contain carbohydrates, their effects
depend on interactions with carbohydrates displayed by pathogens. Therefore the
synthesis of homogeneous carbohydrate structures is also important in the development
of these therapeutic agents.
CHAPTER 1
38
1.2.2 Carbohydrate Based Vaccines
Glycoconjugate vaccines for H. influenzae type b43
, HIV44
and malaria45
have already
proven effective in vivo. Several types of cancers are now also known to express
abnormal amounts of certain glycans when compared to healthy cells allowing for the
production of anticancer vaccines.46
Structure 14 is an example of a vaccine presenting
three identical trisaccharide moieties, designed to mimic the presentation of
carbohydrates on ovarian tumour cells (figure 1.9). The carbohydrates are attached to
keyhole limpet hemocyanin (KLH) which stimulates an immune response and the
production of the desired antibodies.
Figure 1.9 The structure of an anti ovarian cancer vaccine. Carbohydrate component is shown in red.
MUC5AC corresponds to a peptide linker. KLH = Keyhole limpet hemocyanin, a commonly used
immunogenic protein.46
There are also several major bacterial pathogens, including V. cholerae, S. dysenteriae
and certain types of E. coli, whose infection mechanisms rely on the O-glycans they
produce. The study of these interactions is ongoing and is likely to play a crucial role in
the production of future vaccines.47,48
1.2.3 Carbohydrates for Cell Specific Drug Delivery
Targeted drug delivery to a specific organ, tissue or tumour in the body is the current
aim of many medicinal chemists. Such technology would reduce adverse side effects at
CHAPTER 1
39
the same time as increasing the efficacy of the treatment. Using the body’s own
recognition and trafficking pathways would be an efficient way of accomplishing this.
Treatments in development for several diseases now involve gene therapy and with our
understanding of the genome constantly increasing, the prevalence of such treatments is
likely to increase. Often the gene being delivered is only desired in specific cells or
tissues. Carbohydrate modification of virus particles has been explored and would be
one way of increasing the tissue specificity of such treatments.49,50
In 1974 it was discovered that hepatic cells showed specificity for proteins presenting
certain carbohydrate structures. Terminal sialic acid moieties were enzymatically
removed from glycans by treatment with neuraminidase to expose Gal residues. The
modified glycoproteins were then isotopically labelled with 125
I to facilitate detection.
When injected into rabbits these modified glycoproteins were found be quickly removed
from circulation by the liver. Further experiments using proteins with a range of
monosaccharides attached have shown that rabbit liver receptors have specificity
towards glycoproteins displaying terminal Glc or Gal residues.51,52
This has led to the
use of sialic acid to mask terminal sugars in therapeutic glycoproteins, which can
significantly improve circulation half lives.53
It also enabled liver targeting by
modifying therapeutic agents with Glc or Gal.54
A more sophisticated method of liver targeting named lectin-directed enzyme-activated
therapy (LEAPT), has shown great potential in recent years. In this approach an enzyme
that is foreign to the host organism is modified so that it will be preferentially absorbed
by hepatocytes and a prodrug is designed which requires this enzyme to be converted in
to its active form (figure 1.10). In the first example of this approach being successfully
trialled the nonmammalian enzyme rhamnosidase was stripped of its natural
glycosylations using endoglycosidase-H and then synthetically d-galactosylated using
2-imino-2- methoxyethyl 1-thioglycosides. The enzyme was then administered and
taken up by hepatocytes via the ASGPRs. Shortly after this the prodrug, consisting of 5-
flurouracil (15) protected with the nonmammalian sugar rhamnose was administered.
This sugar is also recognised by the ASGPR but is not recognised by mammalian
CHAPTER 1
40
enzymes. Although the drug itself has no specificity, it is only released in cells that
contain the galactosylated enzyme as shown in figure 1.10. This method has been
shown to increase the amount of enzyme delivered to the target cells by over 30 times,
therefore it drastically reduced the delivery of active drug to the rest of the body.55,56
Methods like this show great potential and similar approaches are being considered for
other organs and tissue types.56,57
Figure 1.10 An example of lectin-directed enzyme-activated therapy (LEAPT). In this case the enzyme is
rhamnosidase and the drug released is 5-flurouracil (15).56
Other modern examples of carbohydrate mediated cell targeting include the use of
mannose-6-phosphate (16) (figure 1.11) terminal glycans to derivatise
neoglycoproteins, promoting dramatically increased uptake by muscle cells.58
There
have also been several studies in to the use of carbohydrates to promote uptake of target
molecules or vesicles by macrophages which are important in infection, autoimmune
and antitumour response.59,60
These immune cells could then conceivably deliver the
therapeutic agent to the area they were required.61
Figure 1.11 The structure of mannose-6-phosphate (16).
CHAPTER 1
41
Liposomes and nanoparticles have also been trailed as vectors for drug delivery
including many anticancer agents. Glycosylation of these structures is one method of
enhancing tissue specificity. For example the polysaccharide hyaluronan has been
attached to nanoparticles to improve the delivery of 5-flurouracil (15) to colon cancer
cells and Gal has been used to increase the specificity of artificial vesicles for liver
cancer cells.62-64
This approach could provide a versatile method of delivering
therapeutic agents to their target locations without derivatising them directly, but so far
there are only a few successful examples.62,64
There are many areas in which neoglycoproteins and other glycoconjugates have
therapeutic potential and their medicinal use is likely to increase along with an
increased understanding of their biological roles.65-67
Although there have been
significant advances in the synthesis and analysis of glycoconjugates in the past
decades, new glycoconjugate tools are required for the future study of these important
interactions.
1.3 Carbohydrate Binding Protein Analysis (Considerations)
1.3.1 Polyvalency
Carbohydrate binding proteins (known as lectins) generally display multiple, identical
binding sites, thus giving them the ability to bind more than one copy of the same
carbohydrate simultaneously. This ability is known as polyvalency and can be achieved
by the oligomerisation of multiple subunit monomers, surface presentation of lectins or
extended binding sites that recognise multiple monosaccharides.68
Often a combination
of these strategies is employed; for example Concanavalin A (ConA) has an extended
binding site that recognises multiple Man residues and also forms a tetrameric subunit
formation under native conditions.69,70
When the strength of the average monosaccharide-lectin interaction is considered
(association constants in the mM range), it becomes clear that polyvalency is essential
CHAPTER 1
42
to the biological relevance of carbohydrate-protein interactions. Biological media such
as blood or cytoplasm also contain many carbohydrates in solution; therefore the
affinity of any glycoconjugates for a receptor molecule must be significantly higher than
those of simple sugars to have any unique function. The polyvalent presentation of
carbohydrates is able to enhance binding affinities in to the μM or even nM range where
they are capable of participating in many important binding events such cell adhesion
and microbial capture. Figure 1.12 (A) represents a bacterial cell presenting multiple
lectins binding to a surface presenting multiple glycan structures.69,71
Polyvalent
interactions are also significant enough to allow the separation of different classes of
glycoproteins with a lectin such as ConA, which has different affinities for several
different glycans. This approach, depicted in figure 1.12 (B), is used for glycoprotein
purification.72
Figure 1.12 Examples of the polyvalent presentation of lectins and carbohydrates. A) Cell-cell/cell-
surface interaction. B) The separation of glycoproteins by lectin affinity chromatography.
This binding enhancement is partly achieved by cooperative binding and high ligand
densities can also provide increased local concentrations of ligands, which can achieve
high binding site occupancy despite low affinities. However the effect multivalent
presentation has on binding interactions is still not fully understood. Several studies on
the area have been undertaken and many key factors have been identified such as ligand
density, ligand spacing and ligand flexibility.73-75
A single theory explaining the
phenomenon of polyvalent interactions is unlikely, as lectins with high affinities for the
same ligands can achieve their specificities via different mechanisims.70
Therefore any
CHAPTER 1
43
receptor of interest still requires separate and systematic study, which requires the use of
synthetic ligands.
In addition to the multiplying of ligand affinity, polyvalency also has a dramatic effect
on the specificity of lectins. Many lectins show a broad range of specificity towards
monosaccharides, seemingly with only small preferences between them in solution.
However when presented polyvalently as a cluster, polymer or on a surface, the
specificity is dramatically enhanced. ConA has an affinity for methyl Man, only four
times that of methyl Glc, but when polymerised the preference can be enhanced by over
100 times.76
It has also been shown that carbohydrate ligands with the highest affinity
for a lectin in solution do not always correlate to the structures that have the highest
affinities when presented polyvalently.77
This gives a valuable insight into how natural
glycoconjugates or cell surface glycans achieve their specificity.
Given the dramatic effect multiple ligand presentation (polyvalency) has on
carbohydrate binding any potential therapeutic agent intending to utilise carbohydrates
in their activity must be capable of presenting them multiple times. This can be
achieved by attaching multiple copies of the same monosaccharide ligand, by
synthesising branched ligands that incorporate multiple copies of the monosaccharide
concerned or (as found in nature) a combination of the two. Having a range of
multivalent tools to probe target receptors is essential for the development of future
carbohydrate based therapeutics.
1.3.2 Heterogeneity
In the synthesis of natural N-linked glycoproteins the initial N-glycosidic bond
formation that occurs as proteins are formed in the lumen of the ER is known to be
approximately 90% efficient.16
This leads to some initial heterogeneity in the population
of the glycosylated products (different glycoforms) of the same protein. Subsequent
enzymes also do not have 100% efficiency, therefore increasing the diversity of
glycoforms further. The final structure is the product of a competition between a variety
CHAPTER 1
44
of glucosyltransferases and glycosidases with different specificities and activities.
Additionally, proteins in different regions of a cell or those that have been in circulation
longer will have been modified in different manners. This intrinsically dynamic nature
of glycans in vivo is a source of their versatility; however it makes the purification and
analysis of natural glycans and glycoconjugates extremely problematic. Even when
natural glycans can be purified, the lack of a universal sequencing method for glycans
and the fact that the many different monosaccharides have identical masses makes the
elucidation of complex structures difficult.5,78
Determining the structure activity relationship of therapeutics is essential for assessing
potential harmful effects or developing improved treatments. This means artificial
glycans are required in the development of therapeutic glycoconjugates. They can also
be synthesised at higher purities and on larger scales than would be feasible by the
purification of natural glycoproteins. Furthermore the potential for unwanted side
effects is significantly reduced when therapeutic agents are well defined.
1.3.3 Synthetic Glycoconjugate Scaffolds
There are now a wide range of tools available for the analysis of lectin-carbohydrate
interactions. Due to the importance of presenting multiple copies of the carbohydrate of
interest and doing so in a controlled manner any glycoconjugate designed must be well
defined and have the ability to be multivalent. There are several approaches for creating
multivalent ligands for binding assays or therapeutic applications. Some of the most
successful types include; linear polymers, glycol dendrimers, nanoparticles and
neoglycoproteins. Representations of these structures are shown in figure 1.13.
CHAPTER 1
45
Figure 1.13 Some of the most widely used classes of glycoconjugate scaffold. A) Linear glycopolymers.
B) Glycodendrimers. C) Glycosylated nanoparticles. D) Neoglycoproteins. Green circles represent
glycans.
While there will always be some heterogeneity in a polymer population it is now
relatively straightforward to synthesise glycopolymers with polydispersities lower than
1.1 by a range of different methods.79-81
Having accurate control of both chain length
and polydispersity means the binding affinities of glycopolymers can be tailored to their
use. Glycopolymers and dendrimers offer the most reliable method of creating highly
polyvalent ligands with nM binding affinities.82,83
Their high flexibility, long chain
length and potential for high ligand density can allow them to fit the orientation of any
binding site. These properties have made them extremely useful for cell or surface
labelling and inhibition assays. However these scaffolds are not well suited for the
detailed analysis of carbohydrate binding interactions and require the addition of radio
or fluorescent labels to allow detection.
In the past decades there has been great interest in a variety of carbohydrate coated
nanoparticles for receptor analysis, imaging and drug delivery.84-86
The potential for
carrying therapeutic compounds inside their shell and easy incorporation of labelling
agents make them a desirable delivery mechanism. However the surface presentation of
carbohydrates is not well defined, limiting their use in structure-activity studies.
Synthetic glycoproteins are the most biologically compatible construct for the study of
carbohydrate binding. While they will not easily achieve the high binding affinities
possible using glycopolymers they can be extremely useful. Neoglycoproteins have
been widely used since the 1980’s to probe lectin specificity and to develop
CHAPTER 1
46
carbohydrate based methods for cell targeting.87-89
More recently glycopeptides and
glycoproteins have been seen as strong candidates for future vaccines.90,91
In most cases
neoglycoproteins are nonselectively glycosylated, but a wide variety of selective
methods are now available.92
This means they also have the potential for the systematic
study of the structure-activity relationship of carbohydrate binding proteins.
1.4 Artificial Glycoproteins
Artificial glycoproteins (or neoglycoproteins) have long been used to better understand
the role of carbohydrates in biology. In 1929, simple carbohydrate structures were
attached to immunogenic proteins to determine the effect of small changes to glycan
structure on immunological specificity. Glucoside (17) and galactoside (18) which
differ in structure by the inversion of just one stereocenter, (figure 1.14) were converted
to their diazonium derivatives and then attached to the lysines of the target proteins to
create neoglycoproteins. It was found that changing the carbohydrate attached to a
protein changed the antigenic specificity induced by the glycoprotein. It was also found
that even if different proteins were used the immune response was dictated by the
glycans attached rather than the protein, despite the fact that the glycans alone did not
induce a response. This is a very early example of the importance of carbohydrate
recognition and illustrates how even the smallest possible variation in glycan structure
can lead to significant changes in biological function when glycans are presented as
polyvalent glycoconjugates.93
Figure 1.14 Chemical structures of glucoside (17) and galactoside (18). Hydroxyl group attached to C4
shown in red.
The most common use for neoglycoproteins has been the study of carbohydrate binding
protein interactions. They have been used to identify the specificity of several cell
membrane bound carbohydrate binding proteins both in vivo and in vitro. For example
Man binding of epidermal cell receptors in rats,94
bovine airway smooth muscle cells95
CHAPTER 1
47
and Langerhans cells.88
The specificity of purified intracellular lectins has also been
probed using neoglycoproteins. For example the specificity of some lectins found in cell
nuclei for glucose and N-acetylglucosamine.96
Another use of neoglycoproteins is in the production of tailored monoclonal antibodies.
Synthetic glycans are generally attached to immunogenic proteins so that the host
organism produces antibodies which bind the glycan structures. Antibodies produced in
this way have potential uses in both treatment and diagnosis of disease. However carful
screening is required to ensure the antibodies produced are specific to the glycan of
interest and not to the protein scaffold.97,98
1.4.1 Synthetic Strategies for Neoglycoproteins
The total synthesis of naturally occurring proteins is now possible and has been
achieved for a number of proteins including biologically active enzymes and hormones.
The most common and most versatile approach involves the solid support synthesis of
peptide fragments followed by native chemical ligation (NCL) of these fragments to
produce the final protein.99-102
NCL is particularly useful in the generation of proteins
containing a range of different post translational modifications. This is because different
structures can either be incorporated in to separate peptide fragments during their
synthesis or attached to completed fragments before NCL takes place. This method can
be used to produce high purity glycoproteins containing only natural linkages, which is
why it is the current method of choice for the synthesis of bio mimetic
neoglycoproteins.103-106
CHAPTER 1
48
Figure 1.15 The native chemical ligation (NCL) of two synthetic peptides to form a natural peptide bond.
NCL was first introduced in 1994 using N-terminal cysteine peptides and C-terminal
thioester peptides as shown in figure 1.15.107
In the first step the cysteine residue
performs a reversible transthioesterification with the C-terminal thioester. This could
also take place with any unprotected cysteine side chains, however in the case of the N-
terminal cysteine the next step, which is an irreversible S N acyl transfer, rapidly
takes place to form an amide bond between the two fragments. Although this method
was very successful the linkage was limited to sites containing cysteine; a relatively
uncommon amino acid.
Over the last two decades the NCL methodology has been extended to facilitate the
formation of natural linkages at a variety of sites including, alanine, phenylalanine,
valine, leucine, threonine, lysine, proline and glutamine.108
The first step in this
development was the discovery that peptides could be desulfurized (have their thiol
groups removed) using a Raney’s nickel or palladium/aluminium oxide catalysts. This
was first used to convert cysteine to alanine after NCL to achieve the first “alanine
ligation”.109
Eventually it was found that desulfurization could be achieved without a
CHAPTER 1
49
metal catalysts and in aqueous (aq) media using the reducing agent tris(2-
carboxyethyl)phosphine (TCEP) and a free radical initiator (2,20-azo bis(2-(2-
imidazoline-2-yl)propane) dihydrochloride).110
The general scheme for this approach is
shown in figure 1.16 in which R = H for the “alanine ligation” but can be many other
groups if a different linkage is desired. The distance of the thiol group from the N-
terminus can also be altered, thus broadening the scope of NCL linkages even further.
Figure 1.16 General scheme for the native chemical ligation of peptides to form natural peptide linkages
that do not contain cysteine.
NCL of glycosylated peptide fragments can now be used to create synthetic
glycoproteins identical to their natural counterparts.102,103
This method guarantees the
glycoprotein produced only contains the desired glycoforms and provides total control
over the addition or removal of glycosylation sites. The main drawback of NCL is it still
requires the synthesis of peptide fragments and artificial amino acids required for non
CHAPTER 1
50
cysteine based couplings. It is still faster and cheaper for most recombinant proteins to
purify them once they have been overexpressed in vivo, especially if large quantities are
required. If expressed in an organism lacking the biological machinery for glycosylation
then large amounts of high purity, unglycosylated proteins can be quickly and easily
attained. The desired glycans can then be attached chemically or enzymatically if
required.
Carbohydrates purified from bacterial sources have been used effectively in some
vaccines.111
However the purification of such material is difficult, yields are poor and
the usefulness of the products is often limited by purity. The chemical synthesis of
complex glycosides is by no means easy due to the challenge of selectively
glycosylating a single hydroxyl group on a glycan out of many. However recent
improvements in solution-phase and solid support oligosaccharide synthesis have made
it possible to produce complex glycans on a more practical scale.112-114
For these
reasons a semisynthetic approach in which a synthetic glycan is chemically attached to a
unglycosylated protein produced in vivo is our favoured method of creating
homogeneous neoglycoproteins.
1.4.2 Chemical Glycosylation
There are numerous ways in which carbohydrates have been artificially linked to
proteins, but all methods have disadvantages or limitations. Some early examples are;
diazonium salts or imidates of glycosides being used to glycosylate lysines115,116
or
glycosylamines being used to glycosylate the carboxyl groups of a protein.117
Although
these methods have been used to produce many useful neoglycoproteins they do not
produce well defined products and can alter the physical properties of a protein and
thereby disrupt its function. For example lysines are an attractive target for
glycosylations because they can react with a wide range of functional groups including;
activated esters (such as the NHS-ester), isocyanates, isothiocyanates and sulfonyl
chlorides.118
Examples of these reagents and the linkages they produce are shown in
figure 1.17.
CHAPTER 1
51
Figure 1.17 Examples of reagents which can be used to modify lysine side chains and their products.
Lysine is however a relatively common amino acid and therefore does not provide the
selectivity required for a systematic study of the effect of protein glycosylation. Also
lysine often plays an important role in solubilising proteins so the removal or
derivatisation of several lysine residues could have a detrimental effect on protein
stability and function.
Introducing unnatural amino acids (UAAs) with unique reactivities into peptides is a
good way of ensuring specificity. Ketone containing amino acids have been
incorporated into both solid support and biological syntheses of proteins.119,120
This
group reacts selectively under mild conditions with both with aminooxy compounds or
hydrazine compound to form stable oxime or hydrazone linkages respectively (figure
1.18). Other functional groups also not present in nature have been incorporated in
similar ways; including anilines, aryl halides, and boronic acids.121
Like ketones, these
CHAPTER 1
52
functional groups also react bioorthogonally which means they do not require the use of
any protecting groups.
Figure 1.18 The bioorthogonal reaction of a ketones with aminooxy compounds or hydrazine compound
to form stable oxime or hydrazone linkages respectively.
“Click” chemistry is now a commonly used strategy for protein derivatisation due to its
biocompatibility and selectivity. Generally the reactions used are cycloadditions
between an azide and either an alkene or an alkyne as shown in figure 1.19. Either the
azide or the alkene/alkyne can be incorporated in to a protein’s structure and then the
complimentary functional group required is incorporated in to the molecule which is to
be attached to the protein.
Figure 1.19 Examples of “click” chemistry cycloaddition reactions between azides and alkenes or
alkynes.
Many UAAs containing azide, alkyne or alkene functionality can be incorporated in to
proteins genetically or by simply replacing cell feedstocks.92,121
Structures 19-22 shown
in figure 1.20 are examples of UAAs compatible with “click” chemistry. More recent
examples such as structures 23 and 24 (figure 1.20) utilise strain-promoted reaction
technology and are particularly suited to live cell labelling studies because they do not
require catalysis.122,123
There is now a huge variety of UAAs available for use in a wide
range of organisms.124
However, due to time consuming syntheses and the need for
CHAPTER 1
53
specially engineered cell lines, the use of naturally occurring amino acids is still often
preferable.
Figure 1.20 Examples of unnatural amino acids suitable for “click” chemistry reactions;
azidohomoalanine (19), p-Azido-L-phenylalanine (20), homopropargylglycine (21), homoallylglycine
(22) and two examples of UAAs utilising strain promoted reaction technology (23 and 24).
1.4.3 Glycosylating Cysteines
Cysteines are well suited for potential glycosylation sites because they are the only
amino acid containing a thiol group and are a relatively uncommon amino acid.
Furthermore, many cysteines are internal and form disulphide bridges so will not be
accessible for chemical reactions when the protein is in its native state. With site
directed mutagenesis now a standard laboratory practice and the relatively low cost of
synthetic genes, cysteines can easily be incorporated into any sequenced protein. Many
useful methods have been developed for selectively modifying cysteines, some of which
are summarised in figure 1.21. The more useful include: the formation of disulphides,
the oxidation of cysteine (25) to dehydroalanine (26) and alkylation with
electrophiles.125
A disulphide linkage can be formed by reacting cysteine with another
thiol under oxidising conditions or by disulphide exchange. However, glycoproteins
made by this method are not stable in vivo because they are prone to reduction. This can
CHAPTER 1
54
Figure 1.21 Summary of commonly used methods of glycosylating cysteines.125
be solved by converting the disulphide into a thioether using hexamethylphosphorous
triamide (HMPT),126
in a process known as desulfurization (illustrated in figure 1.21).
O-mesitylenesulfonylhydroxylamine (MSH) can be used to convert cysteine (25) to
dehydroalanine (26) via oxidative elimination. The double bond of dehydroalanine (26)
can then act as a Michael acceptor to nucleophiles, as shown in figure 1.21. Using this
method does result in the loss of stereoselectivity, but has been successfully used to
create glycoproteins stable in vivo. MSH can also oxidise proteins glycosylated via a
thioether linkage back to dehydroalanine, making this method reversible.127
Alkylating cysteine can produce stable linkages in one step and a range of electrophiles
can be used. Michael acceptors (such as maleimides and vinyl sulfones) have been used
to selectively glycosylate128
therapeutic proteins successfully. α-halocarbonyls were one
of the first electrophiles used to modify cysteines129
and are still in use today.
Iodoacetamides are used most often, but reactions with lysine have been reported to
occur.125,130
Chloro or bromoacetamides can be used to increase selectivity at the
expense of the reaction rate. α-halocarbonyl compounds are easy to prepare and can also
mimic the natural N-linked glycosylation structure, displaying glycans in their naturally
CHAPTER 1
55
occurring conformations.131
Glycosides of this type have been successfully used to
prepare many glycoproteins and other protein conjugates.125,132
1.5 Green Fluorescent Protein (GFP)
Wild type GFP is a 26.9 kDa protein consisting of 238 amino acids and is found in
nature expressed by the jellyfish Aequorea victoria along with the protein aequorin. In
the presence of Ca2+
ions and the cofactor coelenterazine, aequorin emits a blue light
(λmax = 470 nm). GFP has a major excitation peak at 395 nm and a minor one at 475
nm allowing it to absorb the light emitted by aequorin and fluoresces at 508 nm to
produce the characteristic bluish-green colour of the jellyfish.133,134
The cloning of the
GFP gene into E. coli in 1992 allowed large quantities of this protein to be produced
quickly for the first time and the mechanism of the fluorophore formation was
elucidated soon after.134,135
As shown in figure 1.22, the fluorophore is formed from
three adjacent amino acids in GFP; Ser 65, Try 66 and Gly 67, which together make
structure 27. In the first step the amido group of Gly 67 performs a nucleophilic attack
on the carbonyl group of Ser 65 to form the tetrahedral intermediate 28. This
intermediate collapses to form the cyclic structure 29 which is then oxidised to form
structure 30 which contains the completed fluorophore (shown in red (figure 1.22)).
Figure 1.22 The formation of GFP’s fluorophore. Fluorophore shown in red.
CHAPTER 1
56
The β-barrel structure of GFP (figure 1.23) consists of a β-sheet, made of 11 parallel β-
strands, with an α-helix running through the centre that contains the fluorophore
sequence. In the centre of the barrel structure this fluorophore is relatively protected
from possible quenching reactions.136,137
Formation of the fluorophore requires no
specialised cofactors or substrates from Aequorea victoria, only that molecular oxygen
to be present during protein expression.
The first GFP fusion protein was expressed by Wang and Hazelrigg in 1994138
and since
then it has been expressed in a variety of organisms including; bacteria, fungi, insects,
fish, plants and mammals as well as in human cells. It is a popular tool for fluorescent
microscopy and live cell imaging because unlike many fluorescent molecules it is
considered to be non toxic in most cases, but can be cytotoxic if exited for extended
periods of time.139,140
It was also found to be relatively resistant to heat, pH (5-12),
detergents, salts and most proteases. GFP is commonly used as a fusion protein because
it rarely affects the mobility or activity of the protein it is attached to. GFP fusion
proteins have been extremely useful for studying protein dynamics, expression and
interactions.141-143
Figure 1.23 Ribbon diagram of GFP showing the β sheets in yellow, α helix in red, fluorophore in blue
and loops in green.
CHAPTER 1
57
A number of improvements have been made to photostability and thermal stability of
wild type (WT) GFP since it was first cloned.144
Enhanced GFP, reported in 1996 was
created by making the single point mutation F64L. This increased the folding
efficiency at 37°C and facilitated the use of GFP in mammalian cells.145
Other
significant improvements include GFPuv146
and superfolded GFP147
in 2005, which
respectively increased the fluorescence and folding efficiency of the protein.
The emission wavelength of GFP’s fluorophore is very sensitive to the hydrogen
bonding and π-stacking interactions with the surrounding amino acids. This has made it
possible to create different coloured GFP mutants, which has widened the scope of their
application. Available GFP mutants now include blue fluorescent protein, cyan
fluorescent protein and yellow fluorescent protein.148
There are also several red
fluorescent proteins commercially available, such as mCherry, however these are
derived from the choral protein DsRed originally found in Discosoma sp.
pH and redox sensitive mutants have allowed GFP to act as a reactive biosensor. These
variants have facilitated the visualisation of biological processes; for example
visualisation of synaptic activity in neurons.149,150
Mutants have also been produced
which act as a detector of metal and halide ions.139
The significance of this remarkable
protein was demonstrated in 2008 when Martin Chalfie, Osamu Shimomurra and Roger
Tsien shared the Nobel Prize in chemistry in for GFP’s discovery and development.
One potential drawback is GFP’s tendency to form dimers at high concentration which
can lead to aggregation of fusion proteins. This has been reported when expressed in a
confined area such as with membrane proteins. However a single point mutation
(K206A) has been found to alleviate this problem with no effect on the fluorescence.151
CHAPTER 1
58
1.5.1 GFPuv
GFPuv was primarily chosen as the protein for our glycosylation studies because
cysteine mutants of it have previously been made and these mutants were successfully
derivatised using iodoacetamides.152
Additionally, the use of a florescent protein as a
model glycoprotein eliminates the need for a separate chemical or radio label.
GFPuv differs from wild type GFP by three point mutations; F99S, M153T and V163A.
These mutations improve the protein’s expression by increasing the efficiency of
folding. The protein produced is more stable to variations in temperature and pH and
also around 45 times more fluorescent than WT GFP.153
Its excitation maximum does
not move from 395 nm, but its emission maximum changes from 509 nm to 508 nm.
Like WT GFP it only contains two cysteines and one of them is internal (C70),
therefore not chemically reactive when the protein is correctly folded. C48 however is
on the surface of the protein and therefore likely to react with α-halocarbonyl reagents.
Mutation of C48 to alanine, E6 to cysteine or I229 to cysteine has been reported to have
no detrimental effect on GFPuv’s fluorescence.152
Therefore there are three potential
Figure 1.24 Ribbon diagram of the crystal structure of GFPuv with three amino acids (E6, C48 and I229)
residues highlighted (in yellow).
CHAPTER 1
59
glycosylation sites already available and a range of mutants with two or three
glycosylation sites should be easy to produce. The relative positions of these previously
mutated residues are shown in figure 1.24.
1.5.2 Glycosylated GFP
The production of glycosylated GFP mutants is not a novel idea. There have been many
cases of glycosylated GFP or GFP fusion proteins being engineered and expressed in
vivo for the use in protein trafficking and enzyme activity studies.154-156
These
neoglycoproteins were generated by incorporating glycosylation sites in to the amino
acid sequence of GFP. However in the majority of cases these glycosylation sites were
incorporated in to either the N or C-terminal sequences, which are not ideally suited due
to increased risk of processing and degradation. More recently there has been an interest
in the incorporation of glycosylation sites in more central positions to produce more
stable glycoconjugates.157
There are limitations to the sites at which natural glycosylation sites can be placed in the
GFP sequence because of the specific three amino acid sequon required. In one study157
only three sites were identified for screening, all on loops protruding from the β-barrel
structure. Of these three one prevented GFP folding and one was not recognised as a
glycosylation site. Position 133 was the only successfully engineered non terminal
glycosylation site, which was still considered a potentially valuable tool for the study of
protein trafficking. Given the restrictions imposed by natural glycosylation sites, it is
unlikely that many more will be discovered for GFP in this way.
The sites in which GFP can be chemically glycosylated are significantly more diverse
because only one amino acid residue requires alteration to create a glycosylation site.
This also makes it easier to glycosylate GFP in central positions (not just at the termini)
to produce more biomimetic neoglycoproteins. GFP modified with a C-terminal
thioester has been attached to a glycopeptides via NCL and the product was successfully
used for lectin binding analysis.158
Additionally GFP-lectin fusion proteins have been
CHAPTER 1
60
successfully used for carbohydrate binding studies.159
In both cases the fluorescence of
the protein was used as a simple way of demonstrating carbohydrate-protein interaction
which suggests our approach should also be successful. Glycosylated GFP mutants
made via chemical glycosylations could be used both for in vitro carbohydrate binding
assays or injected into cells for in vivo protein trafficking studies. However the use of
glycosylated GFP in cells requires careful consideration of the location in which they
are used due to the potential for glycan modification by host enzymes.160
1.6 Project Aims
The interactions of glycoconjugates with carbohydrate binding proteins are responsible
for a wide range of recognition events in vivo; including immune response, cell
adhesion and signal transduction. Glycoconjugates have already found many medicinal
uses as therapeutic and diagnostic agents, but their full potential is yet to be realised.
Access to a variety of homogeneously glycosylated glycoproteins is essential for the
study of these important carbohydrate binding events. However glycoproteins expressed
in vivo are produced in a variety of different glycoforms. Therefore the chemical
synthesis and attachment of biologically relevant glycans to unglycosylated protein
scaffolds is required for a detailed analysis of structure-activity relationships.
The reaction of glycosyl iodoacetamides with cysteines was our chosen method of
selectively glycosylating our protein scaffolds. Cysteines are well suited as
glycosylation sites because they are the only amino acid containing a thiol group and
are relatively uncommon. Iodoacetamides provide an irreversible, one step method of
derivatising proteins, which is selective for cysteines.
We have chosen the green fluorescent protein mutant GFPuv for use as a protein
scaffold due its known tolerance of two cysteine mutations (E6C and I229C) and the
previously successful derivatisation of these cysteines with iodoacetamides.1 The
inherent fluorescence of GFPuv also makes it a useful candidate for fluorescence based
binding assays or cell labelling studies.
CHAPTER 1
61
The general aim of this project was to produce a range of homogenously glycosylated
GFPuv mutants suitable for carbohydrate binding protein analysis and to use then to
analyse the specificity of some carbohydrate binding proteins. The strategic aims were
to first produce a range of active GFPuv mutants containing additional cysteines using
the previously reported mutations E6C and I229C as a starting point. Then a simple
method for the overexpression and purification of these mutants would be optimised to
produce adequate amounts for analysis and derivatisation. The synthesis of a range of
glycosyl iodoacetamides suitable for carbohydrate-binding protein analysis was also
necessary. Simple glycan structures were the initial targets to confirm the validity of
this approach and if this was successful then more complex glycans of biological
significance would be targeted. It was essential that the glycosylation reaction between
synthetic glycosides and GFPuv mutants was optimised to ensure the production of
homogeneous neoglycoproteins. This in turn required a sufficiently sensitive method of
glycoprotein analysis to assess the levels of derivatisation. Finally methods of
measuring carbohydrate binding protein interactions would be needed to be explored to
assess the potential of glycosylated GFPuv as a diagnostic tool.
In summary, the strategic aims for this project are:
To produce a variety of active GFPuv cysteine mutants
To develop a simple, effective method of purifying GFPuv
To synthesise a variety glycosyl iodoacetamides
To optimise the cysteine selective glycosylation of GFPuv mutants
To optimise an efficient method of monitoring neoglycoprotein derivatisation
To screen the neoglycoproteins produced against selected carbohydrate binding
proteins
62
Chapter 2: The Generation, Expression and
Purification of GFPuv Mutants
CHAPTER 2
63
2.1 Generating GFPuv Cysteine Mutants
Our primary goal was the generation of a range of GFPuv cysteine mutants; starting
with those previously reported followed by the screening for additional sites capable of
tolerating mutation to cysteine without effecting activity. However the expression and
purification of both WT GFPuv and any mutants produced was also in need of
consideration.
The ability to quickly produce milligram quantities of molecular GFPuv mutant was
important so overexpression in E. coli followed by purification was the logical method
of production. Several GFP mutants are routinely prepared in this way usually
expressed between 25°C and 37°C. Lower incubation temperatures produce protein at a
reduced rate but have been reported to yield a higher proportion of active protein, which
in most cases is more desirable161,162
The purification of untagged GFP can be achieved
in a number of ways including: hydrophobic interaction chromatography, organic
extraction, high performance liquid chromatography (HPLC) and chromatofocusing on
a pH gradient.162,163
Methods have even been developed for purifying untagged GFPuv
by gradient immobilised metal affinity chromatography (IMAC).161
However this may
not be as effective with all GFPuv mutants.
Figure 2.1 Binding of a hexahistidine tagged protein to an immobilised metal affinity column.
Although not suitable for some biomedical applications and not ideal for larger scale
production IMAC using a hexahistidine tag is still the most reliable method that can be
applied to all conceivable GFP mutants.162
This method relies on the binding of multiple
histidine residues attached to either protein terminus to metal ions immobilised on a
solid phase resin (figure 2.1). The captured protein can subsequently be eluted by
CHAPTER 2
64
washing the resin with a solution containing high concentrations of imidazole.
Therefore our first objective was to introduce a hexahistidine tag on to a terminus of the
GFPuv gene.
2.1.1 Addition of Hexahistidine Tag to GFPuv
The WT GFPuv gene was amplified by polymerase chain reaction (PCR) before being
purified and cloned in to a pET-30a vector using the restriction sites NotI and EcoRI, as
illustrated in figure 2.2 (details section 8.1.2).
Figure 2.2 Schematic representation of the cloning of GFPuv into a pET-30a vector.
The resulting construct was sequenced to confirm the desired product had been
produced (full sequence shown in appendix 1). Figure 2.3 shows the corresponding
amino acid sequence with the hexahistidine tag and GFPuv sections highlighted. This
protein shall be referred to as GFPuv_WT. To maintain the numbering of the GFPuv
residues the hexahistidine tag and other additional amino acids will be assigned
negative values if required (-52M to -1F).
CHAPTER 2
65
MHHHHHHSSGLVPRGSGMKETAAAKFERQ
HMDSPDLGTDDDDKAMADIGSEF-GFPuv
Figure 2.3 The additional amino acid sequence introduced on to the N-terminus of GFPuv. Hexahistidine
tag is shown in red and the GFPuv section is shown in green.
2.2 Site Directed Mutagenesis
To compare the reactivity of the two additional cysteines reported to be tolerated in
GFPuv (E6C and I229C) against the naturally occurring cysteines (C48 and C70) a
range of mutants were required. C48 would be mutated to alanine (the mutation C48A)
and the E6C mutation and the I229C mutation would be introduced separately. For the
C48A mutation a parallel approach was undertaken using inverse PCR with the primers
previously described1 and the more conventional Quickchange method.
2.2.1 Inverse PCR
Figure 2.4 Schematic representation of the inverse PCR method for the C48A mutation.
The GFPuv_WT vector was amplified by a PCR using the primers previously
reported,1,164
the purified PCR product was phosphorylated and then a self ligation was
performed. This procedure is summarised in figure 2.4 (details section 8.1.2). Five of
the resulting plasmids were sequenced and it was confirmed that the ligation and
transformation were successful. However, the sequences showed that the PCR reaction
CHAPTER 2
66
did not produce the desired product. Instead, varying numbers of primer pairs were
inserted into the mutation site. Some examples of the sequences obtained from the
inverse PCR products are shown compared to the GFPuv sequence in figure 2.5. The
repeated sequence of the forward primer used is highlighted in green and is repeated
three times in the sequence Inv 1 and once in the sequence Inv 2. The sequences also
show that the sequence both before and after the primer insertions are in consensus with
the GFPuv sequence. Therefore it was definitely an insertion rather than a mutation
which had taken place and these products were not useful to us.
Figure 2.5 Comparison of the products of the inverse PCR (Inv 1 and Inv 2) with the GFPuv sequence
around the C48A mutation site. Repeat units of the forward primer used in the PCR are highlighted in
green.
It is possible that if enough PCR products from this method were screened then one
would contain the desired mutation or that the PCR reaction could have been optimised.
However, this approach was not pursued further due to our success with the
Quickchange method, which was conducted in parallel to the inverse PCR approach.
CHAPTER 2
67
2.2.2 The Quickchange Method
Figure 2.6 Schematic representation of the Quickchange site directed mutagenesis method for the C48A
mutation.
The C48A mutation was introduced using a modified version of the Quickchange site
directed mutagenesis procedure which is summarised in figure 2.6 (details in section
8.1.2). This new mutant in which the C48A mutation was successful will be referred to
as GFPuv_C48A.
The I229C mutation was introduced to both GFPuv_WT and GFPuv_C48A using the
same PCR conditions found to work for the C48A mutation. In both cases the I229
mutation was introduced successfully. The two new mutants will be referred to as
GFPuv_I229C and GFPuv_C48A_I229C respectively. The GFPuv_C48A_I229C PCR
reaction did alter one base pair that was not intended, but fortunately it was a silent
mutation (aaaaag at L125) (full sequence in appendix 1).
The E6C mutation was introduced to all previously generated mutants (GFPuv_WT,
GFPuv_C48A, GFPuv_I229C and GFPuv_C48A_I229C using a slightly different PCR
(details in section 8.1.2). These new mutants will be referred to as GFPuv_E6C,
GFPuv_E6C_I229C and GFPuv_E6C_C48A_I229C respectively. The GFPuv_EC
contained two silent mutations occurring at P54 (cctccc) and K126 (aaaaag).
CHAPTER 2
68
2.3 Generation of Cysteine Mutants by DNA Shuffling
Many of the improvements to WT GFP have been achieved using a combination of
error prone DNA shuffling and directed evolution.146,147
This technique can also be used
to combine homologous genes with desired characteristics to quickly create libraries of
chimeric proteins (depicted in figure 2.7).165,166
We have demonstrated a method of
using DNA shuffling to rapidly scan several surface residues of GFPuv specifically for
mutation to cysteine. To our knowledge this approach has not been used to generate
cysteine mutants before.
Figure 2.7 Schematic representation of how homologous genes can be combined in a primerless PCR to
create new chimeric genes for screening.
2.3.1 General Considerations
It is of course possible to screen every surface residue of GFPuv for its tolerance to
mutating to cysteine. Then residues which were found to tolerate these mutations could
be screened in pairs or higher order combinations to discover which were still viable.
However, if only ten residues were found to tolerate mutation to cysteines then there
would be 1023 different cysteine mutants combinations to screen. Not only would this
method be time consuming, but it may fail to find mutations which were viable in
combination and were not successful individually. Random mutagenesis of GFP would
CHAPTER 2
69
inevitably produce some useful cysteine mutants if libraries were large enough, but
sifting through all of the non-cysteine mutations would require the analysis of the DNA
sequences of all active mutants produced. Finding sequences containing combinations
of usefully mutated sites would also be extremely unlikely using this approach.
The DNA shuffling of two or more GFPuv genes containing different numbers of
cysteine codons provided a solution to the rapid discovery of new cysteine mutants.
This approach simultaneously screens combinations of cysteines and is potentially
applicable to any protein. Using GFPuv’s inherent fluorescence, large numbers of genes
can be rapidly screened for activity without the need for an additional assay and then
only the sufficiently active products needed sequencing.
2.3.2 Design of Polycysteine Mutants for DNA Shuffling
After analysis of GFPuv’s structure it was found that over 140 of GFPuv’s 238 amino
acids were near the surface of the β-barrel structure. Therefore mutation of any of these
amino acids could yield potentially reactive cysteine mutants. Three synthetic genes
named Shuffle 1-3 were designed to scan 107 of these sites (details in section 8.1.2). In
individual genes the sites to be screened were separated by a minimum of 12 bp (usually
15 or more) to increase the chances of successfully recombining with the unmodified
GFPuv sequence. We initially chose a gene containing 33 surface cysteines (Shuffle 1),
including two sites known to tolerate cysteines (C48 and C229) as controls for the
screen. Shuffle 1 and a GFPuv gene with the mutation C48A (sGFPuv_C48A) were
codon optimised for E. coli and purchased from GeneArt (sequences shown in appendix
2). Both synthetic genes were the cloned in to pET-30a vectors using the EcoRI and
NotI restriction sites included in their design. The amino acid sequence of
sGFPuv_C48A is shown in figure 2.8 with the amino acids corresponding to cysteines
in the sequence of Shuffle 1 highlighted in red.
CHAPTER 2
70
Figure 2.8 Amino acid sequence of the sGFPuv_C48A gene purchased for DNA shuffling experiments.
Amino acids highlighted in red correspond to the residues exchanged for cysteine in the Shuffle 1 gene.
2.3.3 DNA Shuffle of Shuffle 1 and sGFPuv_C48A
Approximately 1 kb fragments containing both synthetic genes (Shuffle 1 and
sGFPuvC48A) were obtained from their respective plasmids by digestion using EcoRI
and NotI. These gene containing fragments were further digested using DNaseI into
random fragments of less than 200 bp for use in the DNA shuffle PCR. A primerless
PCR was then conducted using these fragments under the same conditions as previously
reported (details in section 8.1.2).146,167
The 1 kb fragments obtained were cloned back
in to pET30a and transformed in to an expression vector for screening. Colonies
containing active GFPuv mutants were detectable using UV light (659 nm) a few hours
after induction and visible to the human eye after a few more. As shown in figure 2.9
the colonies expressing active GFPuv mutants were visibly green and the colonies
expressing inactive GFPuv mutants appeared to be the natural brown of E. coli.
Figure 2.9 Example expression plate containing transformants from DNA shuffle products. Highlighted
section on the left is enlarged on the right.
CHAPTER 2
71
Several of the colonies were unusually small, which may be because some of the
mutants produced were harmful to the cells. Around 30% of the normal sized colonies
were visibly green. 94 green colonies were sequenced and out of these ten were found to
be novel mutants whilst the rest were identified as sGFPuv_C48A. The amino acid
sequence of these ten novel GFPuv mutants are summarised in figure 2.10.
Figure 2.10 Summary of the active, GFPuv mutants discovered by the DNA shuffling. Each column
corresponds to a new gene. White sections correspond to segments of sGFP_C48A and the green sections
correspond to segments of Shuffle 1. Naturally occurring C48 has been highlighted yellow. The mutant
names correspond to their library designation.
From these novel GFPuv mutants, ten new sites were found to tolerate mutation to
cysteine and some combinations were also produced. Four new double mutants, one
triple mutant and two quadruple mutants were found in this relatively small library. The
mutants identified in this screen were named corresponding to their library designation.
CHAPTER 2
72
E.g. mutant GFPuv_B10 was the tenth mutant sequenced in row B of the plate
sequenced and refers to the mutations C48A, S202C, N212C and I229C. Interestingly
I229C (previously known to be tolerated) was the most common mutation observed and
the A48C (mutation back to WT) was one of the second most common mutations found
in the genes produced. Also cysteines were less commonly found towards the centre of
the gene (figure 2.10). Further investigation or a larger library size would be needed to
determine if this was due to the protein stability towards mutations, the DNase digest
not being sufficiently random or the PCR conditions.
Due to the fact that so many of the active mutants sequenced were found to match the
DNA sequence of sGFPuv_C48A it was decided to investigate the sequences of some
inactive colonies to determine if they matched the sequence of shuffle one. If this was
the case it would suggest that the DNase digest was not thorough enough. 12 inactive
colonies were sequenced to assess the efficacy of the DNA shuffling experiment and it
was found that all of them had been successfully shuffled. The amino acid sequences of
these inactive mutants are summarised in figure 2.11. These results suggest that if a
larger library of active colonies was screened many more combinations of cysteine
mutations would be discovered.
It was decided that we had a sufficient variety of new mutants to produce a range of
neoglycoproteins for initial screening so no further sequencing or DNA shuffling
reactions were embarked upon. Two further genes were purchased from Gene Art for
screening. One contained all 13 cysteines found to be tolerated in different mutants and
the other contained six cysteines not present in GFPuv_WT (E6C, S30C, T38C, T43C,
K52C and I229C) and C48. The latter turned out to be active and was named
GFPuv_C5+2, but the mutant containing all 13 mutations was not successfully
expressed in an active form.
CHAPTER 2
73
Figure 2.11 Summary of the 12 inactive GFPuv mutants screened. Each column corresponds to a new
gene. White sections correspond to segments of sGFPuv_C48A and the green sections correspond to
segments of Shuffle 1. Naturally occurring C48 has been highlighted yellow. The mutant names
correspond to their library designation.
2.4 Expression of GFPuv
2.4.1 Optimisation of Protein Expression
The optimisation of protein expression was carried out using GFPuv_WT. Induction of
protein production was tested at 22°C, 30°C and 37°C. Samples of each culture were
removed at 1 hour intervals for 8 hours and also after 24 hours. The remaining bulk of
the cell cultures were also harvested stored for quantitative analysis (details in section
8.1.3).
CHAPTER 2
74
Figure 2.12 SDS-PAGE gels of the samples taken from cultures expressing GFPuv_WT at 22, 30 at
37°C. From left to right the lanes correspond to; protein ladder, 0 h, 1 h, 2 h, 3 h, 4 h, 5 h, 6 h, 7 h, 8 h
and 24 h after induction.
Figure 2.12 shows that a protein of approximately the correct mass for GFPuv_WT
(32,500 Da) has been successfully over expressed at all three temperatures tested and is
the main product of the cell cultures. Production of the protein appears to occur faster at
higher temperatures. At 22°C there is a continuous increase in the levels of protein
expressed, but there seems to be little increase in the in the amount of protein produced
at 30°C after 8 hours or at 37°C after 5 hours.
The Western blots in figure 2.13 show that the over expressed protein is successfully
labelled with a hexahistidine tag. It appears that at both 22°C and 30°C there is a
continuous increase in the level of this protein expressed, but at 37°C the increase much
quicker in the first few hours. Also at 37°C there appears to be more prominent bands of
different masses, which suggest an increased amount of incompletely expressed
proteins.
CHAPTER 2
75
Figure 2.13 Western blots of the samples taken from cultures expressing GFP_WT at 22, 30 at 37°C.
From left to right the lanes correspond to 0 h, 1 h, 2 h, 3 h, 4 h, 5 h, 6 h, 7 h, 8 h and 24h after induction.
For a more quantitative analysis of the expression at different temperatures the bulk of
the cells from each culture were analysed for protein content before and after
purification (details in section 8.1.3). The total cell mass, total soluble protein produced
and total purified GFPuv obtained at each temperature are summarised in figure 2.14.
Figure 2.14 The mass of GFPuv_WT purified relative to total cell mass and total soluble protein
produced. Data relating to 400 mL cultures grown at 37°C until an OD of 0.7 was reached followed by
protein expression at different temperatures.
Under the conditions tested there was very little difference in the amount of GFPuv_WT
produced following induction at 30°C and 37°C over 24 hours, although the production
seemed to be much faster at 37°C. Lowering the temperature to 26°C had a more
0
200
400
600
800
1000
1200
26°C 30°C 37°C
Mas
s re
cord
ed
(m
g)
Induction Temperature
Expression of GFP_WT
Total Cell
Total Soluble Protein
Total purified GFPuv
CHAPTER 2
76
significant impact of the amount of GFPuv_WT produced over this time so was not
considered for bulk production. Cysteine mutants are likely to fold less efficiently than
GFPuv rather than more efficiently and that protein folding is generally reported to be
more efficient at lower temperatures. Therefore it was decided that a longer induction
time (up to 24 hours) at 30°C would be more suitable that a shorter induction time (3-5
hours) at 37°C, which would produce a similar amount of protein but be more likely to
contain a higher percentage of incomplete or missfolded protein.
2.4.2 Expression of Mutants
Initially all of the mutants created by site directed mutagenesis were expressed under
the same conditions as optimised for GFPuv_WT. All of these mutants suffered a
reduction in protein yield to some extent (20-50%), but mutants GFPuv_C48A,
GFPuv_C48A_I229C and GFPuv_I229C showed over 95% reduction in yields
compared to GFPuv_WT. Furthermore GFPuv_E6C_C48A and
GFPuv_E6C_C48A_I229C yielded no detectable GFPuv.
DNA sequence analysis revealed that some mutants contained an increased amount of
rare codons when compared to GFP_WT. After transformation in to Rosetta 2
competent cells the yield of GFPuv_I229C was increased over ten fold. The mutants
containing the C48A mutation did not show an improvement in yield upon the same
treatment suggesting the mutation itself was detrimental to expression or folding.
Another possibility is that when the C48A mutation was introduced a mutation in the T7
promoter region occurred and was subsequently passed on to the mutants derived from
it.
As C48 was later found to be unreactive to iodoacetamides, these mutants became
superfluous so no further investigation in to the lower protein yields was undertaken. To
reduce the chances of any further expression problems all further mutants produced
(from DNA shuffling) were derived from codon optimised GFPuv. The expression of
CHAPTER 2
77
these mutants was never as efficient as that of GFPuv_WT under the same conditions,
but was deemed sufficient for our needs. Typical yields of these mutants were 10-30 mg
of purified protein per 400 mL culture, compared to over 100 mg for GFPuv_WT.
2.5 Purification of GFPuv
2.5.1 Gradient Immobilised Metal Affinity Chromatography (IMAC)
GFPuv_WT was purified by nickel chelating chromatography using an imidazole
gradient (details in section 8.1.3). The fractions were collected in a 96-deep well plate
and analysed by SDS-PAGE electrophoresis. The fractions collected from this
purification and the SDS-PAGE Gel of some fractions of interest are shown in figure
2.15.
Figure 2.15 A) SDS gel of selected fractions collected after IMAC column. B) The 96-deep well plate in
which the samples were collected. C) The 96-deep well plate in which the samples were collected
illuminated by blue light and viewed through a light filter.
As shown in figure 2.15 fractions A2-A6 showed some florescence, but contained a
large mixture of protein and little GFPuv_WT. Fractions E12-F3 contained relatively
large amounts of GFPuv_WT with few impurities, however some impurities could be
seen in all fractions.
2.5.2 Anion Exchange Column
It was envisaged that we may need GFPuv_WT mutants at a higher purity than was
achieved using IMAC so other purification methods were investigated. The eluent from
CHAPTER 2
78
wells E12-F3 (figure 2.15) were combined, purified using an anion exchange column
(details in section 8.1.3) and the eluted fractions collected for analysis. Fractions
collected show similar levels of purity to that after IMAC (figure 2.16.A). While the
relative concentration of GFPuv_WT appears higher, the concentrations of impurities
had also increased slightly suggesting this purification step was of little use. Anion
exchange chromatography was not used again for GFPuv purification.
2.5.3 Size Exclusion Chromatography
The GFPuv_WT containing fractions eluted from the anion exchange column were
combined, purified using size exclusion chromatography (details in section 8.1.3) and
the eluted fractions were collected for analysis (figure 2.16.B). The purity of the protein
collected was significantly increased using this method although size exclusion
chromatography takes longer than anionic affinity chromatography and the yields were
significantly reduced.
Figure 2.16 A) SDS-PAGE gel of fluorescent fractions collected from anion exchange chromatography
of GFPuv_WT. B) SDS-PAGE gel of fluorescent fractions collected from size exclusion chromatography
of GFPuv_WT.
2.5.4 Stepwise IMAC
It was noticed that two peaks were often seen in the UV trace when eluting from the
IMAC column with a constant imidazole gradient. Using a stepwise gradient was found
to dramatically improve the efficiency of the IMAC column. The best results were
obtained when 10% elution buffer was used until an initial peak was eluted followed by
CHAPTER 2
79
100% elution buffer until all remaining GFPuv was removed from the column. This
method (details in section 8.1.3) was found to work well with all mutants produced as
shown in figure 2.17. A typical UV trace of a GFPuv purification using this method is
shown in appendix 3.
Figure 2.17 SDS of all mutants purified by stepwise IMAC purification. From left to right the lanes
contain 5 μg of mutants B10, C5, D1, C5, D1, D5, F1, F11, G1, G3, S6, E6C and I229C.
MS of proteins purified in this way showed no significant impurities and no dimmer
formation so it was decided that no further purification was required before
glycosylation reactions were undertaken. Size exclusion chromatography was held as a
reserve method for when additional purification was required before use. An example of
a deconvoluted mass spectrum of GFPuv_E6C_I229C is shown in figure 2.18. No ions
were detected above 8% abundance when compared to the molecular ion peak (detected
at 31857 Da). The absorbance and emission spectra were measured for each mutant
prepared in this way and found to be unaffected by the mutations introduced (spectra
shown in appendix 4).
CHAPTER 2
80
Figure 2.18 Deconvoluted MS of GFPuv_E6C_I229C from 10000-70000 Da after stepwise IMAC.
2.6 Summary
In this chapter we have described the cloning of GFPuv and the addition of a
hexahistidine tag to aid purification. We have also described the production of six
GFPuv mutants by site directed mutagenesis and 11 novel active mutants of GFPuv by
DNA shuffling. Each of the mutants produced differs from WT_GFPuv in the addition
or removal of cysteine residues making them suitable for the synthesis of
neoglycoproteins. These new GFPuv variants include two mutants containing four
additional cysteines and one mutant containing six additional cysteines which are not
present in the WT sequence of GFPuv.
The expression and purification of the GFPuv mutants produced has also been
optimised and a general procedure suitable for the milligram scale production of any of
the mutants produced has been demonstrated. Further details of the prodedures
described in this chapter can be found in section 8.1.
.
81
Chapter 3: Synthesis of Aminoethyl
Glycosides
CHAPTER 3
82
3.1 General Considerations
The conjugation of carbohydrates to biomolecules (such as peptides, lipids and
metabolites) and surface arrays has been achieved through a range of different
methods.168-170
Aminoalkyl linkers have become the most popular due to their stability
and ease of synthesis from commercially available materials. The aminoethyl linker is
one of the most widely used and known to be biocompatible.169,171,172
This linker is also
Figure 3.1 The use of aminoethyl mannoside (27) in carbohydrate arrays and in the synthesis of
glycopeptides. (a) The reaction of amino ethyl mannoside with an activated array surface. (b) The
conversion of amino ethyl mannoside in to an α-halo carbonyl compound capable of reacting with thiols.
(c) The reaction of the activated mannoside (28) with a cysteine (25) containing peptide.
CHAPTER 3
83
easily converted to alpha halo carbonyl compounds which would allow their use on both
carbohydrate arrays and in the synthesis of neoglycoproteins. Figure 3.1 shows the
conversion of aminoethyl mannoside (27) in to mannosyl iodoacetamide (28) by the
reaction with iodoacetic anhydride. The subsequent reaction of 28 with cysteine (25) to
produce a cysteine coupled to a mannose via an amino ethyl linkage is also shown.
An alternative use of the aminoethyl mannoside (27) is in the formation of a mannose
carbohydrate array. This can be achieved by using carboxylic acid terminal linkers on a
gold coated chip as shown in figure 3.1. These carboxylic acid linkers first need
activating using N-ethyl-N’-(dimethylaminopropyl)-carbodiimide (EDC) and N-
hydroxysuccinimide (NHS) before they can successfully be couple to the aminoethyl
linkers.
Aminoethyl glycosides have been prepared via several routes. In many cases the
carbohydrate is first glycosylated with chloroethanol, bromoethanol or azidoethanol
before conversion to the amine.173,174, 175
Alternatively N-Cbz-aminoethanol can be used
directly to avoid a functional group conversion step. Activation of glycosides with
trichloroacetimidate can potentially increases the overall yield (illustrated by steps b-d
in figure 3.2), but the additional two steps involved are difficult to justify in the
synthesis of simple glycosides.176
The direct reaction of a range peracetylated mono and
disaccharides with N-Cbz-aminoethanol (illustrated by step g in figure 3.2) has been
reported with reasonable yields (varying with the glycoside used).177
This is the
approach we have chosen in the synthesis of our aminoethyl glycosides. The products
made by this method also have the desired anomeric configurations that correspond to
naturally occurring terminal glycans, which is essential for a relevant study of
carbohydrate binding interactions.
The example in figure 3.2 shows the synthesis of aminoethyl mannoside (27) from
mannose (29). In the first step (a) the glycoside is protected to form per acetylated
mannose (30). Step g is the coupling of 30 with N-Cbz-aminoethanol to produce the
CHAPTER 3
84
protected aminoethyl mannoside (31) which can be converted to 27 by two deprotection
steps (e and f).
Figure 3.2 Synthesis of aminoethyl mannoside (27). (a) Ac2O in pyridine. (b) BnNH2 in THF. (c)
Cl3CCN, K2CO3 in Dichloromethane (DCM). (d) N-Cbz-aminoethanol, TMSOTf in DCM. (e) NaOMe in
MeOH. (f) Pd/C, H2 in MeOH. (g) N-Cbz-aminoethanol, BF3.Et2O in DCM.
3.2 Synthesis of Monosaccharides
Mannose (29) was peracetylated using acetic anhydride in pyridine to yield 30 as
mixture of both anomers. This mixture of anomers was then coupled with N-Cbz-
aminoethanol in the presence of BF3·Et2O to form the protected aminoethyl mannoside
(31). This glycoside was first deacylated by treatment with sodium hydroxide in
methanol and then the carboxybenzyl group was removed by hydrogenation using a
Pd/C catalyst to yield the deprotected aminoethyl mannoside (27) (details in section
8.2.4).
Figure 3.3 Structures of aminoethyl glucose (32), aminoethyl galactose (33) and aminoethyl GlcNAc
(34).
CHAPTER 3
85
The synthesis of the corresponding aminoethyl glycosides for Glc (32), Gal (33) and
GlcNAc (34) were carried out in a similar fashion to that of the aminoethyl mannoside
(27). In these three cases the enantiomerically pure, β-peracetylated glycosides were
deemed economically viable starting materials so the initial acetylation step was not
required. The final products of these syntheses (32-34) are shown in figure 3.3. The
main variation in the synthesis of glycosides 32-34 was in the coupling step with N-
Cbz-aminoethanol in which the yields were different for each monosaccharide. Also a
different Lewis acid (SnCl4) was used for the coupling of peracetyl glucosamine (35) to
N-Cbz-aminoethanol as it is reported to give preferable yields to BF3·Et2O,177
which
was used in the case of the other monosaccharides. A summary of the coupling
reactions of all of the peracetylated monosaccharides (30, 35-37) to produce the
corresponding protected aminoethyl glycosides (31, 38-40) is shown in figure 3.4. The
final deprotection steps were identical for each of the monosaccharide, aminoethyl
glycosides produced (details in section 8.2.5-7).
Figure 3.4 Summary of glycosylation reactions performed on peracetylated monosaccharides (30, 35-37).
CHAPTER 3
86
3.3 Synthesis of Aminoethyl Trimannoside (41)
The synthesis of trimannoside (41) (shown in figure 3.5) was undertaken to provide a
branched glycoside suitable for attachment to a target protein or a carbohydrate array.
This structure was specifically chosen due to its similarities to the N-glycan trimannose
core and hence the potential for further elaboration with glycotransferases. The
inspiration for this synthesis was taken from previous work undertaken by Kaul and
Hindsgaul in 1991,178
which negated the use of difficult and time consuming protecting
group chemistry. When compared to a traditional chemical synthesis of the same
compound179
this approach was significantly faster, taking only two weeks as opposed
to two months and used much more robust techniques.
The GlcNAc-GlcNAc section of the Man5GlcNAc2 core is reportedly not required for
GlcNAc-transferase I (GnT-I) activity and the substitution of these two sugars for an
alkyl group has been shown to increase the Km of this enzyme for some substrates.180
A
further simplification of the reported method was to disregard the synthetically
challenging β-mannose linkage. It was hoped that this linkage would also be
nonessential for GnT-I activity and would allow the use of aminoethyl mannoside (42)
in the synthesis. Stereoselectivity for α-glycosidic bond formation is inherent in the use
of Man due to both the anomeric and neighboring group participation effects which
further simplifies the synthesis.
Figure 3.5 Retrosynthetic analysis of trimannoside (41), demonstrating how it can be synthesised from
aminoethyl mannoside (42) and a mannosyl donor.
CHAPTER 3
87
The retrosynthetic analysis in figure 3.5 demonstrates how trimannoside 41 could be
made using aminoethyl mannoside 42 and a mannosyl donor. The challenging part of
this synthesis is the selective glycosylation at the 3 and 6 positions of the mannosyl
acceptor (carbon numbering shown in figure 3.6). A similar reaction has been reported
without the use of protecting groups with the overall yield of 17%.178
This is a relatively
low yield for a three step synthesis. However, the alternative is the use of protecting
groups on hydroxyl groups 2 and 4 of the mannosyl acceptor which would require
several additional steps and the purifications which would take significantly longer and
likely not improve the overall yield.
Figure 3.6 Structure of mannosyl acceptor (42) with carbon numbering labelled in red.
This approach depends on the increased reactivity of position 6 and 3 over the other
hydroxyl groups on the mannose acceptor. Reaction at the 6 position is highly favoured
as it is a primary alcohol. Reaction at position 1 would also be favoured, but in this case
it is prevented by the aminoethyl linker. Position 2, 3 and 4 are likely to have similar
reactivities, but position 3 should be favoured due to steric factors affecting the other
two hydroxyl groups. The hydroxyl group at position 2 is hindered by 1,3-diaxial
interactions unlike groups 3 and 4. Also position 4 will experience a larger steric effect
from the mannose attached to the 6 position than position 3. Of course the preference of
the 3 position over 2 and 4 is not enough to produce the desired product only. However
the desired 3,6-linked trimannoside (43) is the only trimannoside which does not
contain vicinal diols. Therefore the two unwanted trimannoside side products (4,6-
linked (44) and 2,6-linked (45)) can be removed by reaction with sodium periodate as
shown in figure 3.7, followed by silica chromatography. This methodology should also
affect the removal of any dimannose compounds which will also be produced.
CHAPTER 3
88
Figure 3.7 The reaction of sodium periodate with a mixture of trimannosides (43-45)
Separation of trimannosides from any tetra or pentamannoside produced would present
the main challenge for this synthesis. The proportion of mannosyl donor could be kept
low enough to prevent the formation of these unwanted side products, removing the
need for further purification, but this would reduce yield of trimannose in favour of
dimannose.
Acetobromo mannose (46) was chosen as the mannosyl donor for the synthesis of
trimannose (41) because it had previously been reported to react successfully in a
similar synthesis.178
46 was successfully synthesised from peracetylated mannose (30)
and HBr as shown in figure 3.8 and was then immediately used for the glycosylation
reaction with 42.
CHAPTER 3
89
Figure 3.8 The synthesis of acetobromo mannose (46) from peracetyl mannose (30).
Aminoethyl mannoside 42 required heating (35°C) and sonication for it to fully dissolve
in acetonitrile before it’s coupling with 46. The reaction was catalysed using HgBr2/
Hg(CN)2 (figure 3.9) and molecular sieves were used to maximise yield. Initial
purification of this reaction by silica chromatography removed the majority of the
unreacted monosaccharide and disaccharide side products.
Figure 3.9 The reaction of aminoethyl mannoside (42) with acetobromo mannose (46) to form a mixture
of mannosides including the trimannoside (43).
This trimannoside rich mixture was treated with sodium periodate for 48 hours and then
further purified by silica chromatography to yield pure trimannoside. The yield of this
synthesis so far with respect to the aminoethyl mannoside (42) was just over 4%.
However considering the large amounts of mannose wasted on forming dimannosides
and the fact that no tetramannosides were detected in the reaction mixture by MS it was
thought that this yield could be improved upon by increasing the proportion of
acetobromo mannose used. The optimisation of this reaction was undertaken by
MChem student Siak Gee Lim, but the yield of 4% with respect to aminoethyl
mannoside (42) was not improved upon.
CHAPTER 3
90
Figure 3.10 Structure of trimannoside (43) with carbohydrate carbon atoms labelled.
To be certain of the success of the synthesis of trimannoside 41 a full characterisation of
the protected derivative 43 was undertaken with the help of Dr R. Sardzik (The
University of Manchester). The numbering system for the carbohydrate ring carbons use
in the NMR assignments is shown in figure 3.10. Conformation of the stereochemistry
of the glycosidic bonds formed was achieved by analysing the coupling of the anomeric
protons with their respective carbons in a non-decoupled HSQC (shown in figure 3.11).
These couplings are all characteristic of α-glycosides, confirming that the predicted
stereochemistry had been achieved. Conformation that the product was glycosylated at
the 3 position and not at position 2 or 4 was found in the coupling of C3 to H1’ in the
HMBC spectrum (relevant section shown in figure 3.12).
CHAPTER 3
91
Figure 3.11 A section of a non-decoupled HSQC of trimannoside (43) showing the coupling of the
anomeric protons with their respective carbons.
Figure 3.12 Section of a HMBC spectrum showing the coupling of C3 to H1’ in trimannoside (43).
CHAPTER 3
92
Figure 3.13 The deacylation of trimannoside (43) to trimannoside (47) using sodium methoxide in
methanol.
Deacetylation of trimannoside (43) to produce 47 was achieved as previously described
for the aminoethyl monosaccharides, using sodium hydroxide in methanol as shown in
figure 3.13. However attempts to hydrogenate trimannoside (47) as previously
described for the aminoethyl monosaccharides (using ethanol or methanol as the
solvent) did not result in a pure product. If both the amount of catalyst and the time of
the reaction were carefully controlled then trimannoside (41) could be obtained as the
main product, but usually with a mixture of adducts also present. Performing the
reaction in water (as shown in figure 3.14) was found to be faster and produced the pure
deprotected mannoside desired. The formation of increasing amounts of side products
with increased reaction times in ethanol or methanol was possibly due to the poor
solubility of trimannoside (41) in these solvents.
Figure 3.14 The hydrogenation of trimannoside (47) to trimannoside (48) using a Pd/C catalyst in water.
CHAPTER 3
93
3.4 Synthesis of Tetramannosides (48 and 49) and Pentamannoside (50)
Increasing the proportion of mannosyl donor (46) compared to mannosyl acceptor (42)
in the glycosylation reaction shown in figure 3.9 increases the probability of forming
tetramannosides (48 and 49) and the pentamannoside (50) shown in figure 3.15. It was
decided to attempt the synthesis of a mixture of these larger branched mannosides as
they could also be used on the carbohydrate arrays and potentially in the production of
glycoproteins. Altering the reaction conditions in this way would also significantly
reduce the amount of material wasted in the formation of unwanted dimannosides which
would increase the overall yield of the synthesis.
Figure 3.15 Structures of tetramannosides (48 and 49) and pentamannoside (50).
The synthesis of 42 and 46 were carried out in the same manner as previously
described, however in the glycosylation step of the synthesis 7 mole equivalents (mol
eq) of acetobromo mannose (46) were used instead of 2.5 mol eq used in the synthesis
of trimannoside (41). The reaction was left stirring for 1 hour at 35°C. Initial
purification of this reaction by silica chromatography resulted in a fraction rich in all of
CHAPTER 3
94
the branched mannosides predicted (43-45, 48-50). This mixture of polymannosides
was treated with sodium periodate to remove the unwanted trimannosides (44 and 45)
and dimannosides, however no solvent system was found to adequately separate the
polymannosides (43, 48-50) on silica.
A small portion of this mixture was deacylated and purified by exclusion
chromatography trials using Bio-Gel. Some fractions did yield purified mannosides
however only very small amounts of product could be separated on each column and the
method was extremely time consuming.
After a solvent system was optimised it was possible to purify each of the
polymannosides simultaneously by HPLC (details in section 8.2.9). This method was
faster than size exclusion chromatography and much larger sample sizes could be
purified per run. A typical trace UV trace for the purification is shown in appendix 5.
This method could also be scaled up and improved upon if a preparative HPLC
instrument was employed. There was some slight separation observed between the two
tetramannosides (48 and 49) at lower column loadings, however for speed of
purification these were collected as a mixture.
Yield of trimannoside (43), tetramannosides (48 and 49) and pentamannoside (50) in
relation to aminoethyl mannoside (42), were 10.2%, 12.9% and 2.7% respectively. The
main losses in yield were through unwanted dimannoside and trimannoside formation.
It is also worth noting that approximately 90% of the acetobromo mannose was lost
through various side products. However, considering the complexity of the products
produced in just one glycosylation step, an overall yield of 25.8% with respect to
aminoethyl mannoside (42) was considered acceptable.
CHAPTER 3
95
3.5 Activation of Glycosides for Glycosylation of Cysteines
Figure 3.16 Conversion of aminoethyl mannoside (27) in to mannosyl iodoacetamide (28) via reaction
with iodoacetic anhydride in sodium bicarbonate buffer.
Each aminoethyl glycoside used for protein derivatisation was activated for reaction
with cysteine by reaction with iodoacetic anhydride as shown in figure 3.16 (full detils
in section 8.2.8). In addition to the glycosides mention in this chapter aminoethyl
lactose (51) donated by Dr R. Sardzik (The University of Manchester), shown in figure
3.17, was also converted to its corresponding glycosyl iodoacetamide.
Figure 3.17 Structure of aminoethyl lactose (51) donated by Dr R. Sardzik (The University of
Manchester).
Product formation was confirmed by HRMS (summarised in table 3.1) and then stored
at -20°C to avoid additional hydrolysis of the products. The structures of all glycosyl
iodoacetamides produced in this way (28, 52-55) are shown in figure 3.18. To maximise
yield it was important to keep the reaction time and time between desalting and
lyophilisation a short as possible to reduce product hydrolysis. Care was taken to
maintain a pH below 8 as a higher pH also resulted in reduced yields. An m/z peak
corresponding to the hydrolysed product was usually observed, but yields of over 90%
were possible.
CHAPTER 3
96
Figure 3.18 structures of glycosyl iodoacetamides produced; mannosyl iodoacetamide (28), glucosyl
iodoacetamide (52), galactosyl iodoacetamide (53), glucosamine iodoacetamide (54) and lactosyl
iodoacetamide (55).
Calculated Mass (Da) Measured Mass (Da) Glycosyl
Iodoacetamide
[M + H]+ = 392.0206
[M + Na]+ = 414.0026
[M + Na]+ = 414.0025 28
[M + H]+ = 392.0206
[M + Na]+ = 414.0026
[M + Na]+ = 414.0025 52
[M + H]+ = 392.0206
[M + Na]+ = 414.0026
[M + Na]+ = 414.0025 53
[M + H]+ = 433.0472
[M + Na]+ = 455.0291
[M + Na]+ = 455.0292 54
[M + H]+ = 554.734
[M + Na]+ = 576.0554
[M + Na]+ = 576.0556 55
Table 3.1 Summary of HRMS data of glycosyl iodoacetamides produced; mannosyl iodoacetamide (28),
glucosyl iodoacetamide (52), galactosyl iodoacetamide (53), glucosamine iodoacetamide (54) and
lactosyl iodoacetamide (55).
3.6 Summary
In this chapter we have described the synthesis of four biologically relevant
monosaccharide aminoethyl glycosides (27, 32-34) and a trisaccharide aminoethyl
glycoside (41) similar to the N-linked glycan trimannose core. All of these glycosides
are suitable for the use on carbohydrate arrays or conversion to their corresponding
glycosyl iodoacetamides for use in the synthesis of neoglycoproteins. All of these
CHAPTER 3
97
glycosides are also biologically relevant due to their anomeric configurations being the
same as naturally occurring terminal glycans. Full details of all of the syntheses
discussed in this chapter can be found in chapter 8.2. The synthesis and purification of
tetramannosides (48 and 49) and pentamannoside (50) was also explored.
98
Chapter 4: The Glycosylation of GFPuv
Mutants
CHAPTER 4
99
4.1 General Considerations
α-halocarbonyl compounds are straightforward to prepare and can be used to create
linkages, similar to that of naturally occurring N-linked glycans.181,182
These compounds
have been widely used in the synthesis of neoglycoproteins and require only that the
acceptor protein contains a cysteine at the desired glycosylation site. Whilst chloro and
bromoacetamides are more stable to hydrolysis and provide an increased
selectivity,125,132,183
we have opted for the use of the more reactive and more commonly
used iodoacetamides.130,181,184
Figure 4.1 shows the reaction of an iodoacetamide with a
peptide containing a cysteine residue under the conditions originally reported.
Figure 4.1 The chemical glycosylation of a peptide with an iodoacetamide, under conditions originally
reported.181
In previous chapters the synthesis of a range of glycosyl iodoacetamides for the
chemical glycosylation of proteins has been described. We have also discussed the
production of multiple cysteine mutants of GFPuv. In this chapter we bring these two
elements together to produce neoglycoproteins for the use in lectin binding assays. It is
essential that we have homogenously glycosylated glycoproteins to obtain meaningful
results, therefore it is crucial that the glycosylation reaction can be accurately
monitored. Electrospray ionisation mass spectrometry (ESI-MS) is a “soft” ionisation
method and very accurate, making it the method of choice for protein and glycoprotein
mass spectrum analysis.
4.2 Analysis of Protein samples by ESI-MS
Benchmark spectra were acquired using horse heart myoglobin (HHM). HHM is a very
stable protein, commercially available and a commonly used calibrant for ESI. A
sample spectrum acquired for this protein is shown in figure 4.2.A and the
CHAPTER 4
100
corresponding deconvoluted mass spectrum, produced by the software is also shown
(figure 4.2.B).
Figure 4.2 A) The acquired spectrum of horse heart myoglobin (HHM), containing the multiply charged
protein peaks of the “charge envelope.” B) The deconvoluted spectrum of HHM produced by MassLynx
4.0.
Sample preparation was very important, to obtain clean, accurate mass spectra of
GFPuv mutants by direct injection MS. The first consideration was that the samples
were sufficiently pure to obtain a clear spectrum. This involved removal of buffer salts,
imidazole (from purification by IMAC) and unreacted iodoacetamides, in the case of
derivatised samples. The second consideration was that the protein concentration was
high enough to provide a sufficient noise to background ratio for the data processing to
be successful.
CHAPTER 4
101
Initially samples of GFPuv were prepared by precipitation in acetone/methanol
followed by resuspended in deionised water (details in section 8.3.2). Direct injection of
the resulting solution resulted in mass spectra which could be used to accurately
determine the molecular mass of GFPuv mutants. Table 4.1 shows examples of the
accuracy achieved with this method and the measured mass of HHM for the
corresponding calibration.
Protein Species Calculated Mass (Da) Measured Mass
(Da) Mono-isotopic Average
WT 32463.86 32484.31 32485.2
E6C 32437.82 32458.34 32458.6
C48A 32431.89 32452.25 32452.7
I229C 31852.49 31882.63 31883.0
HHM 16951.49 (Literature value (Lit.))185
16951.6
Table 4.1 Calculated and measure mass values for GFPuv mutants and horse heart myoglobin (HHM).
Preparing protein samples in this way was useful for initial MS experiments, mainly
because it provided concentrated and significantly desalted samples in one step.
However, these samples often did not completely redissolve and the yields of protein
obtained were therefore inconsistent. The spectra obtained were not as clear as those of
HHM at the same concentration, which was assumed to be partly due to the mass
difference of the proteins, but incomplete desalting was also suspected to be a factor.
To improve the quality of mass spectra obtained of GFPuv samples, different methods
of sample preparation were explored. The use of a PD-10 desalting column was found
to be very successful for buffer exchange of GFPuv samples. Once found to be effective
PD-10 columns were used for buffer exchange before reactions, removal of unreacted
iodoacetamides (to quench reactions) and desalting samples prior to MS analysis. This
method completely removes unwanted salts from samples, which improved the quality
of the mass spectra obtained. The process was reproducible and did not involve the
CHAPTER 4
102
precipitation and problematic resuspension of protein, therefore yields were
quantitative. Buffer exchange also allowed for complete removal of imidazole after
purification so that the concentration of protein could be more easily determined. The
only disadvantage of this method of protein preparation was that a separate
concentration step (using a vivaspin, 10 kDa, spin filter), was often required. Protein
concentration in this manner was often more time consuming than precipitation, but
generally resulted in higher protein yields and was more appropriate for production of
useful glycoproteins.
Reducing Agents
Reducing agents are frequently used on proteins prior to alkylation reactions to prevent
aggregates and maximise the efficiency of derivatisation.1,186
Some of the most
commonly used in the reduction of proteins are dithiothreitol (DDT), 2-mercaptoethanol
and tris(2-carboxyethyl)phosphine (TCEP).
Initial treatment of GFPuv samples with DTT or TCEP showed no change in the mass
spectrum before and after treatment with reducing agent. For this reason it was believed
that reduction of GFPuv mutants before glycosylation was unnecessary. However, after
time it became clear that some samples did not give as clear a spectrum as others and
this could usually be improved by reduction. Example spectra of GFPuv_E6C both
before and after reduction with TCEP are shown in figure 4.3. To maintain consistency
between reactions all samples were treated with TCEP prior to glycosylation.
CHAPTER 4
103
Figure 4.3 A) The ESI mass spectrum of GFPuv_E6C before to treatment with TCEP. B) The ESI mass
spectrum of the same sample of GFPuv_E6C after treatment with TCEP.
TCEP was chosen as our preferred reducing agent because it reacts stoichiometrically
with thiols, in under 10 minutes, at room temperature (r.t.) It is also more stable to
oxidation than other reducing agents such as DTT. TCEP also does not react with metal
ions, making it suitable for use on proteins purified by nickel chelating chromatography.
It was originally believed that TCEP did not react with α-halocarbonyl compounds
although it is now known to do so. However, as such low concentrations are required
for rapid reduction, the removal of TCEP before reactions is not necessary.186-188
CHAPTER 4
104
4.3 Cysteine Reactivity Screen
It was decided to test the reaction of the original GFPuv mutants produced with
commercially available iodoacetamide (56) to conserve the glycosides synthesised in
chapter three. The aim was to determine which of the three cysteines (C6, C48 and
C229) were reactive to α-halocarbonyls and to optimise the reaction (shown in figure
4.4) before undertaking glycoprotein synthesis.
Figure 4.4 The reaction of mutant GFPuv_E6C with iodoacetamide in ammonium carbonate buffer.
Four GFPuv mutants were chosen for this initial screen to assess the reactivity of the
mutants produced by site directed mutagenesis; GFPuv_WT, GFPuv_E6C,
GFPuv_I229C and GFPuv_E6C_I229C. Each of these mutants was purified as
described in section 8.1.3, and treated with TCEP before derivatisation (details in
section 8.3.1). Samples were removed at time points of 0, 1, 2, 5 and 24 hours and
immediately buffer exchanged in to deionised water to quench the reaction. Example
mass spectra taken of samples from the reaction of GFPuv_E6C_I229C with
iodoacetamide (56) are shown in figure 4.5.
CHAPTER 4
105
Figure 4.5 Deconvoluted mass spectra (31700-32100 Da) of samples taken from the reaction of 0.1 mM
GFPuv_E6C_I229C with 1 mM iodoacetamide (56). A) Mass spectrum after 0 hours. B) Mass spectrum
after 1 hour. C) Mass spectrum after 5 hours. D) Mass spectrum after 24 hours.
From these spectra it is evident that only two molecules of iodoacetamide (56) have
reacted with each of the protein species. This suggests two of the cysteines present in
this mutant are unreactive to iodoacetamides under these conditions. It was suspected
that C70 would be unreactive as it is buried in GFPuv’s hydrophobic core. From the
data collected from these initial reactions (summarised in table 4.2) it can be deduced
that C48 is not reactive to iodoacetamides under these conditions either. It should be
noted that the expected mass difference corresponding to the addition reaction of
iodoacetamide (56) would be +57 Da. The fact that none of the additions are exactly
+57 or multiples thereof is due to slight variations in the calibration used before and
after the reactions.
CHAPTER 4
106
GFPuv
Mutant
Measured mass at time point
(Da)
Mass
Difference
(Da)
Cysteines
reacted
0 h 24 h
WT 32486 32488 +2 0
E 32460 32519 +59 1
I 31885 31943 +58 1
EI 31859 31974 +115 2
Table 4.2 Summary of the results of the reaction of four GFP mutants (0.1 mM) with iodoacetamide (56)
(1 mM) over 24 hours. In each case the mass corresponds to the only significant peaks present in the mass
spectra.
4.4 Final Glycosylation Procedure
The unspecific reaction of glycosyl iodoacetamides with proteins has been previously
reported.130
This issue can reportedly be solved by the addition of imidazole to the
reaction buffer. For these reasons and the simplification of our glycoprotein synthesis it
was decided that glycosylation reactions with glycosyl iodoacetamides should be
carried out in the same buffer used to elute proteins from their IMAC columns during
purification. This avoids a buffer exchange step between purification and glycosylation
and also greatly reduces the chances of any unwanted glycosylations which may occur.
The finalised glycoprotein preparation protocol is shown in figure 4.6 and only requires
one buffer exchange step, in which the glycoprotein can be eluted in a suitable buffer
for its desired use (details section 8.3.1).
CHAPTER 4
107
Figure 4.6 The finalised procedure for glycosylation of all GFPuv mutants.
4.5 Analysis of Protein Samples by LCMS
For the preparation of a wide range of neoglycoproteins for screening purposes it was
preferable to have a rapid and reliable method for analysis of products. While the direct
injection method used previously was accurate, it required careful sample preparation
and a large amount of material (50 µg or more) to produce clear spectra. For these
reasons a semi automated liquid chromatography MS (LCMS) method was developed
on a separate instrument.
The Agilent 1100, HPLC system coupled to the Agilent 1100 LC/MSD SL quadrupole
mass spectrometer, allowed for the purification and simultaneous MS and ultraviolet
(UV) measurement of analytes. This allowed buffer salts and other small molecules to
pass through the column prior to MS analysis, increasing the resolution of the spectra
and preventing unnecessary contamination of the detector. Typically only 1 µL of 0.5
CHAPTER 4
108
mg/mL protein sample was required for a clear mass spectrum on this instrument (0.5
μg).
Figure 4.7 is an UV (205 nm) trace from an example LCMS run of a GFPuv mutant in
glycosylation buffer. The UV spectrum shows a large peak (corresponding to buffer
salts) between 1 and 3 minutes after injection of the sample, followed by an interval of
approximately 7 minutes before a second peak is seen. This second peak corresponds to
the protein being eluted from the column into the detector. MS analysis was only carried
out between 10 and 15 minutes after injection of the sample to minimise instrument
contamination.
Figure 4.7 UV (205 nm) trace of a typical LCMS run of a GFPuv mutant.
Data collected in this manner produced a charge envelope of multiply charged protein
species that could be deconvoluted in the same manner as the data obtained by direct
injection. This method removed the need for lengthy sample preparation techniques and
drastically reduced waste protein (over 100 times less protein required for a clear
spectrum). Although each LCMS run took 30 minutes, where as the direct injection
measurements only took 10 minutes per sample, the lack of preparation time and
increased reproducibility more than compensated for this. The masses of each mutant
CHAPTER 4
109
produced were measured using this method (results summarised in table 4.3) and all
future glycosylation reactions were monitored in this way (details in section 8.3.2).
GFPuv
Mutant
Calculated Mass (Da) Measured Mass (Da)
Mono-isotopic Average
WT 32463.86 32484.31 32481.58
E 32437.82 32458.34 32456.00
I 31852.49 31882.63 31878.14
EI 31836.46 31856.65 31854.35
B10 32426.75 32447.33 32447.86
C5 32458.67 32479.41 32477.85
D1 32421.81 32442.23 32438.10
D5 32435.81 32456.32 32453.70
F1 32447.86 32468.31 32467.70
F11 32438.77 32459.28 32456.89
G1 32411.74 32432.21 32430.51
G3 32410.78 32431.27 32432.16
S6 32432.63 32453.32 32451.72
HHM 16951.49 (Lit.)185
16951.28
Table 4.3 Calculated and measure mass values for GFPuv mutants and HHM using an Agilent 1100,
HPLC system coupled to an Agilent 1100 LC/MSD SL quadrupole mass spectrometer.
4.6 Production of Neoglycoprotein Library
For initial lectin binding assays a range of glycoproteins were synthesised containing 1-
4 glycans at various positions on the GFPuv scaffold. All selected mutants were
glycosylated using the method previously described in this chapter (details in section
8.3.1) with both mannosyl iodoacetamide (28) and galactosyl iodoacetamide (53).
Samples were then analysed by LCMS before glycosylation to confirm their purity and
CHAPTER 4
110
again after glycosylation (details in section 8.3.2). If the reactions were found not to be
complete, additional glycosyl iodoacetamide was added and the reaction left to continue
for a further hour before further analysis.
The mass spectra and deconvoluted spectra of GFPuv_C5 before and after glycosylation
with mannosyl iodoacetamide (28) are shown in figure 4.8. The observed mass
difference is 1059 Da, which fit well with the predicted increase of 1052 Da which
would correspond to the addition of four mannosyl moieties. This was the result we
were hoping for because it means that all four cysteine introduce are reactive in this
mutant.
Figure 4.8 A) Measured mass spectrum of GFPuv_C5. B) The deconvoluted molecular ion peak of
GFPuv_C5. B) Measured mass spectrum of GFPuv_C5 after reaction with mannosyl iodoacetamide (28).
D) The deconvoluted molecular ion peak of GFPuv_ C5 after reaction with mannosyl iodoacetamide (28).
CHAPTER 4
111
4.7 Glycosylation of Lysines
Although the main aim of this project was to produce homogeneous neoglycoproteins
by glycosylating GFPuv mutants via their cysteines, it was recognised that
glycosylating all of GFPuv’s lysines could also produce useful products. GFPuv has 16
surface lysines and glycosylating all of these residues would yield a glycoprotein that
could be screened alongside the proteins glycosylated via their cysteines. The
interactions of these unspecifically glycosylated products with carbohydrate binding
proteins could to provide interesting results when compared to the interactions of our
specifically glycosylated glycoproteins. Glycoprotein made in this manner could also
potentially be used in cell labelling studies. It is of course possible to synthesise
glycosides with linkers reactive to primary amines by a variety of methods.121
However,
it was decided to explore methods using our aminoethyl glycosides for reaction with
lysines to avoid additional syntheses.
The possibility of increasing the reactivity of lysine residues towards iodoacetamides by
first activating them with another linker was explored. It was decided that derivatising
lysines so that they become cysteine mimics would be our favoured approach. This
facilitated the used of our existing glycosylation protocol and only required a short
additional step in glycoprotein synthesis. 3,3 -dithiobis (sulfosuccinimidylpropionate)
(DTSSP) (57) is a commercially available, water soluble, reversible, crosslinker
(structure shown in figure 4.9) that was found to suit our requirements.
Figure 4.9 The structure of 3,3 -dithiobis(sulfosuccinimidylpropionate) DTSSP. Produced by
thermo scientific as a reversible protein crosslinker.
GFPuv_WT samples were reacted with DTSSP as recommended in the manufacturer’s
instructions. The upper limit of the recommended DTSSP concentrations were used in
CHAPTER 4
112
conjunction with relatively low GFPuv concentrations to ensure the maximum amount
of lysines were derivatised and reduce the probability of intermolecular crosslinking
which would result in aggregation. Some precipitate was observed, but this always
redissolved once the crosslinkers were reduced.
DTT (50 mM) or 2-mercaptoethanol (5% volume) were recommended to reduce the
crosslinkers, but TCEP (10 mM) was found to be much more effective. However it was
essential that extra buffering capacity be introduced to the solution before the addition
of TCEP or the solution became too acidic and the protein would denature. Typically
the addition of 10% volume of 500 mM Tris buffer was sufficient. After reduction of
the crosslinkers the protein solution could be buffer exchanged into IMAC elution
buffer ready for glycosylation as described previously. The general scheme for the
modification of lysine with TCEP followed by reduction to produce a thio modified
lysine is shown in figure 4.10 (details in section 8.3.1).
Figure 4.10 The reaction of GFPuv with DTSSP followed by the reduction of the disulfides within the
crosslinkers by TCEP to give thiol modified lysines.
Analysis of these products by LCMS was unsuccessful, likely due to a mixture of
several different protein species with varying numbers of crosslinkers and glycosides
attached. Example MALDI spectra of these products are shown in figure 4.11 along
with the MALDI spectra of the starting material (GFPuv_WT) for comparison. The
average increase in mass after treatment with DTSSP followed by reduction with TCEP
was 1079 Da. This corresponds to the modification of an average of 12.16 lysines per
protein. The average increase in mass after this protein sample was subsequently treated
with mannosyl iodoacetamide (28) was 2829 Da. This corresponds to the average
addition of 10.75 mannoside residues to each protein. After each reaction the mass
CHAPTER 4
113
peaks became much broader, indicating an increase in homogeneity, which was
expected given the nature of these modifications. The efficiency of these reactions was
considered acceptable for the production of polyglycosylated GFPuv reference samples,
for use in initial assays. GFPuv samples derivatised with approximately 10 glycosides
were expected to provide a significant binding comparison to the more defined
glycoproteins containing between one and four glycoside modifications. Samples
prepared in this way were termed GFPuv_CL (cross linked GFPuv).
Figure 4.11 MALDI spectra of GFPuv_WT (blue), GFPuv_CL (GFPuv_WT after treatment with DTSSP
followed by reduction with TCEP) (red) and GFP_CL_Man10 (GFPuv_WT derivatised with
approximately 10 mannosides) (green).
4.8 Summary
In this chapter we have described a simple method for the analysis of protein samples
by direct injection MS. We have also described a more sensitive and efficient method of
analysing proteins by LCSM. This method has been shown to be suitable for the
CHAPTER 4
114
monitoring of protein glycosylation reactions and determining the level of glycosylation
present in the products. LCMS analysis greatly reduces the amount of protein needed
for analysis while at the same time eliminating the need for additional purification steps
before analysis of samples.
The optimisation of a reproducible glycosylation procedure of cysteines with
iodoacetamides has been described in this chapter. The final procedure eliminates the
need for buffer exchange after protein purification by IMAC and only requires the use
of a Pd-10 desalting column to quench and purify the reaction mixture. In addition to
the selective glycosylation of cysteines a method for the non specific glycosylation of
lysines using iodoacetamides has also been described. This method has been shown to
produce GFPuv samples with an average of over 10 glycosylations per protein molecule
using GFPuv_WT (which contains no reactive cysteines). Further details of the
procedures described in this chapter can be found in section 8.3.
115
Chapter 5: The Enzymatic Modification of
Glycosides
CHAPTER 5
116
5.1 Glycotransferase Screening on Trimannoside (41)
The synthesis of trimannoside (41) was undertaken to provide polyvalent glycosides
suitable for microarray or glycoprotein attachment if required. Trimannoside (41) also
provided a potential substrate for some glycotransferases. If this glycoside was accepted
as an N-glycan core mimic then it could provide a route to more complex glycan
syntheses or a screening tool for novel enzymes. It is known that the diacetylchitobiose
tail of N-glycans is not essential for N-acetylglucosaminyltransferase I (GnT-I) activity,
but whether a β-linked mannose is required or how the aminoethyl linker would be
tolerated remained unknown.178,180
5.1.1 Screening of Mannosides Against Glycotransferases
Figure 5.1 Structures of aminoethyl mannosides (27 and 41) used for glycotransferase screening.
It was decided to undertake the initial screening of trimannoside (41) and aminoethyl
mannose (27) (structures shown in figure 5.1) using GnT-I (donated by Dr. S.
Gluchowska, Trinity College Dublin) and protein-O-mannose N-
acetylglucosaminyltransferase I (POMGnT-I) (donated by Dr P. Both, University of
Manchester) to assess their potential as glycotransferase substrates. GnT-I occurs
naturally in the Golgi of eukaryotic cells and transfers a GlcNAc moiety from UDP-
GlcNAc to an α-1,3-linked mannose of the Man5GlcNAc2 (58) N-linked glycan core to
produce structure 59 (Man5GlcNAc3) (reaction shown in figure 5.2).189
CHAPTER 5
117
Figure 5.2 The natural action of N-acetylglucosaminyltransferase (GnT-I) on the N-glycan core structure
Man5GlcNAc2 (58).
POMGnT-I is an extracellular enzyme that transfers a GlcNAc moiety from UDP-
GlcNAc to O-linked mannopeptides (reaction shown in figure 5.3).190
Neither enzyme
was expected to react with aminoethyl mannose (27) and POMGnT-I was not expected
to react with either mannoside. However we were hopeful that GnT-I would show some
activity towards the trimannoside (41) as some activity had been observed against
similar trimannosides [mannotriose-di-(N-acetyl-D-glucosamine) (60) and 3α,6α-
mannotriose (61) (figure 5.4)].
Figure 5.3 The natural action of protein-O-mannose N-acetylglucosaminyltransferase I (POMGnT-I) on
an O-linked glycopeptides.
CHAPTER 5
118
Figure 5.4 Known substrates for GnT-I (donated by Dr. S. Gluchowska, Trinity College Dublin).
Mannotriose-di-(N-acetyl-D-glucosamine) (60) and 3α,6α-mannotriose (61) (commercially available
from Sigma).
Matrix assisted laser desorption ionisation-time of flight (MALDI-TOF) MS of
immobilised carbohydrates has previously been used to monitor enzymatic reactions.191
The sensitivity of MALDI-TOF MS means that only small amounts of material are
required for each assay making it well suited for the use of hard to synthesise
compounds such as polysaccharides.
Gold plates coated in alkanethiol spacers (HS-(CH2)17-EG3-OH) and linkers (HS-
(CH2)17-EG6-OCH2COOH) can be used to produce a self assembled monolayer (SAM),
which can then be derivatised with amine fictionalised carbohydrates. These SAMs can
be tailored to produce carbohydrate arrays of different densities by varying proportion
of the carboxylic acid terminal (linker) molecules present in the layer.
Gold plates were coated with SAMs containing alkanethiol spacers and linkers in a
ration 4:1 as described in section 8.4.1. Formation of these monolayers was verified by
MALDI-TOF MS. A sample spectrum shown in figure 5.5 has the expected peaks
corresponding to a successful SAM formation. Peak A corresponds to the homodimer of
two spacer molecules linked by a disulphide bond (m/z = 862). Peak B corresponds to
CHAPTER 5
119
the heterodimer of a spacer molecule linked to a linker molecule by a disulphide bond
(m/z 1052).
Figure 5.5 MALDI-TOF spectrum of a 1:4 (linker:spacer) SAM on gold. A = mass peak corresponding to
a spacer-spacer homodimer. B = mass peak corresponding to a spacer-linker heterodimer.
Trimannoside (41) and mannoside (27) were then immobilised on separate parts of a
SAM coated gold plate as described in section 8.4.1 to produce two different
carbohydrate arrays. The reaction of trimannoside (41) with the activated SAM is
shown in figure 5.6.
CHAPTER 5
120
Figure 5.6 The reaction of trimannoside (41) with an activated SAM on a gold plate to form a
carbohydrate array.
MALDI-TOF MS analysis of the carbohydrate arrays produced showed the expected
mass peaks A (m/z = 862) and B (m/z 1052) corresponding to the SAMs and additional
peaks corresponding to the addition of the carbohydrate structures to some of the linker-
spacer heterodimers. MALDI-TOF MS spectra of the trimannoside (41) and mannoside
(27) arrays are shown in figures 5.7 and 5.8 respectively. Attachment of both glycosides
was found to be successful and clean. In figure 5.7 the hetrodimer coupled to
trimannoside (41) is labeled peak C (m/z = 1581). In figure 5.8 the heterodimer linked
to mannoside (27) is labeled D (m/z = 1257).
CHAPTER 5
121
Figure 5.7 MALDI-TOF MS spectrum of trimannoside (41) carbohydrate array. A = mass peak
corresponding to a spacer-spacer homodimer. B = mass peak corresponding to a spacer-linker
heterodimer. C = peak corresponding to a heterodimer covalently bound to trimannoside (41).
CHAPTER 5
122
Figure 5.8 MALDI-TOF MS spectrum of mannoside (27) carbohydrate array. A = mass peak
corresponding to a spacer-spacer homodimer. B = mass peak corresponding to a spacer-linker
heterodimer. D = peak corresponding to a heterodimer covalently bound to mannoside (27).
To test the enzyme activity of GnT-I and POMGnT-I against the carbohydrate arrays
produced they were incubated with solutions of GnT-I or POMGnT-I in a solution of
UDP-GlcNAc as described in section 8.1.4. The MALDI-TOF MS spectrum of a typical
trimannoside (41) carbohydrate array both before and after treatment with GnT-I is
shown in figure 5.9. In 5.9A the only two peaks correspond to the mass of the
trimannoside (41) attached to the hetrodimer (m/z = 1559) and the corresponding
sodium adduct (m/z = 1581). After treatment with GnT-I and UDP-GlcNAc the
predicted peak corresponding to the addition of a GlcNAc moiety to trimannoside (41)
attached to the heterodimer (m/z = 1762) and its sodium adduct (m/z = 1784) were
observed (Figure 5.9.B). Dr P. Both (The University of Manchester) collaborated with
these enzymatic reaction and assisted with the preparation of the carbohydrate arrays.
CHAPTER 5
123
Figure 5.9 A) MALDI-TOF spectrum of trimannoside (41) attached to a SAM on gold. B) MALDI-TOF
spectrum of trimannoside (41) attached to a SAM on gold after treatment with GnT-I and UDP-GlcNAc.
R= SAM spacer-linker hetrodimer.
POMGnT-I did not modify trimannoside (41) under the conditions used and neither
POMGnT-I or GnT-I modified mannoside (27). Although only a small proportion of the
trimannoside (41) was seen to react (figure 5.9.B), the reactivity was comparable to that
of commercially available mannotriose-di-(N-acetyl-D-glucosamine) (60) under the
same conditions, which suggests the β-mannose linage is not essential for GnT-I
recognition.
5.1.2 Screening of Mannosides Against Yeast Microsomal Extracts
Much is known about the N-linked glycosylation pathways in eukaryotes, but the
specificity of glycosidases and glycotransferases with respect to the trimannoside (41),
when immobilised on a carbohydrate array are unknown. Identifying any other enzymes
which show activity on the trimannoside (41) itself or the product of its reaction with
GnT-1 would provide a synthetic route to more complex glycosides. However,
purifying and screening every known candidate would be extremely time consuming.
CHAPTER 5
124
Crude microsomal extracts from yeasts or higher organisms have previously been used
to modify synthetic glycans and glycoconjugates.192-195
These extracts contain a variety
of glycotransferases and can be used as a relatively quick source of active enzymes. If
modification of the mimetic trimannoside core (41) could be achieved using microsomal
extracts we hoped to investigate the effect of different UDP-sugars being available to
the enzymes contained and potentially the organism that the microsomes were extracted
from. Yeast microsomal extracts were prepared by Dr P. Both (The University of
Manchester), who also helped perform the experiments the subsequent experiments on
the SAM-coated gold plates.
Multiple experiments were performed on the immobilised mannoside (27) and
trimannoside (41), in the presence of UDP-GlcNAc, UDP-Glucose, GDP-Mannose or a
mixture of all three. Microsomal extracts were pretreated with swainsonine to inhibit
mannosidases that would be present in the cell extracts and could cleave mannose
residues from the trimannoside (41) or other potential products. However no
modifications were detected under the conditions used (full details section 8.4.1). This
suggest that either the mannosides (27 and 41) are not substrates for any other the
enzymes present in the yeast microsomal extracts or that active enzymes were not
present in high enough concentrations to produce a detectable amount of product.
5.2 Modification of Lactosylated GFPuv Using Tran-sialidase
Sialic acid-binding immunoglobulin-type lectins (Siglecs) play an important role in a
variety of cellular interactions including inflammatory and immune response.196,197
Sialyllactose coated carbohydrate arrays have previously been shown to bind Siglec
expressing cells,191
therefore GFPuv attached to multiple sialyllactose moieties could
potentially be used to label cells in a similar way.
CHAPTER 5
125
Figure 5.10 Structure of aminoethyl sialyllactose (63).
Chemical synthesis of aminoethyl sialyllactose (63) (structure shown in figure 5.10)
would be significantly more difficult than the synthesis of aminoethyl lactose (51).
However using a trans-sialidase enzyme from Trypanosoma cruzi (TcTs) to selectively
transfer a sialic acid moiety from a donor glycoprotein such as fetuin to a galactose
terminal glycan is relatively straightforward. This reaction forms an α-(2,3)-glycosidic
bond between Gal and sialic acid moieties to produce a recognisable glycan for Siglec
recognition (reaction shown in figure 5.11).191
Figure 5.11 The reaction of immobilised lactose with fetuin in the presence of trans-sialidase (TcTs)
enzyme to produce immobilised sialyllactose.
Optimisation of Trans-sialidase Reaction
GFPuv_I229C was derivatised using lactosyl iodoacetamide (55) as described in section
8.4.2 to produce lactosyl GFPuv (GFPuv_I229C_Lac), on which the TcTs reaction was
optimised. Figure 5.12.A shows the deconvoluted mass spectrum of GFPuv_I229C
before glycosylation and figure 5.12.B shows the corresponding deconvoluted mass
spectrum on GFP_I229C_Lac. The observed mass increase was 429 Da which fit well
CHAPTER 5
126
with the expected increase of 425 Da corresponding to derivitisation with lactosyl
iodoacetamide (55).
Figure 5.12 A) Deconvoluted mass sprectum of GFPuv_I229C. B) Deconvoluted mass spectrum of
GFPuv_I229C_Lac.
GFPuv_I229C_Lac was treated with TcTs (provided by Dr E. Reyes, The University of
Manchester) in the presence of fetuin to produce sialyllactosyl GFPuv
(GFPuv_I229C_Lac_Neu5Ac) as described in section 8.4.2. However, TcTs also
catalyses the hydrolysis of glycosidic bond formed, therefore the reaction time must be
kept relatively low to obtain optimal yields. Samples were removed from the reaction at
half hourly time intervals to determine the optimum length of reaction to maximise the
amount of GFPuv_I229C_Lac_Neu5Ac produced. Samples were denatured using 6 M
guanidine hydrochloride to quench the reaction and then analysed by ESI-MS (results
summarised in figure 5.13). Figure 5.13.A shows the reaction after 30 minutes and
CHAPTER 5
127
illustrates the mass difference between the product and the starting material. Figure
5.13.B shows the reaction after 1 hour and illustrates the structures of
GFPuv_I229C_Lac and GFPuv_I229C_Lac_Neu5Ac. This was also the time point
when the maximum amount of product was observed. Figure 5.13.C shows the reaction
after 2.5 hours when the equilibrium is becoming less favorable to the desired product
(GFPuv_I229C_Lac_Neu5Ac).
Figure 5.13 Mass spectra of samples taken from the reaction of trans-sialidase (TcTs) with
GFPuv_I229C_Lac in the presence of fetuin. A) Reaction after 30 minutes. B) Reaction after 1 hour. C)
Reaction after 2.5 hours.
CHAPTER 5
128
It was decided to limit the length of the TcTs reactions to 1 hour, but for production of
active glycosylated GFPuv samples, the options for quenching the reaction were
limited. Guanidine hydrochloride worked well for quenching samples for analysis but
also rendered the GFPuv inactive. High temperature or extreme pH conditions known to
denature TcTs were also found not to be suitable for use with GFPuv. Size exclusion
chromatography (details in section 8.3.1) was found to separate the two proteins
effectively and GFPuv could be eluted from the column in less than 30 minutes which
was fast enough to effectively quench the reaction before the equilibrium became
unfavorable.
Glycoprotein samples prepared in this manner were analysed by ESI-MS to determine
the efficiency of the derivatisation. Analysis of the spectra suggested at least 60% of the
lactose moieties were successfully modified. This is comparable to previously reported
yields for these types of reactions.198
However it is difficult to determine the exact yield
of this process due to the lack of a truly quantitative method and the lability of the sialic
acid moiety to hydrolysis and ionisation fragmentation.
5.3 Summary
In this chapter we have described the use of enzymes to modify chemically synthesised
aminoethyl glycans on both carbohydrate arrays and when conjugated to GFPuv
mutants. Trimannoside (41) was shown to be successfully glycosylated with a GlcNAc
moiety using the enzyme GnT-I when immobilised on a SAM coated gold plate.
However attempts to used trimannoside (41) to screen glycotransferase activity from
crude yeast microsomal extracts were unsuccessful. Lactosyl iodoacetamide (55) was
shown shown to be successfully in the derivitisation GFPuv_I229C to produce
GFPuv_I229C_Lac which was then enzymatically derivitised using TcTs to produce
GFPuv_I229C_Lac_Neu5Ac. This method was optimised and could potentially be used
to create GFPuv samples suitable for Siglec labeling.
129
Chapter 6: Lectin Binding Assays
CHAPTER 6
130
6.1 General Considerations
Having made a library of neoglycoproteins we wanted to determine the viability of
glycosylated GFPuv as a lectin probe. Several issues needed to be addressed before
more complicated cell based assays were undertaken. Firstly it needed to be verified that
sugars attached to GFPuv were free to interact with carbohydrate binding proteins and
that the unglycosylated mutants did not interact unspecifically with target proteins. Also
it needed to be determined whether a difference could be seen in the interactions of
polyglycosylated GFPuv mutants when compared to the interactions on singly
glycosylated mutants. Additionally we hoped to investigate methods of measuring the
strength of the interaction of glycosylated GFPuv with target lectins. If this could be
done for a lectin with a known binding affinity then we could assess the effect that the
attachment of glycosides to GFPuv has on their binding.
6.2 Fluorescence Based Plate Assay
Initially we wanted a high throughput assay to determine whether or not there were
interactions between glycosylated GFPuv mutants and target lectins. Immobilising the
lectin of choice on a surface of a 96-well plate, incubating different wells with different
glycoproteins, followed by washing and then measuring the fluorescence of each well,
provided a simple method of rapidly obtaining data. Carrying out an assay in this format
allowed the use of a fluorescence plate reader to scan all wells at once and do so quickly
and reproducibly.
CHAPTER 6
131
Figure 6.1 Schematic diagram of the lectin plate assay. The protein avidin (red) is covalently bound to
the surface of a 96-well plate, which enables the capture of biotinylated lectins (blue). Glycosylated
GFPuv mutants can then interact with the immobilised lectins.
The avidin-biotin interaction (affinity constant > 1015
M-1
)199
was the method chosen to
immobilise lectins on to a 96-well plate. This is an efficient method of immobilisation
known not to inhibit the biological activity of proteins. Also several commercially
available lectins are available in biotinylated form, aiding in the reproducibility of initial
assays. This approach (depicted in figure 6.1) is also very versatile, as once optimised
any protein of interest could be biotinylated and applied to the same surface. Whole
cells can also be biotinylated in a similar way,170,200
providing quick access to simple in
vivo assays and eliminating the need for lengthy purifications of the protein of interest.
6.2.2 Lectins Chosen for Initial Screens
Only biotinylated lectins that were commercially available were considered for the
initial screening, to maximise assay reproducibility. All lectins chosen were to be
CHAPTER 6
132
screened against unglycosylated mutants, mannosylated mutants and galactosylated
mutants to provide a protein-protein interaction control and a sugar specificity control.
Concanavalin A (ConA) was chosen due to its relatively strong affinity to
monosaccharides when compared to other lectins available (approximately 1 x 104 M
-1
for Me-α-Man).201,202
Also ConA has previously been the focus of several polyvalent
interaction studies so provides a good bench mark for any data acquired on its binding
preferences.70,203
ConA is a tetrameric lectin (at neutral pH) with one carbohydrate
binding site per subunit. Due to its relative size glycosylated GFPuv (32 kDa) could
only interact with one subunit per ConA molecule (106 kDa). However ConA does
show significantly increased binding affinities for branched polymannosides; for
example it is reported to have an affinity of 2.6 x 105
M-1
(26 times that of monomeric
mannose) for some branched trimannosides.70
Additionally binding affinities of up to
5.1 x 105
M-1
(51 times that of monomeric mannose) have been reported for ConA and
some branched Man derivatised polymers.202
Galanthus nivalis lectin (GNL) is also a tetrameric lectin but it is significantly smaller
than ConA (50 kDa) and contains both a Man and dimannose binding pockets per
subunit.204
This allows for the possibility of several different polyvalent interactions
between glycosylated GFPuv mutants and individual GNL molecules. GNL has been
reported to have binding affinities of up to 3.3 x 103
M-1
for some dimannosides which
is three times less than the affinity of ConA for monomeric Man and 10 times less than
ConA for the same dimmanoside.70
Jacalin is another tetrameric lectin, approximately 64 kDa in size and known to
preferentially bind α-galactose terminal glycans. Like ConA, jacalin only contains one
binding site per subunit, however it is small enough that a polyvalent interaction with
glycosylated GFPuv_B10 may be possible. Jacalin is reported to have an adsorption
coefficient of up to 2.2 x 107
M-1
for Gal coated carbohydrate arrays which is
CHAPTER 6
133
significantly higher than the adsorption coefficient of ConA for its corresponding Man
coated carbohydrate array (5.6 x 106
M-1
).205,206
Jacalin is also known to interact with a
range of other monosaccharides including; N-acetylgalactosamine, Man, Neu5Ac and
Glc.206
6.2.3 Plate Assay Results
Samples of all mutants produced were mannosylated (α-Man) and galactosylated (β-
Gal) in preparation for this initial lectin screen as described in section 8.3.1. Firstly, all
of these samples (10 μM) were assayed against the streptavidin coated 96-well plates as
described in section 8.5.1 with no lectins present, to assess the potential for unspecific
binding between GFPuv mutants and streptavidin (results summarised in figure 6.2).
Error bars in graphs are the standard deviation from the mean over three replicates for
each reading shown.
Figure 6.2 The results of screening the interactions of unglycosylated, mannosylated and galactosylated
GFPuv mutants against streptavidin coated 96-well plates. Samples are grouped according to their
number of glycosylation sites.
CHAPTER 6
134
Under the conditions used there was relatively little unspecific binding between the
GFPuv samples and the streptavidin coated plates. The only sample that showed
significant interaction between protein and streptavidin coated plates was the
unglycosylated GFPuv_CL sample, which suggests that the linkers themselves have
some interaction with the plates. Fortunately neither glycosylated samples modified in
this way (GFPuv_CL_Gal10 and GFPuv_CL_Man10) show the same increased binding.
Therefore these samples could still be used for comparison with the other glycosylated
GFPuv mutants. The GFPuv_I229C mutant show more background binding than
GFPuv_D1 mutant and as they both are glycosylated at the same position (C229),
GFPuv_I was not included in the remainder of the assays.
For the remaining assays, in which one of the chosen lectins was bound to each plate,
the protein concentrations were adjusted so that the glycan concentration was kept
constant (10 μL). Therefore singly glycosylated mutants (E6C, D1, F1 and F11) were
assayed at 10 μM, doubly glycosylated mutants (EI, G1, G3 and D5) were assayed at 5
μM, B10 was assayed at 3.3 μM and C5 and S6 were assayed at 2.5 μM.
GFPuv_CL_Man/Gal10 were estimated to have an average of 10 glycans per protein
molecule by MALDI-TOF MS, so were assayed at 1 μM.
The results from the screening of GFPuv samples against ConA (summarised in figure
6.3) show that all mannosylated mutants interacted significantly more with ConA than
both the unglycosylated and galactosylated mutants. The fact that the F1 mutant
interacts so much more than the other singly glycosylated mutants suggest this mutant
has significantly stronger protein-protein interactions than E6C, D1 and F11. However
the fact that the unglycosylated and galactosylated F1 variants do not show this
increased binding show that the carbohydrate-protein interaction is still the deciding
factor. Most of the polyglycosylated mutants interact more than the singly glycosylated
mutants despite the lower concentrations used.
CHAPTER 6
135
Figure 6.3 The results of screening the interactions of unglycosylated, mannosylated and galactosylated
GFPuv mutants against ConA coated 96-well plates. Samples are grouped according to their number of
glycosylation sites.
Mannosylated G1 and D5 show almost twice the fluorescence of mannosylated G3 and
four times the fluorescence of mannosylated E6C_I229C despite the fact they all have
two glycosylations. This suggests that the spacing of the glycosides is influential in
determining the strength of the interactions between these glycoproteins and ConA.
Additionally both mutants with four mannose residues attached (C5 and S6) and CL
with approximately 10 Man residues attached show no increased fluorescence when
compared to the best doubly mannosylated mutants. This suggests that ConA is not
capable of interacting with more than two mannosides in the way they are presented on
these glycoproteins.
CHAPTER 6
136
Figure 6.4 The results of screening the interactions of unglycosylated, mannosylated and galactosylated
GFPuv mutants against GNL coated 96-well plates. Samples are grouped according to their number of
glycosylation sites.
The results from the screening of GFPuv samples against GNL (summarised in figure
6.4) again show the expected preference of mannosylated GFP mutants. However the
fluorescence levels measured are significantly lower than that with ConA with the same
mutants. This suggests GNL has a lower overall affinity for mannose than ConA. Once
again the F1 mutant shows significantly higher fluorescence than the other singly
glycosylated mutants in both its mannosylated and galactosylated forms, but not in its
unglycosylated form. This suggests there are some important protein-protein
interactions occurring, but that the carbohydrate-protein interactions are again the
discriminating factor.
Interestingly the mutants with the strongest interactions with GNL (G1 and D5) are both
doubly glycosylated. This could be related to the spacing or orientation of the glycans
on these mutants but there is no discernible pattern when comparing the glycosylation
CHAPTER 6
137
site distances. For example the distances between glycosylation sites of the G1 and G3
mutants are almost identical (25 Å) and their affinities for GNL appear to be drastically
different. Potentially this difference could be due to the Man residues attached to G1
being orientated in such a way that they can both fit in to one of GNL’s dimannosides
binding sites or binding two of its binding sites seperatly.70
Surprisingly there seems to
be very little binding of the mannosylated GFPuv samples with more than two Man
residues attached (C5, S6 and CL). It is possible this trend could alter if higher GFPuv
concentrations were used. Unfortunately there were not sufficient amounts of these
samples available to repeat this assay at higher concentrations of GFPuv.
Figure 6.5 The results of screening the interactions of unglycosylated, mannosylated and galactosylated
GFPuv mutants against jacalin coated 96-well plates. Samples are grouped according to their number of
glycosylation sites.
The results from the screening of GFPuv samples against jacalin (summarised in figure
6.5) show a preference towards Man in some cases. Jacalin is known to have a broad
specificity to many different monosaccharides and reportedly binds more strongly to α-
Gal than α-Man. In this assay we are using β-Gal and α-Man and so this result is not
contrary to previous results. For most mutants, the sugar attached had little effect on the
CHAPTER 6
138
fluorescence measured, with glycosylated D5 and CL mutants being the exception. F1,
again stands out as the highest binding singly glycosylated mutant and shows the
expected preference for mannose. G1 and D5 again show higher fluorescence than the
other doubly glycosylated mutants. This suggests that glycosides attached to these
mutants could be more accessible for binding than for the E6C_I229C and G3 mutants.
The mutants with four or more glycosylations (C5, S6 and CL) do show more
fluorescence than the majority of the singly and doubly glycosylated mutants even at
their lower concentrations. The fact the binding is not significantly higher for the CL
mutant suggests that the presentation of the glycosides is the critical factor or that the
concentrations used on the assay are too low to provide an effective comparison.
Additional assays with a larger range of concentrations would be useful in determining
the effect of polyvalent presentation; however there were not sufficient amounts of these
samples available to repeat this assay at higher concentrations of GFPuv.
6.2.4 Fluorescence Based Assay Summary
More could be done to improve the quality of the data gained from the fluorescence
based lectin plate assay. Testing a larger range of concentrations would give more
detailed information on the effect of degree of glycosylation of glycoproteins.
Specifically it would be useful to determine whether the surprisingly low fluorescence
measurements for the mutants B10, C5, S6 and CL were only due to the concentrations
used in the assay. Unfortunately there was insufficient time to prepare new batches of
the glycoproteins required for this.
The results from the assays performed do demonstrate that glycosylated GFPuv mutants
can be used to probe lectin specificity. The attached glycans are evidentially available
for lectin binding because in each assay the glycosylated mutants achieved significantly
higher fluorescence measurements than the unglycosylated mutants. Also discrimination
was clearly shown between Man and Gal by both ConA and GLN. Therefore
CHAPTER 6
139
glycosylated GFPuv mutants do provide a simple and effective method of screening
lectin specificity. Additionally some potential mutants of interest were identified for
further investigation such as the F1, G1 and D5 mutants. These mutants gave
significantly higher readings when compared to other glycoproteins with the same level
of glycosylation with each of the lectins tested.
The sensitivity of this assay in its current format is not good enough to determine the
preference of Jacalin for α-Man over β-Gal with certainty. Both sugars are known to
bind Jacalin, but previous studies suggest Me-α-Man has a significantly higher affinity
than Me-β-Gal.207,208
The main issue is that once bound to the plate, it seems difficult to
remove GFPuv mutants. This is demonstrated by the unexpectedly high binding
observed by singly glycosylated mutants when compared to mutants with more glycans
attached. Whether this is a result of protein-protein interactions or GFPuv aggregation
on the surfaces is unclear. It is possible this could be overcome by including sugars in
the washing buffer used. However we did not have sufficient material available to
undertake these experiments.
The assay in its current form can determine that there are interactions between GFPuv
mutants and the immobilised lectins, but it is not possible to quantify the strength or
type of interactions occurring with the current data. Reproducibility between triplicate
measurements was quite good but could be improved further if GFPuv concentrations
and wash buffer constitution were optimised. Another obstacle to quantifying data from
these assays is the slight variations in fluorescence measurements between GFPuv
batches prepared and GFPuv’s susceptibility to photo bleaching. To minimise this
problem it is very important to maintain consistency with the preparation and storage of
all GFPuv samples.
The results from the fluorescence based assay described do provide a qualitative
comparison of the relative binding strength of the neoglycoproteins produced and the
CHAPTER 6
140
lectins chosen. However, due to the limited amount of time available and the relatively
large amounts of glycosylated GFPuv required for this assay it was decided to
investigate the use of quantitative techniques for assessing lectin binding.
6.3 Isothermal Titration Calorimetry (ITC)
ITC is routinely used technique for the study of reactions and binding thermodynamics
and kinetics via the measurement of minute changes in heat generated by mixing two
solutions. Using ITC it is possible to measure heat changes as small as 0.1 μcal (0.4 μJ),
which allows the determination of binding constants as large as 1 x 109
M-1
.209
The
understanding of several protein-carbohydrate interactions has been improved through
the use of this methodology.70,201,202
ITC can be used not only to generate rate constants
for interactions of interest but also reveals their entropy, enthalpy and stoichiometry.209
Unlike other methods for protein-carbohydrate interaction analysis it does not require
lectin or ligand to be immobilised. Therefore the potential for unspecific binding with
surfaces is reduced and wash or blocking procedures do not need optimising.
The experimental procedure for ITC is relatively straightforward. A ligand solution of
known concentration is added to a receptor solution of known concentration. The heat
change upon mixing is measured and compared to a reference sample. However the
concentrations of each solution need to be extremely accurate and also the composition
of each solution (e.g. buffer salt concentrations) need to be identical to minimise errors.
For truly accurate measurements the heat change of mixing the two buffers alone and
the heat change of mixing either component with the buffer must also be measured and
subtracted from the heat change from mixing the ligand and receptor solutions.
To obtain meaningful thermodynamic data, a series of injections are made over time
until there is no heat change upon mixing ligand and receptor (indicating the receptor
occupancy it saturated). Figure 6.6.A shows an ideal set of responses to a series of
injections. The first injection of ligand solution into the receptor solution should induce
CHAPTER 6
141
the largest response. Each subsequent injection should induce less of a response as more
of the receptor sites are already occupied, therefore there will be a smaller energy
change upon the addition of more ligand. The experiment is over when the addition of
more ligand induces no energy change.
Figure 6.6 Ideal ITC plots. A) The raw data obtained from an ideal series of injections. B) An ideal plot
of the molar energy changes for a bivalent (N = 2) interaction.
The data obtained from a successful series of injections can be used to plot a graph such
as the one in figure 6.6.B. To do this the data is converted into the relative energy
change per Mole which is why the exact concentrations of each solution are required.
The molar enthalpy change (∆H) can be read directly from such a graph as can the
stoichiometry (N) of the interaction. The binding constant (KB) can be found from the
gradient of the graph at its steepest point. This can then be used to find the dissociation
constant (KD) by taking the inverse of this value (1/ KB). The molar entropy change
(∆S) and molar Gibb’s free energy change (∆G) can then be found using the equations
for ∆G shown in figure 6.7.
Figure 6.7 Equations for calculating Gibb’s free energy. R = The molar gas constant 8.314 J mol-1 K-1,
T = The temperature in Kelvin.
CHAPTER 6
142
6.3.1 Titration of Me-α-Man Against ConA
Initial optimisation of conditions was done using ConA and Me-α-Man to establish the
effect of dilution of both ligand and lectin. Previously published data on this interaction
was acquired using lectin and ligand concentrations in the ranges of 0.2-1 mM and 5-50
mM respectively.201,202
The higher range of these concentrations would be problematic
for proteins such as GFPuv due to its large size when compared to the monosaccharides
used in previously reported experiments. GFPuv mutants with multiple glycosylations
were expected to produce useful data at significantly lower concentrations than these.
However it was still crucial to discover the lower limits of the glycoside concentrations
that were needed on the instruments we had available to conserve our glycosylated
GFPuv samples.
The results achieved using 32 μM ConA and 5 mM Me-α-Man are shown in figure 6.8
(full details in section 8.5.2). Figure 6.8.A shows the raw titration data which shows that
each addition of Me-α-Man caused a smaller energy change when added to the ConA
solution. The decrease in the heat change upon mixing is too rapid over the first few
injections to produce a sigmoidal curve (like the graph shown in figure 6.6) in the plot
shown in figure 6.8.B. This is due to the low concentrations of ConA being used.
However the software can still calculate the desired thermodynamic constants by
extrapolating the rest of the curve. The stoichiometry of binding (N) is given at 3.6 (±
0.209), which is relatively close to the expected value of 4 (for tetrameric ConA), but
has a significant margin of error. The affinity constant (K) was calculated at 1.07 x 104
M-1
(±7.91 x 103) is in very good agreement with Lit. (1.03 x 10
4 M
-1).
202 The relatively
large error margins were likely due to the lack appropriate background calorimetric
curves which we did not undertake for these trail titrations, so could not be subtracted
from the raw data.
CHAPTER 6
143
Figure 6.8 Calorimetric data obtained from titration of native ConA (32 μM) with Me-α-Man (5 mM). A)
Raw data from 30 injections of 1 μL each of Me-α-Man into ConA. B) Integrated curve showing the line
of best fit.
The concentration of 5 mM Me-α-Man used in this experiment is still higher than
practical for GFPuv mutants. However the mutants containing multiple glycosylation
sites effectively increase the concentration of sugar in solution and increasing the
volume of each injection would also decrease the concentration required to induce the
same response. Some polyglycosylated mutants were also expected to show increased
CHAPTER 6
144
binding affinities when compared to Me-α-Man, which would also lower the
concentration required to achieve valid results.
6.3.2 Titration of GFPuv Against ConA
Having established a working concentration range from the ConA-Man interaction we
titrated GFPuv_C5_Man4 samples against ConA at a comparable concentration. We
assumed that there would be an increased binding when compared to monomeric Me-α-
Man and that having four mannosides attached to each protein would reduce the
concentration of ligand required to induce the same response by a factor of at least four.
The results achieved with 22 μM ConA and 240 μM GFPuv_C5_Man4 are shown in
figure 6.9.
Although there is significant baseline noise in this titration the software calculates a
stoichiometry (N) at 0.93 (±0.01), which is relatively close to the expected value of 1.
The affinity constant calculated from this data is 2.6 x 106 M
-1 (±3.85 x 10
5), is over 200
times that of Me-α-Man. However this figure is to be taken as a rough approximation,
due to the large margin of error calculated. This large error is likely due to the low
concentrations used and subsequent high back ground noise. The gradual increase in the
background throughout this titration also indicates that a longer interval should have
been left between injections. Unfortunately there was insufficient GFPuv_C5_Man4
available for further titrations at the time.
CHAPTER 6
145
Figure 6.9 Calorimetric data obtained from titration of native ConA (22 μM) with GFPuv_C5_Man4 (240
μM). A) Raw data from 30 injections of 1 μL each of GFPuv_C5_Man4 into ConA. B) Integrated curve
showing the line of best fit.
To determine what was measured in this titration was not protein-protein interactions or
just the heat change of mixing we performed some control titrations. Figure 6.10.A
shows the titration of the buffer used against itself to determine that there are no
unexpected heat changes for this process. Figure 6.10.B shows the titration of
unglycosylated GFPuv_C5 against ConA at the same concentrations used to obtain
binding data for GFPuv_C5_Man4. This was the control titration needed to determine if
there were significant protein-protein interactions between GFPuv and ConA. Both
titrations showed almost no heat change upon mixing which strongly suggest the heat
changes seen in the titration of glycosylated GFP against ConA were the result of
carbohydrate-protein interactions.
CHAPTER 6
146
Figure 6.10 Raw calorimetric data from 30 injections of 1 μL each of: A) ITC buffer into ITC buffer. B)
GFPuv_C5 (240 μM) into ConA (22 μM).
6.3.3 ITC Summary
The approximate concentration range for a successful ITC titration of GFPuv_C5_Man4
with ConA was established (240 μM and 22 μM respectively) and an approximate
affinity constant was determined (2.6 x 106 M
-1). The affinity constant obtained did
have a relatively large margin of error (approximately 15%) so should be treated with
caution. However this data suggests that GFPuv_C5_Man4 is has a significantly larger
binding constant than Me-α-Man of (1.03 x 104 M
-1)202
, which suggests some kind of
polyvalent interaction is occurring. The negligible heat change which occurs when
titrating unglycosylated GFPuv_C5 strongly suggests that the affinity of mannosylated
GFPuc_C5 is due to the carbohydrates attached. Therefore ITC is a valid method of
quantifying the carbohydrate-protein interactions of our neoglycoproteins.
Unfortunately we could not follow up on these initial results due to a lack of time for
preparing additional neoglycoprotein samples.
If a larger microcalorimeter was used (as is more common), a larger volume of GFPuv
could be titrated and the enthalpy changes measured would be significantly higher. This
CHAPTER 6
147
would improve the accuracy of the data obtained. Our instrument choice for initial
experiments was limited due to the volume of material available.
6.4 Summary
In this chapter a fluorescence base assay for measuring the relative binding of
glycosylated GFPuv mutants with immobilised lectins is described. This assay is
capable of rapidly generating large amounts of qualitative data and identifying
glycoproteins for more detailed analysis. It has been shown to be applicable to a three
lectins (ConA, GNL and jacalin) and has the potential to work with several others. In
the ConA and GNL assays a clear discrimination between to monosaccharides (Man
and Gal) was shown. Jacalin was demonstrated to have little discrimination between
Man and Gal, but did show a clear discrimination between glycosylated and
unglycosylated GFPuv mutants.
The fluorescence based assay described was used to identify some potential mutants of
interest for further investigation such as the F1, G1 and D5 mutants. These mutants
gave significantly higher readings when compared to other glycoproteins with the same
level of glycosylation with each of the lectins tested. This suggests that either the
orientation of the glycans on these mutants was favourable or some additional
favourable protein interactions were responsible.
The use of ITC was then investigated as a method for determining quantitative
thermodynamic data for the interactions of glycosylated GFPuv mutants with lectins.
Initial titrations have shown that a GFPuv mutant with four mannose residues attached
(GFPuv_C5_Man4) has a significantly higher affinity for ConA than Me-α-Man, which
suggests a polyvalent interaction is occurring. Unfortunately we could not follow up on
these initial results due to a lack of time for preparing additional neoglycoprotein
samples. However from these initial titrations we have shown that ITC is a valid method
of quantifying the carbohydrate-protein interactions of our neoglycoproteins.
148
Chapter 7: Conclusions and Future
Experiments
CHAPTER 7
149
7.1 Conclusions
The addition of a hexahistidine tag to GFPuv can provide a simple method for the rapid
purification of any mutant. The addition of the 52 amino acid long sequence containing
the hexahistidine tag to the N-terminus of GFPuv does not seem to hinder its use as a
lectin probe. Very high purities of GFPuv mutants are easily attainable using a IMAC
column and a stepwise imidazole gradient. If further enhancements in purity are
required, size exclusion chromatography can be used as a second purification step at the
expense of yield.
Several GFPuv cysteine mutants were created using site directed mutagenesis, but this
was time consuming and in some cases resulted in inactive mutants. DNA shuffling of
synthetic GFPuv genes (one containing 32 additional cysteine codons) provided a
relatively quick route to several new mutants including some with multiple cysteines.
Three mutants with two reactive cysteines (D5, G1 and G3), one mutant containing
three reactive cysteines (B10) and two mutants containing three reactive cysteines (C5
and S6) were generated in this manner. One mutant containing six reactive cysteines
was also generated by the site directed mutagenesis of C5 to include the E6C and I229C
mutations. The use of a fluorescence based screen of the mutants produced avoids the
sequencing of inactive mutants produced by DNA shuffling. Not only could this
approach be used to rapidly generate a much larger library of GFPuv cysteine mutants,
but could be used on other proteins of interest.
The synthesis of the branched trimannoside (41) was achieved without the use of
protecting groups on the glycosyl acceptor (aminoethyl mannoside (27)). While the
yields achieved were low (≤ 10%), we are confident they could be improved upon. The
complexity of the product and the relatively short length of this synthesis makes this
approach viable. There is also the possibility of using a similar approach for alternative
monosaccharide starting materials. Trimannoside (41) was shown to be accepted as a
substrate for GnT-I, proving the β-mannose linkage is not essential for enzyme
recognition in this case.
CHAPTER 7
150
A range of monosaccharide aminoethyl glycosides (27, 32-34) were synthesised and
fully characterised. These glycosides and aminoethyl lactose (51) (donated by Dr R.
Sardzik, The University of Manchester) can be easily converted to glycosyl
iodoacetamides capable of selective glycosylation of cysteines in a one step reaction
with iodoacetic anhydride. The naturally occurring C48 and C70 residues in GFPuv are
unreactive to iodoacetamides, but all other positions of GFPuv screened (6, 15, 30, 38,
43, 52, 105, 188, 202, 212, 221 and 229) were found to be reactive. Several mutants
with up to four reactive cysteine residues were glycosylated selectively using the
glycosides produced. This demonstrates the potential of glycosylated GFPuv constructs
to present glycosides in a polyvalent manner.
Glycosylated GFPuv mutants are suitable for simple, fluorescence based, lectin assays
in which biotinylated lectins are immobilised on avidin coated 96-well plates.
Glycosides attached via the iodoacetamide linker used are capable of interacting with
lectins bound to a surface and GFPuv’s inherent fluorescence provides a simple method
for qualitative binding analysis. Although the slight variation in fluorescence between
mutants may limit the accuracy of the relative binding strengths observed using the
fluorescence based assay described. The approach described was used to successfully
probe the specificity three lectins (ConA, GNL and jacalin) and has the potential to
work with several others. In the ConA and GNL assays a clear discrimination between
to monosaccharides (Man and Gal) was shown. Jacalin was demonstrated to have little
discrimination between Man and Gal, but did show a clear discrimination between
glycosylated and unglycosylated GFPuv mutants. More information could be gained on
the relative strengths of the carbohydrate-protein interactions of the different
neoglycoproteins produced if a larger range of concentrations was assayed.
ITC shows potential for analysing the thermodynamics of binding interactions between
glycosylated GFPuv mutants and lectins. Initial titrations have shown that a GFPuv
mutant with four mannose residues attached (GFPuv_C5_Man4) has a significantly
higher affinity for ConA than Me-α-Man, which suggests a polyvalent interaction is
CHAPTER 7
151
occurring. If a larger range of neoglycoproteins were rigorously tested in this manner
then a quantitative comparison of binding affinities could be achieved using ITC.
7.2 Future Work
The potential of glycosylated GFPuv mutants in the analysis of carbohydrate binding
protein has been explored, but further research would be needed to fully asses this
potential. A larger library of cysteine mutants would be could easily be generated using
the genes designed (Shuffle 1-3, section 8.1.2) and the DNA shuffling technique
demonstrated. Once these mutants were generated they should be screened with the
existing mutants for their tendency to aggregate when glycosylated. It is likely that
mutants containing several reactive cysteines will form intermolecular disulphide
bonds, but once derivatised they may be more or less inclined to dimerise than GFPuv.
This could have a significant impact on their measured binding constants, therefore
should be investigated for each individual mutant produced.
Once a carbohydrate binding protein of interest is selected for a detailed study using
glycosylated GFPuv a wider range of appropriate glycans should be synthesised. These
should include polyvalent glycans related to those known to bind the protein of interest
in vivo to provide a comparison to the monosaccharides currently used. This would
provide more information on the specificity of the carbohydrate binding protein of
interest. The use of glycans with a higher binding affinity for their target proteins would
also increase the sensitivity of any assays undertaken and potentially facilitate the use of
glycosylated GFPuv as a cell labelling tool.
The DNA shuffling approach described could be applied to different proteins to
facilitate their use as glycoprotein scaffolds. Although the fluorescence of GFPuv
makes it an ideal candidate for a range of applications it also has limitations. A larger
protein for example, will have the potential for the inclusion of far more cysteines and
will be able to interact with more widely spaced binding sites on carbohydrate binding
CHAPTER 7
152
proteins. Homodimers of GFPuv or GFP fusion proteins would provide a suitable
candidate for this approach.
153
Chapter 8: Experimental Details
CHAPTER 8
154
8.1 Experimental Details for Chapter 2
Unless stated otherwise, all chemicals and reagents were of analytical grade and used as
received from Sigma-Aldrich. All competent cell lines used were purchased from
Invitrogen and all enzymes used were purchased from New England Biolabs or
Novagen unless otherwise stated. DNA sequences were performed by MWG Biotech.
PCR reactions were carried out using a Mastercycler Gradient PCR-Cycler (Eppendorf).
Affinity columns used in protein purification were purchased from GE Healthcare and
purifications were carried out using an AKTA Explorer 100 FPLC (GE Healthcare).
Agarose gel electropherisis analyses and purifications were carried out using a Mini Sub
Cell GT gelchamber (Biorad) and the gels were analysed with the aid of a Safe Imager
darklite transluminator (Invitrogen).
8.1.1 General Methods
Isolation of Plasmid DNA from E. coli
A single colony containing the plasmids desired for sequencing, amplification or
transformation was used to seed 7 mL of sterile LB medium containing the appropriate
antibiotic. The culture was incubated for 16 hours at 37°C with shaking (250 rpm). The
cells were collected by centrifugation at 3220 g, for 15 minutes. The plasmids were then
purified using a Quiagen spin column DNA purification kit according to the
manufacturer’s instructions.
Transformation of E. coli strains by heat shock
A 50 μL aliquot of competent cells was transferred into a sterile 1.5 mL eppendorf tube
and placed on ice for 15 minutes. 1 μL of DNA solution or 10 μL of ligation mixture
was added and then the tube swirled gently and then left on ice for another 15 minutes.
The tube was then transferred to a 42°C circulating water bath for exactly 45 seconds
and then put back on ice for 2 minutes. 450 μL of SOC medium (pre heated to 42°C)
was added and the cells incubated at 37°C with shaking (250 rpm) for 1 hour to allow
CHAPTER 8
155
expression of the antibiotic resistance gene. The mixture was then spread on agar plates
containing the appropriate antibiotic. The plate was dried in a laminar flow hood for 5
minutes, before being incubated, inverted, at 37°C for16 hours.
Agarose Gel Electrophoresis
Agarose powder was added to 1x TAE Buffer to a final concentration of 1-2% (w/v).
The slurry was heated in a microwave oven until the agarose was completely dissolved.
Any lost volume was replaced with deionised water, the solution cooled to 60°C and
then SYBR Safe was added (10 μL to 100 mL agarose solution). The warm agarose
solution was poured into the tray and a comb was inserted until the gel was set. The gel
was then transferred into the electrophoresis tank and covered with TAE Buffer before
the comb was removed. DNA samples mixed with DNA loading buffer and then loaded
into the wells with a 1kb-ladder reference solution. Electrophoresis was carried out at
110 V for 20 minutes and the gel examined using the Safe ImagerTM
.
DNA Extraction From Agarose Gels
After agarose gel electrophoresis, the required DNA fragments were extracted using a
scalpel and purified using a QIAprep Spin Miniprep kit (Qiagen) as described in
producers manual.
DNA Digestion with Restriction Endonucleases
DNA digestions were carried out according to the manufacturer’s instructions.
Typically 3 μg of DNA was digested with 5 U of restriction enzyme. Specific buffers
provided with the enzymes were used and TE buffer was added to achieve the desired
salt concentrations. Double digestions were preformed when the enzymes optimal
buffer and temperatures were compatible. Restriction digests were generally performed
on a 20-100 μL scale. Although a much larger digest was required prior to the DNA
shuffle PCR (1 mL). Incubation times and temperatures were dependant on the enzymes
CHAPTER 8
156
used. The digested DNA was analysed by agarose gel electrophoresis and isolated by
gel extraction.
Ligation of DNA
A threefold excess of insert DNA was mixed with the digested and purified plasmid to a
final volume of 10 μL (e.g. 3.9 μL insert, 3.9 μL digested plasmid DNA, 1.2 μL of T4
ligase enzyme and 1 μL 10% ligase buffer). T4 ligase enzyme and T4 ligase buffer were
added and the mixture incubated at 16°C for 16 hours.
Amplification of DNA Using PCR
Plasmid DNA were used as templates, DNA polymerase specific buffers were used and
the primers used varied depending on the section of DNA being amplified. 20 μL
reaction mixtures were made up in the appropriate polymerase buffers to the final
concentrations; 1 M Dimethyl Sulfoxide (DMSO), 0.2 mM per dNTP, 1 μM forward
primer, 1 μM reverse primer. PCRs also contained 1-10 ng template plasmid and 0.5-2.5
U of DNA polymerase according to suppliers recommendations. In vitro amplification
PCRs were performed at three different annealing temperature (52°C, 56°C and 60°C)
in a thermocycler using the following temperature-gradient program:
CHAPTER 8
157
Step No. Step Temperature Duration
1 Heating lid 105°C
2 Initial denaturation 95°C 3 minutes
3 Annealing 50°C 30 seconds
4 Elongation 72°C 3 minutes
5 Denaturation 95°C 30 seconds
6 Annealing 51°C 30 seconds
7 Elongation 72°C 3 minutes
8 Denaturation 95°C 30 seconds
9 Annealing 52°C/56°C/60°C 30 seconds
10 Elongation 72°C 3.5 minutes
11 Denaturation 95°C 30 seconds
12 9 repeats of steps 9-11
13 Annealing 52°C/56°C/60°C 30 seconds
14 Elongation 3.5 minutes + 5
seconds/cycle
15 Denaturation 72°C 30 seconds
16 24 repeats of steps 13-15 95°C
17 Final elongation 72°C 7 minutes
Table 8.1 The PCR program used for in vitro DNA amplification.
CHAPTER 8
158
8.1.2 Production of GFPuv Mutant Library
Addition of Hexahistidine Tag to GFPuv
The WT GFPuv gene was amplified by PCR with the forward primer
5’AAAAAAGAATTCATGAGTAAAGGAGAAGAACTTTTCACTGGAGTTGTCCC
and the reverse primer 5’TTTTTTTTTTTTTGCGGCCGCTTATTTGTAGAGC
TCATCCATGCCATGTG using pGFPuv as a template. The amplified gene was
purified by gel electrophoresis and cloned in to a pET-30a vector using the introduced
NotI and EcoRI restriction sites. The resulting construct was transformed into One
Shot® TOP10 competent cells and sequenced using the T7 and pET-RP primers
(acquired sequence show in appendix 1).
Quickchange Site Directed Mutagenesis
Unlike some methods of site directed mutagenesis the quick change method does not
require any specific restriction sites, purification steps or ligation reactions. It introduces
the mutation in one step and unwanted (original) DNA is digested by DpnI restriction
endonuclease, which only digests methylated DNA. Primers are designed by just
altering the desired codon and producing sufficient over lap either side of the mutation
site so that annealing occurs. The reverse primer is simply the reverse compliment of
the forward primer. Once the PCR reaction mixtures are digested by DpnI they can be
directly transformed into the XL1-Blue super competent cells recommended by the
Quickchange manual. Plasmid DNA was used as the DNA template, the primers were
varied according to the site targeted for mutagenesis, a variety of different polymerase
enzymes were used (pfu, pfu Ultra, KOD, KOD XL) with their corresponding buffers.
The successful site directed mutations were achieved in 20 μL reaction volumes
containing 0.5 U of KOD XL DNA polymerase and 10 μg template plasmid in the
recommended polymerase buffer made up to the final concentrations; 750 mM DMSO,
0.2 mM per dNTP, 0.2 μM forward primer, 0.2 μM reverse primer. To test the
efficiency of the restriction enzyme, control PCRs containing no polymerase enzyme
were also performed. Annealing temperatures were varied (50°C, 55°C, 58°C, 62°C,
64°C, 68°C and 72°C) and different numbers of cycles were tried (18-25) until optimal
CHAPTER 8
159
conditions were discovered. All successful site directed mutations were carried out
using the following PCR program:
Step No. Step Temperature Duration
1 Heating lid 105°C
2 Initial denaturation 94°C 2 minutes
3 Denaturation 94°C 30 seconds
4 Annealing/elongation 72°C 6 minutes
5 25 repeats of steps 3 and 4
6 Final elongation 72°C 10 minutes
Table 8.2 The PCR program used for site directed mutagenesis.
Template DNA was removed by addition of DpnI restriction endonuclease and
incubation for 5 hours at 37°C. 1 μL of the mixture was used for transformation into
XL1 Blue super competent cells.
Quickchange Mutation C48A
The C48A mutation was introduced by PCR using GFPuv_WT as the template, forward
primer 5’GGAAAACTTACCCTTAAATTTATTGCCACTACTGGAAAACTACCT
GTTCC and reverse primer 5’GGAACAGGTAGTTTTCCAGTAGTGGCAATAAAT
TTAAGGGTAAGTTTTCC.
CHAPTER 8
160
Quickchange Mutation I229C
The I229C mutation was introduced to both GFPuv_WT and GFPuv_C by a PCR using
the forward primer 5’CTTGAGTTTGTAACTGCTGCTGGGTGTACACATGGCA
TGGGATGAGC and reverse primer 5’GCTCATCCCATGCCATGTGTACA
CCCAGCAGCAGTTACAAACTCAAG.
Quickchange Mutation E6C
The E6C mutation was introduced to GFP_WT, GFPuv_C48A, GFPuv_I229C and
GFPuv_C48A_I229C using forward primer 5’CCGAATTCATGAGTAAAGGAGAA
TGTCTTTTCACTGGAGTTGTCCC and reverse primer 5’GGGACAACTCCA
GTGAAAAGCATTCTCCTTTACTCATGAATTCGG.
Inverse PCR
For the inverse PCR GFPuv_WT was used as the template with the primers previously
reported (forward = 5’CGCGACTACTGGAAAACTACCTGT and reverse =
5’ATAAATTTAAGGGTAAGTTT).1,164
The 50 μL inverse PCR mixtures contained
10-50 μg template plasmid, 2.5 U KOD XL polymerase in the recommended
polymerase buffer made up to the final concentrations; 750 mM DMSO, 0.2 mM per
dNTP, 0.2 μM forward primer, 0.2 μM reverse primer.The following program was used
for amplification:
CHAPTER 8
161
Step No. Step Temperature Duration
1 Heating lid 105°C
2 Initial denaturation 95°C 3 minutes
3 Annealing 68°C 30 seconds
4 Elongation 72°C 6 minutes
5 Denaturation 95°C 30 seconds
6 35 repeats of steps 3-5
7 Final elongation 72°C 7 minutes
Table 8.3 The PCR program used for inverse PCR site directed mutagenesis.
The PCR product was isolated by agarose gel electrophoresis and purified using a
QIAquick Gel Extraction Kit. The DNA was then phosphorylated using by adding 2 μL
of a polynucleotide kinase enzyme for 1 hour at 37°C before ligation. The resulting
plasmids were transformed into Top10 competent cells and isolated for sequencing.
DNA Shuffling
Three genes were designed for DNA shuffling with condon optimised GFPuv_C48A
(sGFP_C48A). Table 8.4 shows the residues that were substituted for cysteine in each
gene. Shuffle 1 contained cysteines at positions 6, 48 and 229 to provide a comparison
with existing mutants.
CHAPTER 8
162
Shuffle 1 Shuffle 2 Shuffle 3
Shuffle 1
cont.
Shuffle 2
cont.
Shuffle 3
cont.
6 1 2 149 137 140
15 5 7 157 142 146
24 9 12 164 147 151
30 17 19 170 153 156
38 21 26 175 159 162
43 28 32 182 166 168
48 34 39 188 172 173
52 41 45 195 176 178
76 47 49 202 181 183
80 51 73 208 186 190
90 77 79 212 194 197
97 81 86 221 200 204
105 93 96 229 206 210
111 99 101 236 214 215
118 104 107 219 223
124 109 113 225 227
128 117 122 230 232
138 126 129 237 238
144 131 133
Table 8.4 The three genes (Shuffle 1-3) designed for DNA shuffle cysteine screen of GFPuv. Numbers
correspond to the amino acids to be substituted for cysteine in each gene.
CHAPTER 8
163
DNase Digest
Plasmids containing the synthetic genes Shuffle 1 and sGFPuv_C48A (purchased from
GeneArt) were transformed in to XL1 blue super competent cells for DNA production.
An approximately 1 kb fragment containing each gene was obtained by digestion using
EcoRI and NotI and purified by electrophoresis using 2% low melting point agarose. 5
μg of each purified gene was diluted in 70 μL of DNase buffer and incubated at 37°C
for 10 minutes. 0.5 U of DNase was then added and the reaction incubated at 37°C for 1
minute. The digestion was stopped by heating the reaction to 80°C for 15 minutes and a
portion of the mixture removed for analysis by electrophoresis. This method was found
to produce DNA fragments of the desired length (less than 200 bp).
Precipitation of Digested DNA
5 µL of 3 M sodium acetate (pH 5.2), 25 µL of 5 M ammonium acetate and 150 µL of
ethanol were added to a 50 µL sample of quenched DNase digest. The mixture was
stored at -20°C over night, centrifuged at 16100 g for 30 minutes (4°C) and the
supernatant discarded. The precipitated DNA was air dried for 15 minutes before being
resuspended in 50 µL polymerase buffer.
DNA Shuffle PCR
A PCR was conducted using the digested fragments of the synthetic genes Shuffle 1 and
sGFPuv_C48A without primers or template under the same conditions as previously
reported.146,167
The 50 µL PCR mixture contained 8 ng digested Shuffle 1, 4 ng digested
sGFPuv_C48A, 0.5 U Phusion DNA polymerase made up in the appropriate polymerase
buffer to a final concentration of 0.2 mM per dNTP. The following program was used
for the DNA shuffle reaction:
CHAPTER 8
164
Step No. Step Temperature Duration
1 Heating lid 105°C
2 Initial denaturation 98°C 3 minutes
3 Annealing 45°C 30 seconds
4 Elongation 72°C 30 seconds
5 Denaturation 98°C 30 seconds
6 35 repeats of steps 3-5
7 Final elongation 72°C 10 minutes
Table 8.5 The PCR program used for the DNA shuffle.
The products of this reaction were then diluted 40x into a new PCR, using the forward
primer 5’GGAAAACTTACCCTTAAATTTATTGCCACTACTGGAAAACTA
CCTGTTCC and the reverse primer 5’GGAACAGGTAGTTTTCCAGTAGTGGCA
ATAAATTTAAGGGTAAGTTTTCC. This 50 μL PCR mixture contained 1.25 μL of
the DNA suffle reaction mixture and 0.5 U Phusion DNA polymerase made up in the
appropriate polymerase buffer to the final concentrations; 0.2 μM forward primer, 0.2
μM reverse primer and 0.2 mM per dNTP. The following program was used for the
amplification:
CHAPTER 8
165
Step No. Step Temperature Duration
1 Heating lid 105°C
2 Initial denaturation 98°C 3 minutes
3 Annealing 50°C 30 seconds
4 Elongation 72°C 30 seconds
5 Denaturation 98°C 30 seconds
6 25 repeats of steps 3-5
7 Final elongation 72°C 10 minutes
Table 8.6 The PCR program used for the amplification of the DNA shuffle products.
After purification using a PCR cleanup kit (Promega), a restriction digest was
performed on the products using EcoRI and NotI. The 1 kb fragments corresponding to
the shuffled GFPuv genes were purified from 2% low melting point agarose, cloned
back in to pET30a and transformed in to XL1 blue super competent cells. Several
hundred colonies were suspended in 100 mL LB media containing kanamycin (50
ug/mL) and incubated at 37°C for 8 hours. The DNA from this culture was extracted
and transformed in to BL21 (DE3) competent cells for screening. After 16 hours growth
at 37°C the colonies were transferred using a Hybond membrane to expression plates
precoated with dilute IPTG and incubated at 30°C for a further 20 hours. Active mutants
were visibly green after this length of expression (figure 2.8), but were easily detectable
using UV light (659nm) after a few hours.
Preparation of Expression Plates
Agar plates containing 50 μg/mL kanamycin were coated with 100 µL of 25 mM IPTG
and dried for 10 minutes in a laminar flow cabinet immediately before use.
CHAPTER 8
166
8.1.3 Protein Expression and Purification
A single colony was picked from agar plates containing freshly transformed E. coli
BL21(DE3) cells and used to seed 7 mL of sterile LB medium containing the
appropriate antibiotic. This culture was incubated for 16 hours at 37°C with shaking
(250 rpm). 400 μL of this culture was then transferred into 400 mL of sterile LB
medium and incubated at 37°C with shaking (250 rpm) until an OD of 0.6-0.7 was
reached. The culture was then induced using IPTG (final concentration 1 mM) and
incubated for 1-24 hours at varying temperatures (22°C, 30°C and 37°C) with shaking
(250rpm). 100 μL aliquots of the cell suspension were taken at time points to analyze
expression levels and to determine the optimal induction time. Each of these samples
was centrifuged for 10 minutes at 16,000 g and the supernatant discarded. The pellets
formed were frozen to aid cell lysis and stored for SDS-PAGE electrophoresis and
western blot analysis.
Bulk cells cultures were harvested by centrifugation at 11325 g for 15 minutes. After
resuspension in IMAC binding buffer the cells were lysed by sonication (20 minutes 10s
on/ 10s off-cycles). The resulting mixture was centrifuged at 39191 g for 30 minutes
and the supernatant discarded. The pellets formed were resuspended in IMAC binding
buffer then stored at -20°C ready for protein purification.
Gradient Immobilised Metal Affinity Chromatography (IMAC)
A 5 mL His Trap column was washed with 10 mL of filtered H2O, 10 mL of NiSO4
solution (100 mM), again with 10 mL of filtered H2O and then with 10 mL of IMAC
binding buffer (50 mM Tris Base, 500 mM NaCl, 10 mM imidazole, pH 7.4). The
sterile, filtered protein sample was then loaded on to the column using a syringe. The
column was inserted into the AKTA-FPLC purification system. A buffer gradient was
set up so that increasing proportions of IMAC elution buffer (50 mM Tris Base, 500
mM NaCl, 500 mM imidazole, pH 7.4) were passed through the column and the eluent
was collected in 1 mL fractions in 96 deep well plates with a flow rate of 2 mL/min.
The protein containing fractions were analysed by SDS-PAGE and Western blot. After
CHAPTER 8
167
finishing the purification, the column was washed with 10 mL of 0.3 M Na-EDTA
solution, 10 mL of sterile water and 10 mL of 20% ethanol before being stored at 4°C.
Optimised IMAC
The sample and column were prepared in the manner previously described and inserted
into the AKTA-FPLC purification system. After 5 minutes of 4 mL/min IMAC binding
buffer flowing through the column the buffer was altered to include 10% IMAC elution
buffer. After a further 5 minutes the buffer was changed to 100% IMAC elution buffer
until all remaining protein had been eluted (approximately 5 minutes). The column was
washed and stored as previously described. A typical UV trace is shown in appendix 3.
Anionic Affinity Chromatography
A 5 mL Hi Trap Q FF column was washed with 10 mL of filtered H2O and then with 10
mL of anionic affinity binding buffer (20 mM NaH2PO4/Na2HPO4, pH 6). The sterile,
filtered, desalted protein sample was then loaded on to the column using a syringe. The
column was then inserted into the AKTA-FPLC purification system. A buffer gradient
was set up so that increasing proportions of anionic affinity elution buffer (20 mM
NaH2PO4/Na2HPO4, 500 mM NaCl, pH 6) were passed through the column and the
eluent was collected in 1 mL fractions in 96 deep well plates. The protein containing
fractions were analysed by SDS-PAGE and Western blot. After finishing the
purification, the column was washed with 10 mL of sterile water and 10 mL of 20%
ethanol before being stored at 4°C.
Size Exclusion Chromatography
A Hi Load Superdex 200 column was inserted into the AKTA-FPLC purification
system and then equilibrated by washing with 128 mL of size exclusion buffer (20 mM
NaH2PO4/Na2HPO4, pH 6). 1 mL of concentrated protein solution was loaded on to the
column using a filling loop and the eluent was collected in 1 mL fractions in 96 deep
CHAPTER 8
168
well plates as another 128 mL of the size exclusion buffer was passed through the
column. The column was then ready to be stored at 4°C. The protein containing
fractions were analysed by SDS-PAGE and Western blot.
SDS-PAGE of Proteins
SDS-PAGE gels were purchased from Biorad and used in the appropriate gel chambers
manufactured by Biorad. Protein samples were prepared by heating to 90°C for 10
minutes in SDS loading buffer. The samples were then loaded into wells along with a
protein standard solution. Gels were run at 150V for 1 hour. After electrophoresis,
proteins were fixed and stained with EZBlue Staining reagent (Sigma) and the excess
dye was then washed from the gel using H2O.
Western Blot Analysis of Proteins
All materials and buffers used for Western blot analysis were purchased from Biorad in
transfer tanks produced by Biorad. After completing SDS-PAGE gel electrophoresis,
the gel was incubated on the orbital shaker for 10 minutes in blotting buffer. Before
assembling the “transfer-sandwich” (bottom-blotting paper-membrane-gel-blotting
paper-top) the blotting paper and nitrocellulose membrane were soaked in blotting
buffer for 5 minutes. Blotting was performed at 15 V for 25 minutes. The membrane
was then incubated on the orbital shaker, with blocking buffer for 1 hour, then
incubated on the orbital shaker for another hour with the anti-His Tag antibody (1/2000
(v/v) diluted in blocking buffer). The membrane was then washed three times by
incubating on the orbital shaker with TTBS buffer for 5 minutes. The blot was then
incubated with aqueous 3,3’-diaminobenzidine solution until staining occurred.
Membranes were rinsed with distilled water, dried and then stored for analysis.
CHAPTER 8
169
8.2 Experimental Details for Chapter 3
Unless stated otherwise, all chemicals were of analytical grade and used as received
from Sigma-Aldrich. All solvents used were from commercial suppliers (Sigma-
Aldrich, Fisher Scientific or Romil). NMR spectra were recorded on Bruker 400
UltrashieldTM
or 600 UltrashieldTM
spectrometers at room temperature and calibrated
according to the chemical shift of tetramethysilane or 3-(trimethylsilyl)propionic-
2,2,3,3-d4 acid sodium salt for samples in D2O (δ = 0 ppm). All compound NMR
spectra were assigned by Dr R. Sardzik (The University of Manchester) using 1H,
13C,
DEPT, COSY, HSQC, HMQC and HMBC NMR experiments as appropriate. Chemical
shifts are given in ppm, coupling constants in Hertz (Hz) and multiplicities indicated
with the appropriate abbreviations: singlet (s), doublet (d), triplet (t), double doublet
(dd), double double doublet (ddd) and multiplet (m). The determination of
diastereomeric ratios are based on comparison of signal intensities of separated signal
pairs in 13
C NMR spectra. ES+ mass spectra were obtained with Waters Micromass
spectrometer. MALDI spectra were obtained using a Bruker Ultraflex TOF/TOF
spectrometer. IR spectra were measured and recorded using a PerkinElmer Sprctrum
RX I FT-IR Spectrometer. Optical activity was measured using an Optical Activity Ltd
AA-1C00 polarimeter. Melting points were measured with a Gallenkamp apparatus and
are not corrected.
8.2.1 General Procedure 1: Peracetylation with Acetic Anhydride and Pyridine177
Glycoside was dissolved in pyridine, 5 mol eq of acetic anhydride was added slowly
and the reaction was monitored by thin layer chromatography (TLC) and more acetic
anhydride was added if required. Most of the pyridine was removed using a rotary
evaporator and the resulting slurry was then dissolved in ethyl acetate. This solution
was washed with CuSO4 (aq), then water, then brine and dried with magnesium
sulphate. The resulting solution was then filtered and the solvent removed to yield the
product.
CHAPTER 8
170
8.2.2 General Procedure 2: Deacetylation with Sodium Methoxide177
Acetylated 2-(benzyloxycarbonyl)aminoethyl glycoside was dissolved in methanol and
NaOMe in methanol (0.33 mol eq) was added. The reaction was then stirred for 16
hours at r.t. The base was neutralised with activated Amberlite IR-120, the resin was
then removed via filtration and the solvent evaporated in vacuo to yield the product.
8.2.3 General Procedure 3: Hydrogenolysis of N-Cbz-protecting Groups177
2-(Benzyloxycarbonyl)aminoethyl glycoside was dissolved in MeOH and Pd/C (10%)
was added. The reaction was then stirred under a H2 atmosphere for 16 hours. The
solution was then filtered through Celite and the solvent removed in vacuo to yield the
free amine.
8.2.4 Synthesis of Aminoethyl Mannoside (27)177
1,2,3,4,6-Penta-O-acetyl-D-mannopyranose (30)177
Figure 8.1 The reaction of mannose (29) with acetic anhydride to form peracetyl mannose (30).
5.00 g of D-(+)-Mannose (29) (27.9 mmol, 1 mol eq) was reacted with 13 mL of acetic
anhydride (138.5 mmol, 5 mol eq) in pyridine as described in general procedure 1
(section 8.2.1) to yield 10.6 g 30 (mixture of both anomers α/β 33:67) as a clear viscous
oil (27.1 mmol, 97%).
1H NMR (400 MHz, CDCl3, mixture of both anomers): signals of β-anomer δ (ppm) =
1.98 (s, 3H, COCH3), 2.07 (s, 3H, COCH3), 2.08 (s, 3H, COCH3), 2.15 (s, 3H, COCH3),
2.19 (s, 3H, COCH3), 3.99–4.05 (m, 1H, 5-H), 4.07 (dd, J = 2.4, 12.4 Hz, 1H, 6-Ha),
4.26 (dd, J = 4.9, 12.4 Hz, 1H, 6-Hb), 5.23 (dd, J = 2.0, 3.1 Hz, 1H, 3-H), 5.31–5.34 (m,
CHAPTER 8
171
2H, 3-H, 4-H), 6.06 (d, J = 2.0 Hz, 1H, 1-H); signals of α-anomer δ (ppm) = 1.98 (s,
3H, COCH3), 2.03 (s, 3H, COCH3), 2.07 (s, 3H, COCH3), 2.15 (s, 3H, COCH3), 2.16 (s,
3H, COCH3), 3.79 (ddd, J = 2.4, 5.3, 9.9 Hz, 1H, 5-H), 4.11 (dd, J = 2.4, 12.4 Hz, 1H,
6-Ha), 4.28 (dd, J = 5.3, 12.4 Hz, 1H, 6-Hb), 5.11 (dd, J = 3.3, 10.0 Hz, 1H, 3-H), 5.27
(t, J = 10.0 Hz, 1H, 4-H), 5.46 (dd, J = 1.2, 3.3 Hz, 1H, 2-H), 5.84 (d, J = 1.2 Hz, 1H, 1-
H); 13
C NMR (101 MHz, CDCl3, mixture of both anomers): signals of β-anomer: δ
(ppm) = 20.7, 20.7, 20.8, 20.8, 20.9 (5 COCH3), 62.2 (C-6), 65.6 (C-4), 68.4 (d, C-2),
68.8 (d, C-3), 70.7 (d, C-5), 90.7 (d, C-1), 168.2, 169.6, 169.8, 170.1, 170.7 (5 COCH3);
signals of α-anomer: δ (ppm) = 20.6, 20.7, 20.8, 20.8, 20.9 (5 COCH3), 62.1 (C-6), 65.4
(C-4), 68.3 (C-2), 70.7 (C-3), 73.4 (C-5), 90.5 (C-1), 168.5, 169.7, 169.9, 170.3, 170.8
(5 COCH3).
2-(Benzyloxycarbonyl)aminoethyl 2,3,4,6-tetra-O-acetyl-α-D-mannopyranoside
(31)177
Figure 8.2 The reaction of peracetylated mannose (30) with benzyl N-(2-hydroxyethyl)-carbamate to
form mannoside (31).
5.62 g of peracetylated mannose (30) (14.41 mmol) and of 8.35 g of benzyl N-(2-
hydroxyethyl)-carbamate (17.29 mmol, 1.2 mol eq) were dissolved in 75 mL of dry
dichloromethane (DCM) under nitrogen. The solution was cooled to 0 °C and 17.9 mL
BF3.Et2O (72.05 mmol, 5 mol eq) was added slowly. The reaction was stirred for 30 min
at 0°C and then for 16 hours at r.t. The reaction was quenched with 10 mL of H2O and
then concentrated in vacuo. The residue was re-dissolved in DCM then washed with sat.
(saturated) NaHCO3, then water and then brine. The organic layers were dried over
MgSO4, the solvent removed under reduced pressure and then purified using column
chromatography on silica (EtOAc/hexane 40:60 to 50:50) to yield 4.22 g 31 as a clear
oil (8.21 mmol, 57%).
CHAPTER 8
172
20
D = +165 (c 1.8, CH2Cl2); 1H NMR (400 MHz, CDCl3): δ (ppm) = 2.00 (s, 3H,
COCH3), 2.04 (s, 3H, COCH3), 2.09 (s, 3H, COCH3), 2.16 (s, 3H, COCH3), 3.36–3.53
(m, 2H, CH2NH), 3.58 (ddd, J = 3.6, 6.8, 10.2 Hz, 1H, CHaHbCH2NH), 3.78 (ddd, J =
3.9, 6.2 10.2 Hz, 1H, CHaHbCH2NH), 3.97 (ddd, J = 2.3, 5.7 9.5 Hz, 1H, 5-H), 4.08 (dd,
J = 2.3, 12.2 H 1H, 6-Ha), 4.26 (dd, J = 5.7, 12.2 Hz, 1H, 6-Hb), 4.82 (d, J = 1.7 Hz, 1H,
1-H), 5.12 (s, 2H, CH2Ph), 5.20 (t, J = 5.8 Hz, 1H, NH), 5.25 (dd, J = 1.7, 3.2 Hz, 1H,
2-H), 5.26 (dd, J = 9.5, 10.1 H), 5.31 (dd, J = 3.2, 10.0 Hz, 1H, 3-H), 7.29–7.39 (m, 5H,
C6H5); 13
C NMR (101 MHz, CDCl3) δ (ppm) = 20.7, 20.8, 20.9 (4 COCH3), 40.7
(CH2NH), 62.5 (C-6), 66.1 (C-4), 66.9 (CH2Ph), 67.8 (CH2CH2NH), 68.8 (C-5), 69.0
(C-3), 69.4 (C-2), 97.8 (C-1), 128.2, 128.6 (o-, m-, p-C from C6H5), 136.4 (i-C from
C6H5), 156.4 (NCOO), 169.8, 170.0, 170.1, 170.7 (4 COCH3); IR: ~ (cm−1
) = 3391 (N-
H stretch), 2936 (C-H stretch), 1748 (C=O stretch), 1531 (C=C aromatic stretch), 1367,
1227, 1140, 1088, 1047, 980; HRMS (ESI+): m/z calcd for C24H31NO12 [M+H]+
526.1925, found 526.1913.
2-(Benzyloxycarbonyl)aminoethyl α-D-mannopyranoside (42)177
Figure 8.3 The deprotection of mannoside (31) with NaOH and MeOH to give mannoside (27).
Aminoethyl mannoside (31) (4.66 g, 8.87 mmol) was deacetylated using NaOMe in
MeOH as described by to General Procedure 2 (section 8.2.2) to yield 3.01 g 42 as a
colourless oil (8.42 mmol, 95%).
20
D = +34.7 (c 2.3, MeOH); 1H NMR (400 MHz, MeOD): δ (ppm) = 3.27–3.39 (m,
2H, CH2NH), 3.47–3.55 (m, 2H, 5-H, CHaHbCH2NH), 3.60 (t, J = 9.5 Hz, 1H, 4-H),
3.68 (dd, J = 5.8, 11.7 Hz, 1H, 6-Ha), 3.69 (dd, J = 3.4, 9.3 Hz, 1H, 3-H), 3.74 (ddd, J =
4.9, 6.4, 10.2 Hz, 1H, CHaHbCH2NH), 3.80 (dd, J = 1.7, 3.4 Hz, 1H, 2-H), 3.81 (dd, J =
2.3, 11.7 Hz, 1H, 6-Hb), 4.75 (d, J = 1.6 Hz, 1H, 1-H), 5.06 (s, 2H, CH2Ph) , 7.24–7.36
CHAPTER 8
173
(m, 5H, C6H5); 13
C NMR (101 MHz, MeOD) δ (ppm) = 41.7 (CH2CH2NH), 62.8 (C-6),
67.5 (CH2Ph, CH2CH2NH), 68.5 (C-4), 72.0 (C-2), 72.5 (C-3), 74.7 (C-5), 101.6 (C-1),
128.8, 129.0, 129.5 (o-, m-, p-C from C6H5), 138.3 (i-C from C6H5), 158.9 (NCOO); IR:
~ (cm−1
) = 3593-3000 (O-H stretch), 2929 (C-H stretch), 1698, 1535 (C=C aromatic
stretch), 1451, 1409, 1335, 1262, 1136, 1094, 1060, 1027, 975, 912, 880; HRMS
(ESI+): m/z calcd for C16H23NO8 [M+Na]+ 380.1321, found 380.1316.
2-Aminoethyl α-D-mannopyranoside (27)177
Figure 8.4 The hydrogenation of mannoside (42) using a Pd/C catalyst and hydrogen gas to give
mannoside (27).
Prepared from aminoethyl glycoside (42) (2.85 g, 7.98 mmol) by hydrogenation using
Pd/C and H2 in MeOH as described in General Procedure 3 (section 8.2.3) to give 1.66
g 27 as a colourless oil (7.44 mmol, 93%).
20
D = +66 (c 2.7, MeOD); 1H NMR (400 MHz, MeOD): δ (ppm) = 2.82–2.86 (m, 2H,
CH2NH2), 3.48 (ddd, J = 4.7, 5.9, 10.2 Hz, 1H, CHaHbCH2NH2), 3.56 (ddd, J = 2.1, 5.8,
9.7 Hz, 1H, 5-H), 3.63 (t, J = 9.4 Hz, 1H, 4-H), 3.73 (dd, J = 5.8, 11.8 Hz, 1H, 6-Ha),
3.74 (dd, J = 3.4, 9.1 Hz, 1H, 3-H), 3.79 (ddd, J = 4.7, 5.9, 10.2 Hz, 1H,
CHaHbCH2NH2), 3.86 (dd, J = 1.7, 3.4 Hz, 1H, 2-H), 3.86 (dd, J = 2.1, 11.8 Hz, 1H, 6-
Hb), 4.80 (d, J = 1.7 Hz, 1H, 1-H); 13
C NMR (101 MHz, MeOD) δ (ppm) = 101.7 (C-1),
74.6 (C-5), 72.5 (C-4), 72.0 (C-2), 70.0 (CH2CH2NH2), 68.6 (C-3), 62.8 (C-6), 42.0
(CH2NH2); IR: ~ (cm−1
) = 3605-3100 (O-H stretch), 2928 (C-H stretch), 1645 (N-H
bend), 1598, 1454, 1418, 1361, 1320, 1258, 1207, 1134, 1062, 975, 877, 805; HRMS
(ESI+): m/z calcd for C8H17NO6 [M+H]+ 224.1134, found 224.1138.
CHAPTER 8
174
8.2.5 Synthesis of Aminoethyl Glucoside (32)177
2-(Benzyloxycarbonyl)aminoethyl 2,3,4,6-tetra-O-acetyl-β-D-glucopyranoside
(39)177
Figure 8.5 The reaction of Peracetyl β-D-glucopyranose (36) and N-Cbz-ethanolamine to produce
glucoside (39).
Peracetyl β-D-glucopyranose (36) (1.00 g, 2.56 mmol) and N-Cbz-ethanolamine (1.2
mol eq) were dissolved in dry acetonitrile under nitrogen. The solution was cooled to
0°C and BF3.Et2O (5 mol eq) was added slowly. The reaction was stirred for 30 min at
0°C and then overnight at r.t. The reaction was quenched with water and concentrated in
vacuo, the residue re-dissolved in dichloromethane then washed once with sat.
NaHCO3, water and then brine. The organic layers were dried using MgSO4, the solvent
removed in vacuo and the product purified using column chromatography on silica
(EtOAc/petroleum ether 40:60) to yield 471 mg 39 as a clear oil (0.90 mmol, 35%).
21
D = −4.1 (c 1.0, CHCl3); 1H NMR (400 MHz, CDCl3): δ (ppm) = 2.00 (s, 6H, 2
COCH3), 2.03 (s, 3H, COCH3), 2.06 (s, 3H, COCH3), 3.37–3.41 (m, 2H, CH2NH), 3.68
(ddd, J = 2.5, 4.8, 9.9 Hz, 1H, 5-H), 3.69–3.74 (m, 1H, OCHaHbCH2), 3.87 (ddd, J =
4.1, 5.5, 10.0 Hz, 1H, OCHaHbCH2), 4.14 (dd, J = 2.4, 12.3 Hz, 6-Ha), 4.14 (dd, J = 4.8,
12.4 Hz, 1H, 6-Hb), 4.48 (d, J = 8.0 Hz, 1H, 1-H), 4.93 (dd, J = 8.0, 9.6 Hz, 1H, 2-H),
5.05 (dd, J = 9.4, 9.7 Hz, 1H, 4-H), 5.09 (s, 2H, CH2Ph), 5.17 (m, 1H, NHCBz), 5.19
(dd, J = 9.4, 9.6 Hz, 3-H), , 7.33–7.36 (m, 5H, C6H5); 13
C NMR (101 MHz, CDCl3) δ
(ppm) = 20.9, 21.0 (4 COCH3), 41.1 (CH2NH), 62.2 (C-6), 67.1 (CH2Ph), 68.7 (C-4),
69.8 (OCH2CH2), 71.6 (C-2), 72.3 (C-5), 73.0 (C-3), 101.4 (C-1), 128.5, 128.5, 128.9
(o-, m-, p-C from C6H5), 136.8 (i-C from C6H5), 156.7 (NCOO), 169.7, 169.8, 170.5,
170.9 (4 COCH3); IR: ~ (cm−1
) = 3379 (N-H stretch), 2946 (C-H stretch), 1755 (C=O
stretch), 1723, 1529 (C=C aromatic stretch), 1431, 1369, 1226, 1039; HRMS (ESI+):
m/z calcd for C24H31NO12 [M+H]+ 526.1925, found 526.1920.
CHAPTER 8
175
2-(Benzyloxycarbonyl)aminoethyl β-D-glucopyranoside (64)177
Figure 8.6 The deprotection of glucoside (39) with NaOH in MeOH to produce glucoside (64).
Aminoethyl glucoside (39) (370 mg, 0.705 mmol) was deacetylated as described in
general procedure 2 (section 8.2.2) to yield 227 mg 64 as clear oil, (0.635 mmol, 90%).
21
D = −11.7 (c 1.0, MeOD); 1H NMR (400 MHz, MeOD): δ (ppm) = 3.17 (dd, J = 7.9,
9.0 Hz, 1H, 2-H), 3.22–3.29 (m, 3H, 4-H, 5-H, CHaHbNH), 3.32–3.38 (m, 1H,
CHaHbNH), 3.34 (t, J = 8.7 Hz, 3-H), 3.58 (ddd, J = 4.2, 6.9, 10.4 Hz,, 1H,
CHaHbCH2NH), 3.62 (dd, J = 5.2, 12.0 Hz, 1 H, 6-Hb), 3.81 (dd, J = 2.0, 11.9 Hz, 1H, 6-
Hb), 3.86 (ddd, J = 5.6, 7.7, 10.2 Hz, 1H, CHaHbCH2NH), 4.22 (d, J = 7.8 Hz, 1H, 1-H),
5.02 (s, 2H, CH2Ph), 7.23–7.29 (m, 5H, C6H5); 13
C NMR (101 MHz, MeOD) δ (ppm) =
42.7 (CH2NH), 63.3 (C-6), 68.2 (CH2Ph), 70.7 (CH2CH2NH), 72.2 (C-5), 75.7 (C-2),
78.5 (C-3, C-4), 105.1 (C-1), 128.2, 128.7, 129.6 (o-, m-, p-C from C6H5), 138.9 (i-C
from C6H5), 159.6 (NCOO); IR: ~ (cm−1
) = 3592-3000 (O-H stretch), 2946 (C-H
stretch), 1702, 1535 (C=C aromatic stretch), 1454, 1337, 1262, 1076, 1031; HRMS
(ESI+): m/z calcd for C16H23NO8 [M+Na]+ 380.1321, found 380.1318.
2-Aminoethyl β-D-glucopyranoside (32)177
Figure 8.7 The hydrogenation of glucoside (39) using a Pd/C catalyst and hydrogen gas to produce
glucoside (32).
CHAPTER 8
176
2-Benzyloxycarbonylaminoethyl-glucopyranoside (39) (182 mg, 0.51 mmol) was
hydrogenated overnight as described in general procedure 3 (section 8.2.3) to yield 110
mg 32 as a white solid (0.49 mmol, 96%).
21
D = +7.2 (c 1.0, MeOH); 1H NMR (400 MHz, MeOD): δ (ppm) = 2.84–2.86 (m, 2H,
CH2NH2), 3.15 (dd, J = 7.8, 9.2 Hz, 1H, 2-H), 3.21–3.24 (m, 2H, 4-H, 5-H), 3.31 (dd, J
= 9.0, 9.1 Hz, 1H, 3-H), 3.58–3.63 (m, 2H, 6-Ha, CHaHbCH2NH), 3.81 (dd, J = 1.3,
11.9 Hz, 1H, 6-Hb), 3.88 (ddd, J = 5.0, 7.7, 9.9 Hz, 1H, CHaHbCH2NH), 4.22 (d, J =
7.8 Hz, 1H, 1-H); 13
C NMR (101 MHz, MeOD) δ (ppm) = 42.7 (CH2NH2), 63.5 (C-6),
71.5 (CH2CH2NH2), 72.4 (C-5), 75.9 (C-2), 78.7 (C-3), 78.8 (C-4), 105.2 (C-1); IR: ~
(cm−1
) = 3584-3000 (O-H stretch), 2945 (C-H stretch), 1645 (N-H bend), 1319, 1078,
1039, 614; HRMS (ESI+): m/z calcd for C8H17NO6 [M+H]+ 224.1134, found 224.1135.
8.2.6 Synthesis of Aminoethyl Galactoside (33)177
2-(Benzyloxycarbonyl)aminoethyl 2,3,4,6-tetra-O-acetyl-β-D-galactopyranoside
(40)177
Figure 8.8 The reaction of peracetyl β-D-galactose (37) with N-Cbz-ethanolamine to produce galactoside
(40).
Peracetyl β-D-Gal (37) (1.95 g, 5.00 mmol) and N-Cbz-ethanolamine (6 mmol, 1.2 mol
eq) were dissolved in dry acetonitrile under nitrogen. The solution was cooled to 0°C
and BF3.Et2O (25 mmol, 5 mol eq) was added slowly. The reaction was stirred for 30
min at 0°C and then overnight at r.t. The reaction was quenched with water and
concentrated in vacuo, the residue re-dissolved in DCM then washed once with sat.
NaHCO3, water and brine. The organic layers were dried using MgSO4, the solvent
removed in vacuo and the product purified using column chromatography on silica
(ethyl acetate/hexane 50:50) to yield 1.58 g 40 as a clear oil (3 mmol, 60% yield).
CHAPTER 8
177
21
D = −1.4 (c 1.0, CHCl3), Lit.210
22
D = +4.4 (c 1.2, CH2Cl2), Lit.211
20
D = +20.7 (c
1, CH2Cl2); 1H NMR (500 MHz, CDCl3): δ (ppm) = 1.90 (s, 3H, COCH3), 1.93 (s, 3H,
COCH3), 1.95 (s, 3H, COCH3), 2.07 (s, 3H, COCH3), 3.32 (m, 2H, CH2NH), 3.62 (ddd,
J = 3.6, 7.1, 10.2 Hz, 1H, CHaHbCH2NH), 3.80–3.82 (m, 2H, 5-H, CHaHbCH2NH),
4.06 (d, J = 6.6 Hz, 2H, 6-H2), 4.38 (d, J = 7.9 Hz, 1H, 1-H), 4.93 (dd, J = 3.4, 10.5 Hz,
1H, 3-H), 5.02 (s, 2H, CH2Ph), 5.10 (dd, J = 8.0, 10.4 Hz, 1H, 2-H), 5.19 (t, J = 5.4 Hz,
1H, NH), 5.31 (dd, J = 0.7, 3.4 H, 1H, 4-H), 7.22–7.30 (m, 5H, C6H5); 13
C NMR
(126 MHz, CDCl3) δ (ppm) = 20.5–20.7 (4 COCH3), 40.8 (CH2NH), 61.3 (C-6), 66.7
(CH2Ph), 67.0 (C-4), 68.8 (C-2), 69.4 (OCH2CH2NH), 70.7 (C-3, C-5), 101.5 (C-1),
128.1, 128.5 (o-, m-, p-C from C6H5), 136.5 (i-C from C6H5), 156.3 (NCOO), 169.6,
170.1, 170.2, 170.4 (4 COCH3); IR: ~ (cm−1
) = 3393 (N-H stretch), 2947 (C-H stretch),
1743 (C=O stretch), 1714, 1524 (C=C aromatic stretch), 1371, 1222, 1048; HRMS
(ESI+): m/z calcd for C24H31NO12 [M+Na]+ 548.1744, found 548.1747.
2-(Benzyloxycarbonyl)aminoethyl β-D-galactopyranoside (65)177
Figure 8.9 The deprotection of galactoside (40) using NaOH in MeOH to form galactoside (65).
Aminoethyl galactoside 40 (1.62 g, 3.08 mmol) was deacetylated using general
procedure 2 (section 8.2.2) to yield 1.10 g 65 (3.08 mmol, 90%) as a clear oil.
21
D = +1.8 (c 1.0, MeOH); 1H NMR (400 MHz, MeOD): δ (ppm) =3.30 (ddd, J = 4.2,
6.8, 14.2 Hz, 1H, CHaHbNH), 3.40 (ddd, J = 4.1, 6.2, 14.2 Hz, 1H, CHaHbNH), 3.46 (dd,
J = 3.2, 9.7 Hz, 1H, 3-H), 3.50 (ddd, J = 1.0, 5.3, 6.8 Hz, 1H, 5-H), 3.52 (dd, J = 7.3,
9.8 Hz, 1H, 2-H), 3.63 (ddd, J = 4.0, 6.8, 10.5 Hz, 1H, CHaHbCH2NH), 3.70 (dd, J =
5.3, 11.4 Hz, 1H, 6-Ha), 3.75 (dd, J = 6.9, 11.3 Hz, 1H, 6-Hb), 3.82 (dd, J = 1.0, 3.2 Hz,
1H, 4-H), 3.91 (ddd, J = 4.2, 6.2, 10.4 Hz, 1H, CHaHbCH2NH), 4.22 (d, J = 7.3 Hz, 1H,
1-H), 5.06 (s, 2H, CH2Ph), 7.24–7.37 (m, 5H, C6H5); 13
C NMR (101 MHz, MeOD) δ
CHAPTER 8
178
(ppm) = 42.0 (CH2NH), 62.4 (C-6), 67.4 (CH2Ph), 69.9 (CH2CH2NH), 70.2 (C-4), 72.5
(C-2), 74.8 (C-3), 76.6 (C-5), 105.0 (C-1), 128.8, 129.0, 129.5 (o-, m-, p-C from C6H5),
138.3 (i-C from C6H5), 158.9 (NCOO); IR: ~ (cm−1
) = 3580-3000 (O-H stretch), 2945
(C-H stretch), 1532 (C=C stretch), 1073,1042; HRMS (ESI+): m/z calcd for C16H23NO8
[M+Na]+ 380.1316, found 380.1308.
2-Aminoethyl β-D-galactopyranoside (33)177
Figure 8.10 The hydrogenation of galactoside (65) using a Pd/C catalyst and hydrogen gas to produce
galactoside (33).
Aminoethyl galactoside 65 (1.00 g, 3.89 mmol) was hydrogenated as described in
general procedure 3 (section 8.2.3) to yield 806 mg 33 as a colourless oil (3.62 mmol,
93%).
20
D = −12.9 (c 2.4, MeOH), Lit.212
20
D = −11.3 (c 0.23, MeOH); 1H NMR
(400 MHz, MeOD): δ (ppm) = 2.80 (ddd, J = 4.2, 6.3, 13.4 Hz, 1H, CHaHbNH2), 2.84
(ddd, J = 4.4, 5.5, 13.4 Hz, 1H, CHaHbNH2), 3.45 (dd, J = 3.3, 9.7 Hz, 1H, 3-H), 3.49
(ddd, J = 1.0, 5.3, 7.0 Hz, 1H, 5-H), 3.52 (dd, J = 7.5, 9.7 Hz, 1H, 2-H), 3.61 (ddd, J =
4.4, 6.3, 10.5 Hz, 1H, CHaHbCH2NH2), 3.69 (dd, J = 5.3, 11.3 Hz, 1H, 6-Ha), 3.73 (dd, J
= 7.0, 11.3 Hz, 1H, 6-Hb), 3.80 (dd, J = 1.0, 3.3 Hz, 1H, 4-H), 3.91 (ddd, J = 4.2, 5.5,
10.3 Hz, 1H, CHaHbCH2NH2), 4.21 (d, J = 7.5 Hz, 1H, 1-H); 13
C NMR (101 MHz,
MeOD) δ (ppm) = 42.2 (CH2NH2), 62.5 (C-6), 70.3 (C-4), 71.9 (CH2CH2NH2), 72.6 (C-
2), 74.9 (C-3), 76.7 (C-5), 105.1 (C-1); IR: ~ (cm−1
) = 3575-3100 (O-H stretch), 2929
(C-H stretch), 1644 (N-H bend), 1264, 1077; HRMS (ESI+): m/z calcd for C8H17NO6
[M+H]+ 224.1134, found 224.1133.
CHAPTER 8
179
8.2.7 Synthesis of Aminoethyl N-Acetyl glucosamine (34)177
2-(Benzyloxycarbonyl)aminoethyl 2-acetamido-3,4,6-tri-O-acetyl-2-deoxy-β-D-
gluco-pyranoside (38)177
Figure 8.11 The reaction of β-D-Glucosamine pentaacetate (35) with of N-Cbz-ethanolamine to produce
glucoside (38).
1.00 g of β-D-Glucosamine pentaacetate (35) (Acros Organics, 2.57 mmol) and 1.25 g
of N-Cbz-ethanolamine (6.43 mmol, 2.5 mol eq) were dissolved in 10 mL of dry DCM,
under nitrogen. The reaction was cooled to 0°C and 360 µL of SnCl4 (3.08 mmol, 1.2
mol eq) was added slowly. The reaction was heated to 75°C for 16 h. The reaction was
then allowed to cool to r.t., quenched with 2 mL Et3N and concentrated in vacuo. The
residue was dissolved in DCM and washed with water. The organic phase was then
dried over MgSO4 and reduced in vacuo. The product was isolated using column
chromatography on silica (EtOAc/hexane 80:20), yielding a white solid which was
recrystallised from chloroform/ethyl acetate to obtain 512 mg 38 as colourless crystals
(2.44 mmol, 38%).
m.p. = 124–127 °C; 20
D = −70.1 (c 2.1, CH2Cl2), Lit.213
24
D = −15 (c 1.0, CHCl3);
1H NMR (500 MHz, CDCl3): δ (ppm) = 1.82 (s, 3H, COCH3), 1.96 (s, 6H, 2 COCH3),
1.98 (s, 3H, COCH3), 3.21–3.29 (m, 1H, CHaHbNH), 3.33–3.42 (m, 1H, CHaHbNH),
3.61 (m, 2H, 5-H, CHaHbCH2NH), 3.80 (ddd, J = 3.5, 5.9, 9.9 Hz, 1H, CHaHbCH2NH),
3.84 (dt, J = 8.6, 10.2 Hz, 1H, 2-H), 4.06 (dd, J = 2.0, 12.3 Hz, 1H, 6-Ha), 4.16 (dd, J =
4.8, 12.3 Hz, 1H, 6-Hb), 4.54 (d, J = 8.3 Hz, 1H, 1-H), 4.98 (dd, J = 9.5, 9.8 Hz, 1H, 4-
H), 5.02 (s, 2H, CH2Ph), 5.12 (dd, J = 9.8, 10.2 Hz, 1H, 3-H), 5.31 (t, J = 5.2 Hz, 1H,
CH2NH), 5.87 (d, J = 8.8 Hz, 1H, 2-NH), 7.23–7.31 (m, 5H, C6H5); 13
C NMR
(126 MHz, CDCl3) δ (ppm) = 20.7, 20.7, 20.8, 23.2 (4 COCH3), 40.7 (CH2NH), 54.4
(C-2), 62.1 (C-6), 66.7 (CH2Ph), 68.5 (C-4), 69.1 (CH2CH2NH), 71.8 (C-5), 72.4 (C-3),
CHAPTER 8
180
101.1 (C-1), 128.1, 128.2, 128.6 (o-, m-, p-C from C6H5), 136.6 (i-C from C6H5), 156.6
(NCOO), 169.5, 170.7, 170.9, 171.0 (4 COCH3); IR: ~ (cm−1
) = 3317 (N-H stretch),
2949 (C-H stretch), 1743 (C=O stretch), 1701, 1660, 1547 (C=C aromatic stretch),
1433, 1377, 1243, 1171, 1150, 1047; HRMS (ESI+): m/z calcd for C24H32N2O11 [M+H]+
525.2084, found 525.2077.
2-(Benzyloxycarbonyl)aminoethyl 2-acetamido-2-deoxy-β-D-glucopyranoside
(66)177
Figure 8.12 The deprotection of glucoside (38) using NaOH and MeOH to produce glucoside (66).
Aminoethyl pyranoside 38 (4.14 g, 7.89 mmol) was deacetylated as described by
general procedure 2 (section 8.2.2) to yield 3.11 g 66 as a white foam (7.81 mmol,
99%).
m.p. = 71-74 °C 20
D = −9.6 (c 2.1, MeOH), Lit.213
24
D = −21.5 (c 1.0, CHCl3); 1H
NMR (400 MHz, MeOD): δ (ppm) = 1.93 (s, 3H, COCH3), 3.23–3.34 (m, 4H, 4-H, 5-H,
CH2NH), 3.43 (dd, J = 8.4, 10.3 Hz, 1H, 3-H), 3.58 (ddd, J = 5.4, 5.6, 10.6 Hz, 1H,
CHaHbCH2NH), 3.65 (dd, J = 8.4, 10.3 Hz, 1H, 2-H), 3.66 (dd, J = 5.6, 11.9 Hz, 1H, 6-
Ha), 3.84 (m, 1H, CHaHbCH2NH), 3.86 (dd, J = 2.2, 12.0 Hz, 1H, 6-Hb), 4.38 (d, J =
8.4 Hz, 1H, 1-H), 5.05 (s, 2H, CH2Ph), 7.35–7.25 (m, 5H, C6H5); 13
C NMR (101 MHz,
MeOD) δ (ppm) = 23.0 (COCH3), 41.9 (CH2NH), 57.2 (C-2), 62.7 (C-6), 67.4 (CH2Ph),
69.5 (CH2CH2NH), 71.9 (C-4), 75.9 (C-3), 77.9 (C-5), 102.8 (C-1), 128.9, 129.0, 129.5
(o-, m-, p-C from C6H5), 138.2 (i-C from C6H5), 158.8 (NCOO), 174.0 (COCH3); IR:
~ (cm−1
) = 3612–3000 (O-H stretch), 2938 (C-H stretch), 2886 (C-H stretch), 1700
(C=O stretch), 1644, 1546 (C=C stretch), 1459, 1421, 1372, 1312, 1258, 1149, 1111,
CHAPTER 8
181
1073, 1035; HRMS (ESI+): m/z calcd for C18H26N2O8 [M+Na]+ 421.1587, found
421.1583.
2-Aminoethyl 2-acetamido-2-deoxy-β-D-glucopyranoside (34)177
Figure 8.13 The hydrogenation of glucoside (66) using a Pd/C catalyst and hydrogen gas to produce
glucoside (34).
Aminoethyl pyranoside 66 was hydrogenated as described by general procedure 3
(section 8.3.3) to yield 1.78 g 34 as a pale yellow foam (6.74 mmol, 96%).
m.p. = 83–86 °C 20
D = −28.3 (c 1.87, MeOH); 1H NMR (400 MHz, D2O): δ (ppm) =
2.05 (s, 3H, COCH3), 2.69–2.85 (m, 2H, CH2NH2), 3.34–3.52 (m, 2H, 4-H, 5-H), 3.52–
3.59 (m, 1H, 3-H), 3.59–3.67 (m, 1H, CHaHbCH2NH2), 3.71–3.80 (m, 2H, 2-H, 6-Ha),
3.87–4.01 (m, 2H, 6-Hb, CHaHbCH2NH2), 4.53 (d, J = 8.4 Hz, 1H, 1-H); 13
C NMR
(101 MHz, D2O) δ (ppm) = 25.0 (COCH3), 43.0 (CH2NH2), 58.5 (C-2), 63.6 (C-6), 72.8
(C-5), 74.6 (CH2CH2NH2), 76.6 (C-3), 78.7 (C-4), 104.3 (C-1), 177.6 (COCH3); IR: ~
(cm−1
) = 3610-3100 (O-H stretch), 2929 (C-H stretch), 2876 (C-H stretch), 2361, 1698
(C=O stretch), 1651 (N-H bend), 1551, 1451, 1430, 1372, 1315, 1262, 1152, 1115,
1073, 1036, 947, 900; HRMS (ESI+): m/z calcd for C10H20N2O6 [M+H]+ 265.1400,
found 265.1404.
CHAPTER 8
182
8.2.8 Activation of Glycosides for Glycosylation of Cysteines130
Figure 8.14 Activation of aminoethyl mannoside (27), for reaction with cysteines, via reaction with
iodoacetic anhydride to produce glycosyl iodoacetamide (28).
The aminoethyl glycoside was dissolved in 1 M sodium bicarbonate (pH 8) and
iodoacetic anhydride (2 mol eq) was added. The reaction was stirred at r.t., in the dark
and additional 1 M sodium bicarbonate was added if the pH of the reaction fell below 7.
The reaction was monitored by MS after 30 minutes and more iodoacetic anhydride was
added if required. When complete the reaction mixture was desalted by passing over
acidic ion exchange resin (DOWEX® 50WX8-100) followed by basic ion exchange
resin (DOWEX®
1x8-100(Cl)). The eluent was collected and lyophilise to yield the
glycosyl iodoacetamide in over 90% yield as a white powder. Due to the reactivity of
the products produced they were generally contaminated with the hydrolysis product so
a full characterisation was not undertaken. Product formation was verified by HRMS,
the details of which are shown in table 3.1.
8.2.9 Synthesis of Polymannosides178
2,3,4,6-Tetra-O-acetyl-α-D-mannopyranosyl Bromide (23)178
Figure 8.15 The reaction of peracetyl mannose (30) with HBr to produce acetobromo mannoside (46).
2 g (5.12 mmol) of peracetylated mannose (30) was dissolved in 15 mL DCM and
cooled to 0°C. 200 μL of acetic acid anhydride was added, followed by 4 mL of HBr
CHAPTER 8
183
(33% in acetic acid) dropwise. After 30 minutes the reaction was allowed to warm to
room temperature and was left stirring overnight. The reaction mixture was then diluted
with DCM, washed with sodium bicarbonate (until neutral), then water and then brine.
The organic phase was the dried with MgSO4, filtered and the solvent removed in vacuo
to give the product 46 in over 90% yield.
19
D = +122.0 (c 1, CHCl3), Lit.214 20
D = +118.3 (c 1, CHCl3); 1H NMR (400MHz,
CDCl3): δH 2.02 (s, 3H, COCH3), 2.08 (s, 3H, COCH3), 2.11 (s, 3H, COCH3), 2.18 (s,
3H, COCH3), 4.14 (dd, J = 2.0, 12.4 Hz, 1H, 6-Ha) 4.23 (ddd, J = 2.0, 4.8, 10.0 Hz, 1H,
5-H) 4.34 (dd, J = 4.8, 12.4 Hz, 1H, 6-Hb) 5.37 (dd, J = 10.4, 10.4 Hz, 1H, 4-H) 5.45
(dd, J = 1.6, 3.6 Hz, 1H, 2-H) 5.71 (dd, J = 3.6, 10.4 Hz, 1H, 3-H) 6.32 (d, J = 1.6 Hz,
1H, 1-H). 13
C NMR (400MHz, CDCl3) δ (ppm) = 20.5, 20.6, 20.7 (3s, 4 x COCH3),
61.3 (s, CH2, C-6), 65.2 (s, CH, C-4), 67.8 (s, CH, C-3), 72.0 (s, CH, C-5), 72.7 (s, CH,
C-2), 83.0 (s, CH, C-1), 169.4, 169.6, 170.4 (3s, 4 x COCH3). IR: ~ (cm−1
) = 2927 (C-
H stretch), 1749 (C=O stretch), 1368, 1220, 1128, 1051, 908; HRMS (ESI+): m/z calcd
for C14H20BrO9 [M+H]+ 411.0291, found 411.0286.
2-(Benzyloxycarbonyl)aminoethyl 3,6-di-O-(2,3,4,6-tetra-O-acetyl-α-D
mannopyranosyl)-α-D-mannopyranoside (43)
Figure 8.16 The reaction of acetobromo mannoside (46) and mannoside (42) to produce trimannoside
(43).
CHAPTER 8
184
0.5 g (1.4 mmol) of 2-(Benzyloxycarbonyl)aminoethyl -D-mannopyranoside (42) was
dissolved in dry acetonitrile by heating the solution to 35°C and sonicating , under
nitrogen. 2 g of 4 Å molecular sieves was added and the solution stirred for 10 minutes.
Then HgBr2 (1.01 g, 2.8 mmol, 2 mol eq), Mg(CN)2 (709 mg, 2.8 mmol, 2 mol eq) and
a solution of acetobromo mannoside (46) (1.15 g, 2.8 mmol, 2 mol eq) in dry
acetonitrile were added sequentially. The reaction mixture was stirred at 35C for 1
hour then filtered through celite. Evaporation of the solvent gave a residue, which was
extracted 3 times with DCM. The extracts were combined and washed successively with
sat. KCl (aq), sat. NaHCO3 (aq), and water. The organic layer was dried, filtered, and
then concentrated in vacuo to yield syrup containing a mixture of glycosides. The purity
of this syrup was enhanced by column chromatography on silica (EtOAc/cyclohexane
80:20 to 90:10) to give a trimannoside rich fraction.
This mixture was dissolved in 5 mL methanol, mixed with an aqueous solution of
sodium periodate (250 mg) and stirred for 48 hours at r.t. The reaction mixture was
diluted with water and then extracted with DCM three times. The combined organic
layers were then washed with water, dried with MgSO4, filtered and concentrated in
vacuo to yield 192 mg pale yellow syrup. 58 mg (57 μmol) of a clear syrup was isolated
by column chromatography on silica (EtOAc/cyclohexane 60:40 to 80:20), giving 4.1%
yield of trimannoside (43) with respect to aminoethyl mannoside (42).
Due to the increased challenge in characterising this product compared to the
monosaccharide glycosides produced, NMR spectra were obtained in CDCl3 at 293 K
on a Bruker AVANCE 600 MHz spectrometer. Correlations were determined using
HSQC_TOCSY and HMBC spectra (appendices 6 and 7 respectively), whilst chemical
shift assignments were made using multiplicity edited HSQC (figure 8.18) and 13
C
spectra by Dr. Robert Sardzik (The University of Manchester). A full assignment of the
carbon and proton chemical shifts of each mannose moiety in trimannoside (43) is given
in table 8.7 and the numbering scheme used for the assignment is shown in figure 8.1.7.
CHAPTER 8
185
20
D = +42 (c 1.4, CH2Cl2); 1H NMR (600 MHz, CDCl3): δH (ppm) = 1.98 (3H,
COCH3), 2.01 (3H, COCH3), 2.04 (3H, COCH3), 2.07 (3H, COCH3), 2.08 (3H,
COCH3), 2.11 (3H, COCH3), 2.15 (3H, COCH3), 2.16 (3H, COCH3), 3.40 (1H,
CHaHCH2NH), 3.46 (1H, CHHbCH2NH), 3.55 (1H, CH2HaHNH), 3.75 (1H,
CH2HHbNH), 5.10 (H2, CH2Ph), 5.58 (1H, NH), 7.35, 7.36 (5H, C6H5); 13
C NMR (50
MHz, CDCl3): δC (ppm) = 20.7, 20.7, 20.9, 20.9 (8 COCH3), 40.5 (CH2CH2NH), 66.8
(CH2Ph), 67.0 (CH2CH2NH), 128.1, 128.2, 128.5, 136.5 (6 C6H5), 156.4 (NCOO),
169.8, 169.8, 170.1, 170.1, 170.2, 170.6, 170.9 (8 COCH3). IR: ~ (cm−1
) = 3570-3000
(O-H stretch), 2926 (C-H stretch), 2854 (C-H stretch), 1746 (C=O stretch), 1528 (C=C
aromatic stretch), 1434, 1370, 1224, 1136, 1045, 979, 939, 912; HRMS (ESI+): m/z
calcd for C44H59NO26 [M+Na]+ 1040.3223, found 1040.3203.
Figure 8.17 The numbering scheme used for the assignment of NMR spectra of trimannoside (43).
CHAPTER 8
186
Sugar
Residue
1H/
13C Chemical Shifts (ppm)
1 2 3 4 5 6a 6b
Man 4.81 4.10 3.83 3.93 3.75 3.80 3.91
99.95 69.76 81.78 65.59 71.50 66.43
Man’
(α-1,3-Man)
5.12 5.35 5.37 5.25 4.32 4.14 4.25
99.04 69.50 69.05 66.35 69.18 63.05
Man”
(α-1,6-Man)
4.93 5.25 5.35 5.28 4.12 4.16 4.26
97.24 69.78 69.00 66.19 68.55 62.54
Table 8.7 1H and
13C chemical shifts of the atoms found in the carbohydrate constituent of trimannoside
(43). Additional signals are listed below a long with the remaining characterisation undertaken.
Figure 8.18 A section of a multiplicity edited HSQC of trimannoside (43).
CHAPTER 8
187
The regiochemistry of the product could be confirmed by the coupling of C3 to H1’ in
the HMBC spectrum (figure 3.12). Similarly the stereochemistry of the product could
be confirmed by the interactions of the anomeric protons with their respective carbons
in a non-decoupled HSQC (Figure 3.9).
2-(Benzyloxycarbonyl)aminoethyl 3,6-di-O-(α-D-mannopyranosyl)-α-D-
Mannopyranoside (47)
Figure 8.19 The deprotection of trimannoside (43) with NaOH and MeOH to produce trimannoside (47).
50 mg aminoethyl trimannoside (43) (0.49 mmol) was deacetylated using NaOMe in
MeOH as described by General Procedure 2 (section 8.2.2) to yield 33 mg 47 (0.45
mmol, 92%) as a clear syrup.
20
D = +61.4 (c 1.4, CH2Cl2); 1H NMR (400 MHz, CDCl3): δH (ppm) = 8.39 (s, 1H),
7.47-7.36 (m, 5H), 5.17-5.06 (m, 3H), 4.10-4.05 (m, 2H), 3.95-3.91 (m, 2H), 3.90-3.85
(m, 5H), 3.84-3.80 (m, 1H), 3.79-3.71 (m, 5H), 3.70-3.63 (m, 4H), 3.62-3.54 (m, 2H);
13C NMR (50 MHz, CDCl3): δC (ppm) = 171.0, 158.3, 136.3, 128.8, 128.3, 127.6,
102.3, 100.0, 99.7, 99.4, 78.5, 73.3, 72.6, 71.2, 70.5, 70.3, 70.0, 69.9, 69.6, 66.8, 66.6,
66.4, 65.5, 60.9, 48.8, 40.1; IR: ~ (cm−1
) = 3612-3000 (O-H stretch), 2933 (C-H
stretch), 2821 (C-H stretch), 1703 (C=O stretch), 1596 (C=C aromatic stretch), 1455,
1352, 1266, 1134, 1062, 1029, 980; HRMS (ESI+): m/z calcd for C28H43NO18 [M+Na]+
704.2378, found 704.2393.
CHAPTER 8
188
3,6-di-O-(α-D-mannopyranosyl)-α-D-mannopyranoside (41)
Figure 8.20 The hydrogenation of trimannoside (47) using a Pd/C catalyst and hydrogen gas to produce
trimannoside (41).
33 mg (48 mmol) of 2-(Benzyloxycarbonyl) aminoethyl mannoside (47) was dissolved
in water and Pd/C (10%) was added. The reaction was then stirred under a H2
atmosphere for 16 hours. The solution was then filtered through Celite and the solvent
removed in vacuo to yield 31 mg aminoethyl glycoside 41 as a clear syrup (46 mmol,
95%).
m.p. 68-71°C; 20
D = +78.9 (c 1.4, CH2Cl2); 1H NMR (400 MHz, CDCl3): δH (ppm) =
8.31 (s, 2H), 4.99-4.97 (m, 1H), 4.77-4.75 (m, 1H), 4.04-3.97 (m, 1H), 3.95-3.92 (m,
1H), 3.92-3.46 (m, 19H), 3.03-2.97 (m, 1H); 13
C NMR (50 MHz, CDCl3): δC (ppm) =
171.0, 102.5, 100.0, 99.3, 78.5, 73.3, 72.7, 71.2, 70.5, 70.3, 69.9, 69.8, 69.5, 66.8, 66.6,
65.4, 65.2, 65.1, 61.0, 60.9, 39.2; IR: ~ (cm−1
) = 3620-3000 (O-H stretch), 2937 (C-H
stretch), 2923 (C-H stretch), 2844 (C-H stretch), 2360, 2343, 1591 (N-H bend), 1346,
1132, 1055, 1033, 979; HRMS (ESI+): m/z calcd for C20H37NO16 [M+H]+ 548.2191,
found 548.2178.
CHAPTER 8
189
Synthesis of 2-(Benzyloxycarbonyl)aminoethyl 3,4,6-tri-O-(2,3,4,6-tetra-O-acetyl-α-
D-mannopyranosyl)-α-D-mannopyranoside (48), 2-(Benzyloxycarbonyl)aminoethyl
2,3,6-tri-O-(2,3,4,6-tetra-O-acetyl-α-D-mannopyranosyl)-α-D-mannopyranoside
(49), and 2-(Benzyloxycarbonyl)aminoethyl 2,3,4,6-penta-O-(2,3,4,6-tetra-O-acetyl-
α-D-mannopyranosyl)-α-D-mannopyranoside (50)
Figure 8.21 The structures of polymannoside side products 48, 48 and 50.
2.12 g (5.94 mmol) of mannoside (42) was dissolved in dry acetonitrile by heating the
solution to 35°C and sonicating , under nitrogen. 10 g of 4 Å molecular sieves was
added and the solution stirred for 10 minutes. Then HgBr2 (14.93 g, 41.58 mmol, 7 mol
eq), mercuric cyanide (10.58 g, 41.58 mmol, 7 mol eq) and a solution of acetobromo
mannoside (46) (17.04 g 41.58 mmol, 7 mol eq) in dry acetonitrile were added
sequentially. The reaction mixture was stirred at 35C for 1 hour then filtered through
celite. Evaporation of the solvent gave a residue, which was extracted three times with
DCM. The extracts were combined and washed successively with sat. KCl (aq), sat.
CHAPTER 8
190
NaHCO3 (aq), and water. The organic layer was dried, filtered, and then concentrated in
vacuo to yield syrup containing a mixture of glycosides. The purity of this syrup was
enhanced by column chromatography on silica (EtOAc/cyclohexane 80:20 to 90:10) to
give a tetramannoside rich fraction.
This mixture was dissolved in 25 mL methanol, mixed with an aqueous solution of
sodium periodate (750 mg) and stirred for 48 hours at r.t. The reaction mixture was
diluted with water and then extracted with DCM three times. The combined organic
layers were then washed with water, dried with MgSO4, filtered and concentrated in
vacuo to yield 3.76 g syrup. This was purified by reverse phase HPLC using a Luna
C18 250 x 15 mm column.
HPLC methods were run at 4 mL/minute flow rate with a gradient of 20% acetonitrile
(aq) to 60 % acetonitrile (aq) over 5 minutes, followed by a gradient of 60 %
acetonitrile (aq) to 95 % acetonitrile (aq) over 20 minutes. A period of 10 minutes at 95
% acetonitrile (aq) to fully clean the column followed by 10 minutes of 20 %
acetonitrile (aq) to reequilibrate the system was added to the end of each run giving a
total run time of 50 minutes. All solvents used contained 0.1 % formic acid and were of
HPLC grade. Fractions were collected at elution times 21.25-22.15 minutes, 24.10-
25.30 minutes and 26.70-27.35 minutes. The fractions were reduced in vacuo then
lyophilised to yield purified 43, 48/49 and 50 respectively, each appearing as a
colourless foam. Typically 65 µL of 100 mg/mL polymannoside mixture was injected
per run. From 1 g of the polymannoside mixture purified in this way 228 mg of 43, 383
mg of 48/49 and 101 mg of 50 was isolated, corresponding to yields in relation to
aminoethyl mannoside (42) of 10.2%, 12.9% and 2.7% respectively.
Characterization of Side Products
Due to the complexity of the side products 48, 49 and 50 and the fact that 48 and 49
were isolated as a mixture, a full NMR assignment was not undertaken. However it was
CHAPTER 8
191
deemed prudent to include as much data as possible on these products for future
reference.
2-(Benzyloxycarbonyl)aminoethyl 2,3,6-tri-O-(2,3,4,6-tetra-O-acetyl-α-D-
mannopyranosyl)-α-D-mannopyranoside (49) and 2-
(Benzyloxycarbonyl)aminoethyl 3,4,6-tri-O-(2,3,4,6-tetra-O-acetyl-α-D-
mannopyranosyl)-α-D-mannopyranoside (48) mixture
m.p. 69-87°C; 20
D = +44.4 (c 1.4, CH2Cl2); 1H NMR (400 MHz, CDCl3): δH (ppm) =
7.31-7.21 (m, 5H), 5.31-5.12 (m, 10H), 2.02 (s, 2H), 4.97-4.69 (m, 2H), 4.32-3.61 (m,
16H), 3.51-3.27 (m, 3H), 2.11-1.86 (m, 36H); 13
C NMR (50 MHz, CDCl3): δC (ppm) =
170.8, 170.6, 170.4, 170.1, 170.1, 169.9, 169.9, 169.9, 169.9, 169.8, 169.8, 169.7,
156.5, 136.5, 128.5, 128.2, 128.1, 99.2, 99.0, 98.3, 97.2, 79.0, 78.4, 71.5, 69.8, 69.7,
69.6, 69.3, 69.2, 69.0, 68.9, 68.8, 68.5, 67.1, 67.0, 66.7, 66.6, 66.2, 66.1, 66.1, 62.9,
62.7, 62.4, 40.4, 20.9, 20.9, 20.8, 20.8, 20.8, 20.7, 20.7, 20.7; IR: ~ (cm−1
) = 3560-
3000 (O-H stretch), 2925 (C-H stretch), 1744 (C=O stretch), 1526 (C=C aromatic
stretch), 1434, 1369, 1220, 1136, 1043, 978, 939, 916; HRMS (ESI+): m/z calcd for
C58H77NO35 [M+H]+ 1348.4354, found 1348.4337.
2-(Benzyloxycarbonyl)aminoethyl 2,3,4,6-penta-O-(2,3,4,6-tetra-O-acetyl-α-D-
mannopyranosyl)-α-D-mannopyranoside (50)
m.p. 74-79°C; 20
D = +41.1 (c 1.4, CH2Cl2); 1H NMR (400 MHz, CDCl3): δH (ppm) =
7.30-7.22 (m, 5H), 5.32-5.12 (m, 17H), 5.07-4.99 (m, 4H), 4.86 (s, 1H), 4.26-4.17 (m,
3H), 4.13-3.90 (m, 12H), 3.85-3.65 (m, 5H), 3.53-3.34 (m, 3H), 2.11-1.86 (m 48H); 13
C
NMR (50 MHz, CDCl3): δC (ppm) = 170.6, 170.4, 170.2, 170.0, 170.0, 169.8, 169.7,
169.6, 169.5, 169.3, 128.5, 128.1, 96.9, 69.7, 69.4, 69.3, 68.9, 68.8, 68.5, 67.3, 66.7,
66.2, 65.9, 62.7, 62.0, 53.5, 29.7, 20.9, 20.8, 20.7, 20.7, 20.7; IR: ~ (cm−1
) = 2955 (C-
H stretch), 2924 (C-H stretch), 2853 (C-H stretch), 1743 (C=O stretch), 1543 (C=C
aromatic stretch), 1524, 1457, 1369, 1217, 1136, 1039, 977, 937, 917; HRMS (ESI+):
m/z calcd for C72H95NO44 [M+H]+ 1678.5305, found 1678.5288.
CHAPTER 8
192
8.3 Experimental Details for Chapter 4
Initial analysis of the GFPuv mutants produced was carried out on a Waters®
Micromass LTC, TOF-MS, using MassLynx™ 4.0 software for the data analysis.
Calibrations were made using commercially available HHM (Sigma) dissolved in
deionised water (0.25 mg/mL) and protein masses generally measured to 1 d.p. A
sample spectrum acquired for HHM is shown in figure 4.2.A and the corresponding
deconvoluted spectrum, produced by the software is also shown (figure 4.2.B). For
monitoring protein derivatisation reactions, calibrations were accepted in the range
16951-16952 for HHM (m/z 16951.49185
).
8.3.1 Glycosylation of GFPuv Mutants
GFPuv_WT purified by step wise IMAC (details in section 8.1.3) was diluted to a final
concentration of 0.1 mM in IMAC elution buffer (concentrations determined by BCA
assay). TCEP was then added to a final concentration of 0.1 mM per free cysteine and
the pH checked by spotting 1 µL of the reaction mixture on pH paper. If the pH was
below 7 then 10x IMAC elution buffer was added. If the pH was satisfactory the
solution was mixed using a rotary mixer for 15 minutes at r.t. Glycosyl iodoacetamide
was then added to a final concentration of 1 mM per free cysteine and the reaction
incubated for 2 hours at 20°C, in the dark with shaking. After 2 hours a 10 µL sample
was taken from the reaction and diluted in 20 µL deionised water before being analysed
by LCMS using method 8.3.2. If the reaction was found to be complete then the protein
sample would be passed through a Pd-10 column to quench the reaction and buffer
exchange in to the desired assay buffer. If the reaction was incomplete then additional
glycosyl iodoacetamide would be added according to the progress of the reaction and
the mixture reanalysed by LCMS after a further hour.
Glycosylation of Lysines
DTSSP (final concentration 5 mM) was added to a solution of the GFPuv_WT mutant
to be modified (50 µmol in PBS). This solution was mixed on a rotary mixer for 30
minutes at r.t. 10 % volume of 10x IMAC elution buffer was then added to the protein
CHAPTER 8
193
suspension before the addition of TCEP (final concentration 10 mM). The pH of the
mixture was checked and if found to be bellow 7 then additional buffer was added.
After a further 15 minutes mixing on the rotary mixer at r.t. the protein was redissolved
and could be passed through a Pd-10 column to remove unwanted small molecules and
buffer exchange into IMAC elution buffer ready for glycosylation.
8.3.2 MS Analysis of Proteins
Precipitation of GFPuv Mutants for ESI-MS Analysis
Initially samples of GFP were prepared by precipitation with 15 volumes of
acetone/methanol (1:1) followed by 1 hour at -20°C. These samples were centrifuged
and the solvent removed before being resuspended in deionised water to a final
concentration of 2 mg/mL before being analysed by mass spectrometry.
LCMS Analysis of Proteins
All LCMS analysis was conducted using an Agilent 1100 series HPLC system fitted
with a C4 Supelcosil LC-304 column, coupled to an Agilent 1100 LC/MSD SL
quadrupole mass spectrometer. Solvents used were HPLC grade or above and had 0.1 %
formic acid added upon opening.
Methods were run at 0.5 mL/minute flow rate with a gradient of 10% acetonitrile (aq) to
60 % acetonitrile (aq) over 15 minutes, followed by a gradient of 60 % acetonitrile (aq)
to 90 % acetonitrile (aq) over 5 minutes. A period of 5 minutes at 90 % acetonitrile (aq)
to fully clean the column followed by 5 minutes of 10 % acetonitrile (aq) to
reequilibrate the system was added to the end of each run giving a total run time of 30
minutes. MS measurements were only collected between 10 and 20 minutes in to the
run to avoid unnecessary contamination of the detector. Typically 1 µL of 0.5 mg/mL
protein sample was injected per run. However if this was insufficient to obtain a clear
CHAPTER 8
194
spectrum 2-5 µL could be tried. After every 5-10 runs it was beneficial to run a blank
sample (deionised water) to minimise the MS back ground.
MALDI-TOF Analysis of GFP Sample
All MALDI-TOF spectra were obtained using a Ultraflex TOF/TOF (Bruker) and
samples were analysed in linear positive mode. Aqueous GFP samples (1-3 mg/mL)
were mixed with Sinapinic acid (15 mg/mL, acetonitrile, 0.1 % formic acid) solution in
a 1:1 ratio. 1 µL of this mixture was spotted on a MALDI plate and left to dry naturally.
8.4 Experimental Details for Chapter 5
8.4.1 Enzymatic Screening of Mannosides
Alkanethiol spacers (HS-(CH2)17-EG3-OH) and linkers (HS-(CH2)17-EG6-OCH2COOH)
used for SAM formation, were purchased from Prochimia Surfaces (Poland). THAP
solutions were prepared by dissolving 5 mg of THAP in 300 µL water/acetone (1:1).
These solutions were prepared fresh weekly and stored in the fridge between uses.
SAM Formation on Gold Coated Plates
Gold plates were washed with Piranha solution (3:1, 96% H2SO4 (aq): 30% H2O2 (aq))
by submerging for 15 minutes, followed by thorough rinsing with deionised water,
rinsing with ethanol and drying with N2 gas. 0.2 mg/mL solutions of carboxylic acid-
terminated (SAM linker) and tri(ethylene glycol) (SAM spacer) alkanethiols were
prepared in DMSO. These solutions were mixed in a 1:4 (linker:spacer) ratio and 1 µL
of this mixture was spotted on to each well of the gold coated plate. The plate was
sealed in a Petri dish with Parafilm and left overnight in the dark to form a mixed SAM.
The plate was then washed with ethanol and dried with N2 gas. Sample wells were then
spotted with 1 µL THAP solution, allowed to dry and then analysed by MALDI-TOF
MS to determine effective SAM formation. After analysis the plates were again washed
with ethanol and dried with N2 gas.
CHAPTER 8
195
Activation and Glycosylation of SAMs on Gold
A 0.1 M EDC/NHS solution was prepared in dry DMF. 1 µL of this solution was
spotted on each well of the gold coated plate. The plate was sealed in a Petri dish and
left for 2 hours before washing with ethanol and drying with N2 gas. 25 mM solutions of
each glycoside to be immobilised were made up in PBS (pH 7.4) and 1 µL of each
solution was spotted on to the desired wells. The plate was sealed in a Petri dish and left
overnight before being washed with ethanol and dried with N2 gas. Sample wells were
then spotted with 1 µL THAP solution, allowed to dry and then analysed by MALDI-
TOF MS to determine the effectiveness of the glycosylations. After analysis the plates
were again washed with ethanol and dried with N2 gas.
GnT-I Reaction on Trimannoside (41)
GnT-I and POMGnT-I reactions were carried out in 50 mM MES buffer containing 2
mM UDP-GlcNAc, 10 mM MnCl and 25% enzyme preparation. 2 µL of this solution
was spotted on each well, the plate placed on a damp paper towel, sealed in a Petri dish
and left overnight at 37°C. The plates were washed with ethanol, acetone and DCM
sequentially before being dried with N2 and spotted with THAP solution and analysed
by MALDI-TOF MS as previously described.
Screening of Mannosides Against Yeast Microsomal Extracts
Yeast microsomes (from P. Pastoris G5115) were prepared by Dr P. Both (The
University of Manchester). A 400 mL cell culture was harvested, by centrifugation (600
g, 20 minutes, 4°C) during the exponential growth phase. Cells were then washed twice
with ice cold buffer (50 mM Tris/HCl, 5 mM MgCl2, 10 mM 2-mercaptoethanol, pH
7.5) and resuspended in 100 mL of the same buffer. Cells were then disrupted in a
French Press (276 MPa) and deoxyribonuclease I (10 µg/mL) was added. Cells were
centrifuged again (600 g, 20 minutes, 4°C) and the supernatant removed for further
centrifugation (18500 g, 50 minutes, 4°C) to yield a microsome rich fraction, which was
flash frozen and stored at -80°C.195
CHAPTER 8
196
Microsome Activity Assay
Mannosides were screened using solutions containing UDP-GlcNAc, UDP-Glucose and
GDP-Mannose separately and with all three glycosyl donors together. For all screens
undertaken the microsomal extract was diluted 50% in to a pH 7.5 buffered solution to
the final concentrations; 25 mM Tris.HCl, 50 µM Swainsonine, 2.5 mM MgCl2, 5 mM
MnCl2, 2 mM per glycosyl donor.
2 µL of this solution was spotted on each well, the plate placed on a damp paper towel,
sealed in a Petri dish and left overnight at 37°C. The plates were washed with ethanol,
acetone and DCM sequentially before being dried with N2 and spotted with THAP
solution and analysed by MALDI-TOF MS as previously described.
8.4.2 Transialidase (TcTs) Reactions
Reactions were carried out on a 1 mL scale or less, facilitating the direct loading of the
reaction mixture on to the size exclusion column. All reactions were carried out in 50
mM phosphate buffer (pH 7.4) with lactosylated GFP concentrations of 0.1 mM. For
singly and doubly lactosylated GFP samples (0.1-0.2 mM lactose), 5 % TcTs solution
was added to the reaction mixture followed by 10 mg of fetuin. For polylactosylated
GFP samples 10 % TcTs solution was added to the reaction mixture followed by 20 mg
fetuin. Reactions were left shaking (250 rpm) at 30°C for 1 hour, centrifuged for 5
minutes at 16100 g and then directly loaded on to a SephadexTM
size exclusion column.
Purifications were performed as described in section 8.2.2.
8.5 Experimental Details for Chapter 6
Unless stated otherwise, all chemicals were of analytical grade and used as received
from Sigma-Aldrich. Fluorescence measurements were carried out using a M200
infinite plate reader (TECAN). Streptavidin coated 96-well plates were purchased from
Thermo scientific and all lectins used were purchased from Vector Laboratories. ITC
measurements were carried out using an ITC-200 microcalorimeter from microcal.
CHAPTER 8
197
8.5.1 Lectin 96-Well Plate Assay
Streptavidin coated 96-well plates (Thermo Scientific) were stored a 4°C until use. Each
well was first washed three times with 200 μL wash buffer (25 mM Tris, 150 mM NaCl,
0.1 % BSA, 0.05 % Tween® -20, pH 7.2) before the addition of 100 μL of the
appropriate biotinylated lectin (20 μg/mL). 100 μL of wash buffer was added to each
control well and the plates incubated for 2 hours with shaking (150 rpm). Each well was
again washed three times with wash buffer (200 μL) and then 100 μL of the appropriate
protein samples were added to each well. Non glycosylated and singly glycosylated
samples were used at 10 μM, whereas multiply glycosylated mutants’ concentrations
were altered so that the concentration of glycoside was 10 μM. Plates were then
incubated for 1 hour in the dark with shaking (150 rpm). After washing wells 10 times
with wash buffer (200 μL) and twice with deionised water (200 μL), 200 μL of
deionised water was added to each well for the fluorescence measurements. The
fluorescence was measured at 508 nm when exited at 395 nm. Blank measurements,
from wells containing lectins but not incubated with GFP samples were subtracted.
8.5.2 ITC Measurements
The calorimeter cell held a volume of 200 μL and had a maximum injection capacity of
35 μL. All lectin and ligand solutions were made using ITC buffer (10 mM HEPES, 154
mM NaCl, 1 mM MnCl2, 1 mM CaCl2, pH 7.5), which had been filtered and degasses.
For all titrations the lectin was placed in the cell and ligand was injected. Typically 30
injections of 1 μL, 1 second in duration, were made with 2 minute intervals. After each
titration the cell was washed three times with SPR buffer and three times with deionised
water. If this was insufficient then the cell was filled with 1 M NaOH (aq) and heated to
65°C for 30 minutes before washing.
References
198
REFERENCES
(1) Suzuki, M.; Ito, Y.; Savage, H. E.; Husimi, Y.; Douglas, K. T. Chem. Lett. 2003,
306-307.
(2) Eylar, E. H. J. Theor. Biol. 1966, 10, 89-113.
(3) Varki, A. Glycobiology 1993, 3, 97-130.
(4) Moremen, K. W.; Tiemeyer, M.; Nairn, A. V. Nature Rev. Mol. Cell Biol. 2012,
13, 448-462.
(5) Gamblin, D. P.; Scanlan, E. M.; Davis, B. G. Chem. Rev. 2009, 109, 131-163.
(6) Taylor, M. E.; Drickamer, K. Introduction to Glycobiology; Oxford university
press: New York, 2006; Vol. 255.
(7) Gagneux, P.; Varki, A. Glycobiology 1999, 9, 747-755.
(8) Dwek, R. A. Chem. Revs. 1996, 96, 683-720.
(9) Singh, R.; Barden, A.; Mori, T.; Beilin, L. Diabetologia 2001, 44, 129-146.
(10) Masterson, W. J.; Raper, J.; Doering, T. L.; Hart, G. W.; Englund, P. T. Cell
1990, 62, 73-80.
(11) Apweiler, R.; Hermjakob, H.; Sharon, N. Biochim. Biophys. Acta. 1999, 1473,
4-8.
(12) Helenius, A.; Aebi, M. Science 2001, 291, 2364-2369.
(13) Gibbs, B. S.; Coward, J. K. Bioorg. Med. Chem. 1999, 7, 441-447.
(14) Schwarz, F.; Aebi, M. Curr. Opin. Struc. Biol. 2011, 21, 576-582.
(15) Lizak, C.; Gerber, S.; Numao, S.; Aebi, M.; Locher, K. P. Nature 2011, 474,
350-355.
(16) Gavel, Y.; Heijne, G. Protein Eng. 1990, 3, 433-442.
(17) Hammond, C.; Braakman, I.; Helenius, A. PNAS 1994, 91, 913-917.
(18) Klausner, D.; Sitia, R. Cell 1990, 62, 611-614.
(19) Kopito, R. R. Cell 1997, 88, 427-430.
(20) Balch, W. E.; McCaffery, J. M.; Plutner, H.; Farquhar, M. G. Cell 1994, 76,
841-852.
(21) Helenius, A. Mol. Biol. Cell 1994, 5, 253-265.
(22) Imperiali, B.; O'Connor, S. E. Curr. Opin. Chem. Biol. 1999, 3, 643-649.
(23) Beintema, J. J. Biosci. Rep. 1986, 6, 709-714.
(24) Bosques, C. J.; Tschampel, S. M.; Woods, R. J.; Imperiali, B. J. Am. Chem. Soc.
2004, 126, 8421-8425.
(25) Jensen, P. H.; Kolarich, D.; Packer, N. H. FEBS J. 2010, 277, 81-94.
(26) Chen, Y.; Tang, Y.; Sheng, Z.; Zhang, Z. Bioinformatis 2008, 101, 1471-1483.
References
199
(27) Harris, R. J.; Ling, V. T. Glycobiology 1993, 3, 219-225.
(28) Steen, P. V. d.; Rudd, P. M.; Dwek, R. A.; Opdenakker, G. Crit. Rev. Anal.
Chem. 1998, 33, 151-208.
(29) Hang, H. C.; Bertozzi, C. R. Bioorgan. Med. Chem. 2005, 13, 5021-5034.
(30) Gill, D. J.; Clausen, H.; Bard, F. Trends Cell. Biol.2011 , 21, 149-158.
(31) Hounsell, E. F.; Davies, M. J.; Renouf, D. V. Glycoconjugate J. 1996, 13, 19-26.
(32) Ismail, M. N.; Stone, E. L.; Panico, M.; Lee, S. H.; Luu, Y.; Ramirez, K.; Ho, S.
B.; Fukuda, M.; Marth, J. D.; Haslam, S. M. Glycobiology, 2011, 21, 82-98.
(33) Kufe, D. W. Nat. Revs. Cancer 2009, 9, 874-885.
(34) Kozarsky, K. F.; Call, S. M.; Dower, S. K.; Krieger, M. Mol. Cell. Biol. 1988, 8,
3357-3363.
(35) Guzman-Aranguez, A.; Argueso, P. Ocul. Surf. 2010, 8, 8-17.
(36) Bao, M.; Booth, J. L.; Elmendorf, B. J.; Canfield, W. M. J. Biol. Chem. 1996,
271, 31437-31445.
(37) Dwek, R. A. Chem. Rev. 1996, 96, 683-720.
(38) Pang, P. C.; Chiu, P. C. N.; Lee, C. L.; Chang, L. Y.; Panico, M.; Morris, H. R.;
Haslam, S. M.; Khoo, K. H.; Clark, G. F.; Yeung, W. S. B. Science 2011, 333,
1761-1764.
(39) Wong, C. H.; Bryan, M. C.; Nyffeler, P. T.; Liu, H.; Chapman, E. Pure Appl.
Chem. 2003, 75, 179-186.
(40) Campos, V. R.; dos Santos, E. A.; Ferreira, V. F.; Montenegro, R. C.; de Souza,
M. C. B. V.; Costa-Lotufo, L. V.; de Moraes, M. O.; Regufe, A. K. P.; Jordao,
A. K.; Pinto, A. C. RSC Advances 2012, 2, 11438-11448.
(41) Henkel, T.; Ciesiolka, T.; Rohr, J.; Zeeck, A. J. Antibiot. 1989, 42, 299.
(42) Nativi, C.; Francesconi, O.; Gabrielli, G.; De Simone, I.; Turchetti, B.; Mello,
T.; Mannelli, L. D. C.; Ghelardini, C.; Buzzini, P.; Roelens, S. Chem-Eur J.
2012, 18, 5064-5072.
(43) Verez-Bencomo, V.; Fernandez-Santana, V.; Hardy, E.; Toledo, M. E.;
Rodriguez, M. C.; Heynngnezz, L.; Rodriguez, A.; Baly, A.; Herrera, L.;
Izquierdo, M. Science 2004, 305, 522-525.
(44) Mandal, M.; Dudkin, V. Y.; Geng, X.; Danishefsky, S. J. Angew. Chem. Int.
Edit. 2004, 43, 2557-2561.
(45) Schofield, L.; Hewitt, M. C.; Evans, K.; Siomos, M. A.; Seeberger, P. H. Nature
2002, 418, 785-789.
(46) Zhu, J.; Wan, Q.; Ragupathi, G.; George, C. M.; Livingston, P. O.; Danishefsky,
S. J. J. Am. Chem. Soc. 2009, 131, 4151-4158.
(47) Beddoe, T.; Paton, A. W.; Le Nours, J. R. M.; Rossjohn, J.; Paton, J. C. Trends
Biochem. Sci. 2010, 35, 411-418.
(48) Spicer, C. D.; Davis, B. G. Chem. Commun. 2013, 49, 2747-2749.
References
200
(49) Pearce, O. M. T.; Fisher, K. D.; Humphries, J.; Seymour, L. W.; Smith, A.;
Davis, B. G. Angew. Chem. Int. Edit. 2005, 44, 1057-1061.
(50) Zhang, H.; Ma, Y.; Sun, X. L. Med. Res. Rev. 2010, 30, 270-289.
(51) Ashwell, G.; Morell, A. G. Adv. Enzymol. 1974, 41, 99-127.
(52) Krantz, M. J.; Holtzman, N. A.; Stowell, C. P.; Lee, Y. C. Biochem. J. 1976, 15,
3963-3969.
(53) Egrie, J. C.; Dwyer, E.; Browne, J. K.; Hitz, A.; Lykos, M. A. Exp. Hemato.
2003, 31, 290-299.
(54) Wadhwa, M. S.; Rice, K. G. J. Drug Target. D. 2003, 11, 255-268.
(55) Garnier, P.; Wang, X. T.; Robinson, M. A.; van Kasteren, S.; Perkins, A. C.;
Frier, M.; Fairbanks, A. J.; Davis, B. G. J. Drug Target. 2010, 18, 794-802.
(56) Robinson, M. A.; Charlton, S. T.; Garnier, P.; Wang, X. T..; Davis, S. S.;
Perkins, A. C.; Frier, M.; Duncan, R.; Savage, T. J.; Wyatt, D. A. P. Natl. Acad.
Sci. USA 2004, 101, 14527-14532.
(57) Minko, T. Adv. Drug Deliver. Rev. 2004, 56, 491-509.
(58) Zhu, Y.; Li, X.; McVie-Wylie, A.; Jiang, C.; Thurberg, B. L.; Raben, N.;
Mattaliano, R. J.; Cheng, S. H. Biochem. J. 2005, 389, 619-628.
(59) Opanasopit, P.; Higuchi, Y.; Kawakami, S.; Yamashita, F.; Nishikawa, M.;
Hashida, M. BBA-Biomembranes 2001, 1511, 134-145.
(60) Van Patten, S. M.; Hughes, H.; Huff, M. R.; Piepenhagen, P. A.; Waire, J.; Qiu,
H.; Ganesa, C.; Reczek, D.; Ward, P. V.; Kutzko, J. P. Glycobiology 2007, 17,
467-478.
(61) Ikehara, Y.; Niwa, T.; Biao, L.; Ikehara, S. K.; Ohashi, N.; Kobayashi, T.;
Shimizu, Y.; Kojima, N.; Nakanishi, H. Cancer Res. 2006, 66, 8740-8748.
(62) Malam, Y.; Loizidou, M.; Seifalian, A. M. Trends Pharmacol. Sci. 2009, 30,
592-599.
(63) Jones, M. N. Adv. Drug Deliver. Rev. 1994, 13, 215-249.
(64) Jain, K.; Kesharwani, P.; Gupta, U.; Jain, N. K. Biomaterials 2012, 33, 4166-
4186.
(65) Seeberger, P. H.; Werz, D. B. Nature 2007, 446, 1046-1051.
(66) Jefferis, R. Nat. Rev. Drug Discov. 2009, 8, 226-234.
(67) Ghazarian, H.; Idoni, B.; Oppenheimer, S. B. Acta Histochem. 2011, 113, 236-
247.
(68) Lis, H.; Sharon, N. Chem. Rev. 1998, 98, 637-674.
(69) Pieters, R. J. Org. Biomol. Chem. 2009, 7, 2013-2025.
(70) Chervenak, M. C.; Toone, E. J. Biochemistry 1995, 34, 5685-5695.
(71) Collins, B. E.; Paulson, J. C. Curr. Opin. Chem. Biol. 2004, 8, 617-625.
References
201
(72) Reichelt, S.; Elsner, C.; Prager, A.; Naumov, S.; Kuballa, J.; Buchmeiser, M. R.
Analyst 2012, 137, 2600-2607.
(73) Polizzotti, B. D.; Maheshwari, R.; Vinkenborg, J.; Kiick, K. L. Macromolecules
2007, 40, 7103-7110.
(74) Kane, R. S. Langmuir 2010, 26, 8636-8640.
(75) Schwefel, D.; Maierhofer, C.; Beck, J. G.; Seeberger, S.; Diederichs, K.; Moller,
H. M.; Welte, W.; Wittmann, V. J. Am. Chem. Soc. 2010, 132, 8704-8719.
(76) Mortell, K. H.; Weatherman, R. V.; Kiessling, L. L. J. Am. Chem. Soc. 1996,
118, 2297-2298.
(77) Liang, R.; Loebach, J.; Horan, N.; Ge, M.; Thompson, C.; Yan, L.; Kahne, D.
PNAS 1997, 94, 10554-10559.
(78) Seeberger, P. H.; Werz, D. B. Nature 2007, 446, 1046-1052.
(79) Ladmiral, V.; Melia, E.; Haddleton, D. M. Eur. Polym. J. 2004, 40, 431-449.
(80) Kainthan, R. K.; Hester, S. R.; Levin, E.; Devine, D. V.; Brooks, D. E.
Biomaterials 2007, 28, 4581-4590.
(81) Parry, A. L.; Clemson, N. A.; Ellis, J.; Bernhard, S. S. R.; Davis, B. G.;
Cameron, N. R. J. Am. Chem. Soc. 2013, 135, 9362-9365.
(82) Imberty, A.; Chabre, Y. M.; Roy, R. Chem-Eur. J. 2008, 14, 7490-7499.
(83) Kitov, P. I.; Mulvey, G. L.; Griener, T. P.; Lipinski, T.; Solomon, D.;
Paszkiewicz, E.; Jacobson, J. M.; Sadowska, J. M.; Suzuki, M.; Yamamura, K.
PNAS 2008, 105, 16837-16842.
(84) Sinha, R.; Kim, G. J.; Nie, S.; Shin, D. M. Mol. Cancer Ther. 2006, 5, 1909-
1917.
(85) Kikkeri, R.; Lepenies, B.; Adibekian, A.; Laurino, P.; Seeberger, P. H. J. Am.
Chem. Soc. 2009, 131, 2110-2112.
(86) Lin, C. C.; Yeh, Y. C.; Yang, C. Y.; Chen, G. F.; Chen, Y. C.; Wu, Y. C.; Chen,
C. C. Chem. Commun. 2003, 2920-2921.
(87) Baenziger, J. U.; Fiete, D. Cell 1980, 22, 611-620.
(88) Stambach, N. S.; Taylor, M. E. Glycobiology 2003, 13, 401-410.
(89) Taylor, M. E. Mammalian Carbohydrate Recognition Systems; Springer: Berlin,
2001, p 105-121.
(90) Galan, M. C.; Dumy, P.; Renaudet, O. Chem. Soc. Rev. 2013, 42, 4599-4612.
(91) Wang, H.; Huang, W.; Orwenyo, J.; Banerjee, A.; Vasta, G. R.; Wang, L. X.
Bioorgan. Med. Chem. 2013, 21, 2037-2044.
(92) Chalker, J. M.; Bernardes, G. J. L.; Davis, B. G. Accounts Chem. Res. 2011, 44,
730-741.
(93) Avery, O. T.; Goebel, W. F. J. Exp. Med. 1929, 50, 533-550.
(94) Kuchler, S.; Graff, M. N.; Gobaille, S.; Vincendon, G.; Roche, A. C.; Delaunoy,
J. P.; Monsigny, M.; Zanetta, J. P. Neurochem. Int. 1994, 24, 43-55.
References
202
(95) Lew, D. B.; Songu-Mize, E.; Pontow, S. E.; Stahl, P. D.; Rattazzi, M. C. J. Clin.
Invest. 1994, 94, 1855.
(96) Felin, M.; Doyennette, Moyne, M. A. S.; Hadj-Sahraoui, Y.; Aubery, M. l.;
Hubert, J.; Seve, A. P. J. Cell. Biochem. 1994, 56, 527-535.
(97) Kitamura, K.; Stockert, E.; Garin-Chesa, P.; Welt, S.; Lloyd, K. O.; Armour, K.
L.; Wallace, T. P.; Harris, W. J.; Carr, F. J.; Old, L. J. PNAS 1994, 91, 12957-
12961.
(98) Kerehgyarto, M.; Fekete, A.; Szurmai, Z.; Kerehgyarto, J.; Takacs, L.; Kurucz,
I.; Guttman, A. Electrophoresis 2013, 34, 2379-2386.
(99) Kent, S. B. H. Chem. Soc. Rev. 2009, 38, 338-351.
(100) Harpaz, Z.; Siman, P.; Kumar, K. S.; Brik, A. ChemBioChem, 2010, 11, 1232-
1235.
(101) Shang, S.; Tan, Z.; Danishefsky, S. J. PNAS 2011, 108, 5986-5989.
(102) Asahina, Y.; Kamitori, S.; Takao, T.; Nishi, N.; Hojo, H. Angew. Chem. Int.
Edit. 2013, 52, 9733-9737.
(103) Wang, P.; Dong, S.; Brailsford, J. A.; Iyer, K.; Townsend, S. D.; Zhang, Q.;
Hendrickson, R. C.; Shieh, J.; Moore, M. A. S.; Danishefsky, S. J. Angew.
Chem. Int. Edit. 2012, 124, 11744-11752.
(104) Kumar, K. S. A.; Spasser, L.; Ohayon, S.; Erlich, L. A.; Brik, A. Bioconjugate
Chem. 2011, 22, 137-143.
(105) Dang, B.; Kubota, T.; Mandal, K.; Bezanilla, F.; Kent, S. B. H. J. Am. Chem.
Soc., 2013, 135, 11911-11919.
(106) Kajihara, Y.; Izumi, M.; Hirano, K.; Murase, T.; Macmillan, D.; Okamoto, R.
Isr. J. Chem. 2011, 51, 917-929.
(107) Dawson, P. E.; Muir, T. W.; Clark-Lewis, I.; Kent, S. B. H. Science 1994, 266,
776-776.
(108) Wong, C. T. T.; Tung, C. L.; Li, X. Mol. BioSyst. 2013, 9, 826-833.
(109) Yan, L. Z.; Dawson, P. E. J. Am. Chem. Soc. 2001, 123, 526-533.
(110) Wan, Q.; Danishefsky, S. J. Angew. Chem. Int. Edit. 2007, 46, 9248-9252.
(111) Whitney, C. G.; Farley, M. M.; Hadler, J.; Harrison, L. H.; Bennett, N. M.;
Lynfield, R.; Reingold, A.; Cieslak, P. R.; Pilishvili, T.; Jackson, D. New Engl.
J. Med. 2003, 348, 1737-1746.
(112) Grayson, E. J.; Bernardes, G. a. J. L.; Chalker, J. M.; Boutureira, O.; Koeppe, J.
R.; Davis, B. G. Angew. Chem. Int. Edit. 2011, 50, 4127-4132.
(113) Seeberger, P. H. Chem. Soc. Rev. 2008, 37, 19-28.
(114) Zhu, X.; Schmidt, R. R. Angew. Chem. Int. Edit. 2009, 48, 1900-1934.
(115) Lee, Y. C.; Stowell, C. P.; Krantz, M. J. Biochemistry 1976, 15, 3956-3963.
(116) McBroom, C. R.; Samanen, C. H.; Goldstein, I. J. Methods Enzymol. 1972, 28,
212-219.
References
203
(117) Yoshida, T.; Lee, Y. C. Carbohyd. Res. 1994, 251, 175-186.
(118) Sletten, E. M.; Bertozzi, C. R. Angew. Chem. Int. Ed. 2009, 48, 6974-6998.
(119) Wang, L.; Brock, A.; Herberich, B.; Schultz, P. G. Science 2001, 292, 498-500.
(120) Liu, H.; Wang, L.; A.Brock; Wong, C.; Schultz, P. G. J. Am. Chem. Soc. 2003,
125, 1702-1703.
(121) Sletten, E. M.; Bertozzi, C. R. Angew. Chem. Int. Ed. 2009, 48, 6974-6998.
(122) Baskin, J. M.; Bertozzi, C. R. QSAR & Comb. Sci. 2007, 26, 1211-1219.
(123) Plass, T.; Milles, S.; Koehler, C.; Szymanski, J.; Mueller, R.; Wiebler, M.;
Schultz, C.; Lemke, E. A. Angew. Chem. Int. Ed. 2012, 51, 4166-4170.
(124) Xie, J.; Schultz, P. G. Nat. Rev. Mol. Cell Bio. 2006, 7, 775-782.
(125) Chalker, J. M.; Bernardes, G. J. L.; Lin, Y. A.; Davis, B. G. Chem. Asian J.
2009, 4, 630-640.
(126) Harpp, D. N.; Gleason, J. G.; Snyder, J. P. J. Am. Chem. Soc. 1968, 90, 4181-
4182.
(127) Bernardes, G. J. L.; Chalker, J. M.; Errey, J. C.; Davis, B. G. J. Am. Chem. Soc.
2008, 130, 5052-5053.
(128) Zhang, Y.; Bhatt, V. S.; Sun, G.; Wang, P. G.; Palmer, A. F. Bioconjug. Chem.
2008, 19, 2221-2230.
(129) Goddard, D. R.; Michaelis, L. J. Biol. Chem. 1935, 112, 361-371.
(130) Macmillan, D.; Bill, R. M.; Sage, K. A.; Fern, D.; Flitsch, S. L. Chem. Biol.
2001, 8, 133-145.
(131) Davis, N. J.; Flitch, S. L. Tetrahedron Lett. 1991, 32, 6793-6796.
(132) Bhat, S.; Vaino, A. R.; Szarek, W. A. Isr. J. Chem. 2000, 40, 335-341.
(133) Shimomura, O.; Johnson, F. H.; Saiga, Y. J. Cell. Compar. Phyl. 1962, 59, 223-
239.
(134) Inouye, S.; Tsuji, F. I. FEBS Lett. 1994, 341, 277-280.
(135) Prasher, D. C.; Eckenrode, V. K.; Ward, W. W.; Prendergast, F. G.; Cormier, M.
J. Gene 1992, 111, 229-233.
(136) Ormö, M.; Cubitt, A. B.; Kallio, K.; Gross, L. A.; Tsien, R. Y.; Remington, S. J.
Science 1996, 273, 1392-1395.
(137) Yang, F.; Moss, L. G.; Phillips, G. N. Nat. Biotech. 1996, 14, 1246-1251.
(138) Wang, S.; Hazelrigg, T. Nature 1994, 369, 400-403.
(139) Zimmer, M. Chem. Rev. 2002, 102, 759-782.
(140) Koike, M.; Yutoku, Y.; Koike, A. FEBS Open Bio 2013, 3, 46-50.
(141) Miura, H.; Inoko, H.; Inoue, I.; Tanaka, M.; Sato, M.; Ohtsuka, M. Anal.
Biochem. 2011, 416, 237-239.
References
204
(142) Manicassamy, B.; Manicassamy, S.; Belicha-Villanueva, A.; Pisanelli, G.;
Pulendran, B.; Garca-Sastre, A. PNAS 2010, 107, 11531-11536.
(143) Weimer, R. M.; Hill, T. C.; Hamilton, A. M.; Zito, K. Cold Spring Harbor P.
[online] 2012, 7, pdb. prot070029, (accessed March 5, 2013).
(144) Siemering, K. R.; Golbik, R.; Sever, R.; Haseloff, J. Curr. Biol. 1996, 6, 1653-
1663.
(145) Yang, T.-T.; Cheng, L.; Kain, S. R. Nucleic Acids Res. 1996, 24, 4592-4593.
(146) Crameri, A.; Whitehorn, E. A.; Tate, E.; Stemmer, W. P.; Kitts, P. A. Nat.
Biotechnol. 1996, 14, 315-319.
(147) Pédelacq, J. D.; Cabantous, S.; Tran, T.; Terwilliger, T. C.; Waldo, G. S. Nat.
biotechnol. 2005, 24, 79-88.
(148) Day, R. N.; Davidson, M. W. Chem. Soc. Rev. 2009, 38, 2887-2921.
(149) Hanson, G. T.; Aggeler, R.; Oglesbee, D.; Cannon, M.; Capaldi, R. A.; Tsien, R.
Y.; Remington, S. J. J. Biol. Chem. 2004, 279, 13044-13053.
(150) Spergel, D. J.; Kruth, U.; Shimshek, D. R.; Sprengel, R.; Seeburg, P. H. Prog.
Neurobiol. 2001, 63, 673-686.
(151) Day, R. N.; Schaufele, F. J. Biomed. Opt. 2008, 13, 1-6.
(152) Suzuki, M.; Ito, Y.; Savage, H. E.; Husimi, Y.; Douglas, K. T. Chem. Lett. 2003,
32, 306-398.
(153) Baffour-Awuah, N. Y.; Fedeles, F.; Zimmer, M. Chem. Phys. 2005, 310, 25-31.
(154) Sohn, E. J.; Rojas-Pierce, M.; Pan, S.; Carter, C.; Serrano-Mislata, A.; Madueao,
F.; Rojo, E.; Surpin, M.; Raikhel, N. V. PNAS 2007, 104, 18801-18806.
(155) Bulbarelli, A.; Sprocati, T.; Barberi, M.; Pedrazzini, E.; Borgese, N. J. Cell Sci.
2002, 115, 1689-1702.
(156) Warner, J. B.; Thalhauser, C.; Tao, K.; Sahagian, G. G. J. Biol. Chem. 2002,
277, 41897-41905.
(157) Paris, N.; Saint-Jean, B.; Faraco, M.; Krzeszowiec, W.; Dalessandro, G.;
Neuhaus, J.-M.; Di Sansebastiano, G. P. Plant Cell Rep. 1998, 29, 79-86.
(158) Richardson, J. P.; Chan, C.-H.; Blanc, J.; Saadi, M.; Macmillan, D. Org. Biomol.
Chem. 2010, 8, 1351-1360.
(159) Ding, S.; Xu, Q.; Ali, M. K.; Baker, J. O.; Bayer, E. A.; Barak, Y.; Lamed, R.;
Sugiyama, J.; Rumbles, G.; Himmel, M. E. Biotechniques 2006, 41, 435.
(160) Snapp, E. L. Trends Cell Biol. 2009, 19, 649-655.
(161) Li, Y.; Agrawal, A.; Sakon, J.; Beitle, R. R. J. Chromatogr. A 2001, 909, 183-
190.
(162) Zhou, X.; Shi, Q.; Xing, X.; Sun, Y. Chinese J. Chem. Eng. 2006, 14, 229-234.
(163) Yakhnin, A. V.; Vinokurov, L. M.; Surin, A. K.; Alakhov, Y. B. Protein Expres.
Purif. 1998, 14, 382-386.
References
205
(164) Imai, Y.; Matsushima, Y.; Sugimura, T.; Terada, M. Nucleic Acids Res. 1991,
19, 2785.
(165) Stemmer, W. P. PNAS 1994, 91, 10747-10751.
(166) Crameri, A.; Raillard, S.-A.; Bermudez, E.; Stemmer, W. P. C. Nature 1998,
391, 288-291.
(167) Stemmer, W. P. C. Nature 1994, 370, 389-391.
(168) Bejugam, M.; Flitsch, S. L. Org. Lett. 2004, 6, 4001-4004.
(169) Blixt, O.; Han, S.; Liao, L.; Zeng, Y.; Hoffmann, J.; Futakawa, S.; Paulson, J. C.
J. Am. Chem. Soc. 2008, 130, 6680-6681.
(170) Hartmann, M.; Horst, A. K.; Klemm, P.; Lindhorst, T. K. Chem. Commun. 2009,
46, 330-332.
(171) Huflejt, M. E.; Vuskovic, M.; Vasiliu, D.; Xu, H.; Obukhova, P.; Shilova, N.;
Tuzikov, A.; Galanina, O.; Arun, B.; Lu, K.; Bovin, N. Mol. Immunol. 2009, 46,
3037-3049.
(172) Zhi, Z.-L.; Laurent, N.; Powell, A. K.; Karamanska , R.; Fais , M.; Voglmeir, J.;
Wright, A.; Blackburn, J. M.; Crocker, P. R.; Russell, D. A.; Flitsch, S.; Field ,
R. A.; Turnbull, J. E. ChemBioChem 2008, 9, 1568-1575.
(173) Chernyak, A. Y.; Sharma, G. V. M.; Kononov, L. O.; Krishna, P. R.; Levinsky,
A. B.; Kochetkov, N. K.; Rama Rao, A. V. Carbohyd. Res. 1992, 223, 303-309.
(174) Park, S.; Shin, I. Org. Lett. 2007, 9, 1675-1678.
(175) Cheshev, P. E.; Khatuntseva, E. A.; Tsvetkov, Y. E.; Shashkov, A. S.; Nifantiev,
N. E. Russ. J. Bioorg. Chem. 2004, 30, 60-70.
(176) Orlandi, S.; Annunziata, R.; Benaglia, M.; Cozzi, F.; Manzoni, L. Tetrahedron
2005, 61, 10048-10060.
(177) Šardzík, R.; Noble, G. T.; Weissenborn, M. J.; Martin, A.; Webb, S. J.; Flitsch,
S. L. Beilstein J. Org. Chem. 2010, 6, 699-703.
(178) Kaur, K.; Hindsgaul, O. Glycoconjugate J. 1991, 8, 90-94.
(179) Kaur, K. J.; Alton, G.; Hindsgaul, O. Carbohyd. Res. 1991, 210, 145-153.
(180) Vella, G. J.; Paulsen, H.; Schachter, H. Can. J. Biochem. Cell B. 1984, 62, 409-
417.
(181) Davis, N. J.; Flitsch, S. L. Tetrahedron Lett. 1991, 32, 6793-6796.
(182) Goddard, D. R.; Michaelis, L. J. Biol. Chem. 1935, 112, 361-371.
(183) Nielsen, M. L.; Vermeulen, M.; Bonaldi, T.; Cox, J. r.; Moroder, L.; Mann, M.
Nat. Methods 2008, 5, 459-460.
(184) Wong, S. Y.; Guile, G. R.; Dwek, R. A.; Arsequell, G. Biochem. J. 1994, 300,
843.
(185) Zaia, J.; Annan, R. S.; Biemann, K. Rapid Commun. Mass Sp. 1992, 6, 32-36.
(186) Eggler, A. L.; Liu, G.; Pezzuto, J. M.; van Breemen, R. B.; Mesecar, A. D. P
Natl. Acad. Sci. USA 2005, 102, 10070-10075.
References
206
(187) Getz, E. B.; Xiao, M.; Chakrabarty, T.; Cooke, R.; Selvin, P. R. Anal. Biochem.
1999, 273, 73-80.
(188) Shafer, D. E.; Inman, J. K.; Lees, A. Anal. Biochem. 2000, 282, 161.
(189) Fujiyama, K.; Ido, Y.; Misaki, R.; Moran, D. G.; Yanagihara, I.; Honda, T.;
Nishimura, S.-I.; Yoshida, T.; Seki, T. J. Biosci. Bioeng. 2001, 92, 569-574.
(190) Voglmeir, J.; Kaloo, S.; Laurent, N.; Meloni, M. M.; Bohlmann, L.; Wilson, I.
B. H.; Flitsch, S. L. Biochem. J. 2011, 436, 447-455.
(191) Sardzik, R.; Sharma, R.; Kaloo, S.; Voglmeir, J.; Crocker, P. R.; Flitsch, S. L.
Chem. Commun. 2011, 47, 5425-5427.
(192) Zhang, Y. N.; Le, X. C.; Dovichi, N. J.; Compston, C. A.; Palcic, M. M.;
Diedrich, P.; Hindsgaul, O. Anal. Biochem. 1995, 227, 368-376.
(193) Trimble, R. B.; Byrd, J. C.; Maley, F. J. Biol. Chem. 1980, 255, 11892-11895.
(194) Wilson, I. B.; Webberley, M. C.; Revers, L.; Flitsch, S. L. Biochem. J. 1995,
310, 909.
(195) Wilson, I. B.; Taylor, J. P.; Webberley, M. C.; Turner, N. J.; Flitsch, S. L.
Biochem. J. 1993, 295, 195.
(196) Crocker, P. R.; Paulson, J. C.; Varki, A. Nat. Rev. Immunol. 2007, 7, 255-66.
(197) Sperandio, M.; Gleissner, C. A.; Ley, K. Immunol. Rev. 2009, 230, 97-113.
(198) Amaya, M. F.; Watts, A. G.; Damager, I.; Wehenkel, A.; Nguyen, T.;
Buschiazzo, A.; Paris, G.; Frasch, A. C.; Withers, S. G.; Alzari, P. M. Structure
2004, 12, 775-84.
(199) Green, N. M. Biochem. J. 1963, 89, 585-591.
(200) Ofek, I.; Courtney, H. S.; Schifferli, D. M.; Beachey, E. H. J. Clin. Microbiol.
1986, 24, 512-516.
(201) Mandal, D. K.; Kishore, N.; Brewer, C. F. Biochemistry 1994, 33, 1149-1156.
(202) Kizhakkedathu, J. N.; Creagh, A. L.; Shenoi, R. A.; Rossi, N. A. A.; Brooks, D.
E.; Chan, T.; Lam, J.; Dandepally, S. R.; Haynes, C. A. Biomacromolecules
2010, 11, 2567-2575.
(203) Dam, T. K.; Roy, R.; Page, D.; Brewer, C. F. Biochemistry 2002, 41, 1359-1363.
(204) Hester, G.; Wright, C. S. J. Mol. Biol. 1996, 262, 516-531.
(205) Smith, E. A.; Thomas, W. D.; Kiessling, L. L.; Corn, R. M. J. Am. Chem. Soc.
2003, 125, 6140-6148.
(206) Arockia Jeyaprakash, A.; Jayashree, G.; Mahanta, S. K.; Swaminathan, C. P.;
Sekar, K.; Surolia, A.; Vijayan, M. J. Mol. Biol. 2005, 347, 181-188.
(207) Bourne, Y.; Astoul, C.; Zamboni, V. r.; Peumans, W.; Menu-Bouaouiche, L.;
Van Damme, E.; Barre, A.; Rouge, P. Biochem. J. 2002, 364, 173-180.
(208) Dalmau, S. R.; Freitas, C. S. Braz. J. Med. Biol. Res. 1989, 22, 601.
(209) Freyer, M. W.; Lewis, E. A. Method. Cell Biol. 2008, 84, 79-113.
References
207
(210) Orlandi, S.; Annunziata, R.; Benaglia, M.; Cozzi, F.; Manzoni, L. Tetrahedron
2005, 61, 10048-10060.
(211) Denoyelle, S. v.; Polidori, A.; Brunelle, M. l.; Vuillaume, P. Y.; Laurent, S.;
Eiazhary, Y.; Pucci, B. New journal of chemistry 2006, 30, 629-646.
(212) Susaki, H.; Suzuki, K.; Ikeda, M.; Yamada, H.; Watanabe, H. K. Chem. Pharm.
Bull. 1994, 42, 2090-2096.
(213) King, R. R.; Cooper, F. P.; Bishop, C. T. Carbohyd. Res. 1977, 55, 83-93.
(214) Micheel, F.; Micheel, H. Ber. Dtsch. Chem. Ges. 1930, 63, 386-393.
Appendices
208
APENDICIES
Appendix 1: DNA Sequences of GFPuv_WT and GFPuv_C48A_I229C
Appendices
209
Appendix 2: The DNA Sequences of sGFPuv_C48A and Shuffle 1
Appendices
210
Appendix 3: Screen Capture of a Typical Stepwise IMAC GFPuv_WT Purification
Blue = UV 280 nm, Brown = conductance, Green = % of elution buffer.
Appendices
211
Appendix 4: The Fluorescence Spectra of GFPuv Mutants
Appendices
212
Appendix 5: Screen Capture of a Typical Polymannoside Purification
Blue = UV 280 nm. Peak 1 = fraction collected containing trimannoside (43). Peak 2 = fraction collected
containing teteramannosides (48 and 49). Peak 3 = fraction collected containing pentamannoside (50).
Peak 1 Peak 2 Peak 3
Appendices
213
Appendix 6: HSQC-TOCSY of Trimannoside (43)
Appendices
214
Appendix 7: HMBC of Trimannoside (43)