University of Southern California Dissertations and Theses

Characterizing and manipulating homology-directed gene editing in human cells

(USC Thesis Other)

Characterizing and manipulating homology-directed gene editing in human cells

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content
Characterizing and manipulating homology-directed gene
editing in human cells

Student: Chun Huang
Professor: Paula Cannon

August 2019

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

1

Table of Contents
Table of Contents ...................................................................................................................................... 1
List of Figures ........................................................................................................................................... 3
Abstract ..................................................................................................................................................... 4
Chapter 1. Introduction............................................................................................................................ 6
1.1 Gene therapy and gene editing ................................................................................................... 6
1.2 Editing at-a-distance ................................................................................................................... 8
1.3 DNA repair pathways ................................................................................................................. 8
1.4 Indel signatures and HDR-mediated editing efficiency. ......................................................... 10
Chapter 2. Materials and Methods ........................................................................................................ 12
2.1 Cell culture ................................................................................................................................ 12
2.2 mRNA preparation ................................................................................................................... 12
2.3 Homology donor design ............................................................................................................ 12
2.4 Nucleofection (electroporation) ................................................................................................ 14
2.5 Cel1 assay................................................................................................................................... 15
2.6 Restriction fragment length polymorphism (RFLP) assay .................................................... 15
2.7 Sequencing & ICE (Inference of CRISPR Edits) assay .......................................................... 16
2.8 Western blotting ........................................................................................................................ 16
2.9 HSPCs (CD34
+
cells) purification ............................................................................................ 17
Chapter 3. Specific Aims ........................................................................................................................ 18
Aim 1. Mapping the landscape of HDR editing efficiency at Trim5α. ........................................ 18
Aim 2. Improving the efficiency of editing at-a-distance. ............................................................ 18
Aim 3. Using indel signatures to predict and optimize HDR-mediated gene editing. ................ 18
2

Chapter 4. Results and Discussion for editing at-a-distance ................................................................ 19
4.1 Construction of homology donors to measure HDR-mediated editing efficiency. ............... 19
4.2 The landscape of HDR efficiency is not symmetric. ............................................................... 19
4.3 Knocking down factors dissociating D-loop decreased HDR-mediated editing efficiency... 20
4.4 Overexpressing Rad51 has a positive effect on editing at-a-distance .................................... 21
4.5 Increasing long-tract end resection can help editing at-a-distance. ...................................... 21
Chapter 5. Results and Discussion for using indel signature as a predictor for HDR outcomes ...... 28
5.1 Each gRNA has a unique indel signature ................................................................................ 28
5.2 Indel signature for one gRNA is same across cell types ......................................................... 28
5.3 NHEJ inhibitor has a smaller impact on MMEJ-dominant gRNA. ...................................... 28
5.4 Optimizing homology donors and method to quantify HDR outcome. ................................. 29
5.5 Majority of HDR outcomes come from MMEJ pool. ............................................................. 30
5.6 MMEJ-dominant gRNAs can create higher HDR outcome. .................................................. 30
5.7 Why indel signatures vary among gRNAs............................................................................... 31
Chapter 6. Future direction ................................................................................................................... 40
6.1 Future direction for enhancing editing at-a-distance project ................................................ 40
6.2 Future directions for predicting HDR outcomes by indel signature. .................................... 40
Supplemental data .................................................................................................................................. 42
References................................................................................................................................................ 45

3

List of Figures
Figure 1. Overview of different DNA repair mechanism. 7
Figure 2. the classical HR pathway. 9
Figure 3 Homology templates designed for Trim5α ZFN & Cas9/gRNA. 14
Figure 4. Landscape of Trim5α. 22
Figure 5. The effect of knocking down RTEL1 and RECQ5 on editing at-a-distance. 23
Figure 6. The effect of overexpressing Rad51 on editing at-a-distance. 24
Figure 7. The effect of overexpressing BLM on editing at-a-distance. 26
Figure 8. The effect of overexpressing BLM and Exo1 on editing at-a-distance. 27
Figure 9. The panel of gRNAs available in lab with indel signature. 32
Figure 10. Dominant indels are conserved in K562 and HSPC. 33
Figure 11. low dosage of Nu7441 has lower impact on MMEJ-dominant gRNAs. 34
Figure 12. Nu7441 can increase % of MMEJ outcomes in HSPCs. 35
Figure 13: ssODN design for each gRNA target site. 36
Figure 14: Effect of ssODN length and PAM versus non-PAM donors on HDR-mediated editing
efficiency. 37
Figure 15. ssODN-mediated HDR events come from MMEJ pool. 38
Figure 16 gRNAs with higher MMEJ ratio can create more HDR outcome. 39

4

Abstract
Gene editing has great potential to cure many genetic and infectious diseases through the
modification of specific genome sequences. The process is initiated by the action of targeted
nucleases such as CRISPR/Cas9 and ZFN which can create double-strand breaks (DSB) at the
desired genomic edit sites. After a DSB is formed, subsequent repair of the DSB can create
different gene editing outcomes, such as gene disruption (caused by the non-homologous end
joining (NHEJ) pathway) or, in the presence of a homology donor template, precise gene editing
(caused by the homology-directed repair (HDR) pathway). In mammalian cells, NHEJ is the
dominant repair pathway. Although this prevalence of NHEJ can be beneficial if the desired
outcome is gene knock-out, the potential to disrupt a targeted gene becomes a deleterious effect
for applications where the desired outcome is HDR-mediated gene editing. Therefore, to limit the
impact caused by unwanted NHEJ outcomes in mammalian cells, a better editing approach is
needed.
In order to do this, I propose an alternative approach for gene editing which I call “editing at-a-
distance”. Instead of creating a DSB close to the intended edit site, the nuclease-induced DSB is
instead created in a nearby intron, and HDR repair proceeds to edit the gene at the more distant
site. I hypothesize that even if cells repair the DSB by NHEJ, the less critical sequences in an
intron will mean that the effect is more tolerable. However, the major challenge of this idea is that
limited studies have shown that HDR-mediated editing efficiency drops dramatically as the
nuclease-induced DSB is separated from the intended edit site. Moreover, editing can also be
polarized, with one direction being favored. Therefore, to make editing at-a-distance more
practical, I hypothesize that it is essential to understand the “landscape of repair” ie how HDR
efficiency varies with distance from a DSB. My result shows that the landscape of HDR-mediated
editing efficiency can be asymmetric and vary both upstream and downstream from the DSB. Also,
manipulating the expression of several HR-related factors can enhance long-tract HDR-mediated
editing efficiency in both Cas9 and ZFN system. In addition, since at a distance editing relies on
long tract processing, I also hypothesize that it may be possible to enhance this process by
manipulating the DNA repair factors involved in HDR.
Also, in this research, I evaluated the characteristics of gRNAs that made them more or less
suitable to promote HDR-mediated gene editing. Traditionally, an ideal gRNA is chosen based on
5

bulk cutting efficiency, because it is expected that higher rates of DSB formation will increase the
rates of an HDR outcome. However, unpublished data shows that this current approach may not
be optimal. I propose instead that the indel signature for a gRNA can predict HDR-mediated
editing efficiency. My results show that indel signatures for each gRNA are consistent across
different cell types, including in primary cells. Importantly, a gRNA producing fewer NHEJ-
mediated outcomes leads to higher HDR-mediated editing efficiency when combined with a
homology donor. With these findings, the indel signature should be considered when selecting an
ideal gRNA for HDR-mediated gene editing.
6

Chapter 1. Introduction
1.1 Gene therapy and gene editing
The idea of gene therapy is to deliver a gene sequence to cure or at least suppress diseases which
are usually inherited or infected diseases. Predominantly, classical gene therapy involves gene
transfer mediated by randomly integrating lentiviral or retroviral vectors. Although it has
successfully cured some monogenic disease such as Severe Combined Immunodeficiency (SCID)
or Wiskott-Aldrich Syndrome, people blame it for safety issue (Cox et al., 2015). Several SCID
patients treated with gene therapy were diagnosed leukemogenesis because of off target effect due
to random insertion by gamma-retroviral vectors (Howe et al., 2008; Hacein-Bey-Abina et al.,
2003). Compared to the classical approach, gene editing is a gene therapy approach that can
perform precise gene modification without causing random transgenic mutation. Therefore, gene
editing is a major technological breakthrough in the field of gene therapy. It provides the chance
to treat both infectious and genetic diseases through the precise modification of disease-related
sequences.
The initiation of gene editing is creation of a double-strand break (DSB) at specific genomic loci
by targeted nucleases such as zinc finger nuclease (ZFN) or clustered regularly interspaced short
palindromic repeat (CRISPR)-associated nuclease Cas9. ZFN can recognize a specific DNA
sequence by zinc finger domains and cut one strand of dsDNA by a FokI nuclease domain. Two
ZFN monomers can work together to form a DSB. On the other hand, CRISPR/Cas9 can bind to
the target sequence by guide RNA (gRNA). A protospacer adjacent motifs (PAM) sequence right
after the gRNA target sequence is required for Cas9 protein to cut the target site. After sensing the
DSB, cells will recruit factors involving in the DNA repair system to repair the DNA lesion. The
subsequent repair outcome can be achieved by either non-homologous end join (NHEJ), homology
directed repair (HDR) or other DNA repair pathways (fig. 1) (Prakash et al., 2016).
In mammalian cells, it is widely believed that NHEJ is the dominant pathway to repair DSB (Mao
et al., 2008). NHEJ can repair the DSB without a homology template but leads to insertions and
deletions, called indels. Indels can lead to frameshifts and disrupt the open reading frame (ORF)
in targeted genes. While the prevalence of NHEJ is beneficial if gene disruption (gene knockout)
can result in therapeutic benefit, it can be detrimental to gene editing if the therapeutic goal is to
7

precise gene editing (Cox et al., 2015). HDR, a template-based repair pathway, can use a homology
template as a blueprint to repair the DNA lesion scarlessly. By providing an exogenous homology
template containing the desired DNA sequence, the HDR pathway may be hijacked to precisely
introduce a change into the targeted gene without disrupting the ORF (fig. 1).

Figure 1. Overview of different DNA repair mechanism. C-NHEJ, classical NHEJ. Alt-EJ,
alternative end join. HDR: homologous directed repair.
Although HDR can implement precise gene editing, HDR activation is restrained and easily
supplanted by NHEJ. First, many HDR-relating factors are highly regulated and suppressed in G1
phase. For instance, CtIP and Exo1 are only phosphorylated in G2/S phase by cyclin-dependent
kinase (CDK) (Tomimatsu et al, 2013; Wang et al., 2013). Brca1 is also targeted by ubiquitin
ligase and unable to recruit other HDR-relating factors to the DSB site during G1 (Orthwein et al.,
2015). Even in the mid-S phase with maximum HDR activity, on average less than 50% of DNA
repair events utilize HDR (Karanam et al., 2012). Because NHEJ is such a dominant pathway and
8

can lead to gene disruption during gene editing, an alternative gene editing approach needs to be
proposed to avoid the deleterious effects of NHEJ.
1.2 Editing at-a-distance
To avoid the potentially deleterious effects of indels, I propose an alternative approach for gene
editing which is called “editing at-a-distance.” Traditionally, to enhance HDR-mediated editing
efficiency, a nuclease-induced DSB is created close to the desired edit site. However, as previously
described, if the DSB is repaired by error-prone NHEJ rather than HDR, it can lead to gene knock
out. Therefore, I have devised an alternative approach whereby the nuclease-induced DSB is
created in a nearby intron. I hypothesize that the indels created in the intron will not affect target
gene expression because they can be spliced out during mRNA maturation.
Although the idea of editing at-a-distance may increase the safety of gene editing, the major
challenge with this idea is that previous studies have shown that HDR-mediated editing efficiency
drops dramatically as the nuclease-induced DSB separates from the intended edit site. The first
paper investigating the gene conversion tract observed that 80% of editing has relatively short tract
gene conversion (less than 60 base-pair (bp)) (Elliott et al., 1998) in mouse embryonic stem cells.
Other studies have also shown that HDR mediated editing efficiency decreases by 60% to 80%
when the edit site is 100 bp away from the cut site caused by I-SecI restriction enzyme (Hollywood
et al., 2016; LaRocque et al., 2010; Brenneman et al., 2002). Therefore, improving editing at-a-
distance efficiency is critical to make editing at-a-distance practical. In this thesis, I try to improve
the HDR-mediated editing efficiency by using different approaches to manipulate classical HDR
repair pathway—homologous recombination (HR).
1.3 DNA repair pathways
After a DSB is formed by targeted nuclease, p53-binding protein 1 (53BP1) can recruit Ku70/80
complex binding to the DSB and inhibit end resection, which is necessary for HDR. Then, DNA-
PK and Artemis form a complex and act as scaffold to recruit other repair factors. Then, other
factors such as XRCC4 and DNA ligase IV bind to Ku70/80 and ligate the two broken DNA ends
(Chang et al., 2017). Because NHEJ is dominant, and it can inhibit HDR, inhibiting NHEJ is a
well-known method to increase HDR. In recent years, different kinds of NHEJ inhibitors have
claimed to be able to improve HDR-mediated editing efficiency. For example, the small molecules
9

Scr7 and Nu7441 can inhibit NHEJ by directly binding to the DNA-binding domain on DNA ligase
IV and blocking DNA-PK auto-phosphorylation respectively (Li et al., 2017; Tavecchio et al.,
2012). Also, 53BP1 inhibitor (i53), a ubiquitin mimic, can inhibit 53BP1 and block its activation
(Canny et al., 2018). However, many NHEJ inhibitors may not work well in primary cells. In our
previous results, we showed that out of many DNA-repair manipulators, only i53 could increase
the HDR ratio in a clinically relevant cell type, hematopoietic stem and progenitor cells (HSPCs).

Figure 2. the classical HR pathway
Aside from NHEJ inhibitor, I try to propose other possible approach that may be able to increase
editing at-a-distance efficiency--manipulating HR-relating factors. The overall HR pathway is
shown in fig. 2. The initiation factor for HR is Breast cancer type 1 protein (Brca1). If Brca1 out
competes its antagonist i53, it can interact with the resection complex MRE11-RAD50-NBS1
10

(MRN), and CtIP. The complex can initiate end resection and leave a short 3’ single-strand DNA
(ssDNA) overhang (Symington et al., 2014). Then, the 5’-3’ exonuclease Exo1, and a RecQ family
helicase, BLM, can form a complex to stimulate extended resection. The ssDNA binding protein
RPA can bind and stabilize the 3’ ssDNA overhang. Then, Brca1, Brca2, and PALB2 can work
together to replace RPA with Rad51 (Carreira et al., 2011). Rad51 is a critical protein in the HR
pathway. After binding to ssDNA, Rad51 can stimulate strand invasion into a homology template,
resulting in the formation of a DNA loop (D-loop). The invaded ssDNA sequence is used as a
primer polymerase-mediated DNA synthesis based on the homology template. Then, RecQ5 and
RTEL1 protein can dissociate the D-loop (Ceccaldi et al., 2016). The invaded overhang leaves the
homology template, and the broken site is re-ligated. I hypothesize that the editing at-a-distance is
restricted by (1) Resection tract length, (2) repair tract length and (3) frequency of strand invasion.
Therefore, I predict that enhancing editing at-a-distance can be reached by the following way. First,
increase extensive resection so that more ssDNA can be created and more room for long-tract gene
conversion. Second, stimulate more strand invasion and stabilize the D-loop structure. By doing
so, more DNA copy can be synthesized by the polymerase, and have more chance to reach the
desired editing site.
Aside from HDR and NHEJ, another DNA repair pathway called alternative end joining (alt-EJ)
or microhomology end joining (MMEJ) has recently attracted my focus because it shares the same
initial step, end resection, with HDR. After end resection, polymerase θ (Polθ) can prompt template
insertion starting from the microhomology sequence (2 to 20 bp) on both ssDNA overhang (fig 1)
(Wyatt et al., 2016). The sequence between DSB end and annealed sequence will be spliced out,
and then the DNA lesion is ligated. Because a sequence is spliced out, the indel patterns caused by
MMEJ are large deletion (>2 bp) (Taheri-Ghahfarokhi1 et al., 2018). Although I have not tried to
manipulate MMEJ to affect HDR in this thesis, the prevalence of MMEJ may be an important
indication to choose an ideal targeted nuclease cutting site for gene editing. More details will be
discussed below.
1.4 Indel signatures and HDR-mediated editing efficiency.
To find an ideal gRNA for gene editing, a common approach is to screen a panel of candidate
gRNAs, and the gRNA with highest cutting efficiency will be chosen. However, unpublished data
in the lab suggests that only considering cutting efficiency may not be a good predictor to choose
11

an ideal gRNA for gene editing. In the data, a gRNA with 60% cutting efficiency can edit 10%
better than the other gRNA which yields 80% cutting efficiency. The similar outcome is also
observed in other experiments in our lab. By looking into the indel outcome for each gRNA, I find
out that the gRNA with more large deletions—indicative of MMEJ—is prone to have higher HDR-
mediated editing efficiency. Based on this outcome, I believe that the indel signature of each gRNA
should be considered to make a more precise choice.
In the DNA repair system, MMEJ shares some common features with HDR. First, both MMEJ
and HDR share the same initial step—end resection. They both need a 3’ ssDNA overhang to
induce template insertion or strand invasion respectively. Also, they are both more prevalent in
G2/S phase (Truong et al., 2013; Karanam et al., 2012). In contrast, canonical NHEJ (c-NHEJ) is
resection-independent and results in the inhibition of end-resection (Symington et al., 2014). Also,
studies show that both MMEJ frequency and HDR-mediated editing increase when NHEJ inhibitor
presents (Brinkman et al., 2018; Jayavaradhan et al., 2019). Because of these shared features, I
hypothesize a gRNA with more MMEJ-mediated indels should stimulate higher HDR-mediated
editing efficiency.
The standard to define an indel is MMEJ- or NHEJ-mediated is the overhang created by targeted
nucleases. Because Cas9 can create blunt end or +1 overhang DSB, small indels (<2 bp) are
commonly viewed as NHEJ-mediated indels while large deletions (>2 bp) are categorized into
MMEJ-mediated indels (Taheri-Ghahfarokhi1 et al., 2018). The following reasons can strengthen
the statement. First, the frequency of small indels declines in the presence of NHEJ inhibitor
(Nu7441), while the frequency of large deletions reduces significantly in Polθ or MRE11 knocked-
out cell line (Taheri-Ghahfarokhi1, 2018; Brinkman, 2018). Second, Brinkman et al show that the
formation of large deletion is much slower than small indels. This result echoes the theory that
resection-dependent MMEJ operates at a slower speed than NHEJ (Brinkman, 2018; Biehs et al.,
2017; Lobrich et al., 1995). Isn conclusion, in this thesis, I define NHEJ can cause small indels (<
2 bp) while MMEJ results in large deletions (>2 bp).

12

Chapter 2. Materials and Methods
2.1 Cell culture
CEM-ss (T cell line), K562 (hematopoietic cell line) and human hematopoietic stem and
progenitor cells (hHSPCs) were used for in vitro study. Both CEM-ss cells and K562 cells were
cultured in Roswell Park Memorial Institute medium (RPMI), supplemented with 10% fetal bovine
serum (FBS) and penicillin-streptomycin (P/S). Cells were passaged every two or three days when
the cell density reached around 8e5 cell/ml. Cells were cultured in T75 flasks and placed in 37°C
5% CO 2 incubator. To keep cells under the optimum condition, cells were split 1 to 2 one day
before experiments.
hHSPCs were cultured in Serum-Free Expansion Medium II (SFEM-II) supplemented with SFT
(including Stem cell factors, Fms-related tyrosine kinase 3 ligand and Thrombopoietin) and
penicillin-streptomycin-amphotericin B (P/S/A). HSPCs growth in T25 or T75 flask with one
million cells per milliliter in 37°C 5% CO 2 incubator.
2.2 mRNA preparation
I use in vitro transduction (IVT) from Cellscript to produce mRNAs (for Rad51 and ZFN) from
plasmids. The plasmid for Rad51 is ordered from Addgene, while the plasmid of ZFN comes from
Sangamo. First, a restriction enzyme (XhoI or XbaI) was used to linearize the plasmid containing
T7 promoter along with the desired gene sequence. The linearized plasmid was purified, mixed
with IVT reagents and put in 37°C for 2 hours. After RNA production, DNase was added to
degrade the remaining DNA inside the sample. Then, RNA was purified by RNA Clean &
Concentrator kit (Zymo). To form a matured mRNA, 5’ cap and 3’ poly A tail are necessary for
RNA. Therefore, purified RNA was loaded into capping mixture and incubated at 37°C 1 hour.
Then, reagents for polyA reaction were added into samples and incubated at 37°C 1 hour. After
that, matured mRNA was purified by RNA Clean & Concentrator kit (Zymo), and RNA
concentration is measured by a nanodrop.
2.3 Homology donor design
Two homology donor systems were designed for this project. One is to work with ZFNs (called
ZFN donor), and the other type is for Cas9 (called Cas9 donor). For ZFN donor, the homology
13

sequence for Trim5α (chr11: 5,665,985-5,666,993 bp) was amplified from hHSPCs by PCR. The
homology sequence was inserted into the backbone of TOPO plasmid by TOPO cloning. After
constructing the donor, a 6-bp restriction site (XhoI) was inserted at various positions to create a
panel of homology donors, with the restriction site at positions -380, -265, -212, -106, -53, +20,
+53, +106, +212-bp relative to the ZFN cut site (fig 3 A).
For Cas9 donor, the homology sequence for Trim5α (chr11: 5,666,550-5,665,483 bp) was ordered
as a synthetic double-strand DNA from IDT (Integrated DNA Technology). Two point-mutations
(GGG to GCG) were introduced at two PAM sites which are respectively recognized by exon and
intron gRNA/Cas9 complex so that these two gRNA/Cas9 complexes cannot recognize and cut
the donor. The PAM site point-mutations also created two different restriction factors (HhaI and
BlpI). The homology sequence was inserted into the backbone of pVAX plasmid. Then, ten
different Cas9 donors were made. Each donor contains a 6-bp XhoI insertion at different positions
(+/- 40, 80, 120, 180, 300, 480, 600-bp) from the middle site between the intron and exon gRNA
target sites (fig 3 B).
(A) Donor construct (ZFN)

14

(B) Donor construct (Cas9)

Figure 3 Homology templates designed for Trim5α ZFN & Cas9/gRNA.
2.4 Nucleofection (electroporation)
Nucleofection was used to deliver required RNA or DNA into cells. For the electroporation
delivering Rad51 mRNA, BLM/ Exo1 plasmid (Addgene) or siRNA for RTEL1/ RECQ5
(Dharmacon), one million K562 cells were resuspended in 100 μl pre-warmed nucleofector
solution (SF buffer, Amaxa). The cell solution was mixed to 6 μg Rad51 mRNA, 5 μg BLM
plasmid, 6 μg Exo1 plasmid or 100-200 μM siRNA targeting RTEL1 or RECQ5 (see fig 5). Then,
the solution was transferred to a cuvette and electroporated with appropriated nucleofector
program (suggested by Amaxa). Cells were mixed with 2 ml 10% FBS RPMI medium and
incubated at 37°C 5% CO 2 incubator. Then, cells may be used to western blotting, or second
nucleofection was done to deliver targeted nucleases and homology donors.
To deliver targeted nucleases and corresponding homology donors, 200,000 K562 or 400,000
CEM-ss cells resuspended in 20 µl pre-warmed SF buffer. For the ZFN donor system, the cell
solution was mixed with 2 μg ZFN mRNA (1 μg for each ZFN monomer) and 1 μg ZFN donor
(plasmid). For the Cas9 system, the cell solution was mixed with 9 μg Cas9/gRNA complex and 1
μg Cas9 donor (plasmid) (or 100 pmol ssODN homology donor). Then, samples were transferred
into 16-well shuttle. Samples were electroporated with appropriate nucleofector program. After
electroporation, each sample was mixed with 500 µl 10% FBS RPMI medium, split into 24- or 96-
well plate and incubated in 37°C 5% CO
2
incubator. Cells were harvested in day 3 post-
nucleofection with Cel1 or RFLP protocol.
15

For HSPCs, one million HSPCs were washed and resuspended in 100 µl BTX buffer. Then,
samples were added into cuvettes and electroporated with the following condition: 250V and 5
msec. Samples were mixed with 900 µl 10% FBS SFEM-II/SFT medium and split into 96 well
plates. Cells were cultured in 37°C 5% CO 2 incubator and harvested on day five post-nucleofection
with Cel1 protocol.
2.5 Cel1 assay
Cel1 assay was used to analyze ZFN cutting efficiency. To harvest cell pellet, samples were
centrifuged, and the supernatant was discarded. Cells were resuspended with cell lysis buffer and
protein degrader (GeneArt kit, Invitrogen). Then, samples were placed in the thermal cycler and
run a cell lysis program. After cell lysis, PCR (Ampli-Taq Gold 360, Thermo Fisher) was used to
amplify the targeted gene sequence. Then, the quality of PCR products was checked on agarose
gel. PCR products were mixed with detection reaction buffer and proceed re-anneal protocol in a
thermal cycler. In this step, PCR product denatured in high temperature (95°C) and reannealed to
double-strand DNA (dsDNA) as the temperature slowly decreases. A proportion of re-annealed
dsDNA will have several bp mismatches at the nuclease cutting site because of various indel
patterns. Detection enzyme which can recognize the mismatch site was added into re-annealed
PCR product and incubated in 37°C 1hr. Samples were loaded and run on 10% polyacrylamide
gel. PCR products with mismatch sequences can form two smaller bands because they were cut
into two shorter sequences. Then, image the gel without saturated bands, and use ImageJ to
quantify the darkness of each band. The nuclease cutting efficiency = 1- ((1-fraction cleaved)
1/2
).
2.6 Restriction fragment length polymorphism (RFLP) assay
RFLP is used to analyze HDR-mediated editing efficiency. Because restriction site was added into
my designed homologous template, the targeted gene will carry the restriction site if it is edited
successfully. RFLP can measure HDR-mediated editing efficiency by quantifying the proportion
of DNA cut by the specific restriction enzyme.
After harvesting cells, extract DNA by DNeasy blood & tissue kit (Qiagen). Then, the targeted
gene was amplified by PCR (Accuprime HiFi). After examining the quality of PCR on agarose gel,
PCR products were incubated with the restriction enzyme that can recognize the specific restriction
site and incubated at 37°C overnight in a thermal cycler. Restriction enzyme-digested PCR
16

products were loaded and run on 1.5% agarose gel. Uncut band (larger band) was viewed as
unedited DNA. Edited DNA can be cut and form two smaller band on the gel. The darkness of
each band was measured by ImageJ. The HDR-mediated editing efficiency = (sum of two small
bands)/ (sum of three bands).
2.7 Sequencing & ICE (Inference of CRISPR Edits) assay
ICE assay is a software provided by Synthego which can be used to analyze Cas9 cutting efficiency
by analyzing Sanger sequencing data. Recently, ICE-beta system has also been provided to analyze
HDR editing efficiency. To analyze Cas9 cutting or editing efficiency with ICE (beta) assay, DNA
was extracted and amplified by Cel1 protocol. After PCR, DNA was purified with PCR
purification kit (Qiagen) and sent to Genewiz for Sanger sequence. After receiving sequencing
data, both sample data and control (un-touch) files were uploaded to the ICE website. ICE assay
can analyze the cutting/editing efficiency based on sequencing data and the gRNA sequence.
2.8 Western blotting
Cells were harvested and resuspended in RIPA (radio-immunoprecipitation assay) buffer (final
concentration: 1 million cells per 70 µl RIPA buffer). After shaking 15 minutes on ice, cell lysate
was centrifuged 14000g for 15 minutes. The supernatant was transferred into new eppendorf and
store in -20°C.
The supernatant was mixed with 1x laemmli buffer and incubated in 95°C for 10 minutes to
denature. Then, using protein electrophoresis (Criterion Precast Gel) to separate different sizes
proteins. After electrophoresis, proteins were transferred from gel to a PVDF membrane (Bio-Rad)
by the trans-blot system (Bio-Rad). Then, incubate the membrane with blocking buffer (5% milk
PBST) overnight in cold room. Next day, primary antibody was properly diluted in blocking buffer
and incubated with the membrane for 2 hours. The membrane was washed by PBST to get rid of
excessive primary antibody. Incubate membrane with 1:10000 diluted conjugated secondary
antibody for 1 hour. After that, the membrane was washed and incubated in western blotting
detection buffer (Amersham) for ten minutes. Then, images were acquired with
chemiluminescence.

17

2.9 HSPCs (CD34
+
cells) purification
HSPCs were harvested from fetal livers. Livers were obtained from Advanced Bioscience
Resources (Alameda, CA) or Novogenix Laboratories (Los Angeles, CA) as anonymous waste
products. The liver tissue was first processed by removal of any large pieces of connective tissue
and the remaining sample was incubated in RPMI-CD (RPMI with collagenase/ dispase), mashed
up until it looks homogenous and incubated at 37°C for 20 minutes. Digested tissue was passed
through a strainer to remove remaining connective tissue. Then, remaining tissue was centrifuged
1000 rpm 5 minutes, and supernatant was removed. Pellet was resuspended with 1x RBC (red
blood cell) lysis buffer for 10 minutes to remove red blood cells. The reaction was stopped by
adding RoboSep buffer (STEMCELL) and centrifuged 1000 rpm 5 minutes to separate unbroken
cells and RBC lysate. The cell pellet was mixed with human CD34 positive selection cocktail
(EasySep). A magnet was used to isolate CD34
+
cells from other types of cells. Then, purified
CD34
+
cells were cultured in SFEM-II medium in 37°C 5% CO
2
incubator. Final cell concentration
was one million cells per milliliter. Gene editing for CD34
+
cells was done 1-day post-purification.
18

Chapter 3. Specific Aims
Aim 1. Mapping the landscape of HDR editing efficiency at Trim5α.
It is known that gene editing efficiency declines dramatically as the nuclease is targeted further
from the editing site. I will extend this work to human cells by determining the ‘landscapes’ of
homology directed gene editing at a therapeutically relevant locus.
Aim 2. Improving the efficiency of editing at-a-distance.
Because HDR-mediated editing efficiency drops as the cut site gets further from the editing site,
enhancing HDR-mediated editing efficiency is critical for the concept of editing at-a-distance. I
intend to focus on improving the efficiency of editing at a distance by manipulating the expression
of key HR factors expression. I choose several key factors involving in different steps of HR,
including BLM, Exo1, Rad51, RTEL1, RECQ5.
Aim 3. Using indel signatures to predict and optimize HDR-mediated gene editing.
Traditionally, an ideal gRNA for gene editing is chosen based on its cutting efficiency. However,
cutting efficiency does not always predict HDR-mediated editing efficiency. I believe that a
targeted nuclease’s indel signature can be used to pre-select for HDR efficient reagents.

19

Chapter 4. Results and Discussion for editing at-a-distance
4.1 Construction of homology donors to measure HDR-mediated editing efficiency.
Trim5α gene was chosen to study editing at-a-distance because it is a promising factor which can
be used to against HIV infection through gene editing approach (Pham et al., 2010; Jung et al.,
2015; Nakayama et al., 2015). To understand the change of HDR-mediated editing efficiency
among different distances from the nuclease-mediated DSBs, two homology template systems
were constructed. One system was designed for ZFN called ZFN donor, and the other system was
designed for Cas9 called Cas9 donor. For ZFN donors, eight different ZFN donors were
constructed. A 6-bp restriction site was inserted into each donor at various distance from the ZFN-
mediated DSB (Fig.3A). On the other hand, for Cas9 donor, the following requirements were set
for this system. First, the homology donors can be used to both intron and exon gRNA. Then, the
donor should be as symmetric as possible to accurately compare the change of HDR-mediated
editing efficiency between upstream and downstream homology arms. To achieve this goal, the
donor was symmetric to the middle site between two gRNA target sites which were 120 bp away
from each other (Fig. 3B). Two point mutations (GGG to GCG) were introduced to PAM sites for
intron and exon gRNA to avoid the template recognized by these two Cas9/gRNA complexes. Also,
these two point mutations can be recognized by different restriction enzymes, so that at-site HDR-
mediated editing efficiency can be quantified. Finally, a 6-bp restriction site was inserted into each
donor at varying distances from the gRNA target sites to quantify editing at-a-distance efficiency.
4.2 The landscape of HDR efficiency is not symmetric.
As mentioned in the introduction, previous studies showed that HDR-mediated editing efficiency
drops asymmetrically between upstream and downstream from the DSB, but most of these
researches were done with DSB made by restriction enzyme rather than clinically relevant
nucleases such as ZFN or CRISPR/Cas9. I believe that mapping out the landscape of HDR-
mediated editing efficiency among various distance from the targeted nuclease target site is vital
to editing at-a-distance. By understanding the landscape, researchers can decide that the targeted
nuclease should create the DSB at the upstream or downstream intron from the desired edit site.
Therefore, I try to make the landscape for both ZFN and Cas9/gRNA in a human cell line.
The result in the exon ZFN system showed that, HDR-mediated editing efficiency drops around
20

70% as the editing site was 100 bp away from the DSB. Also, the HDR-mediated editing efficiency
drops slower in the left homology arm. The HDR-mediated editing efficiency at the -53-bp site
was 5%, while +53-bp site only has 2% HDR-mediated editing efficiency. The data suggests that
the at-a-distance editing caused by exon ZFN in this locus is asymmetric and prefer the upstream
from the cutting site (Fig 4A).
The similar result showed in the landscape made by exon gRNA (Fig 4B). The landscape for exon
gRNA showed that the HDR-mediated editing efficiency was also higher in the left homology arm.
The HDR-mediated editing efficiency only drops 40% when the editing site was -120-bp away
from the exon gRNA target site, while it loses about 70% at +120-bp editing site. In contrast, the
landscape for intron gRNA shows a nearly symmetric landscape. Interestingly, the HDR-mediated
editing efficiency in +/- 20-bp editing site does not drop compared to the at-site HDR-mediated
editing efficiency. By comparing the landscapes of HDR-mediated editing efficiency in fig 4, I
suggest that the landscape might not be symmetric, and the pattern may be distinct among
nuclease-mediated cut sites. However, only one locus was tested in this thesis. More loci should
be scanned to see if there will be some general preference for the landscape of editing at-a-distance.
4.3 Knocking down factors dissociating D-loop decreased HDR-mediated editing efficiency.
I hypothesize that editing at-a-distance efficiency may be enhanced by prolonging the existence of
D-loop. Therefore, I try to knock down factors involving in D-loop dissociation (Fig. 5). First,
RTEL1 was knocked down by siRNA. The RFLP result showed that HDR-mediated editing
efficiency drops significantly at 50-bp and 100-bp away from the ZFN cutting site in the presence
of siRTEL1. Although no significant effect as the distance gets further to 200-bp, the HDR-
mediated editing efficiency was below the limitation for RFLP correct detection. Second, as
RECQ5 was partially knocked down by siRNA, editing at-a-distance efficiency also dropped
though it was not statistically significant. Based on these results, I propose that knocking down
factors involving D-loop dissociation can decrease editing at-a-distance efficiency. This result may
be able to be explained by classical HR mechanism. After D-loop formation and new DNA
synthesis, there are two different mechanisms to finish the repair pathway. One is synthesis-
dependent strand annealing (SDSA), and the other one is DSB repair (DSBR) (Sebesta et al., 2016).
In SDSA, the new synthesized DNA dissociates from the homology template and re-ligates to the
original DNA sequence, resulting in a non-crossover outcome. However, in DSBR, the D-loop can
21

be cleaved and result in a crossover or non-crossover outcome. Because RTEL1 and RECQ5 can
promote SDSA and the non-crossover result (Uringa et al., 2010; Paliwal et al., 2013), more
crossover events may occur without these two factors. In my experiment, cells with crossover
results may die because chromosome is mistakenly linked with exogenous homology template.
Therefore, knocking down RTEL1 and RECQ5 can decrease HDR-mediated editing efficiency.
4.4 Overexpressing Rad51 has a positive effect on editing at-a-distance
D-loop formation is critical for template-mediated precise gene editing. If D-loop formation is
disrupted, the repair system can go through MMEJ rather than HDR after end resection (Mateos-
Gomez et al., 2015). Therefore, another potential approach to enhance HDR outcome should be
promoting D-loop formation by overexpressing Rad51, which is the key factor to form D-loop.
The result in fig 6 A shows that HDR-mediated editing efficiency at 50-bp away from the DSB
was enhanced when Rad51 was overexpressed. The same effect was also observed in the Cas9
system (Fig. 6 B). In the Cas9 system, only at-site editing was significantly enhanced by Rad51
overexpression However, it is clear that editing at-a-distance efficiency in Rad51 samples was
higher than samples treated with non-targeted RNA. In conclusion, though the fold increasing
HDR-mediated editing efficiency was modest, editing at-a-distance can be enhanced in the
presence of exogenous Rad51 mRNA.
4.5 Increasing long-tract end resection can help editing at-a-distance.
Next, factors involving extensive resection—BLM and Exo1—was tested. Until now, it is not clear
what mechanism decides when D-loop should dissociate and stop synthesizing new DNA. I
hypothesize that more extended resection may leave more space for new copied DNA. Also, a
previous study showed that longer ssDNA overhang has higher chance to form D-loop and
attenuate dissociation phase in vitro (Wright et al., 2014). As expected, HDR-mediated editing
efficiency was slightly increased as BLM was overexpressed, but it was not statistically significant
in the ZFN system (Fig 7 C). Then, in the Cas9 system, overexpressing BLM can significantly
increase at-site HDR-mediated editing efficiency (Fig 8). Also, editing at-a-distance efficiency
(from 20 to over 200 bp away) was higher in the samples overexpressing BLM. Moreover,
overexpressing Exo1 alone or co-overexpressing with BLM also shows a similar outcome to BLM
overexpression. In summary, though the exact mechanism is unknown, inducing extensive
resection is helpful for editing at-a-distance.
22

(A) Landscape of HDR% made by ZFN

(B) Landscape of HDR% made by Cas9

Figure 4. Landscape of Trim5α (N=2)
K562 were electroporated with (A) ZFN mRNA and ZFN donors or (B) exon/ intron Cas9 RNP
pre-complexing gRNA and Cas9 donors on day 1 and harvested on day 4. HDR% was measured
by RFLP assay. Cutting efficiency of ZFN was analyzed by Cel1 assay while Cas9/gRNA cutting
efficiency was measured by ICE assay. (limit of detection (LOD) for RFLP is 1.5%)

23

(A) The function of RTEL1 and RECQ5

(B) Knocking down RTEL1 inhibit HDR-mediated editing efficiency

(C) Knocking down RECQ5 slightly decrease HDR-mediated editing efficiency

Figure 5. The effect of knocking down RTEL1 and RECQ5 on editing at-a-distance
(A) The mechanism of RTEL1 and RECQ5. K562 were electroporated with siRNA targeting (B)
RTEL1 or (C) RECQ5 on day 1. Then, the second electroporation was done to deliver exon ZFN
mRNA and corresponding ZFN homology donors on day 3. Cells were harvested on day 6. HDR-
mediated editing efficiency were detected by RFLP assay (N=3).

24

(A) The function of Rad51 (B) Western blot for Rad51

(C) Cutting and HDR editing% with exon ZFN (N=3)

(D) Cutting and HDR editing % with intron gRNA (N=3)

25

Figure 6. The effect of overexpressing Rad51 on editing at-a-distance
(A) The function of Rad51. (B) Rad51 mRNA was electroporated into K562 on day 1. Cells were
harvested on day 2 for western blot. Rad51 or non-targeted mRNA were electroporated into K562
on day 1. Then, the second electroporation was used to deliver (C) Exon ZFN mRNA and
corresponding ZFN donors (plasmid) or (D) Cas9 RNP pre-complexing with intron gRNA and
Cas9 donors (plasmid) on day 2. Cells were harvested three days later. RFLP assay was used to
analyze HDR-mediated editing efficiency. ZFN cutting efficiency was analyzed by Cel1 assay,
while Cas9 cutting efficiency was measured by Sanger sequencing and ICE assay.

26

(A) The function of BLM (B) Western for BLM

(C) Cutting editing % (by Cel1) (N=1) (D) HDR Editing % (by RFLP) (N=3)

Figure 7. The effect of overexpressing BLM on editing at-a-distance
(A) The function of BLM. (B) K562 were electroporated with BLM plasmid on day 1 and
harvested on day 3 for western blot. (C, D) K562 were electroporated with plasmid which carries
targeted protein (BLM or non-targeted) genome sequence on day 1. Then, the second
electroporation was done to deliver exon ZFN mRNA and ZFN homology donors on day 3. Cells
were harvested on day 6. Cutting efficiency was analyzed by Cel1 assay, while HDR-mediated
editing efficiency were detected by RFLP assay.

27

(A) The function of BLM & Exo1 (B) Exo1 western

(C) Cutting efficiency (by ICE) (D) HDR-mediated Editing efficiency (by RFLP)

Figure 8. The effect of overexpressing BLM and Exo1 on editing at-a-distance.
(A) The function of BLM and Exo1. (B) different dosage of Exo1 plasmid was electroporated into
K562 (1 million cells) on day 1. Cells were harvested on day 3 for western blot. (C, D) K562 were
electroporated with plasmid which carry BLM or Exo1 genome sequence on day 1. Then, the
second electroporation was done to deliver Cas9 RNP pre-complexing with intron gRNA and Cas9
homology templates on day 3. Cells were harvested on day 6. Cutting efficiency were analyzed by
Sanger sequencing and ICE assay, while HDR-mediated editing efficiency were detected by RFLP
assay. (N=3)

28

Chapter 5. Results and Discussion for using indel signature as a predictor for
HDR outcomes
Traditionally, an ideal gRNA for gene editing was chosen based on its cutting efficiency because
higher cutting efficiency means that there is more chance to occur HDR for precise gene editing.
However, our previous data shows that a gRNA with lower cutting efficiency can create higher
HDR-mediated editing efficiency than the other gRNA with higher cutting efficiency. Therefore,
aside from cutting efficiency, other factors should be considered to choose an ideal gRNA. By
comparing the indel signature for each gRNA in the previous experiment, the gRNA with higher
HDR-mediated editing efficiency has much more MMEJ-mediated indels. Therefore, I
hypothesize that an MMEJ-dominant gRNA can promote higher HDR-mediated editing efficiency.
5.1 Each gRNA has a unique indel signature
To fully understand the relation between indel signature and HDR outcome, ten different gRNAs
were used to scan their indel signatures. Among them, six gRNAs target Trim5α locus, named
from T5-1 to T5-6. The rest of four gRNAs recognizing Tetherin locus, named as Tet-1 to Tet-4.
Fig. 9A shows the represented indel signature analyzed by ICE assay. By comparing to next-
generation sequencing data, ICE assay can correctly quantify indels higher than 2%. Fig. 9B shows
the indel signature for all ten gRNAs. The result suggested that each gRNA has a different
frequency of MMEJ-mediated indels.
5.2 Indel signature for one gRNA is same across cell types
The previous study shows that the indel signature for the same gRNA is similar across different
cell types (Overbeek et al., 2016). However, the paper only scanned the indel signature across cell
lines. I want to explore further in clinically relevant primary cells--CD34+ cells (hematopoietic
stem and progenitor cells (HSPCs)) (Fig 10). Indel signatures for four gRNAs were scanned in
HSPC, and the relative MMEJ and NHEJ ratio is similar to the result in K562. Though Tet-3 and
Tet-4 has similar MMEJ ratio in HSPC, Tet-3 has higher MMEJ ratio (about 20%) than Tet-4.
More replicate should be done in HSPCs to make the result more solid. But basically, the relative
indel signature and MMEJ ratio is consistent in K562 and HSPCs.
5.3 NHEJ inhibitor has a smaller impact on MMEJ-dominant gRNA.
29

Traditionally, a common way to boost HDR efficiency is treating samples with NHEJ inhibitors.
Previous studies mentioned that NHEJ inhibitors could affect indel signature by promoting MMEJ-
mediated indels (Jayavaradhan et al., 2019). Because my result showed that the MMEJ and NHEJ
ratio was different among gRNAs, I hypothesize that NHEJ inhibitor has less effect on MMEJ-
dominant gRNAs because NHEJ is less common to those gRNAs. The result showed that Nu7441
could inhibit NHEJ outcome thoroughly as its concentration reaches 5 µM in K562 cells. Higher
concentrations of Nu7441 (10 and 20 µM) induced high toxicity and decrease cutting efficiency
(Supplemental fig 2). Therefore, 1 µM Nu7441 was used to test the impact of Nu7441 on indel
signature with ten different gRNAs. The result showed that 1 µM Nu7441 can increase two to
three-fold MMEJ-mediated indels when the natural MMEJ ratio was about 25%; however, it has
almost no effect on Tet-3 which intrinsic MMEJ ratio was high (88%) (Fig 11 A). The impact of
Nu7441 on indel signature decreased as the gRNA is more MMEJ-dominant. The result was
analyzed by non-linearized regression analysis (R
2
= 0.76). The same result was observed in CEM-
ss cells (Fig 11 B). However, in CEM-ss cells, 5 µM Nu7441cannot thoroughly eliminate NHEJ
(Fig 15). It also causes massive CEM-ss cell death and decreased cutting efficiency. Furthermore,
the effect of Nu7441 was also tested on HSPCs. MMEJ-mediated indels increased from 60% to
80% as the concentration of Nu7441 increases from 0 µM to 3 µM (Fig 12). Among the three cell
types I tested, K562 can tolerate higher concentration of Nu7441. It may because K562 is p53
deficient, which can lead to apoptosis when facing DSB (Roos et al., 2006; Law et al.,1993).
Therefore, Nu7441 blocks NHEJ and may prolong DSB exists and causing more cell death in P53
+
cells. In conclusion, because the impact of NHEJ inhibitor on indel signature was various,
increasing HDR-mediated editing efficiency by NHEJ inhibitor may not be efficient to every
gRNA.
5.4 Optimizing homology donors and method to quantify HDR outcome.
Homology donor for each gRNA was designed to evaluate the potential relationship between indel
signature and HDR outcome. Donors with single-stranded oligonucleotide (ssODN) was used
because of easier construct and high HDR-mediated editing efficiency. Previous studies showed
that ssODN with short homology sequence (around 80-bp) could give the best HDR-mediated
editing efficiency (Liang et al., 2017). However, while Liang et al suggested that using ssODN
complementary to the DNA strand carrying the PAM site was better, Richardson et al proposed an
30

opposite result (Liang et al., 2017; Richardson et al., 2016). Because there is still no solid
conclusion for the best construction of ssODN, I designed three different types of ssODN (Fig 13)
to examine their efficiency. The result showed that short ssODN (80-bp donor) has little higher
HDR-mediated editing efficiency than long ssODN (180-bp donor) (fig 14 A). Then, two short
ssODNs (complementary to the PAM or the non-PAM sequence) were tested. Unlike the result
showed by previous studies, non-PAM donor and PAM-donor created similar HDR-mediated
editing efficiency (Fig 14 B).
Also, the Sanger sequence and ICE-beta assay was used to quantify HDR-mediated editing
efficiency in this part. A comparison between ICE-beta assay and RFLP assay was shown in
supplementary fig 3. The HDR efficiency analyzed by ICE-beta and RFLP assay is similar.
Therefore, I suggest using ICE-beta to analyze HDR-mediated editing efficiency since it can show
not only the HDR outcome but the indel signature change to the gRNA.
5.5 Majority of HDR outcomes come from MMEJ pool.
In fig 15, indel signature and HDR-mediated editing efficiency were tested in T5-1 and Tet-4 with
or without Nu7441. In CEM-ss cells, cutting efficiency dropped as the concentration of Nu7441
increases. After normalizing indel outcomes to the cutting efficiency, MMEJ outcome increases
in the presence of Nu7441 in CEM-ss cells. The total ratio of MMEJ and HDR in samples with
ssODN was similar to the ratio of MMEJ in “Cas9 only” samples (Fig 15 A). This result shows
that HDR outcome comes from the MMEJ pool. The similar result was observed in K562 cells
(Fig 15 B). Therefore, it seems obvious that the majority of HDR outcomes is pulled from MMEJ
pool.
5.6 MMEJ-dominant gRNAs can create higher HDR outcome.
HDR outcome is higher in MMEJ-dominant gRNAs. In figure 16, gRNAs with more MMEJ ratio
can induce more HDR outcome in the presence of homology donors. The MMEJ ratio and HDR
outcome share positive correlation (R
2
= 0.76). The result echoes our hypothesis that a gRNA with
higher MMEJ ratio can create higher HDR-mediated editing efficiency.
Furthermore, a high dosage of Nu7441 can largely increase MMEJ ratio and push HDR-mediated
editing efficiency to over 80% in K562 (Fig 15 B). However, the fold that HDR and MMEJ ratio
increase in the presence of Nu7441 was mild in CEM-ss cells. In CEM-ss cells, the sample with
31

MMEJ-dominant gRNA and Nu7441 treatment can have the best HDR outcome. In conclusion,
because the impact of NHEJ inhibitors is various among cell types, using NHEJ inhibitors and
MMEJ-dominant gRNA together can maximize HDR outcome, especially in those cell types more
vulnerable to Nu7441 such as CEM-ss cells.
5.7 Why indel signatures vary among gRNAs
My result shows that the ratio of NHEJ and MMEJ outcomes is distinct among different gRNAs.
Moreover, these different ratios also predicted which gRNA would result in higher ratio of HDR
mediated gene editing. I speculate that the reason for these observations is due to the cutting pattern
made by Cas9/gRNA. Cas9 protein can create a DSB by the action of two different catalytic
domains, HNH and RuvC. Traditionally, it is believed that both domains can cut at 3-bp upstream
of the PAM sequence and leave a blunt end DSB (Gasiunas et al., 2012). However, recently, studies
have shown that RuvC is flexible, and can cut DNA at either 3 or 4-bp upstream of the PAM
sequence, and thereby also leave a 1-bp 5’ overhang at the DSB. Also, some papers even shown
that the RuvC domain has exonuclease activity and can create a several bp short ssDNA overhang
at the break site in vitro. What mechanism decides the cutting pattern for each gRNA has not been
fully understood, but one possible factor may be related to gRNA/target sequence (Zuo et al., 2016;
Taheri-Ghahfarokhi et al., 2018). Furthermore, other papers also mentioned that the post-
translational modification of histones at the target sequence could affect the interaction between
DNA repair factors and the DNA lesion and thereby affect the choice of repair system (Clouaire et
al.,2019). Based on these reasons, the indel signature may vary at different gRNAs.

32

(A) Indel signature for T5-1

(B) Indel signature for Tet-4

(C) The indel signature is different across different gRNA (N=2)

Figure 9. The panel of gRNAs available in lab with indel signature
Ten Cas9/gRNAs were complexed. Among them, four gRNAs target tetherin gene (tet-1, 2, 3, 4),
and the other six target Trim5α locus (T5-1, 2, 3, 4, 5, 6). Indel signatures for each Cas9/gRNA
complex were analyzed by Sanger sequencing and ICE assay. (A, B) Representative indel
signatures for guides T5-1 and Tet-4. (C) The frequency of predicted MMEJ-mediated and NHEJ-
mediated indels for each guide RNA were calculated and normalized to total indels% at greater
than 2% of the population (because our previous result show that indels greater than 2% can be
calculated correctly by ICE assay).
baseline MMEJ for each gRNA (N=2)
T5-1
T5-3
T5-2
Tet-1
T5-6
T5-4
Tet-2
T5-5
Tet-4
Tet-3
0
50
100
MMEJ
NHEJ
%DSB repair outcomes
33

(A) Indel signatures in K562

(B) Indel signatures in HSPC

Figure 10. Dominant indels are conserved in K562 and HSPC.
(A, B) Indel signatures observed in HSPC and K562 (quantified by ICE assay).

indel signature change in K562
T5-1
Tet-2
Tet-4
Tet-3
0
50
100
MMEJ
NHEJ
indel signature change in HSC
T5-1
Tet-2
Tet-4
Tet-3
0
50
100
MMEJ
NHEJ
34

(A) The impact of Nu7441 is weaker as initial MMEJ gets higher (K562)

(B) The impact of Nu7441 is weaker as initial MMEJ gets higher (CEM-ss)

Figure 11. low dosage of Nu7441 has lower impact on MMEJ-dominant gRNAs.
K562 were electroporated with Cas9 pre-complexing to indicated gRNAs (mentioned in fig 9).
After electroporation, K562 were incubated with 1 μM Nu7441 and harvested on day 3 post-
electroporation. Indel signatures were analyzed by Sanger sequencing and ICE assay. (A, B) The
fold that MMEJ% increases at different gRNAs in the presence of 1 μM Nu7441 in K562 and
CEM-ss cells respectively. The MMEJ increasing-fold is calculated by following steps: First,
normalized MMEJ% = MMEJ%/ (MMEJ% + NHEJ%) in each sample. Then, MMEJ increasing-
fold = (normalized MMEJ% in 0 μM Nu7441)/ (normalized MMEJ% in 1 μM Nu7441).

35

Figure 12. Nu7441 can increase % of MMEJ outcomes in HSPCs.
HSPCs were electroporation with Cas9 RNP pre-complexed with Tet-4 gRNA. After
electroporation, HSPCs was treated with various dosage of Nu7441 (0, 0.5, 1, 2, 3µM). twenty-
four hours later, samples were washed by PBS to get rid of Nu7441 and prevent high toxicity.
HSPCs were harvested on day 5 post-electroporation. Indel outcome were analyzed by Sanger
sequencing and ICE assay.

36

Figure 13: ssODN design for each gRNA target site
To scan the HDR-mediated editing efficiency for each gRNAs, ssODN homology templates were
designed. Each template contains a 6-bp restriction site inserted between 3- and 4-bp upstream the
PAM sequence. Two 80 donors were designed. One was complementary to PAM sequence and the
other was complementary to non-PAM sequence.

37

(A) 80-bp or 180-bp ssODN donor can have similar HDR-mediated editing efficiency

(B) Non-PAM and PAM ssODN create similar HDR-mediated editing efficiency (N: non-PAM
donor; P: PAM donor)

Figure 14: Effect of ssODN length and PAM versus non-PAM donors on HDR-mediated
editing efficiency.
K562 were electroporated with Cas9 pre-complexed to indicated gRNAs and corresponding
ssODN templates (80-donor or 180-donor). Cells were harvested on day 3-post electroporation.
Sanger sequencing and ICE-beta assay was used to analyze indel and editing outcome.

38

(A) HDR events pulls from available MMEJ pool (CEM-ss)
(B) HDR events pulls from available MMEJ pool (K562)

Figure 15. ssODN-mediated HDR events come from MMEJ pool.
Cells (CEM-ss or K562) were electroporated with Cas9 RNP pre-complexing with gRNAs and
corresponding ssODN homology templates (80bp non-PAM sequence). After electroporation, cells
were incubated with different dosage of Nu7441 (0, 1, 5 μM) and harvested on day 3 post-
electroporation. Indels were analyzed by Sanger sequencing and ICE-beta assay. (A) The panel
shows the indel patterns in CEM-ss. Nu7441 decreased cutting efficiency in CEM-ss. Therefore,
a normalized version of the panel is shown. (B) The panel shows that the indel patterns in K562.
wt: wild type.

39

Figure 16 gRNAs with higher MMEJ ratio can create more HDR outcome
K562 cells were electroporated with Cas9 RNP pre-complexing with gRNAs and corresponding
ssODN homology templates. Cells were harvested on day 3 post-electroporation. Indels were
analyzed by Sanger sequencing and ICE-beta assay.

Data 10
0 20 40 60 80
0
20
40
60
80
R
2
=0.76
y=x
MMEJ%
HDR%
40

Chapter 6. Future direction
6.1 Future direction for enhancing editing at-a-distance project
To date, editing at-a-distance is a new idea. Although it can prevent deleterious effects of NHEJ,
the HDR-mediated editing efficiency is too low. My current data suggests that the efficiency of
editing at-a-distance can be enhanced by manipulating some HR-related factors, involving D-loop
formation or extensive resection. However, to further improve the efficiency, more factors should
be scanned. There are many other factors directly or indirectly involve in the HDR pathway;
however, their effects on editing at-a-distance have not been understood yet. Also, previous studies
show that manipulating some factors may increase the average tract for gene conversion but
dramatically decrease the frequency of HDR (Brenneman et al., 2002; Wu et al., 2015). To
efficiently find out useful factors, CRISPR interference (CRISPRi) and CRISPR activation
(CRISPRa) library should be used to scan hundreds of HDR-related gene and find out which
factors have more impact (Richardson et al., 2018). After screening CRISPRi/a library, factors
with highest impact will be overexpressed or knocked down individually to enhance the efficiency
of editing at-a-distance. By doing so, I hope I can get a optimal protocol to apply editing at-a-
distance into practice.
6.2 Future directions for predicting HDR outcomes by indel signature.
To further investigate the link between HDR efficiency and indel signature, the following topic
should be studied. First, my research will include other types of homology donor which are widely
used—double-strand DNA (dsDNA) donor and donor delivered by adeno-associated viral (AA V)
vector. In this thesis, HDR-mediated editing efficiency for each gRNA was scanned by ssODN
donors. However, the mechanism of ssODN-mediated HDR pathway has not been fully understood
and is believed to be distinct to the classical HR pathway (Richardson et al., 2018; Bothmer et al.,
2017). Therefore, to make my study more convincing, other types of donor should be tested if their
HDR outcomes also pull from MMEJ pool, and how many proportions of MMEJ will become
HDR outcome in the presence of homology donors. Second, because some cell types (like HSPCs)
naturally express more MMEJ factors (Kawamura et al., 2004), I suspect that it would HDR may
be more difficult to compete with MMEJ. Therefore, I will scan if inhibiting MMEJ can further
improve HDR outcome. I will use PARP inhibitor (PARPi) which is a small molecule MMEJ
41

inhibitor used to cure three negative breast cancer (Liu et al., 2017). Also, I will try to inhibit Polθ
by siRNA or CRISPRi because it is an essential factor to promote MMEJ and suppress HDR (Wyatt
et al., 2016; Mateos-Gomez et al., 2015). Finally, my current research focus on Streptococcus
pyogenes Cas9 (spCas9). The research will be expanded to other types of targeted nuclease such
as Staphylococcus aureus Cas9 (saCas9) and CRISPR/Cas12a (CpfI). saCas9 has the similar
cutting pattern as spCas9, while CpfI can leave a 4 to 5 bp overhang at the cut site (Murovec et al.,
2017; Zhang et al., 2017). I try to understand how different cutting patterns affect indel signature
and HDR-mediated editing efficiency. By exploring more in these topics, I should be able to
enhance HDR more efficiently than before.

42

Supplemental data

Supplemental figure 1. Example of gel image for RFLP assay.
The image comes from the experiment of landscape Trim5α (Fig. 4). Samples were digested by
BlpI. (A) samples were treated with Cas9 RNP pre-complexing with intron and exon gRNAs and
plasmid form Cas9 donor. (B) samples were treated with Cas9/gRNA complex only. (C) the sample
was treated with a homology donor only. (D) un-touch sample. The yellow arrow shows uncut
band, and the two blue arrows show bands cut by BlpI restriction enzyme.

43

(A) The impact of Nu7441 reaches the limit at 5 μM.

(B) High dosage of Nu7441 can decrease cutting efficiency

Supplemental figure 2. Optimizing Nu7441 dose with K562 cells.
K562 were electroporated with Cas9 RNP pre-complexed with T5-1 gRNA. DSB outcomes were
detected by Sanger sequencing and ICE assay 3 days post-electroporation. (A) Nu7441 can switch
indel pattern from NHEJ-dominant to MMEJ-dominant. NHEJ were completed suppressed as the
concentration of Nu7441 passes 5 μM. (B) Cas9 cutting efficiencies was declined in high dosages
of Nu7441 (10 and 20 μM).

44

Supplemental figure 3. HDR-mediated Editing efficiency analyzed by ICE-beta and RFLP
are similar.
K562 were electroporated with different length of ssODN (80 or 180 bp) and corresponding Cas9
RNP pre-complexing gRNA. Cells were harvested on day 3 post-electroporation. Cutting
efficiency was detected by Sanger sequencing and ICE-beta. HDR editing efficiency was detected
by both ICE-beta and RFLP assay.

45

References
[1] Cox, D. B. T., Platt, R. J., & Zhang, F. (2015). Therapeutic genome editing: prospects and
challenges. Nature medicine, 21, 121.
[2] Howe, S. J., Mansour, M. R., Schwarzwaelder, K., Bartholomae, C., Hubank, et al., (2008).
Insertional mutagenesis combined with acquired somatic mutations causes leukemogenesis
following gene therapy of SCID-X1 patients. The Journal of clinical investigation, 118.
[3] Hacein-Bey-Abina, S., Garrigue, A., Wang, G. P., Soulier, J., Lim, A., Morillon, E., et al.,
(2008). Insertional oncogenesis in 4 patients after retrovirus-mediated gene therapy of SCID-
X1. The Journal of clinical investigation, 118, 3132-3142.
[4] Prakash, V., Moore, M., & Yáñez-Muñoz, R. J. (2016). Current progress in therapeutic gene
editing for monogenic diseases. Molecular therapy, 24, 465-474.
[5] Mao, Z., Bozzella, M., Seluanov, A., & Gorbunova, V. (2008). Comparison of
nonhomologous end joining and homologous recombination in human cells. DNA repair, 7,
1765-1771.
[6] Tomimatsu, N., Mukherjee, B., Hardebeck, M. C., Ilcheva, M., Camacho, C. V., et al., (2014).
Phosphorylation of EXO1 by CDKs 1 and 2 regulates DNA end resection and repair pathway
choice. Nature communications, 5, 3561.
[7] Wang, H., Shi, L. Z., Wong, C. C., Han, X., Hwang, P. Y. H., et al., (2013). The interaction
of CtIP and Nbs1 connects CDK and ATM to regulate HR–mediated double-strand break
repair. PLoS genetics, 9, e1003277.
[8] Orthwein, A., Noordermeer, S. M., Wilson, M. D., Landry, S., Enchev, R. I., et al., (2015).
A mechanism for the suppression of homologous recombination in G1 cells. Nature, 528, 422.
[9] Karanam, K., Kafri, R., Loewer, A., & Lahav, G. (2012). Quantitative live cell imaging
reveals a gradual shift between DNA repair mechanisms and a maximal use of HR in mid S
phase. Molecular cell, 47, 320-329.
[10] Elliott, B., Richardson, C., Winderbaum, J., Nickoloff, J. A., & Jasin, M. (1998). Gene
conversion tracts from double-strand break repair in mammalian cells. Molecular and cellular
biology, 18, 93-101.
[11] Beumer, K. J., Trautman, J. K., Mukherjee, K., & Carroll, D. (2013). Donor DNA utilization
during gene targeting with zinc-finger nucleases. Genes, Genomes, Genetics.
[12] LaRocque, J. R., & Jasin, M. (2010). Mechanisms of recombination between diverged
sequences in wild-type and BLM-deficient mouse and human cells. Molecular and cellular
biology, 30, 1887-1897.
46

[13] Brenneman, M. A., Wagener, B. M., Miller, C. A., Allen, C., Nickoloff, J. A., et al., (2002).
XRCC3 controls the fidelity of homologous recombination: roles for XRCC3 in late stages
of recombination. Molecular cell, 10, 387-395.
[14] Chang, H. H., Pannunzio, N. R., Adachi, N., & Lieber, M. R. (2017). Non-homologous DNA
end joining and alternative pathways to double-strand break repair. Nature reviews Molecular
cell biology, 18, 495.
[15] Li, G., Zhang, X., Zhong, C., Mo, J., Quan, R., Yang, J., & Wu, Z. (2017). Small molecules
enhance CRISPR/Cas9-mediated homology-directed genome editing in primary cells.
Scientific reports, 7, 8943.
[16] Tavecchio, M., Munck, J. M., Cano, C., Newell, D. R., & Curtin, N. J. (2012). Further
characterisation of the cellular activity of the DNA-PK inhibitor, NU7441, reveals potential
cross-talk with homologous recombination. Cancer chemotherapy and pharmacology, 69,
155-164.
[17] Canny, M. D., Moatti, N., Wan, L. C., Fradet-Turcotte, A., Krasner, D., et al., (2018).
Inhibition of 53BP1 favors homology-dependent DNA repair and increases CRISPR–Cas9
genome-editing efficiency. Nature biotechnology, 36(1), 95.
[18] Symington, L. S. (2014). End resection at double-strand breaks: mechanism and
regulation. Cold Spring Harbor perspectives in biology, 6, a016436.
[19] Carreira, A., & Kowalczykowski, S. C. (2011). Two classes of BRC repeats in BRCA2
promote RAD51 nucleoprotein filament function by distinct mechanisms. Proceedings of the
National Academy of Sciences, 108, 10448-10453.
[20] Ceccaldi, R., Rondinelli, B., & D’Andrea, A. D. (2016). Repair pathway choices and
consequences at the double-strand break. Trends in cell biology, 26, 52-64.
[21] Wyatt, D. W., Feng, W., Conlin, M. P., Yousefzadeh, M. J., Roberts, S. A., et al., (2016).
Essential roles for polymerase θ-mediated end joining in the repair of chromosome breaks.
Molecular cell, 63(4), 662-673.
[22] Taheri-Ghahfarokhi, A., Taylor, B. J. M., Nitsch, R., Lundin, A., Cavallo, A. L., et al., (2018).
Decoding non-random mutational signatures at Cas9 targeted sites. Nucleic acids
research, 46, 8417-8434
[23] Truong, L. N., Li, Y., Shi, L. Z., Hwang, P. Y. H., He, J., et al., (2013). Microhomology-
mediated End Joining and Homologous Recombination share the initial end resection step to
repair DNA double-strand breaks in mammalian cells. Proceedings of the National Academy
of Sciences, 110, 7720-7725
47

[24] Brinkman, E. K., Chen, T., de Haas, M., Holland, H. A., Akhtar, W., & van Steensel, B.
(2018). Kinetics and fidelity of the repair of Cas9-induced double-strand DNA breaks.
Molecular cell, 70, 801-813.
[25] Jayavaradhan, R., Pillis, D. M., & Malik, P. (2019). A versatile tool for the quantification of
CRISPR/Cas9-induced genome editing events in human hematopoietic cell lines and
hematopoietic stem/progenitor cells. Journal of molecular biology, 431, 102-110.
[26] Biehs, R., Steinlage, M., Barton, O., Juhász, S., Künzel, J., Spies, J., et al., (2017). DNA
double-strand break resection occurs during non-homologous end joining in G1 but is distinct
from resection during homologous recombination. Molecular cell, 65, 671-684.
[27] Löbrich, M., Rydberg, B., & Cooper, P. K. (1995). Repair of x-ray-induced DNA double-
strand breaks in specific Not I restriction fragments in human fibroblasts: joining of correct
and incorrect ends. Proceedings of the National Academy of Sciences, 92, 12050-12054.
[28] Pham, Q. T., Bouchard, A., Grütter, M. G., & Berthoux, L. (2010). Generation of human
TRIM5α mutants with high HIV-1 restriction activity. Gene therapy, 17, 859.
[29] Jung, U., Urak, K., Veillette, M., Nepveu-Traversy, M. É., Pham, Q. T., et al. (2015).
Preclinical assessment of mutant human TRIM5α as an anti-HIV-1 transgene. Human gene
therapy, 26, 664-679.
[30] Nakayama, E. E., & Shioda, T. (2015). Impact of TRIM5α in vivo. AIDS (London, England),
29, 1733.
[31] Mateos-Gomez, P. A., Gong, F., Nair, N., Miller, K. M., et al., (2015). Mammalian
polymerase θ promotes alternative NHEJ and suppresses recombination. Nature, 518, 254.
[32] van Overbeek, M., Capurso, D., Carter, M. M., et al., (2016). DNA repair profiling reveals
nonrandom outcomes at Cas9-mediated breaks. Molecular cell, 63, 633-646.
[33] Antony, J. S., Haque, A. A., Lamsfus‐Calle, A., Daniel‐Moreno, A., et al., (2018).
CRISPR/Cas9 system: A promising technology for the treatment of inherited and neoplastic
hematological diseases. Advances in Cell and Gene Therapy, 1, e10.
[34] Liang, X., Potter, J., Kumar, S., Ravinder, N., & Chesnut, J. D. (2017). Enhanced
CRISPR/Cas9-mediated precise genome editing by improved design and delivery of gRNA,
Cas9 nuclease, and donor DNA. Journal of biotechnology, 241, 136-146.
[35] Richardson, C. D., Ray, G. J., DeWitt, M. A., Curie, G. L., & Corn, J. E. (2016). Enhancing
homology-directed genome editing by catalytically active and inactive CRISPR-Cas9 using
asymmetric donor DNA. Nature biotechnology, 34, 339.
[36] Hollywood, J. A., Lee, C. M., Scallan, M. F., & Harrison, P. T. (2016). Analysis of gene
repair tracts from Cas9/gRNA double-stranded breaks in the human CFTR gene. Scientific
reports, 6, 32230.
48

[37] Sebesta, M., & Krejci, L. (2016). Mechanism of Homologous Recombination. In DNA
Replication, Recombination, and Repair (pp. 73-109). Springer, Tokyo.
[38] Uringa, E. J., Youds, J. L., Lisaingo, K., Lansdorp, P. M., & Boulton, S. J. (2010). RTEL1:
an essential helicase for telomere maintenance and the regulation of homologous
recombination. Nucleic acids research, 39, 1647-1655.
[39] Paliwal, S., Kanagaraj, R., Sturzenegger, A., Burdova, K., & Janscak, P. (2013). Human
RECQ5 helicase promotes repair of DNA double-strand breaks by synthesis-dependent strand
annealing. Nucleic acids research, 42, 2380-2390.
[40] Wright, W. D., & Heyer, W. D. (2014). Rad54 functions as a heteroduplex DNA pump
modulated by its DNA substrates and Rad51 during D loop formation. Molecular cell, 53,
420-432.
[41] Brenneman, M. A., Wagener, B. M., Miller, C. A., Allen, C., & Nickoloff, J. A. (2002).
XRCC3 controls the fidelity of homologous recombination: roles for XRCC3 in late stages
of recombination. Molecular cell, 10, 387-395.
[42] Wu, Y., Lee, S. H., Williamson, E. A., Reinert, B. L., et al., (2015). EEPD1 rescues stressed
replication forks and maintains genome stability by promoting end resection and homologous
recombination repair. PLoS genetics, 11, e1005675.
[43] Richardson, C. D., Kazane, K. R., Feng, S. J., Zelin, E., Bray, N. L., et al., (2018). CRISPR–
Cas9 genome editing in human cells occurs via the Fanconi anemia pathway. Nature genetics,
50, 1132.
[44] Gasiunas, G., Barrangou, R., Horvath, P., & Siksnys, V. (2012). Cas9–crRNA
ribonucleoprotein complex mediates specific DNA cleavage for adaptive immunity in
bacteria. Proceedings of the National Academy of Sciences, 109, E2579-E2586.
[45] Zuo, Z., & Liu, J. (2016). Cas9-catalyzed DNA cleavage generates staggered ends: Evidence
from molecular dynamics simulations. Scientific reports, 6, 37584.
[46] Taheri-Ghahfarokhi, A., Taylor, B. J. M., Nitsch, R., Lundin, A., et al., (2018). Decoding
non-random mutational signatures at Cas9 targeted sites. Nucleic acids research, 46, 8417-
8434.
[47] Clouaire, T., Legube, G. (2019). A Snapshot on the Cis Chromatin Response to DNA Double-
Strand Breaks. Trends in Genetics.
[48] Roos, W. P., Kaina, B. (2006). DNA damage-induced cell death by apoptosis. Trends in
molecular medicine, 12(9), 440-450.
[49] Law, J. C., Ritke, M. K., Yalowich, J. C., Leder, G. H., et al., (1993). Mutational inactivation
of the p53 gene in the human erythroid leukemic K562 cell line. Leukemia research, 17, 1045-
1050.
49

[50] Kawamura, K., Bahar, R., Seimiya, M., Chiyo, M., Wada, A., et al., (2004). DNA polymerase
θ is preferentially expressed in lymphoid tissues and upregulated in human cancers.
International journal of cancer, 109, 9-16.
[51] Bothmer, A., Phadke, T., Barrera, L. A., Margulies, C. M., Lee, C. S., et al., (2017).
Characterization of the interplay between DNA repair and CRISPR/Cas9-induced DNA
lesions at an endogenous locus. Nature communications, 8, 13905.
[52] Liu, Y., Burness, M. L., Martin-Trevino, R., Guy, J., Bai, S., Harouaka, R., et al., (2017).
RAD51 mediates resistance of cancer stem cells to PARP inhibition in triple-negative breast
cancer. Clinical Cancer Research, 23, 514-522.
[53] Mateos-Gomez, P. A., Gong, F., Nair, N., Miller, K. M., et al., (2015). Mammalian
polymerase θ promotes alternative NHEJ and suppresses recombination. Nature, 518, 254.
[54] Murovec, J., Pirc, Ž., Yang, B. (2017). New variants of CRISPR RNA‐guided genome editing
enzymes. Plant biotechnology journal, 15, 917-926.
[55] Zhang, Y., Long, C., Li, H., McAnally, J. R., Baskin, K. K., et al., (2017). CRISPR-Cpf1
correction of muscular dystrophy mutations in human cardiomyocytes and mice. Science
advances, 3, e1602814.

Abstract (if available)

Abstract Gene editing has great potential to cure many genetic and infectious diseases through the modification of specific genome sequences. The process is initiated by the action of targeted nucleases such as CRISPR/Cas9 and ZFN which can create double-strand breaks (DSB) at the desired genomic edit sites. After a DSB is formed, subsequent repair of the DSB can create different gene editing outcomes, such as gene disruption (caused by the non-homologous end joining (NHEJ) pathway) or, in the presence of a homology donor template, precise gene editing (caused by the homology-directed repair (HDR) pathway). In mammalian cells, NHEJ is the dominant repair pathway. Although this prevalence of NHEJ can be beneficial if the desired outcome is gene knock-out, the potential to disrupt a targeted gene becomes a deleterious effect for applications where the desired outcome is HDR-mediated gene editing. Therefore, to limit the impact caused by unwanted NHEJ outcomes in mammalian cells, a better editing approach is needed. ❧ In order to do this, I propose an alternative approach for gene editing which I call “editing at-a-distance”. Instead of creating a DSB close to the intended edit site, the nuclease-induced DSB is instead created in a nearby intron, and HDR repair proceeds to edit the gene at the more distant site. I hypothesize that even if cells repair the DSB by NHEJ, the less critical sequences in an intron will mean that the effect is more tolerable. However, the major challenge of this idea is that limited studies have shown that HDR-mediated editing efficiency drops dramatically as the nuclease-induced DSB is separated from the intended edit site. Moreover, editing can also be polarized, with one direction being favored. Therefore, to make editing at-a-distance more practical, I hypothesize that it is essential to understand the “landscape of repair” ie how HDR efficiency varies with distance from a DSB. My result shows that the landscape of HDR-mediated editing efficiency can be asymmetric and vary both upstream and downstream from the DSB. Also, manipulating the expression of several HR-related factors can enhance long-tract HDR-mediated editing efficiency in both Cas9 and ZFN system. In addition, since at a distance editing relies on long tract processing, I also hypothesize that it may be possible to enhance this process by manipulating the DNA repair factors involved in HDR. ❧ Also, in this research, I evaluated the characteristics of gRNAs that made them more or less suitable to promote HDR-mediated gene editing. Traditionally, an ideal gRNA is chosen based on bulk cutting efficiency, because it is expected that higher rates of DSB formation will increase the rates of an HDR outcome. However, unpublished data shows that this current approach may not be optimal. I propose instead that the indel signature for a gRNA can predict HDR-mediated editing efficiency. My results show that indel signatures for each gRNA are consistent across different cell types, including in primary cells. Importantly, a gRNA producing fewer NHEJ-mediated outcomes leads to higher HDR-mediated editing efficiency when combined with a homology donor. With these findings, the indel signature should be considered when selecting an ideal gRNA for HDR-mediated gene editing.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Rational selection of CRISPR/Cas9 guide RNAs for homology directed genome editing and its utility in the development of gene therapies

PDF

Engineering B cells to express fully customizable therapeutic antibodies with enhanced Fc functions

PDF

Design and characterization of multiplex anti-HIV single domain antibodies for genome editing of the immunoglobulin locus

PDF

Current models of non-homologous end joining and their implications in gene therapy

PDF

Improving adeno-associated viral vector for hematopoietic stem cells gene therapy

PDF

Mechanisms of nucleases in non-homologous DNA end joining

PDF

Identification of molecular mechanism for cell-fate decision in liver; &, SARS-CoV replicon inhibitor high throughput drug screening

PDF

Mechanistic basis for chromosomal translocations at the E2A gene

Asset Metadata

Creator Huang, Chun (author)

Core Title Characterizing and manipulating homology-directed gene editing in human cells

School Keck School of Medicine

Degree Master of Science

Degree Program Molecular Microbiology and Immunology

Publication Date 07/11/2019

Defense Date 05/21/2019

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

Tag CRISPR/Cas9,editing efficiency,indel outcome,MMEJ,NHEJ: HDR,OAI-PMH Harvest,precise gene editing

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Cannon, Paula (committee chair), Comai, Lucio (committee member), Machida, Keigo (committee member)

Creator Email chunhuan@usc.edu,jeff91741@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c89-182378

Unique identifier UC11660409

Identifier etd-HuangChun-7541.pdf (filename),usctheses-c89-182378 (legacy record id)

Legacy Identifier etd-HuangChun-7541.pdf

Dmrecord 182378

Document Type Thesis

Format application/pdf (imt)

Rights Huang, Chun

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