University of Southern California Dissertations and Theses

The function of Rpd3 in balancing the replicaton initiation of different genomic regions

(USC Thesis Other)

The function of Rpd3 in balancing the replicaton initiation of different genomic regions

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content i
THE FUNCTION OF RPD3 IN BALANCING THE REPLICATION
INITIATION OF DIFFERENT GENOMIC REGIONS
by
Yiwei He
A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(MOLECULAR BIOLOGY)
December 2023
ii
ACKNOWLEDGEMENTS

Pursuing a Ph.D. has been both challenging and transformative for me. At the heart of
this experience is a network of individuals whose support, guidance, and encouragement
propelled me forward. First and foremost, I would like to express my profound gratitude to my
mentor, Dr. Oscar Aparicio. His deep passion for science and genuine kindness have been a
guiding light throughout my journey. Under his mentorship, I have learned not just how to
conduct research, but also how to evolve as a person. I will always remember the time we were
working with radioactively labeled DNA. Recognizing my apprehension, Dr. Aparicio took the
initiative to handle potentially hazardous materials, emphasizing the importance of
environmental responsibility even in minor details. His dedication to both the rigors of research
and being a caring and eco-conscious individual has left an indelible mark on me. Joining the
Aparicio Lab is a decision I have never regretted, and the memories and lessons gained will be
cherished for a lifetime. Secondly, I want to express my gratitude to my dissertation committee
members: Dr. Irene Chiolo, Dr. Lin Chen, Dr. Ian Ehrenreich, and Dr. Pinghui Feng. A special
mention goes to Dr. Chiolo, who saw potential in me that I often doubted and encouraged me to
present at the SoCal Genome Stability Symposium. Her encouragement bolstered my
confidence, making that presentation one of my most memorable achievements. Dr. Lin Chen
provided critical insights into the second chapter of my dissertation. I sincerely thank each
member of my defense committee for their academic guidance, dedication during my Ph.D.
qualifying exams, progress report meetings, and invaluable feedback on my dissertation.
Special thanks to my senior lab mates, Dr. Zachary Ostrow, Dr. Jared Peace, and Dr.
Sandra Villwock, who made my initial days in the lab both welcoming and educational. Each of
them was kind and went out of their way to help me understand our research and tackle
iii
challenges in my experiments. I spent the majority of my Ph.D. journey working alongside my
peer lab mates, Dr. Meghan Petrie and Dr. Haiyang Zhang. We supported and encouraged each
other, and together we created numerous cherished memories. Beyond our professional
interactions, we also formed close friendships outside the lab. I deeply appreciate their support
and companionship. Our lab manager, Yan Gan, has been the core of our lab's operations. Her
professionalism and patience ensured our projects ran smoothly. Yan is also someone I deeply
respect in this foreign land, and I am grateful for the life lessons she has shared with me. I would
also like to thank Elliot Kim for his great contribution to my computational project.
Next, I want to thank my 2017 cohort members: Dr. Mezmur Belew, Caleb Ghione,
Joseph Hale, Dan Ma, Joshua Park, and Dr. Nicole Stuhr. Your companionship has truly
enriched my PhD experience. Every birthday celebration and every gathering underscored the
sense of community we have built together. I would also like to extend my thanks to the
members of other labs in the RRI building, especially Yingfei Wang, Dr. Chetan Rawal, Dr.
Kujin Tang, and many others who provided me with considerable academic assistance.
Lastly, but most importantly, I'd like to deeply thank my family, especially my parents,
for their unwavering emotional encouragement and vital financial support. Their continuous trust
in me has been the cornerstone of my journey. To my partner, Kai Xue, your consistent faith in
me and endless encouragement have been more than I could ask for. I cherish the time you've
accompanied me through my PhD journey. For my friends, notably Shiyue Huang and Canwen
Wang, the memories we've built during my Ph.D. journey will always be close to my heart. The
good times we've shared have added so much to my experience.

iv
Table of Contents
ACKNOWLEDGEMENTS II
LIST OF FIGURES VI
ABSTRACT VIII
INTRODUCTION 1
REPLICATION ORIGINS AND THE INITIATION OF DNA REPLICATION IN EUKARYOTES………….1
THE TEMPORAL AND SPATIAL REGULATION OF DNA REPLICATION INITIATION………………..3
THE ROLE OF FKH1 AND RPD3 IN REGULATING DNA REPLICATION INITIATION………………. 5
CHROMATIN IMMUNOPRECIPITATION AND ITS CONTROLS……………………………………… 7
CHAPTER I 10
RPD3 REGULATES SINGLE-COPY ORIGINS INDEPENDENTLY OF THE RDNA
ARRAY BY OPPOSING FKH1-MEDIATED ORIGIN STIMULATION 10
RESULTS 13
RPD3 HAS DIFFERENTIAL EFFECTS ON SINGLE-COPY ORIGINS………………………………… 13
SIR2 DELETION IS EPISTATIC TO RPD3 DELETION FOR RDNA ORIGIN FIRING……………….. 17
RPD3 REGULATES SINGLE-COPY ORIGINS INDEPENDENTLY OF THE RDNA ARRAY…………... 20
RPD3 SUPPRESSES PAN-S PHASE RDNA ORIGIN EFFICIENCY………………………………….. 21
ORIGIN DE-REPRESSION IN RPD3∆ CELLS REQUIRES FKH1/2………………………………….. 23
RPD3 MODULATES FKH1 BINDING TO ORIGINS………………………………………………… 25
RPD3 DELETION SUPPRESSES DBF4 DEFICIENCY IMPLICATING THE FKH1-DDK PATHWAY.. 28
DISCUSSION…………………………………………………………………………………... 29
MATERIALS AND METHODS……………………………………………………………… 33
YEAST STRAIN CONSTRUCTION…………………………………………………………………. 33
OTHER METHODS……………………………………………………………………………….. 35
COMPUTATION AND STATISTICS………………………………………………………………... 36
CHAPTER II 37
BROADLY APPLICABLE CONTROL APPROACHES IMPROVE ACCURACY OF
CHIP-SEQ DATA 37
INTRODUCTION…………………………………………………………………………….. 38
RESULTS………………………………………………………………………………………. 40
EXPRESSION OF AN EPITOPE-TAGGED PROTEIN AS NORMALIZATION CONTROL……………. 40
NON-SPECIFIC SIGNALS ARE PERVASIVE IN CHIP……………………………………………. 42
v
RATIO NORMALIZATION BY EPITOPE-TAGGED CONTROL REFINES DATA QUALITY FOR
TARGET PROTEINS……………………………………………………………………………… 45
FKH1 ANALYSIS IN G1 PHASE TO TEST KNOWN ENRICHMENT AT REPLICATION ORIGINS….. 49
ANALYSIS OF REPLICATION ORIGIN BINDING PROTEINS VALIDATE APPROACH FOR HA…… 53
CONTROLS ENHANCE ANALYSIS OF POTENTIAL HYPERCHIPABLE LOCI……………………. 58
DNA-BINDING MUTANT MAY BE IDEAL CONTROL…………………………………………… 59
DISCUSSION…………………………………………………………………………………... 64
EXPRESSION OF A DECOY PROTEIN TO CONTROL FOR NON-SPECIFIC SEQUENCE
ENRICHMENT……………………………………………………………………………………. 64
MULTIPLE FACTORS CONTRIBUTE TO NON-SPECIFIC CHIP-SIGNAL ENRICHMENT………… 65
NO CONTROL, NO EXPERIMENT……………………………………………………………….. 66
MATERIALS AND METHODS……………………………………………………………… 67
PLASMID CONSTRUCTIONS…………………………………………………………………….. 67
YEAST STRAIN CONSTRUCTIONS………………………………………………………………. 67
OTHER METHODS………………………………………………………………………………. 68
APPENDIX 70
REFERENCES 74
SUPPLEMENTAL FIGURES AND TABLES 88

vi
LIST OF FIGURES
CHAPTER I
FIGURE 1. 1. RPD3 AND SIR2 HAVE DIFFERENT EFFECTS ON SINGLE-COPY ORIGINS. ................ 15
FIGURE 1. 2. SIR2 DELETION IS EPISTATIC TO RPD3 DELETION FOR RDNA ORIGIN FIRING. ...... 18
FIGURE 1. 3. RPD3 SUPPRESSES RDNA ORIGIN FIRING. ........................................................................ 22
FIGURE 1. 4. ORIGIN DE-REPRESSION IN RPD3∆ CELLS REQUIRES FKH1/2. .................................... 25
FIGURE 1. 5. RPD3 MODULATES FKH1 BINDING TO ORIGINS. ............................................................. 27
FIGURE 1. 6. RPD3 DELETION SUPPRESSES DBF4 DEFICIENCY. .......................................................... 29

CHAPTER II
FIGURE 2. 1. CONTROL CONSTRUCTS AND ANALYSIS SCHEME. ....................................................... 40
FIGURE 2. 2. ENRICHMENT OF NON-SPECIFIC CHIP SIGNALS IN CONTROL IPS. ............................ 44
FIGURE 2. 3. RATIO NORMALIZATION USING CONTROL REDUCES NON-SPECIFIC
ENRICHMENT. ......................................................................................................................................... 48
FIGURE 2. 4. IMPROVED DETECTION OF FKH1 BINDING LOCI. ........................................................... 52
FIGURE 2. 5. IMPROVED ACCURACY IN DETECTION OF ORC AND MCM BINDING LOCI. ............ 56
FIGURE 2. 6. SIR2 DETECTION WITHIN HYPERCHIPABLE RDNA LOCUS AND GENOME-WIDE. . 61
FIGURE 2. 7. DNA-BINDING MUTANT CONTROL MAY BE IDEAL. ...................................................... 63

APPENDIX
FIGURE A 1. THE DEMONSTRATION OF THE “CONSECUTIVE-BIN METHOD” IN PEAK
DETECTION. ............................................................................................................................................ 71
FIGURE A 2. THE DEMONSTRATION OF THE HMM MODEL IN PEAK DETECTION. ........................ 72
FIGURE A 3. THE COMPARISON OF THE PEAK DETECTION RESULTS BETWEEN THE TWO
APPROACHES. ......................................................................................................................................... 73

SUPPLEMENTAL FIGURES AND TABLES
FIGURE S1. 1. DNA REPLICATION ANALYSIS OF WT, RPD3∆, SIR2∆, AND RPD3∆ SIR2∆ STRAINS.
.................................................................................................................................................................... 88
FIGURE S1. 2. ANALYSIS OF RDNA COPY NUMBERS AND QUANTIFICATION OF 2D GELS. ........ 89
FIGURE S1. 3. RPD3 DELETION DE-REPRESSES SIMILAR ORIGINS WITH OR WITHOUT HU. ...... 89
FIGURE S1. 4. FKH1 OVEREXPRESSION STIMULATES FIRING OF RPD3-REPRESSED ORIGINS. . 90
FIGURE S1. 5. MCM4 ORIGIN ASSOCIATION UNAFFECTED BY RPD3 DELETION. .......................... 91

FIGURE S2. 1. SUPPLEMENT TO FIGURE 1. .............................................................................................. 101
FIGURE S2. 2. SUPPLEMENT TO FIGURE 2. .............................................................................................. 102
FIGURE S2. 3. SUPPLEMENT TO FIGURE 3. .............................................................................................. 103
FIGURE S2. 4. SUPPLEMENT TO FIGURE 4. .............................................................................................. 104
FIGURE S2. 5. SUPPLEMENT TO FIGURE 5. .............................................................................................. 105
FIGURE S2. 6. SUPPLEMENT TO FIGURE 6. .............................................................................................. 106
FIGURE S2. 7. SUPPLEMENT TO FIGURE 7. .............................................................................................. 107

TABLE 1. 1. SEQUENCES OF DNA OLIGONUCLEOTIDES USED IN THIS STUDY. 92
vii
TABLE 1. 2. GENOTYPES OF S. CEREVISIAE STRAINS USED IN THIS STUDY. 93
TABLE S1. 1. LIST OF RPD3-REPRESSED ORIGINS. .................................................................................. 95
TABLE S1. 2. LIST OF CEN-PROXIMAL ORIGINS. ..................................................................................... 96
TABLE S1. 3. LIST OF FKH1-OE-ACTIVATED ORIGINS. ........................................................................ 100

TABLE S2. 1. RESULTS OF STATISTICAL ANALYSIS OF DATA PRESENTED IN FIGURE 2. .......... 108
TABLE S2. 2. RESULTS OF STATISTICAL ANALYSIS OF DATA PRESENTED IN FIGURE 3. .......... 108
TABLE S2. 3. SEQUENCES OF DNA OLIGONUCLEOTIDES USED IN THIS STUDY. ......................... 111
TABLE S2. 4. GENOTYPES OF S. CEREVISIAE STRAINS USED IN THIS STUDY. ............................. 112

viii
ABSTRACT

Eukaryotic chromosomes are organized into structural and functional domains with
characteristic replication timings, which are thought to contribute to epigenetic programming and
genome stability. Differential replication timing results from epigenetic mechanisms that
positively and negatively regulate the competition for limiting replication initiation factors. In
the budding yeast Saccharomyces cerevisiae, histone deacetylase Sir2 negatively regulates
initiation of the multi-copy (~150) rDNA origins while Rpd3 histone deacetylase negatively
regulates the firing of single-copy origins. However, Rpd3’s effect on single-copy origins might
derive indirectly from a positive function for Rpd3 in rDNA origin firing shifting the competitive
balance. Our quantitative experiments support the idea that origins compete for limiting factors;
however, our results show that Rpd3’s effect on single-copy origin is independent of rDNA
number and of Sir2’s effects on rDNA origin firing. Whereas RPD3 deletion and SIR2 deletion
alter the early S phase dynamics of single-copy and rDNA origin firings in an opposite fashion,
unexpectedly only RPD3 deletion suppresses overall rDNA origin efficiency across the S phase.
Increased origin activation in rpd3∆ requires Fkh1/2 suggesting that Rpd3 opposes Fkh1/2-origin
stimulation, which involves the recruitment of Dbf4-dependent kinase (DDK). Indeed, Fkh1
binding increases at Rpd3-regulated origins in rpd3∆ cells in G1, supporting a mechanism
whereby Rpd3 influences the initiation timing of single-copy origins directly through modulation
of Fkh1-origin binding. Genetic suppression of a DBF4 hypomorphic mutation by RPD3
deletion further supports the conclusion that Rpd3 impedes DDK recruitment by Fkh1, revealing
a mechanism of Rpd3 in origin regulation.
In the study mentioned above, we extensively utilized the Chromatin
Immunoprecipitation followed by sequencing (ChIP-Seq) technique, particularly when
ix
investigating whether Rpd3 affects Fkh1-origin binding. In fact, ChIP-Seq is a widely used
technique for the analysis of protein-DNA interactions in vivo. However, ChIP has pitfalls,
particularly false-positive signal enrichment that permeates the data. We have developed a new
approach to control for non-specific enrichment in ChIP that involves the expression of a non-
genome-binding protein targeted in the IP alongside the experimental target protein due to the
sharing of epitope tags. ChIP of the protein provides a “sensor” for non-specific enrichment that
can be used for the normalization of the experimental data, thereby correcting for non-specific
signals and improving data quality as validated against known binding sites for several proteins
that we tested, including Fkh1, Orc1, Mcm4, and Sir2. We also tested a DNA binding mutant
approach and showed that, when feasible, ChIP of a site-specific DNA binding mutant of the
target protein is likely an ideal control. These methods vastly improve our ChIP-seq results in S.
cerevisiae and should be applicable in other systems.

1

INTRODUCTION

Replication origins and the initiation of DNA replication in eukaryotes

The fundamental nature of life and its continuous propagation lies in the faithful
replication of the genome, a crucial process that ensures genomic integrity throughout all
organisms. A deep comprehension of the regulatory mechanisms of the DNA replication process
is not only essential for advancing our knowledge of cellular biology, but it also offers insights
into deciphering the complex profiles of diseases associated with genomic instability. The
regulation of replication timing, in particular, is especially critical, given its role in ensuring the
accurate and flawless replication of genetic material. Indeed, studies have indicated that the
abnormalities in DNA replication timing could potentially be associated with a range of
pathological conditions, emphasizing the importance of further investigation in this field
(Hiratani & Gilbert, 2009; Rhind & Gilbert, 2013).
To ensure the faithful replication of their genetic material, where the entire genome is
duplicated just once in each cell cycle, eukaryotes have developed a tightly controlled and cell-
cycle-dependent mechanism for the initiation of DNA replication. In the S. cerevisiae,
replication initiation takes place at discrete chromosomal loci, first discovered as Autonomously
Replicating Sequences (ARSs) (Stinchcomb et al., 1979). The incorporation of these 100 - 200
base pair sequences into plasmids enabled their mitotic propagation and thus can be maintained
in a cell. Subsequent studies on ARS within their native chromosomal locations validated their
function as the replication origins (Cvetic & Walter, 2005; Newlon & Theis, 1993). More precise
sequence analyses revealed that ARS is characterized by a conserved A/T-rich element, the ARS
consensus sequence (ACS), and several additional components identified as B1 through B4
2
elements (Rao & Stillman, 1995). The ACS, facilitated by the B1 element, serves as a docking
site for the Origin Recognition Complex (ORC), a six-subunit complex that identifies and binds
to the ACS, thus marking the potential sites of the replication origins across the genome (Rao &
Stillman, 1995). As the cell transitions from M to the early G1 phase, the ORC binding enables
the assembly of the pre-replicative complex (pre-RC) by the recruitment of two replication
factors, Cdc6 and Cdt1, which subsequently facilitate the loading of the minichromosome
maintenance (Mcm2-7, or MCM) helicase complex in its inactive form. The process of loading
the MCM is referred to as origin “licensing”, as it essentially readies the origin for the impending
DNA replication (Bell & Dutta, 2002). During the transit from the G1 to S phase, Cyclin-
dependent kinases (CDKs) and Dbf4-dependent kinases (DDKs) are activated and the
coordinated action of CDKs and DDKs leads to a cascade of events (Labib, 2010). CDK-
mediated phosphorylation of Cdc6 leads to its inactivation and subsequent degradation,
preventing the reloading of the MCM complex onto already licensed origins and helping
maintain the strict regulation of DNA replication, ensuring that each origin is replicated only
once per cell cycle. While DDK, a serine-threonine kinase composed of Cdc7 kinase and its
regulatory subunit Dbf4, plays a crucial role in the activation of the licensed origins by activating
previously loaded MCM complex by the phosphorylation of the Mcm4 and Mcm6 subunits.
Subsequently, Cdc45 and Sld3 are recruited to the origins. CDK also phosphorylates Sld3 and
Sld2, leading to the recruitment of GINS. Upon initiation, the MCM complex acts as an ATP-
dependent DNA helicase to promote the bidirectional progression of replication forks, with
additional factors, including DNA polymerases to assemble a complete replisome and initiate
DNA synthesis. Replisomes progress along parental DNA until they encounter converging forks
(O. M. Aparicio, 2013; Heller et al., 2011; Zou & Stillman, 2000).
3
The temporal and spatial regulation of DNA replication initiation.
Interestingly, despite the identification of over 800 potential replication origins in the S.
cerevisiae genome, only about 500 are utilized during a given S phase, and even among the
origins that do fire, they do not initiate simultaneously. Replication origins in close proximity to
centromeres tend to fire early to facilitate the timely kinetochore assembly, while origins located
in the telomeric and subtelomeric regions, as well as those associated with the heterochromatic
silencers, are typically late firing and/or remain dormant (Dubey et al., 1991; Ferguson &
Fangman, 1992; Friedman et al., 1996). Most of the dormant origins are not used in normal
growth conditions and are “passively” replicated by replication forks from the neighboring active
origins. However, normally dormant origins can be triggered when DNA replication is
challenged with genotoxic drugs and in response to checkpoint signals or when the replication
fork encounters obstacles in passing through higher-order chromatin structures (Ge et al., 2007;
Ibarra et al., 2008). Origin redundancy appears therefore an important mechanism preventing
genomic instability during S phase. In other chromosomal regions, origins demonstrate a wide
range of initiation timings and efficiencies (Heun et al., 2001; Yabuki et al., 2002). In fact,
chromosome conformation capture experiments revealed that chromosomes are organized into
distinct self-interacting domains, referred to as topologically associated domains (TADs), which
can range in size from hundreds to thousands of kilobases. Chromosomal regions within a TAD
tend to have more contacts within the domain than across domain boundaries, and origins within
a TAD are much more likely to fire synchronously than origins in different TADs (Zhao et al.,
2017).
To achieve the synchronized activation of hundreds of origins, each with its characteristic
timings and efficiencies, the regulation of replication initiation is governed by a multi-layered
4
temporal and spatial mechanism. In S. cerevisiae, over-expression of certain replication initiation
factors, particularly those that are DDK-dependent, such as Cdc45 plus its binding partners Sld3
and Sld7, leads to accelerated timing of some origins, which suggests that these factors are
present in limited quantities and are thus referred as limiting initiation factors (Lynch et al.,
2019; Mantiero et al., 2011). The preferential recruitment of limiting factors to specific origins
leads to advanced timing, while other licensed origins remain uninitiated. As replication
progresses, two replication forks originating from flanking early-firing origins meet and
terminate the replication, allowing the limiting factors to be recycled and redirected toward other
licensed origins that are yet to initiate replication. As a consequence, the establishment of
replication initiation timing is influenced by the accessibility of limiting factors to the origins
within a cell, and its regulation is mediated by factors such as chromatin structure, chromosomal
location, and epigenetic modifications, which collectively contribute to the precise control of
replication timing (O. M. Aparicio, 2013).
Euchromatic regions, characterized by a more open chromatin structure, provide better
access to limiting factors, thereby favoring earlier initiation. Conversely, heterochromatin
regions distinguished by their more condensed chromatin structure, present obstacles to the
accessibility of limiting factors, resulting in postponed replication initiation. Serial ARS
relocation experiments showed that the relocation of the early-firing origin ARS1 from its native
centromeric proximal position to a heterochromatic, subtelomeric region proximal to the late-
firing origin ARS501 resulted in a delay in its replication timing. In the reciprocal experiment,
when ARS501 was translocated to a plasmid devoid of its surrounding chromatin context, its
timing was accelerated (Ferguson & Fangman, 1992). In fact, the dynamic relocation of
replication origins may occur within a cell during replication. Early-firing origins could cluster
5
into the replication foci and be relocated to the interior of the cell where limiting initiation
factors are more concentrated. As replication progresses, unfired origins might be relocated
closer to the replication foci, potentially capturing limiting factors as these are released from
terminating early replicons. Eventually, telomeres and other sequences associated with the
nuclear periphery are released from the nuclear periphery to gain access to replication factors
(Duan et al., 2010; Kitamura et al., 2006; Meister et al., 2007). Other studies found that the
deletion of the histone deacetylase Rpd3L results in the earlier activation of approximately one-
third of active origins in the yeast genome. Conversely, the tethering of the histone acetylase
GCN5 promotes the earlier initiation of a nearby origin (J. G. Aparicio et al., 2004; S. R. V.
Knott et al., 2009; Vogelauer et al., 2002). Additionally, the deletion of the gene responsible for
producing the silencing protein SIR3 leads to the advanced timing of sub-telomeric origins.
These findings highlight the influence of chromosomal location, the local chromatin
environment, and epigenetic modifications, on the temporal regulation of origin firing
(Stevenson & Gottschling, 1999).

The role of Fkh1 and Rpd3 in regulating DNA replication initiation

In the S. cerevisiae, the forkhead box (Fox) transcription factors, Forkhead 1 (Fkh1) and
Forkhead 2 (Fkh2), are well-known for their regulatory role over the G2/M phase-specific
transcription of a group of genes known as the CLB2 cluster (Murakami et al., 2010). Expanding
upon their known functions, our lab discovered additional functions of Fkh1 in the regulation of
replication initiation, especially in advancing the timing of early-firing origins (in the absence of
Fkh1, Fkh2 can also compensate for this regulatory role) (S. R. V. Knott et al., 2012). During the
G1 phase, Fkh1 stimulates the DDK-dependent step in replication initiation, resulting in the
6
relocation of specific replication origins away from the nuclear periphery. This repositioning
consequently leads to the formation of distinct spatial clusters of early-firing origins, separating
from the areas where late-firing origins reside. Enriched at these TAD boundaries, Fkh1
effectively enhances the interactions among origins within these chromosomal territories. The
congregation of early-firing origins increases the probability of physical interactions with
limiting initiation factors, such as Dbf4, Sld3, and Cdc45, significantly enhancing the efficiency
of DNA replication. Indeed, Fkh1 facilitate the recruitment of Dbf4 to early-firing origins
dispersed across chromosome arms, an action that propels the early replication of centromere-
distal regions (S. R. V. Knott et al., 2012; Ostrow et al., 2017; Zhang et al., 2019).
Significant challenges can arise during the replication of functionally distinct
chromosomal regions, notably when replicating the highly expressed and repeated DNA
sequences such as the ribosomal DNA genes (rDNA). In the S. cerevisiae, the rDNA gene cluster
is constituted of approximately 150 tandem repeats. Each repeat is around 9.1 Kb in length and
contains one potential origin, leading to a higher origin density in the rDNA than that observed
in other genomic regions (Nomura, 2001). In balancing the replication between rDNA and other
unique genomic regions, it has been suggested that the histone deacetylase Sir2 plays a crucial
role in suppressing unnecessary origin firing within rDNA repeats, allowing only ~20% of the
origins to be activated in a given S phase, preventing excessive replication and potential genome
instability (Pasero et al., 2002). The histone deacetylase Rpd3 is known for its role as a
transcriptional regulator. The binding of Ume6, a sequence-specific DNA-binding transcriptional
repressor, recruits Rpd3 to the promoters of these targeted genes. This recruitment consequently
prompts a remodeling of the chromatin structure as Rpd3 removes acetyl groups from
surrounding histones, making it more condensed and less accessible to the transcription
7
machinery, thus repressing the transcriptional activity of genes in the affected region. Rpd3 is
also associated with the delayed initiation of late-firing, single-copy origins dispersed throughout
the genome. Deletion of RPD3 leads to the advanced firing of many late origins, while
concurrently increasing histone acetylation levels at these sites. In support of the notion that
histone acetylation regulates initiation timing, targeting the Gcn5 histone acetylase to a late
replication origin resulted in an earlier firing time. However, specific recruitment of Rpd3 to
origins has not been demonstrated, so it has been suggesting that Rpd3 might function at single-
copy origins through a nontargeted/global interaction mechanism (J. G. Aparicio et al., 2004; S.
R. V. Knott et al., 2009; Vogelauer et al., 2002).
A contemporary model proposed that Sir2 and Rpd3 distinctly govern the initiation of
rDNA origins, with Sir2 functioning as an inhibitor and Rpd3 acting as an activator. The
influence of Rpd3 on single-copy origins might be an indirect outcome of the competition
between single-copy and rDNA origins for limiting initiation factors (Yoshida et al., 2014). In
our most recent research, we examined the specificity of Rpd3's role in regulating single-copy
origins versus rDNA origins. Our findings align with the concept that origins compete for
limiting factors. However, we suggest that Rpd3’s impact on single-copy origins is independent
of the rDNA copy number. Instead, Rpd3 directly regulates the initiation timing of single-copy
origins through the modulation of Fkh1-origin binding, highlighting a key mechanism of Rpd3 in
origin regulation (Chapter I).

Chromatin immunoprecipitation and its controls

Chromatin Immunoprecipitation followed by sequencing (ChIP-seq) is a widely used
technique for investigating chromatin-associated proteins on a genome-wide scale. As protein-
DNA interactions are fundamental to the regulation of genome functions, ChIP-seq provides
8
valuable insights into gene regulation, epigenetics, and chromatin dynamics. The main workflow
of ChIP-seq involves cross-linking cells or tissues to preserve protein-DNA interactions,
fragmenting chromatin into smaller pieces, immunoprecipitating the protein of interest using
specific antibodies, reversing cross-linking to purify DNA, performing high-throughput
sequencing, and subsequently analyzing the data to identify enriched genomic regions (Nakato &
Sakata, 2021; Park, 2009).
From these experimental steps, it becomes evident that the success of ChIP-seq is highly
dependent on the specificity of antibodies used to target the protein of interest. In practice,
antibodies can occasionally bind non-specifically to chromatin, resulting in false-positive signals
and presenting significant challenges in distinguishing true interactions from false positives. This
non-specific binding may occur due to factors such as loose chromatin structure, which can
inadvertently trap antibodies, and similarities between certain proteins and the protein of interest,
leading to unintended antibody binding. In a previous study, researchers observed that highly
transcribed regions frequently yield signals across unrelated experiments, potentially leading to
false-positive results. This phenomenon, termed "hyper-ChIPability," has raised concerns
regarding potential technical challenges associated with immunoprecipitations (Teytelman et al.,
2013).
To address this issue, several controls have been developed for ChIP-seq, including the
Input DNA control, the Mock IP control, and the Non-specific IP control. However, these
controls have not substantially resolved the problem. Our lab recently developed a novel ChIP-
seq control method, utilizing epitope tag sharing between the protein of interest and a non-
genome-binding protein. We have also evaluated the validity of this approach and demonstrated
9
through practical applications that it improves ChIP-seq accuracy by removing false-positive
signals (Chapter II).

10
CHAPTER I

Rpd3 regulates single-copy origins independently of the rDNA array
by opposing Fkh1-mediated origin stimulation

Adapted from:
He, Y., Petrie, M. V., Zhang, H., Peace, J. M., & Aparicio, O. M. (2022). Rpd3 regulates single-
copy origins independently of the rDNA array by opposing Fkh1-mediated origin
stimulation. Proceedings of the National Academy of Sciences, 119(40), e2212134119.

As the lead author of the following work, I contributed to the experimental design and execution,
data collection and analyses, and all the data visualizations.

11
INTRODUCTION

The replication of functionally distinct chromosomal regions presents challenges to
genome stability. For example, repeated and highly expressed DNA sequences such as the
ribosomal DNA genes appear to be destabilized by transcription-replication collisions, and
unequal sister chromatid exchange or intrachromosomal recombination (Tsang & Carr, 2008). In
S. cerevisiae, the ribosomal DNA (rDNA) gene cluster consists of about 150 direct repeats of the
~9.1kbp rDNA sequence, each containing one potential origin (Nomura, 2001), which represents
a significantly higher origin density than the genome average amongst unique sequences
(MacAlpine & Bell, 2005). Thus, rDNA origins have the potential to impose a significant load
on the replication system and impede the replication of other chromosomal regions, as several
replication initiation factors are present in limiting quantities (O. M. Aparicio, 2013). However,
only ~20% of the rDNA origins fire in each S phase in wild-type (WT) cells (Fangman &
Brewer, 1991; Linskens & Huberman, 1988), and this has been attributed to an intrinsically weak
autonomously replicating sequence (ARS) (Kwan et al., 2013; Miller, 1999), and to suppression
of origin firing by the histone deacetylase Sir2 (Pasero et al., 2002). Deletion of SIR2 results in a
higher proportion of the rDNA origins initiating replication, which has been suggested to
contribute to instability of the rDNA repeats. Histone deacetylase Rpd3, in contrast, has been
implicated in the delayed initiation of late-firing, single-copy origins dispersed throughout the
genome. RPD3 deletion advances the firing of many late origins while increasing the levels of
histone acetylation at these origins (J. G. Aparicio et al., 2004; S. R. V. Knott et al., 2009;
Vogelauer et al., 2002). Targeting of Gcn5 histone acetylase to a late replication origin advanced
its timing, supporting the idea that histone acetylation regulates initiation rate (Vogelauer et al.,
2002). However, specific DNA-binding protein-mediated recruitment of Rpd3 to origins has not
12
been demonstrated (nor ruled out), so Rpd3 has been suggested to act directly at single-copy
origins through an untargeted (or “global”) interaction mechanism (J. G. Aparicio et al., 2004;
Vogelauer et al., 2002).
More recently, however, this model of Rpd3 function has been challenged by a study
claiming that Rpd3 acts directly to stimulate rDNA origins, and that effects of RPD3 deletion on
single-copy origins are an indirect consequence of reduced competition from rDNA origins for a
limited pool of replication initiation factors (Yoshida et al., 2014). Indeed, recent studies have
identified several initiation factors as rate-limiting for origin firing leading to models in which
unequal competition for limiting factors amongst all licensed origins underlies their differential
activities (Mantiero et al., 2011; Patel et al., 2008; Tanaka et al., 2011). In the Yoshida study,
deletion of SIR2 increased early firing of rDNA origins while decreasing early firing of single-
copy origins, consistent with the competition model. Additional deletion of RPD3 suppressed the
effects of SIR2 deletion, consistent with these factors acting in opposition at the rDNA. Thus, it
was concluded that advanced firing of single-copy origins in rpd3∆ cells is a consequence of
reduced Rpd3-stimulated rDNA origin firing competing for limiting factors, rather than a direct
repressive effect of Rpd3 on single-copy origins. A previous study implicating Rpd3 in
transcriptional silencing of the rDNA was also cited as supporting evidence for the idea that
Rpd3 acts directly on rDNA origins (Smith et al., 1999).
The main conclusions of the earlier study relied on quantitative comparisons of
independently generated BrdU-IP-seq datasets with WT, sir2∆, rpd3∆, and sir2∆ rpd3∆ strains.
Such comparisons require signal normalization between the independently produced
experimental datasets, which can be problematic, typically requiring arbitrary, if not dubious,
assumptions about the data. More recently we developed qBrdU-seq (or “QBU”) to overcome
13
this barrier to an unambiguous, quantitative comparison of independent BrdU-IP-seq samples
(Peace et al., 2016). Our method involves sample-specific barcoding of BrdU-labeled genomic
DNA, followed by pooling for immunoprecipitation (IP), amplification and high-throughput
sequencing. Additionally, prior to IP, a small aliquot of the pooled “Input” sample is retained for
amplification and sequencing in parallel, providing a direct reference for normalization of IP
sequencing read counts based on actual input. QBU eliminates and/or corrects for most sources
of technical or experimental error that may hinder (or at least cast doubt upon) quantitative
comparison between individually processed samples.
In this study, we have re-examined the impacts of Rpd3 and Sir2 on genome replication,
with particular focus on the origin competition model and Rpd3 function at single-copy versus
rDNA origins. Our findings support the idea of competition between rDNA and single copy origins,
but not the notion that Rpd3 acts positively through rDNA origins to squelch single-copy origins.
We have also probed more deeply into the mechanism of Rpd3 action at origins and present
evidence that Rpd3 antagonizes single-copy origin firing by regulating Fkh1 binding and hence,
recruitment of DDK to origins. In accord, deletion of RPD3 rescues defective DDK function.

RESULTS

Rpd3 has differential effects on single-copy origins

A rigorous test of the competition model between rDNA and single-copy origins and of the
roles of Rpd3 and Sir2 in regulating these origins requires an unambiguous, quantitative
comparison of origin firing levels. Therefore, we performed QBU analysis of WT, rpd3∆, sir2∆,
and sir2∆ rpd3∆ strains. Cells were synchronized in G1 phase with α-factor and released into S
phase in the presence of hydroxyurea (HU) to block replication and distinguish between early
14
origins, which fire efficiently in HU, and late origins, which do not due to intra-S checkpoint
inhibition (Santocanale & Diffley, 1998). DNA content analysis of cells similarly synchronized
and released into S phase without HU shows similar overall timing of S phase execution for the
four strains, with rpd3∆ cells completing S phase slightly more rapidly than WT as observed
previously (Fig. S1A) (Vogelauer et al., 2002).
Distribution of the QBU values comparing the individual replicates of each strain show
high correlations demonstrating the high experimental reproducibility of the QBU analysis (Fig.
S1B). Chromosomal plots of the QBU data show qualitatively similar results as observed
previously in comparison of WT and rpd3∆ cells, with increased BrdU incorporation at several
single-copy, later-firing origins in rpd3∆ cells (Fig. 1A). We determined the numbers of origins
with significant BrdU incorporation and identified 235 in WT cells and 301 in rpd3∆ cells; a
Venn diagram shows that almost all the origins identified in WT cells were also identified in
rpd3∆ cells (Fig. 1B). Origins were assigned to timing quartiles based on their previously
determined replication timings (TRep) amongst 625 confirmed or likely origin in the oriDB
(Nieduszynski et al., 2007) and displayed as stack graphs of origin numbers and pie charts of
origin proportions (Fig. 1C). These results show that almost all of the additional origins detected
to fire in rpd3∆ cells fall into later (Q2, Q3, and Q4) timing quartiles (in WT cells), consistent
with Rpd3 acting to delay origin firing.

15

Figure 1. 1. Rpd3 and Sir2 have different effects on single-copy origins.
Figure 1.1. Rpd3 and Sir2 have different effects on single-copy origins. Strains CVy43 (WT), CVy44 (rpd3∆),
YHy3 (sir2Δ), and YHy6 (sir2Δ rpd3∆) were synchronized in G1 phase and released into S phase without HU for
DNA content analysis (Fig. S1A) or with HU for 60 min for QBU analysis. Strains with reduced rDNA copy
number PP1758 (WT rDNA20), and JPy115 (rpd3∆ rDNA20) were synchronized in G1 phase and released into S
phase for 120 min in the presence of HU for QBU analysis. (A) QBU values averaged for three replicates and scale-
normalized across strains are shown for representative chromosome XI; origins and origin sub-groups are indicated
with color-coded circles below the x-axis. (B) Venn diagrams depicting overlaps between origins identified as active
in each strain; the union of 311 origins detected was used in subsequent analyses. (C) Origin distributions according
to TRep quartiles shown as pie charts with proportions of identified origins in each strain, and as stack graphs
showing total numbers of origins identified in each TRep quartile. (D) Two-dimensional scatter plots comparing
QBU signals for 500bp regions centered on 311 individual origins plus the rDNA origins, which are represented by
two (overlapping) data points; origins and sub-groups are color-coded as indicated. (E) Average QBU signals for
311 origins according to their TRep quartile assignments. (F) Boxplot distributions of QBU counts for 500bp regions
aligned on origins of the indicated sub-groups; the number of origins in each group is indicated within parentheses;
two-sided t-tests were performed on all pairs of strains and significant results are indicated as: *=p<0.05,
**=p<0.01.
16
As BrdU signals at origins in HU-blocked cells inversely correlate with TRep (S. R. V.
Knott et al., 2009), we also examined the relationship between the effect of RPD3 deletion and
origin timing by plotting QBU signals for individual origins in rpd3∆ cells against their
corresponding values in WT cells (Fig. 1D). The two-dimensional scatter plot reveals a
relationship between WT BrdU levels and the effect of RPD3, with origins having higher QBU
values in WT being decreased in rpd3∆ cells and origins having lower QBU values in WT being
increased in rpd3∆ cells. To directly examine the relationship between the QBU values and
origin timing, we plotted average QBU values for origins divided according to their assigned
TRep quartiles (Fig. 1E). The results show that in rpd3∆ cells QBU values are lower for origins
in the earliest TRep quartile while QBU values are slightly higher for origins in the later quartiles,
although only the difference in the first quartile is statistically significant (Fig. S1C). We also
examined the effects of RPD3 deletion on specific origin classes by highlighting these in the
scatter plots (Fig. 1D), and by producing distribution box-plots for those groups (Fig. 1F); t-tests
of the distributions show that CEN-proximal origins, which are early-firing, are significantly
decreased in activity while Rpd3-repressed origins, which are later-firing, show significantly
increased replication activity in rpd3∆ cells (Fig. 1F). The relatively minor, but significant
reduction in early origin firing in rpd3∆ cells has not been previously reported, likely because
normalization methods obscured the difference. The opposite effect on earlier versus later origins
is consistent with the idea that origins compete for limiting factors and that increased
competition from normally later-firing origins in rpd3∆ cells consequently reduces activation of
normally earlier-firing origins.
In contrast to the results in rpd3∆ cells, the replication profiles of sir2∆ cells showed
decreased BrdU incorporation at virtually all single-copy origins (Fig. 1A, D, E, F), though the
17
effect was smaller amongst origins with higher BrdU incorporation values (Fig. 1D). Only 192
origins were called as active in sir2∆ cells, and there was a decrease in the number of origins called
in each TRep quartile (Fig. 1C). Overall, these results suggest that Sir2 acts non-specifically with
respect to single-copy origins because almost all single-copy origins are decreased in activity.
Combined deletion of SIR2 and RPD3 resulted in a replication profile intermediate to the
single mutant strains, and similar to WT (Fig. 1A, D), as supported by correlation analysis (Fig.
S1D). 226 origins were detected in rpd3∆ sir2∆ cells with almost all these origins also being
detected in WT cells and having a similar TRep distribution as in WT cells (Fig. 1B, C). Overall,
QBU signals at individual origins were also more like WT than either single mutant, although more
origins were reduced in activity as in sir2∆ cells than increased in activity as in rpd3∆ cells (Fig.
1D, F, S1D). Comparison of QBU origin signals between rpd3∆ sir2∆ and sir2∆ cells and between
rpd3∆ sir2∆ and rpd3∆ cells shows that RPD3 deletion in sir2∆ cells and SIR2 deletion in rpd3∆
cells affect origins similarly as they do in WT cells, consistent with RPD3 and SIR2 acting through
independent mechanisms (Fig. 1D, F, S1D).

SIR2 deletion is epistatic to RPD3 deletion for rDNA origin firing

Next we examined the effects of RPD3 and SIR2 deletion on initiation levels of rDNA
origins. In WT cells, the QBU signal for rDNA origins (QBU corrects for copy number so the
signal represents the average firing level per rDNA copy) was between that of single-copy origins
in the first and second timing quartiles (Fig. 2A, compare with Fig. 1E), consistent with the average
mid-S phase peak replication timing of these origins (Peace et al., 2016). In rpd3∆ cells, the rDNA
QBU signal was strongly diminished (Fig. 2A, B). In contrast, the rDNA QBU signal in sir2∆ cells
was about twice the level in WT cells (Fig. 2A, B), in rough agreement with previous reports on
the effect of SIR2 deletion on rDNA origin firing levels (Pasero et al., 2002; Yoshida et al., 2014).
18
Unlike the additive effect of double deletion of RPD3 and SIR2 on single-copy origins, the rDNA
QBU signal in rpd3∆ sir2∆ was almost identical to the sir2∆ cells (Fig. 2A, B). This result suggests
that SIR2 deletion is epistatic to, not additive with RPD3 deletion with respect to rDNA origin
firing, differing from the previous conclusion (Yoshida et al., 2014). Whereas the significantly
reduced early firing of rDNA origins in rpd3∆ cells is consistent with the idea that Rpd3 is a direct
stimulator of rDNA origin firing, the high level of rDNA origin firing in sir2∆ rpd3∆ cells conflicts
with this idea. Moreover, the effect of RPD3 deletion on single-copy origins in rpd3∆ sir2∆ cells
(Fig. 1D) occurred independently of any effect on rDNA origin firing (Fig. 2B), suggesting a direct
effect of Rpd3 on single-copy origins.

Figure 1. 2. SIR2 deletion is epistatic to RPD3 deletion for rDNA origin firing.
Figure 1.2. SIR2 deletion is epistatic to RPD3 deletion for rDNA origin firing. See Figure 1 legend for strains and
experimental description. (A) Average QBU signal for the rDNA depicted as a single locus. (B) Cumulative (+/-s.d.)
QBU signal (500bp windows) for 311 single-copy origins (left panel), and 100 rDNA origins (center panel) was
determined and plotted; two-sided t-tests were performed on all pairs of strains and significant results are indicated as:
*=p<0.05, **=p<0.01. Total QBU signal associated with rDNA origins was divided by total QBU signal associated
with 311 single-copy origins plus the rDNA origins (right panel). (C) Two-dimensional gel electrophoresis analysis
probing for the rDNA was performed on the strains as described above. Two replicate sets are shown; the images
19
framed with broken lines are darker images of the corresponding rpd3∆ image in each set. The arrows in the WT
panels indicate the bubble arcs (filled arrowhead) and 1N spots (unfilled arrowhead) that are used for quantification
(see Fig. S2B for images with exact areas measured for quantification). The graph indicates quantification of the 2D
gels by measuring intensity of “bubble arcs” in relation to the corresponding 1N spots; values are relative to WT
arbitrarily set to 1.

We also quantified QBU signal from single-copy origins versus rDNA origins, after
confirming that all four strains had similar rDNA copy numbers, estimated at ~100 in WT (Fig.
S2A). BrdU incorporation into single-copy origins was significantly lower in the sir2∆ and sir2∆
rpd3∆ strains versus WT, while rpd3∆ cells were like WT (Fig. 2B). We also calculated the absolute
proportion of rDNA origin activity as a proportion of total genome origin activity for each strain,
which shows the substantial shifts in balance between replication of the rDNA in relation to the
genome resulting from deletion(s) of RPD3 and SIR2 (Fig. 2B, right panel). Overall, the finding
that RPD3 deletion in sir2∆ cells affects single-copy (Rpd3-repressed) origins without affecting
rDNA origins argues against the idea that the effect of RPD3 deletion on single-copy origin firing
derives indirectly from diminished rDNA origin firing.
To confirm that the results of QBU at the rDNA accurately reflect origin initiation levels,
we used two-dimensional agarose gel (2D gel) electrophoresis to directly examine replication
initiation structures at rDNA origins. As above, cells were synchronized in G1 phase with α-factor
and released into S phase in the presence of hydroxyurea (HU) to evaluate early replication events.
Comparison of the intensity of replication “bubble arcs” representing initiation structures versus
the 1N spots representing total molecules yields a relative measurement of origin firing activity
(Fig. 2C, S2B). The 2D gel results mirror the QBU results with rDNA origin firing decreased
significantly in rpd3∆ cells, increased significantly in sir2∆ cells, and similarly increased in sir2∆
rpd3∆ cells (Fig 2C, S2B). These results reinforce the conclusion that SIR2 deletion is epistatic to,
not additive with, RPD3 deletion with respect to rDNA origin activation.
20
RPD3 regulates single-copy origins independently of the rDNA array

Yoshida et al. elegantly tested their model of competition between rDNA and single-copy
origins by evaluating the effect of a reduction in rDNA copy number on the suppression of single-
copy origin firing resulting from SIR2 deletion, the premise being that fewer rDNA origins would
create less competition. In support of the competition model and a role for Sir2 in regulating the
rDNA directly and single-copy origin indirectly, they showed that in a strain with only ~20 copies
of rDNA the effect of SIR2 deletion on single-copy origins was largely suppressed, as predicted.
We adopted the same rationale and approach to test whether Rpd3 acts on single-copy origins
through effects on the rDNA by evaluating the effect of RPD3 deletion in the strain background
with the reduced rDNA copy number; we confirmed that the strains with reduced rDNA copy
numbers contained about 25% of the rDNA complement as the normal copy number strains (Fig.
S2A). Chromosome plots show that similar de-repression of dormant/late-firing origins occurred
in rpd3∆ cells compared with WT cells in the strains with ~20 rDNA copies (Fig. 1A). Analysis
of QBU values for individual origins in the scatter plots revealed a similar relationship of RPD3
deletion on origins according to their QBU levels, with the greatest increases among those origins
that normally have low QBU values, reflecting later-firing origins (Fig. 1D). Distribution box-
plots of QBU signals affirm the global increase in firing of Rpd3-repressed origins in rpd3∆ cells
relative to WT (Fig. 1F), and correlation analysis demonstrates closer similarity of the rpd3∆ strain
with reduced rDNA copy number to the rpd3∆ strain with normal rDNA than the WT strain with
normal rDNA (Fig. S1D). Notably, rDNA origin activity level was only modestly reduced relative
to WT in the rpd3∆ rDNA20 strains. These results indicate that the effect of RPD3 deletion on
single-copy origins does not rely on repression of a large array of rDNA origins. Instead, the
21
findings support the conclusion that Rpd3 regulates origin firing directly at single-copy origins,
rather than indirectly through the rDNA.

RPD3 suppresses pan-S phase rDNA origin efficiency

The analyses above were all conducted in the presence of HU to evaluate the levels of
origin firing occurring in an early S phase interval. To determine the impacts of RPD3 and SIR2
on the overall efficiency of rDNA origin firing throughout the duration of S phase, we used 2D
gels to examine and quantify replication intermediates at the rDNA origins in unperturbed cycling
cells (pan-S phase); here we compared the intensities of initiation bubble structures to replication
fork structures to provide a relative measure of origin efficiency (Fig. 3A, S3A). In striking
contrast to the 2D gel results in early S phase (HU) described above, pan-S phase rpd3∆ cells
exhibited modest, but significantly increased overall efficiency of rDNA origin firing relative to
WT cells (Fig. 3A). The overall increased firing of rDNA origins in rpd3∆ cells is inconsistent
with the notion that RPD3 is a direct activator of rDNA origins. Interestingly, sir2∆ cells showed
no difference in pan-S phase rDNA origin efficiency compared to WT (Fig. 3A), which contrasts
with a previous report(Pasero et al., 2002). Moreover, rpd3∆ sir2∆ cells showed similar pan-S
phase rDNA origin efficiency as sir2∆ and WT cells (Fig. 3A). Thus, the striking, and opposite,
effects of SIR2 deletion and RPD3 deletion on rDNA origin firing levels in early S phase are not
reflected in their overall efficiencies through the duration of S phase.
22

Figure 1. 3. Rpd3 suppresses rDNA origin firing.
Figure 1.3. Rpd3 suppresses rDNA origin firing. (A) Strains described in Figure 1 legend were grown
logarithmically and harvested for 2D gel analysis and probing for the rDNA. Three replicate sets are shown. The
arrows in the WT panels indicate the bubble arcs (filled arrowhead) and early Y-arcs (unfilled arrowhead) that are
used for quantification (see Fig. S3A for images with exact areas measured for quantification). The graph shows
quantification of the 2D gels by measuring intensity of “bubble arcs” in relation to the corresponding Y-arcs
(replication fork structures); values are relative to WT arbitrarily set to 1; two-sided t-tests were performed on all pairs
of strains and results are indicated as: *=p<0.05, **=p<0.01. (B) WT, rpd3∆, and sir2 strains described above were
synchronized in G1 phase, released into S phase, and incubated with BrdU for the following time intervals for QBU
analysis: pulse 1=15-30min, pulse 2=25-40 min, pulse 3=35-50 min, and pulse 4=45-60min; 5 min overlap is included
to account for lag time in BrdU entry to cells and incorporation into DNA. The stack graphs show total QBU signal
for 625 potential single-copy origins (left) and total QBU signal for all rDNA origins during each time interval. (C)
Boxplot distributions of QBU signals (500 bp regions) in the first pulse interval of the time-course for the indicated
origin groups.

The differences in origin firing in the early-S versus pan-S phase analysis, particularly in
rpd3∆ cells, suggests that alterations in rDNA origin usage early are subsequently compensated.
For example, more early-firing of normally later-firing, single-copy, Rpd3-regulated origins,
should leave fewer such origins unfired in late S phase in competition with unfired rDNA origins,
thus shifting the competitive balance. To more precisely examine the temporal dynamics of
replication in the rDNA versus single-copy origins in rpd3∆ cells, we analyzed BrdU incorporation
through a time series of pulse-labeling in cells progressing synchronously through S phase. This
23
analysis also eliminates possible perturbations due to hydroxyurea. The analysis shows relative
levels of BrdU incorporation at single-copy versus rDNA origins over the time-course (Fig. 3B).
Replication origin usage profiles are similar in the initial time-pulse as in HU (Fig. S3B). As
expected, rDNA replication occurs throughout S phase in WT cells. In rpd3∆ cells, the relative
proportions of replication at rDNA versus single copy origins shifts in favor of single-copy origins
in earlier pulse periods, and conversely, in favor of rDNA origins in the later periods (Fig. 3B).
The results are consistent with the idea that reduced usage of rDNA origins in early S phase
preserves more rDNA origins to initiate later, likely with fewer unfired, single-copy origins in
competition. Analysis of single-copy origin firing in the earliest interval shows significantly
increased activity of Rpd3-repressed origins, showing that the effect of RPD3 deletion is similar
in early S phase with or without HU (Fig. 3C).
Although deletion of SIR2 had no effect on overall origin efficiency in pan-S phase analysis,
examination of BrdU incorporation over time shows a striking shift toward earlier replication of
the rDNA in sir2∆ compared with WT cells (Fig. 3B), consistent with the early S phase results in
HU. SIR2 deletion also results in notably delayed replication of the total genome, while RPD3
deletion’s effect on the genome is minor and in the same direction as its effect on the rDNA in the
earliest pulse (Fig. 3B). We also note that this type of global BrdU incorporation analysis does not
distinguish active versus passive replication of origin sequences.

Origin de-repression in rpd3∆ cells requires Fkh1/2

FKH1 and/or FKH2 (FKH1/2) are responsible for the early activation of most non-
centromere-proximal, early-firing origins throughout the genome (S. R. V. Knott et al., 2012).
Fkh1 promotes early origin firing by recruiting limiting factor Dbf4 to stimulate execution of the
DDK-dependent Sld3-Cdc45 loading step of replication initiation (Fang et al., 2017; Zhang et al.,
24
2019). Overexpression of Fkh1 stimulates earlier firing of many later-firing origins including
Rpd3-repressed origins (Fig. S4A) (Peace et al., 2016), suggesting an antagonistic relationship
between Fkh1 and Rpd3. Because Fkh1 levels can modulate firing of Rpd3-repressed origins, we
tested whether Fkh1 protein level was altered in rpd3∆ cells and found no difference relative to
WT (Fig. S4B). To gain further insight into the relationship between Rpd3 and Fkh1/2 function in
replication timing, we tested for genetic interaction by determining the effect of their combined
mutation.We performed QBU with WT, rpd3∆, fkh1∆ fkh2-dsm, and rpd3∆ fkh1∆ fkh2-dsm strains;
the fkh2-dsm allele was used rather than fkh2∆ because it is virtually null in replication timing
control but functional in regulating gene expression, and therefore does not, in combination with
FKH1 deletion, exhibit more severe phenotypes of fkh1∆ fkh2∆ cells, such as pseudohyphal growth,
which complicates analysis and potential interpretation (Ostrow et al., 2017). QBU values
comparing the individual replicates of each strain are well correlated (Fig. S4C).
The results show that de-repression of Rpd3-repressed origins in rpd3∆ cells was largely
eliminated in rpd3∆ fkh1∆ fkh2-dsm cells (Fig. 4). Analysis of Rpd3-repressed origins shows no
significant increase in BrdU incorporation in rpd3∆ fkh1∆ fkh2-dsm versus fkh1∆ fkh2-dsm cells
(Fig. 4B). This finding suggests that Rpd3 opposes origin stimulation by Fkh1/2. As origin
stimulation by Fkh1/2 involves its direct binding to DNA sequences at origins, we posited that
repression by Rpd3 likely involves direct effect on the origin.
25

Figure 1. 4. Origin de-repression in rpd3∆ cells requires Fkh1/2.
Figure 1.4. Origin de-repression in rpd3∆ cells requires Fkh1/2. Strains CVy43 (WT), CVy44 (rpd3∆), OAy1114
(fkh1Δ fkh2-dsm), and YHy13 (fkh1Δ fkh2-dsm rpd3∆) were synchronized in G1 phase and released into S phase with
HU for 90 min for QBU analysis. (A) Average QBU values are shown for representative chromosome XI; origins
and origin sub-groups are indicated with color-coded circles below the x-axis. (B) Boxplot distributions of QBU
counts for 500bp regions aligned on origins of the indicated sub-groups; the number of origins in each group is
indicated within parentheses; two-sided t-tests were performed on all pairs of strains and results are indicated as:
*=p<0.05, **=p<0.01. (C) Two-dimensional scatter plots comparing QBU signals centered on 311 individual origins
plus the rDNA origins represented by two (overlapping) data points; origins and sub-groups are color-coded as
indicated.

Rpd3 modulates Fkh1 binding to origins

Given the opposing activities of Fkh1/2 and Rpd3 in origin activation, we considered the
possibility that Rpd3 might regulate binding of Fkh1/2 to origins, particularly Rpd3-regulated
origins, including the possibility of differential regulation of Fkh1/2 binding to rDNA origins
versus single-copy origins. Our previous studies have shown Fkh1/2 binding to Fkh-activated
origins in G1 phase (Ostrow et al., 2014). In that study, we also observed more extensive origin
binding in S phase-arrested cells suggesting that Fkh1/2 may bind additional origins during S phase
to stimulate activation of later origins. Therefore, we hypothesized that Rpd3 regulates Fkh1
26
binding to later-firing origins in G1 and/or early S phase and performed ChIP-Seq of Fkh1-
3xFLAG in WT and rpd3∆ cells synchronized in G1 to test this idea. Comparison of ChIP signal
from experimental replicates shows high correlation demonstrating high reproducibility of the data
(Fig. S5A). In WT cells in G1, heatmaps of average ChIP signal show local Fkh1 enrichment at
previously identified Fkh1 binding sites, including Fkh-activated origins, as expected, as well as
some enrichment at Rpd3-repressed origins, previously unnoticed (Fig. 5A). In rpd3∆ cells in G1,
Fkh1 profiles are similar, however box-plot analysis and statistical testing shows there is a modest,
but significant, increase in Fkh1 enrichment at Rpd3-repressed origins, but not at other origin
groups or Fkh1 binding sites more generally (Fig. 5A, B). Regression analysis of the ChIP
enrichment values for Rpd3-repressed origin correlates with the increased activation of Rpd3-
repressed origins in early S phase in rpd3∆ cells (Fig. 5C). These results are consistent with the
idea that the mechanism of Rpd3 in regulation of origin initiation is to oppose origin binding of
Fkh1.

27

Figure 1. 5. Rpd3 modulates Fkh1 binding to origins.
Figure 1.5. Rpd3 modulates Fkh1 binding to origins. Strains OAy1112 (FKH1-3xFLAG) and YHy17 (rpd3∆,
FKH1-3xFLAG) were synchronized in G1 phase and subjected to ChIP-seq. (A) Heatmaps of ChIP-seq enrichment
at selected origin and Fkh1 binding loci; the number in each group is indicated within parentheses. The “Non-origins”
group refers to Fkh1 binding sites called as peaks by MACS in the Fkh1 ChIP-seq dataset that do not overlap with a
replication origin. (B) Boxplots distributions of ChIP-seq enrichments for 500bp regions aligned on origins of the
indicated sub-groups; the number of origins or Fkh1 binding loci in each group is indicated within parentheses. Two-
sided t-tests were performed on all pairs of strains and results are indicated as: *=p<0.05, **=p<0.01. (C) Two-
dimensional scatter plots of ChIP-seq signals (500bp regions) for 174 Fkh1 binding sites (called as peaks by MACS
as described above but including origins) plus the indicated origin sub-groups, color-coded as indicated. A linear
regression deriving the best-fit line is shown for each origin group.

Fkh1 binding at origins occurs in G1 and S phases and is dependent on the pre-replication
complex (pre-RC), defined by origin-loaded MCM complexes (Ostrow et al., 2014; Reinapae et
al., 2017). Thus, a plausible mechanism for the increased Fkh1 binding to Rpd3-repressed origins
in rpd3∆ cells might derive from an effect of Rpd3 on pre-RC levels. To address this possibility,
we performed ChIP-seq of Mcm4-3xHA in WT and rpd3∆ cells synchronized in G1. The results
28
show robust enrichment for Mcm4 at origins of various classes, as expected, however, no
significant difference in Mcm4 enrichment is detected at Rpd3-repressed origins, nor at the other
groups examined (Fig. S5). Taken together, these results indicate that Rpd3 modulates Fkh1
binding without altering pre-RC levels and are consistent with the idea that Rpd3 acts downstream
of pre-RC assembly to impede Fkh1 binding.

RPD3 deletion suppresses DBF4 deficiency implicating the Fkh1-DDK pathway

As a further test of our hypothesis that Rpd3 acts in opposition to Fkh1/2-dependent
stimulation of Dbf4 recruitment, we tested whether deletion of RPD3 suppresses the temperature-
sensitivity of a C-terminally truncated DBF4 allele (dbf4∆C) (Jones et al., 2010). Deletion of
FKH1 (or FKH1 and FKH2) does not cause temperature-sensitivity, indicating that Dbf4∆C is
defective in function beyond its proposed defect in interaction with Fkh1. In attempting to isolate
a fkh1∆ dbf4∆C strain through meiosis and sporulation of a heterozygous diploid, we determined
that FKH1 deletion causes lethality in combination with dbf4∆C (at normally permissive
temperature for dbf4∆C), strongly suggesting that Dbf4∆C depends on Fkh1 for recruitment to
origins (Fig. 6A). Thus, we created a diploid heterozygous for RPD3/rpd3∆ and DBF4/dbf4∆C
and induced meiosis and sporulation at permissive temperature. Individual spores from mature
tetrads were dissected onto rich medium, incubated at 30°C (restrictive temperature for dbf4∆C),
and imaged and genotyped. The results show that RPD3 deletion suppresses the lethality of dbf4∆C
at 30°C as virtually no dbf4∆C segregants were recovered while the expected numbers of rpd3∆
dbf4∆C segregants were recovered, although these formed much smaller colonies than WT or
rpd3∆ isolates (Fig. 6B). Out of 13 tetrads dissected, 37 viable spores were genotyped, resulting
in 15 WT, 9 rpd3∆, 12 rpd3∆ dbf4∆C, and 1 dbf4∆C viable isolates. Fisher’s exact test indicates
that the RPD3 and DBF4 genotypes being tested are not independent (p=0.00164), strongly
29
supporting the notion that loss of Rpd3 function restores viability of Dbf4-deficient cells, likely
by facilitating Dbf4∆C recruitment to origins by Fkh1.

Figure 1. 6. RPD3 deletion suppresses DBF4 deficiency.
Figure 1.6. RPD3 deletion suppresses DBF4 deficiency. (A) Diploid strain YHy38 (FKH1/fkh1∆ DBF4/dbf4∆C)
was induced into meiosis and sporulation. Tetrads were dissected onto rich medium and incubated at 23°C and imaged
after three days and genotyped. Tetrads are arranged in columns; determined haploid genotypes are indicated on the
panel to the right as F/f for FKH1/fkh1∆ and D/d for DBF4/dbf4∆C; FKH1 and DBF4 are unlinked. Genotypes in red
for inviable spores were inferred based on assumption of 2:2 segregation of alleles. The “fd” hypotheticals are circled.
(B) Diploid strain OAy1188 (RPD3/rpd3∆ DBF4/dbf4∆C) was induced into meiosis and sporulation. Tetrads were
dissected onto rich medium and incubated at 30°C and imaged and genotyped after three days. Tetrads are arranged
in columns; determined haploid genotypes are indicated on the panel to the right as R/r for RPD3/rpd3∆ and D/d for
DBF4/dbf4∆C; RPD3 and DBF4 are unlinked. Genotypes in red for inviable spores were inferred based on assumption
of 2:2 segregation of alleles. The “rd” isolates and hypotheticals are circled; the “Rd” hypotheticals are boxed.

DISCUSSION

A full understanding of genome-wide replication dynamics requires the application of
methods that accurately quantify the process and that may be accurately compared across
independent samples. The QBU method has now allowed us to re-examine how the Sir2 and Rpd3
histone deacetylases affect genome replication, particularly the balance between dispersed, single-
copy origins and the tandemly repetitive rDNA origins, which has been identified as a key factor
in genome stability (Lindström et al., 2018). Overall, our results support the current concept of
30
differential replication timing amongst origins as the consequence of the differential accrual of
limiting replication initiation factors. Deletions of regulators such as Sir2 and Rpd3 that affect the
differential accrual of factors disrupt the existing balance, potentially creating conflicts between
different genomic regions. Thus, understanding how different chromatin regulators help to balance
the activity of origins intrinsically in competition with each other is an important goal.
A clear, new conclusion of our study is that Rpd3 acts independently of Sir2 and of the
rDNA in exerting its effects on single-copy origins. By quantifying and normalizing for the
absolute levels of replication associated with single-copy versus the rDNA origins, we showed that
the effects of RPD3 deletion on single-copy origins are additive with, and hence independent of
SIR2 deletion, whereas SIR2 deletion’s effect on the rDNA origins is epistatic to RPD3 deletion.
Specifically, RPD3 deletion in sir2∆ cells affects single-copy origins without altering rDNA origin
firing, which remains elevated in the sir2∆ background. This is not consistent with Rpd3 acting as
an rDNA origin activator but is consistent with Rpd3 acting as a repressor of single-copy origins.
Indeed, the increased pan-S phase efficiency of rDNA origins in rpd3∆ cells is consistent with
Rpd3 acting as a repressor of rDNA origins also. Moreover, we showed that the effect of RPD3
deletion on single-copy origins occurred in strains with much reduced rDNA complement, even
while the effect on rDNA origins was also reduced. Thus, deregulation of single-copy origins in
rpd3∆ cells does not appear to derive from primary effects on rDNA origins, though we cannot
rule out the possibility of direct effect(s) of Rpd3 on rDNA origins.
Our quantitative conclusions on the effects of Rpd3 and Sir2 on rDNA origin firing from
QBU were confirmed by the “gold standard” method of 2D gels for analysis of origin initiation
levels. First, we showed that the differences in early firing levels (in HU) matched those
determined by QBU, with significantly higher initiation of the rDNA origins in sir2∆ and sir2∆
31
rpd3∆ cells but lower initiation in rpd3∆ cells compared with WT. Remarkably, we found that
rpd3∆ cells showed a higher level of rDNA origin firing than WT in a pan-S phase culture, and
conversely found that sir2∆ and sir2∆ rpd3∆ cells were unchanged from WT in this analysis. In
both cases, it appears that changes in origin usage subsequent to the levels observed in the early S
phase-arrested cells, compensate for the earlier imbalances. We used a BrdU pulse-labeling
approach throughout S phase that provided evidence in support of the idea that the early reduction
in rDNA origin firing in rpd3∆ cells is followed by an increase in later rDNA origin firing relative
to WT cells. We note that the levels of early versus late origin usage are intrinsically linked in that
reduced usage early reserves a larger complement of origins for later usage, and vice-versa. It is
also anticipated that as replication progresses, the abundance of replication factors will increase in
relation to the number of remaining unfired origins, while passive replication of potential origins
by active forks simultaneously plays an important role in these dynamics (Hyrien & Goldar, 2010;
Rhind & Gilbert, 2013). In considering passive replication of the rDNA, it is notable that most
replication is unidirectional due to the presence of the replication fork barrier(Brewer et al., 1992),
so passive replication of rDNA origins is reduced, potentially reserving more origins for later firing.
Interestingly, the reduced early replication of rDNA in rpd3∆ cells ultimately is associated with a
higher overall efficiency, which we think is quite unusual, and are unaware of similar examples.
The exact mechanism involved in the repression of origin firing by Rpd3 has remained
elusive and is generally thought to act through deacetylation to reduce access to chromatin
(Soriano et al., 2014). Several potential targets have been proposed among replication initiation
factors, including those acting in both CDK-dependent and DDK-dependent initiation pathways
(J. G. Aparicio et al., 2004; Mantiero et al., 2011; Vogelauer et al., 2002). For example, we showed
previously that RPD3 deletion partially suppresses the late origin initiation defect of clb5∆ cells
32
(J. G. Aparicio et al., 2004), presumably facilitating access of Clb6-Cdk1, which we showed was
limiting in clb5∆ cells (Gibson et al., 2004). Mantiero and colleagues showed that RPD3 acts in
opposition to simultaneous overexpression of multiple initiation factors on both initiation branches
(Mantiero et al., 2011). More recently, we showed that overexpression of Fkh1/2 alone stimulates
significantly more origins to detectably fire earlier (Peace et al., 2016). As origin stimulation by
Fkh1/2 involves Dbf4 recruitment directly to origins (Fang et al., 2017; Zhang et al., 2019), we
tested whether RPD3 affects the DDK pathway through impedance of Fkh1 access to bind origin
chromatin. Chromatin immunoprecipitation analyses indicate that RPD3 deletion leads to a modest
but significantly increased enrichment of Fkh1 at Rpd3-repressed origins in G1 phase, anticipating
the modest increases in QBU signals at these origins in early S phase. Together with the
dependence of increased origin firing in rpd3∆ cells on FKH1/2, these findings support a
mechanism whereby RPD3 regulates origin function by impeding Fkh1 binding, to varying
degrees at different loci. The relatively subtle effects of RPD3 deletion on Fkh1 binding may
reflect the limiting nature of Fkh1 being distributed across numerous potential binding loci
differing in the number and/or distribution of Fkh1/2 binding sites and/or in local recruitment of
Rpd3. In fact, Fkh1/2 were originally identified as potential origin regulators due to anti-
correlation of their binding with Rpd3-regulated origin loci (S. R. V. Knott et al., 2012). We also
showed previously that Fkh1 binding is associated with decreased nucleosome occupancy (Ostrow
et al., 2014), and Soriano et al. reported changes in nucleosome occupancy at Fkh1 binding sites
in rpd3∆ cells (Soriano et al., 2014).
Finally, we presented evidence that Fkh1 and Rpd3 have opposing functions in relation to
Dbf4 as deletion of FKH1 causes lethality of dbf4∆C cells, whereas elimination of RPD3 function
rescues the viability of cells having the dbf4∆C allele, which is hypomorphic for DDK function.
33
The implication is that RPD3 deletion facilitates Fkh1 recruitment to origins, which becomes
critical in Dbf4-deficient cells. This finding reinforces the model we have proposed wherein Rpd3
acts by impeding Fkh1, though the results do not exclude other factors from being similarly
affected, such as implied for the CDK pathway by suppression of clb5∆ defects as discussed above.
Given the pervasive role of histone deacetylases in most aspects of genome regulation, the
mechanistic insights gained here in yeast are expected to serve as paradigms for derived functions
in more complex systems.
Concluding on the possible evolutionary and developmental roles of Fkh1/2 in regulation
of rDNA origin activity, we note that rDNA dynamics, such as expansion and contraction of
repeats, and regulation of gene expression are both potentially influenced by rDNA origin
activity and replication timing. For example, higher initiation levels are expected to lead to
higher numbers of blocked forks and thus increased potential for DNA breaks leading to changes
in rDNA copy numbers. Increased early firing of the rDNA may also contribute to higher rRNA
expression levels by increasing template numbers early in the cell cycle and confer a growth
advantage. Intriguingly, deletion of both FKH1 and FKH2 leads to pseudohyphal growth, which
is normally a starvation response, on rich medium, suggesting that FKH1/2 play a central role to
integrate growth signals with cell cycle progression and DNA replication, which will be
interesting to study further.

MATERIALS AND METHODS
Yeast strain construction. Strain constructions were carried out by genetic crosses or lithium
acetate transformations with linearized plasmids or PCR products generated with hybrid
oligonucleotide primers having homology to target loci (Ito et al., 1983; Longtine et al., 1998);
34
primer sequences for strain constructions are given in Table 1. Yeast strains are described in
Table 2. All strains used for experiments are congenic with W303 background and most are
derived from BrdU-incorporating strains CVy43, CVy61, and CVy68, which are derived from
SSy161 (Viggiani & Aparicio, 2006). OAy1096 was derived from a cross of OAy1069 and
OAy1070 (Peace et al., 2016). OAy1100 was created by transformation of SSy161 with PCR
product of primers 2xL-3xFLAG-FKH1-F and 2xL-3xFLAG-FKH1-R using plasmid p2xL-
3xFLAG (TRP1) as template (gift from T. Tsukiyama); expression of Fkh1-3xFLAG was
confirmed by Western blotting with anti-FLAG M2 at 1:1000 (Sigma F1804). OAy1102 is an
ADE2 derivative of OAy1100. The clean replacement of FKH2 with fkh2-dsm was constructed in
two steps: FKH2 was deleted in CVy68 using primers pAG61-Fkh2∆5′ and pAG61-Fkh2∆3′
with URA3MX to create strain OAy1107, which was transformed with fkh2-dsm DNA
amplified using primers Fkh2-up400bp and Fkh2-down375bp from p405-Fkh2-dsm(Ostrow et
al., 2017), followed by selection of 5-FOA to yield strain OAy1109. OAy1114 is haploid
segregant from a cross of strains OAy1109 and OAy1096. Strain YHy36 was constructed by
PCR amplification of HIS3MX from pFA6a-HIS3MX6(Longtine et al., 1998) with primers HIS-
F and HIS-R and transformation into JHy4. RPD3 and SIR2 deletions were constructed by PCR
amplification of: rpd3∆::KanMX from the CVy44 or JHy4 genomes or rpd3∆::HIS3MX from the
YHy36 genome using primers RPD3-del-up and RPD3-del-down, or sir2∆::KanMX from the
JHy3 genome using primers SIR2-del-up and SIR2-del-down, followed by transformations into
PP1758, CVy61, MPy102, OAy1102, OAy1114, and YHy3 to create strains JPy115, YHy33,
YHy37, YHy17, YHy13, and YHy6 respectively (see Table 2). HMLα deletions were
constructed by transformation of YHy33 with PCR-amplified hmlɑΔ::URA3MX from pAG61
(Goldstein et al., 1999) using primers HMLalpha-del-F and HMLalpha-del-R to create YHy3.
35
MCM4 was C-terminally tagged with 3xHA epitope by transformation with PCR product of
primers 3HA-MCM4-F and 3HA-MCM4-R and template pFA6-3HA-kanMX6 to create strain
MPy102; expression of Mcm4-3xHA was confirmed by Western blotting with anti-HA (16B12
or 12CA5) at 1:2000. DBF4 C-terminal truncation (dbf4∆C) was created using primers Dbf4∆C-
F1 and Dbf4∆C-R1 to amplify a DNA fragment for insertion of a stop codon with HIS3MX into
a diploid strain based on crossing CVy43 and CVy68. Haploid segregants were generated to
yield YHy19. CVy44 was backcrossed with CVy68 to generate OAy1186, which was mated
with YHy19 to create diploid OAy1188. Genomic alterations were confirmed by PCR analysis
and/or DNA sequence analysis as appropriate.

Other methods. Cultures were grown at 23°C and synchronization was performed as described
previously(Haye-Bertolozzi & Aparicio, 2018). 5μg/mL α-factor was used for synchronization of
BAR1 strains; 5ng/mL for bar1∆ strains. QBU analysis was performed as described (Haye-
Bertolozzi & Aparicio, 2018), using KAPA Hyper Prep Kit (KK8504). DNA content analysis by
flow cytometry (FACScan) has been described previously(J. G. Aparicio et al., 2004). 2D gels
were performed as described previously (Villwock & Aparicio, 2014), except that only the rDNA
fraction in the cesium chloride gradient was isolated and BND cellulose enrichment was omitted.
1μg rDNA was digested with HindIII; the probe for the 2D gel was generated as described
previously by radioactive labeling of a DNA fragment produced by PCR with primers Probe-F and
Probe-R using genomic DNA as template. Radioactive signal was captured on phosphor screens,
scanned with Typhoon scanner and imaged and quantified with ImageQuant software (BioRad).
ChIP-seq was performed as described(Ostrow et al., 2015), using KAPA Hyper Prep Kit (KK8504)
using anti-FLAG M2 monoclonal antibody (Sigma F1804) at 1:100, anti-HA monoclonal
36
antibodies 12CA5 (Millipore-Sigma ROAHA) and 16B12 (BioLegend MMS-101P) at 1:100.
High-throughput DNA sequencing was performed on one of several Illumina platforms at several
different facilities.

Computation and statistics. All sequencing data were binned (50bp) and median-smoothed over
a 1kb window. BrdU peaks were called by MACS 1.4.2 with no-model mode (p<0.01). The called
peaks in each set were cross-referenced against origins listed in OriDB to eliminate any peaks not
aligning with a “confirmed” or “likely” origin. Overlapping origins between the different datasets
were determined by bedtools 2.25.0 using intersect function; from these results, the union of
origins identified in WT and rpd3∆ cells was determined for use as the total origins list of 306 only
for the following purpose: Rpd3-repressed origins were called by applying a two-sample, two-
tailed t-test (p<0.001) comparing WT and rpd3∆ QBU signals within a 5kb window centered on
these 306 replication origins; 80 differential origins were detected, of which 74 were increased in
rpd3∆ and thus denoted Rpd3-repressed (Table S1). The list of origins and their TRep values and
set assignments (FKH-activated; n=94, excludes rDNA origin) are from (S. R. V. Knott et al., 2012;
Nieduszynski et al., 2007). CEN-proximal origin refers to the closest origin on each side of each
centromere (Table S2). FKH1-OE-activated origins were taken from (Peace et al., 2016) (Table
S3). Matlab was used for generation of most data displays and analyses. ChIP-seq data were
normalized against data untagged control strains. Fkh1 ChIP peaks were called by MACS 2.2.7.1
with no-model mode and instead an extension size of 100bp. ChIP data files were compared to
control data files (untagged strain) allowing for background to be estimated (false discovery rate
adjusted p< 0.05).

37
CHAPTER II

Broadly Applicable Control Approaches Improve Accuracy of
ChIP-Seq Data

Adapted from:
Petrie, M. V.*, He, Y.*, Gan, Y., Ostrow, A. Z., & Aparicio, O. M. (2023). Broadly Applicable
Control Approaches Improve Accuracy of ChIP-Seq Data. International Journal of Molecular
Sciences, 24(11), 9271.

As one of the lead authors of the following work, I contributed to the experimental design and
execution, data collection and analyses, and data visualizations.

38
INTRODUCTION

Protein-DNA interactions control all aspects of genome function, from the histones that
compactly organize and regulate access to DNA to trans-acting factors binding DNA to regulate
cis-elements and also to factors interacting indirectly with DNA through chromatin. Chromatin
immunoprecipitation analyzed using DNA sequencing (ChIP-seq) is a widely used and relied
upon method for the determination of the genomic binding sites of chromatin-associated proteins
(Figure 1A) (Nakato & Sakata, 2021; Nakato & Shirahige, 2016; Park, 2009). The efficacy of
ChIP relies on the antibody-mediated immunoprecipitation (IP) of the intended target protein.
Many available antibodies are ineffective or unavailable in sufficient quantities, so a common
approach is the use of well-characterized monoclonal antibodies against commonly used peptide
epitopes such as HA, MYC, and FLAG, which may be encoded into target proteins, usually
without disruption to the native function. In addition to the key challenge of detecting bona fide
interactions (false negatives), one of the persistent problems associated with ChIP-seq is false-
positive signals, which can be pervasive (Fan & Struhl, 2009; Teytelman et al., 2013; Wreczycka
et al., 2019). While several factors might contribute to false positives, one genomic feature that
has been correlated with false-positive ChIP signals is the level of transcription, with highly
transcribed regions frequently exhibiting signals across unrelated experiments, this phenomenon
being referred to as hyperChIPability (Teytelman et al., 2013).
Because the enrichment for hyperChIPable sequences is associated with the IP step,
neither an input chromatin sample nor immunoprecipitation in the absence of antibody are
adequate controls that enable the subtraction of this false signal. Furthermore, ChIP with an
antibody versus epitope tag in a strain lacking an epitope-tagged protein exhibited variability in
the detection of hyperChIPable signals (Teytelman et al., 2013). In this previous study, a more
39
useful control to identify hyperChIPable signals for subtraction from total signals appeared to be
ChIP of cells expressing an unrelated, non-DNA-binding protein, nuclear-expressed green
fluorescent protein, the enrichment of which correlated with locus expression levels. These
results suggest that a combination of factors, including the nature and expression of the target
epitope/protein as well as the characteristics of specific antibodies, influence ChIP-enrichment of
false binding loci. To solve this problem, we aimed to develop a universal approach based on
epitope tag sharing by the protein of interest and a control protein, which is not expected to bind
the genome specifically. Hence, ChIP in parallel with a strain expressing only the control protein
would provide data on non-specific enrichment for background subtraction using normalization.
We also tested an approach utilizing a specific DNA-binding mutant of the protein of interest.
Both methods vastly improve the accuracy of our ChIP-seq datasets, particularly in removing
likely false positives.

40
RESULTS

Expression of an Epitope-Tagged Protein as Normalization Control

Figure 2. 1. Control constructs and analysis scheme.
Figure 2.1. Control constructs and analysis scheme. (A) Standard ChIP-seq procedure performed here. Red stars
represent the protein of interest and blue stars represent other proteins. Red rectangles represent unique barcode
sequences and orange rectangles indicate universal indexes. (B) Schematic representation of DNA elements
comprising the different control constructs, not showing the plasmid vectors. The ADE2 and TRP1 sequences target
integration of the constructs as shown upstream of the promoter regions of these loci; P and T represent promoter
and terminator sequences and LexA-OP2/4 represent two or four LexA binding sequences. (C) Schematic
representation of ratio normalization (RN) to remove background signal.

In the absence of more definitive controls, the examination of highly expressed and other
hyperChIPable loci is fraught with ambiguity. More broadly, the variability in ChIP
enrichment, due to multiple potential factors indicated above, emphasizes the continuing need for
better control approaches. We devised a scheme to generate ChIP controls that allow for the
identification of false-positive signals while simultaneously providing control for internal and/or
external quantitative normalizations. The scheme was designed for the analysis of epitope-tagged
proteins, focusing here on the use of the common MYC, HA, and FLAG epitopes. We generated
41
constructs that express bacterial LexA protein with a C-terminal fusion of one of the three
epitopes and contain lexA operator sites (lexA-Op) to serve as sequence-specific binding site(s)
for the LexA-epitope-tagged protein (Figure 1B). These constructs, expressing LexA-3xHA,
LexA-13xMYC, and LexA-3xFLAG, were stably integrated into an otherwise untagged, wild-
type yeast strain and referred to as HOP, MOP, and FLOP, respectively. These control strains
would be modified by the introduction of the same epitope-tag onto the gene encoding the
protein of interest. The expressions of tagged proteins were confirmed using immunoblot
analysis (Figure S1).
We envisioned that a strain bearing the MOP control construct would be analyzed using
ChIP alongside the identical strain also expressing “your-favorite-gene” (YFG) tagged with
MYC (YFG-MYC (MOP)), with the expectation that differential analysis will reveal bona fide
ChIP-enrichment for the target protein (Figure 1C). In addition, tagged-LexA binding
enrichment observed at the lexA-Op sites was anticipated to serve as a standard for quantitative
normalization directly between experimental replicates or as an internal standard versus other
queried sites in the genome. As such, the control locus also should permit the evaluation and
potential correction of technical variance in the IP between individual samples and replicates.
To begin to evaluate the utility of the approach, we performed ChIP-seq of untagged
strains subject to IP with anti-MYC, anti-FLAG, or anti-HA. Triplicates of each ChIP sample
were produced and, following mapping to the genome and binning, compared to each other in
two-dimensional scatter plots to confirm a high degree of correlation between replicates (Figure
S2A). Triplicates were read-count normalized and averaged. For comparisons, averaged datasets
were scale-normalized using the ChIP signal at the 20th percentile expected to represent
unenriched, background signals across all samples.
42
We analyzed ChIP signals at the lexA binding sites within the integrated control
constructs. Three different constructs were tested: LexA-13xMYC and LexA-3xHA were
expressed from the ADH1 promoter, and these vectors contained one lexA binding array with
four semi-palindromic binding sites for the LexA dimer (Figure 1B); MOP and FLOP were
integrated next to ADE2 on chr XV, while HOP was integrated next to TRP1 on chromosome IV
(Figures 1B and 2A). LexA-3xFLAG was expressed from the stronger TEF1 promoter, and the
integrating vector contained two separate lexA binding arrays, one with two and one with four
lexA binding sequences (Figure 1B). We plotted data for the MOP, FLOP, and HOP loci in the
control ChIPs showing binding at these loci, specifically in the strain expressing the
corresponding epitope-tagged LexA (Figure 2A); data for individual replicates show good
reproducibility (Figure S2A). The FLOP construct that contains two lexA binding arrays
additionally provides a useful demonstration of the resolution of the data, showing a clear
separation between the two data peaks corresponding to the two lexA arrays, separated by
~1.6kb (Figure 2A). These control loci can be used for normalization across sample replicates or
internal standardization. However, to compare independent ChIP samples across different IPs,
with and without target protein expression, we find that a normalization approach based on the
data distributions, such as the 20th percentile background we have applied, is more reliable and
established, and we have used that method throughout this paper. Hereon, we focus on the utility
of MOP/FLOP/HOP expression as normalization controls for background subtraction.

Non-Specific Signals Are Pervasive in ChIP

We examined averaged ChIP signals for 500 bp bins of consecutive, non-overlapping
sequences for the whole genome using quartile boxplot analysis. An examination of the ChIP
signal distribution across the genome amongst the different control IPs in untagged strains shows
43
a wide range of signals and numerous outliers (3–8% of all bins), strongly suggesting false
enrichment (Figure 2B). In the strains expressing a MOP, FLOP, or HOP internal control
construct, the ChIP signal distributions showed significant differences in comparison with the
corresponding IP in the untagged strain (Figure 2B, Table S1). Moreover, two-dimensional
scatter plots show that sequences are differentially enriched depending on the IP; for example,
the sequences with the greatest enrichment in one IP usually did not show the greatest
enrichment in different IPs (Figure 2C). Overall, the results show that significant, non-specific
signals are differentially enriched in these ChIPs, and that the sequence composition of this
enrichment is highly dependent on the specific antibody, as well as the presence of a target
protein (e.g., MOP, FLOP, HOP).

44

Figure 2. 2. Enrichment of non-specific ChIP signals in control IPs.

Figure 2.2. Enrichment of non-specific ChIP signals in control IPs. Strains MPy105 (MOP), MPy35 (FLOP),
and MPy39 (HOP) were grown to log-phase, harvested, and analyzed using ChIP-seq.(A) Plots of ChIP signals
across chromosomes XV and IV, which harbor the MOP, FLOP, and HOP loci, respectively indicated by red dots;
panels to the right show zoomed-in view. (B) Distribution boxplots of ChIP signals (500 bp bins) across the whole
genome and for specific sets of loci; results of Mann–Whitney tests of difference in distributions indicated by
asterisks ** p < 0.01. Outliers are indicated with + and the number of positive outliers is indicated for each in red.
45
(C) Two-dimensional scatter plots of ChIP signals for all bins for Untag versus MOP, FLOP, or HOP for each set of
IPs.

To further examine the nature of these non-specifically IP’d sequences, we analyzed the
ChIP signal distribution of 238 loci previously defined as hyperChIPable (Teytelman et al.,
2013), as well as tRNA genes, which are highly expressed, a characteristic that has been
correlated with non- specific ChIP enrichment. In comparison with the overall signal
distributions from each IP, signals for hyperChIPable and tRNA genes were significantly under-
enriched in untagged strains, except for hyperChIPable in IP-FLAG, and over-enriched in the
MOP, FLOP, and HOP strains, except for hyperChIPable in IP-MYC (Figure 2B, Table S1).
While this lack of significant enrichment for most of these controls might seem contrary to
expectations, we find that much variability exists in those sequences that are non-specifically
enriched in different IPs. Importantly, as shown below, enrichments in the HOP, MOP, and
FLOP strains more effectively represent non-specific enrichment present in the experimental IPs,
while the lack of enrichment of these sequences in the Untag control IPs undermines its
usefulness as a control.

Ratio Normalization by Epitope-Tagged Control Refines Data Quality for Target Proteins

Having established IPs from untagged and control strains as containing non-specifically
enriched sequences, we moved to the analysis of a target protein. We chose Fkh1 due to our
continuing interest in maximizing the accuracy of our data and, hence, our understanding of the
chromatin dynamics of this protein that has been implicated in the regulation of replication,
transcription, and recombination during mating-type switching in yeast (Jin et al., 2020). In this
regard, existing knowledge of Fkh1 binding at specific replication origins, gene promoters, and
the recombination enhancer provides bona fide loci against which to validate our new approach.
46
Additionally, we had previously analyzed Fkh1 using ChIP-chip and reported its association with
multiple loci, including Pol III-transcribed loci such as tRNAs, and considered these previous
results ripe for re-evaluation as potential hyperChIPable artifacts (Ostrow et al., 2014).
Fkh1-MYC was expressed from its endogenous locus in an otherwise untagged strain and
in the MOP control strain; similarly, Fkh1-FLAG was expressed in untagged and the FLOP
strain. These strains were subjected to ChIP-seq alongside the corresponding control strains
described above. Analysis of the replicates showed high reproducibility (Figure S3A), and
replicates were counts-normalized and averaged, after which datasets were scale-normalized at
the 20th percentile of the distributions. To apply the normalization controls against their
respective experimental sample, the Fkh1-MYC and Fkh1-FLAG experimental results were
divided by the corresponding control datasets, which we refer to as ratio normalization (RN)
(Figure 1C). We created distribution boxplots (500 bp) and heatmaps (5 kb) of the regions
surrounding chromosome features, such as “Forkhead- activated” replication origins and “CLB2-
cluster” genes, where Fkh1 is known to function and a subset of which have previously been
shown to bind Fkh1 (S. R. V. Knott et al., 2012; Ostrow et al., 2014; Zhu et al., 2000). We also
analyzed ChIP signals at tRNAs, where we previously detected the binding of Fkh1, though no
function has been ascribed, and at hyperChIPable loci, we expected to represent false- positive
artifacts of ChIP.
We begin with an analysis of Fkh1-MYC, which we analyzed previously using ChiP-
chip (Ostrow et al., 2014). Boxplots and heatmaps show the enrichment of Fkh1-MYC at
hyperChIPable and tRNA loci in both otherwise untagged and MOP strains (Figure 3).
Notably, this enrichment was also present in the MOP strain but not in the untagged strain
(Figures 2B and 3B, Table S2), such that RN effectively eliminates the enrichment of these
47
dubiously enriched loci in the MOP strain background (FKH1-MYC MOP-RN) but not in the
untagged strain background (FKH1-MYC Untag-RN) (Figure 3A,B, Table S2). ChIP signals at
tRNA genes were not significantly enriched in the untagged and MOP control strains, but RN
resulted in the significant enrichment of tRNA genes in the untagged set (UntagRN), as in the
previous study, but not in the MOP-controlled set (MOP-RN) (Figure 3A,B, Table S2). In
contrast, an examination of the sequences expected to bind Fkh1 yielded different results,
showing significant enrichments of Fkh-activated origins in both Fkh1-MYC strains that were
not eliminated but enhanced by RN (Figure 3A,B, Table S2), while CLB2-cluster genes showed
enrichment in the heatmaps after RN but did not meet statistical significance for the CLB2-
cluster genes as a group (Figure 3A,B, Table S2).
48

Figure 2. 3. Ratio normalization using control reduces non-specific enrichment.
Figure 2.3. Ratio normalization using control reduces non-specific enrichment. Strains SSy161 (WT), MPy105
(MOP), MPy166 (FKH1-9xMYC), MPy108 (FKH1-9xMYC(MOP)), MPy35 (FLOP), OAy1100
(FKH1-3xFLAG), and MPy55 (FKH1-3xFLAG(FLOP)) were grown to log-phase, harvested, and an-
alyzed using ChIP-seq. (A) Distribution boxplots of ChIP signals (500 bp bins) across the whole genome
and for specific loci; RN is the ratio normalized signal. (B) Heatmaps of averaged ChIP signal across 5
kb regions centered on the indicated features.

49

Identical analysis of the Fkh1-FLAG data yielded similar overall results with hyper-
ChIPable signals being enriched in the Fkh1-FLAG datasets and being eliminated by RN, in this
case with both corresponding control datasets, whereas tRNAs remained significantly enriched
only in the Fkh1-FLAG UntagRN (Figure 3A,B, Table S2). As above, RN did not eliminate
Fkh1-FLAG signal enrichment from the expected binding loci, including Fkh-activated origins
and CLB2-cluster genes, and again, signal enrichment at Fkh-activated genes was highly
significant but not at CLB2-cluster genes (Figure 3A,B, Table S2). Together with the MYC
epitope results, these data suggest that the application of our control paradigm significantly
improves the quality of ChIP data by eliminating likely false positives while retaining likely true
positives. In this regard, it is interesting that both Fkh1-MYC UntagRN and Fkh1-FLAG
UntagRN showed significant enrichment for tRNA genes reproducing our previously published
ChIP-chip results, whereas these significant enrichments were eliminated by MOP- and FLOP-
RN, suggesting that Fkh1 binding at tRNA genes was a ChIP artifact (Ostrow et al., 2014). We
observed similar results for snoRNAs were another category of features strongly enriched in
previous ChIP-chip analysis of Fkh1-MYC with an untagged strain as a subtraction control. As
with tRNAs, snoRNAs showed enrichment in Fkh1-MYC Untag-RN in our current experiments,
essentially re-producing the previous results; however, MOP-RN showed no enrichment for
snoRNAs, suggesting that it too was an artifact of ChIP (Figure S3C).

Fkh1 Analysis in G1 Phase to Test Known Enrichment at Replication Origins

We created chromosome plots to view the data along chromosomes III-L and IX, which
contain several Fkh1-binding loci of interest indicated on the plots (Figure S3B). In these data
from unsynchronized cultures, Fkh-activated replication origins and CLB2-cluster genes showed
50
relatively minor peaks of enrichment in comparison to the recombination enhancer (RE), which
contains numerous Fkh1 consensus binding sequences. We generated data that we could subject
to a more stringent analysis by synchronizing the FKH1-MYC(MOP) and FKH1-FLAG(FLOP)
and MOP and FLOP control strains in G1 phase with α-factor, when Fkh1 binds to a subset of
replication origins, presenting an expectation of more robust detection than in unsynchronized
cells. Experimental replicates show a high correlation (Figure S4), and an analysis of the control
IPs shows that different sequences are non-specifically enriched in the ChIP of G1 versus
asynchronous cells (Figure S2B). We generated chromosome plots to visualize the uncontrolled
and normalized data along chromosomes III-L and IX, which contain several Fkh1-binding loci
of interest indicated on the plots (Figure 4A). The plots also indicate peaks called by MACS,
which are summarized in the chart (Figure 4B). As expected, the G1 data show more called
peaks that overlap with the Fkh-activated replication origins than in the unsynchronized data.
Peak calling by MACS also independently verifies the specificity of the results and enhancement
of the data through our control approach. Peak calling is notoriously challenging and imperfect
but presents an orthogonal analysis of the data based on characteristics expected of ChIP data.
The results show that RN reduces the number of peaks called (Figure 4B). MOP- and FLOP-RN
result in sets of peaks that contain a higher proportion of expected positives, such as replication
origins, with overlaps between sets depicted by Venns (Figure 4B,C). Fewer peaks are
consistently called in the FLOP-RN versus MOP-RN datasets. We think this is due to the higher
expression of LexA-FLAG from the FLOP construct, resulting in more stringent competition for
IP. Note that virtually all Fkh1-FLAG-controlled peaks are included in the Fkh1-MYC-
controlled peak set, supporting the greater stringency of the Fkh1-FLAG set (Figure 4C).
Additionally, the larger Fkh1-MYC set captures a proportionately larger number of replication
51
origins, indicating that the larger set contains many true positives missed in the more stringent
set. These conclusions are further validated below through an alternative control approach. The
heatmaps largely recapitulate the results with unsynchronized cells, eliminating the enrichment
of hyperChIPables and tRNAs, while maintaining enrichments for CLB2-cluster genes and Fkh-
activated origins (Figure 4D).
52

Figure 2. 4. Improved detection of Fkh1 binding loci.

Figure 2.4. Improved detection of Fkh1 binding loci. Strains MPy105 (MOP), MPy108 (FKH1- 9xMYC(MOP)),
MPy35 (FLOP), and MPy55 (FKH1-3xFLAG(FLOP)) were synchronized in G1 phase, harvested, and analyzed
using ChIP-seq. (A) Plots of ChIP signal across chromosomes III-L and IX with potential Fkh1 binding sites
indicated as colored circles on one track and called peaks indicated on the lower track. (B) Stack graphs of peak
calls overlapping with potential origins according to their categorization in oriDB or with other loci presumed not to
contain replication origins. (C) Venn diagrams showing overlap of called peaks with origin sets and the MYC/MOP
53
versus FLAG/FLOP sets. (D) Heatmaps of averaged ChIP signal across 5 kb regions centered on the indicated
features.
Analysis of Replication Origin Binding Proteins Validate Approach for HA

To evaluate our approach with the HA epitope, we turned to two proteins we previously
analyzed successfully using ChIP and ChIP-chip, the replication origin binding proteins Orc1
and Mcm4, members of the ORC (Origin Recognition Complex) and MCM (Mini-Chromosome
Maintenance) complexes, respectively (O. M. Aparicio et al., 1997; Wyrick et al., 2001). We
epitope-tagged these proteins with 3xHA in the HOP strain background and performed ChIP-seq
with G1-synchronized cultures of HOP, ORC1-HA (HOP), and MCM4-HA (HOP) strains.
Experimental replicates show a high correlation (Figure S5A). Because replication origins in
yeast have been mapped and many functionally confirmed by multiple approaches, the known
origins comprise a powerful set to validate our ChIP of these proteins. We used the extensive
OriDB list of 829 origins categorized therein as confirmed (410), likely (216), and dubious (203)
(Nieduszynski et al., 2007).
We generated chromosome plots of the results before and after RN and indicated the
positions of confirmed origins on these plots as well as peaks called by MACS and generated a
chart of overlap between called peaks and origins in different categories (Figure 5A,B). Venn
diagrams show the overlaps between uncontrolled and RN datasets with origins and compare
Orc1 versus Mcm4 overlaps with and without RN (Figure 5C and Figure S5B). The peak-calling
analysis reveals a substantial reduction in the number of dubious origins called relative to the
other categories with the application of HOP control normalization (Figure 5B). For Orc1, a
similar number of confirmed origins were called peaks in the controlled (254) and uncontrolled
(268) dataset, whereas the number of “other” peaks was reduced nearly 3-fold in the RN (99)
versus uncontrolled (279) set. Similarly, in the RN versus uncontrolled data, dubious origins
were also reduced (from 56 to 22), while origin in the likely category were reduced
54
intermediately to the confirmed and dubious, as expected. Thus, the application of HOP-RN to
Orc1-HA increased the proportion of total called peaks accounted for by a confirmed or likely
origin from 52% (356/691) to 73% (319/440), while only reducing the total number of called
origins in this grouping by 10%.
For Mcm4, the overall results followed a similar pattern with the application of HOP-
RN reducing calls at “other” and “dubious” loci more than “confirmed” and “likely” origin loci.
For example, peak calls overlapping with confirmed origins decreased by 20% (from 247 to
197), while dubious and other calls decreased by 81% (from 402 to 75) (Figure 5B). There was a
lower overall number of Mcm4 than Orc1 peaks called in HOP-RN sets, which may reflect loci
that are inefficient at MCM loading (Wyrick et al., 2001). We tested whether a relationship exists
between CHIP signals for Orc1 and/or Mcm4, presumed to represent the binding occupancy of
these proteins, and the efficiencies of individual origins, as MCM stoichiometry at replication
origins has been proposed to regulate replication timing (Dukaj & Rhind, 2021). However, no
significant correlation was found between origin efficiencies and Orc1 or Mcm4 ChIP signals
(Figure S5C).
We also generated heatmaps of hyperChIPable loci, tRNAs, and origins, and observed a
modest but significant enrichment for hyperChIPable and tRNAs in ORC1-HA HOP-RN, but not
with MCM4-HA HOP-RN. Enrichments may reflect the co-localization of origins with tRNAs,
which has been previously reported (Wyrick et al., 2001). Regardless, the application of RN
reduces signal enrichment at hyperChIPable and tRNA loci, while maintaining, if not enhancing,
origin signals for both ORC1-HA and MCM4-HA (Figure 5D).
Examination of the chromosome plots of the control anti-HA IPs shows the strong enrichment of
centromere (CEN) sequences, which is also shown by a heatmap of averaged data at CEN
55
sequences (Figure 5A,D). We also noticed that signals at replication origins were under-enriched
in the HOP control dataset, contributing to an improvement in the signal upon RN (Figure 5A,D
and Figure S2C, Table S1). Though modest, a similar hypoChIPability of origins signals was
significant in the anti-MYC and anti-FLAG controls (Figure S2C, Table S1). In contrast to the
anti-HA IPs, CEN sequences were significantly under-enriched in the anti-MYC and anti-FLAG
controls (Figure S2C, Table S1).

56

Figure 2. 5. Improved accuracy in detection of ORC and MCM binding loci.

Figure 2.5. Improved accuracy in detection of ORC and MCM binding loci. Strains MPy39 (HOP), MPy199
(ORC1-3xHA(HOP)), and MPy102 (MCM4-3xHA(HOP)) were synchronized in G1 phase, harvested, and analyzed
using ChIP-seq. (A) Plots of ChIP signal across chromosomes III-L and IX with potential origins indicated as gray
circles on one track and called peaks indicated on the lower track in red and CEN indicated by arrowhead. (B) Stack
graphs of peak calls overlapping with potential origins according to their categorization in oriDB or with other loci
presumed not to contain replication origins. (C) Venn diagrams showing origin overlap of peaks called using RN
57
and uncontrolled datasets against confirmed origins. (D) Heatmaps of averaged ChIP signal across 5 kb regions
centered on the indicated features.
58
Controls Enhance Analysis of Potential HyperChIPable Loci

Given the utility of our approach in detecting and eliminating hyperChIPable signals, we
wondered how it would perform at a potentially hyperChIPable locus of interest such as the
highly expressed rDNA, which has indeed been identified as hyperChIPable (Teytelman et al.,
2013). Each rDNA repeat contains a potential replication origin, and we functionally identified
the rDNA origin(s) as Fkh-activated (S. R. V. Knott et al., 2012). In the course of these studies,
we also analyzed Sir2-3xFLAG(FLOP), which has been reported to bind and function within the
rDNA to regulate gene expression and recombination (Kueng et al., 2013). Unlike the largely
sequence-specific DNA binding of Orc1 and Fkh1, Sir2 is recruited to specific loci by other
proteins, including Orc1 indirectly, and then spreads in cis along chromatin to deacetylate
histone tails. We plotted the G1-phase ChIP-enrichment of Sir2-FLAG, Orc1-HA, Mcm4-HA,
and Fkh1-MYC at the rDNA before and after RN (Figure 6A). With the exception of Fkh1-
MYC, each of the uncontrolled ChIPs shows substantial signal enrichment across the entire
rDNA locus with several peaks and valleys. In contrast, RN normalized data show sharp, narrow
peaks at specific loci, and the substantial reduction in the signal across most of the region, with
the exception of the Sir2-FLAG data, which show a combination of sharp peaks above a high
baseline of enrichment across the rDNA region. Orc1, Mcm4, and Fkh1 peaks co- localize to the
origin sequences (ARS1200-1,2), while Sir2 spreads across the region with a major peak at the I
element and a secondary peak aligning with the replication fork barrier (RFB) (Figure 6A).
These findings are consistent with available knowledge (Huang & Moazed, 2003) and suggest
that our controls enhance the detection of bona fide protein binding even within highly expressed
loci such as the rDNA, which may exhibit non-specific enrichment. Inherent differences in
protein binding modes may also be evident in the unique spreading of the Sir2 signal not
59
observed for Orc1, Mcm4, and Fkh1 in these experiments. We also plotted Sir2 as a function of
the distance from telomeres, highlighting other known Sir2 binding loci, such as HML and HMR
(Figure 6B), and we plotted full data for chromosome III, which contain the silent loci (Figure
6C). These data show that most Sir2 enrichment occurs near telomeres, and in the vicinity of
HML and HMR where Sir2 enrichment appears to be maximal between the silencer elements.
Enrichment at HML extends to the telomere, while enrichment appears isolated from the
telomere at the more telomere-distant HMR (Figure 6C).

DNA-Binding Mutant May Be Ideal Control

We tried another approach that in principle should provide the perfect control for ChIP:
the expression of a DNA-binding mutant (dbm) of the target DNA-binding protein. Such a
control would have the same expression level, avidity for the antibody, and similar non-specific
chromatin binding properties that might contribute to a non-specific ChIP signal. The approach
requires prior knowledge and/or availability of such a mutation in the target gene. As the
Forkhead DNA-binding domain structure has been previously determined (Clark et al., 1993;
Stroud et al., 2006), it was possible for us to create a mutant allele with specific amino acid
changes in the DNA sequence recognition and binding surface, that we termed fkh1-dbm; we
confirmed the defect in DNA binding by EMSA (Figure S7A). We constructed MYC- and
FLAG-tagged versions of fkh1-dbm and replaced FKH1 and conducted ChIP-seq in
unsynchronized FKH1-MYC, fkh1-dbm-MYC, FKH1-FLAG, and fkh1-dbm-FLAG (lacking
MOP or FLOP constructs). Data were processed and analyzed as above; correlation of replicates
was confirmed (Figure S7B).
Chromosome plots and heatmaps show the complete elimination of enrichment at
specifically enriched loci, confirming the fkh1-dbm defect in vivo (Figure 7A and Figure S7C).
60
Furthermore, fkh1-dbm exhibited continued enrichment for hyperChIPable and tRNA loci. Thus,
RN using fkh1-dbm (dbm-RN) data results in the elimination of hyperChIPable signals while
retaining the enrichment of the signal at replication origins and CLB2-cluster genes (Figure 7A).
Peak-callings were performed on the RN data and compared as above. For both IPs, data
normalized by fkh1-dbm as opposed to MOP or FLOP yielded more peak calls, 795 versus 501
for MYC and 474 versus 286 for FLAG (Figure 7B). Most peaks in the MOP- or FLOP-
controlled sets were also called in the larger dbm-RN sets, which performed similarly to MOP-
RN and better than FLOP-RN at capturing Fkh-activated origins and CLB2 cluster genes in
proportion to set sizes. We conclude that the DBM control is likely a superior approach when
available.

61

Figure 2. 6. Sir2 detection within hyperChIPable rDNA locus and genome-wide.

Figure 2.6. Sir2 detection within hyperChIPable rDNA locus and genome-wide. (A) Plots of ChIP signals
overlaid across the rDNA region for G1-synchronized strains MPy108 (FKH1-9xMYC(MOP)), MPy199 (ORC1-
HA(HOP)), MPy102 (MCM4-3xHA(HOP)), and YHy29 (SIR2-3xFLAG(FLOP)) before and after RN. (B) Plot of
ChIP signals for all bins according to distance from the nearest telomere; bins mapping to the indicated loci are
highlighted. (C) Plots of ChIP signals across chromosome III, with expanded view of HML and HMR silencer
regions and silencer elements (E and I) indicated; correlation of replicates shown in Figure S6. Because the
62
reference genome is MAT α, while the analysis strain is MATa, we deleted α gene sequences at the MAT locus in
the reference sequence to properly map all α gene sequence reads to HML α.
63

Figure 2. 7. DNA-binding mutant control may be ideal.

Figure 2.7. DNA-binding mutant control may be ideal. Strains MPy166 (FKH1-9xMYC), MPy169 (fkh1- dbm-
9xMYC), MPy172 (fkh1-dbm-3xFLAG), and OAy1100 (FKH1-3xFLAG) were grown to log-phase, harvested, and
analyzed using ChIP-seq. (A) Heatmaps of averaged ChIP signal across 5 kb regions centered on the indicated
features. (B) Venn diagrams showing overlap of called peaks with origin and CLB2 cluster sets comparing dbm-
controlled against MOP- and FLOP-controlled sets analyzed in Figure 3.

64
DISCUSSION

Expression of a Decoy Protein to Control for Non-Specific Sequence Enrichment

Our results show that the in vivo expression of a heterologous or non-DNA-binding
protein sharing an epitope tag with the experimental target protein provides an effective sensor
for ChIP to identify non-specific signals for use in data normalization, and thereby can greatly
enhance the quality of ChIP-seq analysis. Our approach requires data from two ChIPs (each of
which should be replicated): the control strain expressing the “mock” binding protein (e.g.,
MOP, HOP, FLOP) and the experimental strain also expressing the protein of interest tagged
with the same epitope as the control protein. These datasets are scale-normalized, and the
experimental set is divided by the control set, yielding ratio- normalized data. We validated our
data through a series of experiments using proteins with known targets. Our results yield new,
high-quality data for Fkh1 in unsynchronized cells, and Fkh1, Orc1, Mcm4, and Sir2 in G1-
synchronized cells. More broadly, our results validate an experimental control paradigm and
provide constructs for working with epitope-tagged proteins in ChIP.
We chose LexA due to its natural absence from yeast along with its established use as a
robust and specific DNA-binding protein to its specific target sequence (Schnarr et al., 1991); it
was also convenient as we already possessed DNA sequences containing elements for expression
and binding to use for the control constructs. We see no reason why other proteins could not be
similarly used, and they actually might function better. In fact, we show that a potentially ideal
control, if available knowledge and/or resources exist, is the expression of a DNA/chromatin-
binding defective version of the target protein to use for ratio normalization.
65
Our findings strongly emphasize the value of the efforts required to implement these
controls. For proteins that are recruited to chromatin by other factors, ChIP in cells depleted of
the recruitment factor(s) should provide a rigorous test of specificity. Inclusion in the expression
vector of lexA DNA-binding sequences provides a positive control for the expressed LexA
protein in the ChIP with potential information about the resolution of the data. In addition, these
internal controls can serve as internal standards against which other loci may be measured, or to
confirm alternative normalization methods.

Multiple Factors Contribute to Non-Specific ChIP-Signal Enrichment

Analysis of non-specific signal enrichment showed that numerous factors influence such
enrichment, including the antibody used for IP, the expression of antibody target protein(s), both
control and experimental, and the epitope tags. The Fkh1 analysis suggests that IPs against MYC
were more specific than against FLAG; however, it should be recalled that the number of repeats
for the tags varied. In addition, the cell cycle stage of analysis also impacted non-specific
sequence enrichment. For example, the hyper-enrichment of CENs in anti-HA IPs was much
enhanced in G1 phase. In contrast, CENs were hypo-enriched in the FLAG and HA IPs and
replication origins also in all three IPs. Overall, we showed that carrying out ChIP with
appropriate control strains identifies these biases in the data, enabling their effective subtraction
from the data to yield more accurate results.
Our results indicate that the presence of a target protein for the antibody alters non-
specific IP, probably at least in part due to competition for binding the antibody. Compared to an
untagged strain, the strains expressing the heterologous epitope-tagged protein appear to more
greatly enrich hyperChIPable or other non-specific loci that are also likely to be enriched in the
ChIP against the protein of interest, and this appears to be key to their action here. It is possible
66
and likely that the control protein IP enriches for additional, non-specific sequences than the
target protein IP, which is why the experimental strain also expresses the control protein. In this
regard, we think the dbm approach is most ideal because the dbm protein is biochemically
identical to the target protein of interest, save a few amino acid changes disrupting high-avidity,
sequence-specific binding. Thus, other interactions that might contribute to non-specific IP will
likely be retained, allowing their subtraction, with less risk of adding new, non-specific targets of
IP.

No Control, No Experiment

The powerful insight provided by bona fide ChIP results in identifying a protein-
DNA/chromatin interaction in vivo, which has made such analysis de rigueur. However, ChIP
applications, especially for new discoveries, are fraught with uncertainty regarding absent
stringent attention to controls to validate results and, as we have shown, to correct for intrinsic,
yet unpredictable sequence enrichment biases. Even with a previous application of untagged
strain controls, we failed to correct for biased enrichment, an error we have now rectified. The
obvious controls are critical to attempt, but the best controls are not always obvious or feasible.
With the application of effective controls as we describe here, ChIP datasets can be filtered of
contaminants and made more inherently valuable for all downstream uses.

67
MATERIALS AND METHODS

Plasmid Constructions

Primer sequences and descriptions are given in Table S3. Plasmids pADE2-MOP, pTRP1-HOP,
and pADE2-FLOPv2x2 (herein otherwise referred to as pADE2-FLOP) were constructed by
Gibson assembly (SGI #GA1200) and/or standard restriction endonuclease digestions and
ligations (New England Biolabs, Ipswich, MA, USA), according to the manufacturer’s protocols.
13xMYC, 3xHA epitope tags, and TEF1 promoter sequences were PCR amplified from pFA6
vectors (Longtine et al., 1998); 3xFLAG was amplified from p2L-3Flag-TRP1 (T. Tsukiyama,
Seattle, WA, USA); and ADH1 sequences and lexA protein-coding and lex Operator binding
sequences were amplified from pJL4 and pEC15 (Li et al., 2012). Partial ADE2 and TRP1 target
sequences (from BY4741) for genome targeting and selection were designed to integrate the
constructs stably into the homologous regions, bearing the auxotrophic mutations in the W303
background (i.e., ade2-1 and trp1-1). His6-tagged protein expression and the construction of
p405-FKH1 has been described previously (Ostrow et al., 2017). Site-directed mutagenesis was
used to produce pET28a-Fkh1-dbm using Quickchange Multi kit Agilent, Santa Clara, CA,
USA). Plasmid sequences were confirmed by DNA sequencing (Retrogen Inc.).

Yeast Strain Constructions

Primer sequences are given in Table S3. Genotypes of all yeast strains are given in Table S4.
ADE2-MOP/FLOP and TRP1-HOP constructs were liberated from plasmid vectors by digestion
with SacI + KpnI and transformed into yeast with the selection of -ade or -trp medium as
appropriate. DNA was introduced into yeast by lithium acetate transformation with an
appropriate selection (Gietz & Schiestl, 2007) and was confirmed by PCR. The expression of all
68
epitope-tagged proteins was confirmed by Western blot analysis of whole cell protein extracts
produced by TCA precipitation, using the following monoclonal antibodies: anti-FLAG (Sigma
M2), anti-MYC (BioLegend 9E10), or anti-HA (12CA5 or 16B12).

Other Methods

Yeast cultures were grown at 23°C in YEPD and harvested during logarithmic growth or after 3
h of growth in the presence of 5 nM α-factor. ChIP-Seq was performed in triplicate for each
strain using the antibodies listed above at 1:200 according to our protocol and analysis pipeline,
with an additional DNA shearing performed of the purified DNA before library preparation
(Haye-Bertolozzi & Aparicio, 2018; Ostrow et al., 2015). Amplified libraries were subjected to
final quality control and quantifications for high-throughput sequencing by Illumina technology
(150bp paired-end), carried out by Novogene. Sequencing data is available at GEO
(GSE230475). The list of hyperChIPable sequences is from Knott et al., (2012); the list of
origins is from oriDB (cere-visiae.oridb.org); the list of Fkh-activated origins is from Knott et al.,
(2012). S. cerevisiae genome sequence (last modified date: 2019-10-25T18:55:47.000Z) and
element assign-ments (e.g.: tRNAs, CENs) are from Saccharomyces Genome Database
(yeastgenome.org). Origin efficiencies are from McGuffee, Smith, and Whitehouse (2013).
Electrophoretic Mobility Shift Analysis (EMSA) was performed as follows: TEM1 DNA probe
was produced by annealing oligonucleotide primers (Table S3). On ice, 10ng of DNA probe was
combined with protein in 20 uL total volume in 20 mM Tris-HCl pH 7.9, 50mM KCl, 5mM
MgCl2, 3mM DTT, 0.1mg/ml BSA, 10% (v/v) glycerol. Binding reactions were incubated on ice
for 15 min followed by incubation at room temperature (22°C) for 15 min. Samples were loaded
onto a 10% (w/v) (21:1, acrylamide:bis-acrylamide) polyacrylamide gel and separated in a 0.5x
TBE buffer at 120V for 90 min in a 4°C environ-mental chamber. The gel was stained in
69
0.5xTBE, 0.2nM SYBR Green I (Molecular Probes) 10 min at 22°C with gentle mixing. The gel
was de-stained by incubating in H2O 10 min at 22°C with gentle mixing and repeating. The gel
image was captured on a BioRad FX scanner.

70
APPENDIX

Peak detection using ratio normalized ChIP-seq dataset
As discussed in Chapter II, the ChIP targeting the non-genome-binding protein LexA
roughly mimics the non-specific enrichment observed in the ChIP of the protein of interest, due
to the shared epitope tags. Therefore, when subjecting the LexA-carrying strain to ChIP-seq
alongside the corresponding experimental strain, it serves as a control for normalizing the
experimental data. This normalization process involves dividing the experimental data by the
corresponding control data, referred to as ratio normalization (RN). Through this normalization
approach, the background non-specific signals are effectively eliminated, allowing us to obtain a
relatively accurate protein binding profile across the genome.
The current ChIP-seq peak callers available are not fully compatible with data processed
using our ratio normalization (RN) method. Therefore, in order to leverage the data obtained
through this new RN approach and to enhance the accuracy of our existing protein binding list, I
have built upon the foundation of our lab's previous research and developed two distinct peak
calling strategies (S. R. Knott et al., 2009). The initial approach concentrates on defining the
shape of ChIP peaks, whereas the second approach involves fitting Hidden Markov machine-
learning Models (HMMs) to the processed data. These methodologies establish a crucial
connection between our pioneering data normalization technique and the overarching goal of
achieving precise peak identification.
The first approach is typically implemented by employing a sliding window across the
smoothed and ratio-normalized whole-genome sequencing data. This sliding window scans the
ChIP-seq data to identify segments that exhibit a continuous increase followed by a continuous
decrease in signals, while also permitting dips within the window. Such segments, due to their
71
adherence to the shape of a peak, can potentially represent actual ChIP enrichments. Because the
core function of this method revolves around detecting consecutive upward and downward trends
in data points, it is commonly referred to as the "consecutive-bin method." In practical
application, we begin by employing a grid search to explore various hyperparameters, such as the
sliding window size, the number of consecutive bins, the permissible number of dips, and the
extent of smoothing, in order to identify an optimal combination. Once these parameters are set,
we can apply the consecutive-bin method to uncover potential peaks. Subsequently, these peaks
undergo a series of filtration steps. If they meet specific criteria, such as exhibiting experimental
signals that are statistically significantly higher than the local background signal and surpassing a
predefined threshold, they are recorded as true ChIP enrichments (Fig. A1).

Figure A 1. The demonstration of the “consecutive-bin method” in peak detection.
Figure A1. The demonstration of the “consecutive-bin method” in peak detection.

The second approach employs Hidden Markov Models (HMMs), a statistical framework
well-suited for modeling sequential data and useful for identifying hidden states, which are not
directly observed, within an observable sequence. HMMs are particularly applicable to peak
calling tasks, where the objective is to distinguish enriched regions from background noise
72
(unenriched regions) by leveraging the sequential ChIP signals (Fig. A2). In this approach,
similar to the previous method, a sliding window is applied across the ratio-normalized data
segments. This adaptation is necessary because HMM models are highly sensitive to the length
of input information. During each iteration of the window, the data encompassed by the window
are inputted into the HMM. Subsequently, the HMM classifies the window by assigning it to
either an enriched or unenriched category based on the probabilities of its hidden states. The
segments of the identified enriched regions are then logged as ChIP peaks.

Figure A 2. The demonstration of the HMM model in peak detection.
Figure A2. The demonstration of the HMM model in peak detection.

We have validated these two peak detection approaches using several ChIP-seq datasets.
One of the datasets that is especially suitable for this validation purpose is the Mcm4 ChIP in the
G1 phase. Since Mcm4 is recruited to potential origins during the G1 phase, and the positions of
origins on the yeast genome are well studied and documented, we can employ the known origin
positions for validation to assess the accuracy of our peak detection methods. Results showed
that the peak detection results largely agree with each other (Fig A3).
73

Figure A3. The comparison of the peak detection results between the two approaches.
Figure A 3. The comparison of the peak detection results between the two approaches.

74
REFERENCES

Aparicio, J. G., Viggiani, C. J., Gibson, D. G., & Aparicio, O. M. (2004). The Rpd3-Sin3
Histone Deacetylase Regulates Replication Timing and Enables Intra-S Origin Control in
Saccharomyces cerevisiae. Molecular and Cellular Biology, 24(11), 4769–4780.
https://doi.org/10.1128/MCB.24.11.4769-4780.2004

Aparicio, O. M. (2013). Location, location, location: It’s all in the timing for replication origins.
Genes & Development, 27(2), 117–128. https://doi.org/10.1101/gad.209999.112

Aparicio, O. M., Weinstein, D. M., & Bell, S. P. (1997). Components and Dynamics of DNA
Replication Complexes in S. cerevisiae: Redistribution of MCM Proteins and Cdc45p
during S Phase. Cell, 91(1), 59–69. https://doi.org/10.1016/S0092-8674(01)80009-X

Bell, S. P., & Dutta, A. (2002). DNA Replication in Eukaryotic Cells. Annual Review of
Biochemistry, 71(1), 333–374.
https://doi.org/10.1146/annurev.biochem.71.110601.135425

Brewer, B. J., Lockshon, D., & Fangman, W. L. (1992). The arrest of replication forks in the
rDNA of yeast occurs independently of transcription. Cell, 71(2), 267–276.
https://doi.org/10.1016/0092-8674(92)90355-G

Clark, K. L., Halay, E. D., Lai, E., & Burley, S. K. (1993). Co-crystal structure of the HNF-
3/fork head DNA-recognition motif resembles histone H5. Nature, 364(6436), 412–420.
https://doi.org/10.1038/364412a0
75
Cvetic, C., & Walter, J. (2005). Eukaryotic origins of DNA replication: Could you please be
more specific? Seminars in Cell & Developmental Biology, 16(3), 343–353.
https://doi.org/10.1016/j.semcdb.2005.02.009

Duan, Z., Andronescu, M., Schutz, K., McIlwain, S., Kim, Y. J., Lee, C., Shendure, J., Fields, S.,
Blau, C. A., & Noble, W. S. (2010). A three-dimensional model of the yeast genome.
Nature, 465(7296), 363–367. https://doi.org/10.1038/nature08973

Dubey, D. D., Davis, L. R., Greenfeder, S. A., Ong, L. Y., Zhu, J., Broach, J. R., Newlon, C. S.,
& Huberman, J. A. (1991). Evidence Suggesting that the ARS Elements Associated with
Silencers of the Yeast Mating-Type Locus HML Do Not Function as Chromosomal DNA
Replication Origins. Molecular and Cellular Biology, 11(10), 5346–5355.
https://doi.org/10.1128/mcb.11.10.5346-5355.1991

Dukaj, L., & Rhind, N. (2021). The capacity of origins to load MCM establishes replication
timing patterns. PLOS Genetics, 17(3), e1009467.
https://doi.org/10.1371/journal.pgen.1009467

Fan, X., & Struhl, K. (2009). Where Does Mediator Bind In Vivo? PLoS ONE, 4(4), e5029.
https://doi.org/10.1371/journal.pone.0005029

Fang, D., Lengronne, A., Shi, D., Forey, R., Skrzypczak, M., Ginalski, K., Yan, C., Wang, X.,
Cao, Q., Pasero, P., & Lou, H. (2017). Dbf4 recruitment by forkhead transcription factors
defines an upstream rate-limiting step in determining origin firing timing. Genes &
Development, 31(23–24), 2405–2415. https://doi.org/10.1101/gad.306571.117
76
Fangman, W. L., & Brewer, B. J. (1991). Activation of Replication Origins within Yeast
Chromosomes. Annual Review of Cell Biology, 7(1), 375–402.
https://doi.org/10.1146/annurev.cb.07.110191.002111

Ferguson, B. M., & Fangman, W. L. (1992). A position effect on the time of replication origin
activation in yeast. Cell, 68(2), 333–339. https://doi.org/10.1016/0092-8674(92)90474-Q

Friedman, K. L., Diller, J. D., Ferguson, B. M., Nyland, S. V., Brewer, B. J., & Fangman, W. L.
(1996). Multiple determinants controlling activation of yeast replication origins late in S
phase. Genes & Development, 10(13), 1595–1607.
https://doi.org/10.1101/gad.10.13.1595

Ge, X. Q., Jackson, D. A., & Blow, J. J. (2007). Dormant origins licensed by excess Mcm2–7 are
required for human cells to survive replicative stress. Genes & Development, 21(24),
3331–3341. https://doi.org/10.1101/gad.457807

Gibson, D. G., Aparicio, J. G., Hu, F., & Aparicio, O. M. (2004). Diminished S-Phase Cyclin-
Dependent Kinase Function Elicits Vital Rad53-Dependent Checkpoint Responses in
Saccharomyces cerevisiae. Molecular and Cellular Biology, 24(23), 10208–10222.
https://doi.org/10.1128/MCB.24.23.10208-10222.2004

Gietz, R. D., & Schiestl, R. H. (2007). High-efficiency yeast transformation using the LiAc/SS
carrier DNA/PEG method. Nature Protocols, 2(1), 31–34.
https://doi.org/10.1038/nprot.2007.13
77
Goldstein, A. L., Pan, X., & McCusker, J. H. (1999). Heterologous URA3MX cassettes for gene
replacement in Saccharomyces cerevisiae. Yeast, 15(6), 507–511.
https://doi.org/10.1002/(SICI)1097-0061(199904)15:6<507::AID-YEA369>3.0.CO;2-P

Haye-Bertolozzi, J. E., & Aparicio, O. M. (2018). Quantitative Bromodeoxyuridine
Immunoprecipitation Analyzed by High-Throughput Sequencing (qBrdU-Seq or QBU).
In M. Muzi-Falconi & G. W. Brown (Eds.), Genome Instability (Vol. 1672, pp. 209–
225). Springer New York. https://doi.org/10.1007/978-1-4939-7306-4_16

Heller, R. C., Kang, S., Lam, W. M., Chen, S., Chan, C. S., & Bell, S. P. (2011). Eukaryotic
Origin-Dependent DNA Replication In Vitro Reveals Sequential Action of DDK and S-
CDK Kinases. Cell, 146(1), 80–91. https://doi.org/10.1016/j.cell.2011.06.012

Heun, P., Laroche, T., Raghuraman, M. K., & Gasser, S. M. (2001). The positioning and
dynamics of origins of replication in the budding yeast nucleus. The Journal of Cell
Biology, 152(2), 385–400. https://doi.org/10.1083/jcb.152.2.385

Hiratani, I., & Gilbert, D. M. (2009). Replication timing as an epigenetic mark. Epigenetics, 4(2),
93–97. https://doi.org/10.4161/epi.4.2.7772

Huang, J., & Moazed, D. (2003). Association of the RENT complex with nontranscribed and
coding regions of rDNA and a regional requirement for the replication fork block protein
Fob1 in rDNA silencing. Genes & Development, 17(17), 2162–2176.
https://doi.org/10.1101/gad.1108403

78
Hyrien, O., & Goldar, A. (2010). Mathematical modelling of eukaryotic DNA replication.
Chromosome Research, 18(1), 147–161. https://doi.org/10.1007/s10577-009-9092-4

Ibarra, A., Schwob, E., & Méndez, J. (2008). Excess MCM proteins protect human cells from
replicative stress by licensing backup origins of replication. Proceedings of the National
Academy of Sciences, 105(26), 8956–8961. https://doi.org/10.1073/pnas.0803978105

Ito, H., Fukuda, Y., Murata, K., & Kimura, A. (1983). Transformation of intact yeast cells
treated with alkali cations. Journal of Bacteriology, 153(1), 163–168.
https://doi.org/10.1128/jb.153.1.163-168.1983

Jin, Y., Liang, Z., & Lou, H. (2020). The Emerging Roles of Fox Family Transcription Factors in
Chromosome Replication, Organization, and Genome Stability. Cells, 9(1), 258.
https://doi.org/10.3390/cells9010258

Jones, D. R., Prasad, A. A., Chan, P. K., & Duncker, B. P. (2010). The Dbf4 motif C zinc finger
promotes DNA replication and mediates resistance to genotoxic stress. Cell Cycle, 9(10),
2018–2026. https://doi.org/10.4161/cc.9.10.11752

Kitamura, E., Blow, J. J., & Tanaka, T. U. (2006). Live-Cell Imaging Reveals Replication of
Individual Replicons in Eukaryotic Replication Factories. Cell, 125(7), 1297–1308.
https://doi.org/10.1016/j.cell.2006.04.041

Knott, S. R. V., Peace, J. M., Ostrow, A. Z., Gan, Y., Rex, A. E., Viggiani, C. J., Tavaré, S., &
Aparicio, O. M. (2012). Forkhead Transcription Factors Establish Origin Timing and
79
Long-Range Clustering in S. cerevisiae. Cell, 148(1–2), 99–111.
https://doi.org/10.1016/j.cell.2011.12.012

Knott, S. R. V., Viggiani, C. J., Tavaré, S., & Aparicio, O. M. (2009). Genome-wide replication
profiles indicate an expansive role for Rpd3L in regulating replication initiation timing or
efficiency, and reveal genomic loci of Rpd3 function in Saccharomyces cerevisiae. Genes
& Development, 23(9), 1077–1090. https://doi.org/10.1101/gad.1784309

Knott, S. R., Viggiani, C. J., Aparicio, O. M., & Tavaré, S. (2009). Strategies for analyzing
highly enriched IP-chip datasets. BMC Bioinformatics, 10(1), 305.
https://doi.org/10.1186/1471-2105-10-305

Kueng, S., Oppikofer, M., & Gasser, S. M. (2013). SIR Proteins and the Assembly of Silent
Chromatin in Budding Yeast. Annual Review of Genetics, 47(1), 275–306.
https://doi.org/10.1146/annurev-genet-021313-173730

Kwan, E. X., Foss, E. J., Tsuchiyama, S., Alvino, G. M., Kruglyak, L., Kaeberlein, M.,
Raghuraman, M. K., Brewer, B. J., Kennedy, B. K., & Bedalov, A. (2013). A Natural
Polymorphism in rDNA Replication Origins Links Origin Activation with Calorie
Restriction and Lifespan. PLoS Genetics, 9(3), e1003329.
https://doi.org/10.1371/journal.pgen.1003329

Labib, K. (2010). How do Cdc7 and cyclin-dependent kinases trigger the initiation of
chromosome replication in eukaryotic cells? Genes & Development, 24(12), 1208–1219.
https://doi.org/10.1101/gad.1933010
80
Li, J., Coïc, E., Lee, K., Lee, C.-S., Kim, J.-A., Wu, Q., & Haber, J. E. (2012). Regulation of
Budding Yeast Mating-Type Switching Donor Preference by the FHA Domain of Fkh1.
PLoS Genetics, 8(4), e1002630. https://doi.org/10.1371/journal.pgen.1002630

Lindström, M. S., Jurada, D., Bursac, S., Orsolic, I., Bartek, J., & Volarevic, S. (2018).
Nucleolus as an emerging hub in maintenance of genome stability and cancer
pathogenesis. Oncogene, 37(18), 2351–2366. https://doi.org/10.1038/s41388-017-0121-z

Linskens, M. H. K., & Huberman, J. A. (1988). Organization of Replication of Ribosomal DNA
in Saccharomyces cerevisiae. Molecular and Cellular Biology, 8(11), 4927–4935.
https://doi.org/10.1128/mcb.8.11.4927-4935.1988

Longtine, M. S., Mckenzie Iii, A., Demarini, D. J., Shah, N. G., Wach, A., Brachat, A.,
Philippsen, P., & Pringle, J. R. (1998). Additional modules for versatile and economical
PCR-based gene deletion and modification in Saccharomyces cerevisiae. Yeast, 14(10),
953–961. https://doi.org/10.1002/(SICI)1097-0061(199807)14:10<953::AID-
YEA293>3.0.CO;2-U

Lynch, K. L., Alvino, G. M., Kwan, E. X., Brewer, B. J., & Raghuraman, M. K. (2019). The
effects of manipulating levels of replication initiation factors on origin firing efficiency in
yeast. PLOS Genetics, 15(10), e1008430. https://doi.org/10.1371/journal.pgen.1008430

MacAlpine, D. M., & Bell, S. P. (2005). A genomic view of eukaryotic DNA replication.
Chromosome Research, 13(3), 309–326. https://doi.org/10.1007/s10577-005-1508-1

81
Mantiero, D., Mackenzie, A., Donaldson, A., & Zegerman, P. (2011). Limiting replication
initiation factors execute the temporal programme of origin firing in budding yeast:
Limiting replication factors execute temporal programme. The EMBO Journal, 30(23),
4805–4814. https://doi.org/10.1038/emboj.2011.404

Meister, P., Taddei, A., Ponti, A., Baldacci, G., & Gasser, S. M. (2007). Replication foci
dynamics: Replication patterns are modulated by S-phase checkpoint kinases in fission
yeast. The EMBO Journal, 26(5), 1315–1326. https://doi.org/10.1038/sj.emboj.7601538

Miller, C. (1999). The inefficient replication origin from yeast ribosomal DNA is naturally
impaired in the ARS consensus sequence and in DNA unwinding. Nucleic Acids
Research, 27(19), 3921–3930. https://doi.org/10.1093/nar/27.19.3921

Murakami, H., Aiba, H., Nakanishi, M., & Murakami-Tonami, Y. (2010). Regulation of yeast
forkhead transcription factors and FoxM1 by cyclin-dependent and polo-like kinases.
Cell Cycle, 9(16), 3253–3262. https://doi.org/10.4161/cc.9.16.12599

Nakato, R., & Sakata, T. (2021). Methods for ChIP-seq analysis: A practical workflow and
advanced applications. Methods, 187, 44–53.
https://doi.org/10.1016/j.ymeth.2020.03.005

Nakato, R., & Shirahige, K. (2016). Recent advances in ChIP-seq analysis: From quality
management to whole-genome annotation. Briefings in Bioinformatics, bbw023.
https://doi.org/10.1093/bib/bbw023
82
Newlon, C. S., & Theis, J. F. (1993). The structure and function of yeast ARS elements. Current
Opinion in Genetics & Development, 3(5), 752–758. https://doi.org/10.1016/S0959-
437X(05)80094-2

Nieduszynski, C. A., Hiraga, S., Ak, P., Benham, C. J., & Donaldson, A. D. (2007). OriDB: A
DNA replication origin database. Nucleic Acids Research, 35(suppl_1), D40–D46.
https://doi.org/10.1093/nar/gkl758

Nomura, M. (2001). Ribosomal RNA Genes, RNA Polymerases, Nucleolar Structures, and
Synthesis of rRNA in the Yeast Saccharomyces cerevisiae. Cold Spring Harbor
Symposia on Quantitative Biology, 66(0), 555–566.
https://doi.org/10.1101/sqb.2001.66.555

Ostrow, A. Z., Kalhor, R., Gan, Y., Villwock, S. K., Linke, C., Barberis, M., Chen, L., &
Aparicio, O. M. (2017). Conserved forkhead dimerization motif controls DNA replication
timing and spatial organization of chromosomes in S. cerevisiae. Proceedings of the
National Academy of Sciences, 114(12). https://doi.org/10.1073/pnas.1612422114

Ostrow, A. Z., Nellimoottil, T., Knott, S. R. V., Fox, C. A., Tavaré, S., & Aparicio, O. M.
(2014). Fkh1 and Fkh2 Bind Multiple Chromosomal Elements in the S. cerevisiae
Genome with Distinct Specificities and Cell Cycle Dynamics. PLoS ONE, 9(2), e87647.
https://doi.org/10.1371/journal.pone.0087647

Ostrow, A. Z., Viggiani, C. J., Aparicio, J. G., & Aparicio, O. M. (2015). ChIP-Seq to Analyze
the Binding of Replication Proteins to Chromatin. In S. Vengrova & J. Dalgaard (Eds.),
83
DNA Replication (Vol. 1300, pp. 155–168). Springer New York.
https://doi.org/10.1007/978-1-4939-2596-4_11

Park, P. J. (2009). ChIP–seq: Advantages and challenges of a maturing technology. Nature
Reviews Genetics, 10(10), 669–680. https://doi.org/10.1038/nrg2641

Pasero, P., Bensimon, A., & Schwob, E. (2002). Single-molecule analysis reveals clustering and
epigenetic regulation of replication origins at the yeast rDNA locus. Genes &
Development, 16(19), 2479–2484. https://doi.org/10.1101/gad.232902

Patel, P. K., Kommajosyula, N., Rosebrock, A., Bensimon, A., Leatherwood, J., Bechhoefer, J.,
& Rhind, N. (2008). The Hsk1(Cdc7) Replication Kinase Regulates Origin Efficiency.
Molecular Biology of the Cell, 19(12), 5550–5558. https://doi.org/10.1091/mbc.e08-06-
0645

Peace, J. M., Villwock, S. K., Zeytounian, J. L., Gan, Y., & Aparicio, O. M. (2016). Quantitative
BrdU immunoprecipitation method demonstrates that Fkh1 and Fkh2 are rate-limiting
activators of replication origins that reprogram replication timing in G1 phase. Genome
Research, 26(3), 365–375. https://doi.org/10.1101/gr.196857.115

Rao, H., & Stillman, B. (1995). The origin recognition complex interacts with a bipartite DNA
binding site within yeast replicators. Proceedings of the National Academy of Sciences,
92(6), 2224–2228. https://doi.org/10.1073/pnas.92.6.2224

Reinapae, A., Jalakas, K., Avvakumov, N., Lõoke, M., Kristjuhan, K., & Kristjuhan, A. (2017).
Recruitment of Fkh1 to replication origins requires precisely positioned Fkh1/2 binding
84
sites and concurrent assembly of the pre-replicative complex. PLOS Genetics, 13(1),
e1006588. https://doi.org/10.1371/journal.pgen.1006588

Rhind, N., & Gilbert, D. M. (2013). DNA Replication Timing. Cold Spring Harbor Perspectives
in Biology, 5(8), a010132–a010132. https://doi.org/10.1101/cshperspect.a010132

Santocanale, C., & Diffley, J. F. X. (1998). A Mec1- and Rad53-dependent checkpoint controls
late-firing origins of DNA replication. Nature, 395(6702), 615–618.
https://doi.org/10.1038/27001

Schnarr, M., Oertel-Buchheit, P., Kazmaier, M., & Granger-Schnarr, M. (1991). DNA binding
properties of the LexA repressor. Biochimie, 73(4), 423–431.
https://doi.org/10.1016/0300-9084(91)90109-E

Smith, J. S., Caputo, E., & Boeke, J. D. (1999). A Genetic Screen for Ribosomal DNA Silencing
Defects Identifies Multiple DNA Replication and Chromatin-Modulating Factors.
Molecular and Cellular Biology, 19(4), 3184–3197.
https://doi.org/10.1128/MCB.19.4.3184

Soriano, I., Morafraile, E. C., Vázquez, E., Antequera, F., & Segurado, M. (2014). Different
nucleosomal architectures at early and late replicating origins in Saccharomyces
cerevisiae. BMC Genomics, 15(1), 791. https://doi.org/10.1186/1471-2164-15-791

Stevenson, J. B., & Gottschling, D. E. (1999). Telomeric chromatin modulates replication timing
near chromosome ends. Genes & Development, 13(2), 146–151.
https://doi.org/10.1101/gad.13.2.146
85
Stinchcomb, D. T., Struhl, K., & Davis, R. W. (1979). Isolation and characterisation of a yeast
chromosomal replicator. Nature, 282(5734), 39–43. https://doi.org/10.1038/282039a0

Stroud, J. C., Wu, Y., Bates, D. L., Han, A., Nowick, K., Paabo, S., Tong, H., & Chen, L. (2006).
Structure of the Forkhead Domain of FOXP2 Bound to DNA. Structure, 14(1), 159–166.
https://doi.org/10.1016/j.str.2005.10.005

Tanaka, S., Nakato, R., Katou, Y., Shirahige, K., & Araki, H. (2011). Origin Association of Sld3,
Sld7, and Cdc45 Proteins Is a Key Step for Determination of Origin-Firing Timing.
Current Biology, 21(24), 2055–2063. https://doi.org/10.1016/j.cub.2011.11.038

Teytelman, L., Thurtle, D. M., Rine, J., & Van Oudenaarden, A. (2013). Highly expressed loci
are vulnerable to misleading ChIP localization of multiple unrelated proteins.
Proceedings of the National Academy of Sciences, 110(46), 18602–18607.
https://doi.org/10.1073/pnas.1316064110

Tsang, E., & Carr, A. (2008). Replication fork arrest, recombination and the maintenance of
ribosomal DNA stability. DNA Repair, 7(10), 1613–1623.
https://doi.org/10.1016/j.dnarep.2008.06.010

Viggiani, C. J., & Aparicio, O. M. (2006). New vectors for simplified construction of BrdU-
Incorporating strains of Saccharomyces cerevisiae. Yeast, 23(14–15), 1045–1051.
https://doi.org/10.1002/yea.1406

86
Villwock, S. K., & Aparicio, O. M. (2014). Two-Dimensional Agarose Gel Electrophoresis for
Analysis of DNA Replication. In J. S. Smith & D. J. Burke (Eds.), Yeast Genetics (Vol.
1205, pp. 329–340). Springer New York. https://doi.org/10.1007/978-1-4939-1363-3_19

Vogelauer, M., Rubbi, L., Lucas, I., Brewer, B. J., & Grunstein, M. (2002). Histone Acetylation
Regulates the Time of Replication Origin Firing. Molecular Cell, 10(5), 1223–1233.
https://doi.org/10.1016/S1097-2765(02)00702-5

Wreczycka, K., Franke, V., Uyar, B., Wurmus, R., Bulut, S., Tursun, B., & Akalin, A. (2019).
HOT or not: Examining the basis of high-occupancy target regions. Nucleic Acids
Research, 47(11), 5735–5745. https://doi.org/10.1093/nar/gkz460

Wyrick, J. J., Aparicio, J. G., Chen, T., Barnett, J. D., Jennings, E. G., Young, R. A., Bell, S. P.,
& Aparicio, O. M. (2001). Genome-Wide Distribution of ORC and MCM Proteins in S.
cerevisiae: High-Resolution Mapping of Replication Origins. Science, 294(5550), 2357–
2360. https://doi.org/10.1126/science.1066101

Yabuki, N., Terashima, H., & Kitada, K. (2002). Mapping of early firing origins on a replication
profile of budding yeast: Genome-wide mapping of early firing origins. Genes to Cells,
7(8), 781–789. https://doi.org/10.1046/j.1365-2443.2002.00559.x

Yoshida, K., Bacal, J., Desmarais, D., Padioleau, I., Tsaponina, O., Chabes, A., Pantesco, V.,
Dubois, E., Parrinello, H., Skrzypczak, M., Ginalski, K., Lengronne, A., & Pasero, P.
(2014). The Histone Deacetylases Sir2 and Rpd3 Act on Ribosomal DNA to Control the
Replication Program in Budding Yeast. Molecular Cell, 54(4), 691–697.
https://doi.org/10.1016/j.molcel.2014.04.032
87
Zhang, H., Petrie, M. V., He, Y., Peace, J. M., Chiolo, I. E., & Aparicio, O. M. (2019). Dynamic
relocalization of replication origins by Fkh1 requires execution of DDK function and
Cdc45 loading at origins. ELife, 8, e45512. https://doi.org/10.7554/eLife.45512

Zhao, P. A., Rivera-Mulia, J. C., & Gilbert, D. M. (2017). Replication Domains: Genome
Compartmentalization into Functional Replication Units. In H. Masai & M. Foiani (Eds.),
DNA Replication (Vol. 1042, pp. 229–257). Springer Singapore.
https://doi.org/10.1007/978-981-10-6955-0_11

Zhu, G., Spellman, P. T., Volpe, T., Brown, P. O., Botstein, D., Davis, T. N., & Futcher, B.
(2000). Two yeast forkhead genes regulate the cell cycle and pseudohyphal growth.
Nature, 406(6791), 90–94. https://doi.org/10.1038/35017581

Zou, L., & Stillman, B. (2000). Assembly of a Complex Containing Cdc45p, Replication Protein
A, and Mcm2p at Replication Origins Controlled by S-Phase Cyclin-Dependent Kinases
and Cdc7p-Dbf4p Kinase. Molecular and Cellular Biology, 20(9), 3086–3096.
https://doi.org/10.1128/MCB.20.9.3086-3096.2000

88
SUPPLEMENTAL FIGURES AND TABLES

Figure S1. 1. DNA replication analysis of WT, rpd3∆, sir2∆, and rpd3∆ sir2∆ strains.

Figure S1.1 DNA replication analysis of WT, rpd3∆, sir2∆, and rpd3∆ sir2∆ strains. (A) Strains CVy43 (WT),
CVy44 (rpd3∆), YHy3 (sir2Δ), and YHy6 (sir2Δ rpd3∆) were synchronized in G1 phase and released into S phase
without HU for DNA content analysis. (B) QBU values (1kbp bins) for individual experimental replicates described
in Figure 1 legend are plotted against each other. (C) Boxplot distributions of QBU counts across for 500bp regions
aligned on origins of the indicated TRep quartiles; the number of origins in each group is indicated within
parentheses; two-sided t-tests were performed on all pairs of strains and results are indicated as: *=p<0.05,
**=p<0.01. (D) Heatmaps of correlation coefficients based on QBU signals for 5kbp regions centered on 625
potential origins.

A B
90 (min)
75
60
45
30
15
0
WT rpd3Δ sir2Δ sir2Δ rpd3Δ
C
rpd3Δ-1
rpd3Δ-2
R = 0.9938
2
R = 0.9946
2
R = 0.9934
2
rpd3Δ-3
rpd3Δ-3
rpd3Δ-1 rpd3Δ-2
WT-1
WT-2
5 10 0 15
5
10
0
15
WT-1 WT-2
WT-3
WT-3
R = 0.9888
2
R = 0.9876
2
R = 0.9889
2
5 10 0 15 5 10 0 15
5
10
0
15
5
10
0
15
5
10
0
15
5
10
0
15
5
10
0
15
5 10 0 15 5 10 0 15 5 10 0 15
sir2Δ-2
sir2Δ rpd3Δ-1
sir2Δ rpd3Δ-2
R = 0.9870
2
R = 0.9896
2
R = 0.9868
2
R = 0.9917
2
R = 0.9885
2
R = 0.9900
2
sir2Δ-1
sir2Δ-3
sir2Δ-1
sir2Δ-3
sir2Δ-2
sir2Δ rpd3Δ-1
sir2Δ rpd3Δ-3
sir2Δ rpd3Δ-2
sir2Δ rpd3Δ-3
5 10 0 15 5 10 0 15 5 10 0 15
5
10
0
15
5
10
0
15
5
10
0
15
5 10 0 15
5
10
0
15
5
10
0
15
5 10 0 15 5 10 0 15
5
10
0
15
2
4
0
6
8
2 8 0 6 4
1
2
0
3
2 3 0 1
WT-rDNA20-2
WT-rDNA20-1
rpd3Δ-rDNA20-2
rpd3Δ-rDNA20-1
R = 0.9856
2
R = 0.9603
2
4
4
D
Quartile1 (132) Quartile2 (91)
**
**
**
**
**
**
**
**
**
WT rpd3Δ sir2Δ
sir2Δ
rpd3Δ
20
12
8
0
16
4
12
8
0
16
4
WT rpd3Δ sir2Δ
sir2Δ
rpd3Δ
Quartile3 (45) Quartile4 (43)
**
**
**
**
*
12
8
6
0
10
4
2
7
5
4
0
6
3
2
1
WT rpd3Δ sir2Δ
sir2Δ
rpd3Δ
WT rpd3Δ sir2Δ
sir2Δ
rpd3Δ
QBU signal QBU signal
1
0.9
0.8
WT
rpd3Δ
sir2Δ
sir2Δ
rpd3Δ
WT
rDNA20
rpd3Δ
rDNA20
WT
rpd3Δ
WT rpd3Δ sir2Δ sir2Δ
rpd3Δ
WT rpd3Δ WT rpd3Δ
rDNA20
Correlation coefficient
89

Figure S1. 2. Analysis of rDNA copy numbers and quantification of 2D gels.
Figure S1.2. Analysis of rDNA copy numbers and quantification of 2D gels. (A) Total genomic DNA isolated
from G1-arrested cells of strains described in Figure 1 were subjected to high-throughput sequencing, and read counts
for the rDNA region were divided by the total read counts mapping to the nuclear genome; results of ANOVA test
comparing the first four strains is given, and result of two-sided t-test for the rDNA20 strains are given. (B) Images
of 2D gels (one set of two replicates) showing the actual areas analyzed for quantification shown in Figure 2.

Figure S1. 3. RPD3 deletion de-represses similar origins with or without HU.

Figure S1.3. RPD3 deletion de-represses similar origins with or without HU. (A) Images of 2D gels (one set of
three replicates) showing the actual areas analyzed for quantification shown in Figure 3. (B) Chromosome plots
comparing QBU origin-firing profiles in HU versus the first interval of the time-course.

WT rpd3Δ sir2Δ sir2Δrpd3Δ
0.10
0.04
0.02
0.06
0.08
0
WT rpd3Δ
rDNA20
p = 0.3513
p = 0.3721
Proportion of rDNA
B A
WT rpd3Δ
sir2Δ sir2Δ rpd3Δ
WT rpd3Δ
sir2Δ sir2Δ rpd3Δ
Bubble arc 1N spot
Replication origins Rpd3-repressed
Normalized QBU Signal
WT
rpd3Δ
Chromosome XI Coordinate (bp/10 )
5
1 2 4 5 6 3
1
0
0.5
WT
rpd3Δ
HU
1
0
0.5
1
0
0.5
1
0
0.5
First interval
B A
WT rpd3Δ
sir2Δ sir2Δ rpd3Δ
Bubble arc and Y arc
90

Figure S1. 4. Fkh1 overexpression stimulates firing of Rpd3-repressed origins.

Figure S1.4. Fkh1 overexpression stimulates firing of Rpd3-repressed origins. (A) Venn diagram shows
intersection of Fkh1-overexpression-activated origins (Table S3) with Rpd3-repressed origins called in this study
(Table S1). (B) Immunoblot analysis of Fkh1-3xFLAG levels in independent transformants of WT and rpd3∆ strains
(see key); blank lanes are transformants lacking Fkh1-3xFLAG, and outside lanes contain molecular weight markers.
Quantification of the Fkh1-3xFLAG band was determined against the band labeled Internal control; result of t-test
comparing protein levels in WT versus rpd3∆ is shown. (C) QBU values (1kbp bins) for individual experimental
replicates described in Figure 4 legend are plotted against each other.

rpd3Δ-1
rpd3Δ-2
WT-1
WT-2
R = 0.9958
2
R = 0.9904
2
R = 0.9935
2
R = 0.9395
2
fkh1Δ fkh2-dsm-2
fkh1Δ fkh2-dsm-1
fkh1Δ fkh2-dsm
rpd3Δ-1
fkh1Δfkh2-dsm
rpd3Δ-2
2 0 4 6
4
6
0
2
4
6
0
2
2 0 4 6
4
6
0
2
4
6
0
2
2 0 4 6 2 0 4 6
180 61 74
Fkh1-OE-activated
Rpd3-repressed
Fkh1-3FLAG
Internal control
1 2 3 4 5 6
1: rpd3Δ - 1
2: WT -1
3: rpd3Δ - 2
4: WT -2
5: WT -3
6: rpd3Δ - 3
p = 0.5924
rpd3Δ WT
Fkh1 expression level
B A
C
91

Figure S1. 5. Mcm4 origin association unaffected by RPD3 deletion.

Figure S1.5. Mcm4 origin association unaffected by RPD3 deletion. (A, B) ChIP enrichment values (1kbp bins)
for individual experimental replicates described in Figure 5 legend and below are plotted against each other. (C-E)
Strains MPy102 (MCM4-3xHA) and YHy37 (rpd3∆, MCM4-3xHA) were synchronized in G1 phase and subjected to
ChIP-seq. (C) Heatmaps of ChIP-seq enrichment at selected origin loci; the number in each group is indicated within
parentheses. (D) Boxplots distributions of ChIP-seq enrichments for 500bp regions aligned on origins of the indicated
sub-groups; the number of origins in each group is indicated within parentheses. Two-sided t-tests were performed
on all pairs of strains and results are indicated as: *=p<0.05, **=p<0.01. (E) Two-dimensional scatter plots of ChIP-
seq signals for 500bp regions centered on 626 origins plus the indicated origin groups; origins and sub-groups are
color-coded as indicated. A linear regression deriving the best-fit line is shown for each origin group.

WT-1
WT-2
WT-1 WT-2
WT-3
WT-3
R = 0.9991
2
R = 0.9995
2
R = 0.9995
2
5 10 0 15
5
10
0
15
5
10
0
15
5 10 0 15 5 10 0 15
5
10
0
15
rpd3Δ-1
rpd3Δ-2
rpd3Δ-1 rpd3Δ-2
rpd3Δ-3
rpd3Δ-3
R = 0.9978
2
R = 0.9984
2
R = 0.9997
2
5 10 0 15
5
10
0
15
5
10
0
15
5 10 0 15 5 10 0 15
5
10
0
15
B A
WT-1
WT-2
WT-1 WT-2
WT-3
WT-3
R = 0.9961
2
R = 0.9979
2
R = 0.9960
2
1
0
2
3
4
5
1
0
2
3
4
5
1
0
2
3
4
5
1 3 0 5 2 4 1 3 0 5 2 4 1 3 0 5 2 4
R = 0.9987
2
R = 0.9987
2
R = 0.9983
2
1
0
2
3
4
5
1
0
2
3
4
5
1
0
2
3
4
5
1 3 0 5 2 4 1 3 0 5 2 4 1 3 0 5 2 4
rpd3Δ-1
rpd3Δ-2
rpd3Δ-1 rpd3Δ-2
rpd3Δ-3
rpd3Δ-3
C
E
D
rpd3Δ
WT
0 2 4 5
5
3
1
0
4
2
1 3
All origins
Fkh-Activated
Rpd3-Repressed
CEN-proximal
rDNA
Distance from the origin (kb)
Fkh-activated (94)
WT
rpd3Δ
0 2.5 -2.5
Rpd3-repressed (74)
0 2.5 -2.5
CEN-proximal (32)
0 2.5 -2.5
All origins (311)
0 2.5 -2.5
4.5
4.0
2.0
3.5
3.0
2.5
1.5
ChIP signal
All origins (311)
WT rpd3Δ WT rpd3Δ WT rpd3Δ WT rpd3Δ
Rpd3-repressed (74) Fkh-activated (94) CEN-proximal (32)
12
6
4
2
10
8
0
12
6
4
2
10
8
0
12
6
4
2
10
8
0
12
6
4
2
10
8
0
ChIP signal
92
Table 1.1 Sequences of DNA oligonucleotides used in this study.
Oligonucleotides Sequence
SIR2-deletion-F ACATCTAGCACTCCTTCCAAC
SIR2-deletion-R ACCTGCCCTTCTTACATTAAGC
RPD3-deletion-F TCGCGGGCTGAACTGAATC
RPD3-deletion-R GCTTTATCAACAGCGGTGGG
sir2-del-up TTACTTGTAGCCTGCAACTCC
rpd3-del-up TCAGCATAACGAATTGACGG
sir2-del-down ATTCGACTTCTTTCCTTCGTTGT
rpd3-del-down TGCAATTAGAAGAGAGTGAATC
int-sir2-down ATCACAGGGTTCAATGTCGG
int-rpd3-down TAGTGTTCAGTTGAATCACAC
SIR2-del-TRP-F ATTCAAACCATTTTTCCCTCATCGGCACATTAAAGCTGGGACAAGTAACTGCAGG
AATT
SIR2-del-TRP-R ATATTAATTTGGCACTTTTAAATTATTAAATTGCCTTCTACACTATAGGGCGAAT
TGGGT
HIS-F AGAATACCCTCCTTGACAGTC
HIS-R TAGTATCGAATCGACAGCAG
HML-alpha-del-F TATAGGGCAGTGTGTGACTTATGAATTGTTGTAGAAGGACgacatggaggcccagaatac
HML-alpha-del-R ATGGCACAAGGAACACGCATTTTCCCAAGGCTTAGTATACcggcgttagtatcgaatcga
extend-HML-F TTTTGGGACGATATTGTCATTATAGGGCAGTGTGTGACTT
extend-HML-R TTTTATGAAGTAGCTTTCGGATGGCACAAGGAACACGCAT
probe-F CTCACACTTGTACTCCATGA
probe-R TGATGTGGAGAATAAGGTGC
2xL-3xFLAG-FKH1-
5'
TTTCTACTACGACATCCATGGACGTAACAACAAACGCAAACGTGAACAATTCCT
CTCTGAGTAGGGAACAAAAGCTGGAG
2xL-3xFLAG-FKH1-
3'
TTCTTAACGGGTCTTTGTTCTTTATTGTTTAATAATACATATGGGTTCGACGACGC
TGAATTCTATAGGGCGAATTGGGT
pAG61-Fkh1∆5' AATAATAGTGTGTAAATTGTGCGTTCAATTAGCAAAGAAAgacatggaggcccagaatac
pAG61-Fkh1∆3' TATTGTTTAATAATACATATGGGTTCGACGACGCTGAATTcggcgttagtatcgaatcga
Fkh2-up400bp CATTACCGAAAATCTTCGATTTCGC
Fkh2-down375bp CCGAAGCGTTGAGAAACAGC
Dbf4∆C-F1 AGCACAGACAGCACAGCCGGTGAAGAAAGAAACGGTAtgacggatccccgggttaattaa
Dbf4∆C-R1 GATTTTATCACTAAAAGCTACTGCACTTTACGTCGTGTCCcggcgttagtatcgaatcga
3HA-MCM4-F CGAGGGTGTAAGGAGATCAGTTCGCCTGAATAACCGTGTCCGGATCCCCGGGttaa
ttaac
3HA-MCM4-R GATTTTATCACTAAAAGCTACTGCACTTTACGTCGTGTCCcggcgttagtatcgaatcga
Table 1. 1.Sequences of DNA oligonucleotides used in this study.

93
Table 1.2. Genotypes of S. cerevisiae strains used in this study.
Name Genotype Source
SSy161 MATa ade2-1 ura3-1 his3-11,15 trp1-1 leu2-3,112 can1-100
bar1∆::hisG
Viggiani et al. 2006
CVy43 ura3::BrdU-inc(URA3) "
CVy61 trp1::BrdU-inc(TRP1) "
CVy68 MATα leu2::BrdU-inc(LEU2) "
CVy44 rpd3Δ::KanMX ura3::BrdU-inc(URA3) Knott et al. 2009
JHy3 his3-1 leu2-0 ura3-0 met15-0 sir2Δ::KanMX Mark Rose
JHy4 his3-1 leu2-0 ura3-0 met15-0 rpd3Δ::KanMX "
PP1758 fob1∆::URA3 ura3::7xTK (URA3) rDNA20 (~20copies) Philippe Pasero
JPy115 rpd3Δ::KanMX fob1∆::URA3 ura3::7xTK (URA3) rDNA20
(~20copies)
This study
MPy55 FKH1-3xFLAG(TRP1) ADE2::FLOPv2x2 "
MPy102 MCM4-3HA(KanMX) TRP1::HOPv1 "
MPy188 fkh1Δ::URA3 "
OAy1096 fkh1Δ::KanMX fkh2∆::HIS3MX ars305Δ::BrdU-inc(URA3) "
OAy1100 FKH1-3xFLAG(TRP1) "
OAy1102 FKH1-3xFLAG(TRP1) ADE2::FLOPv1 "
OAy1107 MATα fkh2∆::URA3MX leu2::BrdU-inc(LEU2) "
OAy1109 MATα fkh2-dsm leu2::BrdU-inc(LEU2) "
OAy1114 fkh1Δ::KanMX fkh2-dsm ars305Δ::BrdU-inc(URA3) "
OAy1186 MATα rpd3∆KanMX leu2::BrdU-inc(LEU2) "
OAy1188 OAy1186 x YHy19 diploid "
OAy1189 dbf4∆C::HIS3MX leu2::BrdU-inc(LEU2) "
YHy3 sir2Δ::KanMX hmlɑΔ::URA3MX trp1::BrdU-inc(TRP1) "
YHy6 sir2Δ::KanMX rpd3Δ::HIS3MX hmlɑΔ::URA3MX trp1::BrdU-
inc(TRP1)
"
YHy13 fkh1Δ::KanMX fkh2-dsm rpd3Δ::HIS3MX ars305Δ::BrdU-inc(URA3) "
YHy17 rpd3Δ::HIS3MX FKH1-3xFLAG(TRP1) ADE2::FLOPv1 "
YHy19 dbf4∆C::HIS3MX ura3::BrdU-inc(URA3) "
YHy27 rpd3Δ::HIS3MX FKH1-3xFLAG(TRP1) ADE2::FLOPv2x2 "
YHy33 sir2Δ::KanMX trp1::BrdU-inc(TRP1) "
YHy36 his3-1 leu2-0 ura3-0 met15-0 rpd3Δ::HIS3MX "
YHy37 rpd3Δ::HIS3MX MCM4-3HA(KanMX) TRP1::HOPv1 “
YHy38 OAy1189 x MPy188 diploid "
Table 1. 2. Genotypes of S. cerevisiae strains used in this study.

Table S1.1. List of RPD3-repressed origins.
94
Chr Start End
1 124350 124599
2 773918 774348
2 389245 390368
2 801930 802617
2 378434 379194
4 46181 46237
4 316719 317111
4 1353494 1353667
4 232140 232618
4 1057828 1058076
4 1302579 1302819
4 123617 123902
5 212381 212630
5 301565 302061
5 429861 431061
6 68690 68869
6 216344 216692
7 352695 352917
7 653611 654091
7 607176 607619
7 318745 319455
7 64279 64528
7 659809 660054
7 163180 163447
8 392148 392391
8 245719 245968
8 168531 168773
8 380153 382157
8 45589 46076
8 501751 501992
8 56650 59172
9 136094 136335
9 310583 311070
9 80058 80557
10 654069 654309
10 161435 161860
10 353463 355722
10 67467 67949
11 152934 153173
11 416822 417055
11 213080 213385
11 98329 98568
11 611874 612107
11 257390 257839
11 196038 196284
12 622672 623123
95
12 1007180 1007470
12 76711 77163
12 947908 948358
12 794020 794269
13 554392 554750
13 836823 838167
13 94216 94463
13 420429 421629
14 279875 280108
14 449343 449588
14 61597 61894
14 498987 499232
14 89528 89802
15 514258 515458
15 766617 766862
15 759221 766221
15 72636 72872
15 85195 85444
15 489645 490129
15 854735 855228
15 154972 155462
16 584037 584486
16 427273 428757
16 456557 456805
16 89758 92975
16 384536 384784
16 687133 694133
16 684383 684632
Table S1. 1. List of RPD3-repressed origins.

96
Table S1.2. List of CEN-proximal origins.
Chr Start End
1 146703 147690
1 159906 160127
2 237644 237879
2 254890 255136
3 114314 114933
3 108775 109291
4 462430 462700
4 435056 435388
5 145539 145782
5 173636 173874
6 135979 136080
6 167606 168041
7 484932 485160
7 508729 508978
8 115683 117257
8 111293 111766
9 357156 357393
9 341853 342096
10 442248 442658
10 416888 417134
11 447657 447892
11 454453 459197
12 150914 151421
12 139293 140447
13 263062 263296
13 286782 287067
14 635660 635901
14 609458 609706
15 337279 337528
15 308969 309462
16 563822 564061
16 552403 554287
Table S1. 2. List of CEN-proximal origins.

97
Table S1.3. List of FKH1-OE-activated origins.
Chr Start End
1 42208 42209
1 70349 70350
1 124576 124577
1 137379 137380
2 28963 28964
2 93681 93682
2 177531 177532
2 198366 198367
2 209246 209247
2 407858 407859
2 486848 486849
2 802115 802116
2 379233 379234
2 170386 170387
2 631830 631831
3 166850 166851
4 123689 123690
4 253815 253816
4 316820 316821
4 476950 476951
4 505554 505555
4 567568 567569
4 629330 629331
4 703096 703097
4 878261 878262
4 1302607 1302608
4 1486966 1486967
4 639849 639850
4 86193 86194
4 753440 753441
4 1057719 1057720
4 46066 46067
4 1462344 1462345
4 407794 407795
4 1353196 1353197
4 213203 213204
4 1017480 1017481
4 228790 228791
5 301986 301987
5 316847 316848
5 353874 353875
5 287925 287926
5 438519 438520
5 407472 407473
98
5 213535 213536
5 164297 164298
6 216425 216426
6 118532 118533
6 68203 68204
7 204083 204084
7 977842 977843
7 574963 574964
7 64219 64220
7 163609 163610
7 319635 319636
7 568305 568306
7 660411 660412
7 999032 999033
7 765799 765800
8 133557 133558
8 297363 297364
8 380923 380924
8 502050 502051
8 168397 168398
8 245566 245567
8 359199 359200
8 391532 391533
8 78716 78717
9 73807 73808
9 80481 80482
9 105993 105994
9 136324 136325
9 247698 247699
9 310658 310659
9 196640 196641
10 67886 67887
10 161644 161645
10 228315 228316
10 337217 337218
10 354353 354354
10 374791 374792
10 299049 299050
10 455549 455550
10 654623 654624
11 153151 153152
11 98573 98574
11 612115 612116
11 642646 642647
11 195981 195982
11 302607 302608
11 388987 388988
99
11 517034 517035
11 213518 213519
11 417191 417192
11 581850 581851
11 16317 16318
11 329776 329777
11 258184 258185
11 530926 530927
12 76746 76747
12 450643 450644
12 730434 730435
12 928208 928209
12 947991 947992
12 1007319 1007320
12 623203 623204
12 659711 659712
12 745314 745315
12 889006 889007
12 412469 412470
12 794525 794526
12 688195 688196
12 290047 290048
12 139853 139854
12 198446 198447
13 40138 40139
13 94377 94378
13 158926 158927
13 421301 421302
13 433282 433283
13 554466 554467
13 689069 689070
13 837710 837711
13 878637 878638
13 865281 865282
13 898047 898048
13 31678 31679
13 611499 611500
13 535533 535534
13 370902 370903
13 804926 804927
13 226807 226808
13 504481 504482
14 126850 126851
14 169687 169688
14 343561 343562
14 412385 412386
14 449430 449431
100
14 546038 546039
14 691523 691524
14 322229 322230
14 61919 61920
14 89827 89828
14 498922 498923
14 279620 279621
15 72714 72715
15 85214 85215
15 155155 155156
15 227853 227854
15 348277 348278
15 354524 354525
15 371940 371941
15 436914 436915
15 464305 464306
15 490091 490092
15 854790 854791
15 908318 908319
15 874443 874444
15 308932 308933
15 766504 766505
15 981900 981901
15 566938 566939
15 783043 783044
16 90544 90545
16 179437 179438
16 210534 210535
16 241741 241742
16 261434 261435
16 684383 684384
16 881098 881099
16 289715 289716
16 842632 842633
16 819129 819130
16 116468 116469
16 418083 418084
16 584547 584548
16 456925 456926
16 42825 42826
16 384955 384956
16 72806 72807

Table S1. 3. List of FKH1-OE-activated origins.

101

Figure S2. 1. Supplement to Figure 1.

Figure S2.1. Supplement to Figure 1. Western blot analysis of epitope-tagged strains used throughout this study:
“Untag” is SSy161 throughout, A) OAy1146 (FKH1-3xFLAG) and MPy55 (FKH1-3xFLAG FLOP). B) MPy166
(FKH1-9xMYC) and MPy108 (FKH1-9xMYC MOP). C) OAy503 (ORC1-3x-HA), MPy39 (HOP), MPy199
(ORC1-3xHA HOP), MPy184 (MCM4-3xHA HOP). D) YHy29 (SIR2-3xFLAG FLOP) MPy55 (FKH1-3xFLAG
FLOP).

102

Figure S2. 2. Supplement to Figure 2.
Figure S2.2. Supplement to Figure 2. A) Two-dimensional scatter plots analyzing correlation of sample
replicates. B) Two-dimensional scatter plots of unsynchronized vs G1-synchronized MOP/FLOP/HOP data. C)
Boxplots of ChIP signals at replication origins and CENs.

103

Figure S2. 3. Supplement to Figure 3.

104
Figure S2.3. Supplement to Figure 3. A) Two-dimensional scatter plots analyzing correlation of replicates. B)
Plots of ChIP signal across chromosomes III-L and IX with potential Fkh1 binding sites indicated as colored circles
on one track and called peaks indicated on the lower track. C) Heatmaps of averaged ChIP signal at snoRNAs.

Figure S2. 4. Supplement to Figure 4.

Figure S2.4. Supplement to Figure 4. Two-dimensional scatter plots analyzing the correlation of replicates.

105

Figure S2. 5. Supplement to Figure 5.

Figure S2.5. Supplement to Figure 5. A) Two-dimensional scatter plots analyzing correlation of replicates. B)
Venn diagrams showing origin overlap of peaks called using RN and uncontrolled datasets against confirmed and
likely origins. C) Two-dimensional scatter plots comparing Orc1 and Mcm4 ChIP signals against origin
efficiencies.

106

Figure S2. 6. Supplement to Figure 6.

Figure S2.6. Supplement to Figure 6. Two-dimensional scatter plots analyzing correlation of replicates.

107

Figure S2. 7. Supplement to Figure 7.

Figure S2.7. Supplement to Figure 7. A) EMSA analysis of Fkh1-dbm versus Fkh1 binding to TEM1 DNA
probe. B) Two-dimensional scatter plots analyzing correlation of sample replicates. C) Plots of ChIP signal across
chromosomes III-L and IX with potential Fkh1 binding sites indicated as colored circles on one track and called
peaks indicated on the lower track.

108

Table S2.1. Results of statistical analysis of data presented in Figure 2.
Results of Mann-Whitney tests are shown; ns=not significant (p>0.05). Element loci were tested against all bins and
all bins were tested across the same IP.

Table S1 *=negative z-
value

Mann Whitney Hyper-
ChIPable
tRNA origins CEN All bins
Asynchronous Untag 2.34x10
-20
* 2.05x10
-33
* – – 1.86x10-245
MOP 1.08x10
-21
8.24x10
-27
5.33x10
-7
* 5.12x10
-7
*
Untag ns 1.72x10
-30
* – – 5.56x10-37
FLOP 6.71x10
-10
9.05x10
-8
2.73x10
-
30
*
4.49x10
-11
*
HA 8.75x10
-16
* 1.07x10
-36
* – – 1.37x10-21
HOP ns 3.98x10
-2
2.92x10
-
58
*
2.51x10
-2

G1 MOP – – 1.67x10
-
12
*
2.39x10
-11
* –
FLOP – – 6.10x10
-
49
*
4.16x10
-11
* –
HOP – – 1.22x10-
65
*
5.36x10
-12
–
Table S2. 1. Results of statistical analysis of data presented in Figure 2.

Table S2.2. Results of statistical analysis of data presented in Figure 3.
Results of Mann-Whitney tests are shown; ns=not significant (p>0.05). Element loci were tested against all bins.

Table S2
Mann Whitney Hyper-ChIPable tRNA CLB2 Fkh-Activated Origins
Fkh1-MYC 5.51x10
-14
9.44x10
-8
3.26x10
-5
2.04x10
-14

Fkh1-MYC(MOP) 3.77x10
-16
3.59x10
-8
2.18x10
-5
2.37x10
-6

Fkh1-MYC-UntagRN 2.61x10
-31
6.66x10
-30
ns 7.29x10
-32

Fkh1-MYC MOP-RN ns ns ns 4.38x10
-16

Fkh1-FLAG 3.67x10
-6
ns 4.08x10
-7
2.04x10
-14

Fkh1-FLAG(FLOP) 8.78x10
-8
3.31x10
-5
7.68x10
-4
2.37x10
-6

Fkh1-FLAG -UntagRN ns 1.47x10
-29
ns 7.29x10
-32

Fkh1-FLAG FLOP-RN ns ns ns 4.38x10
-16

Table S2. 2. Results of statistical analysis of data presented in Figure 3.

109
Table S2.3. Sequences of DNA oligonucleotides used in this study.

Internal
ID
Name Sequence
1306 Fkh1-up CAGAAACGGTATAGAGAGAACAGG
1307 Fkh1-down CACAGAGGGTACAGAAGTCATAAAG
1793 ADE2-int-R aatacgactcactatagggcgaattggagctcCTTTACAACGAAGTTACCTCTTCCATC
G
1794 ADE2-far-up-F ccctcactaaagggaacaaaagctgggtaccCCTTTTGATGCGGAATTGACTTTTTCTT
G
1795 ADE2-up-F-
overlap
AACTTTTAATTAAGATACATTTCTTACGTCATGATTGATTATTACAG
CTATGC
1796 ADE2-up-R-
overlap
GACGTAAGAAATGTATCTTAATTAAAAGTTGTTAAAGATGTCAGTG
TTATGTTGGTG
1797 TRP1-far-up-F atacgactcactatagggcgaattggagctcGAATGAACGTATACGCGTATATTTCTA
CC
1798 TRP1-int-R ccctcactaaagggaacaaaagctgggtaccTGTCCCACCTGCTTCTGAATC
1799 TRP1-up-R TGCAAAGTACACtTaATTAaCGATGCTGTTCTATTAAATGCTTCC
1800 TRP1-up-F AGAACAGCATCGtTAATtAaGTGTACTTTGCAGTTATGACGC
1801 ADH1-T CCGTTTCTGACAGAGTAAAATTCTTGAGcggccggtagaggtgtggtcaataagag
1802 ADH1-P-
ADE2targ
CATAACACTGACATCTTTAACAACTTTTAATTAAggcggcccaacttcttttctt
1803 lexA-Op-up-
ADE2targ
caatcatgacgtaagaaatgtatcttaattaaCATCAACACATAAATCCTTGCTTAGCTC
1804 lexA-Op-down ctcttattgaccacacctctaccggccgCTCAAGAATTTTACTCTGTCAGAAACGG
1805 ADH1-P-
TRP1targ
TCTGGCGTCATAACTGCAAAGTACACTTAATTAAggcggcccaacttcttttctt
1806 lexA-Op-up-
TRP1targ
GAAGCATTTAATAGAACAGCATCGTTAATTAACATCAACACATAAA
TCCTTGCTTAGCTC
1807 lexA-5'FLAG GGCGGTTGGGGTTATTCGCAACGGCGACTGGCTGcccgggCTCGGTGG
ATCTGGCGGTAG
1808 lexA-3'FLAG cataagaaattcgcccggaattagcttggCTGCAGcctgcaGTTACTTGTCATCGTCATC
1809 lexA-5'HA TTGGGGTTATTCGCAACGGCGACTGGCTGcccggggtgagctcgttttacccatacga
tg
1810 lexA-3'HA cataagaaattcgcccggaattagcttggCTGCAGcagggagctcactgagcagc
1811 lexA-HA1 GGGTTATTCGCAACGGCGACTGGCTGcccggggtAagctcgttttacccatacgatgttc
1812 lexA-HA2 ggctatccctatgacgtcccggactatgcaggatcctatccatatgacgttccagattac
1813 lexA-HA3 agtccgggacgtcatagggatagcccgcatagtcaggaacatcgtatgggtaaaacgagc
1814 lexA-HA4 cgcccggaattagcttggCTGCAGtcactgagcagcgtaatctggaacgtcatatggata
1815 lexA-seq1 CATTATGGATGGTGACTTGCTG
1816 lexA-seq2 GTATCATAGACATAGGTGGTGTG
1817 lexA-OP-seq AGCTGGACCACCTACTG
1822 lexA-OP-seq2 TGACTCTTAGGTTTTAAAACG
1823 ADE2-int-R2 CAAAGCTTCCGGAATCATTTCC
1824 ADE2-far-up-F2 GCCTAGTTTCATGAAATTTTAAAGCA
1825 ADH1-prom-intR AAGCTTGGAGTTGATTGTATGC
110
1826 TRP1-far-up-F2 GAAGAGGAGTAGGGAATATTACTGG
1827 TRP1-int-R2 CAGTCAGAAATCGAGTTCCAATC
1828 lexA-OP-int-R GTTATATGTACAGTACGTCGACTG
1846 lexA-5'MYC#2 GGTTGGGGTTATTCGCAACGGCGACTGGCTGCCCGGGaaCGGTGAAC
AAAAGCTAATCTC
1847 3'MYC-ADH-
TER#2
AAGAAATTCGCCCGGAATTAGCTTGGCTGCAGctaGTGATTGATTAAT
TTTTGTTCACCG
1848 lexA-5'MYC#3 GGTTGGGGTTATTCGCAACGGCGACTGGCTGCCCGGGttaattaaCGGTG
1849 3'MYC-ADH-
TER#3
AAGAAATTCGCCCGGAATTAGCTTGGCTGCAGAAGTggcgcgaattcACT
1864 3'MYC13tag-seq AGAAGTggcgcgaattcACT
1865 TRP1-up-EagI AAGGCCGTTTCTGACAGAGTAAAATTCTTGAGCGGCCGgtgtactttgcagtt
atgacgc
1868 ADH1-ter-Eag-
EcoRI
GATTATAAAGATGACGATGACAAGTAACTGCAgaattcCGGGCGAATT
TCTTATG
1874 TRP1-LOP-Seq1 GCACGGCAGAGACCAATCAG
1875 ADH1-TER-R-
Seq
GAGTCACTTTAAAATTTGTATACAC
1889 ADE2-up-EagI #2 TTTCTGACAGAGTAAAATTCTTGAGCGGCCGTTGTTAAAGATGTCAG
TGTTATGTTGGTG
2244 5'TEFpro-PacI gagcatttaattaaggcgcgccagatc
2245 3'TEFpro-HindIII gagcataagcttggttgtttatgttcggatgtgatgtg
2249 3'TEFpro-Gibson ACGCTTTCATTCCGCCCGGAATTAATTCaagcttggttgtttatgttcggatgtgatgtg
2250 5'TEFpro-Gibson CCAACATAACACTGACATCTTTAACAACTTttaattaaggcgcgccagatc
2351 lexOP-PacI-up CAATCATGACGTAAGAAATGTATC
2352 lexOP-PacI-int TACTTAATTAACGTCGACTGCTGTATATAAAAC
2353 lexOP-int-Seq-R CCAGTGGTTATATGTACAGTACTGCTG
2354 lexOp2x-5'ADE2-
Gib
ccaacataacactgacatctttaacaacttttaatCATCAACACATAAATCCTTGCTTAG
2355 lexOp2x-ADH1p-
Gib
aaaaaaaagaaaagaagttgggccgccttaattaaGTACGTCGACTGCTGTATATAAAAC
2356 lexOp2x-TEFp-
Gib
aggcaagctaaacagatctggcgcgccttaattaaGTACGTCGACTGCTGTATATAAAAC
2386 F MCM4 Tagging CGAGGGTGTAAGGAGATCAGTTCGCCTGAATAACCGTGTCCGGATC
CCCGGGttaattaaC
2387 R MCM4 Tagging TTATTAATTGTTACGCAGGGAATGATTGTAGTAGACAGCATCGATG
AATTCGAGCTCGTT
2454 FKH1-dbm GGCGTGGCAGgcCTCCGTGgcGgcTAATCTGgcCgcAAATAAAGCCT
2513 Fwd_dbm_tail GCAGgcCTCCGTGgcGgcTAATCTGgcCgcAAATAAAGCCTTCGAGAAG
GTGC
2514 Rev_dbm_tail ATTTgcGgcCAGATTAgcCgcCACGGAGgcCTGCCACGCCATTTGAGAA
AAC
2526 Sir2-FLAG-F CGTGTATGTCGTTACATCAGATGAACATCCCAAAACCCTCAGGGAA
CAAAAGCTGGAG
2527 Sir2-FLAG-R GATATTAATTTGGCACTTTTAAATTATTAAATTGCCTTCTACCTATAG
GGCGAATTGGGT
2528 pro-TEM1-for AGCAAAAAAGGTAAACAAGAAAAGGAA
2529 pro-TEM1-rev TTCCTTTTCTTGTTTACCTTTTTTGCT
111

Table S2. 3. Sequences of DNA oligonucleotides used in this study.
Table S2.4 Genotypes of S. cerevisiae strains used in this study.

Name Genotype (only differences
from SSy161 are indicated)
Source Purpose/Construction
SSy161 MATa ade2-1 ura3-1 his3-
11,15 trp1-1 leu2-3,112 can1-
100 bar1∆::hisG
Viggiani
et al.
2006
W303 background, parent strain
SSy162 MAT⍺ " W303 background, parent strain
CVy68 MAT⍺ leu2::BrdU-Inc(LEU2) "
OAy503 ORC1-3xHA(URA) Aparicio
et al.
1997

OAy1100 FKH1-3xFLAG(TRP1) He et al.
2022

ZOy3 FKH1-9xMYC (TRP1) Ostrow
et al.
2014

BY4741 his3∆1 leu2∆0 met15∆0
ura3∆0
Euroscarf Not W303 background; source of WT ADE2 and TRP1 DNA
for plasmid constructions
MPy35 ADE2::FLOPv2x2 This
Study
pADE2-FLOPv2x2 digested and transformed into SSy161
MPy39 TRP1::HOPv1 " pTRP1-HOPv1 digested and transformed into SSy161
MPy41 MAT⍺ TRP::HOPv1 " pTRP1-HOPv1 digested and transformed into SSy162
MPy55 FKH1-3xFLAG(TRP1)
ADE2::FLOPv2x2
" FKH1-3xFLAG TRP1 amplicon (from OAy1146)
transformed into MPy35
MPy57 MAT⍺ MCM4-
3xHA(KanMX6)
TRP1::HOPv1
" PCR amplicon from pFA6a-3HA-KanMX6 transformed into
MPy41
MPy100 MAT⍺ ADE2::MOPv1 " pADE2-MOPv1 digested and transformed into SSy162
MPy102 MCM4-3xHA(KanMX6)
TRP1::HOPv1
" haploid segregant from mating of MPy39 and MPy57
MPy105 ADE2::MOPv1 " Haploid segregant from mating of MPy100 and ZOy3
MPy108 FKH1-9xMYC(TRP1)
ADE2::MOPv1
" Haploid segregant from mating of MPy100 and ZOy3
MPy166 FKH1-9xMYC(TRP1) " FKH1-9xMYC-TRP1 amplicon (from MPy108) transformed
into SSy161
MPy169 fkh1-dbm-9xMYC(TRP1) " FKH1-9xMYC-TRP1 amplicon (from MPy108) with
overlapping mutagenic primers to incorporate dbm mutations
transformed into SSy161
MPy172 fkh1-dbm-3xFLAG(TRP1) " FKH1-3xFLAG-TRP1 amplicon (from MPy54) with
overlapping mutagenic primers to incorporate dbm mutations
transformed into SSy161
MPy184 MCM4-3xHA(KanMX6)
TRP1::HOPv1
" Haploid segregant from mating of MPy102 and OAy1106
MPy187 MAT⍺ fkh1∆::URA3(Ca)
TRP1::HOPv1
" Haploid segregant from mating of MPy102 and OAy1106
MPy199 ORC1-3xHA(URA)
TRP1::HOPv1
" Haploid segregant from mating of MPy187 and OAy503
OAy1106 MAT⍺ leu2::BrdU-Inc(LEU2)
fkh1∆::URA3(Ca)
" FKH1 deletion with URA3 (Candida albicans) marker from
pAG61 in CVy68
112
OAy1146 MAT⍺ leu2::BrdU-Inc(LEU2)
FKH1-3xFLAG(TRP1)
" PCR amplicon from p2L-3FLAG-TRP1 transformed into
CVy68
YHy29 SIR2-3xFLAG(TRP1)
ADE2::FLOPv2x2
" Sir2-3xFLAG TRP1 amplicon (from MPy55) transformed
into MPy35
Table S2. 4. Genotypes of S. cerevisiae strains used in this study.

Abstract (if available)

Abstract Eukaryotic chromosomes are organized into structural and functional domains with characteristic replication timings, which are thought to contribute to epigenetic programming and genome stability. Differential replication timing results from epigenetic mechanisms that positively and negatively regulate the competition for limiting replication initiation factors. In the budding yeast Saccharomyces cerevisiae, histone deacetylase Sir2 negatively regulates initiation of the multi-copy (~150) rDNA origins while Rpd3 histone deacetylase negatively regulates the firing of single-copy origins. However, Rpd3’s effect on single-copy origins might derive indirectly from a positive function for Rpd3 in rDNA origin firing shifting the competitive balance. Our quantitative experiments support the idea that origins compete for limiting factors; however, our results show that Rpd3’s effect on single-copy origin is independent of rDNA number and of Sir2’s effects on rDNA origin firing. Whereas RPD3 deletion and SIR2 deletion alter the early S phase dynamics of single-copy and rDNA origin firings in an opposite fashion, unexpectedly only RPD3 deletion suppresses overall rDNA origin efficiency across the S phase. Increased origin activation in rpd3Δ requires Fkh1/2 suggesting that Rpd3 opposes Fkh1/2-origin stimulation, which involves the recruitment of Dbf4-dependent kinase (DDK). Indeed, Fkh1 binding increases at Rpd3-regulated origins in rpd3Δ cells in G1, supporting a mechanism whereby Rpd3 influences the initiation timing of single-copy origins directly through modulation of Fkh1-origin binding. Genetic suppression of a DBF4 hypomorphic mutation by RPD3 deletion further supports the conclusion that Rpd3 impedes DDK recruitment by Fkh1, revealing a mechanism of Rpd3 in origin regulation. In the study mentioned above, we extensively utilized the Chromatin Immunoprecipitation followed by sequencing (ChIP-Seq) technique, particularly when investigating whether Rpd3 affects Fkh1-origin binding. In fact, ChIP-Seq is a widely used technique for the analysis of protein-DNA interactions in vivo. However, ChIP has pitfalls, particularly false-positive signal enrichment that permeates the data. We have developed a new approach to control for non-specific enrichment in ChIP that involves the expression of a nongenome-binding protein targeted in the IP alongside the experimental target protein due to the sharing of epitope tags. ChIP of the protein provides a “sensor” for non-specific enrichment that can be used for the normalization of the experimental data, thereby correcting for non-specific signals and improving data quality as validated against known binding sites for several proteins that we tested, including Fkh1, Orc1, Mcm4, and Sir2. We also tested a DNA binding mutant approach and showed that, when feasible, ChIP of a site-specific DNA binding mutant of the target protein is likely an ideal control. These methods vastly improve our ChIP-seq results in S.cerevisiae and should be applicable in other systems.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

The role of Rpd3 and Sir2 in regulation of replication initiation in budding yeast: Rpd3 acts directly on single-copy origins while Sir2 works through ribosomal DNA origins

PDF

Forkhead transcription factors regulate replication origin firing through dimerization and cell cycle-dependent chromatin binding in S. cerevisiae

PDF

Distinct mechanisms of DDK recruitment to Fkh-activated and CEN-proximal origins control replication timing program in S. cerevisiae

PDF

Computational analysis of DNA replication timing and fork dynamics in S. cerevisiae

PDF

Forkhead transcription factors control genome wide dynamics of the S. cerevisiae replication timing program

PDF

Maintenance of genome stability at fragile sites in Schizosaccharomyces pombe

PDF

New tools for whole-genome analysis of DNA replication timing and fork elongation in saccharomyces cerevisiae

PDF

Quantitative modeling of in vivo transcription factor–DNA binding and beyond

PDF

The kinetic study of engineered MBD domain interactions with methylated DNA: insight into binding of methylated DNA by MBD2b

PDF

Dps contributes to typical growth, survival, and genome organization in E. coli

PDF

Genome-wide studies reveal the function and evolution of DNA shape

PDF

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

PDF

The role of Hic-5 in glucocorticoid receptor binding to chromatin

PDF

Machine learning of DNA shape and spatial geometry

PDF

The role of Cdc7 in replication fork progression in response to DNA damage

PDF

Using novel small molecule modulators as a tool to elucidate the role of the Myocyte Enhancer Factor 2 (MEF2) family of transcription factors in leukemia

PDF

Genome-wide studies of protein–DNA binding: beyond sequence towards biophysical and physicochemical models

PDF

Biochemical mechanism of TopBP1 recruitment to sites of DNA damage

PDF

Genome-scale insights into the underlying genetics of background effects

PDF

Simulating the helicase motor of SV40 large tumor antigen

Asset Metadata

Creator He, Yiwei (author)

Core Title The function of Rpd3 in balancing the replicaton initiation of different genomic regions

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Molecular Biology

Degree Conferral Date 2023-12

Publication Date 09/06/2023

Defense Date 08/22/2023

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

Tag chromatin domains,chromatin immunoprecipitation,controls,DNA-binding proteins,histone deacetylase,OAI-PMH Harvest,repetitive DNA,replication origin

Format theses (aat)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Aparicio, Oscar (committee chair), Chen, Lin (committee member), Chiolo, Irene (committee member), Ehrenreich, Ian (committee member), Feng, Pinghui (committee member)

Creator Email heyiwei0727@gmail.com,yiweihe@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-oUC113303122

Unique identifier UC113303122

Identifier etd-HeYiwei-12320.pdf (filename)

Legacy Identifier etd-HeYiwei-12320

Document Type Dissertation

Format theses (aat)

Rights He, Yiwei

Internet Media Type application/pdf

Type texts

Source 20230906-usctheses-batch-1091 (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.

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