University of Southern California Dissertations and Theses

Do the ZFX and ZFY transcription factors have redundant or unique functions?

(USC Thesis Other)

Do the ZFX and ZFY transcription factors have redundant or unique functions?

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

Do the ZFX and ZFY Transcription Factors Have Redundant or Unique Functions?

by

Yao Liu
Mentor: Dr. Peggy J. Farnham

A Thesis Presented to the
FACULTY OF THE USC KECK SCHOOL OF MEDICINE
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
MASTER OF SCIENCE
(BIOCHEMISTRY AND MOLECULAR MEDICINE)

December 2022

Copyright 2022 Yao Liu
ii

Acknowledgment

Several days ago, I recognized a new student who will come to Keck School of Medicine from China.
When I talked with her, I asked her a question, “what do you feel now before you take an airplane leaving
from your hometown?” She said, “Fear, worry, nervousness and curiosity.” I asked her because I want to
remind what I felt when I was at the similar moment.

As an international student, I also tried to know every information here as much as possible and I
always thought my preparation was not enough and always imagined the possible awkward situation. My
English was poor, how could I communicate with people and set up my life here; What if I got in trouble
and could not register; What should I say and perform in my lab; Would I disappoint people who believe
me… From time to time, those things popped up in my mind. However, the truth was I am so lucky that I
met many kind persons here.

I remember when I was going to here, I asked Dr. Farnham about renting in the campus and even
where people can get food around the campus. Now, I still “report” everything I encountered to her, as if I
tell my family everything interesting or the trouble in the school after coming back to home. Of course,
most cases are “Hey Dr. Farnham, I am in trouble…” Therefore, not only did I get and learn knowledge,
new horizon and how to be a good scientist like previous students, but also I felt the warmth from this
family and shelter of this home. In China, we treat teachers who give us their entire knowledge and lifelong
inspiration as master (Shifu). The word, master, in Chinese contains the meaning of “teacher” and “parents”.
In my deep heart, I would like to call Dr. Farnham my master.

In this family, Dr. Charles has a good sense of humor and can always bring happiness to us. I learned
the optimistic attitude, which is very valuable for me. Sometimes I remembered Dr. Charles and thought
how he would face up to the thing if he were in my situation. To be honest, Charles is the best one I would
like to chat with lying on the beach. In the lab, he can always give me thorough guidance and share his
profound experience just like my wise old teacher. He is also my master, but I also would like to be good
friend with him.

Emily, Shannon and Katie are my sister in the family. They provided me their support and taught what
they have without reservation. Sometimes, I was like an annoying little brother asking them for help.
“Shannon, where is the reagent? I can’t find it.” “Emily, do you know how to do this?” “Katie, have you
tried this method before?” Well, your annoying brother have another year to bother you now.

I also want to express my appreciation to my thesis committee meeting members, Dr. David Craig, Dr.
Judd Rice and Dr. Suhn Rhie. You gave me many valuable suggestions, expertise and their support. You are
shining examples to me!

So, if I had a chance to talk with me at that time, I would say, “Don’t worry. You are lucky. You’ll meet
iii

the best people there. It will be your memorable memory.”

My friends here also gave me many supports at a variety of perspectives. I tried to add all your name
here, but there are too many names, too many your supports and too many thanks I want to write down.
However, do not hesitate, you, who are reading the acknowledgement are the one I want to embrace and
appreciate here.

Finally, my parents, Feng Liu and Hui Kong, and all my relatives, you are my powerful backing,
making me always have the courage to blaze the way forward through all the difficulties and advance
victoriously.

iv

Table of Contents

ACKNOWLEDGMENT ..................................................................................................................................... ii
LIST OF FIGURES ........................................................................................................................................... v
ABSTRACT ................................................................................................................................................... vii
CHAPTER 1. INTRODUCTION .......................................................................................................................... 1
1.1 Mechanisms of transcriptional regulation. .......................................................................................... 1
1.2 Zinc finger proteins are the largest class of DNA-binding transcription factors. ................................ 1
1.3 Introduction to the ZFX gene family................................................................................................... 3
1.4 Study design ........................................................................................................................................ 6
CHAPTER 2. MATERIALS AND METHODS ....................................................................................................... 8
Cell culture ................................................................................................................................................ 8
Plasmid construction ................................................................................................................................. 8
Overexpression and knockdown of ZFY and ZFX ................................................................................... 9
RNA-seq .................................................................................................................................................... 9
ChIP-seq .................................................................................................................................................. 10
CHAPTER 3. FUNCTIONAL ANALYSIS OF ZFY IN MULTIPLE HUMAN CELL LINES. ..................................... 12
3.1 Abstract ............................................................................................................................................. 12
3.2 Identification of genes regulated by ZFY and ZFX .......................................................................... 16
3.3 ZFY regulates genes in a cell type-specific manner .......................................................................... 21
3.4 Identification of genes directly regulated by ZFY but not ZFX ........................................................ 25
CHAPTER 4. FUNCTIONAL ANALYSIS OF KNOCKDOWN OF ZFY VS. ZFX .................................................... 29
4.1 Abstract ............................................................................................................................................. 29
4.2 Effects of knockdown of ZFY and ZFX in U87 and HepG2 cells .................................................... 29
4.3 Comparison of genes in knockdown vs overexpression of ZFX and ZFY ....................................... 29
CHAPTER 5. SUMMARY AND DISCUSSION ................................................................................................... 35
5.1 Summary of Results .......................................................................................................................... 35
5.2 Future Studies: Identification of ZFY-interacting proteins using TurboID ....................................... 36
REFERENCES ............................................................................................................................................... 39

v

List of Figures

Figure 1.1 A model of a C2H2 zinc finger and a set of consecutive C2H2 zinc fingers ............................. 2
Figure 1.2 Four classes of C2H2 zinc finger proteins classified by the number and arrangement of zinc
fingers ........................................................................................................................................................... 3
Figure 1.3 The ZFX gene family .................................................................................................................. 4
Figure 1.4 ZFX and ZFY gene structure comparison ................................................................................... 5
Figure 1.5 Amino acid alignment of ZFX and ZFY...................................................................................... 5
Figure 1.6 Expression levels of ZFY and ZFX in multiple human tissues ................................................... 6

Figure 2.1 Plasmid constructs ....................................................................................................................... 8
Figure 2.2 RT-qPCR assay to test the effect of overexpression and knockdown experiments ................... 10
Figure 2.3 ChIP qPCR assay to test ChIP-seq enrichment ......................................................................... 11

Figure 3.1 ZFY and ZFX expression level in several human cell lines ...................................................... 13
Figure 3.2 Effects of overexpression of ZFX or ZFY on the transcriptomes of DKO HEK293T, C42B,
LN229, U87, and 22Rv1 cells ..................................................................................................................... 15
Figure 3.3 Identification of genes regulated by ZFY but not ZFX ............................................................. 17
3.3.A Venn plots of comparison of ZFY&ZFX upregulated genes in each cell lines
3.3.B Venn plots of comparison of ZFY&ZFX down-regulated genes in each cell lines
Figure 3.4. GO analysis of genes regulated by ZFY but not ZFX .............................................................. 20
3.4.A Classification of genes commonly regulated by ZFX and ZFY
3.4.B Classification of genes uniquely regulated by ZFY
Figure 3.5 ZFY regulates genes in a cell type-specific manner .................................................................. 22
3.5.A Comparison of ZFY-upregulated genes in 4 different cell lines
3.5.B Comparison of ZFY-downregulated genes in 4 different cell lines
Figure 3.6 Go analysis of ZFY regulated genes in multiple cell lines ........................................................ 24
3.6.C. Gene ontology analysis of genes upregulated by ZFY in the different cell lines
3.6.D Gene ontology analysis of genes downregulated by ZFY in different cell lines
Figure 3.7 Comparison of ZFX vs ZFY genomic binding patterns ............................................................ 26
3.7.A Genomic binding patterns of ZFY and ZFX in 22Rv1 cells
3.7.B Genomic binding patterns of ZFY and ZFX in DKO cells
3.7.C Heatmap showing 22Rv1 ChIP-seq data from endogenous ZFX and ZFY centered on the
genomic location of endogenous ZFY peaks
3.7.D Heatmap showing DKO ChIP-seq data from FLAG-tagged ZFX and ZFY centered on the
genomic location of FLAG-tagged ZFY peaks
3.7.E Tag density plots for ZFX and ZFY in DKO cells
Figure 3.8 Identification of genes uniquely directly regulated by ZFY , but not ZFX ................................. 27
3.8.A Comparison of RNA-seq and ChIP-seq data in 22RV1 cells
3.8.B Comparison of RNA-seq and ChIP-seq data in DKO cells
vi

Figure 4.1. Effects of knockdown of ZFY and ZFX in U87, HepG2.......................................................... 30
Figure 4.2 Comparison of genes regulated by ZFY and ZFX in knockdown U87 and HepG2 cells .......... 31
Figure 4.3 Comparison of genes in knockdown vs overexpression of ZFX and ZFY ................................ 32
Figure 4.4 Comparison of direct regulated genes in knockdown vs overexpression of ZFX in HEK293T,
C42B and U87 cells .................................................................................................................................... 33
Figure 4.5 Tag density plots of new target genes in HEK293T and DKO cells .......................................... 34

Figure 5.1 Using TurboID to identify ZFY-interacting proteins ................................................................. 37
5.1.A. Schematic of Turbo-ZFY construct
5.1.B. RT-qPCR assay to test the function of Turbo-ZFY
Figure 5.2 Schematic of ZFY or ZFX fusion proteins ................................................................................ 38

vii

Abstract

Gene expression is regulated by transcription factors (TFs), Zinc finger proteins are the largest class
of DNA-binding transcription factors. ZFX gene family has 3 family members, ZFX, ZFY and ZNF711.
ZFX and ZFY have very identical structure, and ZFX is related to many kinds of cancer according to
previous study. Although many evidence support the hypothesis that ZFY may have similar function as
ZFX, here are no publications showing an association of ZFY expression with cancer and the function of
ZFY has not been well studied yet. This project aims to identify if ZFX and ZFY have redundant or unique
function. Basically, I performed overexpression and knockdown of ZFY or/and ZFX in multiple human cell
lines, and then analyzed the transcriptome changes by RNA-seq analysis and identified ZFX/ZFY target
genes by ChIP-seq analysis. Here, I reported that, similar to ZFX, ZFY acts as a transcription activator. And
although ZFY and ZFX can regulate a common set of genes, ZFY also has uniquely regulated genes. By
combining RNA-seq and ChIP-seq data, I identified some promoters uniquely upregulated by ZFY that had
ZFY, but not ZFX, bound. Comparing overexpression and knockdown experiments, I also identified a set
of “new” target promoter that is regulated by the increased expression of ZFX. These findings suggest that
ZFY does have some functions different than ZFX. Because the DNA binding domains of ZFY and ZFX
are almost identical, it is likely that differences in protein-protein interactions and the N-terminal region
between ZFY and ZFX contribute to the ability of ZFY to uniquely regulate a set of promoters not regulated
by ZFX. To further analyze the function of ZFY, we have created a TurboID-ZFY construct and are using
mass spectrometry to identify other transcription regulators that interact with ZFY , but not with ZFX. I will
also create the fusion ZFY protein that is changed its NTD into ZFX NTD and then repeat the transactivation
experiments and then using Turbo-ID to identify their interacted proteins.
1

Chapter 1. Introduction

1.1 Mechanisms of transcriptional regulation.
Site-specific DNA-binding transcription factors (TFs) can recognize specific DNA sequences,
commonly known as cis-regulatory elements (CREs), and help to assemble a complex that controls the
transcription process. CREs can be near genes at promoters and far from genes at enhancers or silencers.
[1]

Promoters are required for transcription. The core promoter is defined as the minimal DNA sequence that
directs accurate initiation of transcription; it encompasses the transcription start site (TSS).
[2]
During the
RNA synthesis process, RNA polymeraseⅡand general transcriptional factors bind to core promoters at
common elements such as TATA boxes or CpG islands. TATA boxes are located 25-35 base pairs before the
transcription start site of a gene, whereas CpG islands start ~200bp upstream from the transcription start
site and extend for 300–3000 base pairs downstream. During the RNA synthesis process, RNA polymerase
II and associated general transcription factors assemble on the promoters
[3]
. Both types of core promoters
only produce basal levels of mRNA without the help of other cis elements.
[4]
However, site-specific TFs
binding near promoters or at enhancers can significantly increase the production of mRNA. Alternatively,
some site-specific TFs can act as transcription repressors and can decrease the expression of mRNA. The
site-specific TFs generally work by recruiting, via protein-protein interactions, non-DNA binding co-
activators or co-repressors, which can remodel the chromatin to make it more accessible to the
transcriptional machinery.
[5-6] .
A specific TF can regulate different sets of genes depending on cell types and
thus can control different pathways
[9]
and developmental patterns
[10]
in different cells
[11]
. TFs can work
together at different combinations of regulatory elements, which, along with differential expression of the
TFs, allows different patterns of transcription and gives each cell type its unique characteristics.

1.2 Zinc finger proteins are the largest class of DNA-binding transcription factors.
To recognize a specific region on the genome, DNA-binding TFs use specific protein domains to bind to
short sequence motifs. These DNA binding domains are also used for classifying site-specific DNA-binding
TFs.
[7]
The three largest TF families, comprising over 80% of the sequence-specific TFs, are the C2H2 zinc-
finger family, the homeodomain family, and the helix-loop-helix family.
[8]

Zinc fingers are small protein domains that form a secondary structure stabilized by a zinc ion bound to
four or more Cys and/or His residues of the finger.
[12]
In humans, more than 700 proteins contain a zinc
finger, making this the largest class of transcription factors. There are several categories of zinc finger
proteins which are classified by the pattern of residues which interact with the zinc ion, such as C2H2,
C2C2 and C2HC.
[13]
Among these, the C2H2 zinc finger is the most common type of finger structure. In
C2H2 fingers, the amino acids fold to make a finger with 2 β-sheets in N-terminal region and 1 α-helix
in C-terminal region
[13]
. The C2H2-type ZF comprises up to 30 amino acids with the consensus sequence
CX2-4CX12HX2-8H (X refers to any amino acid) (Figure1.1).
2

Figure1.1 A model of a C2H2 zinc finger (top) and a set of consecutive C2H2 zinc fingers (bottom). Each
circle represents an amino acid. The zinc ion forms 4 bonds with 2 Cysteine (yellow) and 2 Histidine
(blue) amino acids, forming a finger-like structure. Green circles represent the linkers between the zinc
fingers. Orange circles represent two large hydrophobic residues, which are structurally important. The 18
circles without characters can be any amino acid. (Adapted from Mackeh, Rafah et al. 2018.)
[44]

According to the number and arrangement of zinc fingers, C2H2-Zinc finger proteins can largely be
classified into 4 classes: single-fingered zinc finger proteins, triple-fingered zinc finger proteins,
separated-paired-fingered zinc finger proteins and multiple adjacent-fingered zinc finger proteins.
[14,43]

Triple-fingered zinc finger proteins, for example KLF (Kruppel-like factors) family
[15]
, have one cluster
of three consecutive zinc fingers. Separated-paired-fingered zinc finger proteins have one or more pairs of
zinc fingers. This type of zinc finger protein with more than one pair of fingers often has widely separated
pairs; this is a small group of proteins that includes TTK and BNC1. The multiple adjacent-fingered zinc
finger proteins have clusters of four or more ZFs, like ZNF394; up to 30 zinc fingers can be present, such
as in ZNF423. Not all C2H2 zinc finger proteins directly bind to DNA. DNA binding generally requires at
least 3 adjacent fingers, with each finger recognizing 3 or more bases in the major groove of the DNA
helix. In addition, many 3-finger DNA-binding C2H2 ZNFs also have a consensus linker sequence
(TGEKP) between the fingers.

3

Figure1.2 Four classes of C2H2 zinc finger proteins classified by the number and arrangement of zinc
fingers. C2H2 zinc fingers can be categorized into 4 subtypes (single-fingered, triple-fingered, separated-
paired-fingered, and multiple adjacent-fingered ZNFs) based on the number and arrangement of their zinc
fingers. Zinc fingers are shown in pink. (Mackeh, Rafah et al. 2018
[43]
)

C2H2 zinc finger proteins have diverse functions.
[19]
As transcription factors, their dysregulation has been
associated with many diseases, including various kinds of cancers.
[21,22]
Many C2H2 zinc finger proteins are
thought to play a role in the process of neurogenesis and brain development.
[23]
For example some members
of the zeb family, like Zeb1 is a transcriptional repressor that can promote the maturation of neurons.
[24]
The
KLF transcription factor family can control hematopoiesis.
[25]
BCL6 has an important function in B-cell
proliferation and differentiation.
[26]

The mechanisms by which ZNFs regulate transcription is not always clear. However, it is likely that
domains outside of the DNA binding domain are required. Zinc finger proteins can also be divided using
conserved domains other than the finger regions, such as KRAB, SCAN and BTB/POZ domains. These
domains are usually located in the N-terminal region of the protein and function in protein-protein
interaction
[16,17]
The KRAB domain is more widespread than others, found in 45% of the annotated human
C2H2 zinc finger proteins.
[18,19]
The KRAB domain is a repressor domain that recruits histone
methyltransferase complexes that create silenced chromatin. Therefore, a large number of the C2H2
proteins are repressors of transcription.
[20]
However, many ZNFs do not have well characterized repression
or activation domains and how they regulate transcription is not yet understood.
1.3 Introduction to the ZFX gene family
The ZFX family is a small family of human C2H2 zinc finger proteins which belong to the class of multiple-
adjacent zinc finger proteins. The ZFX family has 3 family members, ZFX, ZFY and ZNF711, that are
conserved throughout evolution (Figure.1.3). Human ZFX and ZFY clades are sisters on the same
evolutionary branch and may be paralogous genes. ZFX is located on the X chromosome and contains 13
4

C2H2-type zinc fingers in its DNA binding domain. ZFY is encoded on the Y chromosome and thus there
is only one copy in the diploid human genome.
[36]
The two proteins have 96% overall similarity and 99%
similarity in their zinc finger domains, which suggest they may have a similar function (Figure 1.4).
[37]
. In
contrast, ZNF711 (which is also located on the X chromosome) has less overall similarity compared to ZFX
and ZFY and has amino acid changes that eliminate the structure of the third and seventh finger. The N
terminal domain of ZNF711 also is not as highly related to ZFX as is the N terminal domain of ZFY.

Figure 1.3 The ZFX gene family. Shown is a Treefam (http://www.treefam.org/) alignment for ZFX.

Figure 1.4 ZFX and ZFY gene structure comparison. Dashed lines indicate zinc fingers conserved
between ZFX and ZFY. Zinc fingers are shown in blue. (Ni et al., 2020)

5

Figure 1.5 Amino acid alignment of ZFX and ZFY. The identical amino acids between the two proteins
are marked in purple.

Because ZFX and ZFY have 96% overall similarity and 99% similarity in their zinc finger domains, this
suggests they may have a similar function(Figure 1.2).
[37]
A comparison of the fingers in ZFX and ZFY
shows that the last 9 fingers are properly spaced for DNA-binding.
[27]
However, the conserved linker is only
found between finger 5 and 6. Interestingly, work in the Farnham lab has shown that fingers 11-13 of ZFX
are sufficient for binding to the genome.
[37]
The binding pattern of ZFY is very similar to ZFX in human
22RV1 prostate cancer cells, suggesting that fingers 11-13 are also used for ZFY DNA binding; both
proteins bind primarily to CpG island promoters and the CpG island promoters bound by ZFY are basically
the same as those bound by ZFX.
[37]
Both TFs bind at 240 bp downstream from the TSS in active CpG island
promoters, which is a location different than the observed binding locations of many TFs.
[28-30]

The ZFX family members do not have any of the N-terminal domains associated with other ZNFs.
Specifically, these TFs do not contain the KRAB repression domain found in the majority of ZNFs.
Accordingly, previous work in the Farnham lab has shown that ZFX functions as a transcriptional activator:
the expression level of genes with promoters bound by ZFX is higher than the expression level of genes
with promoters not bound by ZFX and overexpression of ZFX results in the activation of a large set of
genes. As stated above, ZFY is very similar in structure to ZFX. Therefore, ZFY may also be a
transcriptional activator.

Previous work from the Farnham lab showed that loss of function of ZFX may cause significant changes
in the transcriptome: the expression of over 2000 genes was influenced by the knockdown of ZFX in human
embryonic kidney cells, with the top functional categories of the affected genes being related to cell cycle
and tumorigenesis.
[28]
In addition, the level of ZFX expression correlates with aggressiveness and severity
in many cancer types, such as glioblastoma, prostate, lung, and colon cancers.
[31-35]
In contrast, very few
studies have been performed that investigate the role of ZFY in transcription or cell growth control. Also,
6

ZFY levels have not been associated with cancer progression or survival, but many studies have linked high
expression of ZFX to poor patient survival
[39-42]
.

The high similarity between ZFX and ZFY raise the question as to whether they have identical functions.
A comparison of the expression patterns of ZFX and ZFY show that they are both ubiquitously expressed
(with the exception that ZFY is not expressed in female-specific tissues); (Figure 1.6). Although ZFX is on
the X chromosome, it is one of the few genes that escapes X inactivation.
[38]
Therefore, in normal female
tissues, there is expression from two copies of the ZFX gene. Perhaps having an almost identical protein on
the Y chromosome allows equal expression of “ZFX+ZFY” in males and females. In this case, ZFX and
ZFY would have identical functions and be able to substitute for one another. However, there are some
differences in the N termini of the two proteins and this may lead to differences in function. Therefore, my
thesis work has been to compare the function of ZFX and ZFY .

Figure 1.6 Expression levels of ZFY and ZFX in multiple human tissues from GTEx
(https://www.gtexportal.org). Shown are the expression levels of ZFX and ZFY separately in different
human tissues. Pink and blue indicate expression in female and male tissues, respectively.

1.4 Study design
My project aims to determine if ZFY and ZFX have redundant or unique functions. The experiments can
be divided into two parts, overexpression of ZFX or ZFY and knockdown of ZFY and/or ZFX in multiple
human cells. To investigate the transcriptomic effect of ZFX and ZFY, I have transfected multiple human
cell lines with ZFX or ZFY expression plasmids and treated the same cell lines with siRNAs targeting ZFX
7

or ZFY. I have compared the ZFX-regulated genes to the ZFY-regulated genes in each cell type. I have also
done a comparative analysis of the sets of ZFX- and ZFY-regulated genes in each cell line. Finally, using
ChIP-seq analysis, I have identified and compared genes directly regulated by ZFX and ZFY .

8

Chapter 2. Materials and Methods

Cell Culture
Human cell lines HepG2 and U87, LN229, DKO HEK293T were cultured in DEME supplemented with
10% fetal bovine serum plus 1% penicillin and streptomycin. The human cell line C42B was cultured in
RPMI medium supplemented with 10% fetal bovine serum plus 1% penicillin and streptomycin. All cell
lines were obtained from ATCC (https://www.atcc.org/). All cell lines were maintained at 37 ℃ with 5%
CO 2. All cells were demonstrated to be free of mycoplasma.

Plasmid Construction
The ZFY expression construct (Origene #RC216037) was obtained from Origene with the wildtype ZFY
coding region and DYKDDDD FLAG-tag being introduced into the pCMV6-Entry plasmid. The ZFX
expression construct (Origene #RC214045) was obtained from Origene with the wildtype ZFX coding
region and DYKDDDD FLAG-tag being introduced into the pCMV6-Entry plasmid. The resulting
constructs were transformed into CopyCutter™ EPI400™ Chemically Competent E. coli (VWR #75927-
876) and induced to a high number of copies using CopyCutter Induction Solution (VWR # 75927-888)
according to the manufacturer’s protocol. Plasmids were purified using the QIAprep Midiprep/Miniprep kit
(QIAGEN #12143, QIAGEN #12123).

Figure 2.1 Expression plasmids used to generate ZFX-FLAG (Left) and ZFY-FLAG (Right) proteins

9

Overexpression and knockdown of ZFY and ZFX
To study the effects of overexpression of ZFY and ZFX, HEK293T DKO, LN229, C42B, and U87 cells
were seeded into 6-well plates and transfected with ZFY or ZFX expression constructs during exponential
phase growth. Cells were transfected at 60-70% confluency using Lipofectamine 3000 reagent (Invitrogen
#L3000015) following the manufacturer's instructions. Cells were then incubated for 24 hours and
harvested. All transfections were performed in triplicate.

To study the effects of reduced expression of ZFX family members, two sequential transfections of the
siRNAs targeting ZFX and/or ZFY were performed in HepG2 and U87 cells using siRNAs targeting ZFX
(Dharmacon #L-006572-00-0050), ZFY (Dharmacon #L-006573-00-0050) and non-targeting control
siRNAs (Dharmacon #D-001810-10-50). Cells were seeded in 6-well plates for the first transfection. The
second transfection, using the same concentration of siRNAs, was performed 24 hours after the first
transfection. Cells were then incubated for another 24 hours and then harvested. All transfections were
performed in triplicate.

RNA-seq
Total RNA was extracted using TRI Reagent (ZYMO RESEARCH #R2050-1-200) and chloroform
following the manufacturers protocol. RNA integrity was checked using an RNA 6000 Nano kit (Agilent
technologies, Cat. No. 50671511) on a 2100 Bioanalyzer (Agilent technologies, Cat. No. G2939AA).
Before sending the samples for RNA-seq, I verified that the transfection experiments were successful by
analyzing one or more known ZFX target genes, as well as negative control genes (Figure 2.2). I also
verified that the knockdowns were successful by analyzing the levels of ZFX and/or ZFY in cells treated
with control siRNAs or siZFY and/or siZFX. The RNA-seq libraries were prepared and sequenced by
Novogene. Samples were sequenced with 150 bp pair-ended reads. RNA-seq results were aligned to the
human genome GRCh38 using STAR (2.7.0e) and counted using FeatureCounts (Subread/2.0.2).
Differentially expressed genes with absolute fold change >1.5 and adjusted p-value cut-off of 0.05 were
determined using DESeq2 (1.30.1). For gene ontology analysis, I used different p-value cut-off depending
on the number of genes in the different sets.

10

Figure 2.2 RT-qPCR assay to test the effects of overexpression and knockdown experiments. Panels A, B,
and C show RT-qPCR of the relative expression levels of target genes (LRRC41, LONRF2 and WNT7B are
known ZFX target genes) in 4 cell lines after transfection with ZFX or ZFY expression vectors, as compared to
transfection of a control plasmid. The same batch of transfected cells was then used for RNA-seq. In panel D, a
representative siRNA knockdown is shown for HepG2 cells; the expression of ZFY and ZFX expression levels
were tested. Bar plots show the average of the triplicate experiments, with error bars showing the standard error
of the mean (SEM).

ChIP-seq
To identify genomic binding sites of ZFY and ZFX, DKO HEK293T cells were transfected with ZFY or
ZFX expression plasmids and harvested 24 hours after transfection. 12.5 ug of FLAG antibody (Sigma-
Aldrich #F1804-200UG) was used with 100-300 ug chromatin for ChIP assays. Each ZFX or ZFY ChIP-
seq experiment was performed using two biological replicates. Before making the library, I confirmed that
the ChIP experiment has worked by performing qPCR using primers to known binding sites and negative
control regions. LRRC41 primer targets the promoter region of LRRC41 that was bound by ZFX and should
also be bound by ZFY. Control primers target a chromosome region where there is no ZFX or ZFY binding.
11

Figure 2.3 ChIP qPCR assay to test ChIP-seq enrichment. Shown is the ChIP-qPCR result of ZFY-
FLAG and ZFX-FLAG transfected DKO HEK293T cells. Two panels are ChIP biological replicates for
each other. The signals of LRRC41 in each replicate were normalized to the signals of a control primers in
corresponding experiments. LRRC41 is a known ZFX target promoter; The control primers used in two
panels target a chromosome region where there is no ZFX or ZFY binding.

ChIP-seq libraries from DKO cells were prepared using the KAPA HyperPrep kit (Roche #KK8502)
following the manufacturer’s protocol. Samples were sequenced using 150bp paired-end reads. Endogenous
ZFX ChIP experiments of 22Rv1 cells, HEK293T cells, U87 cells and C42B cells were performed by
previous Farnham’s lab members. I performed the ChIP-seq analysis for all ChIP datasets. All ChIP-seq
data were processed according to the ENCODE ChIP-seq pipeline and mapped to the human genome hg38
using bowtie2 (2.4.2). ChIP-seq peaks were called using MACS2 (2.2.7.1). Bedtools (2.27.1) was used to
get the reproducible peaks. Peaks are annotated by CEAS package (1.0.2)

12

Chapter 3. Functional Analysis of ZFY in Multiple Human Cell
Lines.

3.1 Abstract
As detailed in Chapter 1, ZFX and ZFY are thought to be transcriptional activators. These two genes are
very similar in structure, especially in the DNA binding domains. However, there are differences in their
N-terminal domains (which are thought to be the activation domain). Therefore, it is possible that these two
highly similar genes have different functions. To test this hypothesis, I compared the transcriptome changes
in 5 different cell lines after ZFX or ZFY overexpression. I found that ZFY or ZFX overexpression increases
the levels of a common set of genes, many of which are affected by both TFs. However, I also identified a
set of genes that only responded to the overexpression of ZFY. To characterize these sets of genes, I
performed a gene ontology analysis and used ChIP-seq analysis to distinguish direct vs indirect target genes.
I performed all experiments described in this chapter except for the 22Rv1 ChIP-seq datasets. I performed
all analyses of the data.

13

To determine the function of the transcription factor, ZFY, I transfected 5 cell lines with ZFX-FLAG
or ZFY-FLAG expression constructs. For my experiments, I wished to use cell lines that express no or very
little ZFY. To achieve maximal effects in the transfection assays. I began by using DKO cells, which are
derived from HEK293T cells. HEK293T cells are human embryonic kidney cells which express ZFX and
ZNF11, but not ZFY (because they are female cells). Previous studies in the Farnham lab used CRISPR to
knockout expression of ZFX and ZNF711; these double knockout cells are called DKO
[37]
. In addition to
the DKO cells, I also used other cells which express relatively low levels of ZFY; I used prostate cancer
cells and glioblastoma cells. (Figure 3.1).

Figure 3.1 ZFY & ZFX expression level in multiple cell lines. Relative expression levels in 5 tested cell
lines. ZFX and ZFY expressions were normalized to the same housekeeping gene, HPRT1, and then
compared to the ZFX expression in 22Rv1 cells.

After transfection of DKO cells with ZFX and ZFY expression constructs, I harvested the cells and
prepared RNA. I first determined that my transfection worked by using RT-PCR to examine genes
previously identified to respond to ZFX (see Methods Chapter). Upon confirming that the experiment
worked, and the RNA quality was high, RNA-seq analyses were performed. I prepared and analyzed 3
replicates each of DKO cells transfected with ZFX, ZFY , or a control vector. Overexpression of ZFX or
ZFY resulted in significant transcriptome changes in DKO HEK293T cells. Volcano plots showing the
differentially expressed genes after transfection of ZFX and ZFY into DKO cells are shown in (Figure 3.2).
As expected upon introducing an activator in cells, there are more genes going up than down, which
indicates that, like ZFX, ZFY is a transcriptional activator.

I also used two prostate cancer cell lines, C42B and 22RV1 cells. Although C42B are male cells, they
have lost their Y chromosome, either during the tumorigenesis process or during the creation of the cell line.
14

Therefore, these cells do not make any ZFY (Figure 3.1). Although they have retained their X chromosome,
they make ZFX but not ZNF711. On the other hand, 22Rv1 cells have retained their Y chromosome and
express all 3 ZFX family members. The same treatment was performed as described above. ZFY and ZFX
constructs were transfected into C42B or 22Rv1 cells. After 24 hours, I harvested the cells and extracted
the RNA. RT-qPCR was used to monitor the effects of overexpression by examining genes respond to ZFX.
RNAs integrity were checked on the Bioanalyzer before sequencing. 3 replicates of C42B and 22Rv1 cells
were transfected with ZFX, ZFY or a control plasmid. In C42B and 22Rv1 cells, the overexpression of both
ZFX and ZFY had an impact on cell transcriptome (Figure 3.2), but the effect was not as large as what was
observed in DKO cells. This is likely because of the existence of endogenous ZFX expression in C42B cells
and endogenous ZFX and ZFY expression in 22Rv1 cells. As with the DKO experiments, the number of
upregulated genes in both cell lines is much more than the number of downregulated genes.

U87 and LN229 are two human glioblastoma cell lines used in this project. U87 are male cells but
express a very small amount of ZFY and do not express ZNF711 (Figure 3.1). LN229 were taken from a
female patient. Both ZFY and ZNF711 are not expressed in LN229 cells. The same treatment was performed
as described above. Transfected cells were harvested after 24 hours. The effects of overexpression and the
integrity of the RNA were confirmed by RT-qPCR and the Bioanalyzer before sequencing. I found both
U87 and LN229 cells have a significant response to the overexpression. Although the effects are still not as
large as the effects in DKO cells, compared to C42B cells, they have a larger impact on their transcriptome.
In each cell line, between overexpression of ZFX and ZFY, there was a very similar number of regulated
genes. LN229 cells have a larger impact on its transcriptome than U87 cells. A small amount of ZFY in
U87 cells may be the reason why LN229 cells have a more significant impact than U87 cells.

15

Figure 3.2 Effects of ZFX or ZFY overexpression on the transcriptomes of DKO HEK293T, C42B,
LN229, U87, and 22Rv1 cells. Differentially expressed genes are shown for ZFX or ZFY overexpression
vs. control. Significantly downregulated genes are shown in blue and upregulated genes are shown in red.
(cutoff: adjusted p-value (padj)<0.05, |FC|>1.5)
16

3.2 Identification of genes regulated by ZFY and ZFX

In the experiments described above, I showed that both ZFY and ZFX can regulate a large number of
genes in human cells. To determine if they regulate the same genes, I compared the sets of genes whose
expression was increased or decreased by ZFX or ZFY in each cell line. Because ZFX is a transcriptional
activator, the downregulated genes will be indirect targets and the upregulated genes will contain both direct
and indirect targets. By comparing the transcriptome changes driven by ZFX and ZFY in DKO cells, I
identified a large set of genes regulated by both TFs. However, ZFY has a number of unique regulated genes.
To determine the function of ZFY unique regulated genes, a GO analysis was performed for the sets of ZFY
unique regulated genes and the common regulated genes (Figure3.4). The analysis showed that ZFY
uniquely regulates some neural processes such as neurogenesis and neuron differentiation in DKO
HEK293T cells. A common function of ZFX and ZFY is related to cell fate and cell morphology.

I next compared the gene sets regulated by ZFX and ZFY in C42B prostate cells, LN229 and U87
human glioblastoma cell lines. Similar to the results using DKO cells, I found that most genes that showed
increased expression were similar for ZFX and ZFY but that a larger percentage of the down-regulated
genes were different in C42B and U87 cells. The putative directly ZFY uniquely regulated gene in all 3 cell
lines involved in cell fate, development, and differentiation. GO analysis showed that the genes uniquely
regulated by ZFY in C42B cells were also involved in organ development and cell morphogenesis. The
common regulated genes participate in the response to type Ⅰ interferon and its pathway. In LN229 and
U87 cells, the ZFY uniquely regulated genes regulate regionalization. However, most processes have
presented in GO analysis on both ZFY-specific genes and common regulated genes.

In 22Rv1 prostate cancer cells, the gene regulated by ZFX and ZFY were compared. Like other cell
lines, most genes that upregulated by ZFX are also regulated by ZFY. But in 22Rv1 cells, there is a slightly
larger proportion of ZFY unique regulated genes. As for the down-regulated genes, a large proportion of
them are different in 22Rv1 cells, similar to the results using C42B and U87 cells. Common regulated genes
and ZFY unique regulated genes are involved in similar biological processes such as cell fate, neural
processes, and recognition.

17

Figure 3.3. Identification of genes regulated by ZFY but not ZFX. Venn plots show comparison of genes
regulated by ZFY and ZFX. A) Venn plots of comparison of ZFY&ZFX upregulated genes in each cell line.
B) Venn plots of comparison of ZFY&ZFX downregulated genes in each cell line.

18

A

19

B

20

Figure 3.4. GO analysis of genes regulated by ZFY but not ZFX.
A) Classification of genes commonly regulated by ZFX and ZFY. GO analysis on commonly regulated
genes by ZFX and ZFY , the biological processes within the top 10 results from GeneRatio are shown on
each graph. B) Classification of genes uniquely regulated by ZFY. GO analysis on ZFY specific genes, the
biological processes within the top 10 results from GeneRatio are shown on each graph.

21

3.3 ZFY regulates genes in a cell type-specific manner

To determine if the genes affected by overexpression of ZFY were the same or different in the different
cell lines, I compared the up-regulated and down-regulated genes across 5 cell lines (Figure 3.5 A&B). As
described above, the upregulated genes are identified as putative direct targets. In the ZFY upregulated gene
Venn plot, there are a set of cell type-specific gene in all 5 cell lines. DKO cells have over 60 percent of
cell type-specific putatively direct regulated genes among all the upregulated genes. In LN229, U87 and
C42B cells, though the proportion of putative direct regulated genes is not as large as it is in DKO cells,
there are considerable cell type-specific genes in the three cell lines. It seems that 22Rv1 cells have only
11.59% cell type-specific genes, but 164 genes are commonly regulated only by 22Rv1 and DKO cells. The
number of up-regulated genes in DKO may be so huge that covers some possible cell type-specific genes
in other 4 cell lines. Therefore, in other cell lines there might be more cell type-specific genes that can be
identified.

The down-regulated genes are considered as indirect genes. In the ZFY-down-regulated gene Venn
plot, the percentage of cell type-specific genes are much larger than the percentage in ZFY-upregulated
genes Venn plot. All cell lines have almost a half of down-regulated genes are cell type-specific. In addition,
two human glioblastoma, LN229 cells and U87 cells have a larger percentage of overlap compared to any
other combination of 2 cell lines except DKO cells in both up-regulated Venn plot and down-regulated Venn
plot.

To deeply investigate the regulation pattern of ZFY, I then performed GO analysis of ZFY up-regulated
genes and down-regulated genes in each cell line (Figure 3.6 A&B). GO analysis of ZFY up-regulated
genes show ZFY plays a role in the processes such as development, cell fate and neuron projection in most
cell lines. Because GO analysis just showed biological processes within the top 10 GeneRatio, no
significantly specific biological process appears across the 5 cell lines. It is worth mentioning that GO
analysis of U87 and LN229 cells have similar top 10 related biological process. GO analysis of ZFY down-
regulated genes were performed in DKO cell, LN229 cell and U87 cell. In DKO and LN229 cell, the ZFY
involved processes are largely related to all kinds of RNA process.

22

Figure 3.5. ZFY regulates genes in a cell type-specific manner. A) Comparison of ZFY-upregulated
genes in 5 different cell lines. On the Venn plot, the numbers indicate the number of genes for a
corresponding set. The bar graph shows the total number of ZFY up-regulated genes in each corresponding
cell line. The percentages above each bar indicate the percentage of cell type-specific genes in each cell
line. B) Comparison of ZFY-downregulated genes in 5 different cell lines. On the Venn plot, the numbers
indicate the number of genes for a corresponding set. The bar graph shows the total number of ZFY down-
regulated genes in corresponding cell lines. The percentages above each bar indicate the percentage of cell
type-specific genes in each cell line.

23

A

24

B

Figure 3.6 Go analysis of ZFY regulated genes in multiple cell lines. A) Gene ontology analysis of genes
upregulated by ZFY in the different cell lines. B) Gene ontology analysis of genes downregulated by ZFY
in the different cell lines

25

3.4 Identification of genes directly regulated by ZFY but not ZFX

As shown above, I identified a large set of genes that are affected by overexpression of ZFY, but not
ZFX, in DKO cells. These genes are likely to include both direct and indirect targets. The genes that are
directly regulated by ZFY, in contrast to the genes that show expression changes due to effects on signaling
pathways, can be identified by combining the ZFY RNA-seq results and ZFY ChIP-seq results. Genes that
show expression changes and have ZFY bound to their promoters are the direct targets. To begin these
analyses, I first compared the ZFX and ZFY binding patterns in 22Rv1 cells. Previous work in the Farnham
lab had shown that the general pattern of ZFX and ZFY in 22Rv1 cells was similar. However, a detailed
comparison had not been performed. Therefore, I have called peaks and compared binding patterns of ZFX
and ZFY in 22Rv1 cells. None of the other cell lines I used have robust expression of endogenous ZFY.
Therefore, I transfected DKO cells with the FLAG-tagged ZFX or ZFY expression plasmids and performed
ChIP-seq using the FLAG antibody. I confirmed that my ChIP experiments worked by performing ChIP-
PCR using a known ZFX target promoter (see Methods chapter). I then created ChIP libraries and sent the
libraries for sequencing. I analyzed the sequencing files using Bowtie2 and MACS2 and identified
reproducible ZFX and ZFY peaks.

The genomic binding patterns of ZFX and ZFY throughout the human genome are very similar.
Browser tracks from each replicate of ZFX and ZFY ChIP-seq in DKO and 22Rv1 cells are shown in Figure
3.7.A and B. The binding profile shows that the majority of ZFY in DKO cells bind at around +240bp of
the TSS (Figure 3.7.E), which is the same as our previous study that ZFX binds to +240bp to the TSS. To
further investigate the binding pattern of ZFY and compare the binding pattern to ZFX, I identified the
binding sites of ZFY and ZFX in 22Rv1 and DKO cell and made heatmaps for comparison (Figure
3.7.C&D). The binding sites of ZFY and ZFX in both cell lines are very similar.

26

A Genomic binding pattern of ZFY and ZFX in 22Rv1 cells (chr8)

B Genomic binding pattern of ZFY and ZFX in DKO cells (chr7)

Figure 3.7 Comparison of ZFX vs ZFY genomic binding patterns. A) Genomic binding patterns of ZFY
and ZFX in 22Rv1 cells. B) Genomic binding patterns of ZFY and ZFX in DKO cells. C) Heatmap showing
22Rv1 ChIP-seq data from endogenous ZFX and ZFY centered at TSSs. Figure is sorted from highest to
lowest ZFY binding. D) Heatmap showing DKO ChIP-seq data from FLAG-tagged ZFX and ZFY centered
on TSSs. Figure is sorted from highest to lowest ZFY binding. E) Tag density plots for ZFX and ZFY in
DKO cells. Average signal of ZFY (Top) and ZFX (Bottom) ChIP-seq reads in DKO cells at ±2000 bp from
TSSs. One of two replicates were used for the Heatmaps and Tag density plots.

27

The RNA-seq results of overexpression experiments above identified a large set of ZFY unique up-
regulated genes in DKO cells. But not all of them are ZFY directly regulated genes. Some of them have
altered expression because they are indirectly affected due to being downstream of some pathway. In the
ZFY unique down-regulated genes there also should be some direct regulated genes. Therefore, to identify
the ZFY unique directly targets, I combined the RNA-seq and ChIP-seq analysis. First, to get the
reproducible peaks from ChIP-seq in each experiment, I intersected peaks from two replicates and selected
the reproducible peaks between -2000bp to 2000bp of TSS. After annotation, each peak was assigned a
gene ID. These data sets were used as ZFY directly binding gene sets in the following analysis. Then I
combined the genes with RNA-seq data sets to identify the ZFY unique directly target genes in DKO cells.
The same method was used to identify ZFX unique directly target genes.

Figure 3.8 Identification of unique genes directly regulated by ZFY. A) Comparison of RNA-seq and
ChIP-seq data in DKO cells. ZFY binding genes (DKO_Y) are represented by the red circle. ZFX binding
gene (DKO_X) are represented by the yellow circle. ZFY uniquely responding genes (DKO_RNA_Yunique)
are represented by the blue circle. B) Comparison of RNA-seq and ChIP-seq data in 22Rv1 cells. ZFY
binding genes (22Rv1_Y) are represented by the red circle. ZFX binding gene (22Rv1_X) are represented
by the yellow circle. ZFY uniquely responding genes (22Rv1_RNA_Yunique) are represented by the blue
circle.

ZFY unique up-regulated genes identified from overexpression RNA-seq data were used to identify
the genes uniquely regulated and bound by ZFY. The set of genes that are bound by ZFY but not ZFX and
28

overlapped with ZFY unique regulated genes are defined as ZFY unique directly targets. In DKO cells,
among 1084 genes bound only by ZFY, 114 ZFY unique directly target genes were found. (Figure 3.8.A)
In 22Rv1 cells, among 1089 genes bound only by ZFY , 40 ZFY unique directly target genes were found.
(Figure 3.8.B) Moreover, in Figure 3.6, ZFX and ZFY have a large set of common binding sites,
meanwhile ZFY has its ~1000 unique binding genes compared to ZFX in both cell lines

29

Chapter 4. Functional analysis of knockdown of ZFY vs. ZFX

4.1 Abstract
Overexpression experiments indicate ZFY may have both unique and redundant functions compared to
ZFX. However, these results may be influenced by the higher levels of expression of the transcription
factors in the transfection assay. A knockdown assay may provide gene expression information that is more
biologically relevant. And it is a good complement to the overexpression data. Therefore, to further
determine the function of ZFY , I transfected HepG2 cells with siRNAs targeting ZFY or ZFX and performed
an RNA-seq analysis. Previous Farnham lab members did siZFX/siZFY in U87 cells, I used the sequencing
data to perform the RNA-seq analysis. I identified a larger proportion of genes that uniquely responded to
ZFX or ZFY than what I found from overexpression experiments.

4.2 Effects of knockdown of ZFY and ZFX in U87 and HepG2 cells
I chose HepG2 cells, human hepatocellular carcinoma cells, because as male cells, HepG2 express
both ZFX and ZFY but not ZNF711 (Figure 3.1). I performed two sequential transfections of the siRNAs
to make the expression of ZFX or/and ZFY as low as possible. After that, I harvested cells and extracted
RNA. RNA integrity was checked on the Bioanalyzer before sending to be sequenced. Then, RT-qPCR was
used to test the effects of knockdown by checking the expression level of ZFX and ZFY. SiZFX or/and
siZFY and siControl RNAs w used in 3 replicates of HepG2 cells. Knockdown of either TF impacted the
transcriptome in HepG2 cells (Figure 4.1). Knockdown of ZFX resulted in more down-regulated genes
than up-regulated genes. However, there are more up-regulated genes than down-regulated genes after cell
transfected with siZFY or both siZFY and siZFX. The knockdowns in U87 cells were performed by previous
lab members in Dr. Farnham lab; however, I performed the RNA-seq analysis. The knockdowns in U87
cells had an impact on the cell transcriptome (Figure 4.1). In these cells, siZFY caused more down-
regulated effects than up-regulated effects. The effect was larger than we observed in HepG2 cells.

4.3 Comparison of genes in knockdown vs overexpression of ZFX and ZFY
According to the results from the previous chapter, both ZFX and ZFY are transcriptional activators.
The down-regulated genes in knockdown assays are the putative direct targets and indirect targets, whereas
the up-regulated genes in the knockdowns are likely all indirect targets. Different from overexpression data,
the comparison of ZFY and ZFX regulated genes in both cell lines showed significant differences (Figure
4.2).

30

Figure 4.1. Effects of knockdown of ZFY and ZFX in U87 and HepG2. DEGs are shown for siZFX or
siZFY vs. siControl. Significantly downregulated genes are shown in blue and upregulated genes are shown
in red. The expression levels of ZFX/ZFY are marked as yellow dots. (Cutoff: adjusted p-value (padj)<0.05,
|FC|>1.5)

31

Figure 4.2 Comparison of genes regulated by ZFY and ZFX in knockdown U87 and HepG2 cells.
Venn plots of ZFY&ZFX up- and down- regulated genes in each cell lines.

32

Figure 4.3 Comparison of genes in knockdown vs overexpression of ZFX and ZFY in U87 cells. The
down-regulated genes in knockdown experiments and the upregulated genes in overexpression experiments
in U87 cells were used to generate the Venn plots. The left panel shows comparison for ZFX. The blue
circle represents downregulated genes from RNA-seq in ZFX knockdown U87 cells; The green circle
represents upregulated genes from RNA-seq in ZFX overexpressed U87 cells. The left panel shows
comparison for ZFY. The yellow circle represents downregulated genes from RNA-seq in ZFY knockdown
U87 cells; The red circle represents upregulated genes from RNA-seq in ZFY overexpressed U87 cells.

I next wanted to compare the genes regulated by ZFX and ZFY in the overexpression vs the knockdown
experiments. I would expect that genes that require these TFs for high expression would have increased
levels in the overexpression assays and reduced expression in the knockdown assays. However, the
comparison of the entire set of ZFY- or ZFX-regulated genes between knockdown and overexpression
shows small overlaps, indicating very different transcriptome effects between two methods. (Figure 4.3)
The entire set of genes includes both direct and indirect targets. Therefore, I next identified genes directly
upregulated by the TFs by comparing the RNA-seq and ChIP-seq data. Because the ChIP-seq data is not
yet available for ZFY in all the cell lines, I focused on ZFX. Then I compared the two sets of directly
regulated genes in several cell lines. (Figure 4.4)

33

Figure 4.4 Comparison of directly regulated genes in knockdown vs. overexpression of ZFX in
HEK293T, C42B and U87 cells. Blue circles represent direct target genes identified from knockdown
RNA-seq data by combining down-regulated genes and endogenous ZFX ChIP-seq reproducible targets in
corresponding cell lines. Green circles represent direct target genes identified from overexpression RNA-
seq data by combining upregulated genes and ChIP-seq reproducible targets in the corresponding cell lines.
(ChIP experiments of U87 cells, HEK293T and C42B cells were performed by previous Farnham lab
members; I performed the analyses.)

In U87 and C42B cells, I identified a few direct genes that from both overexpression and knockdown
experiments. In HEK293T cells, the overlap is much larger. For knockdown assay, 205 direct regulated
genes among 225 direct regulated genes are overlapped with the direct regulated genes in the
overexpression experiment. These 205 genes that respond correctly to both experimental protocols are
robust ZFX target genes. However, there remain ~1000 genes that are direct targets in the overexpression
assay but not in the knockdown assay. As I mentioned in the Introduction chapter, high expression level of
ZFX can cause many kinds of cancers and poor survival. Although these 1000 genes are bound by ZFX in
endogenous HEK293T cells, it is possible that the amount of bound ZFX was not enough to activate the
promoter. However, overexpression of ZFX could possibly result in more robust binding and thus create a
“new” target promoter that is regulated by the increased expression of ZFX. If so, these “new” target genes
might explain the relationship between high level of ZFX and poor survival rates in cancer patients.
To test the hypothesis that overexpression of ZFX caused increased binding and higher transcription
34

of the “new” target genes, I have further investigated the 1000 “new” target genes. I have plotted the levels
of ZFX binding at the promoters of these 1000 genes in normal HEK293T cells and after overexpression
of ZFX. The tag density plots are shown in Figure 4.5. I found that there is much higher ZFX bound to
those promoters in the transfected cells.

Figure 4.5 Tag density plots of new target genes in HEK293T and DKO cells. Average signal of Flag-
ZFY ChIP-seq reads in DKO cells and endogenous ZFY ChIP-seq reads in HEK293T cells at ±3000 bp
from TSSs. One of two replicates were used for the tag density plot.

35

Chapter 5. Summary and Discussion

5.1 Summary of Results

Overexpression experiments show ZFY has its unique targets compared to ZFX
In overexpression experiments, I introduced ZFX-FLAG or ZFY-FLAG expressional plasmid into 5
cell lines and analyzed the RNA-seq data. The studies suggest that ZFY is a transcriptional activator and
that it can regulate many of the same genes as can ZFX. However, I have also identified some genes that
are regulated by ZFY but not by ZFX in multiple cell lines. Thus, ZFY may have both common and unique
functions compared to ZFX. To determine if the difference in transcriptome effects was due to a distinct
structure that allows ZFY but not ZFX to bind to some promoters, I used ChIP-seq analysis to identify
promoters bound by ZFY and ZFX. Combined with RNA-seq analysis, I found some promoters (114 in
DKO cells, 40 in 22Rv1 cells) uniquely upregulated by ZFY that had ZFY, but not ZFX, bound. These ZFY
uniquely binding genes may be the origin of its unique function. However, it is also possible that ZFY may
have the ability to interact with another protein bound to the promoter to activate transcription (and ZFX
cannot interact with that other protein). Therefore, in the future I will identify other proteins that can interact
with ZFY using Turbo-ID. And then, by comparing them with the proteins interacting with ZFX (which
have already been identified in the Farnham lab), a deeper understanding of the functional distinction
between ZFY and ZFX would be revealed. These future studies are described below in section 5.2.

Knockdown experiments indicate a greater transcriptional difference between ZFX and ZFY
Compared to overexpression experiments, knockdown experiments may provide information about the
function of TFs that is more highly related to their natural role in transcription. Hence, knockdown assays
are appropriate supplementary experiments for overexpression experiments. In this project, I analyzed
RNA-seq knockdown data using 2 cell lines, U87 cells and HepG2 cells. HepG2 cells have both ZFX and
ZFY expressed at similar levels, whereas U87 cells robustly express ZFX but make only a very small
amount of ZFY. After transfection of siZFY and/or siZFX into HepG2 cells and U87 cells, I found larger
differences of transcriptome changes between ZFX and ZFY than in the overexpression experiments, which
reinforces the conclusion that ZFY has a unique function compared to ZFX. Moreover, knockdown of
ZFX/ZFY in HepG2 cells expressing both ZFX and ZFY resulted in fewer transcriptome changes than
knockdown in U87 cells. This could be because U87 cells express a very small amount of ZFY, while
HepG2 cells have both endogenous ZFX and ZFY expression. To some extent, this may indicate the
redundant function between ZFX and ZFY. Although the expression of ZFY in U87 is very low, ZFY
knockdown in U87 cells still has similar impact on the transcriptome compared to ZFX. The overexpression
of ZFY in LN229 and U87 cells in the previous chapter attest even the small amount of ZFY seemly play a
role for U87 transcriptome.

New direct targets bound because of excessive ZFX in ZFX overexpression may be relevant to a poor
survival rate
Interestingly, when combining the overexpression results and the knockdown results, I found that the
36

effects on the transcriptome between these two methods are different. The ZFY- or ZFX-regulated genes in
each pair of overexpression and knockdown experiments have a small overlap. However, this comparison
used all the regulated genes from RNA-seq data (direct and indirect targets). Therefore, I next used ChIP-
seq analysis of ZFX to identify genes directly regulated by ZFX in both methods. In HEK293T cells, most
of the direct targets from knockdown experiments overlapped with direct targets from overexpression
experiments. However, the overlap was smaller in the U87 and C42B experiments. In all the cell lines, there
are many genes from overexpression experiments that are not affected when the normal levels of ZFX are
reduced. As described in the Introduction Chapter, abnormally high expression of ZFX may lead to a low
survival rate for cancer patients. Overexpression of ZFX in the transfection experiments may mimics the
abnormally high expression of ZFX in some cancer patients. The new direct targets identified in the
overexpression experiments may help identify cancer-specific ZFX target genes. I note that I cannot do
these same experiments for ZFY as I do not yet have all the ZFY ChIP-seq data.

Combining endogenous ZFX ChIP-seq results in HEK293T cells and overexpression ZFX ChIP-seq
results in DKO cells, I selected the 1038 new target genes and made the tag density plot. The higher ZFX
bound to the 1038 promoters may indicate that in cancer cells, those 1038 genes may be altered by ZFX
and then inducing malignant and lethal tumor, because increased expression of ZFX by transfection may
mimic the increased levels in cancer cells.

5.2 Future Studies: Identification of ZFY-interacting proteins using TurboID
As described above, I identified genes that are regulated by ZFY but not by ZFX. The specific function of
ZFY compared to ZFX may be caused by different co-factors recruited by ZFX and ZFY. In addition to
comparing the genes regulated by ZFY vs ZFX in each cell lines, I also compared the ZFY up-regulated
genes across these 5 cell lines. The sets of cell type-specific gene in all 5 cell lines indicate ZFY may
regulate genes in a cell type-specific manner. Again, this cell type-specificity may be mediated by protein-
protein interactions. Ongoing studies in the Farnham lab are using the method of TurboID to identify
proteins that interact with, or in the near vicinity of, ZFX. TurboID is a kind of proximity labeling
technology that can be used for the analysis of protein-protein interaction. The principle is to make a fusion
protein containing a biotin ligase that has proximity labeling function. The fusion protein can catalyze
nearby proteins adding biotin to them. Then the marked proteins can be enriched and identified by mass
spectrometry. A similar experimental protocol could be used to identify proteins that interact with, or are in
the near vicinity of, ZFY. To perform these experiments, I have created a ZFY-turbo plasmid (Figure 5.1.A).
I next ensured that fusing the turbo biotinylation domain to ZFY did not affect its function by comparing to
ability of ZFY-Turbo and the starting ZFY expression plasmid to activate a known target gene (Figure
5.1.B). To investigate the possibility that ZFY interacts with unique protein partners in different cell lines,
I will perform the following experiments:

1) Identification of co-activators and co-repressors for ZFY, that are not also bound ZFX.
I will perform TurboID using the ZFY-TurboID clone in diverse cell lines and compare the
identified proteins to those identified using a ZFX-TurboID plasmid. Because ZFY also has
transactivation activity, I expect that I will identify co-activators and perhaps other DNA-binding
37

TFs.

Figure 5.1 Using TurboID to identify ZFY-interacting proteins. A) Schematic of Turbo-ZFY construct.
Based on the construct of ZFY-FLAG plasmid, the miniTurbo sequence was inserted after the FLAG tag.
B) RT-qPCR assay to test the function of Turbo-ZFY. Figure shows relative expression levels of 3 known
targets in each group. Cells are transfected with ZFY-FLAG plasmid, Turbo-ZFY plasmid or Control
plasmid. The expressions were normalized to the same housekeeping gene in corresponding cell lines.

2) Map the domain required for ZFY interaction with the proteins identified using TurboID.. Because
the NTDs regions between ZFX and ZFY are the regions of the protein that are most diverse, this
region may be responsible for the activation of a unique set of promoters. I will design two pairs of
primers at the identical nucleic acid region of ZFX and ZFX and the common up-stream region on
the plasmid. One pair of primers can copy the ZFX NTD sequence and the other can get the
structure without NTD sequence on ZFY expression plasmid. The two pairs of primers add a
sequence overlap that can facilitate the Gibson reaction. This will create a ZFY expression construct
that has the ZFX NTD (Figure 5.2). After I create the new plasmid, I will repeat the
transactivation experiments and the TurboID .

38

Figure 5.2 Schematic of ZFY or ZFX fusion proteins. By exchanging the N-terminal coding region of
ZFY and ZFX, ZFY expression construct will express the ZFX NTD and ZFX expression construct will
express the ZFY NTD.

39

References
[1] Wang, Qiong et al. “Evolution of cis- and trans-regulatory divergence in the chicken genome between two
contrasting breeds analyzed using three tissue types at one-day-old.” BMC genomics vol. 20,1 933. 5 Dec. 2019,
doi:10.1186/s12864-019-6342-5

[2] Danino, Yehuda M et al. “The core promoter: At the heart of gene expression.” Biochimica et biophysica
acta vol. 1849,8 (2015): 1116-31. doi:10.1016/j.bbagrm.2015.04.003

[3] Thomas, Mary C, and Cheng-Ming Chiang. “The general transcription machinery and general
cofactors.” Critical reviews in biochemistry and molecular biology vol. 41,3 (2006): 105-78.
doi:10.1080/10409230600648736

[4] Wittkopp, Patricia J, and Gizem Kalay. “Cis-regulatory elements: molecular mechanisms and evolutionary
processes underlying divergence.” Nature reviews. Genetics vol. 13,1 59-69. 6 Dec. 2011, doi:10.1038/nrg3095

[5] Sperling, Silke. “Transcriptional regulation at a glance.” BMC bioinformatics vol. 8 Suppl 6,Suppl 6 S2. 27
Sep. 2007, doi:10.1186/1471-2105-8-S6-S2

[6] Karin, M. “Too many transcription factors: positive and negative interactions.” The New biologist vol. 2,2
(1990): 126-31.

[7] Luscombe, N M et al. “An overview of the structures of protein-DNA complexes.” Genome biology vol. 1,1
(2000): REVIEWS001. doi:10.1186/gb-2000-1-1-reviews001

[8] Vaquerizas, Juan M et al. “A census of human transcription factors: function, expression and
evolution.” Nature reviews. Genetics vol. 10,4 (2009): 252-63. doi:10.1038/nrg2538

[9] Lambert, Samuel A et al. “The Human Transcription Factors.” Cell vol. 172,4 (2018): 650-665.
doi:10.1016/j.cell.2018.01.029

[10] Lee, Tong Ihn, and Richard A Young. “Transcriptional regulation and its misregulation in disease.” Cell vol.
152,6 (2013): 1237-51. doi:10.1016/j.cell.2013.02.014

[11] Gertz, Jason et al. “Genistein and bisphenol A exposure cause estrogen receptor 1 to bind thousands of sites
in a cell type-specific manner.” Genome research vol. 22,11 (2012): 2153-62. doi:10.1101/gr.135681.111

[12] Pace, Nicholas J, and Eranthie Weerapana. “Zinc-binding cysteines: diverse functions and structural
motifs.” Biomolecules vol. 4,2 419-34. 17 Apr. 2014, doi:10.3390/biom4020419

[13] Gamsjaeger, Roland et al. “Sticky fingers: zinc-fingers as protein-recognition motifs.” Trends in
40

biochemical sciences vol. 32,2 (2007): 63-70. doi:10.1016/j.tibs.2006.12.007

[14] Iuchi, S. “Three classes of C2H2 zinc finger proteins.” Cellular and molecular life sciences : CMLS vol.
58,4 (2001): 625-35. doi:10.1007/PL00000885

[15] Swamynathan, Shivalingappa K. “Krüppel-like factors: three fingers in control.” Human genomics vol. 4,4
(2010): 263-70. doi:10.1186/1479-7364-4-4-263

[16] Urrutia, Raul. “KRAB-containing zinc-finger repressor proteins.” Genome biology vol. 4,10 (2003): 231.
doi:10.1186/gb-2003-4-10-231

[17] Collins, T et al. “All in the family: the BTB/POZ, KRAB, and SCAN domains.” Molecular and cellular
biology vol. 21,11 (2001): 3609-15. doi:10.1128/MCB.21.11.3609-3615.2001

[18] Tadepally, Hamsa D et al. “Evolution of C2H2-zinc finger genes and subfamilies in mammals: species-
specific duplication and loss of clusters, genes and effector domains.” BMC evolutionary biology vol. 8 176. 18
Jun. 2008, doi:10.1186/1471-2148-8-176

[19] Najafabadi, Hamed S et al. “C2H2 zinc finger proteins greatly expand the human regulatory
lexicon.” Nature biotechnology vol. 33,5 (2015): 555-62. doi:10.1038/nbt.3128

[20] Fedotova, A A et al. “C2H2 Zinc Finger Proteins: The Largest but Poorly Explored Family of Higher
Eukaryotic Transcription Factors.” Acta naturae vol. 9,2 (2017): 47-58.

[21] Darnell, James E Jr. “Transcription factors as targets for cancer therapy.” Nature reviews. Cancer vol. 2,10
(2002): 740-9. doi:10.1038/nrc906

[22] Jimenez-Sanchez, G et al. “Human disease genes.” Nature vol. 409,6822 (2001): 853-5.
doi:10.1038/35057050

[23] Al-Naama, Njoud et al. “C2H2-Type Zinc Finger Proteins in Brain Development, Neurodevelopmental, and
Other Neuropsychiatric Disorders: Systematic Literature-Based Analysis.” Frontiers in neurology vol. 11 32. 14
Feb. 2020, doi:10.3389/fneur.2020.00032

[24] Yan, Long et al. “The zinc finger E-box-binding homeobox 1 (Zeb1) promotes the conversion of mouse
fibroblasts into functional neurons.” The Journal of biological chemistry vol. 292,31 (2017): 12959-12970.
doi:10.1074/jbc.M116.771493

[25] Nuez, B et al. “Defective haematopoiesis in fetal liver resulting from inactivation of the EKLF
gene.” Nature vol. 375,6529 (1995): 316-8. doi:10.1038/375316a0

41

[26] Crotty, Shane et al. “Effectors and memories: Bcl-6 and Blimp-1 in T and B lymphocyte
differentiation.” Nature immunology vol. 11,2 (2010): 114-20. doi:10.1038/ni.1837

[27] Brown, Raymond S. “Zinc finger proteins: getting a grip on RNA.” Current opinion in structural
biology vol. 15,1 (2005): 94-8. doi:10.1016/j.sbi.2005.01.006

[28] Rhie, Suhn Kyong et al. “ZFX acts as a transcriptional activator in multiple types of human tumors by
binding downstream of transcription start sites at the majority of CpG island promoters.” Genome research, vol.
28,3 310–320. 2 Feb. 2018, doi:10.1101/gr.228809.117

[29] Blattler, Adam et al. “ZBTB33 binds unmethylated regions of the genome associated with actively expressed
genes.” Epigenetics & chromatin vol. 6,1 13. 21 May. 2013, doi:10.1186/1756-8935-6-13

[30] Jaeger, Savina A et al. “Conservation and regulatory associations of a wide affinity range of mouse
transcription factor binding sites.” Genomics vol. 95,4 (2010): 185-95. doi:10.1016/j.ygeno.2010.01.002

[31] Fang, Xiaoguang et al. “The zinc finger transcription factor ZFX is required for maintaining the tumorigenic
potential of glioblastoma stem cells.” Stem cells (Dayton, Ohio) vol. 32,8 (2014): 2033-47.
doi:10.1002/stem.1730

[32] Jiang, H et al. “Knockdown of zinc finger protein X-linked inhibits prostate cancer cell proliferation and
induces apoptosis by activating caspase-3 and caspase-9.” Cancer gene therapy vol. 19,10 (2012): 684-9.
doi:10.1038/cgt.2012.53

[33] Jiang, Mei et al. “The role of ZFX in non-small cell lung cancer development.” Oncology research vol. 20,4
(2012): 171-8. doi:10.3727/096504012x13548165987493

[34] Jiang, Jin, and Lu-Ying Liu. “Zinc finger protein X-linked is overexpressed in colorectal cancer and is
associated with poor prognosis.” Oncology letters vol. 10,2 (2015): 810-814. doi:10.3892/ol.2015.3353

[35] Zhou, Youxin et al. “The Zfx gene is expressed in human gliomas and is important in the proliferation and
apoptosis of the human malignant glioma cell line U251.” Journal of experimental & clinical cancer research :
CR vol. 30,1 114. 20 Dec. 2011, doi:10.1186/1756-9966-30-114

[36] Decarpentrie, Fanny et al. “Human and mouse ZFY genes produce a conserved testis-specific transcript
encoding a zinc finger protein with a short acidic domain and modified transactivation potential.” Human
molecular genetics vol. 21,12 (2012): 2631-45. doi:10.1093/hmg/dds088

[37] Ni, Weiya et al. “Characterization of the ZFX family of transcription factors that bind downstream of the
start site of CpG island promoters.” Nucleic acids research vol. 48,11 (2020): 5986-6000.
doi:10.1093/nar/gkaa384
42

[38] Navarro-Cobos, Maria Jose et al. “Genes that escape from X-chromosome inactivation: Potential
contributors to Klinefelter syndrome.” American journal of medical genetics. Part C, Seminars in medical
genetics vol. 184,2 (2020): 226-238. doi:10.1002/ajmg.c.31800

[39] Li, Changying et al. “ZFX is a Strong Predictor of Poor Prognosis in Renal Cell Carcinoma.” Medical
science monitor : international medical journal of experimental and clinical research vol. 21 3380-5. 5 Nov. 2015,
doi:10.12659/msm.894708

[40] Song, Xiaoling et al. “ZFX Promotes Proliferation and Metastasis of Pancreatic Cancer Cells via the MAPK
Pathway.” Cellular physiology and biochemistry : international journal of experimental cellular physiology,
biochemistry, and pharmacology vol. 48,1 (2018): 274-284. doi:10.1159/000491727

[41] Weng, Hao et al. “Zinc finger X-chromosomal protein (ZFX) is a significant prognostic indicator and
promotes cellular malignant potential in gallbladder cancer.” Cancer biology & therapy vol. 16,10 (2015): 1462-
70. doi:10.1080/15384047.2015.1070994

[42] Li, Yin et al. “High expression of Zinc-finger protein X-linked is associated with reduced E-cadherin
expression and unfavorable prognosis in nasopharyngeal carcinoma.” International journal of clinical and
experimental pathology vol. 8,4 3919-27. 1 Apr. 2015

[43] Mackeh, Rafah et al. “C2H2-Type Zinc Finger Proteins: Evolutionarily Old and New Partners of the Nuclear
Hormone Receptors.” Nuclear receptor signaling vol. 15 1550762918801071. 24 Oct. 2018,
doi:10.1177/1550762918801071

[44] Mackeh, Rafah et al. “C2H2-Type Zinc Finger Proteins: Evolutionarily Old and New Partners of the Nuclear
Hormone Receptors.” Nuclear receptor signaling vol. 15 1550762918801071. 24 Oct. 2018,
doi:10.1177/1550762918801071

Abstract (if available)

Abstract Gene expression is regulated by transcription factors (TFs), Zinc finger proteins are the largest class of DNA-binding transcription factors. ZFX gene family has 3 family members, ZFX, ZFY and ZNF711. ZFX and ZFY have very identical structure, and ZFX is related to many kinds of cancer according to previous study. Although many evidence support the hypothesis that ZFY may have similar function as ZFX, here are no publications showing an association of ZFY expression with cancer and the function of ZFY has not been well studied yet. This project aims to identify if ZFX and ZFY have redundant or unique function. Basically, I performed overexpression and knockdown of ZFY or/and ZFX in multiple human cell lines, and then analyzed the transcriptome changes by RNA-seq analysis and identified ZFX/ZFY target genes by ChIP-seq analysis. Here, I reported that, similar to ZFX, ZFY acts as a transcription activator. And although ZFY and ZFX can regulate a common set of genes, ZFY also has uniquely regulated genes. By combining RNA-seq and ChIP-seq data, I identified some promoters uniquely upregulated by ZFY that had ZFY, but not ZFX, bound. Comparing overexpression and knockdown experiments, I also identified a set of “new” target promoter that is regulated by the increased expression of ZFX. These findings suggest that ZFY does have some functions different than ZFX. Because the DNA binding domains of ZFY and ZFX are almost identical, it is likely that differences in protein-protein interactions and the N-terminal region between ZFY and ZFX contribute to the ability of ZFY to uniquely regulate a set of promoters not regulated by ZFX. To further analyze the function of ZFY, we have created a TurboID-ZFY construct and are using mass spectrometry to identify other transcription regulators that interact with ZFY, but not with ZFX. I will also create the fusion ZFY protein that is changed its NTD into ZFX NTD and then repeat the transactivation experiments and then using Turbo-ID to identify their interacted proteins.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Characterizing ZFX-mediated gene regulation to reveal possible candidates for clinical intervention

PDF

Identification of target genes and protein partners of ZNF711 in glioblastoma cells

PDF

Characterization of the ZFX family of transcription factors that bind downstream of the start site of CpG island promoters

PDF

Do ZFX and ZNF711 regulate the same genes in HEK293T cells?

PDF

Mapping transcription factor networks linked to glioblastoma multiform: identifying target genes of the oncogenic transcription factor ZFX in glioblastoma multiforme

PDF

Elucidation of MBNL1 function in the nervous system of myotonic dystrophy type 1

PDF

Mechanism of Mbnl1/2-depletion mediated neural defects

PDF

Functional characterization of a prostate cancer risk region

PDF

The relationship between DNA methylation and transcription factor binding in colon cancer cells

PDF

The role of nuclear GRP78 in regulation of EGFR expression in lung cancer

PDF

Impacts of post-translational modifications on interactions between G9a and its N-terminus binding partners

PDF

Exploring the mechanisms that control organellar protein homogeneity

PDF

Understanding the role of APP and DYRK1A in human brain pericytes

PDF

RNA methylation in cancer plasticity and drug resistance

PDF

Identification and characterization of cancer-associated enhancers

PDF

Epigenetic dysregulation in acute myeloid leukemia (AML) with MLL1 aberrations

PDF

Evaluation of preservatives in blood collection tubes for cell-free RNA transcriptional profiles in human plasma

PDF

Evaluating the therapeutic potential of targeting CBX8 in MLLr leukemia

PDF

Using genomics to understand the gene selectivity of steroid hormone receptors

PDF

The polydispersity of human FOXP3 fragment with zinc finger, leucine zipper and forkhead domain

Asset Metadata

Creator Liu, Yao (author)

Core Title Do the ZFX and ZFY transcription factors have redundant or unique functions?

School Keck School of Medicine

Degree Master of Science

Degree Program Biochemistry and Molecular Medicine

Degree Conferral Date 2022-12

Publication Date 09/07/2022

Defense Date 07/26/2022

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

Tag cancer,OAI-PMH Harvest,Transcription,transcription regulation,ZFX,ZFY,zinc finger

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Farnham, Peggy (committee chair), Craig, David (committee member), Rhie, Suhn (committee member), Rice, Judd (committee member)

Creator Email liu_yao0527@163.com,yliu6688@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-oUC111764646

Unique identifier UC111764646

Legacy Identifier etd-LiuYao-11182

Document Type Thesis

Format application/pdf (imt)

Rights Liu, Yao

Type texts

Source 20220908-usctheses-batch-978 (batch), 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 author, as the original true and official version of the work, but does not grant the reader permission to use the work if the desired use is covered by copyright. It is the author, as rights holder, who must provide use permission if such use is covered by copyright. The original signature page accompanying the original submission of the work to the USC Libraries is retained by the USC Libraries and a copy of it may be obtained by authorized requesters contacting the repository e-mail address given.

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

Repository Email cisadmin@lib.usc.edu