UNDERSTANDING AND MANIPULATING GANGLIOSIDE BIOSYNTHESIS A DISSERTATION …qp110yy2867/Final... ·...

UNDERSTANDING AND MANIPULATING GANGLIOSIDE BIOSYNTHESIS

A DISSERTATION

SUBMITTED TO THE DEPARTMENT OF CHEMISTRY

AND THE COMMITTEE ON GRADUATE STUDIES

OF STANFORD UNIVERSITY

IN PARTIAL FULFILLMENT OF THE REQUIREMENTS

FOR THE DEGREE OF

DOCTOR OF PHILOSOPHY

Chad Michael Whitman May 2010

http://creativecommons.org/licenses/by-nc/3.0/us/

This dissertation is online at: http://purl.stanford.edu/qp110yy2867

© 2010 by Chad Michael Whitman. All Rights Reserved.

Re-distributed by Stanford University under license with the author.

This work is licensed under a Creative Commons Attribution-Noncommercial 3.0 United States License.

ii



http://purl.stanford.edu/qp110yy2867

I certify that I have read this dissertation and that, in my opinion, it is fully adequatein scope and quality as a dissertation for the degree of Doctor of Philosophy.

Jennifer Kohler, Primary Adviser


Justin Du Bois, Co-Adviser


Chaitan Khosla

Approved for the Stanford University Committee on Graduate Studies.

Patricia J. Gumport, Vice Provost Graduate Education

This signature page was generated electronically upon submission of this dissertation in electronic format. An original signed hard copy of the signature page is on file inUniversity Archives.

iii

iv

Abstract

Understanding and Manipulating Ganglioside Biosynthesis

By

Chad Michael Whitman

Doctor of Philosophy in Chemistry

Stanford University

The mammalian cell surface is comprised of a heterogeneous mixture of proteins

and lipids decorated by carbohydrates. These glycoproteins and glycolipids are known as

glycoconjugates. Among the many types of glycolipids imbedded within the plasma

membrane resides a class of negatively charged species known as gangliosides.

Characterized by the presence of sialic acid residues, gangliosides are responsible for

regulating the activity of many cell surface proteins and serve as recognition targets for

cell-cell communication and pathogen invasion. In Chapter 1, I introduce the

biosynthetic pathway for mammalian ganglioside synthesis. Next, I describe several

recognized biological roles of gangliosides. Finally, I present recent biochemical

methods used for elucidating biological function at the molecular level.

Gangliosides are synthesized in the endoplasmic reticulum and Golgi by

numerous membrane-bound glycosyltransferases. These glycosyltransferases catalyze

the addition of monosaccharides in a sequential fashion, resulting in a diverse assortment

of gangliosides. Ganglioside biosynthetic enzymes have been shown to associate with

one another, forming biosynthetic clusters. Previous work demonstrated that associations

among these enzymes are controlled by their N-terminal regions, which include a single-

pass transmembrane domain (TM). The formation of homo-oligomeric complexes via

TM domain interactions has also been implicated in regulating glycosyltransferase

function and localization. In Chapter 2, I characterize the interactions among TM

domains of five ganglioside glycosyltransferases. Using an assay that measures TM

domain association in SDS micelles, I discovered that three of the TM domains homo-

oligomerize, including one that forms trimers, pentamers, and higher-order oligomers.

v

To investigate the biological importance of these associations, I employed fluorescence

microscopy and western blot analysis to detect enzyme homo-oligomerization occurring

in transiently transfected mammalian cells.

Gangliosides play critical roles in the regulation of cell signaling and in pathogen-

cell recognition, but the molecular details of many of these processes are poorly

understood. Challenges to biochemical characterization are posed by the transient nature

of ganglioside-mediated interactions and the difficulty in isolating ganglioside-protein

complexes. To surmount these obstacles, I investigate the application of metabolic

oligosaccharide engineering techniques to synthesize photocrosslinking gangliosides in

mammalian cells, as described in Chapter 3. I observed that the photoactivatable GM1

ganglioside analog is recognized by cholera toxin, a well-characterized ganglioside

interactions partner. After demonstration that this complex could be covalently captured

in cells, I have begun preliminary experiments to ascertain the effects complex formation

on the retrograde trafficking mechanisms required for cholera intoxication.

The ability to introduce small structural changes into cell surface glycans through

metabolic oligosaccharide engineering methods has aided in the study of many

carbohydrate-mediated interactions. These techniques have also proved effective at

installing bio-orthogonal reactive groups to visualize glycan structures in both cells and

animals. While previous studies have illustrated the many biological roles of

glycoconjugates in mammalian cells, these studies focused primarily on global levels of

unnatural glycan incorporation and not specific glycoconjugates structures. In Chapter

4, I investigate the substrate selectivity associated with sialic acid engineering of

gangliosides in mammalian cells. I observed substantial differences in unnatural sialic

acid incorporation between cell lines derived from different mammalian species. These

differences may reflect naturally occurring dissimilarities in the ability of different

species to incorporate variant sialic acids into their glycoconjugates.

vi

Acknowledgements

I would like to thank my advisor, Jennifer Kohler, for serving as my mentor for

the past five years. Between our time at Stanford and UT Southwestern, I’ve learned a

great deal about becoming a better scientific researcher and communicator. Her guidance

has allowed me to explore many different scientific techniques in both chemistry and

biology. She has also provided me with the opportunity to mentor fellow labmates and

rotation students. I am forever grateful for the opportunity that she gave me and I only

wish her the best for many years to come.

Being a part of two universities for my graduate career, I have the awesome

opportunity to work with a large number of gifted scientists. Particularly, I would like to

thank Michelle Bond. Through our 4+ years together, we have formed an incredible

bond that has allowed us to work together and develop a deep friendship. I have an

immense amount of respect for her abilities and I hope that we can continue our

friendship as we both move to the opposite sides of the country.

While at Stanford, I had the privilege of working with a talented group of

graduate students and postdocs. Ethan Greenblatt is a great friend who continues to

update me with all of his current experiments, both public and secret, that he is pursuing.

Peter Lee was always willing to help me with my microscopy experiments and was

always the source of great late night conversations. Danielle Dube, our lab’s first

postdoc, was very generous helping me with my presentation skills. Yoshi Tanaka was a

great synthetic chemist who taught me a lot about science and Japanese culture. Meg

Desko started the group with me at Stanford and was a great person to grab coffee with

and chat for hours.

The current members of the group at UT Southwestern have also been equally as

talented as my previous ones. Nam Pham, Randy Parker, and Peter Vu (the first wave of

graduate students in the group) have been a privilege to work with to pass on my

scientific knowledge. I look forward to hearing about their endeavors for years to come

and excited for the future of lab being in their hands. Seokho Yu, Fan Yang, and Bin Li

have bestowed upon me many years of knowledge and experience that have helped me

develop my technical skills. While we have only overlapped for a few months, Rubina

vii

Tuladhar has always been a great source of endless conversation and is a very dedicated

person who will do great things in her upcoming graduate career.

There are several people at Iowa State University who influenced my path to

pursuing my doctorate degree. Prof. Joe Burnett was my instructor for several analytical

chemistry courses and encouraged me to not give up on chemistry. He has also been a

great friend who came and visited me twice while in California. Prof. Victor Lin served

as my research mentor for my senior year and if not for his advisement, I never would

have considered attending graduate school. While he has become a very busy individual,

he still takes time out of his busy schedule to catch up with me and discuss my current

and future endeavours. Brian Trewyn, a graduate student in Victor’s lab, taught me how

to be an effective researcher and how to keep one’s sanity while in graduate school. As

for my classmates, Sassan Sheikholeslami and Ali Mostrom, I’ll never forget our

countless hours spent studying and goofing off at the Cyclone Truck Stop. I’d also like

to thank all of the members of Triangle Fraternity during my undergraduate experience

that helped me develop personally and professionally.

I would also like to thank my parents, Rick and Linda Whitman, for their

incredible support over the years, providing me with countless opportunities to grow and

learn. While I know it was hard for their only son to “leave the nest” and go out West,

I’ve made it a commitment to keep in touch and keep them updated with all my

adventures. Finally, I would like to thank my fiancé Laura McNabb. She has made

many sacrifices for me to complete my degree and I can’t describe into words what she

has meant to me. I’m excited to soon being leaving Texas and returning back to

California with her to begin a new career and start our family.

viii

Table of Contents

Chapter 1 – The biosynthesis and biological roles of gangliosides ..............................1

Introduction.................................................................................................................1 Biosynthesis of gangliosides........................................................................................3 Complexes of glycosyltransferases synthesize gangliosides .........................................4 Gangliosides mediate interactions on the cell surface...................................................6 The role of NeuGc expression in gangliosides .............................................................7 Metabolic oligosaccharide engineering of gangliosides as a potential tool for studying gangliosides.................................................................................................................9 Photocapturing glycosphingolipid interactions in cells...............................................11 Conclusions ...............................................................................................................13 References.................................................................................................................14

Chapter 2 – The biosynthesis and biological roles of gangliosides ............................22

Introduction...............................................................................................................22 Results.......................................................................................................................24

In vitro assay to investigate TM interactions ..........................................................24 Investigating TM interactions in vitro ....................................................................25 Functional analysis of ganglioside glycosyltransferases .........................................29

Discussion .................................................................................................................34 The TM domains of ganglioside glycosyltransferases fail to demonstrate the formation of hetero-oligomeric species with the SN-TM assay ..............................34 The TM domains of several ganglioside glycosyltransferases mediate formation of homo-oligomers.....................................................................................................35 SialT1 and SialT2 enzymes are physically associated in cells.................................37 SialT2 demonstrates the formation of higher-order complexes in cells ...................37

Acknowledgements ...................................................................................................38 Methods ....................................................................................................................39 References.................................................................................................................47

ix

Chapter 3 – Covalently capturing the ganglioside GM1-cholera toxin complex with crosslinking..................................................................................................................52


HPTLC analysis of gangliosides produced by Ac4ManNDAz-treated Jurkat cells ..53 Mass spectrometry analysis of gangliosides produced by Ac4ManNDAz-treated Jurkat cells.............................................................................................................57 Immunofluorescence microscopy analysis of CTxB trafficking in Ac44ManNDAz-treated Jurkat cells exposed to UV light .................................................................58

Discussion .................................................................................................................61 Jurkat cells can synthesize gangliosides with photoreactive chemical groups .........61 Cholera toxin subunit B recognizes GM1-SiaDAz and can be efficiently photocrosslinked....................................................................................................62 The formation of a covalent GM1-CTxB complex does not appear to affect trafficking from the plasma membrane to the TGN ................................................63


Chapter 4 – Exploring the incorporation of sialic acid analogs into gangliosides of mammalian cells ..........................................................................................................75


Incorporation of variant sialic acids into gangliosides ............................................77 Analysis of engineered cell surface sialylation with BJAB and CHO cell lines cultured with ManNAc analogs..............................................................................81

Discussion .................................................................................................................84 Incorporation of modified sialic acids into the GM3 ganglioside of BJAB cells .....84 Incorporation of modified sialic acids into the GM3 ganglioside of CHO cells.......86 The ST3Gal5 enzyme of BJAB cells shows a reduced capacity for engineering NeuGc-containing gangliosides .............................................................................86 Future directions....................................................................................................87


Appendix.................................................................................................................... 105

1H NMR Data.......................................................................................................... 106 13H NMR Data......................................................................................................... 110 ESI-MS Data ........................................................................................................... 114 HPLC Data.............................................................................................................. 118 MALDI-TOF-MS Data............................................................................................ 122

x

List of Illustrations

Chapter 1 – The biosynthesis and biological roles of gangliosides ..............................1

Figure 1. Structure of a typical glycosphingolipid........................................................1 Figure 2. Chemical structure of GM1 ganglioside........................................................2 Figure 3. Biosynthesis of gangliosides .........................................................................4 Figure 4. Ganglioside glycosyltransferases are type II transmembrane proteins............5 Figure 5. Common sialic acid structures ......................................................................7 Figure 6. Examples of unnatural sialic acid analogs incorporated onto cellular glycoconjugates .........................................................................................................10 Figure 7. Common photoreactive groups used in biology...........................................12


Figure 1. Reported associations among glycosyltransferases responsible for ganglioside biosynthesis ............................................................................................23 Figure 2. N-terminal sequences of ganglioside glycosyltransferases implicated in their association.................................................................................................................25 Figure 3. SDS-PAGE analysis of SN-TM homo- and hetero-oligomerization.............26 Figure 4. SDS-PAGE analysis of SN-GalT1 TM WT and mutated fusion proteins.....27 Figure 5. SDS-PAGE analysis of SN-SialT2 TM WT and mutated fusion proteins ....28 Figure 6. KDEL recruitment assay.............................................................................30 Figure 7. Cellular association of ganglioside glycosyltransferases visualized through the KDEL recruitment assay ......................................................................................32 Figure 8. Full length SialT2 enzyme shows homo-oligomerization ............................34

Chapter 3 – Covalently capture the ganglioside GM1-cholera toxin complex with crosslinking..................................................................................................................52

Figure 1. Acetylated analogs used to investigate metabolic oligosaccharide engineering of gangliosides...........................................................................................................54 Figure 2. HPTLC analysis of gangliosides from cultured Jurkat cells.........................56 Figure 3. MALDI-TOF-MS analysis of Jurkat cells cultured with Ac4ManNDAz ......58 Figure 4. Immunofluorescene analysis of Cholera toxin subunit B trafficking in supplemented Jurkat cells ..........................................................................................60

Chapter 4 – Exploring the incorporation of sialic acid analogs into gangliosides of mammalian cells ..........................................................................................................75

Figure 1. Metabolic oligosaccharide engineering of UDP-GlcNAc 2-epimerase deficient cells occurs through the introduction of ManNAc and sialic acid analogs ....78 Figure 2. Panel of ManNAc analogs used to ganglioside incorporation experiments ..79 Figure 3. HPTLC analysis of gangliosides produced by BJAB and CHO cells cultured with ManNAc analogs ...............................................................................................80 Figure 4. Flow cytometry analysis of cells surface display of modified sialic acids ....83

xi

List of Tables


Table 1. Table of primers used to generate SN-TM plasmids .....................................40 Table 2. Table of primers used to generate SN-GalT1 TM mutants............................41 Table 3. Table of primers used to generate SN-SialT2 TM mutants ...........................41 Table 4. Table of primers used to generate KDEL recruitment assay plasmids ...........45

* Several sections in this Chapter were adapted from Whitman, C.M.; Bond, M.R.; Kohler, J.J. In Comprehensive Natural Products II Chemistry and Biology; Mander, L., Lui, H.-W., Eds.; Elsevier: Oxford, 2010; Volume 6, pp 175-224

1

Chapter 1 – The biosynthesis and biological roles of gangliosides*

Introduction

Glycosphingolipids are a heterogeneous class of molecules that decorate the

cellular surface of eukaryotic cells. These amphiphatic structures are anchored into the

membrane by their long hydrocarbon chains and present a variety of carbohydrate

structures to the extracellular environment (Figure 1). Various lipid scaffolds and

monosaccharides combine to form an immense diversity of glycosphingolipids; over 500

different structures have been characterized.1 Glycosphingolipid interactions play critical

roles in all aspects of cell surface recognition. As components of the plasma membrane,

glycosphingolipids utilize their glycan components to modulate the functions of many

cell surface proteins.2 Glycosphingolipids also serve as recognition targets in many cell-

cell interactions,3 Unfortunately, a variety of pathogens, such as cholera toxin and tetanus

toxin, exploit glycosphingolipids to mediate selective binding to their hosts for cell

invasion.4,5

Figure 1. Structure of a typical glycosphingolipid. Glycosphingolipids are amphiphatic molecules that are anchored into the plasma membrane by a hydrophobic ceramide, composed of a sphingosine (blue) and fatty acid (yellow). The ceramide scaffold is decorated with wide variety of glycans, initiating with either glucose or galactose (green).

Glycosphingolipids are divided into seven structural classes based upon their

glycan linkages to ceramide. While the majority of these structures are classified as

neutral glycosphingolipids, the ganglioside class comprises negatively charged species at

2

physiological pH due to the inclusion of sialic acids. While the term “ganglioside” is

often used informally to refer to any sialylated glycosphingolipids, the official

ganglioside class is derived from their common structure of Galβ1,4-Glcα-ceramide, also

known as lactosylceramide (LacCer). Figure 2 shows the chemical structure of the GM1

ganglioside. As detailed below, the biosynthesis of gangliosides starting from LacCer is

quite diverse and will vary by cell type depending on the transcriptional regulation of the

glycosyltransferases responsible for their synthesis. Changes in amount and variety of

gangliosides are observed in many eukaryotic organisms during embryonic brain

development6 and in many types of cancers.7 These changes are critical for facilitating

many ganglioside-based interactions that are occur in these settings. While changes in

ganglioside content are well-documented, the functional implications of these

fluctuations remains unclear.

Figure 2. Chemical structure of GM1 ganglioside. Gangliosides are a class of molecules that share the common structure of Galβ1,4-Glcα-ceramide, known as lactosylceramide (LacCer). Gangliosides contain at least one sialic acid residue in their structure (highlighted in red). Shown above is the chemical structure of a GM1 ganglioside.

Discovery of the genes responsible for ganglioside biosynthesis occurred during

the 1980s and 1990s and facilitated investigations into the many biological roles of these

molecules. With the dawn of chemical biology, chemist and biologists have worked

together to develop an assortment of tools to explore the molecular function and

specificity of gangliosides. For example, the synthesis of photoreactive glycolipids has

provided molecular details for ganglioside binding to tetanus toxin8 and two receptor

tyrosine kinases.9,10 While these studies have yielded information about ganglioside-

mediated interactions, there are still many unanswered questions such as determining

why eukaryotic cells modify their ganglioside composition during tumorigenesis.

3

Nonetheless, the continual advance of chemical and biological techniques will only

further improve our capability to understand the many biological roles of gangliosides.

Biosynthesis of gangliosides

The biosynthesis of gangliosides occurs through the step-wise addition of

nucleotide-activated monosaccharides by glycosyltransferases (Figure 3). Ganglioside

synthesis begins in the endoplasmic reticulum (ER) where ceramide is produced and

displayed on the cytosolic face of the ER membrane. Ceramide is then transported to the

Golgi membrane where it is initially glucosylated (GlcCer) on the cytosolic face to

generate GlcCer. Afterwards, GlcCer is translocated by unidentified mechanisms to the

luminal side of the Golgi for further modification. The attachment of galactose (GalT1)

onto GlcCer generates lactosylceramide (LacCer), the basic building block for the

ganglioside class of glycosphingolipids. Upon reaching the cis-Golgi, LacCer is initially

sialylated (SialT1) to generate GM3. These species can be further modified with

additional sialic acid (SialT2, SialT3) to generate GD3 and GT3. Transfer of these

species towards the trans-Golgi network (TGN) leads to the synthesis of more complex

gangliosides by sequential additions of N-acetylgalactosamine (GalNAcT), galactose

(GalT2), and more sialic acids (SialT4, SialT5). The exact cellular composition of

gangliosides can depend on a large range of factors, including cell type, enzyme levels,

post-translational processing of glycosyltransferases, nucleotide-sugar levels, self-

inhibition by ganglioside products, and pH.11

4

Figure 3. Biosynthesis of gangliosides. Gangliosides are synthesized in step-wise fashion through the addition of monosaccharides (see table insert) by glycosyltransferases (shown in red). Cartoon structures represent the chemical structure of individual gangliosides. Names of ganglioside species are listed below structures. Series classifications of gangliosides families are denoted at the bottom.

Complexes of glycosyltransferaes synthesize gangliosides

All ganglioside glycosyltransferases are type II transmembrane proteins,

consisting of a N-terminal cytoplasmic tail (CT), single-pass transmembrane domain

(TMD), stem region, and a C-terminal catalytic domain (Figure 4). These enzymes are

localized within the various compartments of the secretory pathway, primarily through

their N-terminal region (CT, TMD, and stem).12 Two models describe the retention of

these enzymes to the secretory pathway. The first model, the lipid bilayer thickness or

sorting model,13,14 proposes that glycosyltransferases are retained within the Golgi due to

the smaller size of the TMD compared to plasma membrane proteins. It is known that the

plasma membrane is thicker than Golgi membranes due to the increased presence of

cholesterol, which has been observed to increase lipid bilayer thickness.15 By extending

the length of the TMD domain of a Golgi-resident sialyltransferase (ST6Gal1) from 17 to

5

23 amino acids, Munro showed that the localization was redirected to the plasma

membrane.16 An inverse transposition could be used to relocate plasma membrane

proteins to the Golgi.17 Despite these initial reports, later work showed that the targeting

of ST6Gal1 was not limited to its TMD and could be directed by lumenal sequences

through enzyme homo-oligomerization,18,19 suggesting that the lipid bilayer thickness

model alone could not completely describe Golgi-localization of glycosyltransferases.

Figure 4. Ganglioside glycosyltransferases are type II transmembrane proteins. Type II transmembrane proteins are comprised of an N-terminal cytoplasmic tail, single-pass transmembrane domain, stem region, and C-terminal catalytic domain.

The second model, the oligomerization or aggregation model,20,21 proposes that

formation of oligomeric complexes of glycosyltransferases within the Golgi prevents

their delivery to secretory vessels, thus causing them to be retained. Several recent

reports involving ganglioside glycosyltransferases have provided evidence for this second

model. Work from the Maccioni group identified two independent complexes formed

from ganglioside glycosyltransferases: 1) GalT1, SialT1, and SialT2 in the proximal

Golgi22 and 2) GalNAcT and GalT2 in the TGN23 of CHO cells. The formation of these

complexes was found to be important for their localization within the secretory pathway

6

and could facilitate substrate channeling in ganglioside synthesis. Furthermore, they also

showed that the N-terminal region was sufficient for mediating complex formation. In

separate experiments, the Yu group identified a complex formed from SialT2 and

GalNAcT in the Golgi of F-11A murine neuroblastoma cells.24 These differing reports

should not be assumed as mutually exclusive, rather complex formation may be cell-type

specific and based upon the transcriptional regulation of ganglioside glycosyltransferases.

The study of homo-oligomerization of ganglioside glycosyltransferases and their

potential effects on biological function is explored in Chapter 2.

Gangliosides mediate interactions on the cell surface

Upon delivery to the extracellular plasma membrane of cells, gangliosides have

been shown to play important roles in the regulation of cell-cell interactions. Changes in

cell-surface glycosylation are frequently observed in tumor cells, which are often a target

for an immune response. Natural killer (NK) cells, a specialized lymphocyte cell, will

naturally target and suppress these cells. However, the expression of GD3 on the surface

of tumors cells activates an inhibitory response against NK cell-mediated toxicity through

Siglec-7 binding, allowing tumor cell proliferation to occur.25 Complex gangliosides are

also shown to play critical roles in the central nervous system. Nerve tissues utilize a

myelin-associated glycoprotein (MAG) to stabilize axon-myelin interactions within the

brain. Additionally, this protein is shown to inhibit nerve regeneration after severe

injuries to the central nervous system. The control of these functions is regulated by

highly specific interactions involving the complex gangliosides GD1a and GT1b.26,27

Cell surface gangliosides, such as GM3 and GD1a, also serve as both positive and

negative regulators of several receptor tyrosine kinases. The epidermal growth factor

receptor (EGFR) is a cell surface protein that binds to growth factor ligands and initiates

several downstream signaling cascades. Upregulation of EGFR signaling is thought to be

a critical event in the development of several types of tumors. This response can be

directly inhibited by interactions with GM3 at the cell surface without interfering with

epidermal growth factor (EGF) binding.28 Elevated levels of cell surface GM3 has been

shown to diminish the signaling ability of the insulin receptor (IR), implicating GM3 as a

regulator of insulin sensitivity and perhaps type II diabetes.29 Cancer cell proliferation

7

depends upon angiogenesis through signaling mechanisms involving the vascular

endothelial growth factor receptor (VEGFR). Introduction of GD1a to the surface of

human vascular endothelial cells (HUVECs) has been shown to enhance VEGFR

signaling and cell growth.30 Conversely, this signal can be attenuated by the expression of

GM3, which is the primary ganglioside expressed by these cells.31 These results highlight

the importance of gangliosides in a variety of recognition and signaling events and

highlight ability to modulate protein function through highly specific interactions.

The role of NeuGc expression in gangliosides

Sialic acids are a family of nine-carbon α-keto acids derived from N-

acetylneuraminic acid (NeuAc), N-glycolylneuraminic acid (NeuGc) and deaminated

neuraminic acid (KDN) (Figure 5). Sialic acids are typically found attached to the

nonreducing terminus of glycoproteins (both N-linked and O-linked) and

glycosphingolipids in vertebrates and “higher” invertebrates such as starfish and sea

urchins. This terminal localization allows them to serve as ligands for numerous selectins

and siglecs to mediate a large number of cell-cell adhesion processes in inflammation and

the immune response.32,33 Sialic acid-containing glycoconjugates are commonly modified

by post-glycosylational processing, leading to over 50 naturally occurring variants of

sialic acid.34,35

Figure 5. Common sialic acid structures. Sialic acids are group of nine-carbon α-keto acids derived from N-acetylneuraminic acid (NeuAc), N-glycolylneuraminic acid (NeuGc), and deaminated neuraminic acid (KDN).

The most common forms of sialic acid in mammals are NeuAc and NeuGc, which

are attached onto glycoconjugates by sialyltransferases that utilize cytidine

monophosphate (CMP) activated sugars. CMP-NeuGc is synthesized through the

hydroxylation of CMP-NeuAc by CMP-NeuAc hydroxylase (CMAH). While NeuGc is

8

abundant in most mammals, it is found at extremely low levels in human serum and

organs.7 The genomic source of humans’ inability to synthesize CMP-NeuGc occurs from

an exon deletion/frameshift in the human CMAH gene36,37 estimated to have occurred

~2.5-3 million years ago.38 While this mutation has rendered humans unable to naturally

synthesize NeuGc, it can be acquired from dietary sources, including red meat (lamb,

pork and beef) and milk, and incorporated into human glycoconjugates.39

NeuGc-containing gangliosides are key recognition elements of numerous

pathogens. E. coli infection in mammals is caused by attachment of adhesion proteins

(such as fimbriae or pilli structures) to the intestinal epithelium where the bacteria can

colonize. The K99 fimbriae, generated by E. coli K12, is known to cause severe neonatal

diarrhea to many developing mammals but not in humans; the cause has been elucidated

to occur through a NeuGc-dependent interaction with intestinal gangliosides.40 Subtilase

cytotoxin (SubAb) is an AB5 toxin secreted by Shiga toxigenic E. coli that causes

gastrointestinal disease in humans.41,42 Attachment of this toxin to humans by NeuGc is

greatly enhanced by expression of NeuGc glycoconjugates, presumably through dietary

acquisition of NeuGc. Simian virus 40 (SV40), which utilizes GM1 ganglioside for cell

surface binding, shows significant increases in binding efficiency for GM1-NeuGc over

GM1-NeuAc.43 To better understand the source of NeuGc-containing glycoconjugates in

humans, the NeuGc specificity of the ganglioside glycosyltransferase ST3Gal5 will be

explored in Chapter 4.

Changes in cell surface glycosylation are a hallmark of cancer. Specifically,

many types of tumor cells exhibit altered ganglioside expression, including introduction

of NeuGc.7 As humans do not naturally synthesize NeuGc, its cell surface presentation is

recognized as foreign by the immune system. Consequently, the human immune system

can generate antibodies that specifically recognize the GM3 antigen. In the 1920s,

Hanganutziu44 and Deicher45 independently observed the cellular expression of GM3-

NeuGc in patients treated by injections containing animal antisera. Known today as the

Hanganutziu-Deicher (HD) antigen, the molecular targeting of this antigen by antibodies

is dependent on the presence of NeuGc.46,47 Since these initial studies, researchers have

uncovered a large variety of anti-NeuGc antibodies that are commonly found in healthy

9

humans.48,49 Chickens, which are also unable to naturally synthesize NeuGc, have been

successfully utilized for generating specific anti-ganglioside antibodies.50 Unfortunately,

recent studies have postulated that anti-NeuGc antibody recognition may be a survival

mechanism used by cancer cells to enhance their own propagation.51,52

Metabolic oligosaccharide engineering of gangliosides as a potential tool for

studying gangliosides

The ability to interface chemical techniques and tools from organic chemistry into

the investigation and manipulation of living systems has enabled the expansion of

understanding how biological molecules interact together. The application of this

methodology into protein and lipid glycosylation is known as metabolic oligosaccharide

engineering. Metabolic oligosaccharide engineering refers to the ability to introduce

small structural changes into cellular glycans through the use of unnatural

monosaccharide analogs. To install these glycan modifications, synthetic analogs of

monosaccharide are added to the media of cultured cells or injected into animals.53 Once

inside, endogenous cellular machinery converts the sugar analogs to activated nucleotide

sugar donors, which can be transferred to glycoconjugate substrates. Using this

technology, a diverse class of chemical modifications (extended N-acyl chains, ketones,

azides, thiols, alkynes, diazirines) has been introduced into a variety of sugars, including

sialic acid (Figure 6).53 These non-biological reactive groups have provided unique

chemical reactivity to glycans, which can be used to better understand their function and

localization in biological organisms, without labor-intensive synthetic methods.

Furthermore, these molecules are installed into a wide variety of glycan structures,

allowing for the investigation of a large number of protein interaction partners.

10

Figure 6. Examples of unnatural sialic acid analogs incorporated onto cellular glycoconjugates. Using metabolic oligosaccharide engineering techniques, a large number of chemical modifications have been introduced into cell surface sialic acids (shown in red). The assigned names of each molecule are listed below each structure; attachment to cell surface glycoconjugate is denoted by “R”.

The application of metabolic oligosaccharide engineering as a tool to investigate

the biological roles of gangliosides is a relatively new area of study. The introduction of

bio-orthogonal reactivity into sialic acid structure has been explored as a method for

labeling of mammalian glycans. The incorporation of azides into the N-acyl chain of

sialic acid yields molecular that can be selectively reacted with various reagents inside of

cells and in animals.54 Utilizing the metabolic precursor ManNAz, Bussink et al. showed

successful engineering of GM3 with the metabolized sialic acid analog SiaNAz.55

Interestingly, they observed higher ratios of SiaNAz incorporation into gangliosides over

cell surface glycoproteins. As an initial proof-of-principle experiment, these results

underscore the potential application of monitoring ganglioside levels on the surface of

cells.

Changes in ganglioside expression have been observed in several types of cancer,

including melanoma, colon cancer, and breast cancer.7 These distinct changes in

gangliosides offer a potential target for cancer therapeutics.56,57 An emerging strategy

toward this goal takes advantage of metabolic oligosaccharide engineering to introduce

11

slight structural changes into gangliosides. Even small changes can be sufficient to make

the modified gangliosides appear foreign and immunogenic. These techniques also

ensure that the specific ganglioside response will be active only when unnatural sugar

analogs are administered, minimizing the risk of an autoimmune response. In an attempt

to improve the immunogenicity of GD3, Jennings and co-workers used metabolic

oligosaccharide engineering to introduce a butyryl group to the N-acyl side chain of both

sialic acids that could be used for targeting.58 Upon generation of a highly specific

antibody recognizing this unnatural ganglioside, they demonstrated that they could

selectively target melanoma cells cultured with the N-butyrylated ManNAc precursor,

ManNBut. More remarkably, mice harboring melanoma-derived tumors demonstrated

suppressed tumor growth when treated with combination therapy comprised of the anti-

GD3Bu and ManNBut. While these results appeared promising, the treatment was

unable to eliminate established tumors.

Guo and co-workers have employed a similar approach for development of an

immunotherapy directed against GM3 through introduction of a phenylacetyl group into

the N-acyl side chain.59-61 Using their own highly specific anti-GM3PhAc antibody, they

demonstrated selective targeting of melanoma cells cultured with N-phenylacetyl

ManNAc precursor, ManNPhAc. This targeting was found to be highly cytotoxic to the

ManNPhAc cultured cells. These results suggest that immunotherapy directed against

metabolically engineered gangliosides could provide a potential cancer treatment with

minimal toxic side effects.

Photocapturing glycosphingolipid interactions in cells

Despite some knowledge of the biological roles that glycosphingolipids play in

the plasma membrane, minimal molecular details of the observed interactions are known.

Glycan-dependent binding events are low affinity and suffer from fast off rates. As a

result, these macromolecular complexes tend to dissociate rapidly and are difficult to

isolate by common purification methods. Photocrosslinkers offer a potentially powerful

approach for capturing glycosphingolipid-protein complexes and mapping the binding

surfaces. Upon activation by specific wavelengths, these molecules transform into highly

reactive species that form covalent adducts with nearby atoms. The most common

12

photoreactive groups with demonstrated utility in biological settings include

benzophenones, aryl azides, and diazirines (Figure 7).62,63 The incorporation of

photoreactive functional groups through metabolic oligosaccharide engineering presents

an opportunity to capture and characterize glycosphingolipid-mediated interactions.

Figure 7. Common photoreactive groups used in biology. Benzophenones, aryl azides, and diazirines are some of the most common photoreactive groups that are used in biological samples. Each of these molecules can be activated with UV light to produce reactive species that form covalent adducts with nearby targets.

A variety of approaches have been employed to study ganglioside interactions

with photochemistry. Photoreactive groups have been incorporated into either the

carbohydrate moiety, to detect glycan-based interactions, or in the fatty acid chains, to

detect lipid-mediated interactions.64,65 Caveolae are plasma membrane invaginations that

are implicated in endocytosis and signal transduction pathways. Using radiolabeled,

photoreactive ganglioside analogs, several groups have reported the presence of GM1

within these domains66,67 and have also identified potential binding partners for these

molecules.66 Gangliosides are also known to be important for regulating signal

transduction events on the cell surface. To this end, photoreactive ganglioside analogs

have been successful at capturing complexes with Src-family kinases (c-Src, Lyn)68 and

receptor tyrosine kinases (insulin receptor).9 Additionally, a GM1 probe was

photocrosslinked to α- and β-tubulin,69 suggesting potential mechanistic roles for

gangliosides in the structural remodeling of the cell.

The primary advantage of installing photoreactive groups into gangliosides is

their usefulness at capturing glycan-binding events in situ to uncover underlying

molecular details. Bacterial toxins, such as tetanus toxin, invade the host cell through

molecular recognition of cell surface gangliosides. Starting with naturally occurring

GD1b, Schnaar and coworkers attached an aryl-azide group radiolabeled with 125I onto

the N-acyl chain of sialic acid and were able to capture tetanus toxin, enzymatically

13

digest the complex and isolate the binding region.8 The location of the sialic-acid binding

site was later confirmed by cocrystal structures of the toxin bound to lactose and a GD1a

analog.70,71

The metabolic breakdown of all glycosphingolipid species occurs within the

lysosome. Through the action of many glycosidases and essential cofactors, these

molecules are degraded by step-wise mechanisms into reusable metabolites for a cell.

Using a GM2 analog containing a trifluoromethyl phenyl diazirine and 14C-label,

Sandhoff and coworkers captured the GM2-activator protein (GM2AP) and indentified a

direct interaction occurring between the ceramide of GM2 and a surface loop on

GM2AP.72 These results were consistent with reported crystal structures of GM2AP that

had identified this loop as the most flexible surface of the protein and was most likely to

act as a ligand-binding domain.73,74 Chapter 3 will explore the incorporation of the

photoreactive diazirine group into the gangliosides of Jurkat cells and its applicability in

capturing the glycan-mediated interaction of GM1 to cholera toxin subunit B.

Conclusions

Gangliosides are an important class of glycosphingolipid that regulate numerous

cell surface events. The diversity of structures that can be synthesized allow for an

individual cell to maintain many of these processes simultaneously. Changes in cell

surface ganglioside expression have been shown to play important roles in embryo

development and in cancer progression, but we currently have a limited understanding of

the molecular mechanisms that underlie these important changes. Through the continual

development of biological and chemical tools, we have learned much about gangliosides

but there is still a large amount of information remaining to be uncovered and

understood.

14

References

1 Futerman, A. H. & Hannun, Y. A. The complex life of simple sphingolipids.

EMBO Rep 5, 777-782, (2004).

2 Wennekes, T. et al. Glycosphingolipids--nature, function, and pharmacological

modulation. Angew Chem Int Ed Engl 48, 8848-8869, (2009).

3 Lopez, P. H. & Schnaar, R. L. Gangliosides in cell recognition and membrane

protein regulation. Curr Opin Struct Biol 19, 549-557, (2009).

4 Schengrund, C. L. "Multivalent" saccharides: development of new approaches for

inhibiting the effects of glycosphingolipid-binding pathogens. Biochem

Pharmacol 65, 699-707, (2003).

5 van der Meer-Janssen, Y. P., van Galen, J., Batenburg, J. J. & Helms, J. B. Lipids

in host-pathogen interactions: pathogens exploit the complexity of the host cell

lipidome. Prog Lipid Res 49, 1-26, (2010).

6 Yu, R. K., Nakatani, Y. & Yanagisawa, M. The role of glycosphingolipid

metabolism in the developing brain. J Lipid Res 50 Suppl, S440-445, (2009).

7 Malykh, Y. N., Schauer, R. & Shaw, L. N-Glycolylneuraminic acid in human

tumours. Biochimie 83, 623-634, (2001).

8 Shapiro, R. E. et al. Identification of a ganglioside recognition domain of tetanus

toxin using a novel ganglioside photoaffinity ligand. J Biol Chem 272, 30380-

30386, (1997).

9 Kabayama, K. et al. Dissociation of the insulin receptor and caveolin-1 complex

by ganglioside GM3 in the state of insulin resistance. Proc Natl Acad Sci U S A

104, 13678-13683, (2007).

10 Kawashima, N., Yoon, S. J., Itoh, K. & Nakayama, K. Tyrosine kinase activity of

epidermal growth factor receptor is regulated by GM3 binding through

carbohydrate to carbohydrate interactions. J Biol Chem 284, 6147-6155, (2009).

15

11 Maccioni, H. J., Daniotti, J. L. & Martina, J. A. Organization of ganglioside

synthesis in the Golgi apparatus. Biochim Biophys Acta 1437, 101-118, (1999).

12 Tu, L. & Banfield, D. K. Localization of Golgi-resident glycosyltransferases. Cell

Mol Life Sci 67, 29-41, (2010).

13 Yuan, Z. & Teasdale, R. D. Prediction of Golgi Type II membrane proteins based

on their transmembrane domains. Bioinformatics 18, 1109-1115, (2002).

14 Bretscher, M. S. & Munro, S. Cholesterol and the Golgi apparatus. Science 261,

1280-1281, (1993).

15 Orci, L. et al. Heterogeneous distribution of filipin--cholesterol complexes across

the cisternae of the Golgi apparatus. Proc Natl Acad Sci U S A 78, 293-297,

(1981).

16 Munro, S. Sequences within and adjacent to the transmembrane segment of alpha-

2,6-sialyltransferase specify Golgi retention. EMBO J 10, 3577-3588, (1991).

17 Munro, S. An investigation of the role of transmembrane domains in Golgi

protein retention. EMBO J 14, 4695-4704, (1995).

18 Ma, J., Qian, R., Rausa, F. M., & Colley, K. J. Two naturally occurring alpha2,6-

sialyltransferase forms with a single amino acid change in the catalytic domain

differ in their catalytic activity and proteolytic processing. J Biol Chem 272, 672-

679, (1997).

19 Chen, C., Ma, J., Lazic, A., Backovic, M. & Colley, K. J. Formation of insoluble

oligomers correlates with ST6Gal I stable localization in the golgi. J Biol Chem

275, 13819-13826, (2000).

20 Roseman, S. The synthesis of complex carbohydrates by multiglycosyltransferase

systems and their potential function in intercellular adhesion. Chem Phys Lipids 5,

270-297, (1970).

16

21 Machamer, C. E. Golgi retention signals: do membranes hold the key? Trends

Cell Biol 1, 141-144, (1991).

22 Giraudo, C. G. & Maccioni, H. J. Ganglioside glycosyltransferases organize in

distinct multienzyme complexes in CHO-K1 cells. J Biol Chem 278, 40262-

40271, (2003).

23 Giraudo, C. G., Daniotti, J. L. & Maccioni, H. J. Physical and functional

association of glycolipid N-acetyl-galactosaminyl and galactosyl transferases in

the Golgi apparatus. Proc Natl Acad Sci U S A 98, 1625-1630, (2001).

24 Bieberich, E. et al. Regulation of ganglioside biosynthesis by enzyme complex

formation of glycosyltransferases. Biochemistry 41, 11479-11487, (2002).

25 Nicoll, G. et al. Ganglioside GD3 expression on target cells can modulate NK cell

cytotoxicity via siglec-7-dependent and -independent mechanisms. Eur J Immunol

33, 1642-1648, (2003).

26 Vinson, M. et al. Myelin-associated glycoprotein interacts with ganglioside

GT1b. A mechanism for neurite outgrowth inhibition. J Biol Chem 276, 20280-

20285, (2001).

27 Vyas, A. A. et al. Gangliosides are functional nerve cell ligands for myelin-

associated glycoprotein (MAG), an inhibitor of nerve regeneration. Proc Natl

Acad Sci U S A 99, 8412-8417, (2002).

28 Bremer, E. G., Schlessinger, J. & Hakomori, S. Ganglioside-mediated modulation

of cell growth. Specific effects of GM3 on tyrosine phosphorylation of the

epidermal growth factor receptor. J Biol Chem 261, 2434-2440, (1986).

29 Yamashita, T. et al. Enhanced insulin sensitivity in mice lacking ganglioside

GM3. Proc Natl Acad Sci U S A 100, 3445-3449, (2003).

30 Liu, Y., McCarthy, J. & Ladisch, S. Membrane ganglioside enrichment lowers the

threshold for vascular endothelial cell angiogenic signaling. Cancer Res 66,

10408-10414, (2006).

17

31 Mukherjee, P. et al. Thematic review series: sphingolipids. Ganglioside GM3

suppresses the proangiogenic effects of vascular endothelial growth factor and

ganglioside GD1a. J Lipid Res 49, 929-938, (2008).

32 Varki, A. Glycan-based interactions involving vertebrate sialic-acid-recognizing

proteins. Nature 446, 1023-1029, (2007).

33 Schauer, R. Sialic acids as regulators of molecular and cellular interactions. Curr

Opin Struct Biol 19, 507-514, (2009).

34 Angata, T. & Varki, A. Chemical diversity in the sialic acids and related alpha-

keto acids: an evolutionary perspective. Chem Rev 102, 439-469, (2002).

35 Yu, H. & Chen, X. Carbohydrate post-glycosylational modifications. Org Biomol

Chem 5, 865-872, (2007).

36 Chou, H. H. et al. A mutation in human CMP-sialic acid hydroxylase occurred

after the Homo-Pan divergence. Proc Natl Acad Sci U S A 95, 11751-11756,

(1998).

37 Irie, A., Koyama, S., Kozutsumi, Y., Kawasaki, T. & Suzuki, A. The molecular

basis for the absence of N-glycolylneuraminic acid in humans. J Biol Chem 273,

15866-15871, (1998).

38 Chou, H. H. et al. Inactivation of CMP-N-acetylneuraminic acid hydroxylase

occurred prior to brain expansion during human evolution. Proc Natl Acad Sci U

S A 99, 11736-11741, (2002).

39 Tangvoranuntakul, P. et al. Human uptake and incorporation of an immunogenic

nonhuman dietary sialic acid. Proc Natl Acad Sci U S A 100, 12045-12050,

(2003).

40 Kyogashima, M., Ginsburg, V. & Krivan, H. C. Escherichia coli K99 binds to N-

glycolylsialoparagloboside and N-glycolyl-GM3 found in piglet small intestine.

Arch Biochem Biophys 270, 391-397, (1989).

18

41 Byres, E. et al. Incorporation of a non-human glycan mediates human

susceptibility to a bacterial toxin. Nature 456, 648-652, (2008).

42 Paton, J. C. & Paton, A. W. Pathogenesis and diagnosis of Shiga toxin-producing

Escherichia coli infections. Clin Microbiol Rev 11, 450-479, (1998).

43 Campanero-Rhodes, M. A. et al. N-glycolyl GM1 ganglioside as a receptor for

simian virus 40. J Virol 81, 12846-12858, (2007).

44 Hanganutziu, M. Hemagglutinines heterogenetiques apres injection de serum de

Cheval. Compt Rend Soc Biol 91, 1457-1459, (1924).

45 Deicher, H. Ueber die Erzeugung heterospezifisher Haemagglutinine durch

Injektion artfremden Serums. Z Hyg Infektionskr 106, 561-579, (1926).

46 Higashi, H., Naiki, M., Matuo, S. & Okouchi, K. Angtigen of "serum sickness"

type of heterophile antibodies in human sera: indentification as gangliosides with

N-glycolylneuraminic acid. Biochem Biophys Res Commun 79, 388-395, (1977).

47 Merrick, J. M., Zadarlik, K. & Milgrom, F. Characterization of the Hanganutziu-

Deicher (serum-sicknss) antigen as gangliosides containing N-glycolylneuraminic

acid. Int Arch Allergy Appl Immunol 57, 477-480, (1978).

48 Zhu, A. & Hurst, R. Anti-N-glycolylneuraminic acid antibodies identified in

healthy human serum. Xenotransplantation 9, 376-381, (2002).

49 Padler-Karavani, V. et al. Diversity in specificity, abundance, and composition of

anti-Neu5Gc antibodies in normal humans: potential implications for disease.

Glycobiology 18, 818-830, (2008).

50 Asaoka, H., Nishinaka, S., Wakamiya, N., Matsuda, H. & Murata, M. Two

chicken monoclonal antibodies specific for heterophil Hanganutziu-Deicher

antigens. Immunol Lett 32, 91-96, (1992).

51 Hedlund, M. et al. N-glycolylneuraminic acid deficiency in mice: implications for

human biology and evolution. Mol Cell Biol 27, 4340-4346, (2007).

19

52 Pham, T. et al. Evidence for a novel human-specific xeno-auto-antibody response

against vascular endothelium. Blood 114, 5225-5235, (2009).

53 Du, J. et al. Metabolic glycoengineering: sialic acid and beyond. Glycobiology 19,

1382-1401, (2009).

54 Sletten, E. M. & Bertozzi, C. R. Bioorthogonal chemistry: fishing for selectivity

in a sea of functionality. Angew Chem Int Ed Engl 48, 6974-6998, (2009).

55 Bussink, A. P. et al. N-azidoacetylmannosamine-mediated chemical tagging of

gangliosides. J Lipid Res 48, 1417-1421, (2007).

56 Chapman, P. B. et al. Vaccination with a bivalent G(M2) and G(D2) ganglioside

conjugate vaccine: a trial comparing doses of G(D2)-keyhole limpet hemocyanin.

Clin Cancer Res 6, 4658-4662, (2000).

57 Fredman, P., Hedberg, K. & Brezicka, T. Gangliosides as therapeutic targets for

cancer. BioDrugs 17, 155-167, (2003).

58 Zou, W. et al. Bioengineering of surface GD3 ganglioside for immunotargeting

human melanoma cells. J Biol Chem 279, 25390-25399, (2004).

59 Pan, Y., Chefalo, P., Nagy, N., Harding, C. & Guo, Z. Synthesis and

immunological properties of N-modified GM3 antigens as therapeutic cancer

vaccines. J Med Chem 48, 875-883, (2005).

60 Chefalo, P., Pan, Y., Nagy, N., Guo, Z. & Harding, C. V. Efficient metabolic

engineering of GM3 on tumor cells by N-phenylacetyl-D-mannosamine.

Biochemistry 45, 3733-3739, (2006).

61 Wang, Q., Zhang, J. & Guo, Z. Efficient glycoengineering of GM3 on melanoma

cell and monoclonal antibody-mediated selective killing of the glycoengineered

cancer cell. Bioorg Med Chem 15, 7561-7567, (2007).

62 Brunner, J. New photolabeling and crosslinking methods. Annu Rev Biochem 62,

483-514, (1993).

20

63 Tanaka, Y., Bond, M. R. & Kohler, J. J. Photocrosslinkers illuminate interactions

in living cells. Mol Biosyst 4, 473-480, (2008).

64 Mauri, L. et al. Synthesis of radioactive and photoactivable ganglioside

derivatives for the study of ganglioside-protein interactions. Glycoconj J 20, 11-

23, (2004).

65 Vodovozova, E. L. Photoaffinity labeling and its application in structural biology.

Biochemistry (Mosc) 72, 1-20, (2007).

66 Pitto, M. et al. Use of a photoactivable GM1 ganglioside analogue to assess lipid

distribution in caveolae bilayer. Glycoconj J 17, 215-222, (2000).

67 Fra, A. M., Masserini, M., Palestini, P., Sonnino, S. & Simons, K. A photo-

reactive derivative of ganglioside GM1 specifically cross-links VIP21-caveolin on

the cell surface. FEBS Lett 375, 11-14, (1995).

68 Prinetti, A. et al. Association of Src-family protein tyrosine kinases with

sphingolipids in rat cerebellar granule cells differentiated in culture. Glycoconj J

17, 223-232, (2000).

69 Palestini, P. et al. Tubulin anchoring to glycolipid-enriched, detergent-resistant

domains of the neuronal plasma membrane. J Biol Chem 275, 9978-9985, (2000).

70 Emsley, P. et al. The structures of the H(C) fragment of tetanus toxin with

carbohydrate subunit complexes provide insight into ganglioside binding. J Biol

Chem 275, 8889-8894, (2000).

71 Fotinou, C. et al. The crystal structure of tetanus toxin Hc fragment complexed

with a synthetic GT1b analogue suggests cross-linking between ganglioside

receptors and the toxin. J Biol Chem 276, 32274-32281, (2001).

72 Wendeler, M. et al. Photoaffinity labelling of the human GM2-activator protein.

Mechanistic insight into ganglioside GM2 degradation. Eur J Biochem 271, 614-

627, (2004).

21

73 Wright, C. S., Li, S. C. & Rastinejad, F. Crystal structure of human GM2-

activator protein with a novel beta-cup topology. J Mol Biol 304, 411-422, (2000).

74 Wright, C. S., Zhao, Q. & Rastinejad, F. Structural analysis of lipid complexes of

GM2-activator protein. J Mol Biol 331, 951-964, (2003).

22

Chapter 2 - Exploring the molecular basis for associations among glycosphingolipid

glycosyltransferases

Introduction

In mammalian cells, the endoplasmic reticulum (ER) and Golgi are home to

numerous glycosyltransferases that transfer monosaccharides from nucleotide sugars to

proteins, lipids, and many small molecules. While there have been over 230 human

glycosyltransferases identified to date,1 the process by which complex carbohydrate

structures are assembled onto specific targets has yet to be determined in the same detail

as that of the synthesis of nucleic acids and proteins. In order to explain the regulated

synthesis of complex carbohydrate structures, there have been two important and non-

exclusive models proposed, the assembly line model2 and the multiglycosyltransferase

system.3 In the assembly line model, enzymes are arranged in a linear array that

corresponds to the order in which the modifications occur.4 In the

multiglycosyltransferase system, enzymes that are involved in the particular sequence of

a synthesis are proposed to associate together in complexes, thus allowing for the product

of one enzyme to become the substrate for the next enzyme.3

Evidence consistent with the multiglycosyltransferase model has been obtained

for the enzymes involved in ganglioside biosynthesis (Figure 1A). Gangliosides are a

specific class of glycolipid molecules that contain lactosylceramide (LacCer, Galβ1,4-

Glcα-ceramide) as the core structure and contain at least one sialic acid (Sia) residue

within the structure. Figure 1B shows the structure of the GM3 ganglioside. Three sets

of physical interactions among ganglioside-synthesizing glycosyltransferases have been

reported.5-8 The Maccioni group first reported an interaction between GalNAcT &

GalT2,6 followed by an interaction between GalT1, SialT1, and SialT2.7 Independently,

Bierberich et al. showed an interaction occurring between SialT2 and GalNAcT.5 Further

studies showed that expression of SialT2 in CHO cells causes relocalization of GalT1 and

SialT1 from early ER/Golgi compartments to more distal locations (TGN and recycling

endosomes).8 Additional experiments demonstrated that sequence information contained

within the cytoplasmic tails of these enzymes was responsible for their subcellular

localization within the secretory pathway.9

23

Figure 1. Reported associations among glycosyltransferases responsible for ganglioside biosynthesis. (A) Biosynthetic pathway of gangliosides in mammalian cells. Gangliosides are shown in black font, glycosyltransferases in red font, and reported associations are indicated by the shaded bubbles. Actual gene names for each glycosyltransferases are listed beside pathway figure. (B) GM3 ganglioside. The lactosylceramide (LacCer) core structure is shown in black and the sialic acid residue is shown in red.

Ganglioside glycosyltransferases are type II transmembrane (TM) proteins,

consisting of a N-terminal cytoplasmic tail, single-pass TM domain, stem region and C-

terminal catalytic domain. To better understand the nature of the physical association,

the Maccioni group used truncated sequences of these glycosyltransferases and showed

that the N-terminal domains containing the cytoplasmic tail, the TM domain and all or

part of the stem region were sufficient to mediate association.6-9 Upon further review of

the amino acid sequence of these proteins and considering there are numerous proteins in

nature that associate directly via TM interactions,10,11 I began to speculate that the TM

domain might be responsible for the observed association.

24

Thus, I sought to characterize the molecular basis for these observed physical

interactions, focusing primarily on the TM domains of the enzymes using in vitro assays.

In my efforts to investigate hetero-oligomerization of these enzymes, I discovered that

several of the TM regions formed homo-oligomers in various multiplicities. I then

identified the putative key amino acids that are responsible for homo-oligomerization. In

order to understand the biological relevance of these self-associating TM domains, I

expressed full-length glycosyltransferases in CHO cells and observed their behavior

using an ER retention assay published by our research group12 and by Western blot

analysis. My preliminary results indicate that several ganglioside glycosyltransferases

are able to self-associate inside of cells. Formation of these complexes may be

responsible for the correct localization and production of cellular gangliosides.

Results

In vitro assay to investigate TM interactions

To investigate whether ganglioside glycosyltransferase TM domains can mediate

oligomerization, I used hydropathy analysis13 to identify the putative TM domains of the

five ganglioside glycosyltransferases reported as interaction partners (Figure 2). Using

this information, I decided to use an established dimerization assay that tests the ability of

hydrophobic TM domains to associate within detergent (sodium dodecyl sulfate - SDS)

micelles.14 Each TM domain was expressed as a fusion protein with staphylococcal

nuclease (SN), a non-interacting globular protein (named SN-TM). The use of SN

facilitates the expression and purification of the hydrophobic TM domains (described in

Methods). The purified proteins were then analyzed by SDS-PAGE to determine

whether they associate in the detergent environment. The SDS micelles provide a

suitable mimic of a biological membrane, thus SN-TM fusion proteins that form dimers

in cellular membranes typically migrate as dimers on the SDS gel. The biologically-

relevant dimerization of the TM region of glycophorin A (GpA) has been observed using

this method. An SN-TM construct containing the GpA TM region was used as a positive

control for dimer formation. All purification steps and SDS-PAGE analyses were

performed in the presence of dithiothreitol (DTT) to prevent spurious formation of

intramolecular disulfide bonds between the TM domains.

25

Figure 2. N-terminal sequences of ganglioside glycosyltransferases implicated in their association. Putative transmembrane domains are shaded in pink.

Investigating transmembrane interactions in vitro

The individual SN-TM fusions were analyzed by SDS-PAGE (Figure 3). Both

GalNAcT (Figure 3A – Lane 1) and GalT2 (Figure 3A – Lane 2) migrated almost

exclusively as monomers. The appearance of higher-order oligomers for GalT2 was

determined to be caused by solvent effects and not due to TM association. In contrast,

GalT1 (Figure 3B – Lane 1) produces two very distinct and similarly intense bands on a

gel, correlating to monomer and dimer molecular weights. Similarly, SialT1 (Figure 3B

– Lane 2) produces two distinct bands on a gel, corresponding to the molecular weights

of a monomer and a dimer. SialT2 (Figure 3B – Lane 3) displays multiple

oligomerization states; their molecular weight suggests that these bands represent a trimer

and pentamer, along with other distinct higher-order oligomers. To prove that these

observed interactions were indeed a product of the SN-TM fusions and not due to

unremoved contaminants, I injected these samples onto a reverse phase-high performance

liquid chromatography (RP-HPLC) instrument, isolated the corresponding SN-TM

fusion, and re-analyzed the sample by SDS-PAGE and saw that the observed associations

were reproduced (data not shown).

To investigate hetero-oligomer formation, equimolar quantities of SN-TM fusion

proteins were mixed together in SDS and analyzed by SDS-PAGE.15 Interestingly, I was

unable to observe the predicted complex formation among different TM domains (Figure

3A – Lane 3, Figure 3B – Lanes 4-6). Modifying several experimental parameters,

including mixing time and temperature, also failed to produce any observable hetero-

oligomers.

26

Figure 3. SDS-PAGE analysis of SN-TM homo- and hetero-oligomerization. Molecular weight markers are shown in the left lane of each gel. GpA is the fusion of SN and the transmembrane domain of glycophorin A, which is know to form a homodimer. (A) The transmembranes of GalNAcT and GalT2 fail to demonstrate any homo-oligomerization or hetero-oligomerization when mixed together. The higher order bands present in GalT2 lanes are due to preparation in organic solvent and were demonstrated to be artifacts. (B) The SN-TM fusions of GalT1 and SialT1 are able to form stable dimers while SialT2 shows multiple oligomerization states, including a trimer. The mixing of these individual samples together fails to show any hetero-oligomerization formation.

The discovery of the self-association with several of the TM domains prompted us

to conduct a more detailed investigation into the molecular basis of these interactions.

Close examination of the amino acid sequences of the TM regions revealed the presence

of a number of polar and charged residues located within the mostly hydrophobic

environment. Several reports have demonstrated that the presence of polar residue motifs

in TM regions facilitates their association.16-20 For example, GalT1 contains a sequence

of amino acids (SxxSSxxY) that bears a striking resemblance to a TM dimerization motif

(SxxSSxxT) discovered in a selection experiment.16 Additionally, the spacing of polar

residues three amino acids apart, as seen in the TM domains of GalT1, SialT1 and SialT2,

has been shown to direct TM association.18 Polar residues within the TM domains of

several glycosyltransferases have been suggested to play important roles in dimer

formation and localization (e.g. β1,4-galactosyltransferase,21,22 α2,6-sialyltransferase I,23

and α1,3/4-fucosyltransferase III24). Considering that TM sequences of these ganglioside

27

glycosyltransferases likely form α-helical structures, the arrangement of polar residues

three amino acids apart presents them on the same face of the helix, allowing both

residues to be in simultaneous contact with a neighboring α-helix.

To test whether a similar mechanism is at work in the TM domain of GalT1, I

performed conservative mutations of each of the polar residues to a nonpolar analog

(serine and cysteine to alanine, tyrosine to phenylalanine) of the SN-TM construct and

investigated their effects by SDS-PAGE (Figure 4A). Additionally, I quantified the

relative abundance of monomer and dimers of each mutant (Figure 4B). Mutating

individual amino acids within the TM sequence showed only modest decreases in dimer

formation. Interestingly, the C28A and Y30F mutations actually showed slight increases

in dimer formation. The introduction of two separate double mutations, S23A/S26A or

S25A/C28A, produced dramatic decreases in dimer formation. This result suggests that

the disruption of three amino acid-spaced motifs can modulate the dimer formation of

GalT1. To my surprise, the double mutation of S27A/Y30F actually increased the dimer

formation, corresponding with my previous results from the single mutation Y30F.

These results suggest that the dimer formation of GalT1’s TM domain is largely due to

the presence of properly spaced polar residues.

Figure 4. SDS-PAGE analysis of SN-GalT1 TM WT and mutated fusion proteins. (A) Single and double mutations of GalT1’s TM domain cause mild to significant disruption of the homodimer. Molecular weight markers are shown in the left lane. GpA is the fusion of SN and the transmembrane domain of glycophorin A, which is know to form a homodimer. (B) Quantitative analysis of the relative ratio of monomer to dimer band pixel intensity. Color coding is used to classify mutations as slight-to-moderate dimer disruption (yellow), large dimer disruption (green) or slight increase in dimer formation (red). GalT1 TM WT is highlighted in black and GpA TM (known homodimer) is highlighted in gray.

28

I next explored how similar mutagenesis studies might affect SialT2’s ability to

self-oligomerize into trimers and higher order complexes. The TM domain of SialT2

contains four polar residues that form two pairs, each spaced three amino acids apart.

Conservative mutations of each amino acid of interest were performed as described above

and the resulting proteins were analyzed by SDS-PAGE (Figure 5A). Additionally, I

quantified the percentages of monomer and oligomer species for each mutant to further

analyze their effects (Figure 5B). The single mutation of S18A, C26A or Y29F resulted

in a modest decrease in the fraction of higher order species while the mutation of C21A

produced a dramatic decrease in oligomer formation. When double mutations were

installed, the S18A/C21A species showed a significant diminution of all higher-order

species, including the trimer. This result indicates that the self-association of SialT2’s

TM domain is governed by the cooperative action of S18 and C21. The C26A/Y29F

mutation only showed moderate destabilizing effects, similar to those observed for the

corresponding single mutants.

Figure 5. SDS-PAGE analysis of SN-SialT2 TM WT and mutated fusion proteins. (A) Single and double mutations of SialT2’s TM domain cause mild to significant disruption of the higher order oligomer formation. Molecular weight markers are shown in the left lane. (B) Quantitative analysis of the percentages of monomer and higher order band pixel intensity. Color coding is used to identify the percentage of SialT2 that forms a monomer (green), trimer (blue), pentamer (red), and higher order oligomers (yellow).

The molecular basis for self-oligomerization of SialT1 was not investigated using

the SDS-PAGE method. However, SN-TM constructs of SialT1 containing single

mutations of polar and charged residues in the TM domain were generated. Additionally,

29

I attempted to synthesize peptides of the TM sequences of GalT1, SialT1 and SialT2 but

was unable to efficiently purify them for further study.

Functional analysis of ganglioside glycosyltransferases

To investigate if the in vitro observations of TM oligomerization reflect

interactions that also occur in a cellular environment, I decided to utilize an established

subcellular relocalization assay that measures the physical interaction between two

proteins.25,26 In this assay, the full-length glycosyltransferase is cloned with one of two

different affinity tags (either myc or HA) at the C-terminus to facilitate

immunofluorescence detection of its subcellular location. One form of the enzyme is

fused with the KDEL retention signal (SEKDEL) at the C-terminus causing the protein to

be recognized by the KDEL receptor in the Golgi and become retrograde transported

back into the ER. Both the KDEL-tagged and non-KDEL tagged enzymes are then co-

transfected into CHO cells and localization is determined by immunofluorescence

microscopy. If the glycosyltransferases physically associate, the non-KDEL tagged

protein will become relocalized into the ER (Figure 6).

30

Figure 6. KDEL recruitment assay. Full length glycosyltransferases (depicted in red and purple) were expressed with different C-terminal affinity tags. One form of the enzyme also contains a KDEL retention signal (shown in yellow). Both proteins are synthesized in the ER and trafficked to the Golgi. The KDEL receptor recognizes and retrieves proteins displaying the retention signal back into the ER. If the untagged enzyme interacts with the KDEL-tagged protein, it will also be retrograde trafficked to the ER. If these two proteins do not associate, the untagged enzyme will remain in the Golgi.

Mammalian expression plasmids encoding GalT1, SialT1 and SialT2 were

generated as fusion proteins with the myc or HA affinity tags and with or without the

KDEL retention signal (described in Methods). The non-KDEL-tagged version of all

three enzymes displayed normal Golgi localization when transfected in CHO cells.

Introduction of the KDEL retention sequence onto the C-terminus efficiently relocalized

GalT1, SialT1, and SialT2 into the ER (Figure 7), thus demonstrating that these

ganglioside glycosyltransferases can be used with the KDEL recruitment assay.

To determine if GalT1, SialT1, and SialT2 are able to form self-oligomerizing

complexes inside of cells, I co-transfected CHO cells with a non-KDEL-tagged

31

glycosyltransferase and a KDEL-tagged glycosyltransferase and monitored the location

of the non-KDEL-tagged glycosyltransferase by immunofluorescence microscopy. When

GalT1-HA and GalT1-myc-KDEL were co-expressed, I observed that GalT1-HA

remained strictly in the Golgi and was not recruited into the ER (Figure 7A). When

SialT1-HA and SialT1-myc-KDEL were co-expressed, I observed that SialT1-HA was

recruited into the ER in ~50% of the cells expressing both enzymes (Figure 7B). When

SialT2-HA and SialT2-myc-KDEL were co-expressed, I observed that SialT2-HA was

recruited into the ER in ~90% of the cells expressing both enzymes (Figure 7C). These

results indicate that both SialT1 and SialT2 are likely oligomerizing inside of cells.

Next, I tested the utility of the KDEL-recruitment assay to view the reported

physical associations occurring between GalT1, SialT1, and SialT2 in CHO cells (Figure

7D).7,8 When GalT1-HA and SialT1-myc-KDEL were co-expressed, I observed that

GalT1-HA was localized in the Golgi and not recruited into the ER (Figure 7D). The

CHO cell line contains endogenous levels of GalT1 and SialT1 but is lacking the

presence of SialT2.7 Previous work has demonstrated that the complex formation of these

three enzymes is dependent on the expression of SialT2.8 Without the presence of SialT2

in this experiment, it is very unlikely that SialT1 alone can recruit GalT1. When Sial1-

HA and SialT2-myc-KDEL were co-expressed, I observed that SialT1-HA was being

recruited into the ER in most cells expressing both enzymes (Figure 7D). In this

experiment, all three enzymes are being expressed in CHO cells and I am able to observe

a physical association occurring between the two tagged glycosyltransferases. These

results suggest that the KDEL recruitment assay can show physical associations

occurring between ganglioside glycosyltransferases.

32

Figure 7. Cellular association of ganglioside glycosyltransferases visualized through the KDEL recruitment assay. Addition of the KDEL retention signal causes ganglioside glycosyltransferases to re-localize to the ER. Other glycosyltransferases that co-localize with the KDEL-tagged enzyme will also be recruited to the ER. Localization is assessed by immunofluorescence microscopy and comparison to known organelle markers. (A) GalT1 does not form strong homo-oligomerization. (B) SialT1 shows small levels of recruitment. (C) SialT2 shows recruitment. (D) GalT1 and SialT1 are not interacting with each other in CHO cells that do not contain endogenous SialT2. When SialT2 is co-expressed with SialT1, SialT1 gets recruited to ER.

Since the SialT2 enzyme displayed the most robust level of recruitment with the

KDEL assay, I sought to obtain further evidence of SialT2 homo-oligomerization

occurring within cells. Using the SialT2-HA clone, I transfected CHO cells and

investigated its possible oligomerization. Using SDS-PAGE and western blot (WB)

techniques, I saw the appearance of several higher-order bands occurring with SialT2-HA

(Figure 8A). When the samples were subjected to chemical crosslinking with

formaldehyde, the levels of both monomer and potential higher-order bands decreased.

This suggests that SialT2-HA became highly crosslinked to many other species and the

33

subsequent complex that was formed was insoluble. When I utilized

dithiobis(succinimidyl propionate) (DSP) to perform crosslinking experiments, I

observed significant higher-order oligomerization occurring, corresponding to masses

over 250 kDa. Unfortunately, I was unable to separate out these bands and identify their

exact composition. Since SDS-PAGE can be prone to the dissociation of transmembrane

proteins, I examined the lysates of SialT2-HA transfected CHO cells with perfluoro-

octanoic acid-PAGE (PFO-PAGE). This method has proven to be effective for analyzing

transmembrane proteins in their native quaternary structure.27,28 Using this technique, I

was again able to observe homo-oligomerization of SialT2 (Figure 8B). Since the KDEL

recruitment assay demonstrated a potential interaction occurring between two different

forms of the SialT2 enzyme, I tested whether co-transfection of SialT2-HA and SialT2-

GFP would reproduce these results. While I observed the presence of potential higher-

order oligomers of SialT2-GFP (and SialT2-YFP), I did not see any evidence of a

complex between SialT2-HA and SialT2-GFP by WB analysis (Figure 8C). My data

point toward the formation of a stable SialT2 self-oligomerizing product but further

analysis is required to determine its molecular basis.

34

Figure 8. Full-length SialT2 enzyme shows homo-oligomerization. Protein extracts from CHO cells transiently transfected with SialT2-HA and/or SialT2-GFP/YFP were separated by SDS-PAGE or PFO-PAGE (when noted) and analyzed by Western Blot, as described in methods. (A) SialT2-HA (MW ~ 50 kDa) shows several bands potentially corresponding to self-oligomerization. Treatment of samples with formaldehyde to induce crosslinking of SialT2-HA decreases both monomer band and all higher order bands. Treatment of samples with DSP to induce crosslinking of SialT2-HA also decreases monomer band and generates high molecular weight species. (B) PFO-PAGE analysis of SialT2-HA self-oligomerization shows potential higher-order species. (C) Transfection of SialT2-GFP produces several higher-order species. Co-transfection of SialT2-HA with SialT2-GFP does not show the production of any new higher order species.

Discussion

The TM domains of ganglioside glycosyltransferases fail to demonstrate the formation of

hetero-oligomeric species with the SN-TM assay

My initial investigations were focused toward defining the role that TM domains

play in the formation of several reported enzyme complexes of ganglioside

glycosyltransferases.5-8 My studies were performed using an established dimerization

assay that has been reported to accurately display biologically relevant TM domain

interactions.14 While this assay has proved to be robust for analyzing homo-oligomer

35

formation of TM domains,16,18-20 I did not detect any hetero-oligomer formation between

the individual TM sequences (Figure 3). My attempts to optimize the experiment did not

increase the presentation of hetero-oligomers. The investigation of these interactions

may require more sophisticated biochemical methods, such as analytical ultra

centrifugation (AUC) or multiple angle laser light scattering (MALLS).

The potential of this assay to find these interactions may be hindered for several

reasons. First, the purified SN-TM proteins may be segregated into individual SDS

micelles that are unable to mix with one another on the time scale of my experiment.

Attempting to co-express two or more SN-TM constructs at the same time might be a

possible technique to alleviate this problem. Second, and more importantly, while I am

focusing directly on the TM domain, the initial studies investigating physical associations

among ganglioside glycosyltransferases either used full-length enzymes5 or N-terminal

fragments containing the cytoplasmic tail and part of the stem region.6-8 The cytoplasmic

tails of glycosyltransferases have shown to control their localization within the secretory

pathway,29 including those involved with ganglioside biosynthesis.7,9,30 The amino acid

composition of cytoplasmic tails is also known to be a critical recognition element for

several chaperones of ganglioside glycosyltransferases, including calcium-binding

proteins31 and components of COPII vesicles.32 Stem regions have also been shown to

play critical roles in heterodimer formation of glycosyltransferases.33 The absence of

these elements in my in vitro method may hinder the observation of these enzyme

complexes, requiring more cell-based assays.

The TM domains of several ganglioside glycosyltransferases mediate formation of homo-

oligomers

While I was unable to observe hetero-oligomer formation between the TM

domains of reported ganglioside glycosyltransferase complexes, several of these enzymes

displayed the ability to form homo-oligomer complexes with the SN-TM assay: GalT1

and SialT1 formed stable dimeric complexes and SialT2 formed several higher-order

complexes (Figure 3). When I further investigated the molecular basis for GalT1

dimerization, I found that their associations were due to the presence of properly space

polar amino acids within the hydrophobic sequence; both double mutants S23A/S26A

36

and S25A/C28A showed significant diminution of the observed dimer (Figure 4). This

observation agrees with the report that spacing polar residues three amino acids apart,

thus putting them directly onto the same face of an α-helix, can mediate interactions

between TM domains.18 While the TM domain of GalT1 posses a sequence of polar

amino acids (S23XXS26S27XXY30) similar to a reported dimerization motif

(SXXSSXXT),16 performing the double mutation S27A/Y30F caused GalT1 to form a

slightly stronger dimer than the WT form (Figure 4). This suggests that Y30 is not

involved in the dimerization of GalT1 and potentially inhibits the ability to form a dimer.

When I further investigated the molecular basis for SialT2 oligomerization, I saw

that all four of the polar residues contributed to the observed self-association, with one

residue (C21) being most important (Figure 5). Introduction of the double mutant

S18A/C21A was found to be even more detrimental to SialT2 oligomerization. Cysteine

residues within TM domains have been shown to be critical for the formation of dimeric

species of several glycosyltransferases.21-24 The ability of these cysteines to form

disulfide bonds between enzymes has been shown to be essential for proper cellular

localization and glycan product synthesis. Further work will be needed to determine if

these cysteine residues are critical for these roles.

The ability of hydrophilic amino acids within TM domains to mediate the

formation of homo-oligomers has been well demonstrated and is consistent with my

results. Additionally, there are other biophysical methods that have shown value towards

understanding TM domain interactions in more biologically relevant situations, including

the TOXCAT34 and GALLEX35 assays, and would be suggested as further methods to

fully characterize both homo- and hetero-oligomerization. Overall, I feel confident that

the results seen with GalT1, SialT1, and SialT2 provide evidence that the TM domains

engage as homo-oligomers in SDS micelles. However, these results do not provide

information about oligomerization in a more native membrane context.

37

SialT1 and SialT2 enzymes are physically associated in vivo

To further investigate the potential homo-oligomeric complexes of ganglioside

glycosyltransferases observed by the in vitro method, I utilized an established KDEL-

recruitment assay (illustrated in Figure 6) that has been successful in identifying physical

associations between enzymes involved in N-linked glycan25,26 and poly-N-

acetyllactosamine synthesis.12 To confirm the ability of this assay to function with

ganglioside glycosyltransferases, I demonstrated that SialT2-myc-KDEL can recruit

SialT1-HA into the ER, confirming previous observations with these two enzymes.7,8

When I investigated homo-oligomer formation, I observed several of the ganglioside

glycosyltransferases were able to physically associate together inside of cells: SialT1

demonstrated self-association in ~50% of cells co-expressing both constructs (Figure 7B)

while SialT2 had a much higher level of self-association, being present in ~90% of cells

successfully co-expressing both constructs (Figure 7C). I did not observe any KDEL-

dependent relocalization occurring with transfected GalT1 constructs (Figure 7A). While

these results do not correlate with those found in the SDS-PAGE assay, it is plausible that

the observed TM dimer of GalT1 may be much easier to reproduce within the SN-TM

assay since I am excluding the contributions of all other regions of the enzyme.

However, since it has been shown that the N-terminal regions of ganglioside

glycosyltransferases are able to facilitate hetero-oligomer formation,6-8 I believe that their

TM domains play an undetermined role in facilitating the observed self-oligomerization

interactions observed with SialT1 and SialT2.

SialT2 demonstrates the formation of higher-order complexes in cells

When SialT2-HA was transfected into CHO cells, I was able to observe the

formation of homo-oligomeric species, including the formation of a dimer, by SDS-

PAGE (Figure 7A, C) and PFO-PAGE (Figure 7B). These results were obtained without

the use of crosslinking reagents and performed under reducing conditions (DTT). SialT2

has been observed by other groups to form a dimeric species in transfected CHO-K136

and F-11A5 cells but only under non-reducing conditions. My observations suggest that

oligomerization is not entirely controlled by disulfide bond formation since all

experiments were performed under reducing conditions. My attempts to capture these

38

complexes with common chemical crosslinking reagents provided mixed results: the use

of formaldehyde caused an overall decrease in signal without increased formation of high

molecular weight complexes, while DSP generated large molecular weight complexes

(greater than 250 kDa) that could not be resolved by SDS-PAGE (Figure 7A). These

reagents are likely causing SialT2 to form complexes with other cellular proteins and

molecules that may or may not be physiologically relevant to their actual

oligomerization. The use of crosslinking reagents that are able to trap TM interactions

directly would prove to be more relevant for these studies. One such possibility would be

to introduce photoactivatable amino acids into the transmembrane domain, such as photo-

leucine and photo-methionine, which have been used to capture protein-protein

interactions in living cells.37 Additionally, it would be very interesting to see if the SialT2

TM domain mutations, primarily S18A and C21A, would disrupt the self-association of

full length SialT2 observed in the KDEL recruitment assay and SDS-PAGE/WB

analyses.

In conclusion, I present preliminary evidence that several TM domains of

ganglioside glycosyltransferases are able to self-associate into oligomeric species both in

vitro and inside of cells. The primary molecular basis for these interactions is dependent

upon the presence of properly spaced hydrophilic amino acids. Future work is needed to

investigate the functional consequences of these associations within a cellular

environment, including effects on subcellular localization and ganglioside product

formation.

Acknowledgements

I would like to thank Don Engelman, Jennifer Czlapinski, and Suzanne Pfeffer for

their reagent gifts. I would like to thank Marguerite M. Desko and Peter L. Lee for help

helpful suggestions on microscopy and WB analyses. I would like to thank Steve Kaiser

for helpful advice and assistance with MALDI-TOF-MS analysis. I would like to thank

Kitty Lee, Cell Sciences Imaging Facility at Stanford University, for assistance with

confocal microscopy.

39

Methods

Reagents

All chemicals and reagents were purchased from Fisher Scientific (Waltham,

MA) or Sigma-Aldrich (St. Louis, MO) unless noted. Restriction enzymes, T4 DNA

Ligase, calf intestinal phosphatase were purchased from New England Biolabs (Ipswich,

MA). Mouse anti-myc antibody was purchased from the Development Studies

Hybridoma Bank (Iowa City, IA). Mouse anti-GM130 was purchased from BD

Biosciences (San Jose, CA). Mouse anti-p115 was as a generous gift of Suzanne Pfeffer

(Stanford University School of Medicine, Stanford, CA). Rabbit anti-giantin was

purchased from Covance (Princeton, NJ). Rabbit anti-calreticulin was purchased from

Stressgen (Ann Arbor, MI). Rabbit anti-HA was purchased from Abcam (Cambridge,

MA). Rat anti-HA was purchased from Roche Applied Science (Indianapolis, IN). Goat

anti-calnexin was purchased from Santa Cruz Biotechnology Inc (Santa Cruz, CA). Goat

anti-rat HRP, Goat anti-mouse Alexa 488, Goat anti-mouse Alexa 610, Goat anti-rabbit

Alexa 488, Goat anti-rabbit Alexa 610, and Goat anti-donkey Alexa 488 were purchased

from Invitrogen (Carlsbad, CA). Vectashield Mounting Medium with DAPI was

purchased from Vector Labs (Burlingame, CA).

Cloning of SN-TM constructs

pT7SN/linker and pT7SN/GpA-TM were gifts of Don Engelman (Yale

University, New Haven, CT). pT7SN/GalNAcT-TM and pT7SN/GalT2 were gifts of

Jennifer Kohler (UT Southwestern Medical Center, Dallas, TX). pCR-Blunt II-

TOPO/GalT1-HA and pcDNA3.1-Zeo/SialT2 were gifts of Jennifer Czlapinski

(University of California, Berkeley, CA). All plasmids encoding staphylococcal

nuclease-transmembrane chimeras (SN-TM) were constructed from pT7SN/linker.

Putative transmembrane sequences were identified and selected for cloning using Kyte-

Doolittle hydropathy analysis.13 PCR amplification primers for cloning were acquired

from the Stanford University Protein and Nucleic Acid Facility and are listed in Table 1.

The transmembrane sequence of GalT1 (β4GalT6, amino acids 15-40) was

amplified from the plasmid pCR-Blunt II-TOPO/GalT1-HA by PCR using primers

40

GalT1-TM 5’ and GalT1-TM 3’. The transmembrane sequence of SialT1 (ST3Gal5,

amino acids 6-31) was generated synthetically by PCR first using primers SialT1-TM 1st

Primer 5’ and SialT1-TM 1st Primer 3’, followed by using primers SialT1-TM 2nd Primer

5’ and SialT1-TM 2nd Primer 3’. The transmembrane sequence of SialT2 (ST8Sia1,

amino acids 30-53) was amplified from the plasmid pcDNA3.1-Zeo/SialT2 by PCR using

primers SialT2-TM 5’ and SialT2-TM 3’. The PCR products were cloned into the

plasmid pCR-Blunt II-TOPO using the Zero Blunt TOPO PCR Cloning Kit (Invitrogen),

whose identity was confirmed by sequencing. The inserts were excised using AvrII and

BamH1, then ligated into pT7SN/Linker that had been digested with the same enzymes

and treated with calf intestinal phosphatase. Successful ligation was confirmed by

restriction digest. The plasmids were named SN-GalT1 TM, SN-SialT1 TM and SN

SialT2-TM, respectively.

Table 1. Table of primers used to generate SN-TM plasmids. Primer names and sequences, written from 5’ to 3’, are listed.

Mutation of TM inserts

Site-directed mutagenesis was performed according to the Stratagene QuikChange

Site-Directed Mutagenesis Kit (La Jolla, CA). PCR amplification primers for cloning

were acquired from the Stanford University Protein and Nucleic Acid Facility.

Single mutations on wild-type SN-GalT1 TM were performed using the primers

listed in Table 2. Double mutations to generate S23A/S26A (using SN-GalT1 S23A),

S25A/C28A (using SN-GalT1 C28A), and S27A/Y30F (using SN-GalT1 S27A) were

performed using the primers listed in Table 2.

41

Table 2. Table of primers used to generate SN-GalT1 TM mutants. Primer names and sequences, written from 5’ to 3’, are listed.

Single mutations on wild-type SN-SialT2 TM were performed using the primers

listed in Table 3. Double mutations to generate S18A/C21A (using SN-SialT2 C21A)

and C26A/Y29F (using SialT2-TM C26A) were performed using the primers listed in

Table 3.

Table 3. Table of primers used to generate SN-SialT2 TM mutants. Primer names and sequences, written from 5’ to 3’, are listed.

42

Purification of SN-TM constructs

pT7SN/constructs were transformed into BL21(DE3) competent cells to express

the SN-TM fusion proteins. Individual colonies were taken and grown overnight in LB

media. These cultures were used to inoculate new LB cultures at a dilution of 1:100,

growing at 37 oC. After reaching an OD600 = 0.6, IPTG was added to a final

concentration of 1.0 mM and grown for two additional hours. Cells were harvested by

centrifugation for 20 min at 5,000 g, 4 oC. Cells were resuspended in 1/20 culture

volume of 50 mM Tris-HCl (pH 8.0), 5 mM EDTA (pH = 8.0), 1 mM PMSF, 1 mM

DTT. Cells were lysed with three rounds of freeze-thaw cycles. The lysate was clarified

by centrifugation for 45 min @ 20,000 g, 4 oC. To remove unwanted contaminants, the

remaining pellet was then washed first with a solution of 25 mM Tris-base (pH = 8.0), 1

M NaCl, 1 mM EDTA, 1 mM PMSF, 1 mM DTT followed by 25 mM Tris-base (pH =

8.0), 25 mM NaCl, 2 % Thesit, 1 mM EDTA, 1 mM PMSF. The pellet was washed for

three hours with each solution and clarified by centrifugation for 45 min @ 20,000 g, 4 oC. SN-TM chimeras were extracted using a solution of 25 mM Tris-HCl, 1 M NaCl, 1

mM EDTA, 1 mM PMSF, 1 mM DTT, 4.4 M urea and clarified by centrifugation. The

remaining supernatant was concentrated using an Amicon Ultra-15 centrifugal device,

5,000 MWCO (Millipore, Billerica, MA) and the buffer was exchanged with the

following of 25 mM Tris-base (pH = 8.0), 200 mM NaCl, 1 mM EDTA, 1 mM PMSF, 1

mM DTT.

Further purification was achieved using reverse phase-HPLC. Samples were

filtered with a 0.22 µM syringe filter and injected onto an Alltech Apollo C18u, 250 x 4.6

mm column. Samples were eluted using a gradient of 0 – 75 % Buffer B at a flow rate of

1.00 mL/min with Buffer A as 99.9% H2O, 0.1% TFA and Buffer B as 99.9%

acetonitrile, 0.1% TFA. Collected peaks were verified by MALDI-TOF-MS analysis

before rotary evaporation of solvent in Savant SpeedVac (ThermoFisher Scientific,

Waltham, MA). Expected masses of SN-TM constructs: GalT1 – 21383.67 Da, SialT1 –

21588.06, SialT2 – 21207.63, GalT2 – 21130.51, GalNAcT – 21113.34. All SN-TM

constructs (WT and mutants) displayed masses within 1% of expected mass. Dried

protein samples were redissolved using 25 mM Tris-base (pH = 8.0), 200 mM NaCl, 1

43

mM EDTA, 1 mM PMSF, 1 mM DTT, 2.0% Thesit. Sample concentrations were

determined using the BCA® Protein Assay Kit (Pierce, Rockford, IL).

SDS-PAGE analysis of SN-TM oligomerization

Purified SN-TM samples were mixed 1:1 with 2x SDS-PAGE loading buffer (100

mM Tris-base [pH = 6.8], 20 % glycerol, 4 % SDS, 0.2 % bromophenol blue), boiled for

10 min and separated using 12 % Tris-HCl Ready Gels (Bio-Rad, Hercules, CA). For

hetero-oligomerization experiments, samples were mixed together for 30 minutes at room

temperature before loading. Molecular weights were confirmed using Benchmark

Protein Ladder (Invitrogen). Gels were stained with either Coomassie Brilliant Blue (45

% MeOH, 45% ddH2O, 10% glacial acetic acid, 0.0025% Coomassie Brilliant Blue

R250) or SimplyBlue SafeStain (Invitrogen). Imaging of gels was achieved using an

Alpha Innotech FluorChem HD2 (Santa Clara, CA) and quantification of bands was

performed using ImageJ software (NIH, Bethesda, MD).

Cloning of full length ganglioside glycosyltransferases genes

pcDNA3.1(+)/GalT1-HA, pcDNA3.1(+)/SialT1-myc, and pcDNA3.1-Zeo/SialT2

and were generous gifts of Jennifer Czlapinski (University of California, Berkeley, CA).

All GalT1, SialT1, and SialT2 inserts were cloned with a C-terminal affinity tag (myc or

HA) and with a KDEL tag (when applicable), followed by a stop codon. The amino acid

sequences for the C-terminal epitope tags are: myc tag – EQKLISEEDL, HA tag –

YPYDVPDYA, KDEL tag – SEKDEL. PCR amplification primers for cloning were

acquired from the Stanford University Protein and Nucleic Acid Facility and are listed in

Table 4.

GalT1-myc was produced by PCR from the plasmid pCR-Blunt II-TOPO/GalT1-

HA as a template using the primers GalT1-myc 5’ and GalT1-myc 3’. GalT1-myc-

KDEL was produced by PCR from the plasmid pcDNA3.1(+)/GalT1-myc as a template

using primers GalT1-myc and myc-KDEL 3’. GalT1-HA-KDEL was produced by PCR

from the plasmid pCR-Blunt II-TOPO/GalT1-HA as a template using primers GalT1-myc

5’ and GalT1-HA-KDEL 3’. The PCR products were cloned into the plasmid

pCR4Blunt-TOPO using the Zero Blunt TOPO PCR Cloning Kit (Invitrogen), whose

44

identity was confirmed by sequencing. The inserts were excised using NheI and EcoRI,

then ligated into pcDNA3.1-Zeo that had been digested with the same enzymes and

treated with calf intestinal phosphatase. Plasmids containing the inserts were identified

by restriction digest. The plasmids were named GalT1-myc, GalT1-myc-KDEL, and

GalT1-HA-KDEL, respectively.

SialT1-HA was produced by PCR from the plasmid pcDNA3.1(+)/SialT1-myc as

a template using primers SialT1-HA 5’ and SialT1-HA 3’. SialT1-myc-KDEL was

produced by PCR from the plasmid pcDNA3.1(+)/SialT1-myc as a template using

primers SialT1-HA 5’ and myc-KDEL 3’. SialT1-HA-KDEL was produced by PCR

from the plasmid pcDNA3.1(+)/SialT1-HA as a template using primers SialT1-HA 5’

and SialT1-HA-KDEL 3’. The PCR products were cloned into the plasmid pCR4Blunt-

TOPO using the Zero Blunt TOPO PCR Cloning Kit (Invitrogen), and their identity was

confirmed by sequencing. The inserts were excised using NheI and EcoRI, then ligated

into pcDNA3.1-Zeo that had been digested with the same enzymes and treated with calf

intestinal phosphatase. Positive ligation was tested by restriction digest. The plasmids

were named SialT1-HA, SialT1-myc-KDEL, and SialT1-HA-KDEL, respectively.

SialT2-myc was produced by PCR from the plasmid pcDNA3.1-Zeo/SialT2 as a

template using primers SialT2-myc/HA 5’ and SialT2-myc 3’. SialT2-HA was produced

by PCR from the plasmid pcDNA3.1-Zeo/SialT2 as a template using primers SialT2-

myc/HA 5’ and SialT2-HA 3’. SialT2-myc-KDEL was produced by PCR from the

plasmid pcDNA3.1/SialT2-myc as a template using primers SialT2-myc/HA 5’ and

SialT2-myc-KDEL 3’. SialT2-HA-KDEL was produced by PCR from the plasmid

pcDNA3.1/SialT2-HA as a template using primers SialT2-myc/HA 5’ and SialT2-HA-

KDEL 3’. The PCR products were cloned into the plasmid pCR4Blunt-TOPO using the

Zero Blunt TOPO PCR Cloning Kit (Invitrogen) and confirmed by sequencing. The

inserts were excised using NheI and EcoRV, then ligated into pcDNA3.1-Zeo that had

been digested with the same enzymes and treated with calf intestinal phosphatase.

Positive ligation was tested by restriction digest. The plasmids were named SialT2-myc,

SialT2-HA, SialT2-myc-KDEL and SialT2-HA-KDEL, respectively.

45

Table 4. Table of primers used to generate KDEL recruitment assay plasmids. Primer names and sequences, written from 5’ to 3’, are listed.

Immunofluorescence analysis

One day prior to transfection, CHO cells were seeded at a density of 1 x 105

cells/mL in individual wells of a 6-well plate containing cover slips (10 cm2 surface

area/well). Cells were transfected using Lipofectamine 2000 Reagent (Invitrogen)

according to the manufacturers directions, using Opti-MEM (Invitrogen) as media during

the transfection. After exposure to transfection for four hours, fresh media was added

and the cells were cultured for 48 hours before harvesting.

Cells were fixed with 3.7% paraformaldehyde/PBS for 20 minutes without

washing. After several washes with PBS, the cells were permeabilized with 0.1% Triton

X-100, 1.0% BSA/PBS for 5 minutes without shaking. After several washes with PBS,

the plates were placed at RT and incubated with 1.0% BSA/PBS for 1 hour to block non-

specific binding. After blocking, the cells were incubated with primary antibody (1:500

dilution) for 2 hours at RT. The coverslips were then washed 3x with PBS and the cells

were incubated with secondary antibody (1:500 dilution) for 1 hour at RT. The

coverslips were washed again 3x with PBS. After washing, the coverslips were mounted

onto glass cover slides using Vectashield Mounting Medium with DAPI and sealed with

nail polish. Slides were stored at 4 oC until analysis. Cells were visualized using the

HCX PL APO 63x 1.32-0.60 oil objective of a Leica SP2 AOBS Confocal Laser

Scanning Microscopy equipped with 405 nm, 488 nm, and 594 nm lasers. Image analysis

was performed using Adobe Photoshop and Adobe Illustrator.

46

Western Blot Analyses of SialT2

One day prior to transfection, CHO cells were seeded at a density of 1 x 105

cells/mL in 5 mL of media in 6 cm dishes. Cells were transfected using Lipofectamine

2000 Reagent according to the manufacturers directions, using Opti-MEM as media

during the transfection. After exposure to transfection for four hours, fresh media was

added and the cells were cultured for 48 hours before harvesting. Cells were washed

three times with cold PBS and lysed using 0.5 mL of lysis buffer (50 mM Tris, pH = 7.2,

300 mM NaCl, 1.0 % Triton X-100, 1 mM DTT, 1 mM PMSF) with protease inhibitor

cocktail (antipain, turkey trypsin inhibitor, leupeptin, aprotinin) added for 30 min at 4 oC.

Formaldehyde crosslinking studies were performed directly after the 48 hour

incubation period. Cells were washed 3x with cold PBS and incubate with 0.5 % - 1.0 %

formaldehyde solution (in ddH2O) for 10 min at room temperature. The reaction was

quenched with 0.75 M glycine, incubating for an additional 10 min at room temperature.

Cells were again washed three times with cold PBS and lysed as described above. DSP

crosslinking studies were performed on harvested cell lysates. DSP (Pierce) was added at

0.0 mM, 0.12 mM, 0.25 mM, 0.62 mM and 1.24 mM concentrations to 30 µL aliquots of

lysates and incubated for 20 minutes at room temperature. The reaction was quenched

with 0.75 M glycine, incubating for an additional 10 min at room temperature before

analysis by SDS-PAGE.

For SDS-PAGE analysis, samples were mixed with an equal volume of 2x SDS-

PAGE loading buffer (100 mM Tris-base, pH = 6.8, 20 % glycerol, 4 % SDS, 0.2 %

bromophenol blue), boiled for 10 min at 95 oC and separated using 12 % Tris-HCl Ready

Gels (Bio-Rad, Hercules, CA). PFO-PAGE of SialT2 samples was performed as

described previously.28 Briefly, samples were mixed 1:1 with 2x PFO-PAGE loading

buffer (100 mM Tris-base [pH = 8.0], 20 % glycerol, 2 % PFO [pH = 8.0], 0.005%

bromophenol blue), vortexed briefly, centrifuged for 5 min at 10,000g. 7 % Tris-HCl

Ready Gels were pre-run for 60 min at 50 V (constant voltage), 4 oC before loading

samples. Electrophoresis was performed in the cold room at 140 V (constant voltage) for

2 hours using PFO-PAGE running buffer (25 mM Tris, 192 mM, 0.25% PFO at pH =

8.5), pre-chilled at 4 oC.

47

Proteins were transferred to nitrocellulose membrane and stained briefly with

Ponceau S Staining (0.1 % Ponceau S in 5 % acetic acid) to verify the presence of

transferred proteins. After rinsing with Tris-Buffered Saline with 0.1 % Tween-20

(TBST), the membrane was blocked in 5 % nonfat dry milk in TBST for one hour at

room temperature before incubation with rat anti-HA antibody (1:1250 dilution)

overnight at 4 oC. After rinsing with TBST, the membrane was blocked again with TBST

containing 5 % nonfat dry milk for hour, followed by incubation with goat anti-rat HRP

(1:5000 dilution) for one hour at room temperature. The membrane was developed using

SuperSignal West Pico chemiluminescence substrate (Pierce) and imaged on Pierce CL-

XPosure Film (Pierce). Imaging of gels was achieved using an Alpha Innotech

FluorChem HD2 (Santa Clara, CA).

References

1 Lairson, L. L., Henrissat, B., Davies, G. J. & Withers, S. G. Glycosyltransferases:

structures, functions, and mechanisms. Annu Rev Biochem 77, 521-555, (2008).

2 Roth, J. Localization of glycosylation sites in the Golgi apparatus using

immunolabeling and cytochemistry. J Electron Microsc Tech 17, 121-131,

(1991).

3 Roseman, S. The synthesis of complex carbohydrates by multiglycosyltransferase

systems and their potential function in intercellular adhesion. Chem Phys Lipids 5,

270-297, (1970).

4 de Graffenried, C. L. & Bertozzi, C. R. The roles of enzyme localisation and

complex formation in glycan assembly within the Golgi apparatus. Curr Opin

Cell Biol 16, 356-363, (2004).

5 Bieberich, E. et al. Regulation of ganglioside biosynthesis by enzyme complex

formation of glycosyltransferases. Biochemistry 41, 11479-11487, (2002).

6 Giraudo, C. G., Daniotti, J. L. & Maccioni, H. J. Physical and functional

association of glycolipid N-acetyl-galactosaminyl and galactosyl transferases in

the Golgi apparatus. Proc Natl Acad Sci U S A 98, 1625-1630, (2001).

48

7 Giraudo, C. G. & Maccioni, H. J. Ganglioside glycosyltransferases organize in

distinct multienzyme complexes in CHO-K1 cells. J Biol Chem 278, 40262-

40271, (2003).

8 Uliana, A. S., Crespo, P. M., Martina, J. A., Daniotti, J. L. & Maccioni, H. J.

Modulation of GalT1 and SialT1 sub-Golgi localization by SialT2 expression

reveals an organellar level of glycolipid synthesis control. J Biol Chem 281,

32852-32860, (2006).

9 Uliana, A. S., Giraudo, C. G. & Maccioni, H. J. Cytoplasmic tails of SialT2 and

GalNAcT impose their respective proximal and distal Golgi localization. Traffic

7, 604-612, (2006).

10 Bormann, B. J. & Engelman, D. M. Intramembrane helix-helix association in

oligomerization and transmembrane signaling. Annu Rev Biophys Biomol Struct

21, 223-242, (1992).

11 Senes, A., Engel, D. E. & DeGrado, W. F. Folding of helical membrane proteins:

the role of polar, GxxxG-like and proline motifs. Curr Opin Struct Biol 14, 465-

479, (2004).

12 Lee, P. L., Kohler, J. J. & Pfeffer, S. R. Association of beta-1,3-N-

acetylglucosaminyltransferase 1 and beta-1,4-galactosyltransferase 1, trans-Golgi

enzymes involved in coupled poly-N-acetyllactosamine synthesis. Glycobiology

19, 655-664, (2009).

13 Kyte, J. & Doolittle, R. F. A simple method for displaying the hydropathic

character of a protein. J Mol Biol 157, 105-132, (1982).

14 Lemmon, M. A. et al. Glycophorin A dimerization is driven by specific

interactions between transmembrane alpha-helices. J Biol Chem 267, 7683-7689,

(1992).

15 Stanley, A. M. & Fleming, K. G. The transmembrane domains of ErbB receptors

do not dimerize strongly in micelles. J Mol Biol 347, 759-772, (2005).

49

16 Dawson, J. P., Weinger, J. S. & Engelman, D. M. Motifs of serine and threonine

can drive association of transmembrane helices. J Mol Biol 316, 799-805, (2002).

17 Gratkowski, H., Lear, J. D. & DeGrado, W. F. Polar side chains drive the

association of model transmembrane peptides. Proc Natl Acad Sci U S A 98, 880-

885, (2001).

18 Sal-Man, N., Gerber, D. & Shai, Y. The identification of a minimal dimerization

motif QXXS that enables homo- and hetero-association of transmembrane helices

in vivo. J Biol Chem 280, 27449-27457, (2005).

19 Zhou, F. X., Cocco, M. J., Russ, W. P., Brunger, A. T. & Engelman, D. M.

Interhelical hydrogen bonding drives strong interactions in membrane proteins.

Nat Struct Biol 7, 154-160, (2000).

20 Zhou, F. X., Merianos, H. J., Brunger, A. T. & Engelman, D. M. Polar residues

drive association of polyleucine transmembrane helices. Proc Natl Acad Sci U S A

98, 2250-2255, (2001).

21 Aoki, D., Lee, N., Yamaguchi, N., Dubois, C. & Fukuda, M. N. Golgi retention of

a trans-Golgi membrane protein, galactosyltransferase, requires cysteine and

histidine residues within the membrane-anchoring domain. Proc Natl Acad Sci U

S A 89, 4319-4323, (1992).

22 Yamaguchi, N. & Fukuda, M. N. Golgi retention mechanism of beta-1,4-

galactosyltransferase. Membrane-spanning domain-dependent homodimerization

and association with alpha- and beta-tubulins. J Biol Chem 270, 12170-12176,

(1995).

23 Qian, R., Chen, C. & Colley, K. J. Location and mechanism of alpha 2,6-

sialyltransferase dimer formation. Role of cysteine residues in enzyme

dimerization, localization, activity, and processing. J Biol Chem 276, 28641-

28649, (2001).

50

24 Sousa, V. L. et al. Importance of Cys, Gln, and Tyr from the transmembrane

domain of human alpha 3/4 fucosyltransferase III for its localization and sorting

in the Golgi of baby hamster kidney cells. J Biol Chem 278, 7624-7629, (2003).

25 Munro, S. An investigation of the role of transmembrane domains in Golgi

protein retention. Embo J 14, 4695-4704, (1995).

26 Nilsson, T. et al. Kin recognition between medial Golgi enzymes in HeLa cells.

Embo J 13, 562-574, (1994).

27 Neve, E. P., Lahtinen, U. & Pettersson, R. F. Oligomerization and interacellular

localization of the glycoprotein receptor ERGIC-53 is independent of disulfide

bonds. J Mol Biol 354, 556-568, (2005).

28 Ramjeesingh, M., Huan, L. J., Garami, E. & Bear, C. E. Novel method for

evaluation of the oligomeric structure of membrane proteins. Biochem J 342 (Pt

1), 119-123, (1999).

29 Fenteany, F. H. & Colley, K. J. Multiple signals are required for alpha2,6-

sialyltransferase (ST6Gal I) oligomerization and Golgi localization. J Biol Chem

280, 5423-5429, (2005).

30 Uemura, S., Yoshida, S., Shishido, F. & Inokuchi, J. The cytoplasmic tail of GM3

synthase defines its subcellular localization, stability, and in vivo activity. Mol

Biol Cell 20, 3088-3100, (2009).

31 Quintero, C. A., Valdez-Taubas, J., Ferrari, M. L., Haedo, S. D. & Maccioni, H. J.

Calsenilin and CALP interact with the cytoplasmic tail of UDP-

Gal:GA2/GM2/GD2 beta-1,3-galactosyltransferase. Biochem J 412, 19-26,

(2008).

32 Giraudo, C. G. & Maccioni, H. J. Endoplasmic reticulum export of

glycosyltransferases depends on interaction of a cytoplasmic dibasic motif with

Sar1. Mol Biol Cell 14, 3753-3766, (2003).

51

33 Nilsson, T., Rabouille, C., Hui, N., Watson, R. & Warren, G. The role of the

membrane-spanning domain and stalk region of N-acetylglucosaminyltransferase

I in retention, kin recognition and structural maintenance of the Golgi apparatus in

HeLa cells. J Cell Sci 109 ( Pt 7), 1975-1989, (1996).

34 Russ, W. P. & Engelman, D. M. TOXCAT: a measure of transmembrane helix

association in a biological membrane. Proc Natl Acad Sci U S A 96, 863-868,

(1999).

35 Schneider, D. & Engelman, D. M. GALLEX, a measurement of heterologous

association of transmembrane helices in a biological membrane. J Biol Chem 278,

3105-3111, (2003).

36 Daniotti, J. L., Martina, J. A., Giraudo, C. G., Zurita, A. R. & Maccioni, H. J.

GM3 alpha2,8-sialyltransferase (GD3 synthase): protein characterization and sub-

golgi location in CHO-K1 cells. J Neurochem 74, 1711-1720, (2000).

37 Suchanek, M., Radzikowska, A. & Thiele, C. Photo-leucine and photo-methionine

allow identification of protein-protein interactions in living cells. Nat Methods 2,

261-267, (2005).

52

Chapter 3 – Covalently capturing the ganglioside GM1-cholera toxin complex with

photocrosslinking

Introduction

Cell surface glycans are important recognition factors for a large number of cell-

cell and cell-ligand interactions. Many of these binding events are regulated by the

presence or absence of sialic acid.1,2 Sialic acids are a family of nine carbon α-keto acids

that are found terminally attached to glycoproteins and glycosphingolipids.3 While these

molecules serve important regulatory roles for mammalian cells, they are also utilized as

receptors for invasion by many pathogens. For example, influenza viruses utilize

hemagglutinin to bind to cell surface sialic acid.4 Sialylated glycosphingolipids, known as

gangliosides, are the principal recognition elements for the invasion of several viral

pathogens, such as BK,4 SV40,5 and polyoma virus.5 They are also targets of many toxins

produced by bacterial pathogens, including tetanus toxin,6 E. coli heat-labile

enterotoxins,7 and cholera toxin.8

Explicitly determining ganglioside interaction partners and their molecular detail

remains a challenge as most carbohydrate-mediated interactions are characterized by low

affinities and fast off rates. These properties make traditional purification techniques

impractical, as the unstable complex cannot withstand the washing steps. To assist in

capturing these labile complexes, there is a need to develop new photocrosslinking tools

that capture these complexes in their native states. One conceivable solution to this

problem is to introduce photoreactive groups directly onto the glycan structure using

metabolic oligosaccharide engineering techniques. Metabolic oligosaccharide

engineering methodology has been highly successful at introducing a wide range of

chemical modifications into several monosaccharides, including sialic acid.9 Our group

has recently demonstrated the successful incorporation of a diazirine moiety into the cell

surface sialic acid binding lectin CD22 and used this crosslinker to covalently capture

CD22 homo-oligomers.10 In this case, the modification of sialic acid was achieved by

culturing mammalian cells with a diazirine containing N-acetylmannosamine analog

(ManNDAz – Figure 1), which is metabolized by cells and displayed in sialosides as

SiaDAz. The diazirine was selected instead of other crosslinkers because of its small

53

size, long UV activation wavelength, high reactivity, and non-selectivity.11 If positioned

appropriately, the diazirine will introduce a covalent bond between sialic acid and its

interaction partner upon photoirradiation. This covalently linked complex can then be

isolated, purified, and the partners identified.

While our group’s initial studies were able to capture the oligomerization of a

glycoprotein,10 I wanted to investigate the applicability of metabolic oligosaccharide

engineering to install photocrosslinking sialic acids into gangliosides. I chose to use

Jurkat cells for my investigation as they produce several ganglioside structures and

previous work showed their ability to incorporate unnatural sialic acid analogs into cell

surface gangliosides.12 I observed that Ac4ManNDAz-treated Jurkat cells were able to

produce SiaDAz-modified GM2 and GM1 gangliosides. These results demonstrated that

our unnatural analog effectively competed with endogenous sialic acid for addition into

gangliosides. Next, I verified that the modified GM1 structure bearing the

photocrosslinking moiety could be recognized by one of its known-binding partners,

cholera toxin subunit B (CTxB).8 In cooperative experiments performed with Michelle

Bond (UT Southwestern), we have confirmed that GM1-SiaDAz can be successfully

crosslinked to CTxB. To investigate association with GM1 affects CTxB endocytosis

and retrograde trafficking, I performed preliminary experiments exploring the co-

localization of CTxB in several cellular organelles using immunofluorescence

microscopy. My work shows that SiaDAz can efficiently be incorporated into cell

surface gangliosides of Jurkat cells and can be covalently linked to a known binding

partner with high efficiency.

Results

HPTLC analysis of gangliosides produced by Ac4ManNDAz-treated Jurkat cells

To investigate the potential of mammalian cells to incorporate a diazirine-

modified sugar into gangliosides, Jurkat cells were cultured with fully acetylated

ManNDAz (Ac4ManNDAz) for 72 hours. Mammalian cells are able to take up acetylated

ManNAc analogs by passive diffusion, remove the acetyl groups by non-specific

esterases, convert the ManNAc analogs to their sialic acid counterparts, and incorporate

54

the sialic acid analogs into glycoconjugates that are displayed on the cell surface.10,11,13

While ManNAc analogs serve as committed precursors for sialic acid analogs, analogs of

the 2-epimer of ManNAc, GlcNAc, are not efficiently metabolized to their sialic acid

counterparts.13,14 Thus, I chose to use the diazirine-containing analog GlcNDAz (Figure

1) as control molecule that is not expected to be metabolized to SiaDAz. Additionally,

Jurkat cells were cultured with Ac4ManNAc (Figure 1) to show that the exogenous

addition of the normal metabolites in the sialic acid biosynthesis pathway does not affect

ganglioside production. Acetylated sugars were added to the media at a final

concentration of 100 µM. After harvesting the cultured cells, I extracted the gangliosides

using established methods15-17 (described in Methods) and resolved them using high

performance TLC plates (HPTLC). To visualize gangliosides, the HPTLC plates were

stained with resorcinol, which reacts specifically with gangliosides upon heating or

visualized using CTxB immunostaining.

Figure 1. Acetylated analogs used to investigate metabolic oligosaccharide engineering of gangliosides. Jurkat cells were cultured with Ac4ManNDAz or Ac4GlcNDAz to investigate their incorporation into gangliosides. Ac4ManNAc is used as a control molecule to demonstrate that supplementation with exogenous sugar does not affect ganglioside production. Sugars were fully acetylated to improve cellular uptake; acetyl groups are expected to be cleaved inside the cell by non-specific esterases.

Jurkat cells are able to synthesize several different types of gangliosides,

including GM3, GM2, GM1 and GD1a (Figure 2A, B). The appearance of two distinct

bands for each ganglioside species is due the incorporation of different fatty acids into the

ceramide portion of the molecule. When compared to unsupplemented Jurkat cells, the

Ac4ManNDAz-treated cells exhibit significant changes in their ganglioside composition

as evidenced by the mobility changes on the HPTLC plate. Introduction of the diazirine

side chain onto sialic acid increases the hydrophobicity of the ganglioside causing it to

have a higher Rf value. The Ac4ManNDAz-treated cells continue to generate normal

55

gangliosides but at weaker at reduced levels. As expected, culturing Jurkat cells with

Ac4GlcNDAz did not cause any detectable changes in ganglioside production. This

suggests that one or more of the enzymes responsible for metabolizing GlcNDAz into

ManNDAz cannot accept this change to the N-acyl chain. I also observed that the

introduction of exogenous Ac4ManNAc to Jurkat cells does not appear to increase

production of cellular gangliosides.

To confirm the successful modification of Jurkat gangliosides with ManNDAz, I

separated the gangliosides from untreated, Ac4ManNAc-, Ac4ManNDAz-, and

Ac4GlcNDAz-treated Jurkat cells by HPTLC and probed the HPTLC plate for the GM1

ganglioside using a CTxB-Alexa Fluor 488 conjugate (Figure 2C). As a known binding

partner of GM1, CTxB expectedly binds to GM1 in unsupplemented and Ac4ManNAc-

treated Jurkat cell samples. Importantly, CTxB also recognizes additional bands in the

gangliosides from Ac4ManNDAz-treated cells. I hypothesized that these bands

correspond to GM1-SiaDAz. When gangliosides from the Ac4GlcNDAz-treated cells

were examined for binding to CTxB, there was no additional band to indicate the

presence of GM1-SiaDAz gangliosides, further confirming the inability of Jurkat cells to

metabolize GlcNDAz into gangliosides. These results also provide evidence that CTxB

recognition of GM1 is maintained upon installation of the unnatural diazirine side chain

in sialic acid.

56

Figure 2. HPTLC analysis of gangliosides from cultured Jurkat cells. Jurkat cells were cultured for 72 hours untreated or with Ac4ManNAc, Ac4ManNDAz, or Ac4ManNDAz. The cultured cells were harvested and the gangliosides were extracted. The ganglioside composition from each cultured Jurkat sample was resolved by HPTLC and analyzed by resorcinol (B) or CTxB immunostaining (C). (A) Jurkat cells are able to synthesize several different types of gangliosides, including GM3, GM2, GM1 and GD1a. (B) Ac4ManNDAz-treated Jurkat cells display modified ganglioside patterns compared to unsupplemented cells. The modified gangliosides are recognized by their mobility shift on the HPTLC plate caused by the introduction of the SiaDAz molecule. GlcNDAz, the metabolic precursor to ManNDAz, is not metabolized into SiaDAz. Treatment of cells with Ac4ManNAc does not affect ganglioside composition levels. (C) Cholera toxin subunit B is able to recognize GM1-SiaDAz. Cholera toxin subunit B binds GM1 ganglioside and can be used to confirm the presence of GM1. Ac4ManNDAz-treated Jurkat cells display two different sets of GM1 molecules: the top pair representing GM1-SiaDAz and the bottom pair representing endogenous GM1.

57

Mass spectrometry analysis of gangliosides produced by Ac4ManNDAz-treated Jurkat

cells

To confirm the successful modification of Jurkat gangliosides with SiaDAz,

extracted gangliosides were analyzed by mass spectrometry at the Complex Carbohydrate

Research Center (University of Georgia). Samples were crystallized with

trihydroxyacetophenone monohydrate (THAP) matrix and analyzed by MALDI-TOF-MS

in negative ionization mode. As shown in Figure 3, Jurkat cells cultured with

Ac4ManNDAz showed efficient incorporation of SiaDAz into GM2 and GM1

gangliosides. It was also noted that several of the peaks contained SiaDAz modified

structures without the diazirine group present (denoted by stars in Figure 3). The loss of

N2 by this MALDI-TOF-MS was consistent between several analyses (data not shown)

and is believed to occur upon ionization. These results were compared to untreated and

Ac4ManNAc-treated Jurkat samples to determine the endogenous ganglioside

composition (shown in Appendix). As observed in Figure 2B, Ac4GlcNDAz-treated cells

did not show detectable levels of SiaDAz-containing gangliosides (located in Appendix).

These results confirm that Jurkat cells can metabolize Ac4ManNDAz to SiaDAz and

incorporate SiaDAz into gangliosides at significant levels.

58

Figure 3. MALDI-TOF-MS analysis of Jurkat cells cultured with Ac4ManNDAz. Ganglioside extracts of Jurkat cells cultured with Ac4ManNDAz were crystallized with THAP and analyzed by MALDI-TOF-MS. When cultured with Ac4ManNDAz, Jurkat cells are able to incorporate SiaDAz into GM2 and GM1. Unlabeled gangliosides (GM2 – purple, GM1 – green) and SiaDAz-containing gangliosides (GM2 – orange, GM1 – cyan) are denoted above. SiaDAz-containing gangliosides subject to loss of diazirine are denoted by stars (GM2 – orange, GM1 – cyan).

Immunofluorescence microscopy analysis of CTxB trafficking in Ac4ManNDAz-treated

Jurkat cells exposed to UV light

Cholera toxin, a member of the AB5 group of toxins, is secreted by the bacterium

Vibrio cholerae and is responsible for the massive fluid loss that accompanies cholera

infection.18 The toxin is comprised of two distinct subunits: A and B. Using the B

subunit, cholera toxin binds to the host cell through cell surface GM1 with sub-

nanomolar KD.19-22 Upon binding the plasma membrane, cholera toxin is endocytosed and

trafficked through retrograde mechanisms through the endosomes, trans-Golgi network

(TGN), and finally into the endoplasmic reticulum (ER). Once in the ER, the A subunit

is cleaved in two, after which the A1 subunit is transported to the cytosol. Once there,

the A1 subunit activates the heterotrimeric G protein Gs-α, resulting in the activation of

adenylyl cyclase, generating large amounts of cAMP, which induces massive H2O and

chloride secretion by the cells.23

59

After observing that CTxB is able to recognize the unnatural GM1-SiaDAz

ganglioside synthesized by Jurkat cells, I wanted to investigate whether this interaction

would have any effects on the trafficking of CTxB from the cell membrane to the ER.

CTxB has been shown to traffic from the plasma membrane to the ER without being

complexed to the A subunit.24 Cooperative experiments performed in our group have

demonstrated that GM1-SiaDAz expressing Jurkat cells can be efficiently crosslinked to

CTxB (data not shown). Using this knowledge, I examined the intracellular trafficking of

CTxB when complexed directly to GM1 by immunofluorescence microscopy. Jurkat

cells were untreated or cultured with Ac4ManNDAz for 70 hours and then exposed to

CTxB for 10 min to allow for cell surface binding; cells were kept on ice to prevent

endocytosis. The cells were then exposed to 365 nm UV irradiation for 10 minutes to

photoactivate the diazirine into a reactive carbene to covalently bind CTxB to GM1.

Control experiments without photoirradiation were performed with untreated and

Ac4ManNDAz-treated cells. To induce CTxB endocytosis, the cellular media was

replaced with pre-warmed media and incubated for an additional 90 minutes at 37 oC.

The cells were then fixed, permeabilized and immunostained with antibodies against

CTxB and several intracellular organelles. Confocal microscopy was used to determine

subcellular localization of CTxB in the Jurkat cells.

When I analyzed the non-irradiated/unsupplemented Jurkat cells, I observed that

CTxB colocalized with early endosomes, Golgi, and ER organelles (Figure 4A). This

indicates that CTxB is able to traffic from the plasma membrane and into the secretory

pathway without forming a complex with the A subunit. Ac4ManNDAz treatment of

Jurkat cells displayed similar colocalization observations (Figure 4C), indicating that

SiaDAz-engineering of GM1 does not affect CTxB trafficking. When both untreated

(Figure 4B) and Ac4ManNDAz-treated (Figure 4D) cells were photoirradiated, direct

CTxB colocalization was observed for early endosomes and Golgi but not for the ER; I

observed a clear separation occurring outside of the nucleus between CTxB and the ER. I

also noticed that CTxB formed more punctate structures on the outside of cells after UV

exposure. Since this observation was viewed with untreated and Ac4ManNDAz-treated

cells, this suggests that the likely cause of the behavior is UV irradiation, not

crosslinking. Analysis of Jurkat cells by immunofluorescence is further complicated by

60

the immense size of the Jurkat cell nucleus: this makes it difficult to distinguish CTxB

localization between the cell surface and ER. While Jurkat cells are able to synthesize

photoactivatable gangliosides, these results indicate that use of photoirradiation to

crosslink CTxB to GM1-SiaDAz may prevent investigating intracellular trafficking by

immunofluorescence microscopy.

Figure 4. Immunofluorescence analysis of Cholera toxin subunit B trafficking in supplemented Jurkat cells. Jurkat cells, either untreated or Ac4ManNDAz, were cultured with CTxB for 10 minutes on ice. The cells were then photoirradiated for 10 min with 365 nm light. After culturing the cells for 90 minutes at 37oC, CTxB trafficking through several organelles (ER - Calnexin, cis-Golgi – GM130, trans-Golgi network (TGN46), and early endosomes – EEA1) was monitored by immunofluorescence microscopy. Immunostaining for organelle markers is shown in Column 1. Immunostaining for CTxB is shown in Column 2. Column 3 represents the overlay of the images: organelle markers are shown in red, CTxB is shown in green, and the nuclear stain (DAPI) is shown in blue. Colocalization between organelle markers and CTxB appears yellow in the merged image. (A) Untreated Jurkat cells. (B) Untreatedd Jurkat cells exposed to UV light. (C) Ac4ManNDAz-treated Jurkat cells. (D) Ac4ManNDAz-treated cells exposed to UV light.

61

Discussion

Jurkat cells can synthesize gangliosides with photoreactive chemical groups

The primary goal of my studies was to demonstrate successful engineering of

photoactivatable gangliosides in mammalian cells. Metabolic oligosaccharide

engineering is an acknowledged technique to introduce diverse chemical functionality

into cell surface sialosides.9 While our laboratory has demonstrated the successful

incorporation of a diazirine into a sialylated glycoprotein (CD22),10,13 my objective was

to investigate the applicability of this technology into cell surface gangliosides. Using

Jurkat cells, I was able to observe efficient metabolism of ManNDAz into SiaDAz in

several ganglioside species by HPTLC analysis (Figure 2B). Culturing Jurkat cells with

Ac4ManNDAz caused dramatic changes in the composition of endogenous gangliosides;

the levels of unmodified gangliosides diminished and several new gangliosides appeared.

These new species are likely SiaDAz-containing gangliosides as the introduction of the

diazirine side chain increases the hydrophobicity of the molecule, causing a mobility shift

towards the solvent front. To confirm that treatment of Jurkat cells with exogenous

ManNAc does not increase overall ganglioside production, I cultured cells with their

natural substrate, ManNAc, at the same molar concentration as with ManNDAz. As

shown in Figure 2B, cells that were untreated or Ac4ManNAc-treated showed similar

levels of ganglioside production, validating that our culturing techniques are not forcing

an increased production of gangliosides.

To confirm the existence of SiaDAz-containing gangliosides in Jurkat cells, I

submitted the total ganglioside extracts for MALDI-TOF-MS analysis (Figure 3). I

observed that Jurkat cells are able to synthesize GM2-SiaDAz and GM1-SiaDAz. During

these analyses, peaks were observed for several SiaDAz-containing gangliosides that

corresponded to a loss of N2 after irradiation by the instrument laser (Figure 3 – denoted

by stars). This was determined to be an unavoidable consequence of analyzing these

species by MALDI-TOF-MS (Roberto Sonon, personal communication). Regardless of

the loss of diazirine, this technique validated my HPTLC observations that Jurkat cells

were able to synthesize photoactivatable gangliosides.

62

When Jurkat cells were cutlured with Ac4GlcNDAz, there was no apparent

change in the ganglioside production of these cells (Figure 2B). These results were

consistent with MALDI-TOF-MS analysis of the total ganglioside extract (data not

shown). In mammalian cells, GlcNAc is coupled to UDP to generate an activated

nucleotide sugar that is subsequently epimerized by UDP-GlcNAc 2-epimerase into

ManNAc. Theoretically, this synthetic route could be possible for the conversion of

GlcNDAz into ManNDAz; however, my results suggest that this does not occur in Jurkat

cells. It appears that either the UDP-GlcNAc transferase or UDP-GlcNAc 2-epimerase is

unable to accept the introduction of the diazirine side found in GlcNDAz.

Another important aspect of my studies is that ManNDAz competes effectively

with natural Jurkat cell metabolites for incorporation into gangliosides. This was clearly

demonstrated with my HPTLC (Figure 2B, C) and MALDI-TOF-MS (Figure 3) analyses,

where significant levels of unnatural gangliosides were produced. These results are

consistent with previous work performed in our lab showing that ManNDAz can

effectively compete for sialic acid incorporation into the cell surface CD22.10,13

Additionally, I observed that culturing Jurkat cells with Ac4ManNAz, a widely used

substrate for sialic acid metabolic oligosaccharide engineering, can also compete directly

for ganglioside incorporation into Jurkat cells (data not shown).

Cholera toxin subunit B recognizes GM1-SiaDAz and can be efficiently photocrosslinked

I investigated if a well-established binding partner, CTxB, could efficiently

recognize GM1-SiaDAz. For this experiment, I resolved ganglioside extracts from

Ac4ManNDAz-treated Jurkat cells by HPTLC and immunostained against GM1 with a

CTxB-Alexa Fluor 488 conjugate (Figure 2C). The application of immunostaining

techniques for ganglioside analysis on HPTLC plates is well established and is more

sensitive than resorcinol staining.25,26 I observed that CTxB-Alexa Fluor 488 bound to

two additional bands in Ac4ManNDAz-treated Jurkat cells; these bands are likely GM1-

SiaDAz since introducing the diazirine side chain increases their Rf value. The

introduction of the diazirine side chain into GM1 does not appear to dramatically affect

this interaction, suggesting that we should be able to exploit the photoreactivity of the

diazirine to capture this complex. Indeed, cooperative experiments conducted in our

63

laboratory have demonstrated that Ac4ManNDAz-treated Jurkat cells can be efficiently

photocrosslinked to CTxB (data not shown). These results demonstrate that

photoactivatable gangliosides can be used as crosslinking reagents to capture interaction

partners.

The formation of a covalent GM1-CTxB complex does not appear to affect trafficking

from the plasma membrane to the TGN

After demonstrating that CTxB can recognize GM1-SiaDAz synthesized by

Ac4ManNDAz-treated Jurkat cells, I decided to investigate whether the formation of a

covalent complex would affect the retrograde trafficking of CTxB. CTxB has been shown

to traffic from the plasma membrane to the ER without being complexed to the A

subunit.24 I chose to monitor CTxB trafficking in Jurkat cells by performing

colocalization immunofluorescence microscopy using organelle markers for the early

endosomes, TGN, cis-Golgi, and ER. I observed that CTxB was able to efficiently traffic

from the plasma membrane to the ER within the time frame of my experiment, regardless

of culturing with Ac4ManNDAz (Figures 4A, 4C). This result confirms my observations

seen from CTxB-Alexa Fluor 488 immunostaining (Figure 2C). When I exposed the

Jurkat cells to UV irradiation to activate the diazirine, I did not observe any noticeable

changes in CTxB trafficking between unsupplemented and Ac4ManNDAz-treated cells:

CTxB trafficking was limited to the early endosomes and TGN. There was a clear

separation between the fluorescence emanating from CTxB and the ER (Figures 4B, 4D).

In addition, CTxB appeared to distribute into small puncta near the plasma membrane

that did not colocalize with several organelle markers. These results provide preliminary

evidence that photocrosslinking CTxB to GM1 does not affect its ability to traffic from

the plasma membrane toward the TGN. Due to the complications of CTxB trafficking

into the ER because of UV irradiation, I am unable to make any conclusions regarding

the trafficking of covalently formed CTxB-GM1 complexes at this time.

One of the major limitations encountered within this immunofluorescence

analysis is the small size of the cytoplasm in Jurkat cells; most of the intracellular space

is encompassed by the nucleus (Figure 4). The narrow cytosol makes it extremely

difficult to analyze the subcellular distribution of CTxB between the plasma membrane

64

and ER, as these two organelles appear adjacent. Although Jurkat cells provided a useful

starting point for our experiments because of their ganglioside patterning, they have

proven less useful for immunofluorescence experiments. An epithelial cell expressing

GM1 would be the most biologically relevant cell type and might yield a clearer picture

of CTxB trafficking. To this end, our lab is currently exploring the human intestinal

epithelial cell line T84 for its ability to produce SiaDAz-GM1, as it widely used for

studying the intracellular effects of cholera toxin infection.27

In summary, I have demonstrated that a mammalian cell line can produce

photoactivatable gangliosides that are recognized by an established binding partner.

Irradiation of these molecules can efficiently capture this ganglioside-protein interaction.

My results show promise for using this photoreactive tool to capture and investigate

ganglioside-mediated interactions. Future studies are required to investigate the

biological consequences of covalently linked GM1-CTxB in mammalian cells.

Acknowledgements

I would like to thank Kim Orth (UT Southwestern Medical Center) for sharing

Jurkat cells. I would like to thank Michelle R. Bond and Seokho Yu (UT Southwestern

Medical Center) for providing Ac4GlcNDAz. I would like to thank Yan Li (UT

Southwestern Medical Center Protein Chemistry Technology Center) for assistance with

ESI-MS analysis. I would like to thank Pablo Lopez (Johns Hopkins University) and

Aarnoud C. van der Spoel (University of Oxford) for help and advice with ganglioside

extractions. I would like to thank Parastoo Azadi and Roberto Sonon (University of

Georgia Complex Carbohydrate Research Center) for mass spectrometry of ganglioside

samples. Additionally, I would like to thank Michelle R. Bond for helpful suggestions

with my experiments and concurrent analyses involving photocrosslinking of cholera

toxin subunit B to GM1 targets.

65

Methods

Reagents

All chemicals, reagents, and general supplies were purchased from Fisher

Scientific (Waltham, MA) or Sigma-Aldrich (St. Louis, MO) unless noted. N-

hydroxybenzotriazole (HOBt) was purchased from AnaSpec (Fremont, CA). SepPak

tC18 columns (0.3 g) were purchased from Fisher Scientific. HPTLC plates (20 x 20 cm,

glass backed, 200 µm thickness) were purchased from EMD Chemicals (Gibbstown, NJ).

Matreya ganglioside standards 1508, 1510, and 1511 were purchased from Matreya LLC

(Pleasant Gap, MD). Cholera toxin subunit B-Alexa Fluor 488 conjugate was purchased

from Invitrogen (Carlsbad, CA).

Cell culture reagents, including RPMI 1640 with 2 mM L-glutamine were

purchased from Invitrogen (Carlsbad, CA). Cholera toxin B subunit B (from Vibrio

cholerae) was purchased from Sigma Aldrich. Mouse anti-GM130 and mouse anti-EEA1

were purchased from BD Biosciences (San Jose, CA). Chicken anti-cholera toxin subunit

B, rabbit anti-TGN 46, and goat anti-chicken FITC were purchased from Abcam

(Cambridge, MA). Rabbit anti-calnexin (H-70) was purchased from Santa Cruz

Biotechnology, Inc. (Santa Cruz, CA). Goat anti-mouse Alexa Fluor 546, goat anti-rabbit

Alexa Fluor 546, goat anti-mouse Alexa Fluor 633, and goat anti-rabbit Alexa Fluor 633

were purchased from Invitrogen. Vectashield mounting medium with DAPI was

purchased from Vector Labs (Burlingame, CA). 0.1% poly-L-lysine solution was

obtained from Sigma (St. Louis, MO).

Cell culturing conditions

Jurkat cells were grown and maintained in RPMI 1640 with 2 mM L-glutamine

containing 10% heat-inactivated FBS at 37oC, 5% CO2 in a water-saturated environment.

Cells were cultured at 2.0-2.5 x 105 cells/mL in media and grown for 48 hours before

passaging. Typically, cell densities were maintained between 2.5 x 105 cells/mL and 2.0

x 106 cells/ml. Cell viability was analyzed using Trypan blue dye staining with the

Countess Automated Cell Counter instrument (Invitrogen).

66

General information for the chemical synthesis of N-acetylmannosamine analogs

All chemicals were used as received from commercial suppliers without further

purification. 1-hydroxybenzotriazole hydrate was purchased from AnaSpec. All other

chemicals were purchased from Sigma-Aldrich or Fisher Scientific unless otherwise

noted. Reaction progress was monitored by analytical thin layer chromatography (TLC)

on silica gel 60 F254 glass backed plates (Fisher) and stained with ceric ammonium

molybdate. Flash column chromatography was carried out with silica gel 60 (particle

size 40-63 µm, EMD Chemicals). All 1H-NMR and 13C-NMR spectra were recorded on

a Varian 500 MHz spectrometer and are reported in δ ppm scale. 1H-NMR spectra were

referenced to D2O (4.80 ppm) or CDCl3 (7.26 ppm). 13C-NMR spectra were referenced

to CDCl3 (77.23 ppm). ESI-MS data were collected at the UT Southwestern Medical

Center Protein Chemistry Technology Center. All acetylated sugars were prepared as 10

mM stock solutions in ethanol. The purity of acetylated sugars was confirmed by HPLC

analysis before cellular treatments (spectra located in appendix).

Synthesis of Ac4ManNAc

Ac4ManNAc was synthesized as previously reported.13 Briefly, to a solution of D-

(+)-N-acetylmannosamine (301.7 mg, 1.36 mmol) in pyridine (16.4 mL, 204 mmol),

acetic anhydride (4.72 mL, 54 mmol) was added and stirred overnight on ice. The

reaction mixture was diluted by CH2Cl2 and washed successively by 1.0 M HCl,

saturated sodium bicarbonate, and brine. The organic layer was dried over magnesium

sulfate and evaporated in vacuo. The residue was purified by flash chromatography

(hexanes / ethyl acetate gradient = 5/1, 3/1, 1/1) to afford Ac4ManNAc (282 mg, 53%,

mixture of anomers). 1H-NMR (500 MHz, CDCl3): δ 1.65 (3H, s), 2.02 (3H, s), 2.07

(3H, s), 2.11 (3H, s), 2.18 (3H, s), 4.10 (1H, dd, J = 2.3, 12.5), 4.28 (1H, t, J = 3.7), 4.78

(1H, ddd, J = 1.6, 3.9, 9.1), 5.06 (1H, d, 4.0), 5.13 (1H, t, J = 9.8), 5.33 (1H, d, J = 4.5),

5.79 (1H, d, J = 9.0), 5.86 (1H, d, J = 1.6). 13C-NMR (125 MHz, CDCl3): δ 20.88, 20.90,

20.92, 20.96, 20.97, 21.01, 21.08, 23.56, 23.65, 49.52, 49.74, 62.20, 65.41, 65.62, 68.99,

70.29, 71.56, 73.68, 90.86, 91.90, 168.34, 168.55, 169.92, 169.93, 170.23, 170.32,

67

170.73, 170.74, 170.82. ESI-MS for C16H23NO10 [M], calculated for 389.13, found

389.12. 1H-NMR, 13C-NMR, and ESI-MS spectra are presented in the appendix.

Synthesis of Ac4ManNDAz

ManNDAz was synthesized as previously reported.10 Briefly, to a solution of 4,4-

azo-pentanoic acid28 (128 mg, 1.00 mmol), D-(+)-mannosamine hydrochloride (216 mg,

1.00 mmol) and triethylamine (278 µL, 2.00 mmol) in MeOH (10 mL), 1-ethyl-3-(3-

dimethyllaminopropyl)carbodiimide hydrochloride (383 mg, 2.00 mmol) and 1-

hydroxybenzotriazole hydrate (135 mg, 1.00 mmol) were added. The reaction mixture

was stirred on ice for 10 minutes, followed by stirring at room temperature overnight.

The resulting mixture was concentrated in vacuo, roughly purified by flash

chromatography (CH2Cl2 / MeOH gradient = 1/0, 10/1, 4/1), and used directly to

synthesize the acetylated product, Ac4ManNDAz. The acetylation of ManNDAz was

performed by the same procedure described in the synthesis of Ac4ManNAc to afford

Ac4ManNDAz (84 mg, 28% over two steps, mixture of anomers). 1H-NMR (500 MHz,

CDCl3): δ 1.06 (3H, s), 1.81 (2H, m), 2.02 (3H, s), 2.07 (3H, s), 2.12 (3H, s), 2.19 (3H,

s), 4.05 (2H, s), 4.09 (1H, ddd, J = 6.2, 18, 30.5), 4.29 (1H, t, J = 4.5), 4.78 (1H, dd, J =

2.3, 8.2), 5.06 (1H, d, J = 4.0), 5.20 (1H, t, J = 9.8), 5.32 (1H, d, J = 4.4), 5.80 (1H, d, J =

9.0), 6.04 (1H, s). 13C-NMR (125 MHz, CDCl3): δ 20.18, 20.19, 20.85, 20.88, 20.91,

20.95, 20.97, 21.08, 25.50 25.54, 29.94, 30.05, 30.71, 30.84, 49.55, 49.76, 62.02, 62.15,

65.32, 69.06, 70.32, 71.57, 73.67, 90.80, 91.79, 168.33, 168.53, 169.82, 169.91, 170.20,

170.28, 170.75, 170.78, 171.62, 172.14. ESI-MS for C19H27N3O10 [M], calculated for

457.17, found 457.16. 1H-NMR, 13C-NMR, and ESI-MS spectra are presented in the

appendix.

Exposure of Jurkat cells with N-acetylmannosamine analogs – Ganglioside Analysis

Jurkat cells were seeded at a density of 2.5 x 105 cells/mL in 15 cm tissue culture

plates in 60 mL of media. Prior to the addition of cells to a tissue culture plate, acetylated

sugar in ethanol or ethanol only was added and the solvent was evaporated at ambient

temperature and pressue. For each condition, 2-3 plates of cells were used. After

growing for 72 hours, cells were counted and centrifuged at 220g for 5 min in 50 ml

68

conical tubes. To ensure consistent results among all samples, equal numbers of cells

were collected for every sample; the total number of cells collected for an experiment

ranged between 1.5 – 2.0 x 108 cells. Cell pellets were stored at -80 oC overnight before

proceeding to ganglioside extraction.

Extraction of gangliosides - Total Lipid Extraction

Cell pellets were thawed to room temperature, resuspended with 300 µl of ice

cold ddH2O (W), and dounced 50 times with a Kontes tissue grinder, tube size 20. With

a glass Pasteur pipette and a 2 mL rubber bulb, the cell lysate suspension was transferred

into a 4 mL glass vial containing 800 µL of methanol (M), already stirring. 400 µL of

chloroform (C) was added to the vial and the mixture was stirred thoroughly for 2 hours

at room temperature. Samples were covered in foil to prevent exposure to light. After

stirring, the mixture was transferred by a glass Pasteur pipette into a 13 x 100 mm glass

culture tube and centrifuged at 2800g for 10 min @ 30 oC. The supernatant (containing

the total lipid extract) was transferred by a glass Pasteur pipette into a new 4 mL glass

vial and evaporated to dryness under N2 gas.

Extraction of gangliosides - Phospholipid Extraction

The dried total lipid extract was resuspended with 800 µL butanol and 1200 µL

diisopropyl ether and sonicated in a water bath for 10 minutes. The resuspended lipids

were then transferred into a 13 x 100 mm glass culture tube using a glass Pasteur pipette.

To extract undesired phospholipids from the mixture, 1000 µl of 50 mM NaCl was added

to the tube and mixed vigorously by pipetting up and down repeatedly with a glass

Pasteur pipette. The mixture was then centrifuged at 2800g for 10 min at 30 oC to

separate the two phases. Using a glass Pasteur pipette, the organic phase (top layer) was

carefully removed. The aqueous mixture (bottom layer) was then extracted two more

times using the same ratio of butanol and diisopropyl ether.

Extraction of gangliosides - SepPak purification

After the final extraction, the remaining lipid mixture was loaded onto a SepPak

tC18 column, 0.3g size. The column was first pre-treated with three 2 mL washes of

69

C/M/W (2:43:55) followed by two 2 mL washes of C/M (1:1) and ending with three more

2 mL washes of C/M/W (2:43:55). After loading of the sample, the column was washed

three times with 2 mL of C/M/W (2:43:55) followed by three 2 mL washes of C/M (1:1)

to desalt the sample and remove unwanted contaminants. Elution of gangliosides was

achieved using 2 mL of 100 % methanol. Ganglioside extracts were then transferred into

a new 4 mL glass vial and evaporated to dryness under N2.

HPTLC Analysis of Extracted Gangliosides

Extracted ganglioside samples were redissolved with 30 µL C/M/W (2:1:0.1) and

resolved on HPTLC plates. Ganglioside standards were loaded onto the plate to provide

mass references. Gangliosides were separated with chloroform:methanol:0.2% CaCl2(aq)

(80:45:10) as the running buffer. HPTLC plates were first pre-run before loading 10 µL

of ganglioside extract. After thoroughly drying the plate in a fume hood, gangliosides

were detected by resorcinol staining (0.1% resorcinol, 0.04% CuSO4 in hydrochloric

acid:water [4:1]). Plates were imaged using an Alpha Innotech FluorChem HD2 and

images were processed using Adobe Photoshop.

To specifically detect GM1, extracted gangliosides were separated by HPTLC as

described above. The plate was fully dried under 40 mbar vacuum in a dessicator for 45

minutes. The plate was then treated with 0.5% polyisobutylmethacrylate in hexanes

(diluted from a 2.0% stock solution prepared in CHCl3) for 2 minutes by immersion in

the solution. The plate was then dried thoroughly with a stream of air, followed by

immersion in 1.0% BSA/PBS for 30 minutes at room temperature. The plate was

incubated with cholera toxin subunit B, Alexa Fluor 488 conjugate solution (1:20,000

dilution, in PBS) for 50 minutes at room temperature, followed by several brief washes

with PBS. After fully drying the plate using a stream of air, the plate was imaged by

Typhoon using the 488 excitation laser and 520 emission filter. Image analysis was

performed using Adobe Photoshop. Resorcinol staining was performed (as described

above) after fluorescence imaging to confirm ganglioside composition.

70

Mass Spectrometry Analysis of Extracted Gangliosides

Dried ganglioside extracts were sent to the Complex Carbohydrate Research

Center at the University of Georgia for mass spectrometry analysis. To analyze the

overall composition of extracted gangliosides, MALDI-TOF-MS was performed.

Samples were crystallized onto a MALDI plate with trihydroxyacetophenone

monohydrate (THAP) as a matrix. Analysis of gangliosides was performed in the

negative ion mode using a Bruker microflex instrument. Spectra are presented in the

appendix.

Exposure of Jurkat cells with N-acetylmannosamine analogs – Immunofluorescence

analysis

Jurkat cells were seeded at a density of 2.5 x 105 cells/mL into individual wells of

6-well tissue culture plates in 4 mL of media. Prior to the addition of cells to the tissue

culture plate, acetylated sugar or ethanol was added and evaporated. For each condition,

3-4 individual wells were used. After growing for 70 hours, the cells were counted, spun

down at 200g for 5 min, and resuspended in original media to 5.0 x 106 cells/mL.

Coverslip preparation and poly-L-lysine coating

Coverslips were heated in a loosely covered glass beaker in 1 M HCl at 50oC for 4

hours. After cooling to room temperature, the slides were rinsed several times with

ddH2O to remove all traces of acid. The slides were then sonicated three times for 30

minutes in ddH2O, changing the solution in between. The coverslips were then sonicated

successively with 50%, 70% and 95% ethanol, each for 30 minutes. Slides were then

dried on filter paper and stored. To generate poly-L-lysine coated slides, coverslips were

coated in 0.1% (w/v) poly-L-lysine solution using a Petri dish overnight at 4oC, while

rotating. The coverslips were then washed at least 10x in ddH2O to remove poly-L-lysine

solution. Then the coverslips were rinsed with 100% ethanol and dried with filter paper

in a sterile hood. Cells were added onto the coverslips when dry.

71

Cholera toxin incubation and photocrosslinking

1000 µl of Jurkat cell suspension was transferred into individual wells of new 6-

well plates containing poly-L-lysine coated coverslips tamped onto the bottom surface.

Cells were incubated for 20 min at room temperature to allow the Jurkat cells to bind to

the lysine-coated surface. After incubation, the plates were transferred onto ice for 10

min. To each well, 5 µL of cholera toxin subunit B (1 mg/mL) was added and the plate

was gently swirled, on ice. The plates were then exposed to 365 nm light for 10 min –

approximately 2 cm away from bulb. After irradiation, the culture media was removed

and replaced with fresh pre-warmed media. Cells were incubated at 37oC, 5% CO2 for 90

minutes to allow for endocytosis of cholera toxin subunit B.

Immunofluorescence analysis

After incubation and photocrosslinking, the plates were immediately transferred

back onto ice, the supernatant aspirated, and directly fixed with 3.7%

paraformaldehyde/PBS for 20 minutes without washing. After several washes with PBS,

the cells were permeabilized with 0.1% Triton X-100, 1.0% BSA/PBS for 5 minutes

without shaking. After several washes with PBS, the plates were placed at room

temperature and incubated with 1.0% BSA/PBS for 1 hour to block non-specific binding.

After blocking, the cells were incubated with primary antibody (1:500 dilution; rabbit

anti-calnexin was used at 1:50 dilution) for 2 hours at room temperature. The coverslips

were then washed three times with PBS and the cells were incubated with secondary

antibody (1:500 dilution) for 1 hour at room temperature. The coverslips were washed

again three times with PBS. After washing, the coverslips were mounted onto glass

cover slides using Vectashield Mounting Medium with DAPI and sealed with nail polish.

Slides were stored at 4oC until analysis. Cells were visualized using the HCX PL APO

Lambda Blue 63x 1.40 oil UV objective of a Leica TCS SP5 confocal microscope

equipped with 405 nm, 488 nm, 561 nm, and 633 nm lasers. Image analysis was

performed using ImageJ (NIH), Adobe Photoshop, and Adobe Illustrator.

72

References


proteins. Nature 446, 1023-1029, (2007).

2 Schauer, R. Sialic acids as regulators of molecular and cellular interactions. Curr

Opin Struct Biol 19, 507-514, (2009).

3 Angata, T. & Varki, A. Chemical diversity in the sialic acids and related alpha-

keto acids: an evolutionary perspective. Chem Rev 102, 439-469, (2002).

4 Sylte, M. J. & Suarez, D. L. Influenza neuraminidase as a vaccine antigen. Curr

Top Microbiol Immunol 333, 227-241, (2009).

5 Tsai, B. et al. Gangliosides are receptors for murine polyoma virus and SV40.

EMBO J 22, 4346-4355, (2003).

6 Chen, C., Fu, Z., Kim, J. J., Barbieri, J. T. & Baldwin, M. R. Gangliosides as high

affinity receptors for tetanus neurotoxin. J Biol Chem 284, 26569-26577, (2009).

7 Connell, T. D. Cholera toxin, LT-I, LT-IIa and LT-IIb: the critical role of

ganglioside binding in immunomodulation by type I and type II heat-labile

enterotoxins. Expert Rev Vaccines 6, 821-834, (2007).

8 Bennett, V. & Cuatrecasas, P. Mechanism of activation of adenylate cyclase by

Vibrio cholerae enterotoxin. J Membr Biol 22, 29-52, (1975).


1382-1401, (2009).

10 Tanaka, Y. & Kohler, J. J. Photoactivatable crosslinking sugars for capturing

glycoprotein interactions. J Am Chem Soc 130, 3278-3279, (2008).

11 Tanaka, Y., Bond, M. R. & Kohler, J. J. Photocrosslinkers illuminate interactions

in living cells. Mol Biosyst 4, 473-480, (2008).

73



13 Bond, M. R., Zhang, H., Vu, P. D. & Kohler, J. J. Photocrosslinking of

glycoconjugates using metabolically incorporated diazirine-containing sugars. Nat

Protoc 4, 1044-1063, (2009).

14 Luchansky, S. J., Yarema, K. J., Takahashi, S. & Bertozzi, C. R. GlcNAc 2-

epimerase can serve a catabolic role in sialic acid metabolism. J Biol Chem 278,

8035-8042, (2003).

15 Ladisch, S. & Gillard, B. A solvent partition method for microscale ganglioside

purification. Anal Biochem 146, 220-231, (1985).

16 Schnaar, R. L. Isolation of glycosphingolipids. Methods Enzymol 230, 348-370,

(1994).

17 Nimrichter, L. et al. E-selectin receptors on human leukocytes. Blood 112, 3744-

3752, (2008).

18 Sack, D. A., Sack, R. B., Nair, G. B. & Siddique, A. K. Cholera. Lancet 363, 223-

233, (2004).

19 Cuatrecasas, P. Interaction of Vibrio cholerae enterotoxin with cell membranes.

Biochemistry 12, 3547-3558, (1973).

20 Kuziemko, G. M., Stroh, M. & Stevens, R. C. Cholera toxin binding affinity and

specificity for gangliosides determined by surface plasmon resonance.

Biochemistry 35, 6375-6384, (1996).

21 MacKenzie, C. R., Hirama, T., Lee, K. K., Altman, E. & Young, N. M.

Quantitative analysis of bacterial toxin affinity and specificity for glycolipid

receptors by surface plasmon resonance. J Biol Chem 272, 5533-5538, (1997).

22 Dawson, R. M. Characterization of the binding of cholera toxin to ganglioside

GM1 immobilized onto microtitre plates. J Appl Toxicol 25, 30-38, (2005).

74

23 Chinnapen, D. J., Chinnapen, H., Saslowsky, D. & Lencer, W. I. Rafting with

cholera toxin: endocytosis and trafficking from plasma membrane to ER. FEMS

Microbiol Lett 266, 129-137, (2007).

24 Fujinaga, Y. et al. Gangliosides that associate with lipid rafts mediate transport of

cholera and related toxins from the plasma membrane to endoplasmic reticulm.

Mol Biol Cell 14, 4783-4793, (2003).

25 Schnaar, R. L. & Needham, L. K. Thin-layer chromatography of

glycosphingolipids. Methods Enzymol 230, 371-389, (1994).

26 Lopez, P. H. & Schnaar, R. L. Determination of glycolipid-protein interaction

specificity. Methods Enzymol 417, 205-220, (2006).

27 Wolf, A. A., Fujinaga, Y. & Lencer, W. I. Uncoupling of the cholera toxin-G(M1)

ganglioside receptor complex from endocytosis, retrograde Golgi trafficking, and

downstream signal transduction by depletion of membrane cholesterol. J Biol

Chem 277, 16249-16256, (2002).

28 Church, R. F. R. & Weiss, M. J. Diazirines. II. Synthesis and Properties of Small

Functionalized Diazirine Molecules - Some Observations on Reaction of a

Diaziridine with Iodine-Iodide Ion System. Journal of Organic Chemistry 35,

2465-2471, (1970).

75

Chapter 4 – Exploring the incorporation of sialic acid analogs into gangliosides of

mammalian cells

Introduction

The cell surface expression of sialic acid plays an integral role in facilitating

numerous molecular interactions on the eukaryotic plasma membrane.1 For example,

sialic acid is responsible for mediating adhesion of leukocytes to endothelial cells at sites

of inflammation.2 Sialic acids are commonly expressed into a wide variety of cell surface

glycoconjugates, including glycolipids. Sialylated glycolipids, known as gangliosides,

are responsible for many biological roles, including cell signaling, cell-cell

communication, and pathogen recognition.3 They are also highly expressed in the

developing mammalian brain4 and in malignancy.5 Since gangliosides are anchored

directly into the plasma membrane, studying their function and molecular details in a

native environment is often difficult.6 This hindrance precipitates the need to develop

new biological methods that can be used to examine gangliosides’ behavior in their

native environment.

One approach to investigating the roles of cell surface sialic acid is to introduce

small structural modifications by culturing cells with analogs of sialic acid or N-

acetylmannosamine (ManNAc), which is a metabolic precursor of sialic acid. Known as

metabolic oligosaccharide engineering, this technique relies on the ability of cells to

uptake these analogs through passive diffusion, process them with their natural cellular

machinery, and insert the unnatural metabolite into glycan structures in place of the

natural sugar.7,8 Pioneering work in the early 1990s by Reutter and co-workers

demonstrated that molecules in which the N-acyl chain of sialic acid was lengthened with

additional methylene units could be readily tolerated and incorporated into cell surface

glycan structures in cells and live animals.9,10 This technology demonstrated to be highly

applicable for exploring host-virus interactions,11 neuronal cell differentiation,12 and

inhibiting cell surface polysialylation.13 Since then, metabolic oligosaccharide

engineering has expanded its utility to introduce chemically reactive functional groups

(i.e. ketones,14 azides,15,16 alkynes17) into cell surface sialic acid.18 The incorporation of

these bio-orthogonal groups has provided a chemical handle for labeling sialylated glycan

76

structures in vivo.19-24 Recently, the ability to capture sialic acid-mediated interactions,

which suffer from weak affinity and fast off-rates, has been achieved by metabolic

oligosaccharide engineering techniques through the introduction of photoactivatable

groups.16,25,26

While this technology has proved effective at studying the many biological roles

of sialic acid, the majority of these studies have investigated the global incorporation of

unnatural sialic acid analogs. In humans, sialic acid is attached onto glycan chains in

α2,3-, α2,6-, and α2,8-linkages. Production of sialic acid glycoconjugates, or sialosides,

is controlled by 20 different sialyltransferases, each with individual substrate

specificities. This diversity presents a potential problem in generalizing metabolic

oligosaccharide engineering results for all types of sialosides. Because gangliosides can

be extracted from cells and independently analyzed, these molecules provide an excellent

target to study the application of metabolic oligosaccharide engineering technology into

specific sialosides. Several groups have reported the use of metabolic oligosaccharide

engineering employing ManNAc analogs to produce unnatural gangliosides in

mammalian cells (detailed in Chapter 1). ManNAz, an azide-containing analog, has

been successfully metabolized into SiaNAz, which was incorporated into cell surface

gangliosides of Jurkat cells.27 ManNProp and ManNPhAc have also been successfully

metabolized into cell surface gangliosides in several cancer lines and been utilized as

immunogenic targets for antibodies.28-31 ManNGc, which can be metabolized into

NeuGc, has been used to metabolically produce hydroxylated cell surface gangliosides in

neuronal cells to investigate myelin-axon interactions.32

To further probe the substrate flexibility that is possible with ganglioside

biosynthesis, I synthesized a panel of ManNAc analogs that are known to be metabolized

to sialosides and investigated their incorporation into gangliosides. In addition to above

mentioned ManNAc analogs, my panel was expanded to include ManNBut10,33 and

ManNDAz26 which have been previously demonstrated to become metabolized into cell

surface glycoconjugates. Because of the diversity of ganglioside structures present in

mammalian cells, I focused my experiments on production of the GM3 ganglioside,

which is the precursor for other ganglioside products. To learn more about specificity

77

differences between mammalian species, I conducted my experiments in both human

(BJAB) and hamster (CHO) cell lines. My results reveal significant differences between

these two cell lines’ ability to accept and incorporate modified sialic acids into

gangliosides.

Results

Incorporation of variant sialic acids into gangliosides

To test whether sialic acid analogs are incorporated into the ganglioside GM3, I

relied on a key set of reagents: two cell lines, BJAB K2034,35 and CHO Lec3,36 that are

impaired in sialic acid biosynthesis through their inactive UDP-GlcNAc 2-epimerase

(Figure 1). This enzyme catalyzes the conversion of UDP-GlcNAc into ManNAc and is

vital for production of sialic acid in mammalian cells. When these epimerase-deficient

cell lines are cultured in serum free conditions, they are unable to synthesize gangliosides

due to their inability to produce sialic acid. Two metabolites within the sialic acid

biosynthesis pathway, ManNAc and sialic acid, present an entry point for introducing

chemical modifications into sialylated glycans (Figure 1 – highlighted in green): in this

work, I have chose to use ManNAc analogs due to their synthetic simplicity compared to

generating sialic acid analogs. When cells are cultured with ManNAc, they regain the

ability to generate sialylated glycoconjugates, including gangliosides. Similarly, when

the cells are cultured with unnatural ManNAc analogs, any observation of gangliosides is

interpreted as reflecting successful incorporation of these unnatural analogs into

gangliosides. BJAB cells primarily generate GM3 but also are able to produce small

amounts of GM1 (confirmed by MALDI-TOF-MS analysis, data not shown). CHO cells

are only able to generate GM3 because they lack the enzymes needed to synthesize

downstream ganglioside products.37

78

Figure 1. Metabolic oligosaccharide engineering of UDP-GlcNAc 2-epimerase-deficient cells occurs through the introduction of ManNAc and sialic acid analogs. BJAB K20 and CHO Lec3 cells lack UDP-GlcNAc 2-epimerase activity, rendering them unable to generate sialosides when cultured in serum free media. If these cells are cultured with downstream metabolites in this pathway, they regain the ability to generate sialosides. Two metabolites, ManNAc and sialic acid (highlighted in green), serve as entry points for introducing chemical modifications into sialosides (denoted in red).

I chose to use a panel of ManNAc analogs that have reported to be metabolized to

sialosides in one or more cell types;18 this panel includes several analogs which have

been previously shown to be metabolized to their sialic acid counterparts and

incorporated into gangliosides (Figure 2).27,28,30,31 BJAB K20 and CHO Lec3 cells were

cultured in serum free conditions with each of the ManNAc analogs for 48-72 hours.

Free hydroxyl groups present on each sugar were acetylated with acetic anhydride to

improve diffusion through the membrane.38 The acetyl protecting groups are believed to

be removed by nonspecific esterases inside the cells. After culturing with acetylated

ManNAc analogs, the cells were harvested and the gangliosides were extracted using

established methods (detailed in Methods).39-41 Gangliosides were resolved by high

performance thin layer chromatography (HPTLC) and visualized by resorcinol staining,

which specifically detects sialic acid-containing molecules.

79

Figure 2. Panel of ManNAc analogs used in ganglioside incorporation experiments. In order to investigate the incorporation of unnatural sialic acid analogs into gangliosides, a panel of unnatural N-acyl modifications were introduced into ManNAc and cultured with BJAB K20 and CHO Lec3 cells. All of these analogs have been demonstrated to be metabolized to cell surface sialosides. The color-coding scheme used for these analogs is continued throughout this chapter.

When I examined the ability of BJAB K20 cells to incorporate unnatural

ManNAc analogs into GM3, I observed that these human cells were quite permissive to

unnatural modifications to the N-acyl side chain of sialic acid (Figure 3A). Increasing

the length of the N-acyl side by additional methylene groups can easily be

accommodated, demonstrated by incorporation of ManNProp and ManNBut. This

apparent increase in N-acyl chain length also allows for efficient metabolism of ManNAz

and ManNDAz, introducing chemical functionality in the side chain of sialic acid.

However, the incorporation of ManNGc into GM3 as NeuGc appears to be significantly

reduced in comparison to the natural substrate ManNAc. Only one analog (ManNPhAc)

was not incorporated into GM3 at levels detectable by resorcinol staining. MALDI-TOF-

MS analysis (located in appendix) of extracted gangliosides from BJAB K20 cells

cultured with ManNAc, ManNGc, ManNAz, and ManNDAz confirmed the identity of

the modified GM3 molecules.

When I cultured CHO Lec3 cells with the same panel of ManNAc analogs, I

observed that the hamster cells display restricted substrate tolerance (Figure 3B). In

comparison to BJAB K20 cells, CHO Lec3 cells appear to only tolerate minimal

perturbations to the N-acyl side chain of sialic acid, evidenced from ManNProp culturing

experiments. Increasing the side chain beyond one methylene group led to diminished

80

levels of incorporation for several analogs, ManNBut and ManNAz; longer N-acyl side

chains, such as ManNPhAc and ManNDAz, were not incorporated at detectable levels.

Unlike BJAB K20 cells, CHO cells were capable of metabolizing ManNGc and

incorporating the sialic acid analog into GM3 at levels nearly comparable to ManNAc.

MALDI-TOF-MS analysis of extracted gangliosides from CHO Lec3 cells cultured with

ManNAc, ManNGc, and ManNAz confirmed successful incorporation of their sialic acid

counterparts into GM3 (located in appendix).

Figure 3. HPTLC analysis of gangliosides produced by BJAB and CHO cells cultured with ManNAc analogs. The ganglioside extracts of BJAB K20 and CHO Lec3 cells (epi -) cultured with various ManNAc analogs were resolved by HPTLC and visualized by resorcinol staining. When K20 and Lec3 cells are cultured under serum-free conditions, they lose the ability to synthesize gangliosides; if the cells are cultured with their natural substrate (ManNAc), normal ganglioside production is restored. BJAB K88 and CHO cell lines possessing normal functioning UDP-GlcNAc 2-epimerase activity were used as positive controls (epi +). (A) BJAB cells generate GM3 and GM1 gangliosides. BJAB cells are able to effectively metabolize ManNProp, ManNBut, and ManNAz to their sialic acid counterparts and incorporate them into GM3. Cells cultured with ManNGc and ManNDAz demonstrate significantly lower levels of GM3 production, while cells cultured with ManNPhAc do not produce detectable levels of GM3. (B) CHO cells produce one ganglioside: GM3. CHO cells are effectively metabolize only ManNGc and ManNProp into their sialic acid counterparts and incorporate them into GM3. Cells cultured with ManNBut and ManNAz display significantly lower levels of GM3 production, while cells cultured with ManNPhAc and ManNDAz do not produce detectable levels of GM3.

81

These observations highlight key differences in the utilities of these two cell lines

for performing metabolic oligosaccharide engineering experiments with gangliosides.

The human BJAB cell line demonstrates the substrate flexibility to introduce larger

modifications that include chemical functionalities (such as azides or diazirines) into

gangliosides but shows attenuated ability to introduce other naturally occurring forms of

sialic acid (NeuGc). Conversely, the hamster cell line shows much restricted substrate

scope but, unlike the BJAB cell line, is able to utilize ManGc as a substrate for generating

GM3-NeuGc.

Analysis of engineered cell surface sialylation with BJAB and CHO cell lines cultured

with ManNAc analogs

The HPTLC experiments performed above provide a great deal of information

regarding the ability of BJAB and CHO cells to metabolize ManNAc analogs and

incorporate their sialic acid counterparts into GM3. We were surprised to observe

significant differences between the two cell lines, and decided to perform additional

experiments to determine whether the differences reflected the overall ability of the cells

to metabolize these analogs to sialosides or if the differences were specific to ganglioside

production. To look at global cell surface sialoside production, I cultured BJAB K20 and

CHO Lec3 cells with my ManNAc analog panel and used flow cytometry to measure the

incorporation of these unnatural analogs into cell surface sialosides by lectin binding.

Lectins are a family of plant and animal carbohydrate binding proteins that are readily

used to detect specific glycan epitopes and linkages. For analysis of sialosides, α2,3-

linked sialosides were detected with the Maackia amurensis lectin (MAA) and α2,6-

linked sialosides were detected with the Sambucus nigra lectin (SNA). Currently, there

are no commercially available sources of accurately detecting α2,8-linked sialosides.

Because lectin recognition of sialosides could potentially be hindered by structural

changes to sialic acid and to include α2,8-linked structures, I examined global sialic acid

engineering using the recently published periodate oxidation and aniline-catalyzed oxime

ligation (PAL).42 This novel method uses mild periodate oxidation to convert the glycolyl

chain of sialic acid into a reactive aldehyde that can be covalently labeled by an

82

aminooxy-functionalized biotin for detection. Because the presence of cell surface

aldehydes do not occur naturally, this method can be used to specifically label sialic acid.

Because sialic acid is α2,3-linked in the GM3 ganglioside, I first compared the

ganglioside incorporation experiments in BJAB cells (Figure 3A) to the overall

incorporation of α2,3-linked sialic acid on the cell surface of BJAB cells as measured by

MAA binding (Figure 4A). The ganglioside incorporation experiments and the MAA

binding experiments resulted in similar trends: significant levels of incorporation were

observed with extended N-acyl alkyl chains (ManNProp and ManNBut) and with the

introduction of azide group (ManNAz), while the hydroxyl- and diazirine-modified

analogs (ManNGc and ManNDAz) yielded slightly lower levels of incorporation. Next, I

examined the incorporation of analogs into α2,6-linked sialic acid glycoconjugates on the

surface of BJAB cells (Figure 4B). I observed the same trends for analog incorporation

with one exception: ManNGc. When BJAB K20 cells were cultured with ManNGc, I

observed low levels of MAA binding (Figure 4A) but the levels of SNA binding exceed

those produced by the natural substrate (ManNAc) (Figure 4B). However, the amount of

ManNGc incorporation into the total cell surface sialosides, as measured by PAL, was

substantially lower (Figure 4C). These results suggest NeuGc is efficiently incorporated

into α2,6-linked sialosides, but less efficiently incorporated into α 2,3linked sialosides.

Because I observe robust levels of SNA binding, it seems unlikely that the low levels of

MAA binding result from inefficient metabolism of ManNGc to the nucleotide sugar

donor, CMP-NeuGc. Rather, I speculate that the incorporation differences result from

differences in the ability of individual sialyltransferases to tolerate modifications to sialic

acid.

Next, I used lectin binding to examine the ability of different ManNAc analogs to

be metabolized to sialosides and displayed on the surface of CHO cells (Figure 4D). My

observations were consistent with the ganglioside incorporation experiments conducted

in CHO cells: the highest levels of incorporation were only seen with small modifications

to the N-acyl chain (ManNGc and ManNProp) while further increases in size diminished

the levels of incorporation (ManNBut, ManNAz, and ManNDAz). These results were

also fairly consistent with total cell surface sialoside measurements made by PAL (Figure

83

4E). I was unable to examine α2,6-linked glycoconjugates since CHO cells are devoid of

ST6Gal1 enzyme, the primary enzyme responsible for α2,6-linked glycoconjugates found

on N-linked glycosylation.43

Figure 4. Flow cytometry analysis of cell surface display of modified sialic acids. Cell surface sialylation of engineered BJAB K20 and CHO Lec3 cells (epi -) was analyzed by lectin binding (MAA - α2,3-sialic acid linkages, SNA - α2,6-sialic acid linkages) and periodate oxidation and aniline-catalyzed oxime ligation (PAL – measures total cell surface sialylation). BJAB K88 and CHO cell lines possessing functional UDP-GlcNAc 2-epimerase activity were used as positive controls (epi +). Experiments were performed in biological triplicate; duplication of the entire experiment yielded similar results. (A) MAA lectin analysis of cultured BJAB cells. (B) SNA lectin analysis of cultured BJAB cells. (C) PAL analysis of cultured BJAB cells. (D) MAA lectin analysis of cultured CHO cells. (E) PAL analysis of cultured CHO cells. SNA binding was not assayed in CHO cells because they lack production of α2,6-linked sialosides.

84

Discussion

Metabolic oligosaccharide engineering is an emerging technique that provides the

ability to introduce small structural changes into individual monosaccharides to probe the

biological roles and responsibilities of cell surface glycosylation.18 To date, this

technique has provided significant information regarding glycoproteins while the focus

on glycolipid engineering, specifically gangliosides, has only recently begun to be

explored.28-31 My goal was to explore the substrate flexibility of ganglioside

sialyltransferases with metabolic oligosaccharide engineering using a panel of ManNAc

analogs reported to become metabolized into cell surface sialosides. My experimental

system relied on two UDP-GlcNAc 2-epimerase deficient cell lines (BJAB K20 and CHO

Lec3) that, when devoid of serum rich media, are only able to generate sialic acid-

containing gangliosides when supplemented with ManNAc (Figure 1).34-36 This unique

system provides a direct readout of the successful incorporation of unnatural sialic acid

analogs into gangliosides.

Incorporation of modified sialic acids into the GM3 ganglioside of BJAB cells

I cultured BJAB K20 cells with a panel of ManNAc analogs (Figure 2) and

visualized their incorporation of the resulting sialic acid analog into gangliosides by

HPTLC (Figure 3A). BJAB K20 cells appear to tolerate extension of the N-acyl chain on

GM3 with additional methylene units using ManNProp and ManNBut. This promiscuity

in hsaST3Gal5 allows for the introduction of bio-orthogonal chemical groups into GM3,

as evidenced with BJAB K20 cells cultured with ManNAz and ManNDAz. Interestingly,

the relative production of GM3-SiaNAz in BJAB K20 cells appears to be consistent to

normal GM3 levels. This raises the possibility that many reported ManNAz labeling

experiments3,15,19-24,44 could be labeling gangliosides along with glycoproteins.

Generating SiaDAz-labeled GM3 presents an excellent opportunity to investigate the

molecular basis for many recently reported associations involving this ganglioside, as

described in Chapter 1.3,44 When comparing the relative levels of the GM3 analog being

produced, it appears that increasing the size of the N-acyl chain decreases ganglioside

production (Figure 3). This observation was found to be consistent with the flow

cytometry measurements made on α2,3-linked glycoconjugates (Figure 4A) and α2,6-

85

linked glycoconjugates (Figure 4B). Because the introduction of N-acyl chain

modifications on sialic acid can hinder lectin recognition, I analyzed the cell surface

sialoside production by PAL and observed a consistent global pattern of glycan

engineering. These results indicate that metabolic oligosaccharide engineering of

gangliosides is generally comparable to similar experiments targeting glycoproteins in

human cells.

While BJAB K20 cells were able to metabolize several ManNAc analogs into

GM3, I observed diminished incorporation of ManNGc (discussed below) and was

unable to detect the utilization of ManNPhAc (Figure 3A). ManNPhAc was included

within this panel due to its successful incorporation into GM3 in two human tumor cell

lines: K562 and SK-MEL-28.30 In these experiments, a highly specific antibody

generated against this epitope was used to detect its presence on the cell surface. Because

antibody recognition is more sensitive than resorcinol staining, it is plausible that

ManNPhAc engineering is occurring at extremely low levels. The metabolic engineering

experiments performed here were done with a higher concentration (100 µM) of the

ManNAc analog than used previously with ManNPhAc (40 µM); this decision was based

upon our group’s previous work investigating optimal levels of ManNDAz incorporation

into BJAB K20 cells.25 It is possible that reducing the concentration of ManNPhAc in the

culture media might allow for detectable synthesis of GM3-SiaPhAc. Examination of

α2,3-linked (Figure 4A), α2,6-linked (Figure 4B) and total cell surface sialosides (Figure

4C) demonstrated minimal incorporation levels of ManNPhAc into sialylated glycan

structures, indicating that this trend is likely occurring with all types of sialyltransferases.

Another possibility for the observed differences could involve the type of cancer cell line

that was chosen. Due to the mutational differences that exist between these cancerous

cell lines, it is difficult to identify the exact mechanism behind these differing

observations. BJAB K20 cells were chosen for these studies for their ability to provide a

direct readout of metabolic oligosaccharide engineering without competition from

endogenous metabolites.

86

Incorporation of modified sialic acids into the GM3 ganglioside of CHO cells

I cultured CHO Lec3 cells with a panel of ManNAc analogs (Figure 2) and

visualized their incorporation into gangliosides by HPTLC (Figure 3B). CHO Lec3 cells

appear to tolerate minor extensions onto the N-acyl chain on GM3, observed with

ManNGc and ManNProp. Only trace amounts of gangliosides could be detected with the

longer extensions from ManNProp and ManNAz, while both ManNPhAc and ManNDAz

were unable to be detected. These results were fairly consistent with the flow cytometry

measurements made on α2,3-linked glycoconjugates (Figure 4D) and total cell surface

sialosides (Figure 4E). The ability of CHO cells to introduce ManNAz into cellular

glycans is well documented; CHO cells are able to incorporate ManNAz into cellular

glycoproteins,45,46 including α2,3-linked glycoconjugates 47 (shown here in Figure 4D).

However, CHO cells appear to be inefficient at incorporating the azide functionality into

gangliosides (Figure 3B). It appears that the narrow substrate flexibility of ST3Gal5 in

CHO cells hinders their usage to study ganglioside-based interactions with metabolic

oligosaccharide engineering techniques.

The ST3Gal5 enzyme of BJAB cells shows a reduced capacity for engineering NeuGc-

containing gangliosides

When I cultured BJAB K20 cells with ManNGc, I observed that synthesis of the

corresponding GM3 analog (GM3-NeuGc) was considerably lower than BJAB K20 cells

cultured with ManNAc (Figure 3A). This trend appears to be consistent with the entire

α2,3-sialyltransferase family, as evidenced by MAA-lectin analysis (Figure 4A). One

possible explanation for these results is that BJAB K20 cells are inefficient at producing

adequate levels of CMP-NeuGc for sialoside synthesis. However, NeuGc engineering of

BJAB K20 cells into α2,6-linked glycoconjugates was observed to be higher than BJAB

K20 cells engineered with NeuAc (Figure 4B). This implies that the decreased synthesis

of α2,3-linked sialosides is not likely due to reduced production of CMP-NeuGc, since I

observed the synthesis of α2,6-linked NeuGc structures. Furthermore, the production of

NeuGc-containing sialosides in CHO cells was efficiently recognized by the MAA lectin,

suggested that the reduced level of MAA binding is not simply due to MAA’s inability to

87

recognize NeuGc-containing glycans. Indeed, previous reports have demonstrated that

SNA and MAA lectins do not discriminate between NeuAc and NeuGc.48 Instead, my

results point to an interpretation that human α2,3-sialyltransferases, specifically here

ST3Gal5, are able to utilize CMP-NeuGc as efficiently as α2,6-sialyltransferases. To

confirm my hypothesis, I am currently planning experiments to analyze the production of

CMP-NeuGc in BJAB K20 cells cultured with ManNGc. These cells will be harvested

for their nucleotides and analyzed by high performance anion exchange chromatography

(HPEAC) to quantify the synthesis of CMP-NeuGc. These results will be compared to

BJAB K88 and BJAB K20 cells cultured with ManNAc to determine if there are

adequate levels of CMP-NeuGc being produced. These results will help determine if the

production of GM3-NeuGc is hampered by inefficient synthesis of CMP-NeuGc or by the

substrate specificity of ST3Gal5. More direct measurements of ST3Gal5 specificity are

also underway.

Future directions

Overall, I observed that the human BJAB cells and hamster CHO cells display

significant differences in their ability to incorporate modified sialic acid into

gangliosides. These differences are extremely evident between ManNGc-, ManNAz- and

ManNDAz-treated cells. BJAB K20 cells are able to synthesize GM3-SiaNAz and GM3-

SiaDAz more efficiently than CHO Lec3 cells. Conversely, CHO Lec3 cells are able to

synthesize GM3-NeuGc more efficiently than CHO Lec3 cells. These results could be

caused from subtle structural variations in the catalytic domains and immediate

surrounding areas of ST3Gal5. Therefore, my hypothesis is BJAB K20 cells transfected

with choST3Gal5 would regain the ability to synthesize GM3-NeuGc and CHO Lec3

transfected with hsaST3Gal5 would obtain the ability to synthesize GM3-SiaNAz and

GM3-SiaDAz. To test this theory, I am currently planning to perform these

aforementioned transfections and analyze the production of GM3 by flow cytometry. To

confirm that the transfection of ST3Gal5 into these cells does not simply result in

overproduction of GM3, I am planning to perform control experiments where I transfect

the endogenous gene into each cell type. Due to the incomplete sequencing of the

hamster genome, I was unable to generate a mammalian expression plasmid for hamster

88

ST3Gal5; in its place, I have chosen to use the mouse ST3Gal5 for my investigations.

Additionally, this hypothesis will be tested by recombinantly expressing ST3Gal5

enzymes and performing in vitro experiments with unnatural sialic acid analogs.

Furthermore, it would be worthwhile to perform these experiments at different ManNAc

analog concentrations to see if the efficiency of metabolic oligosaccharide engineering

changes.

In conclusion, I have demonstrated that metabolic oligosaccharide engineering

techniques can be applied to introduce small perturbations into the N-acyl side chain of

the sialic acid residue of the GM3 ganglioside. My results also illustrate significant

differences in the incorporation of sialic acid analogs into GM3 in different cell lines; the

most notable difference was observed with α2,3-NeuGc incorporation. These differences

may reflect species-specific or cell type-specific differences. Further analyses are

currently underway to determine if the attenuated synthesis of GM3-NeuGc in BJAB K20

cells versus CHO cells is caused by an impairment of CMP-NeuGc synthesis or by

differences in ST3Gal5 specificity.

Acknowledgements

I would like to thank Michael Pawlita (German Cancer Research Center) and

James Paulson (The Scripps Research Institute) for sharing BJAB K20 and BJAB K88

cells, Mark Lehrman (UT Southwestern Medical Center) for sharing CHO cells, and

Pamela Stanley (Albert Einstein College of Medicine) for sharing CHO Lec3 cells. I

would like to thank Yan Li (UT Southwestern Medical Center Protein Chemistry

Technology Center) for mass spectrometry analysis of sugars, Parastoo Azadi and

Roberto Sonon (University of Georgia Complex Carbohydrate Research Center) for mass

spectrometry of ganglioside samples, and Angela Mobley (UT Southwestern Medical

Center Flow Cytometry Core Facility) for help with flow cytometry.

89

Methods

General information for the chemical synthesis of N-acetylmannosamine analogs

All chemicals were used as received from commercial suppliers without further

purification. 1-hydroxybenzotriazole hydrate was purchased from AnaSpec. N-butyric

anhydride, phenylacetic acid and propionic anhydride were purchased from TCI America.

All other chemicals were purchased from Sigma-Aldrich or Fisher Scientific unless

otherwise noted. Reaction progress was monitored by analytical thin layer

chromatography (TLC) on silica gel 60 F254 glass backed plates (Fisher) and stained with

ceric ammonium molybdate. Flash column chromatography was carried out with silica

gel 60 (particle size 40-63 µm, EMD Chemicals). All 1H-NMR and 13C-NMR spectra

were recorded on a Varian 500 MHz spectrometer and are reported in δ ppm scale. 1H-

NMR spectra were referenced to D2O (4.80 ppm) or CDCl3 (7.26 ppm). 13C-NMR

spectra were referenced to CDCl3 (77.23 ppm). ESI-MS data were collected at the UT

Southwestern Medical Center Protein Chemistry Technology Center. All acetylated

sugars were prepared as 10 mM stock solutions in ethanol. The purity of acetylated

sugars was confirmed by HPLC analysis before cellular treatments (spectra located in

appendix).

Synthesis of Ac4ManNAc

Ac4ManNAc was synthesized as previously reported.13 Briefly, to a solution of D-

(+)-N-acetylmannosamine (301.7 mg, 1.36 mmol) in pyridine (16.4 mL, 204 mmol),

acetic anhydride (4.72 mL, 54 mmol) was added and stirred overnight on ice. The

reaction mixture was diluted by CH2Cl2 and washed successively by 1.0 M HCl,

saturated sodium bicarbonate, and brine. The organic layer was dried over magnesium

sulfate and evaporated in vacuo. The residue was purified by flash chromatography

(hexanes / ethyl acetate gradient = 5/1, 3/1, 1/1) to afford Ac4ManNAc (282 mg, 53%,

mixture of anomers). 1H-NMR (500 MHz, CDCl3): δ 1.65 (3H, s), 2.02 (3H, s), 2.07

(3H, s), 2.11 (3H, s), 2.18 (3H, s), 4.10 (1H, dd, J = 2.3, 12.5), 4.28 (1H, t, J = 3.7), 4.78

(1H, ddd, J = 1.6, 3.9, 9.1), 5.06 (1H, d, 4.0), 5.13 (1H, t, J = 9.8), 5.33 (1H, d, J = 4.5),

5.79 (1H, d, J = 9.0), 5.86 (1H, d, J = 1.6). 13C-NMR (125 MHz, CDCl3): δ 20.88, 20.90,

90

20.92, 20.96, 20.97, 21.01, 21.08, 23.56, 23.65, 49.52, 49.74, 62.20, 65.41, 65.62, 68.99,

70.29, 71.56, 73.68, 90.86, 91.90, 168.34, 168.55, 169.92, 169.93, 170.23, 170.32,

170.73, 170.74, 170.82. ESI-MS for C16H23NO10 [M], calculated for 389.13, found

389.12. 1H-NMR, 13C-NMR, and ESI-MS spectra are presented in the appendix.

Synthesis of Ac5ManNGc

Monoacetylated ManNGc was synthesized as previously reported.53 Briefly, to a

solution of D-(+)-mannosamine hydrochloride (216 mg, 1.00 mmol) and sodium

bicarbonate (1.68 g, 20 mmol) in water (8.6 mL, 480 mmol) chilled on ice, acetoxyacetyl

chloride (537 µL, 5.00 mmol) was added dropwise and the reaction was stirred for 3

hours on ice, monitoring reaction progress by TLC (ethyl acetate / acetic acid / water =

3/2/1) using nihydrin to detect unreacted starting material and orcinol-sulfuric acid to

detect sugars. After filtering through Celite in a glass Pasteur pipette, the filtrate was

neutralized with 1.0 M HCl (pH ~ 7, added dropwise). The resulting mixture was

concentrated in vacuo, roughly purified by flash chromatography (ethyl acetate /

isopropanol / water = 27/8/4), and used directly to synthesize the fully acetylated product,

Ac5ManNGc. The acetylation of monoacetylated ManNGc was performed by the same

procedure described in the synthesis of Ac4ManNAc to afford Ac5ManNGc (159 mg,

36%, mixture of anomers). 1H-NMR (500 MHz, CDCl3): δ 1.97 (3H, s), 2.03 (3H, s),

2.07 (3H, s), 2.15 (3H, s), 2.18 (3H, s), 4.03 (1H, dd, J = 1.9, 12.6), 4.23 (1H, d, J = 4.7),

4.58 (2H, s), 4.65 (1H, ddd, J = 1.9, 4.4, 9.0), 5.04 (1H, d, J = 3.8), 5.14 (1H, t, J = 10.3),

5.30 (1H, d, J = 4.3), 6.01 (1H, d, J = 1.3), 6.37 (1H, d, J = 9.2). 13C-NMR (125 MHz,

CDCl3): δ 20.84, 20.86, 20.87, 20.87, 20.89, 20.91, 20.96, 21.06, 49.16, 49.61, 61.96,

62.04, 63.25, 63.30, 65.15, 65.26, 68.99, 70.31, 71.40, 73.63, 90.62, 91.63, 167.41,

168.00, 168.32, 168.53, 169.53, 169.58, 169.75, 169.79, 170.29, 170.32, 170.61, 170.62.

ESI-MS for C18H25NO12 [M], calculated for 447.14, found 447.15. 1H-NMR, 13C-NMR,

and ESI-MS spectra are presented in the appendix.

Synthesis of Ac4ManNProp

ManNProp was synthesized as previously reported.52 Briefly, to a solution of D-

(+)-mannosamine hydrochloride (300 mg, 1.39 mmol) in MeOH (20 mL) and 3.0 M

91

NaOH (0.5 mL), propionic anhydride (1.0 mL, 7.80 mmol) was added dropwise on ice

while stirring for several hours, monitoring reaction progress by TLC. After completion,

1.0 M HCl was added dropwise to neutralize the solution (pH ~ 7). After the solvent was

evaporated in vacuo, dried with several washes of toluene, filtered with cotton in Pasteur

pipette, and evaporated under vacuum. The resulting mixture was roughly purified by

flash chromatography (CH2Cl2 / MeOH = 1/0, 10/1, 4/1) and used directly to synthesize

the acetylated product, Ac4ManNProp. The acetylation of ManNProp was performed by

the same procedure described in the synthesis of Ac4ManNAc to afford Ac4ManNDAz

(284 mg, 51%, mixture of anomers). 1H-NMR (500 MHz, CDCl3): δ 1.19 (3H, t, J =

7.9), 2.00 (3H, s), 2.07 (3H, s), 2.11 (3H, s), 2.19 (3H, s), 2.34 (2H, m), 4.06 (1H, dd, J =

1.4, 12.0), 4.28 (1H, t, J = 4.8), 4.80 (1H, ddd, J = 1.6, 4.4, 9.2), 5.06 (1H, d, J = 4.0),

5.18 (1H, t, J = 10.3), 5.22 (1H, d, J = 4.5), 5.65 (1H, d, J = 9.3), 6.04 (1H, d, J = 4.8). 13C-NMR (125 MHz, CDCl3): δ 9.92, 10.11, 20.88, 20.90, 20.92, 20.94, 20.95, 20.99,

21.09, 29.84, 29.99, 49.28, 49.54, 53.64, 54.65, 62.07, 62.19, 65.40, 65.58, 69.06, 70.26,

71.56, 73.62, 90.88, 91.93, 168.36, 168.53, 169.90, 170.22, 170.29, 170.72, 173.94,

174.59. ESI-MS for C17H25NO10 [M-H]-, calculated for 402.15, found 402.15. 1H-NMR, 13C-NMR, and ESI-MS spectra are presented in the appendix.

Synthesis of Ac4ManNBut

ManNBut was synthesized as previously reported.52 Briefly, to a solution of D-

(+)-mannosamine hydrochloride (300 mg, 1.39 mmol) in MeOH (5.0 mL) and 3.0 M

NaOH (0.5 mL), butyric anhydride (1.0 mL, 6.13 mmol) was added dropwise on ice

while stirring for several hours, monitoring reaction progress by TLC. After completion,

1.0 M HCl was added dropwise to neutralize the solution (pH ~ 7). After the solvent was

evaporated in vacuo, dried with several washes of toluene, filtered with cotton in Pasteur

pipette, and evaporated under vacuum. The resulting mixture was roughly purified by

flash chromatography (CH2Cl2 / MeOH = 1/0, 10/1, 4/1) and used directly to synthesize

the acetylated product, Ac4ManNBut. The acetylated of ManNBut was performed by the

same procedure described in the synthesis of Ac4ManNAc to afford Ac4ManNBut (266

mg, 45%, mixture of anomers). 1H-NMR (500 MHz, CDCl3): δ 1.00 (3H, t, J = 7.3), 1.71

(2H, m), 2.01 (3H, s), 2.07 (3H, s), 2.11 (3H, s), 2.19 (3H, s), 2.25 (2H, t, J = 7.5), 4.06

92

(1H, dd, J = 4.9), 4.28 (1H, t, J = 4.8), 4.68 (1H, ddd, J = 4.6, 9.5, 12.6), 5.06 (1H, d, J =

4.0), 5.18 (1H, t, J = 10.3), 5.33 (1H, d, J = 4.4), 5.65 (1H, d, J = 9.4), 6.04 (1H, s). 13C-

NMR (125 MHz, CDCl3): δ 13.72, 13.82, 19.3, 19.49, 20.88, 20.90, 20.93, 20.95, 21.10,

38.73, 38.90, 49.24, 49.56, 62.06, 62.89, 65.37, 65.55, 69.09, 70.29, 71.60, 73.65, 90.84,

91.94, 168.37, 168.51, 169.88, 170.22, 170.72, 173.16, 173.80. ESI-MS for C18H27NO10

[M-H]-, calculated for 416.16, found 416.17. 1H-NMR, 13C-NMR, and ESI-MS spectra

are presented in the appendix.

Synthesis of Ac4ManNPhAc

To a solution of phenylacetic acid (136 mg, 1.00 mmol), D-(+)-mannosamine

hydrochloride (216 mg, 1.00 mmol) and triethylamine (280 µL, 2.00 mmol) in MeOH (10

mL), 1-ethyl-3-(3-dimethyllaminopropyl)carbodiimide hydrochloride (388 mg, 2.00

mmol) was added. The reaction mixture was stirred on ice for 10 minutes, followed by

stirring at room temperature overnight. The resulting mixture was concentrated in vacuo,

roughly purified by flash chromatography (CH2Cl2 / MeOH = 1/0, 10/1, 4/1), and used

directly to synthesize the acetylated product, Ac4ManNPhAc. The acetylation of

ManNPhAc was performed by the same procedure described in the synthesis of

Ac4ManNAc to afford Ac4ManNPhAc (117 mg, 25%, mixture of anomers). 1H-NMR

(500 MHz, CDCl3): δ 1.94 (3H, s), 2.01 (3H, s), 2.04 (3H, s), 2.16 (3H, s), 3.99 (1H, ddd,

J = 2.2, 6.0, 13.0), 4.06 (1H, dd, J = 2.7, 9.7), 4.66 (1H, ddd, J = 4.3, 9.4, 13.7), 4.97 (2H,

s), 4.99 (1H, t, J = 5.6), 5.02 (1H, d, J = 3.7), 5.28 (1H, d, J = 4.4), 5.63 (1H, t, J = 9.0),

5.96 (1H, d, J = 1.6), 7.34 (3H, m), 7.42 (2H, m). 13C-NMR (125 MHz, CDCl3): δ 20.81,

20.83, 20.84, 20.86, 20.89, 20.92, 21.04, 43.80, 43.99, 49.43, 49.46, 61.94, 61.96, 65.19,

65.24, 69.21, 70.16, 71.35, 73.37, 90.58, 91.71, 127.63, 127.78, 129.22, 129.37, 129.41,

129.49, 134.48, 168.33, 168.36, 169.65, 169.70, 170.20, 170.24, 170.68, 170.70, 171.35.

ESI-MS for C22H27NO10 [M], calculated for 465.16, found 465.16. 1H-NMR, 13C-NMR,

and ESI-MS spectra are presented in the appendix.

Synthesis of Ac4ManNAz

ManNAz was synthesized as previously reported.48 Briefly, to a solution of

azidoacetic acid (360 mg, 3.00 mmol), D-(+)-mannosamine hydrochloride (432 mg, 2.00

93

mmol) and triethylamine (560 µL, 4.00 mmol) in MeOH (20 mL), 1-ethyl-3-(3-





chromatography (CH2Cl2 / MeOH = 1/0, 5/1, 3/1), and used directly to synthesize the

acetylated product, Ac4ManNAz. The acetylation of ManNAz was performed by the

same procedure described in the synthesis of Ac4ManNAc to afford Ac4ManNAz (437

mg, 50%, mixture of anomers). 1H-NMR (500 MHz, CDCl3): δ 1.96 (3H, s), 2.03 (3H, s),

2.08 (3H, s), 2.15 (3H, s), 4.02 (2H, s), 4.09 (1H, ddd, J = 2.0, 8.0, 14.4), 4.20 (1H, d, J =

1.8), 4.58 (1H, ddd, J = 4.2, 9.3, 13.5), 5.04 (1h, d, J = 3.9), 5.19 (1H, t, J = 10.1), 5.20

(1H, d, J = 4.3), 6.01 (1H, s), 6.65 (1H, d, J = 18.1). 13C-NMR (125 MHz, CDCl3): δ

20.80, 20.84, 20.87, 20.91, 20.95, 20.98, 21.06, 49.44, 49.90, 52.55, 52.74, 61.84, 61.93,

65.08, 65.27, 69.02, 70.41, 71.62, 73.55, 90.42, 91.47, 166.94, 167.52, 168.30, 168.54,

169.76, 170.30, 170.35, 170.73. ESI-MS for C16H22N4O10 [M], calculated for 430.13,

found 430.14. 1H-NMR, 13C-NMR, and ESI-MS spectra are presented in the appendix.

Synthesis of Ac4ManNDAz

ManNDAz was synthesized as previously reported.10 Briefly, to a solution of 4,4-

azo-pentanoic acid27 (128 mg, 1.00 mmol), D-(+)-mannosamine hydrochloride (216 mg,

1.00 mmol) and triethylamine (278 µL, 2.00 mmol) in MeOH (10 mL), 1-ethyl-3-(3-





chromatography (CH2Cl2 / MeOH gradient = 1/0, 10/1, 4/1), and used directly to

synthesize the acetylated product, Ac4ManNDAz. The acetylation of ManNDAz was

performed by the same procedure described in the synthesis of Ac4ManNAc to afford

Ac4ManNDAz (84 mg, 28% over two steps, mixture of anomers). 1H-NMR (500 MHz,

CDCl3): δ 1.06 (3H, s), 1.81 (2H, m), 2.02 (3H, s), 2.07 (3H, s), 2.12 (3H, s), 2.19 (3H,

s), 4.05 (2H, s), 4.09 (1H, ddd, J = 6.2, 18, 30.5), 4.29 (1H, t, J = 4.5), 4.78 (1H, dd, J =

94

2.3, 8.2), 5.06 (1H, d, J = 4.0), 5.20 (1H, t, J = 9.8), 5.32 (1H, d, J = 4.4), 5.80 (1H, d, J =

9.0), 6.04 (1H, s). 13C-NMR (125 MHz, CDCl3): δ 20.18, 20.19, 20.85, 20.88, 20.91,

20.95, 20.97, 21.08, 25.50 25.54, 29.94, 30.05, 30.71, 30.84, 49.55, 49.76, 62.02, 62.15,

65.32, 69.06, 70.32, 71.57, 73.67, 90.80, 91.79, 168.33, 168.53, 169.82, 169.91, 170.20,

170.28, 170.75, 170.78, 171.62, 172.14. ESI-MS for C19H27N3O10 [M], calculated for

457.17, found 457.16. 1H-NMR, 13C-NMR, and ESI-MS spectra are presented in the

appendix.

Cell culturing experiment reagents

RPMI 1640 with 2 mM L-glutamine, α-minimum Eagle’s medium with

glutamine, ribonucleosides, and deoxyribonucleosides, Opti-MEM, fetal calf serum,

penicillin/streptomycin, dPBS, PBS (pH = 7.4), FITC-streptavidin, aminooxy-biotin, and

propidium iodide were purchased from Invitrogen. Nutridoma SP and BSA Fraction V

were purchased from Roche Applied Science. Aniline was purchased from Sigma-

Aldrich. Glycerol was purchased from Fisher Scientific. SNA-FITC was purchased from

EY Labs. MAA-biotin was purchased from Vector Labs. DTAF-streptavidin was

purchased from Jackson Immunoresearch. Cell counting was performed on the

Invitrogen Countess Automated Cell Counter. Flow cytometry experiments were

performed on a BD Biosciences FACSCaliber flow cytometer

Cell culturing conditions

BJAB K20 and K88 cells were grown and maintained in RPMI 1640 with 2 mM

L-glutamine containing 10% fetal calf serum, 100 U/ml penicillin, and 100 µg/ml

streptomycin at 37oC, 5% CO2 in a water-saturated environment. Cells were cultured at

2.5 x 105 cells/mL in media and grown for 48 hours before passaging. Typically, cell

densities were maintained between 2.5 x 105 cells/mL and 2.0 x 106 cells/mL. Cell

viability was analyzed using Trypan blue dye staining with the Countess Automated Cell

Counter instrument.

CHO and CHO Lec3 cells were grown and maintained in α-minimum Eagle’s

medium w/ glutamine, ribonucleosides, and deoxyribonucleosides containing 10% fetal

calf serum at 37oC, 5% CO2 in a water-saturated environment. Cells were cultured at 2.5

95

x 104 cells/mL in media and grown for 72 hours before passaging. Typically, cell

densities were maintained between 2.5 x 104 cells/ml and 2.0 x 106 cells/mL.

Serum free exposure and supplementation with monosaccharides

Initially, CHO and CHO Lec3 cells were plated at 5.0 x 104 cells/ml and allowed

to grow for 24 hours. Cells were then washed 3 times with PBS, then cultured with

OptiMEM media for serum free exposure. Using 10 mM ethanol stocks of each sugar,

ethanol or acetylated sugar was added to each plate while swirling to make a final

concentration of 100 µM. After growth in the presence of the appropriate

monosaccharides for 48 hours, the cells were removed from the plate by exposure to

trypsin for 5 minutes, counted, and harvested by centrifugation at 220g for 5 min in 50 ml

conical tubes.

To generate serum free conditions for BJAB K20 and K88 cells, the cells were

grown in RPMI 1640 with 2 mM L-glutamine containing 1x Nutridoma SP, 50 U/ml

penicillin, and 50 µg/ml streptomycin. Cells were cultured for two passages at 2.5 x 105

cells/ml in media for 72 hours at a time before supplementation with monosaccharides.

Prior to the addition of cells to a tissue culture plates, acetylated sugar or ethanol were

added and the ethanol was pre-evaporated. After cell counting, BJAB cells were plated at

2.5 x 105 cells/ml in serum free media for each monosaccharide condition. After growing

for 72 hours, cells were counted and harvested by centrifugation at 220g for 5 min in 50

ml conical tubes.

To ensure consistent results among all samples, equal numbers of cells were

collected for every samle; the toal number of cells collected for an experiment ranged

between 3.0 – 4.0 x 107 cells for CHO/CHO Lec3 analysis and 1.0 – 1.2 x 108 cells for

BJAB K88/K20 analysis. Cell pellets were stored at -80 oC overnight before proceeding

to ganglioside extraction.

Extraction of gangliosides - Total Lipid Extraction

Cell pellets were thawed to room temperature, resuspended with 300 µl of ice

cold ddH2O (W), and dounced 50 times with a Kontes tissue grinder, tube size 20. With

96

a glass Pasteur pipette and a 2 mL rubber bulb, the cell lysate suspension was transferred

into a 4 mL glass vial containing 800 µL of methanol (M), already stirring. 400 µL of

chloroform (C) was added to the vial and the mixture was stirred thoroughly for 2 hours

at room temperature. Samples were covered in foil to prevent exposure to light. After

stirring, the mixture was transferred by a glass Pasteur pipette into a 13 x 100 mm glass

culture tube and centrifuged at 2800g for 10 min @ 30 oC. The supernatant (containing

the total lipid extract) was transferred by a glass Pasteur pipette into a new 4 mL glass

vial and evaporated to dryness under N2 gas.

Extraction of gangliosides - Phospholipid Extraction

The dried total lipid extract was resuspended with 800 µL butanol and 1200 µL

diisopropyl ether and sonicated in a water bath for 10 minutes. The resuspended lipids

were then transferred into a 13 x 100 mm glass culture tube using a glass Pasteur pipette.

To extract undesired phospholipids from the mixture, 1000 µl of 50 mM NaCl was added

to the tube and mixed vigorously by pipetting up and down repeatedly with a glass

Pasteur pipette. The mixture was then centrifuged at 2800g for 10 min at 30 oC to

separate the two phases. Using a glass Pasteur pipette, the organic phase (top layer) was

carefully removed. The aqueous mixture (bottom layer) was then extracted two more

times using the same ratio of butanol and diisopropyl ether.

Extraction of gangliosides - SepPak purification

After the final extraction, the remaining lipid mixture was loaded onto a SepPak tC18

column, 0.3g size. The column was first pre-treated with three 2 mL washes of C/M/W

(2:43:55) followed by two 2 mL washes of C/M (1:1) and ending with three more 2 mL

washes of C/M/W (2:43:55). After loading of the sample, the column was washed three

times with 2 mL of C/M/W (2:43:55) followed by three 2 mL washes of C/M (1:1) to

desalt the sample and remove unwanted contaminants. Elution of gangliosides was

achieved using 2 mL of 100 % methanol. Ganglioside extracts were then transferred into

a new 4 mL glass vial and evaporated to dryness under N2.

97

HPTLC Analysis of Extracted Gangliosides

Extracted ganglioside samples were redissolved with 30 µL C/M/W (2:1:0.1) and

resolved on HPTLC plates. Ganglioside standards were loaded onto the plate to provide

mass references. Gangliosides were separated with chloroform:methanol:0.2% CaCl2(aq)

(80:45:10) as the running buffer. HPTLC plates were first pre-run before loading 10 µL

of ganglioside extract. After thoroughly drying the plate in a fume hood, gangliosides

were detected by resorcinol staining (0.1% resorcinol, 0.04% CuSO4 in hydrochloric

acid:water [4:1]). Plates were imaged using an Alpha Innotech FluorChem HD2 and

images were processed using Adobe Photoshop.

Mass Spectrometry Analysis of Extracted Gangliosides

Dried ganglioside extracts were sent to the Complex Carbohydrate Research

Center at the University of Georgia for mass spectrometry analysis. To analyze the

overall composition of extracted gangliosides, MALDI-TOF-MS was performed.

Samples were crystallized onto a MALDI plate with trihydroxyacetophenone

monohydrate (THAP) as a matrix. Analysis of gangliosides was performed in the

negative ion mode using a Bruker microflex instrument.

Flow cytometry - Lectin

Cells were resuspended at 1.875 x 106 cells/ml in PBS and aliquoted into a v-

bottom 96-well plate in 200 µl amounts. Analyses were performed in triplicate with three

separate cultures of each condition. Cells were washed 3 times with 200 µL of 0.1%

BSA/PBS and centrifuged at 2,000 rpm, 4 oC for 4 minutes. For MAA binding

experiments, the cells were incubated with 50 µL of 10 µg/ml of MAA-biotin in 0.1%

BSA/PBS for 30 minutes on ice. After being washed 3 times with 200 µL of 0.1%

BSA/PBS, the cells were incubated with 50 µL of 20 µg/ml of streptavidin-FITC for 30

min on ice. After being washed 3 times with 200 µL of 0.1% BSA/PBS, the cells were

resuspended in 400 µl of 0.1% BSA/PBS. Propidium iodide, used to identify dead cells

versus live cells, was added at a final concentration of 50 µg/ml into each tube before

analysis. The cells were analyzed using a FACSCaliber flow cytometer. Live cells

98

(10,000 cells/sample) were identified by their forward scatter versus side scatter plot; all

propidium iodide positive cells (dead cells) were excluded from analysis. FITC

fluorescence was measured on the FL-1 channel of the instrument.

For SNA binding experiments, the cells were incubated with 10 µg/ml of SNA-

FITC in 0.1% BSA/PBS for 30 minutes on ice. After being washed three times with

0.1% BSA/PBS, the cells were resuspended in 400 µl of 0.1% BSA/PBS. Propidium

iodide, used to identify dead cells versus live cells, was added at 50 µg/ml into each tube

before analysis.

Flow cytometry - PAL

Cells were resuspended at 1.0 x 106 cells/ml in PBS and aliquoted into a v-bottom

96-well plate in 200 µl amounts, in biological triplicate. Cells were washed 2x with PBS

(pH = 7.4) and centrifuged at 2,000 rpm, 4oC for 4 minutes. Cells were then resuspended

in 200 µl 1.0 mM NaIO4/PBS (pH = 7.4) and incubated for 30 min on ice. After

incubation, 50 µl of 5.0 mM glycerol/PBS (pH = 7.4) was added and incubated for 30

min on ice to quench the oxidation reaction. Cells were pelleted by centrifugation at

2,000 rpm, 4oC for 4 min to remove the supernatant and resuspended with 200 µl of 5.0%

FBS/PBS (pH = 6.7). The cells were then washed 2 times with 5.0% FBS/PBS (pH =

6.7) and centrifuged at 2,000 rpm, 4 oC for 4 minutes. After washing, the cells were

resuspended in 200 µl of 0.1 mM aminooxy-biotin, 10.0 mM aniline 5.0 % FBS/PBS (pH

= 6.7) and incubated for 90 min on ice. After incubation, cells were washed three times

with 5.0% FBS/PBS (pH = 6.7) and centrifuged at 2,000 rpm, 4oC for 4 minutes. After

washing, the cells were resuspended in 200 µl of 3.2 µg/ml DTAF-Streptavidin 5.0%

FBS/PBS for 30 min on ice. Cells were then washed three times with 5.0% FBS/PBS

(pH = 6.7) and centrifuged at 2,000 rpm, 4oC for 4 minutes. After being washed three

times with 0.1% BSA/PBS, the cells were resuspended in 400 µl of 5.0% FBS/PBS (pH =

6.7). The cells were analyzed using a FACSCaliber Flow Cytometer. Live cells (10,000

cells/sample) were identified by their forward scatter versus side scatter plot. DTAF

fluorescence was measured on the FL-1 channel of the instrument.

99

References


proteins. Nature 446, 1023-1029, (2007).

2 Lowe, J. B. Glycan-dependent leukocyte adhesion and recruitment in

inflammation. Curr Opin Cell Biol 15, 531-538, (2003).

3 Lopez, P. H. & Schnaar, R. L. Gangliosides in cell recognition and membrane

protein regulation. Curr Opin Struct Biol 19, 549-557, (2009).

4 Yu, R. K., Nakatani, Y. & Yanagisawa, M. The role of glycosphingolipid

metabolism in the developing brain. J Lipid Res 50 Suppl, S440-445, (2009).

5 Malykh, Y. N., Schauer, R. & Shaw, L. N-Glycolylneuraminic acid in human

tumours. Biochimie 83, 623-634, (2001).

6 Lacomba, R. et al. Determination of sialic acid and gangliosides in biological

samples and dairy products: a review. J Pharm Biomed Anal 51, 346-357, (2010).

7 Keppler, O. T., Horstkorte, R., Pawlita, M., Schmidt, C. & Reutter, W.

Biochemical engineering of the N-acyl side chain of sialic acid: biological

implications. Glycobiology 11, 11R-18R, (2001).

8 Wang, Z., Du, J., Che, P. L., Meledeo, M. A. & Yarema, K. J. Hexosamine

analogs: from metabolic glycoengineering to drug discovery. Curr Opin Chem

Biol 13, 565-572, (2009).

9 Kayser, H. et al. Biosynthesis of a nonphysiological sialic acid in different rat

organs, using N-propanoyl-D-hexosamines as precursors. J Biol Chem 267,

16934-16938, (1992).

10 Keppler, O. T. et al. Biosynthetic modulation of sialic acid-dependent virus-

receptor interactions of two primate polyoma viruses. J Biol Chem 270, 1308-

1314, (1995).

100

11 Herrmann, M., von der Lieth, C. W., Stehling, P., Reutter, W. & Pawlita, M.

Consequences of a subtle sialic acid modification on the murine polyomavirus

receptor. J Virol 71, 5922-5931, (1997).

12 Schmidt, C., Stehling, P., Schnitzer, J., Reutter, W. & Horstkorte, R. Biochemical

engineering of neural cell surfaces by the synthetic N-propanoyl-substituted

neuraminic acid precursor. J Biol Chem 273, 19146-19152, (1998).

13 Mahal, L. K. et al. A small-molecule modulator of poly-alpha 2,8-sialic acid

expression on cultured neurons and tumor cells. Science 294, 380-381, (2001).

14 Mahal, L. K., Yarema, K. J. & Bertozzi, C. R. Engineering chemical reactivity on

cell surfaces through oligosaccharide biosynthesis. Science 276, 1125-1128,

(1997).

15 Saxon, E. & Bertozzi, C. R. Cell surface engineering by a modified Staudinger

reaction. Science 287, 2007-2010, (2000).

16 Han, S., Collins, B. E., Bengtson, P. & Paulson, J. C. Homomultimeric complexes

of CD22 in B cells revealed by protein-glycan cross-linking. Nat Chem Biol 1, 93-

97, (2005).

17 Hsu, T. L. et al. Alkynyl sugar analogs for the labeling and visualization of

glycoconjugates in cells. Proc Natl Acad Sci U S A 104, 2614-2619, (2007).


1382-1401, (2009).

19 Prescher, J. A., Dube, D. H. & Bertozzi, C. R. Chemical remodelling of cell

surfaces in living animals. Nature 430, 873-877, (2004).

20 Dube, D. H., Prescher, J. A., Quang, C. N. & Bertozzi, C. R. Probing mucin-type

O-linked glycosylation in living animals. Proc Natl Acad Sci U S A 103, 4819-

4824, (2006).

101

21 Baskin, J. M. et al. Copper-free click chemistry for dynamic in vivo imaging.

Proc Natl Acad Sci U S A 104, 16793-16797, (2007).

22 Laughlin, S. T., Baskin, J. M., Amacher, S. L. & Bertozzi, C. R. In vivo imaging

of membrane-associated glycans in developing zebrafish. Science 320, 664-667,

(2008).

23 Chang, P. V. et al. Metabolic labeling of sialic acids in living animals with

alkynyl sugars. Angew Chem Int Ed Engl 48, 4030-4033, (2009).

24 Chang, P. V. et al. Copper-free click chemistry in living animals. Proc Natl Acad

Sci U S A 107, 1821-1826, (2010).

25 Bond, M. R., Zhang, H., Vu, P. D. & Kohler, J. J. Photocrosslinking of

glycoconjugates using metabolically incorporated diazirine-containing sugars. Nat

Protoc 4, 1044-1063, (2009).

26 Tanaka, Y. & Kohler, J. J. Photoactivatable crosslinking sugars for capturing

glycoprotein interactions. J Am Chem Soc 130, 3278-3279, (2008).



28 Zou, W. et al. Bioengineering of surface GD3 ganglioside for immunotargeting

human melanoma cells. J Biol Chem 279, 25390-25399, (2004).

29 Pan, Y., Chefalo, P., Nagy, N., Harding, C. & Guo, Z. Synthesis and

immunological properties of N-modified GM3 antigens as therapeutic cancer

vaccines. J Med Chem 48, 875-883, (2005).

30 Chefalo, P., Pan, Y., Nagy, N., Guo, Z. & Harding, C. V. Efficient metabolic

engineering of GM3 on tumor cells by N-phenylacetyl-D-mannosamine.

Biochemistry 45, 3733-3739, (2006).

102

31 Wang, Q., Zhang, J. & Guo, Z. Efficient glycoengineering of GM3 on melanoma

cell and monoclonal antibody-mediated selective killing of the glycoengineered

cancer cell. Bioorg Med Chem 15, 7561-7567, (2007).

32 Collins, B. E., Fralich, T. J., Itonori, S., Ichikawa, Y. & Schnaar, R. L.

Conversion of cellular sialic acid expression from N-acetyl- to N-

glycolylneuraminic acid using a synthetic precursor, N-glycolylmannosamine

pentaacetate: inhibition of myelin-associated glycoprotein binding to neural cells.

Glycobiology 10, 11-20, (2000).

33 Jacobs, C. L. et al. Substrate specificity of the sialic acid biosynthetic pathway.

Biochemistry 40, 12864-12874, (2001).

34 Keppler, O. T. et al. Differential sialylation of cell surface glycoconjugates in a

human B lymphoma cell line regulates susceptibility for CD95 (APO-1/Fas)-

mediated apoptosis and for infection by a lymphotropic virus. Glycobiology 9,

557-569, (1999).

35 Keppler, O. T. et al. UDP-GlcNAc 2-epimerase: a regulator of cell surface

sialylation. Science 284, 1372-1376, (1999).

36 Hong, Y. & Stanley, P. Lec3 Chinese hamster ovary mutants lack UDP-N-

acetylglucosamine 2-epimerase activity because of mutations in the epimerase

domain of the Gne gene. J Biol Chem 278, 53045-53054, (2003).

37 Rosales Fritz, V. M., Daniotti, J. L. & Maccioni, H. J. Chinese hamster ovary

cells lacking GM1 and GD1a synthesize gangliosides upon transfection with

human GM2 synthase. Biochim Biophys Acta 1354, 153-158, (1997).

38 Jones, M. B. et al. Characterization of the cellular uptake and metabolic

conversion of acetylated N-acetylmannosamine (ManNAc) analogues to sialic

acids. Biotechnol Bioeng 85, 394-405, (2004).

39 Ladisch, S. & Gillard, B. A solvent partition method for microscale ganglioside

purification. Anal Biochem 146, 220-231, (1985).

103

40 Schnaar, R. L. Isolation of glycosphingolipids. Methods Enzymol 230, 348-370,

(1994).

41 Nimrichter, L. et al. E-selectin receptors on human leukocytes. Blood 112, 3744-

3752, (2008).

42 Zeng, Y., Ramya, T. N., Dirksen, A., Dawson, P. E. & Paulson, J. C. High-

efficiency labeling of sialylated glycoproteins on living cells. Nat Methods 6, 207-

209, (2009).

43 Lee, E. U., Roth, J. & Paulson, J. C. Alteration of terminal glycosylation

sequences on N-linked oligosaccharides of Chinese hamster ovary cells by

expression of beta-galactoside alpha 2,6-sialyltransferase. J Biol Chem 264,

13848-13855, (1989).

44 Parker, R. B. & Kohler, J. J. Regulation of intracellular signaling by extracellular

glycan remodeling. ACS Chem Biol 5, 35-46, (2010).

45 Luchansky, S. J., Argade, S., Hayes, B. K. & Bertozzi, C. R. Metabolic

functionalization of recombinant glycoproteins. Biochemistry 43, 12358-12366,

(2004).

46 Luchansky, S. J. et al. Constructing azide-labeled cell surfaces using

polysaccharide biosynthetic pathways. Methods Enzymol 362, 249-272, (2003).

47 Luchansky, S. J., Goon, S. & Bertozzi, C. R. Expanding the diversity of unnatural

cell-surface sialic acids. Chembiochem 5, 371-374, (2004).

48 Brinkman-Van der Linden, E. C., Sonnenburg, J. L. & Varki, A. Effects of sialic

acid substitutions on recognition by Sambucus nigra agglutinin and Maackia

amurensis hemagglutinin. Anal Biochem 303, 98-104, (2002).

49 Markotic, A., Lumen, R., Marusic, A., Jonjic, S. & Muthing, J. Ganglioside

expression in tissues of mice lacking the tumor necrosis factor receptor 1.

Carbohydr Res 321, 75-87, (1999).

104

50 Pan, Y., Ayani, T., Nadas, J., Wen, S. & Guo, Z. Accessibility of N-acyl-D-

mannosamines to N-acetyl-D-neuraminic acid aldolase. Carbohydr Res 339,

2091-2100, (2004).

51 Church, R. F. R. & Weiss, M. J. Diazirines. II. Synthesis and Properties of Small

Functionalized Diazirine Molecules - Some Observations on Reaction of a

Diaziridine with Iodine-Iodide Ion System. Journal of Organic Chemistry 35,

2465-2471, (1970).

105

Appendix

106

1H NMR Data

110

13C NMR Data

114

ESI-MS Data

118

HPLC Data

122

MALDI-TOF-MS Data

UNDERSTANDING AND MANIPULATING GANGLIOSIDE BIOSYNTHESIS A DISSERTATION …qp110yy2867/Final... ·...

Documents

Transcript of UNDERSTANDING AND MANIPULATING GANGLIOSIDE BIOSYNTHESIS A DISSERTATION …qp110yy2867/Final... ·...