University of Southern California Dissertations and Theses

Using chemical biology approaches to investigate the consequences of protein concentration and activity in cancer cells

(USC Thesis Other)

Using chemical biology approaches to investigate the consequences of protein concentration and activity in cancer cells

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content USING CHEMICAL BIOLOGY APPROACHES TO
INVESTIGATE THE CONSEQUENCES OF
PROTEIN CONCENTRATION AND ACTIVITY
IN CANCER CELLS
by

YU HSUAN LIN

A Thesis Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the Requirements for the Degree
DOCTOR OF PHILOSOPHY
(CHEMISTRY)

August 2016

Copyright 2016 Yu Hsuan Lin

I

Acknowledgement
To my parents (Sabrina Yang and Clark Lin), for your understanding,
support and all the sacrifices made for the family, for giving me the freedom
to choose the major I wanted, to chase the dream of studying abroad. Without
you, I would never come to the U.S. and achieve this goal. To my brother
Aaron, for being my best friend in life, for your powerful support and for
undertaking the actions of filial piety, which is deeply rooted in our mind,
while I’m away from home. To Balyn for being my role model, for teaching me
countless experiment techniques and skills. To Franziska for introducing me
molecular biology techniques. To Tharindu, for being an incredible listener
and a great company. I cannot imagine how difficult it is to achieve this goal
without you. I’m blessed to have had you by my side for the large part of my
PhD. To Kelly, for being the leader, among graduate students, of the lab after
Balyn graduated, for being a great company and a good friend, for patiently
teaching me countless English words and slangs that I hadn’t heard of. To
Chathu, for helping me finish my thesis with his template, for those coffee
breaks and casual conversations in the hallway that brightens up my day. To
all my friends, for all your support, company and understanding. To my
committee member Dr. Wang, for enlightening me during my defense. To my
committee member Dr. Zhang, for your invaluable advice and guidance
throughout my PhD study as well as for my job hunting. Thank you for being a
good listener. And most importantly, to my mentor Matt for taking me on
board into the dream lab that I wanted when I applied USC, for trusting me

II

and handing a molecular biology project to me when I knew nothing about the
subject, for training me as well as guiding me throughout numerous projects,
for his patience with my mistakes and failures, for teaching me how to
examine a question as deep and as broad. And lastly, to my beloved girlfriend
Vicky, for brightening up my days, for your care and company, for your
optimism when I’m in dark, for your smile and your love.

III

Abstract
In Pratt lab, my research focus can be divided into two main topics.
During the first two years of graduate research, I developed a dual small-
molecule rheostat that allows scientists to fine-tune the concentration of their
protein of interests in living mammalian cells. One of the most successful
strategies for controlling protein concentrations in living cells relies on
protein destabilization domains (DD). Under normal conditions, a DD will be
rapidly degraded by the proteasome. However, the same DD can be stabilized
or “shielded” in a stoichiometric complex with a small molecule, enabling
dose-dependent control of its concentration. This process has been exploited
by several labs to post-translationally control the expression levels of proteins
in vitro as well as in vivo, although the previous technologies resulted in
permanent fusion of the protein of interest to the DD, which can affect
biological activity and complicate results. Right before I got recruited, the
Pratt lab reported a complementary strategy, termed traceless shielding
(TShld), in which the protein of interest is released in its native form. I was
involved in the development of an optimized protein concentration control
system, TTShld, which retains the traceless features of TShld but utilizes two
tiers of small molecule control to set protein concentrations in living cells.
These experiments provide the first protein concentration control system that
results in both a wide range of protein concentrations and proteins free from
engineered fusion constructs. The TTShld system has a greatly improved
dynamic range compared to our previously reported system, and the traceless

IV

feature is attractive for elucidation of the consequences of protein
concentration in cell biology.
During the last four years of my PhD, I switched my focus onto
studying how the activity of glutamine fructose-6-phosphate (GFAT)
influences the global level of O-GlcNAc modification in cancer cells. Nutrient-
sensitive O-GlcNAc modification is a type of post-translational modification
that occurs on intracellular proteins. The modification is not only important
for numerous aspects of cell biology but also misregulated in various disease-
relevant signaling pathways, including neurodegeneration diseases as well as
cancer. The modification is up-regulated in virtually all cancer diagnosed so
far and the hyper-O-GlcNAcylation has been shown to be indispensable for a
variety aspects of tumorigenesis. Governing the rate-limiting step of
hexosamine biosynthetic pathway (HBP), GFAT controls the amount of
glucose-derived carbon that enters HBP, the end product of which is the high-
energy donor nucleotide UDP-GlcNAc that can be utilized by O-GlcNAc
transferase (OGT) to catalyze O-GlcNAc addition onto modified substrates.
By decreasing GFAT activity using a small molecule inhibitor, we
demonstrated GFAT as a handle to lower the hyper-O-GlcNAcylation inside
cancer cells, in turn weakened cancer cells to survive oxidative stress.
Furthermore, through lowering GFAT abundance by targeting its mRNA, we
discovered cancer cells originated from different organs have different
capacity to maintain their hyper-O-GlcNAcylation in response to the
decreased GFAT level. Lastly, by silencing GFAT through modifying its DNA

V

in H1299 non-small cell lung cancer cell line, we discovered that GFAT is
indispensable for the cancer cells to maintain their hyper-O-GlcNAcylation as
well as viability, only if we didn’t provide them with GlcNAc, which supports
O-GlcNAcylation by entering HBP. Moreover, losing GFAT expression not
only took away anchorage-independent growth ability away from the H1299
cells, but also inhibited their in vivo tumor growth rate in xenograft tumor
model. Taken together, our data shed lights on a therapeutic window towards
cancer, where lowering hyper-O-GlcNAcylation through inhibition of GFAT
could be a way to slow down tumor progression by targeting cancer that are
not flexible in maintaining their hyper-O-GlcNAcylation.

VI

TABLE OF CONTENTS
Acknowledgement .............................................................................................. I
Abstract ........................................................................................................... III
Table of Contents ............................................................................................ VI
Table of Figures ............................................................................................... IX
Chapter 1. Cancer Metabolism and O-GlcNAc Modification .......................... 1
1.1 Introduction .......................................................................................... 1
1.2 Hexosamine Biosynthetic Pathway ..................................................... 6
1.3 The Biochemical Consequences of O-GlcNAc
Modification in Cancer ............................................................................... 9
1.3.1 O-GlcNAc and cancer cell proliferation ..................................... 9
1.3.2 O-GlcNAc and cancer cell survival and stress resistance ........ 10
1.3.3 O-GlcNAc and cancer cell metastasis and invasion ................. 12
1.4 Reference ............................................................................................ 12
Chapter 2. Approaches to perturb protein functions ................................... 17
2.1 Introduction ........................................................................................ 17
2.2 Approaches at protein level ............................................................... 20
2.3 Approaches at DNA level ................................................................... 24
2.3.1 Knock-out and Knock-in ......................................................... 24
2.3.2 Conditional control of gene function ....................................... 25
2.3.3 Genome editing tools ................................................................ 27
2.4 Approaches at RNA level ................................................................... 32
2.5 Reference ........................................................................................... 34
Chapter 3. A Dual Small-Molecule Rheostat for
Precise Control of Protein Concentration in Mammalian Cells ..................... 38
3.1 Introduction ....................................................................................... 38
3.2 Results ................................................................................................. 41

VII

3.2.1 Generation and characterization of TTshld-GFP cell lines ..... 41
3.2.2 Generation and characterization of
TTshld(F36V)-GFP Hek cell .............................................................. 44
3.2.3 Quantitative analysis of TTshld-GFP ...................................... 46
3.2.4 Demonstration of generality of TTshld(F36V) system ............ 47
3.3 Discussion and Conclusions .............................................................. 49
3.4 Materials and Methods ...................................................................... 50
3.4.1 Plasmid Construction .............................................................. 50
3.4.2 Cell Culture .............................................................................. 50
3.4.3 Transfections and Generation of Stable Cell Lines .................. 51
3.4.4 Cell Culture Experiments ......................................................... 51
3.4.5 Western Blotting ....................................................................... 52
3.4.6 Flow Cytometry Experiments ................................................... 53
3.5 Reference ........................................................................................... 53
Chapter 4. Role of Glutamine Fructose-6-phosphate Amidotransferase
(GFAT) in O-GlcNAc modification and Tumorigenesis ................................. 56
4.1 Introduction ....................................................................................... 56
4.2 Results ................................................................................................ 60
4.2.1 Inhibition of GFAT sensitizes
cancer cells to oxidative stress ........................................................... 60
4.2.2 Generation and Characterization of GFAT KD cell lines ......... 67
4.2.3 Generation and Characterization of GFAT KO H1299 cells .... 71
4.2.3.1 Design of 20-nucleotide guide RNA ........................ 72
4.2.3.2 Transfection and Generation of H1299 ................... 74
4.2.3.3 Isolation and characterization of
clonal knock-out cell lines ……….……………………………….……. 76
4.2.3.4 Effect of GFAT KO on Tumorigenesis ....................... 81

VIII

4.3 Discussion and Conclusions .............................................................. 84
4.4 Materials and Methods ...................................................................... 90
4.4.1 Cell Culture .............................................................................. 90
4.4.2 DON, Diamide and GlcNAc treatment ..................................... 91
4.4.3 Characterization of GFAT KD cell lines ................................... 91
4.4.4 Western blotting analysis of H1299 CRISPR
GFAT1, Mix-target cells and GFAT KO cells ...................................... 91
4.4.5 MTS Cell Proliferation Assay ................................................... 92
4.4.6 Caspase 3/7 Activity Assay ...................................................... 93
4.4.7 Customized GFAT KD plasmids and
Construction of CRISPR GFAT plasmids .......................................... 94
4.4.8 Transfections and Generation of Stable GFAT KD Cell Lines . 95
4.4.9 Transfections and Clonal Selection of H1299 GFAT KO cells 96
4.4.10 Genomic PCR and BP Recombination Reactions .................... 97
4.4.11 Western Blotting ...................................................................... 98
4.4.12 Soft Agar Colony Formation Assay ......................................... 99
4.4.13 H1299 GFAT KO Growth Curve Experiment ......................... 101
4.5 Reference ......................................................................................... 102
Bibliography .................................................................................................. 106

IX

Table of Figures
Figure 1-1. Comparasion between oxidative phosphorylation,
anaerobic glycolysis and aerobic glycolysis ....................................................... 2

Figure 1-2. Hexosamine biosynthetic pathway .................................................. 7
Figure 2-1. Comparasion of techniques in perturbating protein function ...... 19
Figure 3-1. TShld system .................................................................................. 41
Figure 3-2. Analysis of TTshld-GFP in stably transformed cells ................... 43
Figure 3-3. Inducibility of FK506 in TTshld-GFP HEK
and analysis of TTshld(F36V)-GFP HEK ........................................................ 45
Figure 3-4. Quantitative analysis of TTShld-GFP ........................................... 47
Figure 3-5. Analysis of TTShld(F36V)-PKM2 and
TTShld(F36V)-Caspase-3.. .............................................................................. 48
Figure 4-1. Structure of OGT inhibitors ........................................................ 58
Figure 4-2. Effect of DON on H1299 ............................................................... 61
Figure 4-3. Effect of DON and Diamide treatment on
O-GlcNAc level in three different lung cancer cells ........................................ 62
Figure 4-4. Effect of DON and Diamide treatment on
cell proliferation in three different lung cancer cells ..................................... 63
Figure 4-5. Effect of DON and Diamide treatment on
Caspase-3 and/ or Caspase-7 activation in three different lung cancer cells 64
Figure 4-6. Effect of GlcNAc supplementation on
O-GlcNAc modification under DON treatment ............................................... 65
Figure 4-7. Effect of DON, Diamide and GlcNAc treatment
on Caspase-3/ or Caspase-7 activation in three different lung cancer cells .. 66
Figure 4-8. Wetern blotting analysis of MCF7 GFAT1 KD,
MCF7 GFAT2 KD, A549 GFAT1 KD, A549 HFAT2 KD,
H1299 GFAT1 KD and H1299 GFAT2 KD.. .................................................... 69

X

Figure 4-9. Western blotting analysis of MDA-MB-231 GFAT1 KD,
GFAT2 KD and double-knockdown cells ........................................................ 70
Figure 4-10. Schematic of the guide RNA design ............................................ 73
Figure 4-11. Western blotting analysis of H1299 CRISPR GFAT1
and H1299 CRISPR Mix-target cells ................................................................ 75
Figure 4-12. Western blotting analysis of H1299 CRISPR GFAT1
and H1299 CRISPR Mix-target clonal populations ......................................... 76
Figure 4-13. Schematic illustation of two cloning events
in genomic characterization of H1299 GFAT KO cells .................................... 77
Figure 4-14. Characterization of genomic mutations in each
clonal population .............................................................................................. 78
Figure 4-15. Western blotting analysis of H1299
GFAT KO cells under different concentration of GlcNAc supplmentation .... 79
Figure 4-16. Growth curve of H1299 GFAT KO cells
under different amount of GlcNAc suppevaluation lmenentations ................ 81
Figure 4-17. Evaluation of anchorage-independent
growth ability of H1299 GFAT KO cells .......................................................... 82
Figure 4-18. Evaluation of tumor cell growth of H1299 ................................. 83

1

Chapter 1. Cancer Metabolism and O-GlcNAc
Modification
1.1 Introduction
In our body, different tissues execute different types of metabolism.
Supplied with abundant of oxygen, nonproliferating (differentiated) tissues go
through oxidative phosphorylation, in which they first metabolize glucose to
pyruvate through glycolysis in the cytosol and then completely oxidize most of
that pyruvate in the mitochondria to CO2 in the tricarboxylic acid (TCA) cycle.
This is the most efficient way for the cells in the tissues to generate adenosine
5’-triphosphate (ATP) because around 36 moles of ATP are generated for
every mole of glucose consumed. Because oxygen is required as the final
electron acceptor in the oxidative phosphorylation, oxygen is indispensable in
the process. When oxygen is limiting, differentiated tissues then redirect the
pyruvate generated by glycolysis away from mitochondria oxidative
phosphorylation for generation of large amounts of lactate, termed anaerobic
glycolysis. Anaerobic glycolysis results in minimal ATP production when
compared with oxidative phosphorylation since only 2 moles of ATP are
generated for every mole of glucose consumed (Figure 1-1)(Vander Heiden et
al., 2009).
In contrast to normal differentiated tissues, highly proliferative tissues
and most tumors go through a special type of metabolism called aerobic
glycolysis: converting most of the glucose taken up by the cells into lactate

2

regardless of oxygen supply (Figure 1-1), also known as Warburg effect.
Aerobic glycolysis is not an efficient way of generating ATP since only about 4
moles of ATP is produced for every mole of glucose consumed by the cells.
Nobel Laureate Otto Warburg thought the reason cancer cells tend to ferment
glucose into lactate even under abundant oxygen supply is because of their
mitochondria is dysfunctional; however, later research showed that
mitochondrial function is not impaired in most cancer cells(Fantin et al.,
2006; Wallace, 2012), suggesting other explanation for this rewired
metabolism within tumors.
Figure 1-1. Comparison between oxidative phosphorylation, anaerobic glycolysis
and aerobic glycolysis. Cells in differentiated tissues under go two different types of
metabolism. In the presence of abundant oxygen, glucose-derived pyruvate is mainly shuttled
into mitochondria for oxidative phosphorylation. This reaction generates roughly 36 moles of
ATP for every mole of glucose metabolized by the cells. Under low level of oxygen, cells then
go though anaerobic glycolysis, in which they redirect the pyruvate away from mitochondria
in generation of lactate. This reaction can only generate 4 moles of ATP for every mole of
glucose consumed. For proliferating tissues and tumors, the cells tend to direct the pyruvate
generated from glycolysis away from mitochondria and generate a large amount of lactate
regardless of presence of oxygen, so called aerobic glycolysis. Figure adapted from (Vander
Heiden et al., 2009).

3

Why is a less efficient metabolism, at least in terms of ATP production,
preferred in proliferating cells as well as in cancer cells? Vander Heiden et al.
suggested two sound explanations for it. They first suggested that inefficient
ATP production is not a problem for proliferating mammalian cells, since
those cells are always supplied with continuous amount of glucose and
nutrient from the blood stream. In order to compensate for the lower
production rate of ATP, cancer cells upregulate glucose transporters, notably
GLUT1, which results in an increase in glucose uptake from their environment
into cytoplasm(Hsu and Sabatini, 2008). The second explanation they
suggested is that proliferating cells need to utilize glucose-derived carbon
source for biomass production. Therefore, oxidizing most of the glucose to
CO2 via oxidative phosphorylation in mitochondria runs against to the needs
of a carbon source for biomass-generating pathways, such as pentose
phosphate pathway to generate ribose 5-phosphate and NADPH for
nucleotide biosynthesis, lipid synthesis from glucose-derived citrate and
amino acid synthesis from glycolytic intermediates. From this point of view,
this rewired glucose metabolism is actually indispensable for cancer cells to
support its unstoppable cell proliferation and has been proposed to be one of
the emerging hallmarks of cancer (Hanahan and Weinberg, 2011a).
While the molecular mechanism behind the switch from oxidative
phosphorylation to aerobic glycolysis in cancer cells is still an open debate,
many of the predominant mutations observed in cancer as well as loss of
tumor suppressors have been associated with the reprogrammed metabolism

4

because most of them assumed roles in metabolic regulation(Jones and
Thompson, 2009). For example, upregulation of PI3K/Akt signaling
enhanced transcription of genes involved in glycolysis. Among those genes,
M2 isoform of pyruvate kinase (PKM2), which regulates a rate-limiting step of
glycolysis, has shown to be important for cancer cells to execute aerobic
glycolysis as well as their survival under stressed conditions. Enzymatic
activity of PKM2 can be controlled by its alteration between tetrameric (high-
activity) and dimeric (low-activity) states. The ratio of tetramer and dimer of
PKM2 is influenced by multiple factors, including nutrient
availability(Chaneton et al., 2012), growth signaling(Hitosugi et al., 2009)
and cellular stress(Anastasiou et al., 2011). Therefore, coupled with the
reprogrammed glucose metabolism, the expression of PKM2 offers cancer
cells a handle to control the flow of glycolytic metabolites either into biomass
generating pathways branching off from the earlier steps of glycolysis or into
pyruvate for ATP production. Specifically, lower activity of PKM2 has been
shown to correlate with higher ribose phosphate production, which is a key
intermediate for nucleotide production, in human lung cancer
cells(Anastasiou et al., 2012).
Another pathway that the glycolytic metabolites can be shuttled into is
hexosamine biosynthetic pathway (HBP), which branches off from glycolysis
by the rate-limiting enzyme: glutamine fructose-6-phosphate
amidotransferase (GFAT). The end product of the HBP is UDP-N-
acetylglucosamine (UDP-GlcNAc), which can be utilized by O-GlcNAc-

5

transferase (OGT) to install N-Acetyl glucosamine (O-GlcNAc) through a
glycosyl linkage onto serine or threonine residues of nuclear, cytosolic and
mitochondrial proteins in plants and animals(Zachara and Hart, 2002).
Unlike other glycosylation events, such as N-glycosylation on cell surface
proteins, O-GlcNAcylation rapidly cycles on and off modified proteins in a
given time that is shorter than the life span of a protein. The modification is
indispensable for development in mammals and Drosophila because
knockouts of OGT are not viable(Shafi et al., 2000). O-GlcNAc modification
can be deemed as a nutrient sensor because the biosynthesis of UDP-GlcNAc
is regulated by a variety of metabolic pathways, including glucose, amino
acids, fatty acids and nucleotide metabolism(Hardivillé and Hart, 2014). The
modification has been found on large number of proteins and is involved in a
variety of cellular processes. It has been shown to affect enzyme activity(Yi et
al., 2012), stability(Olivier-Van Stichelen et al., 2014) and localization(Zhang
et al., 2014). It also regulates epigenetics(Lewis and Hanover, 2014),
transcription(Ferrer et al., 2014), autophagy(Wang et al., 2012) and ER
protein quality control(Denzel and Antebi, 2015).
The modification is misregulated in a variety of diseases, including
cancer, diabetes and neurodegeneration. Specifically, it has been found to be
upregulated in virtually all cancer types examined so far(Issad, 2013) and the
hyper-O-GlcNAcylation has been thought to play roles in multiple aspects of
tumorigenesis, including cancer cell proliferation, stress resistance,
angiogenesis, invasion and metastasis(Ma and Vosseller, 2013). Further

6

details as to how O-GlcNAc impacts cancer cell growth as well as oncogenic
cellular processes are outlined in 1.3.
1.2 Hexosamine Biosynthetic Pathway
2% to 3% of the glucose that enters the cells is shuttled into the HBP
depending on cell and tissue type(Bouché et al., 2004). Catalyzing the
conversion of fructose-6-phosphate to glucosamine-6-phosphate using
glutamine, GFAT determines the rate for glucose-derived metabolites
entering HBP. Glucosamine 6-phosphate N-acetyltransferas (GAT) next
installs an acetyl group onto the sugar using one equivalent acetyl-CoA to
generate N-acetylglucosamine-6-phosphate. Phosphoglucomutase 3 (PGM3
or AGM) then mutates the sugar to form N-acetylglucosamine-1-phosphate,
the substrate for N-acetylglucosamine pyrophosphorylase (AGX1). AGX1
utilizes uridine triphosphate (UTP) and GlcNAc-1-phosphate to produce UDP-
GlcNAc. Salvage pathways can also introduce glucosamine and GlcNAc into
HBP directly bypassing GFAT for generation of the donor sugar (Figure 1-2).
Extracellular glucosamine can enter the cells through glucose transporter and
then be phosphorylated by hexokinase, entering HBP(Schleicher and Weigert,
2000). Extracellular GlcNAc can be taken up by the cells through bulk
endocytosis(Lau and Dennis, 2008), whereas endogenous pool of GlcNAc can
be replenished from oligosaccharides and glycoproteins through lysosomal
degradation and then phosphorylated by GlcNAc kinase to become GlcNAc-6-
phosphate(Hinderlich et al., 2000), supporting the generation of the donor
sugar by skipping both GFAT and GAT catalysis.

7

Figure 1-2. Hexosamine biosynthetic pathway HBP branches off from glycolysis at the
conversion of fructose-6-phophate to glucosamine 6-phosphate by GFAT. After the
installation of an acetyl group by Glucosamine 6-phosphate N-acetyltransferase (GAT), the
metabolite is then isomerized by phosphoglucomutase 3 (PGM3 or AGM) to yield GlcNAc-1-
phosphate. In the final step of HBP, UDP-GlcNAc pyrophosphorylase (AGX1) uses UTP and
GlcNAc-1-phosphate to produce UDP-GlcNAc. The resulting donor sugar can then be utilized
by OGT to modify its substrates. O-GlcNAcase (OGA) can remove the modification.

Integrating multiple nutrient inputs, the HBP generates different
amounts of UDP-GlcNAc in response to the level of nutrients provided to the
cells. The concentration of the donor sugar can also vary between different
organelles inside the cells, depending on the presence of an active nucleotide
transporter, as well as different permeability associated with each
compartment(Bond and Hanover, 2015). UDP-GlcNAc is an important
substrate for three major types of glycosylation events in the cells: N-
glycosylation, mucin-type O-linked glycosylation as well as the O-
GlcNAcylation. N-glycosylation occurs in endoplasmic reticulum (ER), where
a high-mannose glycan is preassembled on the lipid carrier dolichol
Protein
HO CH
3
(H)
O
O CH
3
(H)
OH
NHAc
HO
HO
Protein
GFAT
O-GlcNAc
transferase O-GlcNAcase
GlcNH2-6-
phosphate
Fructose-6-
phosphate
Glucose
GAT
N-Acetylglucosamine
-6-phosphate
AGM
N-Acetylglucosamine
-1-phosphate
UDP-GlcNAc
(OGT) (OGA)
AGX1
GlcNAc kinase
N-Acetylglucosamine
Hexokinase
Glucosamine

8

phosphate and then transferred to the side-chain asparagine residues that are
specified by consensus Asn-X-Ser/Thr of nascent ER proteins(Aebi, 2013).
UDP-GlcNAc is the substrate for the first two biosynthetic steps in the
construction of the oligosaccharide. N-glycosylation is important for the
structure, as well as the function, of the cell surface proteins, and it also helps
proteins to fold properly during its maturation in the ER. Mucin-type O-
linked occurs in the Golgi apparatus and is initiated by polypeptide N-acetyl-
α-galactosaminyltransferases (ppGalNAcTs) using UDP-N-
acetylgalactosamine (UDP-GalNAc), which is produced from UDP-GlcNAc by
the UDP-galactose-4-epimerase (GALE). Mucin-type O-linked modification of
proteins has been shown to alter glycans on cancer cells, regulates half-life of
serum and modulates signaling pathways(Hang and Bertozzi, 2005).
Unlike N-glycosylation and mucin-type O-linked glycosylation that
both involve in multiple enzymes in the event, O-GlcNAcylation only requires
one enzyme to install the modification as well as one enzyme to remove it.
OGT is the sole enzyme that catalyzes the addition of O-GlcNAc to about 4000
substrates(Ma and Hart, 2014). Three different OGT isoforms are generated
from one single hOGT gene through alternative splicing: nucleocytoplasmic
OGT (ncOGT), mitochondrial OGT (mOGT) as well as the shortest isoform of
OGT (sOGT). The three splice variants have different number of N-terminal
tetratricopeptide repeat (TPR) motifs, giving rise to their ability to modulate
substrate specificity(Bond and Hanover, 2015). OGA is the sole enzyme
responsible for removing O-GlcNAc and it has two splice isoforms: the long

9

and short isoforms (OGA-L and OGA-S). Both isoforms have the same
catalytic domains but have different organelle distribution in the cells.
1.3 The Biochemical Consequences of O-GlcNAc Modification in
Cancer
1.3.1 O-GlcNAc and cancer cell proliferation
Sustaining proliferative signalling and evading growth suppressors are
the fuels of unlimited replicative potential in cancer cells(Hanahan and
Weinberg, 2011b). O-GlcNAcylation has been linked to cancer cell’s
uncontrolled proliferation through its influence on two important cell cycle
regulators, forkhead box protein M1 (FOXM1) and cyclin D1. FOXM1 is a
transcriptional regulator highly expressed in various cancers and it plays an
important role in cell cycle progression by stimulating expression of Skp2 and
Cks1, which are involved in the proteolysis of p27
kip1
to facilitate G1/S
transition(Wang et al., 2005). High level of OGT has been shown to increase
FOXM1 activity in breast cancer MCF-10A cells(Caldwell et al., 2010).
Furthermore, down regulation of OGT through shRNA in a prostate cancer
cell line, PC3-ML, was associated with increased FOXM1 degradation through
a proteasome-dependent process(Lynch et al., 2012). Therefore, O-
GlcNAcylation is implicated in the stability of FOXM1. Cyclin D1 is a key
regulator of the G1 phase that phosphorylates and activates cyclin-dependent
kinase 4 (CDK4) and CDK6. Activated CDK4/CDK6 then phosphorylates the
tumor suppressor RB protein family, promoting the activation of E2F-
responsive genes that are essential for DNA synthesis(Musgrove et al., 2011).

10

Cyclin D1 is down-regulated in siOGT-transfected ovarian cancer cell line
OVCAR-4(Kwei et al., 2012). Furthermore, down regulation of OGT protein
level through siRNA results in decreased global O-GlcNAcylation level in
MCF7 cells and it also delays serum-stimulated cyclin D1 synthesis and cell
proliferation(Stichelen et al., 2012). Taken together, hyper-O-GlcNAcylation
contributes to the deregulated cell cycle control in cancer cells through up-
regulation of positive regulators of cell cycle progression.
1.3.2 O-GlcNAc and cancer cell survival and stress resistance
Tumors reside in a harsh environment where they are constantly
subject to a number of physiologic stresses, including nutrient restriction,
oxidative stress and hypoxia. Therefore, cancer cells evolve not only
capabilities to cope with stressful condition but also mechanisms to resist or
evade cell death, which is recognized as one of the hallmarks of
cancer(Hanahan and Weinberg, 2011b). Notably, cancer cells develop anti-
apoptotic mechanisms through the loss of tumor suppressor TP53, which
activates proteins that are responsible DNA damage response(Junttila and
Evan, 2009). O-GlcNAcylation participates in cancer cell survival by
reinforcing the anti-apoptotic networks. Specifically, Ma and coworkers
showed that lowering hyper-O-GlcNAcylation in pancreatic ductal
adenocarcinoma (PDAC) cell lines by knocking-down OGT decreases the
expression of the anti-apoptotic protein Bcl-xL, and induces cleavage of
caspase-9 and 3, which results in the execution of the intrinsic apoptotic
signals(Ma et al., 2013). Ferrer and coworkers also showed that decreasing O-

11

GlcNAcylation level through down-regulation of OGT leads to ER-mediated
apoptosis in breast cancer MCF7 cell line(Ferrer et al., 2014).
O-GlcNAcylation level are rapidly increase in response to various cell-
and tissue-stressors(Groves et al., 2013; Wang et al., 2014) and the increased
modification level has thought to be protective towards cellular stresses.
Multiple lines of evidence have shown that activation of HBP and subsequent
elevation of O-GlcNAc levels increases the survival rate of cancer cells that are
undergoing stress condition. Notably, the hyper-O-GlcNAcylation level has
been shown to protect cancer cell from hypoxia and glucose deprivation
through O-GlcNAcylation of PFK1, one of the key enzymes that regulates the
flow of glucose-derived carbon in glycolysis(Yi et al., 2012). O-GlcNAc
modification of PFK1 at S529 decreases its activity, thus redirecting
metabolites away from glycolysis into other pathways, such as pentose
phosphate pathway that generates NADPH to maintain a pool of reduced
glutathione (GSH) for combating oxidative stress as well as provides building
blocks for biomass generation. Jones and coworkers also showed that in
osteosarcoma U2OS cell line, metabolic flux through the HBP under glucose-
restricted conditions maintains PI3K/PKB signaling pathway, which
influences long term cell proliferation as well as survival(Jones et al., 2014).
Lastly, activation of the HBP through upregulation of GFAT is necessary for
pancreatic cancer cells to survive hypoxic conditions(Guillaumond et al.,
2013). All in all, in cancer cells, upregulation of O-GlcNAcylation partially

12

orchestrates the adaptive metabolic response in order to survive and
proliferate under stress conditions.
1.3.3 O-GlcNAc and cancer cell metastasis and invasion
One of the important hallmarks of cancer is the ability of cancer cells to
invade and metastasize to distant organs(Hanahan and Weinberg, 2011b),
which accounts for more than 90% of cancer mortality (Gupta and Massagué,
2006). Metastasis is a biological cascade that consists of multiple distinct
steps and the hyper-O-GlcNAcylation is implicated in one of the early
important steps: the loss of cellular adhesion and epithelial to mesenchymal
transition (EMT) (Gupta and Massagué, 2006; Issad, 2013). EMT involves
loss of E-cadherin, a major component of adherent junctions that is
responsible for cell-cell adhesion. Two independent research groups have
showed that reducing hyper-O-GlcNAcylation in cancers increases the
expression of E-cadherin, whereas enhancing hyper-O-GlcNAcylation through
inhibition of OGA has the opposite effect(Gu et al., 2010; Ma et al., 2013).
Furthermore, OGT shRNA dramatically reduced the metastasis potential of
4T1 cells in vivo, and down-regulation of E-cadherin in the OGT knock-down
cells restored 4T1 cells’ invasiveness(Gu et al., 2010). Therefore, hyper-O-
GlcNAcylation assists metastasis by downregulating E-cadherin protein level.
1.4 Reference
Aebi, M. (2013). Biochimica et Biophysica Acta. BBA - Mol. Cell
Research 1833, 2430–2437.

13

Anastasiou, D., Poulogiannis, G., Asara, J.M., Boxer, M.B., Jiang, J.K.,
Shen, M., Bellinger, G., Sasaki, A.T., Locasale, J.W., Auld, D.S., et al. (2011).
Inhibition of Pyruvate Kinase M2 by Reactive Oxygen Species Contributes to
Cellular Antioxidant Responses. Science 334, 1278–1283.
Anastasiou, D., Yu, Y., Israelsen, W.J., Jiang, J.-K., Boxer, M.B., Hong,
B.S., Tempel, W., Dimov, S., Shen, M., Jha, A., et al. (2012). Pyruvate kinase
M2 activators promote tetramer formation and suppress tumorigenesis. Nat.
Chem. Biol. 8, 839–847.
Bond, M.R., and Hanover, J.A. (2015). A little sugar goes a long way:
The cell biology of O-GlcNAc. J. of Cell Biol. 208, 869–880.
Bouché, C., Serdy, S., Kahn, C.R., and Goldfine, A.B. (2004). The
Cellular Fate of Glucose and Its Relevance in Type 2 Diabetes. Endocrine
Reviews 25, 807–830.
Caldwell, S.A., Jackson, S.R., Shahriari, K.S., Lynch, T.P., Sethi, G.,
Walker, S., Vosseller, K., and Reginato, M.J. (2010). Nutrient sensor O-
GlcNAc transferase regulates breast cancer tumorigenesis through targeting
of the oncogenic transcription factor FoxM1. Oncogene 29, 2831–2842.
Chaneton, B., Hillmann, P., Zheng, L., Martin, A.C.L., Maddocks,
O.D.K., Chokkathukalam, A., Coyle, J.E., Jankevics, A., Holding, F.P.,
Vousden, K.H., et al. (2012). Serine is a natural ligand and allosteric activator
of pyruvate kinase M2. Nature 1–7.
Denzel, M.S., and Antebi, A. (2015). ScienceDirectHexosamine
pathway and (ER) protein quality control. Curr. Opin. in Cell Bio. 33, 14–18.
Fantin, V.R., St-Pierre, J., and Leder, P. (2006). Attenuation of LDH-A
expression uncovers a link between glycolysis, mitochondrial physiology, and
tumor maintenance. Cancer Cell 9, 425–434.
Ferrer, C.M., Lynch, T.P., Sodi, V.L., Falcone, J.N., Schwab, L.P.,
Peacock, D.L., Vocadlo, D.J., Seagroves, T.N., and Reginato, M.J. (2014). O-
GlcNAcylation Regulates Cancer Metabolism and Survival Stress Signalingvia
Regulation of the HIF-1 Pathway. Mol. Cell 1–12.
Groves, J.A., Lee, A., Yildirir, G., and Zachara, N.E. (2013). Dynamic O-
GlcNAcylation and its roles in the cellular stress response and homeostasis.
Cell Stress and Chaperones 18, 535–558.
Gu, Y., Mi, W., Ge, Y., Liu, H., Fan, Q., Han, C., Yang, J., Han, F., Lu,
X., and Yu, W. (2010). GlcNAcylation Plays an Essential Role in Breast Cancer
Metastasis. Cancer Research 70, 6344–6351.

14

Guillaumond, F., Leca, J., Olivares, O., Lavaut, M.-N., Vidal, N.,
Berthezène, P., Dusetti, N.J., Loncle, C., Calvo, E., Turrini, O., et al. (2013).
Strengthened glycolysis under hypoxia supports tumor symbiosis and
hexosamine biosynthesis in pancreatic adenocarcinoma. Proc. Natl. Acad. Sci.
U.S.A. 110, 3919–3924.
Gupta, G.P., and Massagué, J. (2006). Cancer Metastasis: Building a
Framework. Cell 127, 679–695.
Hanahan, D., and Weinberg, R.A. (2011a). Hallmarks of Cancer: The
Next Generation. Cell 144, 646–674.
Hanahan, D., and Weinberg, R.A. (2011b). Hallmarks of Cancer: The
Next Generation. Cell 144, 646–674.
Hang, H.C., and Bertozzi, C.R. (2005). The chemistry and biology of
mucin-type O-linked glycosylation. Bio. & Medi. Chem. 13, 5021–5034.
Hardivillé, S., and Hart, G.W. (2014). Nutrient Regulation of Signaling,
Transcription, and Cell Physiology by O-GlcNAcylation. Cell Metabolism 20,
208–213.
Hinderlich, S., Berger, M., Schwarzkopf, M., Effertz, K., and Reutter,
W. (2000). Molecular cloning and characterization of murine and human N-
acetylglucosamine kinase. Eur. J. Biochem. 267, 3301–3308.
Hitosugi, T., Kang, S., Vander Heiden, M.G., Chung, T.W., Elf, S.,
Lythgoe, K., Dong, S., Lonial, S., Wang, X., Chen, G.Z., et al. (2009). Tyrosine
Phosphorylation Inhibits PKM2 to Promote the Warburg Effect and Tumor
Growth. Science Signaling 2, ra73–ra73.
Hsu, P.P., and Sabatini, D.M. (2008). Cancer Cell Metabolism:
Warburg and Beyond. Cell 134, 703–707.
Issad, T. (2013). O-GlcNAcylation: a new cancer hallmark? 1–14.
Jones, D.R., Keune, W.-J., Anderson, K.E., Stephens, L.R., Hawkins,
P.T., and Divecha, N. (2014). The hexosamine biosynthesis pathway and O-
GlcNAcylation maintain insulin-stimulated PI3K-PKB phosphorylation and
tumour cell growth after short-term glucose deprivation. Febs J n/a–n/a.
Jones, R.G., and Thompson, C.B. (2009). Tumor suppressors and cell
metabolism: a recipe for cancer growth. Genes & Development 23, 537–548.
Junttila, M.R., and Evan, G.I. (2009). p53 — a Jack of all trades but
master of none. Nat. Rev. Cancer 9, 821–829.

15

Kwei, K.A., Baker, J.B., and Pelham, R.J. (2012). Modulators of
Sensitivity and Resistance to Inhibition of PI3K Identified in a
Pharmacogenomic Screen of the NCI-60 Human Tumor Cell Line Collection.
PLoS ONE 7, e46518.
Lau, K.S., and Dennis, J.W. (2008). N-Glycans in cancer progression.
Glycobiology 18, 750–760.
Lewis, B.A., and Hanover, J.A. (2014). O-GlcNAc and the Epigenetic
Regulation of Gene Expression. J. Biol. Chem. 289, 34440–34448.
Lynch, T.P., Ferrer, C.M., Jackson, S.R., Shahriari, K.S., Vosseller, K.,
and Reginato, M.J. (2012). Critical Role of O-Linked -N-Acetylglucosamine
Transferase in Prostate Cancer Invasion, Angiogenesis, and Metastasis. J.
Biol. Chem. 287, 11070–11081.
Ma, J., and Hart, G.W. (2014). O-GlcNAc profiling: from proteins to
proteomes. Clin Proteomics 11, 8.
Ma, Z., Vocadlo, D.J., and Vosseller, K. (2013). Hyper-O-GlcNAcylation
Is Anti-apoptotic and Maintains Constitutive NF- B Activity in Pancreatic
Cancer Cells. J. Biol. Chem. 288, 15121–15130.
Ma, Z., and Vosseller, K. (2013). O-GlcNAc in cancer biology. Amino
Acids 45, 719–733.
Musgrove, E.A., Caldon, C.E., Barraclough, J., Stone, A., and
Sutherland, R.L. (2011). Cyclin D as a therapeutic target in cancer. Nat. Rev.
Cancer 11, 558–572.
Olivier-Van Stichelen, S., Dehennaut, V., Buzy, A., Zachayus, J.L.,
Guinez, C., Mir, A.M., Yazidi-Belkoura, El, I., Copin, M.C., Boureme, D.,
Loyaux, D., et al. (2014). O-GlcNAcylation stabilizes -catenin through direct
competition with phosphorylation at threonine 41. The FASEB Journal 28,
3325–3338.
Schleicher, E.D., and Weigert, C. (2000). Role of the hexosamine
biosynthetic pathway in diabetic nephropathy. Kidney International 58, S13–
S18.
Shafi, R., Lyer, S., Ellies, L.G., O'Donnell, N., Marek, K.W., Chui, D.,
Hart, G.W., and Marth, J.D. (2000). The O-GlcNAc transferase gene resides
on the X chromosome and is essential for embryonic stem cell viability and
mouse ontogeny. Proc. Natl. Acad. Sci. U.S.A. 97, 5735–5739.
Stichelen, S.O.-V., Drougat, L., Dehennaut, V., Yazidi-Belkoura, El, I.,
Guinez, C., Mir, A.M., Michalski, J.C., Vercoutter-Edouart, A.-S., and

16

Lefebvre, T. (2012). Serum-stimulated cell cycle entry promotes ncOGT
synthesis required for cyclin D expression. Oncogenesis 1, e36–e36.
Vander Heiden, M.G., Cantley, L.C., and Thompson, C.B. (2009).
Understanding the Warburg Effect: The Metabolic Requirements of Cell
Proliferation. Science 324, 1029–1033.
Wallace, D.C. (2012). Mitochondria and cancer. Nat. Rev. Cancer 12,
685–698.
Wang, I.-C., Chen, Y.J., Hughes, D., Petrovic, V., Major, M.L., Park,
H.J., Tan, Y., Ackerson, T., and Costa, R.H. (2005). Forkhead Box M1
Regulates the Transcriptional Network of Genes Essential for Mitotic
Progression and Genes Encoding the SCF (Skp2-Cks1) Ubiquitin Ligase. Mol.
and Cell. Biol. 25, 10875–10894.
Wang, P., Lazarus, B.D., and Forsythe, M.E. (2012). O-GlcNAc cycling
mutants modulate proteotoxicity in Caenorhabditis elegans models of human
neurodegenerative diseases.
Wang, Z.V., Deng, Y., Gao, N., Pedrozo, Z., Li, D.L., Morales, C.R.,
Criollo, A., Luo, X., Tan, W., Jiang, N., et al. (2014). Spliced X-Box Binding
Protein 1 Couples the Unfolded Protein Response to Hexosamine Biosynthetic
Pathway. Cell 156, 1179–1192.
Yi, W., Clark, P.M., Mason, D.E., Keenan, M.C., Hill, C., Goddard,
W.A., Peters, E.C., Driggers, E.M., and Hsieh-Wilson, L.C. (2012).
Phosphofructokinase 1 glycosylation regulates cell growth and metabolism.
Science 337, 975–980.
Zachara, N.E., and Hart, G.W. (2002). The Emerging Significance of O-
GlcNAc in Cellular Regulation. Chem. Rev. 102, 431–438.
Zhang, Q., Liu, X., Gao, W., Li, P., Hou, J., Li, J., and Wong, J. (2014).
Differential Regulation of the Ten-Eleven Translocation (TET) Family of
Dioxygenases by O-Linked β- N-Acetylglucosamine Transferase (OGT). J.
Biol. Chem. 289, 5986–5996.

17

Chapter 2. Approaches to perturb protein
functions
2.1 Introduction
One of the fundamental ways to investigate the role of a gene is to
increase or decrease its function and observe the response of the system. The
central dogma of biology (DNA makes RNA makes protein) thus offers three
windows for the perturbation to occur. Figure 2-1 lists some of the widely-
used tools for perturbing the function of a gene. Small molecules have a long-
standing history in helping scientists to study the function of proteins. Acting
on proteins directly, small molecules are fast, dose-dependent and reversible.
Some of them can be very specific, but they are not general. The most
common method to find an activator or an inhibitor for a protein of interest
(POI) is to screen it through a collection of compounds in search of a potent,
as well as selective, modulator. The collection of screened compounds usually
consists of natural products or synthetic molecules designed based on
structural scaffold of known substrates or inhibitors. Often times natural
products could be difficult to obtain in large quantities either from natural
resources or through synthetic efforts, and a set of compounds that have
similar core scaffolds would not be ideal in finding a specific inhibitor for a
POI that belongs to a large group of protein family, such as kinases, that has a
high degree of homology in their active site(Bishop et al., 1998). To elevate the
specificity as well as of the generality of the small molecule approach to the
next level, the Shokat group pioneered an alternative approach that combined

18

features of genetics and chemistry, giving rise to the so-called bump-hole
strategy(Shogren-Knaak, Alaimo, & Shokat, 2001). In general, they replaced a
large, hydrophobic and conserved amino acid in the ATP-binding site of a
kinase of interest with either a glycine or alanine, creating a “hole” into the
active site of the protein target. Next, they turned a nonspecific kinase
inhibitor of the kinase of interest into a specific inhibitor towards the mutant
kinase by adding a bulky side-chain substituent, the bump, which specifically
complement the mutation introduced into the active site. Notably, the
modified inhibitor has greatly reduced affinity towards the wild-type kinase
because of steric hindrance with the bulky active site residue. This strategy is
generalizable since it has not only been implemented into small molecule
inhibitors design towards kinases, but has also been used for guanosine
triphosphate (GTP)-binding proteins and seven-pass transmembrane
proteins(Bishop et al., 2000).

19

Figure 2-1. Comparison of techniques in perturbing protein function. Widely used
techniques are plotted as a function of how fast a phenotype can be observed after the
perturbation has occurred as well as how specific the perturbation is. The red hue of their
circumference corresponds to their generality. Blue hue of the dot represents their tunability.
Figure adapted from (Shogren-Knaak et al., 2001). RNAi: RNA interference. CRISPR:
clustered regularly interspaced short palindromic repeats.

On the other side of the spectrum, transcriptional and translational
approaches are specific and general, but their main drawbacks are they lack
good tunability and require a significant amount of time to implement.
Techniques that target at the DNA level (i.e., knock-in and knock-out) have
been useful for discovering genes that when mutated, give rise to identifiable
phenotypes. Those methods are robust and specific but have been difficult to
implement and poorly reversible. The development of genome editing tools in
the past decade or so has highly expediated the process of introducing desired
mutations and insertions as well as performing knock-out of any target genes.
Specifically, type II prokaryotic clustered regularly interspaced short
palindromic repeats (CRISPR)/ Cas9 system has been successfully applied in
editing genome across a wide variety of cell types and whole
Specificity
Timescale
Small molecules
Tet-On/Off
RNAi
Temperature
sensitive
Conditional
Knock-out/in
Mouse
Knock-out
CRISPR

20

organisms(Sander & Joung, 2014). Approaches that target at the RNA level
(i.e., RNAi and riboswitches) are faster and easier to implement compared to
the traditional DNA manipulation techniques; however, their drawbacks are
the efficiency of knockdown and the possibility of off-target effects. In this
chapter, I discuss about the advantages and limitations of some representative
methods for perturbing function of a gene at each level.
2.2 Approaches at protein level
Small-molecule modulators of protein function can overcome some of
the drawbacks of the genetic and post-transcriptional methods. They can be
used in a rapid, dose-dependent and reversible manner to influence a variety
of protein functions, including activity(Bishop et al., 1998), localization(Geda
et al., 2008) and protein-protein interactions(Pruschy et al., 1994). Many
synthetic efforts are dedicated to developing molecules that target every
protein of interest, and we may see one day that the scientific community has
access to a huge library of compounds that enable specific and proteome wide
studies. Until this one molecule-one protein goal is realized, scientists ponder
on one question: how can we increase the generality of using small molecules
to target proteins?
One of the most successful strategies in this field is the utilization of
small molecules as regulators of protein stability. A pioneering work in this
area was demonstrated by Varshavsky and coworkers in 1995(Johnston,
Johnson, Waller, & Varshavsky, 1995). They engineered a dihydrofolate

21

reductase (DHFR) protein, the stability of which is governed by the presence
of a high-affinity ligand of DHFR: methotrexate (MTX). In short, Ubiquitin is
fused to the N-terminus of the mutant DHFR, which has an arginine (Arg) at
its N-terminus. After ubiquitin is cleaved away from the fusion protein, the
destabilizing Arg is revealed resulting in the degradation of fusion protein by
the proteasome, governed by the N-end rule that relates the in vivo half-life of
a protein to the identity of its N-terminal residue(Varshavsky, 1992). Addition
of the MTX inhibited its degradation through stabilization of folded
conformation of DHFR. Later on, Crabtree and co-workers expanded this
notion into the development of destabilization domains (DDs) to confer the
degradation of their POI(Stankunas, Bayle, Gestwicki, & Lin, 2003). In their
work, they combined the high specificity and versatility of genetics with the
rapid and reversible nature of small molecule-mediated effects. Genetic fusion
of a POI to a triple mutant FRB (FKBP12-rapamycin-binding protein, which is
a domain of mTOR kinase), termed FRB*, resulted in degradation of the
fusion protein. However, FRB* can be stoichiometrically stabilized by the
addition of a rapamycin analogue, MaRap, after forming a ternary complex
with the endogenous protein FKBP. This method allows them to control
glycogen synthase kinase-3 protein abundance and activity in genetically
modified mice. The Wandless group at Stanford University further improved
upon this result through the discovery of an mutant FKBP-based
DD(Banaszynski, Chen, Maynard-Smith, Ooi, & Wandless, 2006). Unlike the
FRB*-based system, FKBP-based DD does not require formation of a ternary

22

complex, which potentially minimizes the effects on normal protein function.
The FKBP-based DD was able to control the stability of a variety of proteins as
well as in a variety of in vitro cell lines and organisms in a reversible and
tunable manner by using an engineered small molecule Shld1(Banaszynski,
Sellmyer, Contag, Wandless, & Thorne, 2008; Iwamoto, Björklund, Lundberg,
Kirik, & Wandless, 2010). Despite the successful application of the FKBP-
based DD, it requires the POI to be fused to the destabilizing domain, which
may interfere with its function. To circumvent this issue, Pratt and Muir and a
coworker introduced the SURF (split ubiquitin for the rescue of function)
system, where split ubiquitin technology was implemented into the FRB-
FKBP dimerization pair(Pratt, Schwartz, & Muir, 2007). Addition of
rapamycin results in dimerization between FRB and FKBP as well as
complementation of the ubiquitin moiety, followed by ubiquitin hydrolysis to
release the intact POI, including caspase-3, kinase v-Src and transcriptional
factor Smad3.
The above-mentioned technologies utilized intracellular protein quality
control system, where destabilized proteins are targeted for ubiquitination
and degradation by the proteasome. Instead of manipulating the stability of
the DD-POI fusion, researchers have showed that recruitment of a ubiquitin
E3 ligase to a POI is another efficient way to achieve ubiquitination and
subsequent degradation of POI. Notably, Sakamoto and Crews and coworkers
engineered chimeric molecules, termed bifunctional proteolysis targeting
chimeras (protacs), that target a protein to the Skp1-Cullin-F box complex for

23

ubiquitination and degradation in vitro(Sakamoto et al., 2001). Later on,
Schneekloth and Crews and coworkers developed other protac molecules that
allow them to conditionally control protein function in vivo.
After the attachment of polyubquitin chains to the POI, the protein of
interest is target to the proteasome for degradation. Therefore, methods have
been developed to recruit the proteasome to the POI to achieve its
degradation. Janse and coworkers utilized Fpr1-rapamycin-Tor1 dimerization
pair from S. cerevisiae to show that proteasome localization is sufficient for
degradation of their POI(Janse, Crosas, Finley, & Church, 2004). They fused
26 proteasome subunit Rpn10 to Fpr1 and His3, the POI, to Tor1. They
observed that upon addition of rapamycin to cells expressing Fpr1-Rpn10
fusion, the Tor1-His3 fusion protein localized to the proteasome and was
rapidly degraded. Elegant work from Renicke and colleagues from Germany
illustrates another technique in this class. They fused the light-perceiving
domain of phototropin, light oxygen voltage (LOV2) domain, to a fragment of
ornithine decarboxylase (ODC)(Renicke, Schuster, Usherenko, Essen, &
Taxis, 2013). When attached to the C-terminus of a protein, active ODC
degron induces ubiquitin-independent proteasomal degradation. The
resulting generic photosensitive degron (gpd) was highly versatile, evidenced
by their ability to control the stability of soluble as well as membrane bound
proteins and to manipulate yeast cell cycle and growth. Utilizing the same
LOV2 domain, Bonger and Wandless and coworkers engineered a blue-light
inducible degradation (B-LID) domain, achieving conditional protein control

24

in mammalian cells and in zebrafish embryos using nontoxic blue
light(Bonger, Rakhit, Payumo, Chen, & Wandless, 2014).
2.3 Approaches at DNA level
2.3.1 Knock-out and Knock-in
In 2007, the Nobel Prize in Physiology or Medicine was awarded to
Martin Evans, Mario Capecchi and Oliver Smithies for their discoveries of
“principles for introducing specific gene modification in mice by the use of
embryonic stem cells”, which served as the basis for Knock-out (inactivation
of a certain gene) and Knock-in (introduction of a desired mutation)
technologies. Although those two concepts can also be applied to a variety of
species as well as in-vitro cell lines, historically they refer to the generation of
genetically engineered mice from embryonic stem cells, where the genetic
modification was introduced. In short, stem cells are isolated from mouse
blastocyst, followed by introduction of the targeting vector through
electroporation. Homologous recombination then occurs in between the gene
of interest (GOI) and the vector. In the case of knock-out approach, a positive
selection marker flanked by loxP sequence(Kos, 2004) replaces one of the
exons of the targeted gene. In the case of knock-in approach, a cassette that
contains the positive selection marker, loxP sequence and a mutated exon
replaces the original exon of the targeted gene. The electroporated-cells then
undergo drug selection to enrich the cells that had gone through successful
electroporation as well as the recombination event. Expression of Cre-
recombinase(Kos, 2004) in the selected population allows the removal of the

25

positive selection marker in either case, leaving no traces of the manipulation
behind in the genome of the stem cells. As a result, the genes that undergo
knock-out step loses one exon, whereas the genes that undergo the knock-in
step now possess the desired mutation. The genetically altered stem cells are
then injected into a blastocyst, followed by implantation into the uterus of a
foster mother. Some of the mice that the foster mother gives birth to are
heterozygous, possessing one wild-allele and one mutated-allele of the GOI.
Subsequent breeding of the heterozygous mice would give rise to homozygous
mice, possessing two mutated alleles. Knock-out and Knock-in mice are
powerful tools to study function of certain genes; however, some experimental
limitations exist. Firstly, inactivation of a gene that is indispensable for
development sometimes leads to embryonic lethality, which prevents the
observation of a less drastic phenotype in adult mice. Secondly, the knock-out
or knock-in mice might exhibit cellular or molecular compensation in respond
to the mutated GOI, which would complicate the interpretation of the results
that obtained from these type of experiments.
2.3.2 Conditional control of gene function
In order to circumvent embryonic lethality when perturbing essential
genes at DNA level, scientists have developed methods for the conditional
activation or deactivation genes. One of the most successful methods in this
category is the utilization of tetracycline responsive system, which was
pioneered by Bujard and Gossen in 1992 (Gossen and Bujard, 1992). In E.
Coli., tetracycline resistance gene is constitutively repressed by the

26

tetracycline repressor (tetR), a protein that binds specifically to tetracycline
operator (tetO) sequences within the promoter, resulting in gene silencing.
Upon its addition, tetracycline binds to the tetR and that results in the
dissociation of tetR from the tetO operator, turning on the transcription of the
gene. To apply the system into mammalian cells, Bujard and Gossen fused the
tetR with the activation domains of the herpes simplex virus VP16 protein.
VP16 is a virally encoded transcription factor that recruits cellular
transcription machineries to activate viral transcription that are necessary for
virus replication. Addition of the VP16 domain transforms the TetR from a
transcriptional repressor to a transcriptional activator, and the resulting
hybrid protein is known as the tetracycline-controlled transactivator (tTA).
The second modification that Bujard and Gossen did to the bacterial
expression system was the use of a cytomegalovirus (CMV)- derived minimal
promoter, fused with tetO sequences to control transgene expression.
Therefore, in the absence of tetracycline, tetR-VP12 fusion binds to tetO
sequence, turning on the transcription of downstream GOI. In the presence of
tetracycline, tetracycline-tetR-VP12 complex dissociates from the fused
promoter, silencing GOI. Later on Gossen and co-workers engineered a
reverse tTA (rtTA) system, where the transcription activator binds to tetO in
the presence of doxycycline, an analogue of tetracycline, turning on
expression of GOI (Gossen et al., 1995). The tetracycline/ doxycycline
responsive systems have been applied to in vitro cell lines as well as the

27

generation of transgenic mice where expression of GOI can be turned on
temporally as well as spatially(Ryding, Sharp, & Mullins, 2001).
In contrast to conditionally turning on expression of certain genes, one
can think of a system where a gene is deactivated at a given time or/ and at a
chosen tissue. The most popular way of achieving conditional gene deletion is
the utilization of Cre/LoxP system. The Cre recombinase is a type of
recombination enzyme found in bacteriophage P1. It recognizes a specific 34
base pair nucleotide sequence called a loxP site, which consist of two 13-bp
inverse repeats flanking an 8-bp core sequence that governs the directionality
of the loxP site. When the two loxP sites are in the same orientation, Cre
excises the intervening DNA segment; in the case of opposite orientation, Cre
inverts it. Either one of the cases is sufficient to inactivate the GOI. However,
since the latter approach runs into the possibility of repetitive recombination
events, the former strategy is used in knockout constructs. Ligand-activated
Cre allows time-specific knock out of GOI (Furth et al., 1994; Wunderlich,
Wildner, Rajewsky, & Edenhofer, 2001) and the utilization of tissue-specific
promoters, as well as tetracycline-regulated expression, of Cre enables enables
tissue and spatially localized gene(Saam & Gordon, 1999).
2.3.3 Genome editing tools
Genome editing tools allow researchers to introduce targeted genomic
sequence changes in living cells and organisms to achieve a variety of
applications, including frameshift knockout mutations, sequence insertions

28

for fusing epitope tags as well as specific sequence alterations. In order to
perform targeted genome editing, one has to create a DNA double-stranded
break (DSB) at the genomic locus to be modified. After the introduction of
DSB by nucleases, 2 different repair mechanisms naturally occurs inside the
cells would take place: nonhomologous end-joining (NHEJ) and homology-
directed repair (HDR). NHEJ results in re-ligation of the DNA with
introduction of random insertion/ deletion (indel) mutations. An indel within
an exon of a gene can lead to frameshift in the translational reading frame and
premature stop codons. HDR pathway is more prominent in dividing cells and
occurs less frequently compared to NHEJ, but it can be utilized to introduce
desired mutations in the presence of an exogenously introduced DNA, either a
double-stranded DNA construct or a single-stranded oligonucleotides. The
resulting mutation can then be verified by conventional sequencing. Earlier
methods in this class utilized nucleases that recognized specific DNA
sequences through protein-DNA interactions, such as zinc finger nucleases
(ZFNs)(Porteus, 2003) and transcription activator-like effector nucleases
(TALENs)(Wood et al., 2011). The ZFN is composed of two domains: the
DNA-binding domain of eukaryotic transcription factors zinc finger proteins
and the nuclease domain of the FokI restriction enzyme, both of which are
important to achieve the genome editing purposes. The zinc finger proteins
offer DNA binding specificity as well as flexibility for user to define the
targeted-sequence, whereas FokI catalytic domain provides a robust cleavage
activity towards the target site. Since FokI nuclease functions as a

29

homodimer, a pair of ZFN needs to be deployed at the target site in order to
result in DSBs. Typically, each ZFN contains three fingers to bind a 9-bp DNA
target, so a functional ZFN, which comprises of a dimer of ZNF, offers 18-bp
specificity towards the intended cleavage site. However, during the
construction of engineered zinc finger arrays, researchers need to consider
context-dependent effects between individual finger domains in an
array(Wolfe, Nekludova, & Pabo, 2000), preventing it to become a
mainstream genome editing tool. Like ZFNs, TALENs utilize the FokI catalytic
domain to confer its DNA cleavage activity. But it uses transcription activator-
like effector (TALE) from the largest family of type III effector proteins found
in a group of Gram-negative bacterial plant pathogens, Xanthomonas spp., to
achieve its DNA recognition functionality. A typical TALEN has the following
structure: a single DNA-binding domain (termed repeat 0), 18 repeat variable
di-residues (RVDs), followed by a 12 base spacer and the FokI nuclease
domain. The repeat 0 and each of the RVD recognizes one nucleotide of the
target DNA. Therefore, each TALEN recognizes a target that is 19 bp long. In
order for the FokI nuclease to perform its function, each TALEN targeted
DNA sequence are chosen such that a pair of TALEN are arranged in an
opposing orientation on opposite sides of double-stranded DNA, so that a pair
of FokI nucleases can function as a homodimer to achieve DSBs. The ZFN and
TALEN have allowed researchers to perform genome editing in a variety of
species and mammalian cell lines for numerous applications(Gaj, Gersbach, &
Barbas, 2013; Urnov, Rebar, Holmes, Zhang, & Gregory, 2010). Despite the

30

successful application of these two nucleases in a variety of different cell types
and organisms, the complexity in the design of customizable DNA-binding
specificities in ZFNs(Wolfe et al., 2000) and the repetitive nature of TALEN-
coding sequences(Jorgensen et al., 2009) have hampered their wide adoption
in the scientific community.
Unlike ZFNs and TALENs, RNA-guided Cas9 nuclease from type II
prokaryotic CRISPR adaptive immune system relies on simple Watson-Crick
base-paring rules for the recognition between the nuclease and the targeted
DNA sequence(Cong et al., 2013). CRISPR-Cas is a microbial adaptive
immune system that uses RNA-guided nucleases to destroy foreign invading
DNAs. So far, three types (I-III) of CRISPR systems have been identified
across a wide range of bacteria, whereas type II CRISPR system is the one that
is best characterized. The CRISPR immune system incorporates foreign
invading DNA, known as protospacers, in between a distinctive array of
repetitive elements sequences, together they constitute of CRISPR RNA
(crRNA) array. Within the array, each protospacer is always associated with a
protospacer adjacent motif (PAM), which can vary between different CRISPR
systems(Brouns et al., 2008). The transcript of the crRNA array hybridizes
with a second RNA, known as transactivating CRISPR RNA (tracrRNA). The
hybridized RNAs then forms complex with Cas9 nuclease. Each protospacer
has a 20 nucleotides sequence that directs Cas9 to a 20 bp complementary
DNA sequence. When the target DNA is next to the same PAM sequence that
is associated with the CRISPR system, the Cas9 will cleave the target DNA and

31

result in a double-stranded break. For the type II CRISPR system, the target
DNA sequence must lie immediately 5’ of a PAM sequence that fulfil the NGG
requirement.
The type II CRISPR system has been adapted to induce sequence-
specific double-stranded breaks and subsequent genome editing. The most
widely used application of this technology was developed by Zhang lab at the
Broad institute. They fused the crRNA and tracrRNA together to create a
chimeric, single-guide RNA (sgRNA) and then they cloned the sgRNA
construct into an expression plasmid along with a human codon-optimized
Cas9. They demonstrated that simply by changing the first 20 nucleotides of
the sgRNA, the protospacer region, Cas9 can be directed to almost any DNA
sequence as long as it is next to the NGG PAM(Ran et al., 2013), thus highly
improved the efficiency as well as the versatility of the genome editing tools
previously described. Cas9 has not only been shown to perform genome
editing in a variety cell and organisms, it has also been applied to activate
genes in cells(Perez-Pinera et al., 2013), alter histone modification(Hilton et
al., 2015) as well as imaging of specific genomic loci(B. Chen et al., 2013).
Despite all the above successful applications of CRISPR-Cas9 system, one
have to be aware of its off target effects since mismatches in the guide-RNA
sequence are tolerated depending on their position and hence could result in
off-target modification(Fu et al., 2013).

32

2.4 Approaches at RNA level
In eukaryotes, regulation of gene function at RNA level is achieved by
small (around 20-30 nucleotide) noncoding RNAs and their related effector
proteins. These small RNAs generally have inhibitory roles in gene expression
and control; therefore, their mechanisms of action usually fall under the
classification of RNAi (RNA interference) or RNA silencing. Two main
categories of these small RNAs are short interfering RNA (siRNA) and
microRNAs (miRNA). In general, miRNAs assume roles in regulation of
endogenous genes and they are produced from an organism’s own genome,
whereas siRNAs serve as defence system of genome integrity and are usually
derived from invading viruses, transposons or transgenes. Even though siRNA
and miRNA take on different roles in the cells, their mechanism of actions
shares the same backbone. In short, double-stranded RNA (dsRNA)
precursors derived either from host genome or from invasive nucleic acids are
excised by a protein called Dicer into ~20-30 nucleotide fragments. The
processed dsRNA is then loaded onto a RISC (RNA-induced silencing
complex), where it is unwound to result in the guide strand to associate with a
protein within the RISC called Argonaute (Ago). The single-stranded RNA-
RISC complex then recognize targets through Watson-Crick base paring to
initiate numerous processes, including RNA and DNA degradation as well as
inhibition of transcription and translation(Carthew & Sontheimer, 2009).
One of the tasks that the RNA-RISC complex does is to promote mRNA
degradation to achieve gene silencing. The guide strand of the siRNA

33

recognizes its target mRNA through base paring. In the case of perfect
complementary, Ago cleaves the target mRNA, resulting in its degradation
and no productions of the target protein.
A knock-down strategy involves in the usage of siRNAs or short hairpin
RNAs (shRNAs) to achieve gene silencing through degradation of its
transcript, the mRNA of the GOI. This method harnesses the endogenous
Dicer as well as the RISC, but it exogenously provides the cells with the
double-stranded RNA that is needed for recognizing the target mRNA. In the
case of siRNA approach, double stranded RNA is introduced into the cytosol
by using transfection reagents. The introduced RNA is then processed by
Dicer to enter the siRNA silencing pathway. However, high concentration of
cytoplasmic siRNA might result in off target effects. And the transfected cells
lose the phenotype gradually as they proliferate since the siRNAs are only
transiently present in the cells. In contrast, shRNA approach allows the
generation of stable knock-down cell line, allowing the researchers to study
the phenotype with good reproducibility and consistency. shRNA consist of
two complementary 19-22 base pairs RNA sequences linked by a short loop of
4-11 nucleotides similar to the hairpin found in the natural miRNA processes.
Viral vectors such as adeno-associated viruses or lentiviruses are usually used
to delivery shRNA plasmid into the cells’ genome. Since the shRNA strategy
involves in the introduction foreign DNA into the genome of the organism of
interest, this method can also be perceived as an approach at DNA level.

34

2.5 Reference
Banaszynski, L. A., Chen, L.-C., Maynard-Smith, L. A., Ooi, A. G. L., &
Wandless, T. J. (2006). A Rapid, Reversible, and Tunable Method to Regulate
Protein Function in Living Cells Using Synthetic Small Molecules. Cell,
126(5), 995–1004.
Banaszynski, L. A., Sellmyer, M. A., Contag, C. H., Wandless, T. J., & Thorne,
S. H. (2008). Chemical control of protein stability and function in living mice.
Nat. Medi. 14(10), 1123–1127.
Bishop, A. C., Shah, K., Liu, Y., Witucki, L., Kung, C., & Shokat, K. M. (1998).
Design of allele-specific inhibitors to probe protein kinase signaling. Curr.
Biol. : CB, 8(5), 257–266.
Bishop, A., Buzko, O., Heyeck-Dumas, S., Jung, I., Kraybill, B., Liu, Y., et al.
(2000). Unnatural ligands for engineered proteins: new tools for chemical
genetics. Biophysics and Biomolecular Structure, 29, 577–606.
Bonger, K. M., Rakhit, R., Payumo, A. Y., Chen, J. K., & Wandless, T. J.
(2014). General Method for Regulating Protein Stability with Light. ACS
Chemical Biology, 9(1), 111–115.
Brouns, S. J. J., Jore, M. M., Lundgren, M., Westra, E. R., Slijkhuis, R. J. H.,
Snijders, A. P. L., et al. (2008). Small CRISPR RNAs guide antiviral defense in
prokaryotes. Science, 321(5891), 960–964.
Carthew, R. W., & Sontheimer, E. J. (2009). Origins and Mechanisms of
miRNAs and siRNAs. Cell, 136(4), 642–655.
Chen, B., Gilbert, L. A., Cimini, B. A., Schnitzbauer, J., Zhang, W., Li, G.-W.,
et al. (2013). Dynamic Imaging of Genomic Lociin Living Human Cells by an
Optimized CRISPR/Cas System. Cell, 155(7), 1479–1491.
Cong, L., Ran, F. A., Cox, D., Lin, S., Barretto, R., Habib, N., et al. (2013).
Multiplex Genome Engineering Using CRISPR/Cas Systems. Science,
339(6121), 819–823.
Fu, Y., Foden, J. A., Khayter, C., Maeder, M. L., Reyon, D., Joung, J. K., &
Sander, J. D. (2013). High-frequency off-target mutagenesis induced by
CRISPR-Cas nucleases in human cells. Nat. Biotech., 31(9), 822–826.
Furth, P. A., St Onge, L., Böger, H., Gruss, P., Gossen, M., Kistner, A., et al.
(1994). Temporal control of gene expression in transgenic mice by a
tetracycline-responsive promoter. Proc. Natl. Acad. Sci., 91(20), 9302–9306.

35

Gaj, T., Gersbach, C. A., & Barbas, C. F., III. (2013). ZFN, TALEN, and
CRISPR/Cas-based methods for genome engineering. Trends in
Biotechnology, 31(7), 397–405.
Geda, P., Patury, S., Ma, J., Bharucha, N., Dobry, C. J., Lawson, S. K., et al.
(2008). A small molecule-directed approach to control protein localization
and function. Yeast, 25(8), 577–594.
Hilton, I. B., D'Ippolito, A. M., Vockley, C. M., Thakore, P. I., Crawford, G. E.,
Reddy, T. E., & Gersbach, C. A. (2015). Epigenome editing by a CRISPR-Cas9-
based acetyltransferase activates genes from promoters and enhancers. Nat.
Biotech., 33(5), 510–517.
Iwamoto, M., Björklund, T., Lundberg, C., Kirik, D., & Wandless, T. J. (2010).
A General Chemical Method to Regulate Protein Stability in the Mammalian
Central Nervous System. Chem. Biol., 17(9), 981–988.
Janse, D. M., Crosas, B., Finley, D., & Church, G. M. (2004). Localization to
the Proteasome Is Sufficient for Degradation. J. Biol. Chem., 279(20), 21415–
21420.
Johnston, J. A., Johnson, E. S., Waller, P. R., & Varshavsky, A. (1995).
Methotrexate inhibits proteolysis of dihydrofolate reductase by the N-end rule
pathway. The J. Biol. Chem., 270(14), 8172–8178.
Jorgensen, C., Sherman, A., Chen, G. I., Pasculescu, A., Poliakov, A., Hsiung,
M., et al. (2009). Cell-Specific Information Processing in Segregating
Populations of Eph Receptor Ephrin-Expressing Cells. Science, 326(5959),
1502–1509.
Kos, C. H. (2004). Cre/loxP system for generating tissue-specific knockout
mouse models. Nutrition Reviews, 62(6), 243–246.
Perez-Pinera, P., Kocak, D. D., Vockley, C. M., Adler, A. F., Kabadi, A. M.,
Polstein, L. R., et al. (2013). RNA-guided gene activation by CRISPR-Cas9–
based transcription factors. Nature Methods, 10(10), 973–976.
Porteus, M. H. (2003). Chimeric Nucleases Stimulate Gene Targeting in
Human Cells. Science, 300(5620), 763–763.
http://doi.org/10.1126/science.1078395
Pratt, M. R., Schwartz, E. C., & Muir, T. W. (2007). Small-molecule-mediated
rescue of protein function by an inducible proteolytic shunt. Proc. Natl. Acad.
Sci. of U.S.A., 104(27), 11209–11214.
Pruschy, M. N. M., Spencer, D. M. D., Kapoor, T. M. T., Miyake, H. H.,
Crabtree, G. R. G., & Schreiber, S. L. S. (1994). Mechanistic studies of a

36

signaling pathway activated by the organic dimerizer FK1012. Chem. Biol.,
1(3), 163–172.
Ran, F. A., Hsu, P. D., Wright, J., Agarwala, V., Scott, D. A., & Zhang, F.
(2013). Genome engineering using the CRISPR-Cas9 system. Nature
Protocols, 8(11), 2281–2308.
Renicke, C., Schuster, D., Usherenko, S., Essen, L.-O., & Taxis, C. (2013). A
LOV2 Domain-Based Optogenetic Tool to Control Protein Degradationand
Cellular Function. Chem. Biol., 20(4), 619–626.
Ryding, A. D., Sharp, M. G., & Mullins, J. J. (2001). Conditional transgenic
technologies. Journal of Endocrinology.
Saam, J. R., & Gordon, J. I. (1999). Inducible gene knockouts in the small
intestinal and colonic epithelium. J. Biol. Chem., 274(53), 38071–38082.
Sakamoto, K. M., Kim, K. B., Kumagai, A., Mercurio, F., Crews, C. M., &
Deshaies, R. J. (2001). Protacs: chimeric molecules that target proteins to the
Skp1-Cullin-F box complex for ubiquitination and degradation. Proc. Natl.
Acad. Sci. of U.S.A., 98(15), 8554–8559.
Sander, J. D., & Joung, J. K. (2014). CriSPr-Cas systems for editing,
regulating and targeting genomes. Nat. Biotech., 32(4), 347–355.
Shogren-Knaak, M. A., Alaimo, P. J., & Shokat, K. M. (2001). Recent advances
in chemical approaches to the study of biological systems. Annual Review of
Cell and Developmental Biology, 17, 405–433.
Stankunas, K., Bayle, J. H., Gestwicki, J. E., & Lin, Y. M. (2003). Conditional
protein alleles using knockin mice and a chemical inducer of dimerization.
Mol. Cell.
Urnov, F. D., Rebar, E. J., Holmes, M. C., Zhang, H. S., & Gregory, P. D.
(2010). Genome editing with engineeredzinc finger nucleases. Nature
Publishing Group, 11(9), 636–646.
Varshavsky, A. (1992). The N-End Rule. Cell, 69(5), 725–735.
Wolfe, S. A., Nekludova, L., & Pabo, C. O. (2000). DNA recognition by
Cys2His2 zinc finger proteins. Biophysics and Biomolecular Structure, 29,
183–212.
Wood, A. J., Lo, T. W., Zeitler, B., Pickle, C. S., Ralston, E. J., Lee, A. H., et al.
(2011). Targeted Genome Editing Across Species Using ZFNs and TALENs.
Science, 333(6040), 307–307.

37

Wunderlich, F. T., Wildner, H., Rajewsky, K., & Edenhofer, F. (2001). New
variants of inducible Cre recombinase: a novel mutant of Cre-PR fusion
protein exhibits enhanced sensitivity and an expanded range of inducibility.
Nucleic Acids Research, 29(10), E47.

38

Chapter 3. A Dual Small-Molecule Rheostat for
Precise Control of Protein Concentration in
Mammalian Cells
3.1 Introduction
Careful modulation of protein concentration is crucial in a variety of
biological processes. For example, concentration of hnRNPA1 protein plays an
critical role in controlling the isoform-selective expression of pyruvate kinase
M1 or M2 through mutually exclusive mRNA splicing, (Chen et al., 2012) and
the abundance of tumor suppressor p53 is elevated under stressed condition,
enabling the cells to initiate various cellular responses. (Bargonetti and
Manfredi, 2002) Therefore, tools that can precisely control protein
concentration in living cells are valuable for investigating the contribution of
cellular protein levels to their function. (Banaszynski and Wandless, 2006)
One of the most successful strategies towards this goal relies on protein
destabilization domains (DD). Under normal conditions, a DD will be rapidly
degraded by the proteasome. However, the same DD can be stabilized or
“shielded” in a stoichiometric complex with a small-molecule, enabling dose-
dependent control of its concentration. This process has been exploited by
several labs to posttranslationally control the expression levels of proteins in
vitro as well as in vivo, exemplified by work by the Wandless lab, which used
a synthetic ligand Shld1 to control the stability of a proteins genetically fused
to an FKBP-based DD (FKBP*, FKBP F36V, L106P). (Banaszynski et al.,
2006; 2008; Chu et al., 2008; Iwamoto et al., 2010; Stankunas et al., 2007)
Although these technologies are powerful and generalizable, they result in a

39

permanent fusion of the protein of interest to the DD, which can affect its
biological activity and complicate any results.
To address this issue, we previously reported a complementary
strategy, termed traceless shielding (TShld), in which the protein of interest is
released in its native form. (Lau et al., 2010) The TShld system contained two
constructs (Figure 3-1A). The first construct consisted of FRB, a small domain
of the mTOR protein, fused to a mutant (I13G) of the N-terminal fragment of
ubiquitin (UbN, residues 1-37). The second construct had the following
architecture: the DD (FKBP*) followed by the C-terminal fragment of
ubiquitin (UbC, residues 35-76) and a protein of interest [e.g., green
fluorescent protein (GFP) as a model]. To enable the rapid generation of
stably transformed cell line, we encoded the two constructs into a single
plasmid by using a viral “self-cleaving” 2A peptide that induces a
cotranslational cleavage between the two proteins. (Holst et al., 2006;
Szymczak et al., 2004). Addition of rapamycin stabilized the FKBP*-UbC-GFP
construct, which would otherwise be degraded by the proteasome, and
subsequently recruited FRB-UbN construct, facilitating the complementation
between UbN and UbC and dose-dependent release of the protein of interest
from the DD by endogenous ubiquitin hydrolases (Figure 3-2B). While this
system was able to successfully control the concentration of different proteins
in their native form, we only observed a modest dynamic range of 5- to 7-fold,
due to high background levels of non-degraded protein.

40

Here, we describe an optimized protein concentration control system
that retains the traceless features of TShld but utilizes two tiers of small-
molecule control to set protein concentration in living cells. Specifically, we
placed level of TShld transcription under the control of doxycycline using a
tetracycline-responsive promoter (Tet-On) to create TTShld. Using three
concentrations of doxycyline in combination with a range of rapamycin
concentrations resulted in a 130-fold range of GFP levels as a model protein.
Notably, in this experiment, we observed that induced dimerization of UbN
and UbC using rapamycin was not required for ubiquitin complementation
and release of GFP. Therefore, we next demonstrated that rapamycin could be
replaced with either FK506 or Shld1 to achieve a 87- and 56-fold range of GFP
concentrations, respectively. All three small-molecules display good linearity
with respect to the amounts of shielded protein across the entire range of
concentrations, and the system is generalizable to other proteins of interest.
These experiments provide the first protein-concentration control system that
results in both a wide-range of protein concentrations and proteins free from
engineered fusion constructs.

41

Figure 3-1. TShld system. A) Architecture of TShld system. B) Design of TShld system.
Under normal conditions, the FKBP* destabilizing domain results in proteasomal
degradation of the fusion protein containing the protein of interest (POI). Upon addition of
rapamycin, FKBP* is stablized and FRB-UbN is recruited, causing complementation with
UbC and release of the protein of interest.
3.2 Results
3.2.1 Generation and characterization of TTshld-GFP cell lines
To improve the dynamic range of the TShld system, we added a second
tier of small-molecule control: the Retro-X Tet-On Advanced Inducible
Expression system (Clontech), in which the transcription level of gene of
interest can be controlled by adjusting the concentration of the system’s
inducer, doxycycline. For the ease of quantifying protein concentration, the
green fluorescence protein (GFP) was selected as our model protein of
interest. We cloned TShld-GFP construct into Retro-X Tet-On Advanced
Vector, termed Tet-On TShld-GFP (TTShld GFP). Commercially available Tet-
On HEK and HeLa cells were then stably transfected with TTShld-GFP using
amphotropic retroviral infection. To confirm shielding and release of GFP, the
HEK cells were treated with doxycycline (1000 ng/ mL) and rapamycin (500
A
B

42

nM) or water and DMSO vehicles for 48 hours. Western blotting showed that
a significant amount of GFP was released from the system only in the treated
cells (Figure 3-2A). Notably, we did not observe any uncleaved T2A protein or
full-length FKBP*-UbC-GFP. Next, we ask if the amount of GFP released from
the system will be dose-dependent on both the doxycycline and rapamycin.
Towards this end, stably transfected HEK and HeLa cells were treated with a
combination of three different concentrations of doxycycline (0, 50, 1000
ng/mL) and three different concentrations of rapamycin (0, 50, 500 nM).
Analysis by western blotting showed that the amount of GFP released from
the system was dose dependent upon the two small molecules (Figure 3-2B
and 3-2C). Again, no uncleaved T2A protein was detected.

43

Figure 3-2. Analysis of TTShld-GFP in stably transformed cells. A) TTShld-GFP
HEK cells were treated with doxycycline (1000 ng/ mL) and rapamycin (500 nM) or water
and DMSO vehicles for 48 hours and analyzed by Western blotting (FRB-UbN: anti-FLAG,
FKBP*-UbC-GFP and GFP: anti-HA) B) TTShld-GFP HEK cells were treated with the
indicated concentrations of doxycycline and rapamycin for 48 hours before analyzed by
Western blotting. C) TTShld-GFP HeLa cells were treated with the indicated concentrations
of doxycycline and rapamycin for 48 hours before Western blotting.
TTshld-GFP HeK
Cleaved T2A
FRB-UbN
FKBP*-UbC-GFP
GFP
Actin
Dox 1000 ng/mL, Rap 500 nM
- +
A
TTshld-GFP HeK
Cleaved T2A
FRB-UbN
FKBP*-UbC-GFP
GFP
Actin
Rap (nM)
Doxycycline
0 ng/mL
Doxycycline
50 ng/mL
Doxycycline
1000 ng/mL
0 50 500 0 50 500 0 50 500
B
TTshld-GFP HeLa
Cleaved T2A
FRB-UbN
FKBP*-UbC-GFP
GFP
Actin
Rap (nM)
Doxycycline
0 ng/mL
Doxycycline
50 ng/mL
Doxycycline
1000 ng/mL
0 50 500 0 50 500 0 50 500
C

44

3.2.2 Generation and characterization of TTshld(F36V)-GFP Hek cell
Intriguingly, we observed the release of GFP from the system in the
absence of rapamycin when the cells were treated with doxycycline alone
(Figure 3-2B and 3-2C). Moreover, the amount of GFP released was dose-
dependent on the concentration of doxycycline. Taken together, these data
support a model where rapamycin-induced co-localization of the ubiquitin
fragments (i.e., UbN and UbC) is not absolutely required for their
complementation and GFP release. To test if this is the case, TTShld-GFP
expressing HEK cells were treated with the same concentration of doxycycline
(0, 50, 1000 ng/ mL) and three different concentrations of FK-506 (0, 50,
500 nM), which will bind-to and stabilize FKBP* but not recruit FRB (Kohler
and Bertozzi, 2003). Analysis by western blotting (Figure 3-3A) showed that
GFP was released in a similar dose-dependent manner compared to the cells
that were treated with rapamycin (Figure 3-2B). This demonstrates that the
rate of ubiquitin self-complementation is in large-part driven by the cellular
concentrations of UbN and UbC and not by the rapamycin-mediated physical
interaction between FKBP* and FRB. Importantly, this suggests that
rapamycin and FK506, which both have endogenous cellular targets that
could produce confounding biological effects in a protein concentration
experiment. (Brown et al., 1995; Corradetti and Guan, 2006; Liu et al., 1991),
could be replaced with the orthogonal small-molecule Shld1. (Banaszynski et
al., 2006) To test this possibility, we introduced a F36V mutation into the
FKBP*, enabling the binding of Shld1 and subsequent stabilization of FKBP*

45

and used retroviral transfection to introduce TTShld(F36V)-GFP into Tet-On
HeK cells. We then treated these cells with a combination of three different
concentrations of doxycycline (0, 50, 1000 ng/mL) and four different
concentrations of Shld1 (0, 50, 500, 1000 nM). Analysis by western blotting
demonstrated that the GFP was again released in a dose dependent manner
with no observable uncleaved T2A protein or full-length FKBP*-UbC-GFP
(Figure 3-3B).

Figure 3-3. Inducibility of FK506 in TTshld-GFP Hek and analysis of
TTshld(F36V)-GFP HeK. A) TTShld-GFP HEK cells were treated with the indicated
concentrations of doxycycline and FK506 for 48 hours before Western blotting. B)
TTShld(F36V)-GFP expressing HEK cells were treated with the indicated concentrations of
doxycycline and Shld1 for 48 hours and then analyzed by Western blot.
TTshld-GFP HeK
Cleaved T2A
FRB-UbN
FKBP*-UbC-GFP
GFP
Actin
FK506 (nM)
Doxycycline
0 ng/mL
Doxycycline
50 ng/mL
Doxycycline
1000 ng/mL
0 50 500 0 50 500 0 50 500
A
TTshld(F36V)-GFP HeK
Cleaved T2A
FRB-UbN
FKBP*-UbC-GFP
GFP
Actin
Shld1 (nM)
Doxycycline
0 ng/mL
Doxycycline
50 ng/mL
Doxycycline
1000 ng/mL
0 50 500
B
1000 0 50 500 1000 0 50 500 1000

46

3.2.3 Quantitative analysis of TTshld-GFP
Next, we performed flow cytometry to quantitatively measure GFP
concentration and to compare the different FKBP stabilizing small molecules
(rapamycin, FK506, and Shld1). In combination with three concentrations of
doxycycline, rapamycin precisely controlled the fluorescence intensity of GFP,
measured by flow cytometry, in HEK cells with a dynamic range of 130-fold,
whereas FK-506 in the same cell-line gave a dynamic range of 87-fold (Figure
3-4A and 3-4B). This supports our conclusion that ubiquitin
complementation is driven by the concentrations of UbN and UbC but also
demonstrates that rapamycin-induced physical proximity can further drive
GFP release. Treating Tet-On HEK cells that stably express TTShld(F36V)-
GFP with a range of concentrations of doxycycline and Shld1 resulted in GFP
expression over a 56-fold dynamic range (Figure 3-4C). Because this stable
cell-line is necessarily different from the TTShld cell-line (i.e., it contains the
F36V mutation), it may have inherent differences in absolute transgene
expression. Therefore, to compare the linearity of our control of GFP
concentration, we normalized the highest GFP expression level within the two
cell-lines. Among three FKBP-stabilizing molecules, Shld1 gave the most
linear control of GFP concentration (Data not shown). Therefore, we decided
to investigate the kinetics of GFP expression using these cells. Accordingly, we
treated the cells with 1000 ng/mL doxycycline and 1 mM Shld1 for different
amounts of time before analysis by flow cytometry (Figure 3-4D). GFP

47

expression was observed as early as 4 hours and plateaued at 72, consistent
with transcriptional control.

Figure 3-4. Quantitative analysis of TTShld-GFP. A & B) TTShld-GFP HEK cells were
treated with doxycycline and rapamycin or FK-506 for the indicated concentrations for 48
hours and GFP fluorescence was quantified by flow cytometry. C) TTShld(F36V)-GFP HEK
cells were treated with doxycycline and Shld1 for the indicated concentration before flow
cytometry analysis. D) TTShld(F36V)-GFP HEK cells were treated with doxycycline (1000
ng/mL) and Shld1 (1000 nM) for the indicated lengths of time before flow cytometry analysis.
All experiments were performed in triplicate and error bars represent standard deviation;
MFI: mean fluorescence intensity.
3.2.4 Demonstration of generality of TTshld(F36V) system
15000
TTshld GFP HeK (Rapamycin)
11250
7500
3750
0
Doxycycline
Rapamycin
0
0
0
50
0
500
50
0
50
50
50
500
1000
0
1000
50
1000
500
A
MFI
(ng/mL)
(nM)
8000
TTshld GFP HeK (FK506)
6000
4000
2000
0
Doxycycline
FK506
0
0
0
50
0
500
50
0
50
50
50
500
1000
0
1000
50
1000
500
B
MFI
(ng/mL)
(nM)
8000
TTshld (F36V) GFP Hek
6000
4000
2000
0
Doxycycline
Shld1
C
MFI
0
0
0
50
0
500
0
1000
50
0
50
50
50
500
50
1000
1000
0
1000
50
1000
500
1000
1000
(ng/mL)
(nM)
0 24 48 72 96
100%
75%
50%
25%
Percentage of maximum
GFP fluorescence
D
TTshld (F36V) GFP Hek
t (Hr)

48

Finally, to explore the generality of the TTShld(F36V) system, we
analyzed two additional proteins of interest. We generated Tet-On HeLa cells
that stably express either pyruvate kinase M2 or caspase-3 under control of
the TTShld(F36V) system by retroviral infection. These cell-lines were treated
with 12 different concentration combinations of doxycycline (0, 50, 1000
ng/mL) and Shld1 (0, 50, 500, 1000 nM) for 48 hours and analyzed by
western blotting (Figure 3-5). In both cases we observed good linear
expression level of the protein of interest across the whole spectrum of two
small molecules.
Figure 3-5. Analysis of TTShld(F36V)-PKM2 and TTShld(F36V)-Caspase-3. HeLa
cell-lines stably expressing either construct were treated with indicated concentration of
doxycycline and Shld1 for 48 hours before Western blotting analysis.

Tshld(F36V)-PKM2 HeLa
Uncleaved T2A
FRB-UbN
FKBP*-UbC-PKM2
PKM2
Actin
Shld1 (nM)
Doxycycline
0 ng/mL
Doxycycline
50 ng/mL
Doxycycline
1000 ng/mL
0 50 500 1000 0 50 500 1000 0 50 500 1000
Tshld(F36V)-CASP3 HeLa
FRB-UbN
FKBP*-UbC-CASP3
CASP3
Actin
Shld1 (nM)
Doxycycline
0 ng/mL
Doxycycline
50 ng/mL
Doxycycline
1000 ng/mL
0 50 500 1000 0 50 500 1000 0 50 500 1000
Uncleaved T2A

49

3.3 Discussion and Conclusions
In summary, we have upgraded the protein-concentration tuneability
of the TShld system by placing it under an additional level of small-molecule
control at the transcription level, termed TTShld. Using GFP as a model
protein, TTShld system gives us a 130-fold range access of protein abundance
when using rapamycin as the FKBP*-stabilizing molecule. Interestingly, we
found that formation of the FKBP*-rapamycin-FRB ternary complex and
coincident co-localization of UbN and UbC was not absolutely required for
ubiquitin complementation and release of GFP. Instead, treatment of cells
expressing the TTShld system with FK506 instead of rapamycin, resulted in a
87-fold range of GFP concentrations, demonstrating that UbN and UbC can
self-complement in a concentration dependent fashion. This contradicts
several studies that have used the split-ubiquitin technology in
Saccharomyces cerevisiae to identify and characterize protein-protein
interactions. (Johnsson and Varshavsky, 1994; Kittanakom et al., 2009;
Möckli et al., 2007; Stagljar et al., 1998; Yan and Lennarz, 2005) We attribute
this difference to the inherent differences between mammalian and yeast
cells. Specifically, the higher growth temperature of mammalian cells, 37 ℃
compared to yeast, 30 ℃, could kinetically allow for ubiquitin self-
complementation. Additionally, mammalian cells may express a molecular
chaperone that facilitates split-ubiquitin folding. In either case, this
concentration-dependent self-complementation allowed us to introduce a
F36V mutation in FKBP* to make use of the non-toxic Shld1 as the FKBP*

50

stabilizing molecule, (Banaszynski et al., 2006) yielding a 56-fold range of
GFP concentrations. Therefore, the TTShld system can be applied to systems
where inhibition of the mTOR kinase or calcineurin by rapamycin or FK506,
respectively, should be avoided. (Brown et al., 1995; Corradetti and Guan,
2006; Liu et al., 1991) The TTShld system has a greatly improved dynamic-
range compared to our previously reported system. (Lau et al., 2010), and we
are confident that the traceless feature will be attractive for the elucidation of
the consequences of protein concentration in cell biology.
3.4 Materials and Methods
3.4.1 Plasmid Construction
All constructs were prepared by using standard molecular cloning
techniques. Reference 9 was the source of Tshld-GFP system. PKM2. Caspase
3 were provided by the Muir laboratory at Rockefeller University. Tshld
constructs were cloned into retroviral expression vector pRetroX-Tet-on
vector by using Gateway recombination technology (Clontech). Mutation in
FKBP was introduced by using QuikChange site-directed mutagenesis Kit
(Agilent Technologies).
3.4.2 Cell Culture
All cell lines were maintained in 10 cm tissue culture dishes with
Dulbecco’s modified Eagle’s medium (DMEM, Cellgro) supplemented with 10
% fetal bovine serum (FBS, Atlanta Biologicals, Norcross, GA) at 37℃ and 5%
CO2 according to standard procedures.

51

3.4.3 Transfections and Generation of Stable Cell Lines
AmphoPack-293 cells (Clontech) were transfected with retroviral
expression vectors by using standard calcium phosphate transfection
procedures. The growth medium was then replaced with fresh medium 7 h
post-transfection. Virus-containing medium were collected and filtered with
0.45 µm syringe filters (cellulose, VWR) 24 h post-transfection. Target cell
growth medium was then replaced with the virus-containing medium
supplemented with 4 µg/ mL polybrene (Sigma). Growth medium of the
transfected AmphoPack-293 cells was replenished for subsequent infections.
Infection of the target cells was performed twice a day for a total of 2 days.
Upon completion of the infection, the target cells were allowed to recover in
normal growth medium for 7 h and subsequently selected by incubating in
medium containing 3 µg/mL puromycin (Calbiochem) and 500 µg/mL G418
(Calbiochem) until confluent.
3.4.4 Cell Culture Experiments
For confirming shielding and release of GFP, TTshld GFP HeK cells
were plated around 20% confluence and immediately treated with doxycycline
(1000 ng/ mL) and rapamycin (500nM) for 48 hours and then harvested by
trypsinizing from the plate. For all the concentration course experiments, the
stable cell lines were plated around 20% confluence in 10 cm tissue culture
dishes and immediately treated with indicated concentration of doxycycline

52

and rapamycin/ FK506/ Shld1 for 48 hours before harvested by
trypsinization.
3.4.5 Western Blotting
Cells were washed three times with PBS and then lysed with NP-40
lysis buffer (1% NP-40, 50 mM triethanolamine, 150 mM NaCl, pH 7.4) with
Complete Mini protease inhibitor cocktail (Roche Biosciences). Total protein
concentrations were normalized by Pierce BCA Protein Assay Kit (Thermo
Scientific). 40 µg of lysates were then mixed with 4x SDS loading buffer and
separated by SDS-PAGE before being transferred to PVDF membrane (Bio-
rad) standard western blotting protocols.
All western blots were blocked in TBST (10 mM Tris, 150 mM NaCl,
0.1% Tween-20, pH 8.0) containing 5% non-fat milk for 1 h at RT. Then the
membranes were incubating with the following primary antibody at RT for 1 h
under constant rocking: anti-HA (1:1000 dilution, Covance), anti-FLAG M2
antibody (1:2000 dilution, Sigma). The anti-actin antibodies were used at
1:1000 dilutions. The blots were then washed with TBST three times for 30
mins and then incubated with corresponding horseradish peroxidase (HRP)-
conjugated secondary antibodies (Jackson ImmunoResearch) for 1 h in TBST
containing 5% non-fat milk. After being washed three times with TBST, the
blots were developed by using ECL reagents (Bio-Rad) and imaged by
ChemiDoc XRS+ molecular imager (Bio-Rad).

53

3.4.6 Flow Cytometry Experiments
For the concentration course experiments, cells were plated in 6-well
plates around 20% confluence and then immediately treated with indicated
concentration of doxycycline and rapamycin/ FK506/ Shld1 for 48 hours. The
cells were trypsinized from each well and washed with PBS one time. The
resulting cell pellets were resuspended in 1 mL of PBS and analyzed for GFP
mean fluorescence intensity at the USC FACS Core using the BD LSR II flow
cytometer with 10000 events collected for analysis. For the time course
experiment, TTshld (F36V) HeK cells were treated with 1000 ng/ mL of
doxycycline and 1000 nM of Shld1 for different lengths of time before
undergoing the above mentioned sample preparation procedures for flow
cytometry.
3.5 Reference
Banaszynski, L.A., and Wandless, T.J. (2006). Conditional Control of Protein
Function. Chem. Biol. 13, 11–21.
Banaszynski, L.A., Chen, L.-C., Maynard-Smith, L.A., Ooi, A.G.L., and
Wandless, T.J. (2006). A Rapid, Reversible, and Tunable Method to Regulate
Protein Function in Living Cells Using Synthetic Small Molecules. Cell 126,
995–1004.
Banaszynski, L.A., Sellmyer, M.A., Contag, C.H., Wandless, T.J., and Thorne,
S.H. (2008). Chemical control of protein stability and function in living mice.
Nat. Med. 14, 1123–1127.
Bargonetti, J., and Manfredi, J.J. (2002). Multiple roles of the tumor
suppressor p53. Curr Opin Oncol 14, 86–91.
Brown, E.J., Beal, P.A., Keith, C.T., Chen, J., Shin, T.B., and Schreiber, S.L.
(1995). Control of p70 s6 kinase by kinase activity of FRAP in vivo. Nature
377, 441–446.

54

Chen, M., David, C.J., and Manley, J.L. (2012). Concentration-dependent
control of pyruvate kinase M mutually exclusive splicing by hnRNP proteins.
Nat Struct Mol Biol 19, 346–354.
Chu, B.W., Banaszynski, L.A., Chen, L.-C., and Wandless, T.J. (2008). Bio. &
Medi. Chem. Letters 18, 5941–5944.
Corradetti, M.N., and Guan, K.-L. (2006). Upstream of the mammalian target
of rapamycin: do all roads pass through mTOR? Oncogene 25, 6347–6360.
Holst, J., Vignali, K.M., Burton, A.R., and Vignali, D.A.A. (2006). Rapid
analysis of T-cell selection in vivo using T cell–receptor retrogenic mice.
Nature Methods 3, 191–197.
Iwamoto, M., Björklund, T., Lundberg, C., Kirik, D., and Wandless, T.J.
(2010). A General Chemical Method to Regulate Protein Stability in the
Mammalian Central Nervous System. Chemistry & Biology 17, 981–988.
Johnsson, N.N., and Varshavsky, A.A. (1994). Split ubiquitin as a sensor of
protein interactions in vivo. Proc. Natl. Acad. Sci. U.S.A. 91, 10340–10344.
Kittanakom, S., Chuk, M., Wong, V., Snyder, J., Edmonds, D., Lydakis, A.,
Zhang, Z., Auerbach, D., and Stagljar, I. (2009). Analysis of Membrane
Protein Complexes Using the Split-Ubiquitin Membrane Yeast Two-Hybrid
System. In Methods in Molecular Biology, (Totowa, NJ: Humana Press), pp.
247–271.
Kohler, J.J., and Bertozzi, C.R. (2003). Regulating Cell Surface Glycosylation
by Small Molecule Control of Enzyme Localization. Chemistry & Biology 10,
1303–1311.
Lau, H.D., Yaegashi, J., Zaro, B.W., and Pratt, M.R. (2010). Precise Control of
Protein Concentration in Living Cells. Angew. Chem. Int. Ed. 49, 8458–8461.
Liu, J., Farmer, J.D., Lane, W.S., Friedman, J., Weissman, I., and Schreiber,
S.L. (1991). Calcineurin is a common target of cyclophilin-cyclosporin A and
FKBP-FK506 complexes. Cell 66, 807–815.
Möckli, N., Deplazes, A., Hassa, P., Zhang, Z., Peter, M., Hottiger, M., Stagljar,
I., and Auerbach, D. (2007). Yeast split-ubiquitin-based cytosolic screening
system to detect interactions between transcriptionally active proteins.
Biotech. 42, 725–730.
Stagljar, I., Korostensky, C., Johnsson, N., and Heesen, te, S. (1998). A genetic
system based on split-ubiquitin for the analysis of interactions between
membrane proteins in vivo. Proc. Natl. Acad. Sci. U.S.A. 95, 5187–5192.

55

Stankunas, K., Bayle, J.H., Havranek, J.J., Wandless, T.J., Baker, D.,
Crabtree, G.R., and Gestwicki, J.E. (2007). Rescue of Degradation-Prone
Mutants of the FK506-Rapamycin Binding (FRB) Protein with Chemical
Ligands. ChemBioChem 8, 1162–1169.
Szymczak, A.L., Workman, C.J., Wang, Y., Vignali, K.M., Dilioglou, S., Vanin,
E.F., and Vignali, D.A.A. (2004). Correction of multi-gene deficiency in vivo
using a single “self-cleaving” 2A peptide–based retroviral vector. Nat
Biotechnol 22, 589–594.
Yan, A., and Lennarz, W.J. (2005). Unraveling the Mechanism of Protein N-
Glycosylation. J. Biol. Chem. 280, 3121–3124.

56

Chapter 4. Role of Glutamine Fructose-6-
phosphate Amidotransferase (GFAT) in O-GlcNAc
modification and Tumorigenesis
4.1 Introduction
Tumor and proliferative tissues exhibit a special type of metabolism,
termed aerobic glycolysis, where the glucose taken up by the cells is shuttled
away from mitochondrial oxidative phosphorylation. This altered glucose
metabolism, or energy metabolism, has been considered to be one of the
emerging hallmarks of cancer(Hanahan and Weinberg, 2011). The redirected-
glucose not only fuels the production of biomass but also provides the carbon
source for pathways that diverted from earlier steps of glycolysis, including
hexosamine biosynthetic pathway (HBP). The end product of HBP is the
activated nucleotide sugar donor UDP-GlcNAc, which can be utilized by OGT
to result in O-GlcNAc modification of myriad of proteins.
The rewired glucose metabolism cooperates with increased uptake of
glutamine(Dang, 2010) to maximize flux of metabolites into HBP. In fact, O-
GlcNAcylation has found to be up-regulated in virtually all cancer type
examined so far(Issad, 2013). This hyper-O-GlcNAcylation is driven in part by
oncogenes such as Kras, which not only up-regulate level of glucose
transporters and glycolytic enzymes, but also up-regulate GFAT1(Ying et al.,
2012), the rate limiting enzyme of HBP.
Furthermore, the activation of HBP and the subsequent elevation of O-
GlcNAc level have been shown to expedite cancer cell cycle progression(Lynch

57

et al., 2012), promote cancer cell survival under stress condition
(Guillaumond et al., 2013; Jones et al., 2014; Wang et al., 2014), support
angiogenesis (Lynch et al., 2012; Zhu et al., 2011), contribute to cancer cell
metastasis and invasion(Caldwell et al., 2010; Lynch et al., 2012) and
reinforce the metabolic reprogramming in cancer cells(Yi et al., 2012).
Notably, knocking down OGT inhibits cancer cell proliferation, tumor
invasion and metastasis in vivo and in vitro in breast and prostate cancer
cells(Caldwell et al., 2010).
Therefore, strategies to lower O-GlcNAcylation levels in cancer have
been developed, which includes the development of small molecule OGT
inhibitors. However, current available OGT inhibitors have off-target effects
and poor IC50s, which compromised their usage. Specifically, Vocadlo lab at
Simon Fraser University developed a metabolic precursor, 5S-GlcNAc (Figure
4-1A), that after taken up by the cells, it will be incorporated into HBP in
generating UDP-5S-GlcNAc, which dramatically slows down the transfer rate
of the GlcNAc moiety towards substrates(Gloster et al., 2011). However, as a
substrate mimic, UDP-5SGlcNAc may inhibit other UDP-GlcNAc-dependent
enzymes. Therefore, it is not specific towards OGT. Walker lab at Harvard
developed another two inhibitors for OGT. The first inhibitor contains a five-
heteroatom dicarbamate (Figure 4-1B) core that irreversibly inhibits OGT by
crosslinking Lys842 and Cys917 of OGT(Jiang et al., 2011). The second
inhibitor was discovered through a high-throughput screening campaign,
which result in a inhibitor, OSMI-1 (Figure 4-1C), that has a similar IC50

58

compared to 5S-GlcNAc(Ortiz-Meoz et al., 2015). Even though treating cells
with 50 µM of OSMI-1 dramatically lower O-GlcNAc levels, it activates
apoptosis pathway and result in around 50% of cell death.

Figure 4-1. Structure of OGT inhibitors. A) 5S-GlcNAc B) Benzoxazolinone compound
#2 developed by Walker lab C) OSMI-1
While efforts are pouring in for the development and optimization of
OGT inhibitors, we decide to look up-stream from where the O-GlcNAc
modification takes place, searching for a target within HBP to lower the O-
GlcNAcylation level by restricting the production of UDP-GlcNAc. Elegant
work done by Boehmelt and coworkers showed that inactivation of EMeg32,
murine version of the glucosamine-6-phosphate acetyltransferase, results in
decreased level of UDP-GlcNAc, impaired proliferation as well as embryonic
lethality in mouse embryonic stem cell(Boehmelt et al., 2000). Furthermore,
the authors showed that administration of 10 mM GlcNAc was able to rescue
the defects caused by knock-out of the gene. Their data demonstrates that
S
OH
HO
HO
OH
NH
O
N
H
O
S
O
O
NH
N
O
O
OCH
3
S
B A C
N
O
O
O
O
O
O

59

inhibition of enzymes that situated earlier in the HBP constrains production
of the UDP sugar donor and may result in lethal phenotype. Also,
pharmacological inhibition of GFAT as well as GFAT1 knockdown exacerbates
simulated ischemia/reperfusion (I/R) stress in an in vitro heart-attack
model(Wang et al., 2014) by hampering the activation of HBP and subsequent
elevation of O-GlcNAc level, suggesting GFAT as a potential target to
influence O-GlcNAc modification level in cancer. Taking all the data above
into consideration, we decided to investigate the contribution of GFAT to O-
GlcNAc modification level in cancer cells.
Expressed from two different genes, three different isoforms of GFAT
exist in human: GFAT1 isoform 1 (GFAT1-1), GFAT1 isoform 2 (GFAT1-2) and
GFAT2. GFAT1-1 and GFAT1-2 are expressed from the same gene, GFPT1
located at 2p13, through alternative splicing. GFAT1-1 has 699 amino acids
and weighs 78.806 kilo-Dalton (kD.), whereas GFAT1-2 has 681 amino acids
and has a weight of 76.759 kD.. GFAT2, encoded by GFPT2 located at 5q34-
q35, has 682 amino acids and weighs 76.931 kD.. GFAT1-1 and GFAT1-2,
hereafter GFAT1, are predominantly expressed in heart, placenta, lung, liver,
pancreas and skeletal muscle, whereas GFAT2 is expressed throughout the
central nervous system and ovaries. Despite the fact that GFAT1 is the main
isoform expresse in the pancreas, GFAT1 and GFAT2 were upregulated
differentially during hypoxia in pancreatic cancer cell lines(Guillaumond et
al., 2013), suggesting each isoform might assume different roles during the
stress response in cancer.

60

In this study, we investigate the relationship between O-GlcNAcylation
and GFAT activity as well as protein abundance in cancer cells through three
different approaches. We revealed GFAT as a modulator in controlling not
only the O-GlcNAcylation inside the cancer cells but also the availability of
UDP-GlcNAc, which also supports N-linked as well as O-linked glycosylation
events important for tumorigenesis.
4.2 Results
4.2.1 Inhibition of GFAT sensitizes cancer cells to oxidative stress
To evaluate the contribution of GFAT activity towards O-GlcNAc
modification level in cancer cells, I first took advantage of a known and widely
used small molecule inhibitor of GFAT: 6-diazo-5-oxo-L-norleucine, DON
(Ahluwalia et al., 1990; Dehennaut et al., 2007; Gurel et al., 2014; Herzog et
al., 2014). DON is a antitumor agent isolated from Streptomyces and it
inhibits glutamine-utilizing enzymes through an irreversible alkylation of the
cysteine in their active sites(LaRonde-LeBlanc et al., 2009). To investigate the
effect of DON on global O-GlcNAc modification level in cancer cells, I treated
H1299 cells, a non-small cell lung carcinoma cell line, with 50 µM DON for up
to two days. Western blotting analysis showed that 50 µM DON effectively
lowered O-GlcNAc level in H1299 (Figure 4-2).

61

Figure 4-2. Effect of DON on H1299. H1299 cells were treated with vehicle (left lane), 1
day DON at 50 µM (middle lane) and 2 days DON at 50 µM (right lane) and the lysates were
probed with O-GlcNAc antibody RL2 as well as the Actin antibody to serve as loading control.
Since the upregulation of O-GlcNAc level has proven to be important
for cell survival under various cell and tissue stressors(Ferrer et al., 2014;
Kang et al., 2009; Wang et al., 2014), we next ask how would cancer cells’ O-
GlcNAc modification levels respond to oxidative stress with or without GFAT
inhibition. Towards this end, I treated H1299, A549 and MCF7 cells with
either vehicle, DON, diamide, or a combination of DON and Diamide for 40
hours. Western blotting analysis showed that the three lung cancer cells
responded to the treatment in a similar fashion, where the cells increased the
global O-GlcNAc level under diamide treatment with or without DON
treatment (Figure 4-3).

kDa
150
100
75
50
37
O-GlcNAc (RL2)
Anti-Actin
H1299

62

Figure 4-3. Effect of DON and Diamide treatment on O-GlcNAc level in three
different lung cancer cells. For all three cell lines, the first lane is the vehicle treated
sample. The second lane is the cells that have only been treated with diamide 16 hours prior
to the work up. The third lane is the cells that has treated with DON for 40 hours prior to the
work up. The forth lane is the cells that were first treated with DON for 24 hours followed by
DON and diamide treatment for another 16 hours. The concentration of Diamide is 100 µM
for all three cancer cells. 50 uM DON was applied to A549 and H1299 cells, whereas 100 uM
DON was applied to MCF7 cells. The lysates were probed with O-GlcNAc antibody RL2 as
well as the Actin antibody to serve as loading control.
Next, I utilized MTS cell proliferation assay to evaluate if DON,
diamide or a combination of both small-molecules affected the viability of
cancer cells. The panel of lung cancer cells was then treated with a range
concentration of diamide, 100 µM DON, or with both small-molecules
together for up to 40 hours, followed by the execution of MTS assay
(Promega) based on manufacturer’s protocol. The assay revealed that treating
the cells with either of the small molecule inhibits their proliferation and
treating the cells with both molecules had an additive effect (Figure 4-4).

kDa
150
100
75
50
37
O-GlcNAc (RL2)
A549 H1299
Anti-Actin
MCF7
DON
Diamide
-
-
-
+
+
-
+
+
-
-
-
+
+
-
+
+
-
-
-
+
+
-
+
+

63

Figure 4-4. Effect of DON and Diamide treatment on cell proliferation in three
different lung cancer cells. For all cell lines, the first lane is the vehicle treated sample.
The second lane to the forth lane are the cells that has only been treated with indicated
concentration of diamide 16 hours prior to the work up. The fifth lane is the cells that has
treated with DON for 40 hours prior to the work up. The sixth lane to the eighth lane are the
cells that were first treated with 100 µM DON for 24 hours followed by 100 µM DON and
indicated concentration of diamide treatment for another 16 hours. Signals were all
normalized to vehicle treated sample within each cell line. All condition was performed in
triplicate. Error bars represent s.e.m..(* = P < 0.05 ** = P < 0.01 *** = P < 0.001 **** = P <
0.0001 N.S. = Not significant)
Before the execution of the MTS assay, I observed double small molecules-treated
wells contained a significant amount of adherent cells exhibiting shrunk cell body
as well as blebs (data not shown), features of apoptotic cells(Ziegler, 2004).
Therefore, to gain a clearer picture about how the two small molecules are
affecting cancer cells’ viability, we decided to switch the assay from MTS to one
that would allow us to measure the activity of Caspase-3 and Caspase-7, two of
three important executioner caspases that carry out apoptotic signals for
programmed cell death(McIlwain et al., 2013). Caspase-Glo 3/7 Assay (Promega)
was performed after the same treatment as in Figure 4-4 in the same panel of
lung cancer cells and the result showed that double-molecule treatment had a
A549
***
****
N.S.
40
60
80
100
20
****
**
Diamide (μM)
DON (μM)
0
0
100
0
150
0
200
0
0
100
100
100
150
100
200
100
Normalized Absorption
(MTS)
N.S.
** ** **
*
0
H1299
***
**
**
* ***
****
* ***
N.S.
N.S.
40
60
80
100
20
0
0
100
0
150
0
200
0
0
100
100
100
150
100
200
100
0
****
****
**
****
N.S.
40
60
80
100
20
****
**
MCF7
0
0
100
0
150
0
200
0
0
100
100
100
150
100
200
100
*** *
**
0

64

compounded effect, especially in H1299 and MCF7, on activating Caspase-3 and/
or Caspase-7 compared to either of the single-molecule treatment (Figure 4-5).

Figure 4-5. Effect of DON and Diamide treatment on Caspase-3 and/ or Caspase-
7 activation in three different lung cancer cells. For all cell lines, the first lane is the
vehicle treated sample. The second lane is the cells that has only been treated with indicated
concentration of diamide 16 hours prior to the assay. The third lane is the cells that has
treated with 100 µM DON for 40 hours prior to the assay. The forth lane is the cells that were
first treated with 100 µM DON for 24 hours followed by 100 µM DON and indicated
concentration of diamide treatment for another 16 hours. Signals from each sample were first
normalized to the cell number at the end of the assay in each condition before normalizing to
vehicle treated sample within each cell line. MCF7 is Caspase-3 null; therefore, the signal is
entirely coming from Caspase-7 activity. Error bars represent s.e.m.. (* = P < 0.05 ** = P <
0.01 *** = P < 0.001 **** = P < 0.0001 N.S. = Not significant)
This result demonstrated that inhibition of GFAT by DON sensitized
the three lung cancer cells to oxidative stress, potentially through a
mechanism that involves lowering global O-GlcNAc modification. However,
as a glutamine antagonist, DON theoretically inhibits all glutamine-utilizing
enzymes inside the cells(LaRonde-LeBlanc et al., 2009). Therefore, one can
reasonably argue that the sensitization might not be caused by GFAT
A549
Diamide (μM)
DON (μM)
0
0
200
0
0
100
200
100
Normalized caspase-3/7 activity
(Luminescence)
*** ****
*
****
***
3.0
4.0
2.0
1.0
0.0
H1299
Diamide (μM)
DON (μM)
0
0
100
0
0
100
100
100
Normalized caspase-3/7 activity
(Luminescence)
6.0
8.0
4.0
2.0
0.0
****
N.S.
N.S.
*
MCF7
100
Diamide (μM)
DON (μM)
0
0
150
0
0
100
150
100
Normalized caspase-7 activity
(Luminescence)
150
50
0.0
****
****
***
**

65

inhibition or the down regulation of O-GlcNAc modification but stems from
inhibition of other enzymes. To address this concern as well as to gain insights
into how DON sensitize lung cancer cells to oxidative stress, we decided to
treat the cells with N-acetylglucosamine (GlcNAc) to rescue the O-GlcNAc
levels (Figure 4-2) under DON treatment. Towards this end, I first treated the
same panel of lung cancer cells with 10 mM GlcNAc on top of the DON
treatment to rescue the modification levels. Western blotting analysis showed
that GlcNAc supplementation was able to reverse the effect of DON on
dropping global O-GlcNAc modification level in all cell lines (Figure 4-6).

Figure 4-6. Effect of GlcNAc supplementation on O-GlcNAc modification under
DON treatment. For all cell lines, the first lane is the vehicle treated sample. The second
lane is the cells that have only been treated with 100 µM DON for 40 hours prior to the work
up. The third lane is the cells that were treated with both 100 µM DON and 10 mM GlcNAc
for 40 hours. The lysates were probed with O-GlcNAc antibody RL2 as well as the actin
antibody to serve as loading control.
Having confirmed GlcNAc supplementation was able to restore O-GlcNAc levels
under DON treatment, we next ask if GlcNAc supplementation has any effect on
A549 H1299 MCF7
DON (100 μM)
GlcNAc (10 mM)
-
-
+ +
+ -
-
-
+ +
+ -
-
-
+ +
+ -
kDa
150
100
75
50
37
O-GlcNAc (RL2)
Anti-Actin

66

activation of Caspase-3 and -7 in three lung cancer cell lines under both DON and
diamide treatment. The panel of cancer cell lines was challenged again with the
tw0-small molecules with or without GlcNAc supplementation followed by
analysis with the Caspase-Glo 3/7 Assay. The result showed that the sugar
treatment was able to effectively prevent Caspase-3 and -7 activation in A549 and
H1299 cells, whereas it lowered Caspase-7 activation down to the level of diamide
treated sample in MCF7 (Figure 4-7). This data demonstrated that DON was
mainly on-target in our research setting: during oxidative stress, the vulnerability
of lung cancer cells exerted by DON was largely due to its effect on lowering
global O-GlcNAc modification level potentially through GFAT inhibition.
Figure 4-7. Effect of DON, Diamide and GlcNAc treatment on Caspase-3 and/ or
Caspase-7 activation in three different lung cancer cells. For all cell lines, the first
lane is the vehicle treated sample. The second lane is the cells that have only been treated
with indicated concentration of diamide 16 hours prior to the assay. The third lane is the cells
that were first treated with 100 µM DON for 24 hours followed by 100 µM DON and indicated
concentration of diamide treatment for another 16 hours. The forth lane is the cells that were
first treated with 100 µM DON as well as 10 mM GlcNAc for 24 hours followed by 100 µM
DON, 10 mM GlcNAc and indicated concentration of diamide treatment for another 16 hours.
Signals from each sample were first normalized to the cell number at the end of the assay in
each condition before normalizing to vehicle treated sample within each cell line. Error bars
represent s.e.m..(* = P < 0.05 ** = P < 0.01 *** = P < 0.001 **** = P < 0.0001 N.S. = Not
significant)
MCF7
Diamide (μM)
DON (μM)
0
0
150
0
150
100
150
100
Normalized caspase-7 activity
(Luminescence)
9.0
12.0
6.0
3.0
0.0
****
GlcNAc (mM) 0 0 0 10
A549
Diamide (μM)
DON (μM)
0
0
200
0
200
100
200
100
Normalized caspase-3/7 activity
(Luminescence)
3.0
4.0
2.0
1.0
0.0
***
GlcNAc (mM) 0 0 0 10
H1299
Diamide (μM)
DON (μM)
0
0
100
0
100
100
100
100
Normalized caspase-3/7 activity
(Luminescence)
6.0
8.0
4.0
2.0
0.0
****
N.S.
GlcNAc (mM) 0 0 0 10
*
****
****

67

4.2.2 Generation and Characterization of GFAT KD cell lines
After evaluation of GFAT as a potential cancer target for lowering its O-
GlcNAc modification level through pharmacology, we decided to target GFAT
at its RNA level in various cancer cell lines. We chose to lower GFAT mRNA
level by using short-hairpin RNA (shRNA) knockdown (KD) strategy to
decrease the GFAT protein level in cancer cells. Towards this end, we utilized
2 different shRNA knockdown plasmids, pLKO.1-GFAT1-puromycin and
pLKO.1-GFAT2-neomycin, to target the two isoforms of GFAT separately.
Hek293 cells were transfected with either of the plasmid to generate lentiviral
particles that contain the shRNA sequences for targeting either of the isoform.
Infectious particles were then incubated with target cell lines, followed by
either puromycin or neomycin treatment to give rise to stable knockdown cell
lines for GFAT1 or GFAT2, respectively. I first applied either of the viruses to
MCF7, a breast cancer cell line, as well as to two lung cancer cell lines A549
and H1299, to generate MCF7 GFAT1 KD, MCF7 GFAT2 KD, A549 GFAT1 KD,
A549 GFAT2 KD, H1299 GFAT1 KD and H1299 GFAT2 KD cells. Taking
advantage of an antibody that recognized both isoforms of GFAT, Western
blotting analysis showed that GFAT1 shRNA dramatically decreased total
GFAT protein level in all cell lines, whereas GFAT2 shRNA result in a slight
reduction in GFAT protein level in MCF7, and a moderate reduction in A549
(Figure 4-8). However, in A549 and MCF7 KD cells, the global O-GlcNAc
levels were essentially undisturbed by the knockdown of either of the GFAT
isoforms. This observation made us speculate if the KD cell lines were able to

68

maintain the O-GlcNAc level by upregulation of OGT and/ or down-regulation
of OGA protein level. To test this possibility, I probed the lysate with anti-
OGT as well as anti-OGA antibody in A549 and MCF7 KD cell lines and the
results showed all KD cell lines had a slight increase in OGT protein level
compared to the control KD cells whereas OGA level is largely unchanged
(Figure 4-8). Western blotting analysis of the resulting GFAT knockdown
H1299 cells revealed GFAT1 shRNA was able to lower the global O-GlcNAc
level, whereas GFAT2 shRNA had essentially no effect on either total GFAT
level or global O-GlcNAc level (Figure 4-8). Also, GFAT1 shRNA slightly
increased OGA protein level, whereas GFAT2 shRNA slightly increased both
OGT and OGA protein level compared to the eGFP control knockdown H1299
cells (Figure 4-8).

69

Figure 4-8. Western blotting analysis of MCF7 GFAT1 KD, MCF7 GFAT2 KD,
A549 KD GFAT1 KD, A549 GFAT2 KD, H1299 GFAT1 KD and H1299 GFAT2 KD.
For all cell lines, the eGFP lane corresponds to the control KD cells generated from shRNA
targeting expressed green fluorescent protein (eGFP) through the same lentiviral
transfection/ infection protocol as in GFAT1/2 KD cells. The lysates were probed with O-
GlcNAc antibody RL2, anti-GFAT, anti-OGA, anti-OGT as well as the actin antibody to serve
as loading control.
To exclude any possibility that the effect of GFAT1 and GFAT2
knockdown in MCF7, A549 and H1299 was cell line dependant, I next applied
the same set of lentiviral viral particles to another breast cancer cell line:
MDA-MB-231. Similar to the results that we got from MCF7 GFAT KD cells,
MDA-MB-231 GFAT KD cells maintained their O-GlcNAc modification level
despite successful knockdown of GFAT1 or GFAT2 (Figure 4-9A).
A549
Anti-Actin
Anti-GFAT
Anti-OGA
Anti-OGT
MCF7
KD
GFAT1 eGFP
KD
GFAT2
(Anti-O-GlcNAc)
RL2
(Anti-O-GlcNAc)
RL2
Anti-Actin
Anti-GFAT
Anti-OGA
Anti-OGT
KD
GFAT1 eGFP
KD
GFAT2
kDa
150
100
75
50
37
kDa
150
100
75
50
37
(Anti-O-GlcNAc)
RL2
150
100
75
50
37
kDa
H1299
KD
GFAT1 eGFP
KD
GFAT2
Anti-Actin
Anti-GFAT
Anti-OGA
Anti-OGT

70

Figure 4-9. Western blotting analysis of MDA-MB-231 GFAT1 KD, GFAT2 KD and
double-knockdown cells. All lysates were probed with O-GlcNAc antibody RL2, anti-
GFAT, anti-OGA, anti-OGT as well as the actin antibody to serve as loading control. The eGFP
lane corresponds to the control KD cells generated from shRNA targeting expressed green
fluorescent protein (eGFP) through the same lentiviral transfection/ infection protocol as in
GFAT1/2 KD cells. A) GFAT1 and GFAT2 were first targeted by shRNA separately in MDA-
MB-231 cells. In the anti-GFAT blot, the upper thicker band is GFAT1, whereas the lower
small band is GFAT2 B) The middle lane was the cells that first introduced with GFAT1
shRNA followed by GFAT2 shRNA. The third lane was the cells that first introduced with
GFAT2 shRNA followed by GFAT1 shRNA.
Unlike MCF7 cells that mainly expresses GFAT1 (Figure 4-8), MDA-
MB-231 cells highly express both GFAT1 and GFAT2 (Figure 4-9A). Therefore,
protein levels of the other isoform that was not targeted by shRNA remains
largely the same, especially in the case of MDA-MB-231 GFAT2 KD cells
(Figure 4-9A). To efficiently target both of the GFAT isoforms in MDA-MB-
231 cells, I applied the lentiviral particles that target the remaining GFAT
(Anti-O-GlcNAc)
RL2
(Anti-O-GlcNAc)
RL2
MDA-MB-231
KD
GFAT1 eGFP
KD
GFAT2
MDA-MB-231
Anti-Actin
Anti-GFAT
Anti-OGA
Anti-OGT
eGFP
KD
GFAT1
GFAT2
KD
GFAT2
GFAT1
Anti-Actin
Anti-GFAT
A
B
150
100
75
50
37
kDa
150
100
75
50
kDa
Anti-OGA
Anti-OGT

71

isoform in the GFAT1 KD and GFAT2 KD cells to generate double-knockdown
cells. Western blotting revealed that the O-GlcNAc level was not affected in
the resulting MDA-MB-231 double-knockdown cells; however, the cells
responded to the double GFAT knockdown with upregulation of OGT and
down-regulation of OGA protein level (Figure 4-9B). Our data revealed that
O-GlcNAc modification level responded differently to the expression of
GFAT1 shRNA and GFAT2 shRNA in different type of cancer cells. In other
words, different cancer cell types exhibit different degree of flexibility in
maintaining the O-GlcNAc level in response to GFAT inhibition through the
degradation of its mRNA.
4.2.3 Generation and Characterization of GFAT KO H1299 cells
The data that we obtained in the GFAT KD cell lines suggested that the
cancer cells were able to maintain O-GlcNAc level potentially by adjusting
OGT and/ or OGA level in response to the drop of GFAT protein level.
Another possibility, which is not mutually exclusive to the previous one, is
that the remaining GFAT level, which was barely detectable by western
blotting, was able to maintain the flux of HBP and to provide the cells with
enough UDP-GlcNAc for OGT to modify its substrates. Therefore, to entirely
deprive the cancer cells with GFAT, we turn to genomic editing tools that
would allow us to abolish the expression of GFAT1 as well as GFAT2 genes.
We decided to use clustered regularly interspaced short palindromic repeats
(CRISPR)/ Cas9 system for its facile generation of knockout cell lines as well
as its widespread successful applications in the scientific community(Sander

72

and Joung, 2014). Guidelines and procedures for utilizing CRISPR/ Cas9
system to generate a knockdown cells in vitro can be found in a detailed
protocol (Ran et al., 2013) published by Zhang lab. While the paper provided
us with a informational and useful guidance, certain modifications as well as
adaptations have been made to the protocols based on the resources to which
my lab was accessible as well as the GFAT genes that we are targeting. The
following four sections documented the steps we took to generate as well as to
characterize H1299 GFAT KO cells.
4.2.3.1 Design of 20-nucleotide guide RNA
In human cells, three isoforms of GFAT are expressed from two
different genes: GFPT1, which is located at 2p13, and GFPT2, which is locatd
at 5q34-q35. Therefore, in order to completetly abolish the expression of all
three GFAT isoforms, two genes located at two different genomic loci have to
be targeted by Cas9 nuclease. To ensure facile generation of the GFAT KO
cells, we decided to target both genes with just one guide RNA sequence. In
the search for the single guide RNA sequence, I first aligned the mRNA of
GFAT1 and GFAT2 using MegAlign (DNASTAR) and then I looked for a
stretch of shared nucleotide (nt) sequence in between two genes which has to
fullfill the NGG PAM requirement at its 3’ end. Fortunately, I was able to find
a 14 nucleotides overlapping sequence which fulfuill the PAM requirment
located at exon 2 of each gene. Based on this 14-nt overlapping sequence in
between GFPT1 and GFPT2, I designed 2 different guide RNAs to see if either

73

one of them could achieve our goals. The first 20-nt guide RNA fully matches
the GFPT1 gene within exon 2, termed GFAT1 guide RNA; however, it has
three mismatches towards the GFPT2 gene (Figure 4-10A).

Figure 4-10. Schematic of the guide RNA design. A) The 20-nt sequence preceded the
AGG PAM (colored in green) is GFAT1 guide RNA sequence, which aligns perfectly towards
GFPT1 gene within exon 2, whereas it has 3 mismatches towards GFPT2 gene. B) A 20-nt
Mix-target guide RNA was designed to target both genes by having 1 mismatch towards
GFPT1 gene and 2 mismatches towards GFPT2 gene.
Studies have shown that Cas9 is able to tolerate certain mismatches
between guide RNA and target DNA, depending on the number, location as
well as the identity of the mismatches within the 20-nt guide RNA sequence
(Hsu et al., 2013). Specifically, Hsu et al. dicovered that single-base
mismatches in the PAM-distal region are tolerated to a greater extent
compared to the PAM-proximal region. However, when the number of
mismatches increases, the activity of Cas9 decreases. Therefore, even though
GFAT1 guide RNA has three mismatches towards GFPT2 gene, the cleavage
might not be abolished since all of the mismatches are located away from the
GFAT1 guide RNA sequence : CTT CAG AGA CTG GAG TAC AG AGG
5’-CTTCAGAGA CTGG AGT ACAG-3’
3’-GAAGTCTCT GACC TCA TGTC-5’
5’-CUUCAGAGACUGGAGUACAG-3’
GFPT1 DNA GFPT2 DNA
5’-CTGCAGCGG CTGG AGT ACAG-3’
3’-GACGTCGCC GACC TCA TGTC-5’
5’-CUUCAGAGACUGGAGUACAG-3’
3 mismatches
5’-CTTCAGAGA CTGG AGT ACAG-3’
3’-GAAGTCTCT GACC TCA TGTC-5’
5’-CUUCAGCGACUGGAGUACAG-3’
GFPT1 DNA
1 mismatch
5’-CUUCAGCGACUGGAGUACAG-3’
GFPT2 DNA
2 mismatches
5’-CTGCAGCGG CTGG AGT ACAG-3’
3’-GACGTCGCC GACC TCA TGTC-5’
100% match
Mix-target guide RNA sequence : CTT CAG CGA CTG GAG TAC AG AGG B
A

74

PAM sequence. In case GFAT1 guide RNA is not able to target GFPT2 gene, I
designed a second 20-nt guide RNA that has sequence input from both genes,
termed Mix-target guide RNA. The Mix-target guide RNA has one mismatch
towards GFPT1 gene and two mismatches towards GFPT2 gene (Figure 4-
10B). Since single mismatch and double mismatches located towards the
PAM-distal region are usually tolerated by Cas9(Hsu et al., 2013), we
anticipated Mix-target guide RNA would be able to target both genes at the
same time.
4.2.3.2 Transfection and Generation of H1299
With two promising guide RNA sequences at hand, I generated two
CRISPR GFAT plasmids, termed CRISPR GFAT1 and CRISPR Mix-target, by
cloning the two guide RNA sequences into pSpCas9(BB)-2A-Puro (Addgene
plasmid ID : 48139), a plasmid containing Cas9, sgRNA scaffold that is
important for optimal Cas9 expression and activity in vivo(Hsu et al., 2013),
as well as the puromycin resistance gene that would allow us to perform a
positive selection post transfection. I then used lipofectamine 2000
(Invitrogen) to transiently transfect either CRISPR GFAT1 or CRISPR Mix-
target into H1299 cells, followed by puromycin selection for 2 days. The
positive population from two different trasnfection were allowed to grow to
confluent in complete media supplemented with 10 mM of GlcNAc, followed
by Western blotting analysis. The result revealed that GFAT protein level was

75

no longer detectable by the antibody we used in either H1299 CRISPR GFAT1
or H1299 CRISPR Mix-target cells (Figure 4-11A).

Figure 4-11. Western blotting analysis of H1299 CRISPR GFAT1 and H1299
CRISPR Mix-target cells. A) Lysates of H1299 Wild-type, H1299 CRISPR GFAT1 and
H1299 CRISPR Mix-target cells were probed with anti-GFAT and anti-actin antibody. The
anti-GFAT antibody recognizes both GFAT1 and GFAT2 isoforms B) O-GlcNAc modification
levels were detected by anti-O-GlcNAc antibody RL2 in H1299 wild-type cells as well as in
H1299 CRISPR GFAT1 and H1299 CRISPR Mix-target cells grown under the indicated
GlcNAc supplementation condition.
After confriming the GFAT protein level was depleted in H1299
CRISPR GFAT1 and H1299 CRISPR Mix-target cells, we next asked if O-
GlcNAc modification level was affected in those two cell lines. Towards this
end, H1299 CRISPR GFAT1 and H1299 CRISPR Mix-target cells were grown
in complete media with or without 10 mM GlcNAc supplementation. Nine
days after the two population were grown in the absence of GlcNAc, I
discovered that the growth rate of the two cell lines dropped, compared to
MIX GFAT1
H1299
WT
MIX
H1299
CRISPR
GFAT1
10 mM
GlcNAc
9 days
- + + - -
H1299
WT
H1299
CRISPR
MIX
H1299
CRISPR
GFAT1
Anti-Actin
Anti-GFAT
A
B
(Anti-O-GlcNAc)
RL2
Anti-Actin
150
100
75
50
37
kDa

76

their GlcNAc-fed counterparts. I then collected the cells under all conditions
and subjected them to western blotting analysis. The data revealed that the O-
GlcNAc modification level in both KO cell lines dropped significally after
grown in media without any GlcNAc after nine days compared to the GlcNAc-
fed population, as well as the wild-type H1299 cells (Figure 4-11B).
4.2.3.3 Isolation and characterization of clonal Knock-out cell lines
To reveal the genomic mutation within GFPT1 and GFPT2 in the
H1299 GFAT KO cells, I performed clonal selection in both H1299 CRISPR
GFAT1 and H1299 CRISPR Mix-target cells to obtain clonal populations in
each line. After the clonal populations were grown to confluency in 10 cm
dishes, I first performed western blotting analysis on all the clonal
populations to confirm their GFAT protein levels (Figure 4-12).

Figure 4-12. Western blotting analysis of H1299 CRISPR GFAT1 and H1299
CRISPR Mix-target clonal populations. Lysates of clonal populations from two H1299
KO GFAT cell lines were probed with anti-GFAT antibody. H1299 CRISPR Mix-target clone
number 2, 4 and 6 still had detectable amount of GFAT left, whereas H1299 CRISPR GFAT1
clone number 4, 7 and 8 and H1299 CRISPR Mix-target clone number 3, 5, 7, 8 and 9 did not
have detectable amount of GFAT.
With several GFAT KO clonal populations at hand, I next performed
DNA extraction on the clonal cell lines using FlexiGene DNA Kit (Qiagen) to
obtain genomic DNA from each clonal population, followed by polymerase
H1299
WT
H1299
CRISPR
MIX
-9
Anti-GFAT
MIX
-3
H1299
WT
GFAT1
-7
GFAT1
-4
MIX
-4
H1299 CRISPR
MIX
-5
MIX
-6
MIX
-7
H1299
WT
MIX
-2
GFAT1
-8
MIX
-8
H1299 CRISPR

77

chain reaction (PCR) to amplify exon 2 of the GFPT1 or GFPT2 gene using
KOD Hot Start Master Mix (Novagen) as well as forward and reverse primers
that had attB sequence attached to their 5’ end (Figure 4-13A).

Figure 4-13. Schematic illustration of two cloning events in genomic
characterization of H1299 GFAT KO cells. A) Genomic DNA from individual clones was
subjected to PCR using two sets of primers, dedicated to either GFPT1 or GFPT2 gene. Two
sets of primers have the same design: attB sequence-attached forward primer and attB
sequence-attached reverse primer were designed to align to intron 1 and intron 2,
respectively, of each GFPT gene. B) PCR products from each gene were then subjected to BP
recombination reaction with pDONR 201 plasmid (Gateway Cloning Technology, Thermo
Fisher), which has a ccdB survival gene for negative selection, to obtain pENTRY plasmids
that encodes individual genotype.
The attB-PCR products allowed me to perform BP recombination
reaction using BP Clonase (Gateway Cloning Technology, Thermo Fisher)
(Figure 4-13B) to separate individual mutation that had occurred as a result of
DNA repair mechanism after CRISPR/ Cas9 cleavage of each allele. The
resulting pENTRY plasmids that contain different exon 2 sequences from
either GFPT1 or GFPT2 were then separated and amplified by transformation
into E. Coli., followed by regular sequencing technique (Laragen) to reveal the
genomic mutation. The sequencing result showed that there were different
mutations associated with different clonal populations (Figure 4-14).
Gene
attB
attB
BP Clonase
ccdB
attP
attP
pDONR
Gene
attL
attL
pENTRY
ccdB
attR
attR
GFPT1 or GFPT2 DNA Exon 1 Exon 2 Intron 1
Intron 2
108 bp
attB
attB
A
B

78

Figure 4-14. Characterization of genomic mutations in each clonal population. A)
Comparison of human GFPT1 partial exon 2 sequence between wild-type and H1299 CRISPR
Mix-target clone number 3, 7 and 9. For the WT sequence, PAM sequence AGG is in bold.
Sequence in blue is the 20-nt targeted by either GFAT1 guide RNA or Mix-target guide RNA.
Letters in red indicated either insertion or mutation events, whereas red dashes indicated
deletion events. B) Comparison of human GFPT2 partial exon 2 sequence between WT and
H1299 CRISPR Mix-target clone number 7. Only one of the gnomic loci that encoded GFAT2
was modified as a result of CRISPR/ Cas9.
Specifically, a total of two different genomic signals were collected from
each clonal population for each gene, suggesting that there were 2 alleles of
each gene reside in the H1299 cells. I also discovered that GFPT1 gene was
modified in all the clonal populations, whereas only one copy of GFPT2 gene
was modified in only one of the clonal populations we harvested: Mix-target
clone number 7 (Figure 4-14).
Next, to further characterize the effect of GFAT knock-out in H1299
cells, H1299 CRISPR Mix-target clone number 9 (here after referred to as
H1299 GFAT KO) was used. In an earlier experiment, we saw that the O-
human GFPT1 exon 2 5’-... ...-3’ GACCCTAATCAAAGGCCTTCAGAGACTGGAGTACAGAGGATATGATTCTGCTG
PAM
Clone Mix-target 3
Clone Mix-target 7
Clone Mix-target 9
5’-... ...-3’ GACCCTAATCAAAGGCCTTCAGAGACTGGAGTA A CAGAGGATATGATTCTGCTG
(+1)
(-6)
5’-... ...-3’ GACCCTAATCAAAGGCCTTCAGAGACTGGAGTA - - - - - - ATATGATTCTGCTG
(-5)
(-6)
5’-... ...-3’ GACCCTAATCAAAGGCCTTCAGAGACTGGAGTA - - GA - - - TATGATTCTGCTG
5’-... ...-3’ GACCCTAATCAAAGGCCTTCAGAGACTGGAGT - - - GA - - - TATGATTCTGCTG
(+1)
5’-... ...-3’ GACCCTAATCAAAGGCCTTCAGAGACTGGAGTA A CAGAGGATATGATTCTGCTG
5’-...
...-3’
GACCCTAATCAAAGGCCTTCAGAGACTGGAGTGGAGAAAATGAAGCTCAGATGTCT
GGGAAGGCGAGAGTATCTAAAACCTTTACCAGAGAAGACTGGAGGGCCTCTTTGG
GAGATGAAACTCTCTGTCACCTGGAGATTCACCAAAGACTCTGAGGCTGAGGGGA
ACAGTATGA
ACCCTCATCAAGGGCCTGCAGCGGCTGGAGTACAGAGGCTACGACTCGGCAG 5’-... ...-3’
PAM
human GFPT2 exon 2
Clone Mix-target 7 (WT)
(1N,+1)
ACCCTCATCAAGGGCCTGCAGCGGCTGGAGTACAGAGGCTACGACTCGGCAG 5’-... ...-3’
ACCCTCATCAAGGGCCTGCAGCGGCTGGAGTT TCAGAGGCTACGACTCGGCAG 5’-... ...-3’
(2N, +137)
A
B

79

GlcNAc modification levels decreased dramatically nine days after GlcNAc
deprivation in both H1299 CRISPR GFAT1 and H1299 CRISPR Mix-target
cells (Figure 4-11B). Therefore, we decided to first investigate the correlation
between the concentration of GlcNAc supplementation and the O-GlcNAc
modification level in the H1299 GFAT KO cells. Towards this end, the cells
were incubated with 5, 4, 3, 2 and 1 mM of GlcNAc for 5 days, followed by
western blotting analysis to visualize the resulting O-GlcNAc modification
level. The result showed that the O-GlcNAc modification level correlated to
the GlcNAc concentration in a dose-dependent fashion (Figure 4-15).

Figure 4-15. Western blotting analysis of H1299 GFAT KO cells under different
concentration of GlcNAc supplementation. Lysates of either H1299 WT or H1299
GFAT KO cells grown under indicated concentration of GlcNAc for a course of 5 days were
probed with anti-O-GlcNAc antibody RL2 as well as anti-actin for loading control.
Specifically, we observed a significant decrease of O-GlcNAc level when
changing the concentration of GlcNAc from 4 mM to 3 mM in GFAT KO cells,
5 Days
H1299 GFAT KO
Anti-Actin
GlcNAc (mM)
4 5 0 1 3 2
(Anti-O-GlcNAc)
RL2
150
100
75
50
37
kDa
H1299
WT

80

suggesting there was a cut-off UDP-GlcNAc concentration for OGT to modify
the majority of its substrates within H1299 GFAT KO cells.
During the GlcNAc concentration course experiment, I observed that
the GFAT KO cells exhibiting a different growth rate under different amounts
of GlcNAc supplementation. Therefore, we next sought to investigate the
growth rate of H1299 GFAT KO cells when grown under different
concentration of GlcNAc. Towards this end, H1299 GFAT KO cells were grown
in GlcNAc concentration ranging from 4 to 1 mM of GlcNAc for 48 hours.
Then I plated an even number of H1299 WT cell as well as H1299 GFAT KO
cells into 6 wells plate under GlcNAc concentration ranging from 4 to 1 mM.
After 16 hours of plating, the cell number from each condition was counted
and served as the starting cell number for each condition. Then I counted the
number of cells from each condition every 24 hours for a total of 72 hours.
The result showed that the growth rate of H1299 GFAT KO cells correlated
with the amount of GlcNAc provided (Figure 4-16).

81

Figure 4-16. Growth curve of H1299 GFAT KO cells under different amount of
GlcNAc supplementation. Cell numbers calculated at the indicated time points were
normalized to the cell number at time 0 within each condition, including H1299 WT cells, to
obtain the fold-increase of cell number. Note that at time 0, the H1299 GFAT KO cells had
been incubated with the indicated concentration of GlcNAc for 64 hours. The experiment was
performed in triplicate and the error bars represents s.e.m..
4.2.3.4 Effect of GFAT KO on Tumorigenesis
With a characterized clonal population of GFAT KO cells in hand, we
next thought to evaluate if GFAT knock-out had any effect on tumorigenic
properties of H1299 cells. Firstly, we subjected H1299 GFAT KO cells to a soft
agar colony formation assay. The assay evaluates in vitro ability for cells to
grow in an anchorage-independent manner, which is an important hallmark
of cancer cells, by measuring cell proliferation in a semisolid culture media.
Towards this end, the assay was initiated for H1299 WT cells as well as H1299
GFAT KO cells under different concentration of GlcNAc ranging from 5 mM to
0 mM. Three weeks later, number of colonies under each condition was
counted using light microscope, after which the colonies were stained by
0
5
10
15
20
0 24 48 72
Fold-increase
Cell Number
Time (h)
= H1299 WT
= GFAT KO (4 mM GlcNAc)
= GFAT KO (3 mM GlcNAc)
= GFAT KO (2 mM GlcNAc)
= GFAT KO (1 mM GlcNAc)

82

crystal violet to obtain an image. Surprisingly, H1299 GFAT KO cells didn’t
form any colonies regardless of the amount of GlcNAc that was supplemented
to the cells, whereas H1299 WT cells forms visible colonies (Figure 4-17).

Figure 4-17. Evaluation of anchorage-independent growth ability of H1299 GFAT
KO cells. A) After crystal violet staining, an image of each 6 well plate was taken by
ChemiDoc XRS+ molecular imager (Bio-Rad) through coomassie blue mode. A representative
image from each condition was shown. B) The number of colonies for each condition at the
end of the assay was counted and plotted for comparison. A triplicate experiment was
initiated for each condition and the error bar represents s.e.m..
This data suggested that GFAT knock-out completely disabled the
H1299 cells’ ability to grow in an anchorage-independent manner, despite the
fact that the KO cells had similar level of O-GlcNAc modification (Figure 4-15)
and growth rate in 2D culture compared to WT cells when supplemented with
3 to 4 mM of GlcNAc.
Number of
Colonies
GFAT KO (5 mM GlcNAc)
WT
0
22.5
45.0
67.5
90.0
GFAT KO (4 mM GlcNAc)
GFAT KO (3 mM GlcNAc)
GFAT KO (2 mM GlcNAc)
GFAT KO (1 mM GlcNAc)
5 mM GlcNAc 4 mM GlcNAc
0 mM GlcNAc
WT
3 mM GlcNAc 2 mM GlcNAc 1 mM GlcNAc
B A

83

Next, to determine whether GFAT knock-out affects tumour cell
growth in vivo, we collaborated with Lyssiotis Lab at University of Michigan
to perform xenograft study using H1299 WT and H1299 GFAT KO cells. Nude
mice were injected with same number of either H1299 WT or H1299 GFAT
KO cells, and the tumor growth was monitored for a roughly 6-week period.
During the course of the experiment, H1299 GFAT KO cells showed a delay in
tumor development compared to WT cells. Also, as shown in Figure 4-18,
H1299 GFAT KO cells developed smaller tumors compared to their WT
counterpart at the end of the experiment, judging by their total tumor mass as
well as volume.

Figure 4-18. Evaluation of tumor cell growth of H1299 GFAT KO cells in vivo. A)
Dissected tumors from the nude mice. Five tumors derived from H1299 GFAT KO cells are on
A B

84

the left and 5 tumors derived from H1299 WT cells are from the right. B) Mass and volume of
dissected tumors. Each dot represents the tumor mass or volume from each mice. Error bars
represent standard deviation.
4.3 Discussion and Conclusions
By targeting GFAT at the protein, RNA or DNA level, our data provides
a comprehensive analysis towards the relationship between GFAT activity and
O-GlcNAc modification in various lung and breast cancer cell lines. We
revealed the potential of GFAT to become a cancer target as we showed that
inhibition of GFAT at either protein or DNA level affects cancer cells’ stress
resistance ability or tumorigenic properties, respectively.
To inhibit GFAT at its protein level, I treated A549, H1299 and MCF7
cells with a known GFAT inhibitor, 6-diazo-5-oxo-L-norleucine or DON, and
showed that 100 µM of DON was effective in lowering the O-GlcNAc
modification level in H1299 and A549, whereas the same treatment had a
small effect on the modification level in MCF7 cells. We also showed that
treating the same panel of cells with 10 mM of GlcNAc was able to fully
restore the O-GlcNAc modification level under DON treatment,
demonstrating that DON was mainly on-target in the context of our
experiments. The result also suggests that those three cancer cells were able to
incorporate the sugar into the HBP to support the biosynthesis of UDP-
GlcNAc while DON restricts the flow of metabolites diverging from glycolysis
into HBP. Next, I showed that DON and diamide had an additive effect on cell
proliferation; however, the same panel of cancer cells was sensitized to the
oxidative stress by DON treatment, shown by upregulation of Caspase-3 and/

85

or Caspase-7 activities especially in the case of H1299 and MCF7. We further
showed that GlcNAc supplementation was able to prevent the activation of
Caspase 3 and/ or Caspase-7 under the dual small-molecule challenge,
demonstrating that upregulation of O-GlcNAc level was crucial for the cancer
cells to survive the oxidative stress. Our data aligns well with previous
research where others have showed that global O-GlcNAc modification level
was rapidly increased upon exposure of a variety of cellular stress (Kang et al.,
2009; Wang et al., 2014; Zachara, 2004) and this upregulation has proven to
be largely protective since artificially raising the level of the modification
increases the ability of cells and tissues to survive stress (Liu et al., 2006;
Yang et al., 2006).
To target GFAT at RNA level, we utilized two plasmids that express
short hairpin RNA (shRNA) designed to promote the degradation of mRNA of
either GFAT1 or GFAT2. Our data demonstrates that the shRNA strategy was
neither universal nor effective in lowering the global O-GlcNAc modification
level in cancer cells, proven by the fact that only H1299 cells lowered their O-
GlcNAc modification level in response to GFAT1. In all of the other cell lines
(MCF7, A549 and MDA-MB-231) that I examined, the O-GlcNAc modification
level was essentially the same compared to their control-KD counterpart
despite GFAT protein levels were lowered dramatically through GFAT1
shRNA expression. Expression of GFAT2 shRNA reduced total GFAT protein
level modestly in A549 as well as in MDA-MB-231 cells, whereas it only had a
slight effect on reducing GFAT protein levels in MCF7 and almost no effect in

86

H1299 cells. Therefore, it is fair to say that the amount of GFAT2 that was
expressed in each cell line can be correlated to the amount of reduction in
GFAT protein level caused by the GFAT2 shRNA. Interestingly, we observed
that GFAT1 shRNA was able to reduce the protein level of GFAT2, whereas
GFAT1 protein level was essentially undisturbed by GFAT2 shRNA in MDA-
MB-231 GFAT KD cells. Lastly, we were able to show that in order to maintain
their O-GlcNAc modification level, some cancer cells have the abilities to
adapt to the lose of GFAT protein level by either upregulation of OGT and/ or
downregulation of OGA protein levels, particularly illustrated by the effect of
double-knockdown that we observed in the MDA-MB-231 cells. OGT has an
unique ability to change its effective Km for protein substrates in response to
the local UDP-GlcNAc concentrations(Shen et al., 2012); therefore, it is
reasonable to hypothesize that upregulation of OGT is a direct way to
maintain O-GlcNAc modification level in response to smaller amount of UDP-
GlcNAc as a result of dramatically reduced GFAT protein level. In contrast,
the H1299 cells failed to maintain their O-GlcNAc levels in response to GFAT1
shRNA expression. Taken together, our data suggests that in response to
GFAT inhibition at its mRNA level, different cancer cells exhibit divergent
degree of flexibility in maintaining their global O-GlcNAc modification. And
that flexibility is represented by, but not limited to, the ability to upregulate
OGT protein levels as well as to downregulate OGA protein levels.
Finally, to target GFAT at DNA level, we utilized CRISPR/ Cas9
nucleases to introduce double-stranded breaks and subsequent

87

nonhomologous end joining (NHEJ) DNA damage repair that would
inevitably create an early stop codon in the coding sequence of GFAT1 as well
as GFAT2 genes. Harnessing the promiscuity of Cas9 nucleases towards its
target DNA
(Doench et al., 2014; Fu et al., 2013; Hsu et al., 2013), I designed two guide
RNA sequences and hoping at least one of them would target both GFPT1
gene and GFPT2 gene at the same time. Fortunately, I was able to knock-out
GFAT in H1299 cells, which exclusively express GFAT1 judging by our data
obtained in the H1299 GFAT KD cell lines, using either the GFAT1 guide RNA
or the Mix-target guide RNA. Notably, Mix-target guide RNA was able to
modify both GFPT1 and GFPT2 genes, though we only observed one copy of
the GFPT2 gene was modified in one of the Mix-target clones that we
examined. We attributed this low modification rate of GFPT2 gene to the fact
that it was not being actively transcribed in H1299 cells, hence its chromatin
structure at the genomic loci might not be accessible to Cas9-RNA complex.
Importantly, we were able rescue the phenotype of H1299 GFAT KO cells by
GlcNAc supplementation and subsequently showed that the O-GlcNAcylation
levels and the proliferation rate were positively correlated to the amount of
GlcNAc we provided. Incubating the H1299 GFAT KO cells with at least 4 mM
of GlcNAc enabled the cells to exhibit similar O-GlcNAc modification level
and proliferation rate compared to its wild-type counterpart.

88

Lastly, we were able to reveal tumorigenic defects of the H1299 GFAT
KO cells by testing their colony formation ability in soft agar and in vivo
tumour growth potential in xenograft model. Firstly, GFAT knock-out
completely abolished H1299 cells’ ability to grow in an anchorage-
independent manner, which has been correlated with metastasis potential of
breast and lung tumors as well as murine fibrosarcomas(Cifone and Fidler,
1980; Mori et al., 2009). Despite the fact that H1299 GFAT KO cells could
exhibit similar O-GlcNAcylation level and 2D growth rate as their wild-type
counterpart, GlcNAc supplementation might not be able to achieve sufficient
and consistent supply of UDP-GlcNAc, which is an important substrate for
not just O-GlcNAcylation but also for N- and O-linked glycosylation events.
Numerous lines of evidence have shown that altered cell surface O-linked
glycans and enhancements of N-linked glycans are intimately associated with
cancer development and progression(Cheng et al., 2015; Christiansen et al.,
2014; Pinho and Reis, 2015). One of the factors that result in changes in cell
surface glycosylation patterns is the abundance as well as the availability of
the nucleotide donor sugar for the glycosylation events. Furthermore,
overexpression of mannoside acetyl-glucosaminyltransferase 5 (MGAT5)
gene, which encodes the N-acetylglucosaminyltransferase V (GnT-V), results
in increased β1, 6-N-acetylglucosamine (β1,6GlcNAc)-branched structures in
N-linked glycans on cancer cell surface, which implies a high demand of UDP-
GlcNAc for the malignant transformation. Also, insufficient supply of the UDP
donor sugar might result in a malfunction extracellular matrix (ECM), which

89

is composed of a dynamic and complex array of glycoproteins, collagens and
glycosaminoglycans, which could negatively influence potential of tumor cell
invasion and metastasis(Arnal-Estape and Nguyen, 2015). Therefore, we
speculate that GFAT knock-out in H1299 cells result in a unstable supply of
UDP-GlcNAc, which is under high demand for fueling not only the abnormal
pattern of glycosylation events on the cancer cell surface but also the secreted
glycoproteins. Finally, H1299 GFAT KO cells exhibited a delay in tumor
growth and formed a smaller tumor compared to its wild-type counterpart in
a xenograft model performed by the Lyssiotis Lab at the University of
Michigan. This data seems counter-intuitive to the data that we got in the soft
agar colony formation assay; however, our collaborator was able to detect
GFAT2 expression through quantitative PCR (qPCR) in the dissected tumor
samples formed from H1299 GFAT KO cells (data not shown). We reasoned
that the subcutaneous tumor environment induced the GFAT KO cells to re-
program themselves to induce GFAT2 expression at later stages of the
experiment, reinforcing the importance of GFAT in tumor formation.
In summary, I took three different routes to investigate the
relationship between O-GlcNAc modification level and the activity of GFAT.
Firstly, when GFAT was targeted at protein level, we showed that small
molecule DON sensitized cancer cells to oxidative stress by lowering its O-
GlcNAc modification level rapidly. Furthermore, by artificially augmenting
the O-GlcNAcylation levels through GlcNAc supplementation, we were able to
restore the cancer cells’ ability to survive the stress. Secondly, when GFAT was

90

targeted at RNA level, we discovered that cancer cells originated from
different organs, which results in different metabolism as well as genetic
background, possessed different degree of flexibility in maintaining its O-
GlcNAc modification when the GFAT protein level was decreased through
expression of shRNA. Lastly, we successfully utilized CRISPR/ Cas9 system to
generate GFAT knock-out H1299 cells, resulting in a unique cell line which
exhibited normal 2D culture growth rate but failed to proliferate in 3D culture
condition. Taken together, our data shed lights on the potential of GFAT to
become a cancer target, particularly for the tumor types that rely heavily on a
consistent and up-regulated supply of UDP-GlcNAc to combat harsh
environments, as well as gaining momentum and ability to disseminate into
surrounding healthy tissues.
4.4 Materials and Methods
4.4.1 Cell Culture
AmphoPack-293 (Clontech), MDA-MB-231 and HEK293 cells were
cultured in Dulbecco’s modified Eagle’s medium (DMEM, Corning Cellgro).
H1299 (ATCC) and H1975 (ATCC) were cultured in RPMI 1640 (Corning
Cellgro). A549 cells were cultured in F-12K Nutrient Mixture (Kaighn's Mod.,
Corning Cellgro). MCF7 cells were grown in Minimum Essential Medium
(MEM, Corning Cellgro) supplemented with 0.01 mg/ml recombinant insulin
(Gibco® Insulin-Transferrin-Selenium, Thermo Fisher). All media was
supplemented with 10 % fetal bovine serum (FBS, Atlanta Biologicals,

91

Norcross, GA). All cells were grown at 37℃ and 5% CO 2 according to standard
procedures.
4.4.2 DON, Diamide and GlcNAc treatment
To investigate the effect of DON, Diamide and GlcNAc on cancer cell’s
O-GlcNAc modification level, half a million cells were plated for each
condition in a 10 cm plate 7 hours prior to the start of the treatment. At t= 0
h, the existing media was replaced with fresh media containing the
corresponding treatment. The second treatment was applied by changing the
old media into new with the corresponding drugs, where it was applicable.
4.4.3 Characterization of GFAT KD cell lines
To investigate the effect of GFAT1 and GFAT2 shRNA in cancer cells,
half a million cells were plated for each cell line in a 10 cm plate. The cells
were worked up after 48 hours.
4.4.4 Western blotting analysis of H1299 CRISPR GFAT1, Mix-target cells
and GFAT KO cells
To evaluate the effect of GlcNAc supplementation on the O-GlcNAc
modification level in H1299 CRISPR GFAT1 as well as H1299 CRISPR Mix-
target cells, H1299 WT grown in complete media as well as the two CRISPR
cell lines grown in complete media with or without 10 mM GlcNAc
supplementation were split every 2 days into a new 10 cm dish with two to ten
ratio. The cells under all condition were worked up after 9 days.

92

To investigate the correlation in between GlcNAc concentration and
the O-GlcNAc level of the H1299 GFAT KO cells, a confluent 10 cm plate of
H1299 GFAT KO cells was split 2:10 into 5 plates with complete media
supplemented with 5, 4, 3, 2, 1 mM of GlcNAc. After 2 days, a confluent plate
of H1299 WT cells and the KO cells under each condition was trypsinized and
resuspended into 10 mL of complete media, after which the cell number was
determined by Countess® II Automated Cell Counter (Thermo Fisher). 5 x
10^5 cells from each condition, including H1299 WT control cells, was plated
into a 10 cm dish. The cells were worked up after 3 days followed by western
blotting analysis.
4.4.5 MTS Cell Proliferation Assay
A triplicate experiment was performed for each treatment. At 8 hours
prior to the treatment, 5000 cells were plated into each well in 96 well assay
plate (Costar, black plate, clear bottom with lid). At t = 0 hr, existing media
was replaced with 100 µL of fresh media containing either vehicle or 100 µM
6-Diazo-5-oxo-L-norleucine (DON, BACHEM). At t = 25 hr, 100 µL of fresh
media, containing vehicle or 200 µM DON (2X in water) or 200-400 µM
Diamide (2X in water, Sigma Aldrich) or both 200 µM DON (2X in water) and
200-400 µM Diamide (2X in water), was added into each well without
removing the existing media. At t = 40 hr, existing media was replaced with
100 µL of fresh media immediately before performing the assay according to
manufacturer’s protocol (Promega, CellTiter 96 AQueous Non-Radioactive
Cell Proliferation Assay). The plate was gently swirled for 5 seconds with

93

medium intensity before the absorbance of formazan was measured at 490
nm using Synergy H4 Hybrid reader (BioTek) under normal read speed.
4.4.6 Caspase 3/7 Activity Assay
A quadruplicate experiment was performed for each treatment. At 8
hours prior to the treatment, 5000 cells were plated into each well in 96 well
assay plate (Costar, white plate, clear bottom with lid). At t = 0 hr, existing
media was replaced with 50 uL of fresh media containing either vehicle or
DON. At t = 25 hr, 50 uL of fresh media, containing vehicle or DON (2X in
water) or Diamide (2X in water, Sigma Aldrich) or both DON (2X in water)
and Diamide (2X in water), was added into each well without removing the
existing media. At t = 40 hr, for obtaining the cell count in each treatment, the
media in one of the quadruplicate experiment was removed and transferred
into an eppendorf tube. The well was washed with 50 µL of 1X PBS, which
then was transferred into the same eppendorf tube. Adherent cells were
trypsinized using 50 µL of Trypsin (Corning) and transferred into the same
eppendorf tube. Cell number was calculated using hemocytometer. The
remaining 3 wells of each quadruplicate experiment was dedicated to the
assay (Caspase-Glo 3/7 Assay, Promega) according to manufacturers protocol.
The plate was shook for 2 minutes with fast intensity before luminescence was
measured using Synergy H4 Hybrid reader (BioTek, Gain = 135, integration
time = 5 sec, read height of 1 mm).

94

For the GlcNAc rescue experiment, minor modification was made to
the protocol stated above: At t = 0 hr, existing media was replaced with 50 uL
of fresh media containing either vehicle or DON or DON with 10 mM GlcNAc
(Alfa Aesar). At t = 25 hr, 50 uL of fresh media, containing vehicle or Diamide
(2X in water, Sigma Aldrich) or DON (2X in water, Sigma Aldrich) and
Diamide (2X in water, Sigma Aldrich) or DON (2X in water) and Diamide (2X
in water) and GlcNAc (20 mM), was added into each well without removing
the existing media. The rest of the procedures stayed the same.
4.4.7 Customized GFAT KD plasmids and Construction of CRISPR GFAT
plasmids
pLKO.1-GFAT1-puromycin (Packaging plasmids: pLKO.1-puro, Clone
ID TRCN0000075220) and pLKO.1-GFAT2-neomycin (Packaging plasmids:
pLKO.1-Neo, Clone ID TRCN0000075223) were customized and ordered
from Sigma-Aldrich.
For cloning GFAT1 CRISPR plasmid, a duplex oligo was ordered from
IDT (Integrated DNA Technologies) with 5’- CACC G CTT CAG AGA CTG
GAG TAC AG - 3’ as the top strand and 5’- AAAC CT GTA CTC CAG TCT CTG
AAG C – 3’ as the bottom strand, adding phosphate group to the 5’ end of
both strands. For cloning Mix-target CRISPR plasmid, a duplex oligo was
ordered from IDT with 5’ - CACC G CTT CAG CGA CTG GAG TAC AG - 3’ as
the top strand and 5’- AAAC CT GTA CTC CAG TCG CTG AAG C– 3’ as the
bottom strand, adding phosphate group to the 5’ end of both strands. After

95

BbsI (New England Biolabs, NEB) digestion for one and half hours,
pSpCas9(BB)-2A-Puro parent plasmid was then purified through PCR
purification kit (Qiagen). 20 ng of the purified parent plasmid was then
ligated to 0.4 ng of either GFAT1 guide RNA duplex oligo or Mix-target guide
RNA duplex oligo using ligase and ligase buffer purchased from NEB. One
hour ligation product was transformed directly to One Shot® MAX
Efficiency® DH5α™-T1
R
Competent Cells (Thermo Fisher).
4.4.8 Transfections and Generation of Stable GFAT KD Cell Lines
HEK293 cells were transfected with lentiviral expression vectors
(pLKO.1-GFAT1-puromycin or pLKO.1-GFAT2-neomycin (5 µg each) along
with virus packaging plasmids (RRE 2.5 µg, Rev 2.5 µg and vsvg 2.5 µg) by
using standard calcium phosphate transfection procedures. The growth
medium was then replaced with fresh medium 7 h post-transfection. Virus-
containing medium were collected and filtered with 0.45 µm syringe filters
(cellulose acetate, VWR) 24 h post-transfection. Target cell growth medium
was then replaced with the virus-containing medium supplemented with 8
µg/ mL polybrene (Sigma). Growth medium of the transfected HEK293 cells
was replenished for subsequent infections. Infection of the target cells was
performed twice a day for a total of 2 days. Upon completion of the infection,
the target cells were allowed to recover in normal growth medium for 7 h and
subsequently selected by incubating in medium containing 2 µg/mL
puromycin (Calbiochem) and 500 µg/mL G418 (Calbiochem) until confluent.

96

4.4.9 Transfections and Clonal Selection of H1299 GFAT KO cells
Transient transfection of H1299 with GFAT1 CRISPR plasmid or Mix-
target CRISPR plasmid was achieved by using Lipofectamine 2000 reagent
(Invitrogen) following manufacturer’s protocol. In short, 2.5*10^6 target cells
were plated into a 10 cm dish 24 hours before transfection. Each transfection
mixture contained 20 µg of each plasmid, 60 µL of the Lipofectamine 2000
reagent along side with the Opti-MEM media. 30 hours after transfection, the
cells were moved to a 20 cm dish with fresh media supplemented with 1 µg/
mL of puromycin (Calbiochem) for 2 days, after which the antibiotics was
removed by changing the drug-contained media into drug-free media. The
cells were grown to confluent until further characterization.
To perform clonal population selection, a confluent 10 cm plate of
either H1299 CRISPR GFAT1 or H1299 CRISPR Mix-target cells were
trypsinized and resuspended into 10 mL of complete media. One to 5 µL of the
cell mixture was then plated into a 20 cm plate with 20 mL of complete
media. The media was replaced every three days until colonies of cells were
established throughout the plate. About 10 days after plating, individual
colonies were removed from the 20 cm dish to a 24 well plate by using
PYREX® Cloning Cylinder (Corning) through standard trypsinization
technique. Each clone was then propagated till confluency in a 10 cm dish
before further characterization.

97

4.4.10 Genomic PCR and BP Recombination Reactions
For performing genomic PCR of each individual clones, 1 x 10^6 cells
of each clone were subjected to FlexiGene DNA Kit (Qiagen) to obtain their
genomic DNA. PCR reactions were performed using KOD hot start master mix
(Novagen) following manufacturer’s protocol. Each reaction contained 100 ng
of genomic DNA as well as 1.5 µL of each forward and reverse attB primers at
10 µM specific for either GFPT1 exon 2 or GFPT2 exon 2. The PCR products
were first confirmed by running an agarose gel electrophoresis, followed by
PCR purification step (Qiagen). The GFAT1 attB forward primer has the
following sequence: 5’- GGG GAC AAG TTT GTA CAA AAA AGC AGG CTC
TCC CCC TTG AAA AAG TTC C -3’. The GFAT1 attB reverse primer: 5’- GGG
GAC CAC TTT GTA CAA GAA AGC TGG GTA GCC CAG GCA ACA TAG TGA G
-3’. The GFAT2 attB forward primer: 5’- GGG GAC AAG TTT GTA CAA AAA
AGC AGG CTC TGA GCT TTG AAG TGG GCA G -3’. The GFAT2 attB reverse
primer: 5’- GGG GAC CAC TTT GTA CAA GAA AGC TGG GTC TCA CTG GCT
CAC ACC TAC A -3’.
To perform BP recombination reactions, 15 – 150 ng of PCR products,
60 ng of pDONR 201 plasmid as well as 0.8 µL of BP Clonase (Thermo Fisher)
were introduced into a 4 µL reaction. The reaction was incubated at RT for
one hour, followed by transformation into One Shot® MAX Efficiency®
DH5α™-T1
R
Competent Cells (Thermo Fisher). To sequence the GFPT1 exon
2 sequence within each pENTRY plasmid, pDONR 201 forward primer (5’-
GTA ACA TCA GAG ATT TTG AGA CAC -3’) was used. As for sequencing the

98

GFPT2 exon 2 sequence within each pENTRY plasmid, pDONR 201 reverse
primer (5’-TCG CGT TAA CGC TAG CAT GGA TCT C -3’) was used.
4.4.11 Western Blotting
Cells were washed three times with PBS and then lysed with NP-40
lysis buffer (1% NP-40, 50 mM triethanolamine, 150 mM NaCl, pH 7.4) with
Complete Mini protease inhibitor cocktail (Roche Biosciences) for 15 min and
then centrifuged at 4
0
C for 10 min at 10,000*g. The supernatant (soluble cell
lysate) was collected and the protein concentration was determined by
bicinchoninic acid (BCA) assay (Pierce, Thermo Scientific). Twenty to fourty
µg of lysates were then mixed with 4x SDS loading buffer and separated by
SDS-PAGE using 4-20% Criterion™ Tris-HCl polyacrylamide pre-cast gel
(Biorad) before being transferred to PVDF membrane (Bio-rad) through
standard western blotting protocols. For all RL2 blots, it was critical to load
the amount of lysates at the lower end, say 20 µg, to make the observation of
differences between drug treatments/ cell lines easy.
Western blots were blocked in TBST (10 mM Tris, 150 mM NaCl, 0.1%
Tween-20, pH 8.0) containing either 5% non-fat milk (for anti-GFAT and
anti-Actin) or 5% bovine serum albumin, BSA, (for anti-O-GlcNAc, anti-OGT
and anti-OGA) for 1 h at RT for anti-GFAT, anti-Actin, anti-OGT and anti-
OGA and for overnight at 4
0
C for anti-O-GlcNAc under constant rocking.
Then the membranes were incubated with the primary antibody in its
blocking solution overnight at 4
0
C under constant rocking. The blots were

99

then washed with TBST three times for 30 min and then incubated with
corresponding horseradish peroxidase (HRP)-conjugated secondary
antibodies (Jackson ImmunoResearch) for 1 h in TBST containing 5% non-fat
milk (for anti-GFAT and anti-Actin) or 5% BSA (for anti-O-GlcNAc, anti-OGT
and anti-OGA). After being washed three times with TBST, the blots were
developed by using ECL reagents (Bio-Rad) and imaged by ChemiDoc XRS+
molecular imager (Bio-Rad). The anti-GFAT antibody (IBL-America), anti-
actin antibody (Sigma-Aldrich), anti-O-GlcNAc antibody (Thermo Scientific)
and anti-OGT antibody (Cell Signaling Technology) were all used at 1:1000
dilutions, whereas anti-OGA antibody (Cell Signaling Technology) was used at
1:2000 dilutions.
4.4.12 Soft Agar Colony Formation Assay
A triplicate experiment was performed for each condition. The soft
agar consists of 2 layers of agar with different percentage. For plating the
bottom layer (0.6% agar by weight) for each triplicate, 3 mL of warm 2x
modified RPMI 1640 media, containing the corresponding 2x concentrated
GlcNAc where it was applicable, was mixed with 3 mL of autoclaved warm
1.2% bacto agar (VWR) in water immediately followed by dispensing 1.5 mL of
the media-agar mixture into each well of 6-well plate (BD Falcon) for 3 wells
total. While the bottom layer was solidifying, the cells under investigation
were washed with PBS, then trypsinized and suspended in 1x modified RPMI
1640 media. The cell density was determined by Countess® II Automated Cell
Counter (Thermo Fisher) following manufacturer’s protocol. After the bottom

100

layer solidified, 1.5 mL of top layer (0.4% agar by weight) was plated into each
well on top of the bottom layer immediately after mixing the following
solution: 2 mL of warm autoclaved 1.2% agar in water, 2 mL of 1x modified
RPMI 1640 (with corresponding 1x GlcNAc concentration where it was
applicable) contained 40,000 cells and 2 mL of 2x modified RPMI 1640
media. The plate was incubated at RT for 10 to 15 minutes until the top layer
of agar solidified before putting into incubator set at 37℃ and 5% CO 2.
Starting from the next day of plating, 1 mL of 1x modified RPMI 1640,
contained the concentration of GlcNAc corresponds to the condition of each
well if applicable, was added into each well to prevent the agar from drying.
The media was changed every three days. 3 weeks later, number of colonies in
each well was counted by using light microscope by selecting 7 random fields.
For each field, colonies embedded in the agar at different depths were counted
by adjusting the focus of the microscope. Number of colonies of each well is
represented by the sum of number of colonies of the 7 random fields.
To stain the colonies, the media was removed carefully by p1000
pipette, followed by adding 1 mL of 0.01% crystal violet into each well. The
plate was then incubated at RT with constant rocking for 30 minutes, after
which the staining solution was carefully removed by p1000 pipette. To wash
away the background staining, 2 mL of DI water was added into each well,
followed by constant rocking at RT for 20 minutes. The washing step was
repeated for another 2 times before each well was incubated with 2 mL of DI
water at RT with constant rocking overnight.

101

To take images of the stained colonies, excess water in each well was
removed carefully by p1000 pipette before visualizing the plate using
ChemiDoc XRS+ molecular imager (Bio-Rad) through coomassie blue mode.
2x modified RPMI 1640 contained 2x non-essential amino acid (Gibco,
Thermo Fisher), 2x GlutaMAX™ (Gibco, Thermo Fisher) and 20% FBS
(Atlanta Biologicals, Norcross, GA). 1x modified RPMI 1640 contained 1x non-
essential amino acid, 1x GlutaMAX™ and 10% FBS.
4.4.13 H1299 GFAT KO Growth Curve Experiment
Five x 10^5 cells of either H1299 WT or H1299 GFAT KO cells were
plated into each 10 cm plate for each condition. After 2 days, all cells grown
under each condition were trypsinized and resuspended into 10 mL of
complete media. The cell density from each condition was calculated using
Countess® II Automated Cell Counter (Thermo Fisher). Five x 10^4 cells of
each condition were plated into each well of 6 well in triplicate for a total of 4
time points. 16 hours after the cells had been plated into the 6-well plate, the
cell number in each condition was obtained to serve as the starting cell
number for the following time points. Cell number was obtained every 24
hours after the initial work up for up to 72 hours.

102

4.5 Reference
Ahluwalia, G.S., Grem, J.L., Hao, Z., and Cooney, D.A. (1990). Metabolism
and action of amino acid analog anti-cancer agents. Pharmacol Ther 46, 243–
271.
Arnal-Estape, A., and Nguyen, D.X. (2015). Sweets for a Bitter End: Lung
Cancer Cell-Surface Protein Glycosylation Mediates Metastatic Colonization.
Cancer Discovery 5, 109–111.
Boehmelt, G., Wakeham, A., Elia, A., Sasaki, T., Plyte, S., Potter, J., Yang, Y.,
Tsang, E., Ruland, J., Iscove, N.N., et al. (2000). Decreased UDP-GlcNAc
levels abrogate proliferation control in EMeg32-deficient cells. Embo J 19,
5092–5104.
Caldwell, S.A., Jackson, S.R., Shahriari, K.S., Lynch, T.P., Sethi, G., Walker,
S., Vosseller, K., and Reginato, M.J. (2010). Nutrient sensor O-GlcNAc
transferase regulates breast cancer tumorigenesis through targeting of the
oncogenic transcription factor FoxM1. Oncogene 29, 2831–2842.
Cheng, C., Ru, P., Geng, F., Liu, J., Yoo, J.Y., Wu, X., Cheng, X., Euthine, V.,
Hu, P., Guo, J.Y., et al. (2015). Glucose-Mediated N-glycosylation of SCAP Is
Essential for SREBP-1 Activation and Tumor Growth. Cancer Cell 28, 569–
581.
Christiansen, M.N., Chik, J., Lee, L., Anugraham, M., Abrahams, J.L., and
Packer, N.H. (2014). Cell surface protein glycosylation in cancer. Proteomics
14, 525–546.
Cifone, M.A., and Fidler, I.J. (1980). Correlation of patterns of anchorage-
independent growth with in vivo behavior of cells from a murine
fibrosarcoma. Proc. Natl. Acad. Sci. U.S.A. 77, 1039–1043.
Dang, C.V. (2010). Glutaminolysis: Supplying carbon or nitrogen, or both for
cancer cells? Cc 9, 3884–3886.
Dehennaut, V., Lefebvre, T., Sellier, C., Leroy, Y., Gross, B., Walker, S., Cacan,
R., Michalski, J.C., Vilain, J.P., and Bodart, J.F. (2007). O-Linked N-
Acetylglucosaminyltransferase Inhibition Prevents G2/M Transition in
Xenopus laevis Oocytes. J. Biol. Chem. 282, 12527–12536.
Doench, J.G., Hartenian, E., Graham, D.B., Tothova, Z., Hegde, M., Smith, I.,
Sullender, M., Ebert, B.L., Xavier, R.J., and Root, D.E. (2014). rational design
of highly active sgrNAs for CrIsPr-Cas9–mediated gene inactivation. Nat
Biotechnol 1–8.

103

Ferrer, C.M., Lynch, T.P., Sodi, V.L., Falcone, J.N., Schwab, L.P., Peacock,
D.L., Vocadlo, D.J., Seagroves, T.N., and Reginato, M.J. (2014). O-
GlcNAcylation Regulates Cancer Metabolism and Survival Stress Signalingvia
Regulation of the HIF-1 Pathway. Mol. Cell 1–12.
Fu, Y., Foden, J.A., Khayter, C., Maeder, M.L., Reyon, D., Joung, J.K., and
Sander, J.D. (2013). High-frequency off-target mutagenesis induced by
CRISPR-Cas nucleases in human cells. Nat Biotechnol 31, 822–826.
Gloster, T.M., Zandberg, W.F., Heinonen, J.E., Shen, D.L., Deng, L., and
Vocadlo, D.J. (2011). Hijacking a biosynthetic pathway yields a
glycosyltransferase inhibitor within cells. Nat. Chem. Biol. 7, 174–181.
Guillaumond, F., Leca, J., Olivares, O., Lavaut, M.-N., Vidal, N., Berthezène,
P., Dusetti, N.J., Loncle, C., Calvo, E., Turrini, O., et al. (2013). Strengthened
glycolysis under hypoxia supports tumor symbiosis and hexosamine
biosynthesis in pancreatic adenocarcinoma. Proc. Natl. Acad. Sci. U.S.A. 110,
3919–3924.
Gurel, Z., Zaro, B.W., Pratt, M.R., and Sheibani, N. (2014). Identification of O-
GlcNAc Modification Targets in Mouse Retinal Pericytes: Implication of p53
in Pathogenesis of Diabetic Retinopathy. PLoS ONE 9, e95561.
Hanahan, D., and Weinberg, R.A. (2011). Hallmarks of Cancer: The Next
Generation. Cell 144, 646–674.
Herzog, R., Bender, T.O., Vychytil, A., Bialas, K., Aufricht, C., and Kratochwill,
K. (2014). Dynamic O-Linked N-Acetylglucosamine Modification of Proteins
Affects Stress Responses and Survival of Mesothelial Cells Exposed to
Peritoneal Dialysis Fluids. Journal of the American Society of Nephrology 25,
2778–2788.
Hsu, P.D., Scott, D.A., Weinstein, J.A., Ran, F.A., Konermann, S., Agarwala,
V., Li, Y., Fine, E.J., Wu, X., Shalem, O., et al. (2013). DNA targeting
specificity of RNA-guided Cas9 nucleases. Nat Biotechnol 31, 827–832.
Issad, T. (2013). O-GlcNAcylation: a new cancer hallmark? 1–14.
Jiang, J., Lazarus, M.B., Pasquina, L., Sliz, P., and Walker, S. (2011). a neutral
diphosphate mimic crosslinks the active site of human o-glcnac transferase.
Nat. Chem. Biol. 8, 72–77.
Jones, D.R., Keune, W.-J., Anderson, K.E., Stephens, L.R., Hawkins, P.T., and
Divecha, N. (2014). The hexosamine biosynthesis pathway and O-
GlcNAcylation maintain insulin-stimulated PI3K-PKB phosphorylation and
tumour cell growth after short-term glucose deprivation. Febs J n/a–n/a.

104

Kang, J.G., Park, S.Y., Ji, S., Jang, I., Park, S., Kim, H.S., Kim, S.M., Yook, J.I.,
Park, Y.I., Roth, J., et al. (2009). O-GlcNAc Protein Modification in Cancer
Cells Increases in Response to Glucose Deprivation through Glycogen
Degradation. J. Biol. Chem. 284, 34777–34784.
LaRonde-LeBlanc, N., Resto, M., and Gerratana, B. (2009). Regulation of
active site coupling in glutamine-dependent NAD+ synthetase. Nat Struct Mol
Biol 16, 421–429.
Liu, J., PANG, Y., CHANG, T., BOUNELIS, P., CHATHAM, J., and
MARCHASE, R. (2006). Increased hexosamine biosynthesis and protein O-
GlcNAc levels associated with myocardial protection against calcium paradox
and ischemia. Journal of Molecular and Cellular Cardiology 40, 303–312.
Lynch, T.P., Ferrer, C.M., Jackson, S.R., Shahriari, K.S., Vosseller, K., and
Reginato, M.J. (2012). Critical Role of O-Linked -N-Acetylglucosamine
Transferase in Prostate Cancer Invasion, Angiogenesis, and Metastasis. J.
Biol. Chem. 287, 11070–11081.
McIlwain, D.R., Berger, T., and Mak, T.W. (2013). Caspase Functions in Cell
Death and Disease. Cold Spring Harbor Perspectives in Biology 5, a008656–
a008656.
Mori, S., Chang, J.T., Andrechek, E.R., Matsumura, N., Baba, T., Yao, G., Kim,
J.W., Gatza, M., Murphy, S., and Nevins, J.R. (2009). Anchorage-independent
cell growth signature identifies tumors with metastatic potential. Oncogene
28, 2796–2805.
Ortiz-Meoz, R.F., Jiang, J., Lazarus, M.B., Orman, M., Janetzko, J., Fan, C.,
Duveau, D.Y., Tan, Z.-W., Thomas, C.J., and Walker, S. (2015). A Small
Molecule That Inhibits OGT Activity in Cells. ACS Chem. Biol. 10, 1392–1397.
Pinho, S.S., and Reis, C.A. (2015). Glycosylation in cancer: mechanisms and
clinical implications. Nat. Rev. Cancer 15, 540–555.
Ran, F.A., Hsu, P.D., Wright, J., Agarwala, V., Scott, D.A., and Zhang, F.
(2013). Genome engineering using the CRISPR-Cas9 system. Nat Protoc 8,
2281–2308.
Sander, J.D., and Joung, J.K. (2014). CriSPr-Cas systems for editing,
regulating and targeting genomes. Nat Biotechnol 32, 347–355.
Shen, D.L., Gloster, T.M., Yuzwa, S.A., and Vocadlo, D.J. (2012). Insights into
O-Linked N-Acetylglucosamine ([0-9]O-GlcNAc) Processing and Dynamics
through Kinetic Analysis of O-GlcNAc Transferase and O-GlcNAcase Activity
on Protein Substrates. J. Biol. Chem. 287, 15395–15408.

105

Wang, Z.V., Deng, Y., Gao, N., Pedrozo, Z., Li, D.L., Morales, C.R., Criollo, A.,
Luo, X., Tan, W., Jiang, N., et al. (2014). Spliced X-Box Binding Protein 1
Couples the Unfolded Protein Response to Hexosamine Biosynthetic Pathway.
Cell 156, 1179–1192.
Yang, S., Zou, L.-Y., Bounelis, P., Chaudry, I., Chatham, J.C., and Marchase,
R.B. (2006). GLUCOSAMINE ADMINISTRATION DURING
RESUSCITATION IMPROVES ORGAN FUNCTION AFTER TRAUMA
HEMORRHAGE. Shock 25, 600–607.
Yi, W., Clark, P.M., Mason, D.E., Keenan, M.C., Hill, C., Goddard, W.A.,
Peters, E.C., Driggers, E.M., and Hsieh-Wilson, L.C. (2012).
Phosphofructokinase 1 glycosylation regulates cell growth and metabolism.
Science 337, 975–980.
Ying, H., Kimmelman, A.C., Lyssiotis, C.A., Hua, S., Chu, G.C., Fletcher-
Sananikone, E., Locasale, J.W., Son, J., Zhang, H., Coloff, J.L., et al. (2012).
Oncogenic Kras Maintains Pancreatic Tumors through Regulationof Anabolic
Glucose Metabolism. Cell 149, 656–670.
Zachara, N.E. (2004). Dynamic O-GlcNAc Modification of Nucleocytoplasmic
Proteins in Response to Stress: A SURVIVAL RESPONSE OF MAMMALIAN
CELLS. J. Biol. Chem. 279, 30133–30142.
Zhu, Q., Zhou, L., Yang, Z., Lai, M., Xie, H., Wu, L., Xing, C., Zhang, F., and
Zheng, S. (2011). O-GlcNAcylation plays a role in tumor recurrence of
hepatocellular carcinoma following liver transplantation. Med Oncol 29, 985–
993.
Ziegler, U. (2004). Morphological Features of Cell Death. News in
Physiological Sciences 19, 124–128.

106

Bibliography

Aebi, M. (2013). Biochimica et Biophysica Acta. BBA - Mol. Cell Research
1833, 2430–2437.
Ahluwalia, G.S., Grem, J.L., Hao, Z., and Cooney, D.A. (1990). Metabolism
and action of amino acid analog anti-cancer agents. Pharmacol Ther 46, 243–
271.
Anastasiou, D., Poulogiannis, G., Asara, J.M., Boxer, M.B., Jiang, J.K., Shen,
M., Bellinger, G., Sasaki, A.T., Locasale, J.W., Auld, D.S., et al. (2011).
Inhibition of Pyruvate Kinase M2 by Reactive Oxygen Species Contributes to
Cellular Antioxidant Responses. Science 334, 1278–1283.
Anastasiou, D., Yu, Y., Israelsen, W.J., Jiang, J.-K., Boxer, M.B., Hong, B.S.,
Tempel, W., Dimov, S., Shen, M., Jha, A., et al. (2012). Pyruvate kinase M2
activators promote tetramer formation and suppress tumorigenesis. Nat.
Chem. Biol. 8, 839–847.
Arnal-Estape, A., and Nguyen, D.X. (2015). Sweets for a Bitter End: Lung
Cancer Cell-Surface Protein Glycosylation Mediates Metastatic Colonization.
Cancer Discovery 5, 109–111.
Banaszynski, L.A., and Wandless, T.J. (2006). Conditional Control of Protein
Function. Chemistry & Biology 13, 11–21.
Banaszynski, L.A., Chen, L.-C., Maynard-Smith, L.A., Ooi, A.G.L., and
Wandless, T.J. (2006). A Rapid, Reversible, and Tunable Method to Regulate
Protein Function in Living Cells Using Synthetic Small Molecules. Cell 126,
995–1004.
Banaszynski, L.A., Sellmyer, M.A., Contag, C.H., Wandless, T.J., and Thorne,
S.H. (2008). Chemical control of protein stability and function in living mice.
Nat Med 14, 1123–1127.
Bargonetti, J., and Manfredi, J.J. (2002). Multiple roles of the tumor
suppressor p53. Curr Opin Oncol 14, 86–91.
Bishop, A.C., Shah, K., Liu, Y., Witucki, L., Kung, C., and Shokat, K.M. (1998).
Design of allele-specific inhibitors to probe protein kinase signaling. Curr.
Biol. 8, 257–266.
Bishop, A., Buzko, O., Heyeck-Dumas, S., Jung, I., Kraybill, B., Liu, Y., Shah,
K., Ulrich, S., Witucki, L., Yang, F., et al. (2000). Unnatural ligands for

107

engineered proteins: new tools for chemical genetics. Annu. Rev. Biophys.
Biomol. Struct. 29, 577–606.
Boehmelt, G., Wakeham, A., Elia, A., Sasaki, T., Plyte, S., Potter, J., Yang, Y.,
Tsang, E., Ruland, J., Iscove, N.N., et al. (2000). Decreased UDP-GlcNAc
levels abrogate proliferation control in EMeg32-deficient cells. Embo J 19,
5092–5104.
Bond, M.R., and Hanover, J.A. (2015). A little sugar goes a long way: The cell
biology of O-GlcNAc. J. of Cell Biol 208, 869–880.
Bonger, K.M., Rakhit, R., Payumo, A.Y., Chen, J.K., and Wandless, T.J.
(2014). General Method for Regulating Protein Stability with Light. ACS
Chem. Biol. 9, 111–115.
Bouché, C., Serdy, S., Kahn, C.R., and Goldfine, A.B. (2004). The Cellular Fate
of Glucose and Its Relevance in Type 2 Diabetes. Endocrine Reviews 25, 807–
830.
Brouns, S.J.J., Jore, M.M., Lundgren, M., Westra, E.R., Slijkhuis, R.J.H.,
Snijders, A.P.L., Dickman, M.J., Makarova, K.S., Koonin, E.V., and van der
Oost, J. (2008). Small CRISPR RNAs guide antiviral defense in prokaryotes.
Science 321, 960–964.
Brown, E.J., Beal, P.A., Keith, C.T., Chen, J., Shin, T.B., and Schreiber, S.L.
(1995). Control of p70 s6 kinase by kinase activity of FRAP in vivo. Nature
377, 441–446.
Caldwell, S.A., Jackson, S.R., Shahriari, K.S., Lynch, T.P., Sethi, G., Walker,
S., Vosseller, K., and Reginato, M.J. (2010). Nutrient sensor O-GlcNAc
transferase regulates breast cancer tumorigenesis through targeting of the
oncogenic transcription factor FoxM1. Oncogene 29, 2831–2842.
Carthew, R.W., and Sontheimer, E.J. (2009). Origins and Mechanisms of
miRNAs and siRNAs. Cell 136, 642–655.
Chaneton, B., Hillmann, P., Zheng, L., Martin, A.C.L., Maddocks, O.D.K.,
Chokkathukalam, A., Coyle, J.E., Jankevics, A., Holding, F.P., Vousden, K.H.,
et al. (2012). Serine is a natural ligand and allosteric activator of pyruvate
kinase M2. Nature 1–7.
Chen, B., Gilbert, L.A., Cimini, B.A., Schnitzbauer, J., Zhang, W., Li, G.-W.,
Park, J., Blackburn, E.H., Weissman, J.S., Qi, L.S., et al. (2013). Dynamic
Imaging of Genomic Lociin Living Human Cells by an Optimized CRISPR/Cas
System. Cell 155, 1479–1491.

108

Chen, M., David, C.J., and Manley, J.L. (2012). Concentration-dependent
control of pyruvate kinase M mutually exclusive splicing by hnRNP proteins.
Nat Struct Mol Biol 19, 346–354.
Cheng, C., Ru, P., Geng, F., Liu, J., Yoo, J.Y., Wu, X., Cheng, X., Euthine, V.,
Hu, P., Guo, J.Y., et al. (2015). Glucose-Mediated N-glycosylation of SCAP Is
Essential for SREBP-1 Activation and Tumor Growth. Cancer Cell 28, 569–
581.
Christiansen, M.N., Chik, J., Lee, L., Anugraham, M., Abrahams, J.L., and
Packer, N.H. (2014). Cell surface protein glycosylation in cancer. Proteomics
14, 525–546.
Chu, B.W., Banaszynski, L.A., Chen, L.-C., and Wandless, T.J. (2008). Bio. &
Medi. Chem. Letters 18, 5941–5944.
Cifone, M.A., and Fidler, I.J. (1980). Correlation of patterns of anchorage-
independent growth with in vivo behavior of cells from a murine
fibrosarcoma. Proc. Natl. Acad. Sci. U.S.A. 77, 1039–1043.
Cong, L., Ran, F.A., Cox, D., Lin, S., Barretto, R., Habib, N., Hsu, P.D., Wu, X.,
Jiang, W., Marraffini, L.A., et al. (2013). Multiplex Genome Engineering
Using CRISPR/Cas Systems. Science 339, 819–823.
Corradetti, M.N., and Guan, K.-L. (2006). Upstream of the mammalian target
of rapamycin: do all roads pass through mTOR? Oncogene 25, 6347–6360.
Dang, C.V. (2010). Glutaminolysis: Supplying carbon or nitrogen, or both for
cancer cells? Cc 9, 3884–3886.
Dehennaut, V., Lefebvre, T., Sellier, C., Leroy, Y., Gross, B., Walker, S., Cacan,
R., Michalski, J.C., Vilain, J.P., and Bodart, J.F. (2007). O-Linked N-
Acetylglucosaminyltransferase Inhibition Prevents G2/M Transition in
Xenopus laevis Oocytes. J. Biol. Chem. 282, 12527–12536.
Denzel, M.S., and Antebi, A. (2015). ScienceDirectHexosamine pathway and
(ER) protein quality control. Curr. Opin. in Cell Bio. 33, 14–18.
Doench, J.G., Hartenian, E., Graham, D.B., Tothova, Z., Hegde, M., Smith, I.,
Sullender, M., Ebert, B.L., Xavier, R.J., and Root, D.E. (2014). rational design
of highly active sgrNAs for CrIsPr-Cas9–mediated gene inactivation. Nat
Biotechnol 1–8.
Fantin, V.R., St-Pierre, J., and Leder, P. (2006). Attenuation of LDH-A
expression uncovers a link between glycolysis, mitochondrial physiology, and
tumor maintenance. Cancer Cell 9, 425–434.

109

Ferrer, C.M., Lynch, T.P., Sodi, V.L., Falcone, J.N., Schwab, L.P., Peacock,
D.L., Vocadlo, D.J., Seagroves, T.N., and Reginato, M.J. (2014). O-
GlcNAcylation Regulates Cancer Metabolism and Survival Stress Signalingvia
Regulation of the HIF-1 Pathway. Mol. Cell 1–12.
Fu, Y., Foden, J.A., Khayter, C., Maeder, M.L., Reyon, D., Joung, J.K., and
Sander, J.D. (2013). High-frequency off-target mutagenesis induced by
CRISPR-Cas nucleases in human cells. Nat Biotechnol 31, 822–826.
Furth, P.A., St Onge, L., Böger, H., Gruss, P., Gossen, M., Kistner, A., Bujard,
H., and Hennighausen, L. (1994). Temporal control of gene expression in
transgenic mice by a tetracycline-responsive promoter. Proc. Natl. Acad. Sci.
U.S.A. 91, 9302–9306.
Gaj, T., Gersbach, C.A., and Barbas, C.F., III (2013). ZFN, TALEN, and
CRISPR/Cas-based methods for genome engineering. Trends in
Biotechnology 31, 397–405.
Geda, P., Patury, S., Ma, J., Bharucha, N., Dobry, C.J., Lawson, S.K.,
Gestwicki, J.E., and Kumar, A. (2008). A small molecule-directed approach to
control protein localization and function. Yeast 25, 577–594.
Gloster, T.M., Zandberg, W.F., Heinonen, J.E., Shen, D.L., Deng, L., and
Vocadlo, D.J. (2011). Hijacking a biosynthetic pathway yields a
glycosyltransferase inhibitor within cells. Nat. Chem. Biol. 7, 174–181.
Groves, J.A., Lee, A., Yildirir, G., and Zachara, N.E. (2013). Dynamic O-
GlcNAcylation and its roles in the cellular stress response and homeostasis.
Cell Stress and Chaperones 18, 535–558.
Gu, Y., Mi, W., Ge, Y., Liu, H., Fan, Q., Han, C., Yang, J., Han, F., Lu, X., and
Yu, W. (2010). GlcNAcylation Plays an Essential Role in Breast Cancer
Metastasis. Cancer Research 70, 6344–6351.
Guillaumond, F., Leca, J., Olivares, O., Lavaut, M.-N., Vidal, N., Berthezène,
P., Dusetti, N.J., Loncle, C., Calvo, E., Turrini, O., et al. (2013). Strengthened
glycolysis under hypoxia supports tumor symbiosis and hexosamine
biosynthesis in pancreatic adenocarcinoma. Proc. Natl. Acad. Sci. U.S.A. 110,
3919–3924.
Gupta, G.P., and Massagué, J. (2006). Cancer Metastasis: Building a
Framework. Cell 127, 679–695.
Gurel, Z., Zaro, B.W., Pratt, M.R., and Sheibani, N. (2014). Identification of O-
GlcNAc Modification Targets in Mouse Retinal Pericytes: Implication of p53
in Pathogenesis of Diabetic Retinopathy. PLoS ONE 9, e95561.

110

Hanahan, D., and Weinberg, R.A. (2011). Hallmarks of Cancer: The Next
Generation. Cell 144, 646–674.
Hang, H.C., and Bertozzi, C.R. (2005). The chemistry and biology of mucin-
type O-linked glycosylation. Bio. & Medi. Chem. 13, 5021–5034.
Hardivillé, S., and Hart, G.W. (2014). Nutrient Regulation of Signaling,
Transcription, and Cell Physiology by O-GlcNAcylation. Cell Metabolism 20,
208–213.
Herzog, R., Bender, T.O., Vychytil, A., Bialas, K., Aufricht, C., and Kratochwill,
K. (2014). Dynamic O-Linked N-Acetylglucosamine Modification of Proteins
Affects Stress Responses and Survival of Mesothelial Cells Exposed to
Peritoneal Dialysis Fluids. Journal of the American Society of Nephrology 25,
2778–2788.
Hilton, I.B., D'Ippolito, A.M., Vockley, C.M., Thakore, P.I., Crawford, G.E.,
Reddy, T.E., and Gersbach, C.A. (2015). Epigenome editing by a CRISPR-
Cas9-based acetyltransferase activates genes from promoters and enhancers.
Nat Biotechnol 33, 510–517.
Hinderlich, S., Berger, M., Schwarzkopf, M., Effertz, K., and Reutter, W.
(2000). Molecular cloning and characterization of murine and human N-
acetylglucosamine kinase. Eur. J. Biochem. 267, 3301–3308.
Hitosugi, T., Kang, S., Vander Heiden, M.G., Chung, T.W., Elf, S., Lythgoe, K.,
Dong, S., Lonial, S., Wang, X., Chen, G.Z., et al. (2009). Tyrosine
Phosphorylation Inhibits PKM2 to Promote the Warburg Effect and Tumor
Growth. Science Signaling 2, ra73–ra73.
Holst, J., Vignali, K.M., Burton, A.R., and Vignali, D.A.A. (2006). Rapid
analysis of T-cell selection in vivo using T cell–receptor retrogenic mice.
Nature Methods 3, 191–197.
Hsu, P.D., Scott, D.A., Weinstein, J.A., Ran, F.A., Konermann, S., Agarwala,
V., Li, Y., Fine, E.J., Wu, X., Shalem, O., et al. (2013). DNA targeting
specificity of RNA-guided Cas9 nucleases. Nat Biotechnol 31, 827–832.
Hsu, P.P., and Sabatini, D.M. (2008). Cancer Cell Metabolism: Warburg and
Beyond. Cell 134, 703–707.
Issad, T. (2013). O-GlcNAcylation: a new cancer hallmark? 1–14.
Iwamoto, M., Björklund, T., Lundberg, C., Kirik, D., and Wandless, T.J.
(2010). A General Chemical Method to Regulate Protein Stability in the
Mammalian Central Nervous System. Chemistry & Biology 17, 981–988.

111

Janse, D.M., Crosas, B., Finley, D., and Church, G.M. (2004). Localization to
the Proteasome Is Sufficient for Degradation. J. Biol. Chem. 279, 21415–
21420.
Jiang, J., Lazarus, M.B., Pasquina, L., Sliz, P., and Walker, S. (2011). A neutral
diphosphate mimic crosslinks the active site of human o-glcnac transferase.
Nat. Chem. Biol. 8, 72–77.
Johnsson, N.N., and Varshavsky, A.A. (1994). Split ubiquitin as a sensor of
protein interactions in vivo. Proc. Natl. Acad. Sci. U.S.A. 91, 10340–10344.
Johnston, J.A., Johnson, E.S., Waller, P.R., and Varshavsky, A. (1995).
Methotrexate inhibits proteolysis of dihydrofolate reductase by the N-end rule
pathway. J. Biol. Chem. 270, 8172–8178.
Jones, D.R., Keune, W.-J., Anderson, K.E., Stephens, L.R., Hawkins, P.T., and
Divecha, N. (2014). The hexosamine biosynthesis pathway and O-
GlcNAcylation maintain insulin-stimulated PI3K-PKB phosphorylation and
tumour cell growth after short-term glucose deprivation. Febs J n/a–n/a.
Jones, R.G., and Thompson, C.B. (2009). Tumor suppressors and cell
metabolism: a recipe for cancer growth. Genes & Development 23, 537–548.
Jorgensen, C., Sherman, A., Chen, G.I., Pasculescu, A., Poliakov, A., Hsiung,
M., Larsen, B., Wilkinson, D.G., Linding, R., and Pawson, T. (2009). Cell-
Specific Information Processing in Segregating Populations of Eph Receptor
Ephrin-Expressing Cells. Science 326, 1502–1509.
Junttila, M.R., and Evan, G.I. (2009). p53 — a Jack of all trades but master of
none. Nat. Rev. Cancer 9, 821–829.
Kang, J.G., Park, S.Y., Ji, S., Jang, I., Park, S., Kim, H.S., Kim, S.M., Yook, J.I.,
Park, Y.I., Roth, J., et al. (2009). O-GlcNAc Protein Modification in Cancer
Cells Increases in Response to Glucose Deprivation through Glycogen
Degradation. J. Biol. Chem. 284, 34777–34784.
Kittanakom, S., Chuk, M., Wong, V., Snyder, J., Edmonds, D., Lydakis, A.,
Zhang, Z., Auerbach, D., and Stagljar, I. (2009). Analysis of Membrane
Protein Complexes Using the Split-Ubiquitin Membrane Yeast Two-Hybrid
System. In Methods in Molecular Biology, (Totowa, NJ: Humana Press), pp.
247–271.
Kohler, J.J., and Bertozzi, C.R. (2003). Regulating Cell Surface Glycosylation
by Small Molecule Control of Enzyme Localization. Chemistry & Biology 10,
1303–1311.

112

Kos, C.H. (2004). Cre/loxP system for generating tissue-specific knockout
mouse models. Nutr Rev 62, 243–246.
Kwei, K.A., Baker, J.B., and Pelham, R.J. (2012). Modulators of Sensitivity
and Resistance to Inhibition of PI3K Identified in a Pharmacogenomic Screen
of the NCI-60 Human Tumor Cell Line Collection. PLoS ONE 7, e46518.
LaRonde-LeBlanc, N., Resto, M., and Gerratana, B. (2009). Regulation of
active site coupling in glutamine-dependent NAD+ synthetase. Nat Struct Mol
Biol 16, 421–429.
Lau, H.D., Yaegashi, J., Zaro, B.W., and Pratt, M.R. (2010). Precise Control of
Protein Concentration in Living Cells. Angew. Chem. Int. Ed. 49, 8458–8461.
Lau, K.S., and Dennis, J.W. (2008). N-Glycans in cancer progression.
Glycobiology 18, 750–760.
Lewis, B.A., and Hanover, J.A. (2014). O-GlcNAc and the Epigenetic
Regulation of Gene Expression. J. Biol. Chem. 289, 34440–34448.
Liu, J., Farmer, J.D., Lane, W.S., Friedman, J., Weissman, I., and Schreiber,
S.L. (1991). Calcineurin is a common target of cyclophilin-cyclosporin A and
FKBP-FK506 complexes. Cell 66, 807–815.
Liu, J., PANG, Y., CHANG, T., BOUNELIS, P., CHATHAM, J., and
MARCHASE, R. (2006). Increased hexosamine biosynthesis and protein O-
GlcNAc levels associated with myocardial protection against calcium paradox
and ischemia. Journal of Molecular and Cellular Cardiology 40, 303–312.
Lynch, T.P., Ferrer, C.M., Jackson, S.R., Shahriari, K.S., Vosseller, K., and
Reginato, M.J. (2012). Critical Role of O-Linked -N-Acetylglucosamine
Transferase in Prostate Cancer Invasion, Angiogenesis, and Metastasis. J.
Biol. Chem. 287, 11070–11081.
Ma, J., and Hart, G.W. (2014). O-GlcNAc profiling: from proteins to
proteomes. Clin Proteomics 11, 8.
Ma, Z., Vocadlo, D.J., and Vosseller, K. (2013). Hyper-O-GlcNAcylation Is
Anti-apoptotic and Maintains Constitutive NF- B Activity in Pancreatic
Cancer Cells. J. Biol. Chem. 288, 15121–15130.
Ma, Z., and Vosseller, K. (2013). O-GlcNAc in cancer biology. Amino Acids 45,
719–733.
McIlwain, D.R., Berger, T., and Mak, T.W. (2013). Caspase Functions in Cell
Death and Disease. Cold Spring Harbor Perspectives in Biology 5, a008656–
a008656.

113

Mori, S., Chang, J.T., Andrechek, E.R., Matsumura, N., Baba, T., Yao, G., Kim,
J.W., Gatza, M., Murphy, S., and Nevins, J.R. (2009). Anchorage-independent
cell growth signature identifies tumors with metastatic potential. Oncogene
28, 2796–2805.
Möckli, N., Deplazes, A., Hassa, P., Zhang, Z., Peter, M., Hottiger, M., Stagljar,
I., and Auerbach, D. (2007). Yeast split-ubiquitin-based cytosolic screening
system to detect interactions between transcriptionally active proteins.
Biotech. 42, 725–730.
Musgrove, E.A., Caldon, C.E., Barraclough, J., Stone, A., and Sutherland, R.L.
(2011). Cyclin D as a therapeutic target in cancer. Nat. Rev. Cancer 11, 558–
572.
Olivier-Van Stichelen, S., Dehennaut, V., Buzy, A., Zachayus, J.L., Guinez, C.,
Mir, A.M., Yazidi-Belkoura, El, I., Copin, M.C., Boureme, D., Loyaux, D., et al.
(2014). O-GlcNAcylation stabilizes -catenin through direct competition with
phosphorylation at threonine 41. The FASEB Journal 28, 3325–3338.
Ortiz-Meoz, R.F., Jiang, J., Lazarus, M.B., Orman, M., Janetzko, J., Fan, C.,
Duveau, D.Y., Tan, Z.-W., Thomas, C.J., and Walker, S. (2015). A Small
Molecule That Inhibits OGT Activity in Cells. ACS Chem. Biol. 10, 1392–1397.
Perez-Pinera, P., Kocak, D.D., Vockley, C.M., Adler, A.F., Kabadi, A.M.,
Polstein, L.R., Thakore, P.I., Glass, K.A., Ousterout, D.G., Leong, K.W., et al.
(2013). RNA-guided gene activation by CRISPR-Cas9–based transcription
factors. Nature Methods 10, 973–976.
Pinho, S.S., and Reis, C.A. (2015). Glycosylation in cancer: mechanisms and
clinical implications. Nat. Rev. Cancer 15, 540–555.
Porteus, M.H. (2003). Chimeric Nucleases Stimulate Gene Targeting in
Human Cells. Science 300, 763–763.
Pratt, M.R., Schwartz, E.C., and Muir, T.W. (2007). Small-molecule-mediated
rescue of protein function by an inducible proteolytic shunt. Proc. Natl. Acad.
Sci. U.S.A. 104, 11209–11214.
Pruschy, M.N.M., Spencer, D.M.D., Kapoor, T.M.T., Miyake, H.H., Crabtree,
G.R.G., and Schreiber, S.L.S. (1994). Mechanistic studies of a signaling
pathway activated by the organic dimerizer FK1012. Chemistry & Biology 1,
163–172.
Ran, F.A., Hsu, P.D., Wright, J., Agarwala, V., Scott, D.A., and Zhang, F.
(2013). Genome engineering using the CRISPR-Cas9 system. Nat Protoc 8,
2281–2308.

114

Renicke, C., Schuster, D., Usherenko, S., Essen, L.-O., and Taxis, C. (2013). A
LOV2 Domain-Based Optogenetic Tool to Control Protein Degradationand
Cellular Function. Chemistry & Biology 20, 619–626.
Ryding, A.D., Sharp, M.G., and Mullins, J.J. (2001). Conditional transgenic
technologies. Journal of Endocrinology.
Saam, J.R., and Gordon, J.I. (1999). Inducible gene knockouts in the small
intestinal and colonic epithelium. J. Biol. Chem. 274, 38071–38082.
Sakamoto, K.M., Kim, K.B., Kumagai, A., Mercurio, F., Crews, C.M., and
Deshaies, R.J. (2001). Protacs: chimeric molecules that target proteins to the
Skp1-Cullin-F box complex for ubiquitination and degradation. Proc. Natl.
Acad. Sci. U.S.A. 98, 8554–8559.
Sander, J.D., and Joung, J.K. (2014). CriSPr-Cas systems for editing,
regulating and targeting genomes. Nat Biotechnol 32, 347–355.
Schleicher, E.D., and Weigert, C. (2000). Role of the hexosamine biosynthetic
pathway in diabetic nephropathy. Kidney International 58, S13–S18.
Shafi, R., Lyer, S., Ellies, L.G., O'Donnell, N., Marek, K.W., Chui, D., Hart,
G.W., and Marth, J.D. (2000). The O-GlcNAc transferase gene resides on the
X chromosome and is essential for embryonic stem cell viability and mouse
ontogeny. Proc. Natl. Acad. Sci. U.S.A. 97, 5735–5739.
Shen, D.L., Gloster, T.M., Yuzwa, S.A., and Vocadlo, D.J. (2012). Insights into
O-Linked N-Acetylglucosamine ([0-9]O-GlcNAc) Processing and Dynamics
through Kinetic Analysis of O-GlcNAc Transferase and O-GlcNAcase Activity
on Protein Substrates. J. Biol. Chem. 287, 15395–15408.
Shogren-Knaak, M.A., Alaimo, P.J., and Shokat, K.M. (2001). Recent
advances in chemical approaches to the study of biological systems. Annu.
Rev. Cell Dev. Biol. 17, 405–433.
Stagljar, I., Korostensky, C., Johnsson, N., and Heesen, te, S. (1998). A genetic
system based on split-ubiquitin for the analysis of interactions between
membrane proteins in vivo. Proc. Natl. Acad. Sci. U.S.A. 95, 5187–5192.
Stankunas, K., Bayle, J.H., Gestwicki, J.E., and Lin, Y.M. (2003). Conditional
protein alleles using knockin mice and a chemical inducer of dimerization.
Mol. Cell.
Stankunas, K., Bayle, J.H., Havranek, J.J., Wandless, T.J., Baker, D.,
Crabtree, G.R., and Gestwicki, J.E. (2007). Rescue of Degradation-Prone
Mutants of the FK506-Rapamycin Binding (FRB) Protein with Chemical
Ligands. ChemBioChem 8, 1162–1169.

115

Stichelen, S.O.-V., Drougat, L., Dehennaut, V., Yazidi-Belkoura, El, I., Guinez,
C., Mir, A.M., Michalski, J.C., Vercoutter-Edouart, A.-S., and Lefebvre, T.
(2012). Serum-stimulated cell cycle entry promotes ncOGT synthesis required
for cyclin D expression. Oncogenesis 1, e36–e36.
Szymczak, A.L., Workman, C.J., Wang, Y., Vignali, K.M., Dilioglou, S., Vanin,
E.F., and Vignali, D.A.A. (2004). Correction of multi-gene deficiency in vivo
using a single “self-cleaving” 2A peptide–based retroviral vector. Nat
Biotechnol 22, 589–594.
Urnov, F.D., Rebar, E.J., Holmes, M.C., Zhang, H.S., and Gregory, P.D.
(2010). Genome editing with engineeredzinc finger nucleases. Nature
Publishing Group 11, 636–646.
Vander Heiden, M.G., Cantley, L.C., and Thompson, C.B. (2009).
Understanding the Warburg Effect: The Metabolic Requirements of Cell
Proliferation. Science 324, 1029–1033.
Varshavsky, A. (1992). The N-End Rule. Cell 69, 725–735.
Wang, I.-C., Chen, Y.J., Hughes, D., Petrovic, V., Major, M.L., Park, H.J., Tan,
Y., Ackerson, T., and Costa, R.H. (2005). Forkhead Box M1 Regulates the
Transcriptional Network of Genes Essential for Mitotic Progression and
Genes Encoding the SCF (Skp2-Cks1) Ubiquitin Ligase. Mol. and Cell. Biol.
25, 10875–10894.
Wang, P., Lazarus, B.D., and Forsythe, M.E. (2012). O-GlcNAc cycling
mutants modulate proteotoxicity in Caenorhabditis elegans models of human
neurodegenerative diseases.
Wang, Z.V., Deng, Y., Gao, N., Pedrozo, Z., Li, D.L., Morales, C.R., Criollo, A.,
Luo, X., Tan, W., Jiang, N., et al. (2014). Spliced X-Box Binding Protein 1
Couples the Unfolded Protein Response to Hexosamine Biosynthetic Pathway.
Cell 156, 1179–1192.
Wolfe, S.A., Nekludova, L., and Pabo, C.O. (2000). DNA recognition by
Cys2His2 zinc finger proteins. Annu. Rev. Biophys. Biomol. Struct. 29, 183–
212.
Wood, A.J., Lo, T.W., Zeitler, B., Pickle, C.S., Ralston, E.J., Lee, A.H., Amora,
R., Miller, J.C., Leung, E., Meng, X., et al. (2011). Targeted Genome Editing
Across Species Using ZFNs and TALENs. Science 333, 307–307.
Wunderlich, F.T., Wildner, H., Rajewsky, K., and Edenhofer, F. (2001). New
variants of inducible Cre recombinase: a novel mutant of Cre-PR fusion
protein exhibits enhanced sensitivity and an expanded range of inducibility.
Nucleic Acids Research 29, E47.

116

Yan, A., and Lennarz, W.J. (2005). Unraveling the Mechanism of Protein N-
Glycosylation. J. Biol. Chem. 280, 3121–3124.
Yang, S., Zou, L.-Y., Bounelis, P., Chaudry, I., Chatham, J.C., and Marchase,
R.B. (2006). GLUCOSAMINE ADMINISTRATION DURING
RESUSCITATION IMPROVES ORGAN FUNCTION AFTER TRAUMA
HEMORRHAGE. Shock 25, 600–607.
Yi, W., Clark, P.M., Mason, D.E., Keenan, M.C., Hill, C., Goddard, W.A.,
Peters, E.C., Driggers, E.M., and Hsieh-Wilson, L.C. (2012).
Phosphofructokinase 1 glycosylation regulates cell growth and metabolism.
Science 337, 975–980.
Ying, H., Kimmelman, A.C., Lyssiotis, C.A., Hua, S., Chu, G.C., Fletcher-
Sananikone, E., Locasale, J.W., Son, J., Zhang, H., Coloff, J.L., et al. (2012).
Oncogenic Kras Maintains Pancreatic Tumors through Regulationof Anabolic
Glucose Metabolism. Cell 149, 656–670.
Zachara, N.E. (2004). Dynamic O-GlcNAc Modification of Nucleocytoplasmic
Proteins in Response to Stress: A SURVIVAL RESPONSE OF MAMMALIAN
CELLS. J. Biol. Chem. 279, 30133–30142.
Zachara, N.E., and Hart, G.W. (2002). The Emerging Significance of O-
GlcNAc in Cellular Regulation. Chem. Rev. 102, 431–438.
Zhang, Q., Liu, X., Gao, W., Li, P., Hou, J., Li, J., and Wong, J. (2014).
Differential Regulation of the Ten-Eleven Translocation (TET) Family of
Dioxygenases by O-Linked β- N-Acetylglucosamine Transferase (OGT). J.
Biol. Chem. 289, 5986–5996.
Zhu, Q., Zhou, L., Yang, Z., Lai, M., Xie, H., Wu, L., Xing, C., Zhang, F., and
Zheng, S. (2011). O-GlcNAcylation plays a role in tumor recurrence of
hepatocellular carcinoma following liver transplantation. Med Oncol 29, 985–
993.
Ziegler, U. (2004). Morphological Features of Cell Death. News in
Physiological Sciences 19, 124–128.

Abstract (if available)

Abstract In Pratt lab, my research focus can be divided into two main topics. During the first two years of graduate research, I developed a dual small-molecule rheostat that allows scientists to fine-tune the concentration of their protein of interests in living mammalian cells. One of the most successful strategies for controlling protein concentrations in living cells relies on protein destabilization domains (DD). Under normal conditions, a DD will be rapidly degraded by the proteasome. However, the same DD can be stabilized or “shielded” in a stoichiometric complex with a small molecule, enabling dose-dependent control of its concentration. This process has been exploited by several labs to post-translationally control the expression levels of proteins in vitro as well as in vivo, although the previous technologies resulted in permanent fusion of the protein of interest to the DD, which can affect biological activity and complicate results. Right before I got recruited, the Pratt lab reported a complementary strategy, termed traceless shielding (TShld), in which the protein of interest is released in its native form. I was involved in the development of an optimized protein concentration control system, TTShld, which retains the traceless features of TShld but utilizes two tiers of small molecule control to set protein concentrations in living cells. These experiments provide the first protein concentration control system that results in both a wide range of protein concentrations and proteins free from engineered fusion constructs. The TTShld system has a greatly improved dynamic range compared to our previously reported system, and the traceless feature is attractive for elucidation of the consequences of protein concentration in cell biology. ❧ During the last four years of my PhD, I switched my focus onto studying how the activity of glutamine fructose-6-phosphate (GFAT) influences the global level of O-GlcNAc modification in cancer cells. Nutrient-sensitive O-GlcNAc modification is a type of post-translational modification that occurs on intracellular proteins. The modification is not only important for numerous aspects of cell biology but also misregulated in various disease-relevant signaling pathways, including neurodegeneration diseases as well as cancer. The modification is up-regulated in virtually all cancer diagnosed so far and the hyper-O-GlcNAcylation has been shown to be indispensable for a variety aspects of tumorigenesis. Governing the rate-limiting step of hexosamine biosynthetic pathway (HBP), GFAT controls the amount of glucose-derived carbon that enters HBP, the end product of which is the high-energy donor nucleotide UDP-GlcNAc that can be utilized by O-GlcNAc transferase (OGT) to catalyze O-GlcNAc addition onto modified substrates. By decreasing GFAT activity using a small molecule inhibitor, we demonstrated GFAT as a handle to lower the hyper-O-GlcNAcylation inside cancer cells, in turn weakened cancer cells to survive oxidative stress. Furthermore, through lowering GFAT abundance by targeting its mRNA, we discovered cancer cells originated from different organs have different capacity to maintain their hyper-O-GlcNAcylation in response to the decreased GFAT level. Lastly, by silencing GFAT through modifying its DNA in H1299 non-small cell lung cancer cell line, we discovered that GFAT is indispensable for the cancer cells to maintain their hyper-O-GlcNAcylation as well as viability, only if we didn’t provide them with GlcNAc, which supports O-GlcNAcylation by entering HBP. Moreover, losing GFAT expression not only took away anchorage-independent growth ability away from the H1299 cells, but also inhibited their in vivo tumor growth rate in xenograft tumor model. Taken together, our data shed lights on a therapeutic window towards cancer, where lowering hyper-O-GlcNAcylation through inhibition of GFAT could be a way to slow down tumor progression by targeting cancer that are not flexible in maintaining their hyper-O-GlcNAcylation.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Optimization of chemical reporters of O-GlcNAc for improved specificity and metabolic mapping

PDF

Direct characterization of functional consequences of O-GlcNAc through protein semi-synthesis

PDF

Development of metabolic chemical reporters for the investigation of protein glycosylation

PDF

Investigating the role of O-GlcNAcylation in α-synuclein aggregation

PDF

O-GlcNAc: development of new inhibitors and investigations into fibroblast contraction

PDF

Novel synthesis of β-glycosides for SPPS of GLCNAC glycoproteins and study of their site-specific biochemical and biophysical consequences

PDF

Synthesis, characterization and application of chemical tools for investigating the role of O-GlcNAc modification in the development and survival of mammalian cells

PDF

Development of bioorthogonal probes to reveal proteome modifications

PDF

Using chemistry to reveal the consequences of post translational modifications in cancer

PDF

Developing and exploiting small molecules to study O-GlcNAc modification

PDF

Chemical dissection of monosaccharide metabolic chemical reporter selectivity

PDF

Uncovering the protective role of protein glycosylation in Parkinson's disease utilizing protein semi-synthesis

PDF

Understanding the roles of posttranslational modifications in aggregation using synthetic proteins

PDF

Site-specific effects of ubiquitin and ubiquitin-like modifier proteins on α-synuclein aggregation

PDF

Improving antitumor efficacy of chimeric antigen receptor-engineered immune cell therapy with synthetic biology and combination therapy approaches

PDF

Examination of the effect of oleandrin on head and neck cancer cells

PDF

Medicinal chemistry and chemical biology approaches to improve the selectivity and potency of kinase inhibitors

PDF

Development of selective covalent probes to identify and modulate protein targets

PDF

Synthetic studies of chemical probes for i) DNA, ii) RNA polymerases and iii) tropomyosin receptor kinase

Asset Metadata

Creator Lin, Yu Hsuan (author)

Core Title Using chemical biology approaches to investigate the consequences of protein concentration and activity in cancer cells

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Chemistry

Publication Date 08/01/2016

Defense Date 05/09/2016

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag cancer metabolism,CRISPR,destabilization domains,GFAT,GFP,hexosamine biosynthetic pathway,OAI-PMH Harvest,O-GlcNAc modification,protein concentration,split ubiquitin,tumorigenesis

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Pratt, Matthew R. (committee chair), Wang, Pin (committee member), Zhang, Chao (committee member)

Creator Email chestnut0600@gmail.com,linyuhsu@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c40-289686

Unique identifier UC11280582

Identifier etd-LinYuHsuan-4696.pdf (filename),usctheses-c40-289686 (legacy record id)

Legacy Identifier etd-LinYuHsuan-4696.pdf

Dmrecord 289686

Document Type Dissertation

Format application/pdf (imt)

Rights Lin, Yu Hsuan

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA