1

Title: Loss of histone H3.3 results in DNA replication defects and 1

altered origin dynamics in C. elegans. 2

3

Authors: Maude Strobino, Joanna M. Wenda and Florian A. Steiner 4

5

6

7

Affiliations: 8

Department of Molecular Biology and Institute for Genetics and Genomics in Geneva, 9

Section of Biology, Faculty of Sciences, University of Geneva, 1211 Geneva, 10

Switzerland 11

12

Correspondence to: florian.steiner@unige.ch 13

14

2

Abstract 15

Histone H3.3 is a replication-independent variant of histone H3 with important roles in 16

development, differentiation and fertility. Here we show that loss of H3.3 results in 17

replication defects in Caenorhabditis elegans embryos. To characterize these defects, 18

we adapt methods to determine replication timing, map replication origins, and examine 19

replication fork progression. Our analysis of the spatiotemporal regulation of DNA 20

replication shows that despite the very rapid embryonic cell cycle, the genome is 21

replicated from early and late firing origins and is partitioned into domains of early and 22

late replication. We find that under temperature stress conditions, additional replication 23

origins become activated. Moreover, loss of H3.3 results in impaired replication fork 24

progression around origins, which is particularly evident at stress-activated origins. 25

These replication defects are accompanied by replication checkpoint activation, a 26

prolonged cell cycle, and increased lethality in checkpoint-compromised embryos. Our 27

comprehensive analysis of DNA replication in C. elegans reveals the genomic location 28

of replication origins and the dynamics of their firing, and uncovers a role of H3.3 in the 29

regulation of replication origins under stress conditions. 30

31

Keywords: Histone H3.3, replication, origins, Repli-seq, EdU-seq, ChEC-seq, DNA 32

combing, C. elegans 33

34

35

3

Introduction 36

At every cell cycle, the entire genome is replicated exactly once for accurate transmission of 37

genetic information to the daughter cells. To ensure rapid and uniform duplication of the 38

genome, DNA replication is initiated in a bi-directional way at many different sites in the 39

genome, called replication origins. To prevent re-replication and to control the activity of origins 40

under replication stress, replication initiation is composed of two non-overlapping steps called 41

‘origin licensing’ and ‘origin firing’. Origin licensing occurs during G1 through the formation of the 42

pre-replicative complexes. During this process, the origin recognition complex (ORC) binds to 43

the origins and recruits the replication licensing factors Cdt1 and Cdc6, which together load the 44

complex of the minichromosome maintenance proteins 2-7 (Mcm2-7), a helicase that unwinds 45

DNA ahead of the replication fork (Sonneville et al., 2012). To ensure that each origin fires only 46

once per cell cycle, Cdt1 and Cdc6 are subsequently removed from the nucleus or degraded to 47

prevent the Mcm2-7 complex from rebinding origins that have already fired (Blow and Dutta, 48

2005; Feng and Kipreos, 2003; Kim and Kipreos, 2007). Origin firing occurs upon entry into S 49

phase, through binding of different initiation factors like Sld2 (RecQL4/RecQ4 in humans), Sld3 50

(Treslin/TICRR in humans) and Dpb11 (TopBP1 in humans), and phosphorylation of the Mcm2-51

7 complex (Bruck and Kaplan, 2015; Kanemaki and Labib, 2006; Labib and Diffley, 2001). 52

These firing factors are present in limiting amounts (Mantiero et al., 2011; Tanaka et al., 2011), 53

and all origins therefore do not fire synchronously at the start of S phase. Instead, some origins 54

fire early and some late, resulting in large domains of early and late replication (Desprat et al., 55

2009; Dileep and Gilbert, 2018; Farkash-Amar et al., 2008; MacAlpine et al., 2004; Takahashi et 56

al., 2019). These domains may change during development, and it has been shown in multiple 57

organisms that early replicating domains are associated with transcriptionally active regions 58

(Hiratani et al., 2010, 2008; Schübeler et al., 2002; Therizols et al., 2014; White et al., 2004; 59

Woodfine et al., 2004). 60

61

4

Upon replication origin firing, the replication forks move bi-directionally until replication bubbles 62

of opposing directions meet and the replication of the corresponding stretch of chromosome is 63

complete. Under specific conditions, e.g. conflicts between transcription and replication forks, 64

depletion of nucleotides, or depletion of histones, replication forks can stall or collapse (Bester 65

et al., 2011; Groth et al., 2007; Macheret and Halazonetis, 2018). This replication stress can 66

lead to DNA damage (Macheret and Halazonetis, 2015; Técher et al., 2017), and triggers the 67

activation of the ATR checkpoint, which phosphorylates the Chk1 kinase, leading to cell cycle 68

arrest (Hyun et al., 2016; Liu et al., 2000; Moiseeva et al., 2019). As a response to moderate 69

replication stress, cells can activate additional origins (Técher et al., 2017). These origins, called 70

latent or dormant origins, are inactive and do not fire during the undisturbed S phase, but act as 71

backup origins in case of replication stress (Boos and Ferreira, 2019; Woodward et al., 2006). 72

How firing origins are selected over dormant origins, and under what conditions dormant origins 73

are activated is unclear (Fragkos et al., 2015). 74

75

It is widely accepted that replication origins are not defined by DNA sequences in higher 76

eukaryotes and therefore cannot be predicted from genome sequences, making their 77

identification challenging. Numerous assays have been developed to map replication origins in 78

different organisms (Prioleau and MacAlpine, 2016), mostly by mapping different replication 79

elements such as nascent DNA strands (Besnard et al., 2012), replication bubbles (Mesner et 80

al., 2013), Okazaki fragments (Petryk et al., 2016), replication initiation sites (Langley et al., 81

2016), or initiation factors (Dellino et al., 2013; Miotto et al., 2016). Despite these efforts, the 82

number, location and dynamic regulation of replication origins remains controversial. For 83

example, two studies in C. elegans recently provided genome-wide maps of replication origins 84

using two different methods, and arrived at largely different conclusions (Pourkarimi et al., 2016; 85

Rodríguez-Martínez et al., 2017). The first study used Okazaki fragment sequencing to identify 86

~2’000 origins and concluded that replication origins are defined prior to the broad onset of 87

5

zygotic transcription and maintained during development (Pourkarimi et al., 2016). The second 88

study applied mapping of small RNA-primed nascent DNA strands to identify ~16’000 origins 89

and described a major re-organization of replication origins after gastrulation (Rodríguez-90

Martínez et al., 2017). Additional approaches will therefore be required to understand how 91

origins are defined and dynamically regulated. 92

93

The variant histone H3.3 has been linked to different steps of DNA replication. It is enriched at 94

replication origins in Drosophila melanogaster and Arabidopsis thaliana (Deal et al., 2010; Eaton 95

et al., 2011; MacAlpine et al., 2010; Paranjape and Calvi, 2016; Stroud et al., 2012), and loss of 96

H3.3 results in a slower replication fork in chicken cells (Frey et al., 2014). The association of 97

H3.3 with early replication domains, and its recruitment to sites of DNA repair have also been 98

described in human cells (Adam et al., 2013; Clément et al., 2018). H3.3 is incorporated into 99

transcriptionally active regions as well as in centromeric and telomeric repeats (Buschbeck and 100

Hake, 2017; Henikoff and Smith, 2015), but the link between its genomic location and its 101

function is not always clear. Loss of H3.3 results in lethality or sterility in most organisms (Hödl 102

and Basler, 2009; Jang et al., 2015; Sakai et al., 2009; Tang et al., 2015; Wollmann et al., 103

2017). The C. elegans genome contains five genes encoding H3.3 homologues, and we 104

recently showed that mutation of all five genes resulted in viable worms with a reduced number 105

of viable offspring at higher temperatures (Delaney et al., 2018). 106

107

Given the role of H3.3 in DNA replication in other organisms, we speculated that the reduced 108

brood size observed in C. elegans H3.3 null mutants is caused by defects in DNA replication. 109

Here we show that at higher temperatures, loss of H3.3 causes mild DNA replication defects 110

and activation of the embryonic DNA replication checkpoint. To investigate these defects, we 111

adapt ChEC-seq, Repli-seq, EdU-seq and DNA combing for use in C. elegans, which we apply 112

to define and characterize replication timing and replication origins. We find that during S phase, 113

6

replication and the firing of replication origins are temporally regulated. H3.3 is enriched at 114

replication origins, but is not required for origin firing. Instead, loss of H3.3 results in delayed 115

fork progression around replication origins, in particular around origins that are only active under 116

temperature stress conditions. Together, our results uncover a role of H3.3 in DNA replication in 117

C. elegans and show that replication origin use can be adapted in response to stress. 118

119

120

Results 121

Genome-wide replication timing in C. elegans embryos 122

Histone H3.3 has been linked to the regulation of DNA replication in a variety of organisms 123

(Adam et al., 2013; Clément et al., 2018; Deal et al., 2010; Eaton et al., 2011; Frey et al., 2014; 124

MacAlpine et al., 2010; Paranjape and Calvi, 2016; Stroud et al., 2012). We therefore 125

hypothesized that the reduced brood sizes observed in C. elegans H3.3 null mutants under mild 126

temperature stress are caused by defects in DNA replication (Delaney et al., 2018). Our first aim 127

was to determine whether loss of H3.3 causes global changes in the temporal program of 128

genome replication. To characterize replication timing during the rapid cell cycle in early C. 129

elegans embryogenesis, we adapted Repli-seq for use in this organism (Hansen et al., 2010). 130

This method relies on the incorporation of the thymidine analogue 5-ethynyl-2'-deoxyuridine 131

(EdU) into newly replicated regions of the genome (Figure 1A). To separate early and late 132

replicated genomic regions, we pulse-labelled asynchronous embryonic cell populations with 133

EdU, then sorted the cells according to their DNA content and sequenced EdU-containing 134

fragments. We obtained cells that just entered S phase (early replication), showed ongoing DNA 135

synthesis (mid replication), or were at or near completion of replicating the genome (late 136

replication) (Figure S1A). These Repli-seq experiments revealed that the genome is partitioned 137

into mutually exclusive domains of early and late replication (Figure 1B, C). Surprisingly, 138

different chromosomes display distinct overall replication timing patterns. Domains of early 139

7

replication are enriched on chromosomes I, II and III, while domains of late replication are more 140

prevalent on chromosomes IV, V and X (Figure 1C and Figure S1B). We found that domains of 141

early replication correlate well with the presence of HIS-72, the most highly expressed C. 142

elegans H3.3 homologue (Figure 1B and Figure S1C). However, H3.3 null mutant worms 143

showed largely the same early and late replication domains as wild-type worms, indicating that 144

H3.3 is not a major driver of global DNA replication timing (Figure 1B, C and Figure S1B). 145

Genes in domains of early replication show higher levels of expression in embryo compared to 146

genes in domains of later replication, consistent with the observation from other organisms that 147

regions of open chromatin replicate early (Ekundayo and Bleichert, 2019; Prioleau and 148

MacAlpine, 2016) (Figure 1D). However, there is no enrichment of large genes in late replicating 149

domains that would suggest potential interference between transcription and replication under 150

stress conditions, as observed in human cells (Figure 1E) (Macheret and Halazonetis, 2018). 151

152

Mapping replication origins by ChEC-seq and EdU-seq 153

Given the presence of domains of early and late replication, we reasoned that there should be 154

distinct populations of early and late firing replication origins. With the enrichment of H3.3 at 155

replication origins in other organisms, we speculated that loss of H3.3 may cause subtle defects 156

in the position or timing of origin firing in H3.3 null mutants. We therefore next aimed to define 157

replication origins and their timing during S phase. Previous origin mapping approaches in C. 158

elegans do not allow for the analysis of origin dynamics during the cell cycle (Pourkarimi et al., 159

2016; Rodríguez-Martínez et al., 2017). We therefore took an alternative approach that 160

combines the mapping of replication origin factors and the observation of nucleotide 161

incorporation at origins. We first profiled the genomic binding sites of proteins involved in 162

replication origin licensing and firing. We selected two conserved proteins, CDT-1 and TRES-1, 163

that are involved in different steps of origin activation (Figure 2A). CDT-1 is the homologue for 164

human Cdt1 and is implicated in the licensing of all origins during G1 phase (Zhong et al., 2003) 165

8

(Figure 2A). TRES-1 is the homologue of human Treslin and is subsequently recruited 166

exclusively to the origins that are going to fire (Guo et al., 2015; Kumagai et al., 2010). 167

Immunofluorescence microscopy confirmed that CDT-1 is present on chromatin during 168

anaphase and telophase, but levels become undetectable during S-phase. TRES-1 only 169

appears on chromatin during telophase, but remains present during S phase (Figure 2B). These 170

observations are consistent with their roles in origin licensing and origin firing, respectively. 171

Typically, genomic protein binding sites are profiled by chromatin immunoprecipitation. 172

However, this approach depends on antibody availability and solubility of the chromatin 173

component of interest, all of which can be limiting factors. To circumvent these problems, we 174

adapted Chromatin Endogenous Cleavage (ChEC) to C. elegans. This method relies on the 175

fusion of the protein of interest with micrococcal nuclease (MNase), which cuts DNA around the 176

binding sites of the fusion protein and thereby releases DNA fragments that can be sequenced 177

and mapped (Schmid et al., 2004; Zentner et al., 2015) (Figure 2C). We generated MNase 178

fusion constructs at the endogenous cdt-1 and tres-1 loci using CRISPR/Cas9. MNase 179

insertions resulted in worms that are phenotypically wild-type. To generate genome-wide 180

binding maps of CDT-1 and TRES-1, we performed ChEC-seq experiment on purified 181

embryonic nuclei and identified peaks of CDT-1 and TRES-1 signal. We observed CDT-1 and 182

TRES-1 peaks that are distributed across the genome in both datasets. Moreover, as expected, 183

the CDT-1 peaks mostly overlap with TRES-1 peaks (licensed and firing), but some CDT-1 184

peaks show no corresponding TRES-1 signal (licensed only) (Figure 2D). 185

The CDT-1 and TRES-1 ChEC-seq data allowed us to discriminate between licensed and firing 186

origins, but we could not establish the timing of origin firing within S phase. To identify 187

replication origins that fire immediately after entry into S phase, we modified the Repli-seq 188

protocol. Instead of analyzing asynchronous cell populations, we used hydroxyurea (HU) to 189

synchronize cells at the entry of the S phase. Upon HU treatment, virtually all cells become 190

arrested at the entry of the S phase, and a large part of the cell population proceeds with DNA 191

9

replication upon removal of HU (Figure S2A). We combined the release into S phase with a 192

second HU block in the presence of EdU, allowing the cells to replicate short stretches of DNA 193

around early replication origins before arresting. The EdU-containing DNA fragments were then 194

isolated and sequenced. These EdU-seq experiments revealed discrete peaks that are located 195

in close proximity to with a subset of the CDT-1 and TRES-1 peaks (licensed and firing with 196

EdU signal) (Figure 2D). The EdU-seq signal sometimes appear as double peaks flanking the 197

CDT-1 and TRES-1 peaks, indicating that we detect short fork movements away from the 198

replication origins. We conclude that the sites enriched for EdU-seq, CDT-1 and TRES-1 signal 199

correspond to replication origins that fire upon entry into S phase. 200

201

Classifications of origins genome wide 202

To obtain a map of embryonic replication origins genome-wide, we combined the peak calls 203

from the CDT-1 ChEC-seq, TRES-1 ChEC-seq and EdU-seq experiments. In total, we obtained 204

a map of 1110 origins. Based on their signal in the three datasets, unsupervised clustering 205

distinguished three types of origins, which we classified as “early firing”, “late firing” and 206

“dormant” (Figure 3A). Early firing origins show enrichment in all three datasets, as they are 207

licensed (CDT-1) and fire at the onset of S phase (TRES-1 and EdU-seq). Late firing origins 208

show enrichment in CDT-1 and TRES-1 ChEC-seq datasets, but not EdU-seq, indicating that 209

they fire after the onset of S phase. Dormant origins show enrichment only in the CDT-1 ChEC-210

seq dataset, indicating that they are licensed, but do not fire under standard conditions (Figure 211

3A). While the enrichment of CDT-1 and TRES-1 overlaps within a narrow window at the 212

replication origins, the distribution of EdU-seq signal is more broad, which likely reflects short 213

stretches of fork progression away from the replication origins after the release from the initial 214

HU block. As expected, firing origins are found mostly outside of gene-coding regions, while 215

dormant origins are enriched within genes (Figure S2B). The genomic localization of the 216

replication origins identified here largely overlap with those previously identified (Figure S2C), 217

10

but our mapping approach offers the advantage of capturing origin firing dynamics during the 218

cell cycle, and also includes dormant origins that may act as backup origins for use under 219

specific conditions. Similar to D. melanogaster and A. thaliana (Deal et al., 2010; Eaton et al., 220

2011; MacAlpine et al., 2010; Paranjape and Calvi, 2016; Stroud et al., 2012), H3.3 is enriched 221

around the origins of all three categories (Figure 3B), suggesting that it may play a role in origin 222

identity or activation. 223

224

Altered origin dynamics in H3.3 null mutants 225

We next compared the distribution of firing origins in wild type and H3.3 null mutant worms. For 226

this, we repeated the TRES-1 ChEC-seq and the EdU-seq experiments in wild type and H3.3 227

null mutant strains at 25°C, where a difference in brood-size between the two strains is 228

observed. The TRES-1 ChEC-seq experiments showed that most origins licensed in wild type 229

worms were also utilized in absence of H3.3 (Figure 4A). In both strains, we also observed 230

weak TRES-1 signal at origins that are dormant at 20°C. These origins were initially identified 231

solely based on the presence of CDT-1 under normal conditions, and the presence of TRES-1 232

at 25°C shows that these are functional origins that become activated at higher temperatures. 233

This result confirms the presence of “backup” origins that are inactive during undisturbed S 234

phase, but are activated under replication stress (Boos and Ferreira, 2019; Woodward et al., 235

2006). We did not detect any difference between the firing of early origins in wild type and H3.3 236

null mutant based on EdU-seq signal (Figure 4C, 0min). These results indicate that the origin 237

distribution and the firing of early origins remains unchanged upon loss of H3.3. 238

Although the firing of early origins is not affected by the loss of H3.3, we were wondering if fork 239

progression following origin firing could be altered. The EdU-seq method not only allows one to 240

identify early firing origins, but can also be used to visualize fork progression around these 241

origins, by adding an EdU pulse at 0, 30 or 60 minutes after the release from the HU block 242

(Figure 4B-C). As expected, the EdU signal enrichment at the early firing origins is no longer 243

11

present after 30 minutes, and regions occupied by early firing origins incorporate very little EdU 244

after 60 minutes, which reflects the bi-directional movement of the replication forks away from 245

the origins (Figure 4B-C, WT). This bi-directional fork movement is difficult to observe for late 246

origins, as we have no means to synchronize their firing. Nevertheless, the EdU-seq signal at 247

late origins appears depleted at 30 minutes, suggesting that they fire within the first 30 minutes 248

after S-phase onset. In H3.3 null mutants, as mentioned above, early origins appear to fire 249

normally, and we observe bi-directional fork movement. However, we found that the EdU signal 250

remains higher around early origins after 30 and 60 minutes, indicating that fork progression is 251

delayed (Figure 4B-C, ∆ H3.3). This delay in fork progression appears at origins of all classes 252

and results in an enrichment of EdU incorporation at late origins. The difference between wild 253

type and H3.3 null mutant cells is even more pronounced at origins that are dormant under 254

standard conditions. These origins fire in both strains at 25°C, as evidenced by the increased 255

presence of TRES-1 compared to 20°C (Figure 3A and Figure 4A), but fork progression from 256

the origins is impaired in H3.3 null mutants, resulting in an accumulation of EdU-seq signal at 257

later time points. The enrichment at late and dormant origins is observed within a very narrow 258

window, suggesting that the forks stall or collapse immediately after origins fire. The 259

coincidence of this EdU-seq signal enrichment with the center of late and dormant origins is 260

particularly remarkable, as these origins were identified solely based on a completely 261

independent method, namely the mapping of the CDT-1 and TRES-1 binding sites. The 262

aberrant fork progression in H3.3 null mutant worms is only seen under mild temperature stress 263

conditions, and is not observed at 20°C (Figure S3), indicating that H3.3 is required for 264

regulating origin dynamics under stress. 265

266

Fork speed is unaffected by loss of H3.3 267

12

To establish if loss of H3.3 resulted in a general slowing of the replication fork speed along the 268

chromatin fiber, we adapted DNA combing to C. elegans (Michalet et al., 1997). This method 269

relies on the pulse labeling of DNA by the subsequent incorporation of two thymidine analogues, 270

IdU and CldU. By visualizing the length of IdU and CldU incorporation on stretched individual 271

chromatin fibers, replication fork speed can be estimated (Figure 5A). We found that fork speed 272

is on average 1.4 kb/min (Figure 5B), but did not detect any significant difference between wild 273

type and H3.3 null mutant embryos, indicating that H3.3 is not required for normal progression 274

of the moving replication fork. 275

276

Increased DNA damage in H3.3 null mutant embryos 277

DNA combing assesses the speed of the moving replication fork, but it does not measure the 278

average fork speed genome-wide, which can be affected by fork pausing, stalling or collapse. 279

Our EdU-seq time course analysis suggested that replication origins fire normally upon loss of 280

H3.3, but that the progression of forks is altered around origins. We therefore probed the 281

embryos for the presence of replication stress that might be indicative of fork stalling and 282

collapse. In C. elegans embryos, DNA replication defects are characterized by an increased 283

length of the first embryonic cell cycle (Brauchle et al., 2003; Holway et al., 2006; Stevens et al., 284

2016). We measured cell cycle timing from the appearance of a cleavage furrow in the P0 cell to 285

nuclear envelope breakdown of the AB cell (Figure 6A). We found that the cell cycle time was 286

significantly increased in H3.3 null mutant embryos compared to wild type (Figure 6B). 287

Removing the replication checkpoint through targeting the checkpoint kinase 1 (CHK-1) by RNAi 288

restored normal cell cycle timing, confirming that the increased cell cycle length is linked to a 289

replication defect (Figure 6B). Moreover, removal of CHK-1 resulted in an increase of the 290

embryonic lethality in H3.3 null mutants (Figure 6C). Together, our results indicate that loss of 291

H3.3 results in defects in DNA replication or DNA damage repair, and that these defects 292

become detrimental upon removal of CHK-1. 293

13

294

Taken together, we find that replication of the C. elegans genome is partitioned into early and 295

late domains, and that the global timing of replication and replication origin firing is unchanged 296

in H3.3 null mutant worms. However, loss of H3.3 leads to defective fork progression around 297

origins and DNA replication defects at 25°C. Our results identify the replication dynamics during 298

the embryonic cell cycle and uncover a role of H3.3 in DNA replication in C. elegans embryos. 299

300

301

Discussion 302

The variant histone H3.3 has previously been linked to DNA replication, but its mechanistic role 303

in this process, and how its function relates to its genomic localization are not well understood. 304

In most animal models, loss of H3.3 results in lethality or sterility, making the developmental 305

analysis of H3.3 mutant challenging (Couldrey et al., 1999; Hödl and Basler, 2009; Sakai et al., 306

2009; Tang et al., 2015). Our recent discovery that C. elegans H3.3 null mutants are viable 307

allowed us to analyze potential developmental defects in more detail (Delaney et al., 2018). 308

Here we describe that under stress conditions, H3.3 is required for DNA replication dynamics 309

around replication origins and protects the genome from replication stress. 310

311

For understanding the replication defects upon loss of H3.3 in detail, we first had to develop 312

tools to investigate DNA replication dynamics in C. elegans embryos. DNA replication timing in 313

C. elegans had not been analyzed in detail, and replication origins have only recently been 314

mapped by different techniques, with conflicting outcomes (Pourkarimi et al., 2016; Rodríguez-315

Martínez et al., 2017). We therefore adapted techniques that are based on the incorporation of 316

EdU to determine the genome-wide replication timing (Repli-seq), and to follow the fork 317

progression around replication origins (EdU-seq). These experiments revealed the surprising 318

finding that despite the very short cell cycle during embryogenesis, genomic regions of early 319

14

and late replication are mutually exclusive, and distributed in a non-random fashion. Partitioning 320

of the genome into early and late replicating domains is very well described for mammalian 321

genomes, both in cell population averages (Desprat et al., 2009; Farkash-Amar et al., 2008; 322

Hiratani et al., 2008; MacAlpine et al., 2004) and single cells (Dileep and Gilbert, 2018; 323

Takahashi et al., 2019) but had not yet been observed in C. elegans. Replication timing has 324

been attributed to the 3D organization of chromatin into discrete domains (Marchal et al., 2019). 325

We also observed this robust spatial organization of replication timing in C. elegans, despite 326

absence of strong 3D compartmentalization (Crane et al., 2015). Chromosomes I-III are 327

enriched for early replication domains, while most of chromosomes IV, V and X, and the 328

telomeric regions of all chromosomes replicate late. We note that genes located in 329

chromosomes I-III are on average more highly expressed than genes on chromosomes IV-X, 330

and may therefore constitute an environment of open chromatin that is permissive for the 331

binding of replication initiation factors. The domains of early replication are also enriched for 332

H3.3. A similar correlation between early replication and H3.3 incorporation has been observed 333

in human cells (Clément et al., 2018). Given this conserved correlation, it is maybe surprising 334

that loss of H3.3 does not globally alter the domains of early and late replication. However, 335

replication timing has been found to be remarkably robust, and remains unaltered in a wide 336

range of mutants that affect chromatin organization (Marchal et al., 2019). Our results support 337

the notion that the concentrations of some factors implicated in DNA replication are limiting, and 338

that they will access the regions of open chromatin more readily (Mantiero et al., 2011). The 339

concentration of replication factors is tightly controlled during S phase, and failure to export or 340

degrade factors involved in licensing can lead to over-replication of the genome (Sonneville et 341

al., 2012). 342

343

To comprehensively identify replication origins throughout S phase, we mapped the genomic 344

binding sites of the origin licensing factor CDT-1 and the origin firing factor TRES-1. For 345

15

determining their binding sites genome wide, we adapted the ChEC-seq method for use in C. 346

elegans (Schmid et al., 2004; Zentner et al., 2015). We then combined the results with the early 347

origins identified by EdU-seq. Our origin identification thus combines methods that detect 348

proteins involved in origin identity with methods that rely on the identification of sites of 349

nucleotide incorporation during origin firing. This approach allows the identification of origin 350

firing dynamics during the cell cycle, and also includes dormant origins. C. elegans origins have 351

been mapped previously using Okazaki fragment sequencing or nascent strand sequencing, 352

both of which should identify firing origins (Pourkarimi et al., 2016; Rodríguez-Martínez et al., 353

2017). Our approach identified significantly fewer origins (~1’000 vs ~2’000 or ~16’000, 354

respectively), but the genomic location of the origins identified in this study largely overlap with 355

those previously identified (Figure S2C). Previous approaches utilized the depletion of DNase 356

lig-1 by RNAi (Pourkarimi et al., 2016) or liquid culture (Rodríguez-Martínez et al., 2017) for the 357

cultivation of worms for origin analysis. Both treatments result in worms that are superficially 358

wild type, as is the case for H3.3 null mutant worms. Given our observation that even mild 359

temperature stress can affect origin firing, and that liquid culture can affect gene expression 360

(Çelen et al., 2018), these treatments could alter origin use and contribute to the observed 361

differences in origin numbers. To independently estimate the number of origins required to 362

replicate the entire genome in our system, we measured the speed of the replication fork on 363

single molecules by DNA combing, and determined the time required to reach maximum EdU 364

incorporation within one cell cycle. We found that a plateau of EdU incorporation is reached 365

about 40 minutes after the first observed EdU incorporation, suggesting that cells require about 366

40 minutes to replicate the entire genome under these conditions (Figure S4). This is 367

substantially longer than the S-phase duration observed in developing embryos (Sonneville et 368

al., 2012), and we attribute the delay to the dissociation and the treatment with HU and EdU. 369

Based on this replication time and the fork speed measured in the DNA combing experiment, 370

we estimate the number of firing origins required for replicating the entire genome to be about 371

16

924. This number is remarkably close to the number of firing origins identified by ChEC-seq and 372

EdU-seq (798), and we therefore think that our mapping strategy accurately reflects active 373

replication origins. While our methods allow for the detection of origins that only fire under 374

specific conditions, they still rely on averaging signal from very large populations of cells, and 375

will therefore be less sensitive to detect origins that are only licensed at specific developmental 376

time points or in specific cell types. 377

378

H3.3 is enriched at replication origins in D. melanogaster and A. thaliana, but the functional 379

significance of this enrichment remains unclear, and no defects in origin dynamics have been 380

detected upon loss of H3.3 (Deal et al., 2010; Eaton et al., 2011; MacAlpine et al., 2010; 381

Paranjape and Calvi, 2016; Stroud et al., 2012). We find that H3.3 is enriched at replication 382

origins, but consistent with the findings from D. melanogaster, we found no evidence that loss of 383

H3.3 inhibits replication origin firing (Paranjape and Calvi, 2016). The EdU-seq enrichments at 384

early origins remains unchanged, and H3.3 null mutant worms are able to activate dormant 385

origins similar to wild type under stress conditions, as evidenced by the TRES-1 signal observed 386

at 25°C. However, we found clear differences in EdU incorporation around origins as S phase 387

progresses. Fork progression from origins is impaired, and the EdU-seq signal remains enriched 388

at replication origins, in particular at those that only fire under stress conditions. The speed of 389

the replication fork along the chromatin fiber is not affected by the loss of H3.3, as determined 390

by DNA combing experiments. It is therefore likely that instead the frequency of fork stalling or 391

fork collapse is increased as the fork is progressing away from the replication origins, resulting 392

in the increased EdU-seq signal around replication origins. Our observation of the replication 393

checkpoint activation through CHK-1 is also consistent with increased replication stress rather 394

than a general slowing of fork progression. The defects are only observed at 25°C, and origin 395

dynamics in wild type and H3.3 null mutant worms are indistinguishable at 20°C, consistent with 396

the fertility defects of H3.3 null mutant worms that are only observed at higher temperatures. 397

17

The reasons behind the temperature sensitivity of the replication fork upon loss of H3.3 are 398

currently unclear. 399

400

The increased replication stress in the H3.3 null mutant could be caused directly by altering the 401

chromatin organization, or indirectly by affecting gene expression. Decreased nucleosome 402

density may affect replication fork progression. In D. melanogaster embryos, a decrease in 403

histone concentration results in an increased cell cycle length and CHK-1 activation, similar to 404

our observations in C. elegans embryos (Chari et al., 2019). However, only three genes 405

encoding for H3.3 are expressed during embryogenesis, compared to 16 genes encoding 406

canonical H3 (Boeck et al., 2016). It is therefore unlikely that loss of H3.3 results in a dramatic 407

drop in nucleosome density genome wide. However, replacing H3.3 nucleosomes with H3 408

nucleosomes may alter the chromatin dynamics and also contribute to the observed replication 409

stress. Absence of H3.3 may not cause replication stress directly, but may prevent fork restart 410

after fork pausing or collapse, similar to what has been observed after DNA damage in chicken 411

cells (Frey et al., 2014). Alternatively, loss of H3.3 could alter the transcriptional landscape 412

genome-wide, thus resulting in increased conflicts between transcription and DNA replication. 413

Indeed, the majority of the dormant origins, for which we find the strongest effects, are located 414

within genes (Figure S2B). However, we previously reported that loss of H3.3 only results in 415

minor changes in the transcriptome (Delaney et al., 2018). Loss of H3.3 could also alter the 416

histone marks around origins. C. elegans replication origins were found to be enriched for H3 417

acetylation, H3K4 methylation and H3K27 acetylation (Pourkarimi et al., 2016), and it is possible 418

that the presence of H3.3 aids the deposition of some of these marks. However, the causal 419

relationship between histone modifications and the identity of replication origins is not clear. 420

Many studies have accumulated evidence that chromatin context is important for the activity of 421

replication origins, and we now show that H3.3 plays an important role in the fork progression 422

following origin firing under stress condition. More research will be needed to establish a causal 423

18

relationship between the chromatin context and replication origin formation, to understand how 424

the timing of origin firing is regulated, and to determine how the replication fork interacts with 425

nucleosomes containing different histone modifications or variants. 426

427

428

Materials and Methods 429

Worm culture and strain generation 430

Worms were grown on NGM plates seeded with OP50 for maintenance, and on peptone-rich 431

NGM plates seeded with NA22 for large-scale experiments. The sequence coding for MNase 432

was inserted at the cdt-1 and tres-1 loci using CRISPR/Cas9 as described in (Arribere et al., 433

2014), using plasmids with 1 kb homology arms as repair templates. A list of the strains used in 434

this study is given in Table S1. RNAi experiments were carried out by feeding, using clones 435

from the Ahringer library (Kamath et al., 2003). For experiments measuring effects at 25°C, 436

worms were grown at this temperature for at least 5 generations. 437

438

Phenotype analysis 439

For determining embryonic lethality at 25°C, L4 worms were placed on plates containing RNAi 440

or control food overnight. Single adults were transferred to new plates for 3h and then removed. 441

The number of eggs, hatched larvae and adults was counted on subsequent days to determine 442

the penetrance of embryonic lethality. Plates with 0 or 1 embryos were excluded from the final 443

count. Data from three independent experiments with 10 plates per experiment and condition 444

were used. Total number of embryos: WT control = 445, ∆ H3.3 control = 240, WT chk-1 RNAi = 445

270, ∆ H3.3 chk-1 RNAi = 251. 446

19

For determining the cell cycle length during early embryonic cell divisions at 25°C, L4 worms 447

were placed on plates containing RNAi or control food overnight. Adults were cut in egg buffer 448

(118mM NaCl, 48 mM KCL, 2mM CaCl2, 2mM MgCl2 and 25mM Hepes pH7,5) to release 449

young embryos. These embryos were mounted on 2% agarose pads and recorded every 10 s 450

with DIC settings on a Leica DMI8 microscope, with the temperature maintained at 25°C. Cell 451

cycle length was measured from the appearance of the cleavage furrow of the P0 cell to nuclear 452

envelope breakdown of the AB cell. WT control = 3, ∆ H3.3 control = 5, WT RNAi = 4, ∆ H3.3 453

RNAi = 4. 454

For determining the subcellular localization of CDT-1 and TRES-1, gravid adult worms were 455

washed in PBS containing 0.1% Triton X-100, anesthetized in DB buffer (50 mM sucrose, 75 456

mM HEPES pH 6.5, 60 mM NaCl, 5 mM KCl, 2 mM MgCl2, 10 mM EGTA pH 7.5, 0.1% NaN3) 457

and cut in half to release embryos. The embryos were transferred to poly-L-lysine-coated slides, 458

freeze-cracked and fixed in cold methanol for 5 min. Slides were washed two times in PBS and 459

incubated with anti-HA antibody (mAb 42F13, 1:60) for CDT-1 staining or FLAG antibody 460

(Sigma-Aldrich, F7425-.2MG, 1:200) for TRES-1 staining, overnight at 4°C. After two washes in 461

PBS, slides were incubated with secondary antibodies (Alexa Fluor 488 goat anti-mouse or 462

Alexa Fluor 488 goat anti-rabbit; 1:700) for 1 h at room temperature, washed once in PBS and 463

counterstained with DAPI (2 µg/ml) for 15 minutes. After a final PBS wash, the slides were 464

mounted in VECTASHIELD Antifade Mounting Medium. Images were acquired with Leica SP8 465

confocal microscope. For all the images Z-sections with 0.3 µm step were acquired, and 466

processed with Fiji software (maximum Z-projection, contrast adjustment, Gaussian blur filter 467

r=0.5) (Schindelin et al., 2012). 468

469

Chromatin Endogenous Cleavage (ChEC) 470

20

Worm cultures were synchronized by embryo isolation with sodium hypochlorite for 4 471

generations. Embryos were treated with chitinase (1h, RT , 2 U/ml), washed twice in buffer A 472

(15 mM Tris pH 7.5, 0.1 mM EGTA, 0.34 M sucrose, 0.2 mM spermine, 0.5 mM spermidine, 0.5 473

mM PMSF) and then resuspend in buffer A supplemented with detergents (0.25% NP-40 and 474

0.1% Tx-100). Nuclei were released with a glass dounce homogenizer and 15 stokes using the 475

loose inserting pestle and 15 stokes using the tight inserting pestle. Debris were spun down at 476

100 g for 2 minutes and the supernatant, containing nuclei, was saved. The debris were 477

resuspended in buffer A supplemented with detergents and the nuclei isolation step was 478

repeated. The supernatants containing the nuclei were pooled, and the nuclei were pelleted at 479

1000 g for 10 min. Nuclei were washed once in buffer A, once in buffer B (15 mM Tris pH 7.5, 480

80 mM KCl, 0.1 mM EGTA, 0.2 mM spermine, 0.5 mM spermidine, 1 mM PMSF), and 481

resuspended in buffer B. MNase was activated by the addition of CaCl2 to a final concentration 482

of 2 mM at 30°C. Samples were incubated for 5 to 10 minutes (for CDT-1 ChEC), 5 to 7 minutes 483

(for TRES-1 ChEC) or for 1 h (free MNase control). For the control sample, 0.3 U/ml of MNase 484

(Biolabs, Cat. No. M0247S) was added at the same time as the CaCl2. MNase was inactivated 485

by adding an equal volume of 2x stop buffer (400 mM NaCl, 20 mM EGTA). Samples were 486

treated overnight with Proteinase K and SDS at 55°C. Phenol-chloroform-isoamyl treatment was 487

used to remove Proteinase K prior the RNase treatment (Roche, Cat. No. 11119915001) for 488

1h30 minutes at 37°C. After chloroform extraction, DNA was precipitated overnight with ethanol 489

and 0.25 M NaCl. DNA was resuspended overnight in TE prior to constructing the libraries using 490

NEBNext® Ultra™ II DNA Library Prep. Fifty base pair paired-end read sequencing reactions 491

were then performed on an Illumina Hi-Seq 2500 sequencer at the Genomics Platform of the 492

University of Geneva, except for TRES-1 ChEC-seq samples from 25°C, which were sequenced 493

with one hundred base pair single-end read sequencing reactions. 494

495

Embryonic cell extraction for EdU incorporation or DNA combing 496

21

Cells were extracted from embryos as described in (Bianchi and Driscoll, 2006). Briefly, 497

embryos were isolated using sodium hypochlorite, treated with chitinase (1h, RT, 2U/ml), 498

washed once in M9 and resuspend in 10ml of L-15 medium (supplemented with 10% FBS (heat 499

inactivated, Gibco), 41 mM sucrose and 1% pen/strep (1:100)). Cells were dissociated with 500

three passes through a syringe with a 22 gauge needle and two passes through a syringe with a 501

26 gauge needle. Debris were pelleted by spinning at 80 g for 1 min, and the supernatant, 502

containing cells, was saved. Debris were resuspended in supplemented L-15 medium and cell 503

extraction was repeated twice more. Supernatants were pooled together and filtered using 5 504

micron filters. Cells were pelleted at 500 g for 15 min and resuspended in supplemented L-15 505

medium. 506

507

EdU-seq 508

EdU-seq was carried out as described (Macheret and Halazonetis, 2019) with minor 509

modifications. Embryonic cells were synchronized by addition of 20 mM of HU (Sigma, Cat. No. 510

H8627) during 1h at 20°C or 25°C. Cells were washed twice with PBS to remove HU and 511

resuspended in L-15 medium. At desired time points, 25 µM of EdU and 20 mM of HU were 512

added for 10 minutes. Cells were washed twice with PBS, fixed with 90% methanol and stored 513

at -20°C. Cells were washed with ice-cold PBS and permeabilized with PBS containing 0.2% 514

Triton X-100 for 30 minutes at room temperature in the dark. EdU was coupled to a cleavable 515

biotin-azide linker (Azide-PEG(3+3)-S-S-biotin) (Jena Biosciences, Cat. No. CLK-A2112-10) 516

using the reagents of the Click-it Kit (Invitrogen, Cat. No. C-10424). Cells were washed with 517

PBS and treated overnight at 50°C with proteinase K in the lysis buffer. DNA was extracted by 518

phenol/chloroform/isoamyl extraction followed by chloroform extraction. DNA was precipitated 519

overnight at -20°C with presence of ethanol and 0.25 M NaCl and resuspended overnight in TE. 520

For an estimation of the time required to replicate the entire genome, cells were isolated, 521

synchronized using HU, and released in the presence of EdU as described above. Cell aliquots 522

22

were harvested between 0 and 130 min. Cell permeabilization and click-it reaction were done as 523

described above to couple Alexa Fluor™ 647 Azide (Thermofisher Cat. No. A10277) to the 524

EdU-labeled DNA. Cells were washed twice with PBS and treated with RNase (Roche, Cat. No. 525

11119915001) for 30 minutes at 37°C. PI (Sigma, Cat. No. 81845) was added, and samples 526

incubated overnight at room temperature in the dark. Levels of EdU incorporation were 527

assessed by flow cytometry (Gallios, Beckman Coulter). 528

529

Repli-seq 530

Asynchronous embryonic cells extracted from worms grown at 25°C were incubated for 5 531

minutes with 25 µM EdU (Invitrogen, Cat. No. A10044), washed twice with PBS and fixed with 532

90% methanol overnight. After fixation and click-it reaction as described in the EdU-seq section, 533

cells were washed twice with PBS and treated with RNase (Roche, Cat. No. 11119915001) for 534

30 minutes at 37°C. Propidium iodide (Sigma, Cat. No. 81845) was added, and samples were 535

incubated overnight at room temperature in the dark. Cells were sorted according to their DNA 536

content using a MoFlo Astrios flow sorter (Beckman Coulter) at the Flow Cytometry platform of 537

the Medical Faculty of the University of Geneva. DNA was extracted as described in the EdU-538

seq section. 539

540

Isolation and sequencing of EdU-labelled DNA from EdU-seq and Repli-seq 541

Five to eight µg of DNA were sonicated with a bioruptor sonicator (Diagenode) to obtain 542

fragments of 100-500 bp. EdU-labelled fragments were isolated using Dynabeads MyOne 543

Streptavidine C1 (Invitrogen, Cat. No. 65001) as previously described (Macheret and 544

Halazonetis, 2019). DNA was eluted by the addition of 2% of B-mercaptoethanol (Sigma, Cat. 545

No. M6250) and incubation for 1h at room temperature. The eluted DNA, as well as fragmented 546

total DNA, was used for library preparation using the TruSeq ChIP Sample Prep Kit (Illumina, 547

Cat. No. IP-202-1012). One hundred base pair single-end read sequencing reactions were 548

23

performed on an Illumina Hi-Seq 2500 sequencer at by the Genomics Platform of the University 549

of Geneva. 550

551

Data processing, domain annotation and peak calling 552

Sequencing reads were mapped to the C. elegans reference genome WBcel215 using 553

Novoalign software (default parameters, producing SAM format files). Reads were transformed 554

into 1’000 bins size (ChEC-seq and EdU-seq) or 10’000 bins size (Repli-seq) (JSON format) 555

with a custom script. Biological replicates that passed the quality control were merged (Repli-556

seq : WT early (2), WT mid (1), WT late (3), ∆ H3.3 early (2), ∆ H3.3 mid (2), ∆ H3.3 late (2), 557

ChEC-seq : CDT-1 20°C (5), TRES-1 20°C (4), TRES-1 25°C (3), TRES-1 25°C ∆ H3.3 (3), 558

EdU-seq : WT 0 min 20°C (4), all the other EdU-seq data (2)). A custom R script was used to 559

transform JSON files into bedgraphs. All files were normalized to the same number of reads. 560

Experimental samples were then normalized by subtracting the number of reads present in the 561

corresponding bin in the control file. The control samples were DNA from nuclei treated with 562

Free MNase (ChEC-seq), genomic DNA from cells treated with HU (EdU-seq), or genomic DNA 563

from cells without treatment (Repli-seq). Finally, the data were smoothed with a sliding window 564

averaging 5 consecutive bins using a custom R script. Genome Browser images were obtained 565

by using the Sushi Bioconductor package in R. 566

Domains of early and late replication were defined according to the signal enrichment present in 567

early and late Repli-seq in WT samples. Bins with positive signal in early and negative signal in 568

late datasets were defined as early domains, and vice versa for the ones with negative signal in 569

early and positive signal in late as late domain. The bins with either positive signal in early and 570

late repli-seq or negative signal in both datasets were not assigned to a domain. Domains 571

smaller than 5kb were merged with the previous domain. 572

24

For peak calling, normalized and smoothed data was converted to SGA files using the “ChIP-573

convert” tool in the Vital-it platform (https://ccg.epfl.ch//chipseq/). The output SGA files were 574

used as input for peak calling with the tool “ChIP-peak” of the Vital-it platform with these specific 575

parameters: Cut-off = 1000, Read counts = 10, Window = 3000, Vicinity = 5000, strand = any, 576

Repeat Masker = on, Refine peak position = off, Relative Enrichment 325 (CDT-1), 450 (TRES-577

1) or 350 (EdU). Peak lists from the ChEC-seq and EdU-seq experiments were merged, and 578

duplicate peaks removed. Peaks are considered duplicate if two datasets contains a peak within 579

the same 5kb window. To avoid sequencing artefacts, peaks that were also enriched in the 580

genomic DNA sample were removed. The final list of peaks was then clustered in 15 using 581

deepTools (Ramírez et al., 2016). Clusters with similar enrichment were merged together and 582

one cluster, containing signal only in the EdU-seq sample, was deleted, as the sites in this 583

cluster did not show fork movement after release from the HU block. Heatmaps were created 584

using deepTools (Ramírez et al., 2016). The genomic annotation of early, late and dormant 585

origins was determined using PAVIS website (https://manticore.niehs.nih.gov/pavis2/) (Huang et 586

al., 2013). Sequencing data has been deposited at the Gene Expression Omnibus (GEO) under 587

accession number GSE140804. 588

589

DNA combing and staining 590

DNA combing was performed as described in (Michalet et al., 1997). Embryonic cells were 591

incubated with 20 mM HU for 1h at 20°C or 25°C, washed twice with ice-cold PBS, and 592

incubated with 40 µM CldU for 20 minutes at the desired temperature. Cells were washed twice 593

with L-15 and then incubated in presence of 400 µM IdU for 20 minutes at the desired 594

temperature. Cells were washed with PBS and frozen at -20°C. Cells were embedded into 595

agarose plugs, and DNA was processed using the FiberPrep® DNA extraction kit (Genomic 596

Vision). The combing and antibody incubation was done as described in (Costantino et al., 597

2014). The DNA combing slides were imaged on a Leica DM5000B microscope, and the images 598

25

were analyzed with Fiji software (Schindelin et al., 2012). The analysis was carried out with data 599

from six independent experiments with a total of 49 measurements (WT, 20°C), 32 600

measurements (WT, 25°C), 28 measurements (∆ H3.3, 20°C) and 49 measurements (∆ H3.3, 601

25°C). 602

603

604

Acknowledgements 605

We thank Morgane Macheret and Thanos Halazonetis for sharing unpublished protocols, 606

Jonathan Mailler and Laura Padayachy for discussion and help with the DNA combing 607

experiments, Dario Menendez for help with strain generation, Caroline Gabus for reagent 608

preparation, Nicolas Roggli for help with data analysis and image processing, Florian Huber and 609

Ioannis Xenarios for help with the data analysis and Uli Laemmli and Gabriel Zentner for 610

reagents and advice when establishing the ChEC method. We thank all the Steiner lab 611

members for discussion and comments on the manuscript. We thank Thanos Halazonetis, 612

Damien Hermand, Jonathan Mailler, Laura Padayachy, Reinier Prosee, David Shore and 613

Maksym Shyian for comments on the manuscript. Some strains were provided by the CGC, 614

which is funded by NIH Office of Research Infrastructure Programs (P40 OD010440). We are 615

grateful to the Genomics Platform of the iGE3, the Bioimaging Center of the Faculty of Sciences 616

and the Flow Cytometry platform of the Medical Faculty at the University of Geneva. The work 617

was funded by the Swiss National Science Foundation (Grants 31003A_156774 and 618

31003A_175606) and the Republic and Canton of Geneva. 619

620

621

Competing interests 622

The authors declare no competing interests. 623

26

624

625

Author contributions 626

FAS and MS conceived the study, MS carried out genetics, genomics and DNA combing 627

experiments, JMW carried out staining experiments, MS analyzed the data, MS and FAS 628

prepared the manuscript, FAS acquired funding. 629

630

631

References 632

Adam S, Polo SE, Almouzni G. 2013. Transcription recovery after DNA damage requires 633

chromatin priming by the H3.3 histone chaperone HIRA. Cell 155:94–106. 634

Arribere JA, Bell RT, Fu BXH, Artiles KL, Hartman PS, Fire AZ. 2014. Efficient marker-free 635

recovery of custom genetic modifications with CRISPR/Cas9 in Caenorhabditis elegans. 636

Genetics 198:837–846. 637

Besnard E, Babled A, Lapasset L, Milhavet O, Parrinello H, Dantec C, Marin J-M, Lemaitre J-M. 638

2012. Unraveling cell type-specific and reprogrammable human replication origin signatures 639

associated with G-quadruplex consensus motifs. Nat Struct Mol Biol 19:837–844. 640

Bester AC, Roniger M, Oren YS, Im MM, Sarni D, Chaoat M, Bensimon A, Zamir G, Shewach 641

DS, Kerem B. 2011. Nucleotide deficiency promotes genomic instability in early stages of 642

cancer development. Cell 145:435–446. 643

Bianchi L, Driscoll M. 2006. Culture of embryonic C. elegans cells for electrophysiological and 644

pharmacological analyses. WormBook 1–15. 645

Blow JJ, Dutta A. 2005. Preventing re-replication of chromosomal DNA. Nat Rev Mol Cell Biol 646

6:476–486. 647

Boeck ME, Huynh C, Gevirtzman L, Thompson OA, Wang G, Kasper DM, Reinke V, Hillier LW, 648

27

Waterston RH. 2016. The time-resolved transcriptome of C. elegans. Genome Res 649

26:1441–1450. 650

Boos D, Ferreira P. 2019. Origin Firing Regulations to Control Genome Replication Timing. 651

Genes 10. doi:10.3390/genes10030199 652

Brauchle M, Baumer K, Gönczy P. 2003. Differential Activation of the DNA Replication 653

Checkpoint Contributes to Asynchrony of Cell Division in C. elegans Embryos. Current 654

Biology. doi:10.1016/s0960-9822(03)00295-1 655

Bruck I, Kaplan DL. 2015. Conserved mechanism for coordinating replication fork helicase 656

assembly with phosphorylation of the helicase. Proc Natl Acad Sci U S A 112:11223–657

11228. 658

Buschbeck M, Hake SB. 2017. Variants of core histones and their roles in cell fate decisions, 659

development and cancer. Nat Rev Mol Cell Biol 18:299–314. 660

Çelen İ, Doh JH, Sabanayagam CR. 2018. Effects of liquid cultivation on gene expression and 661

phenotype of C. elegans. BMC Genomics 19:562. 662

Chari S, Wilky H, Govindan J, Amodeo AA. 2019. Histone concentration regulates the cell cycle 663

and transcription in early development. Development 146. doi:10.1242/dev.177402 664

Clément C, Orsi GA, Gatto A, Boyarchuk E, Forest A, Hajj B, Miné-Hattab J, Garnier M, Gurard-665

Levin ZA, Quivy J-P, Almouzni G. 2018. High-resolution visualization of H3 variants during 666

replication reveals their controlled recycling. Nat Commun 9:3181. 667

Costantino L, Sotiriou SK, Rantala JK, Magin S, Mladenov E, Helleday T, Haber JE, Iliakis G, 668

Kallioniemi OP, Halazonetis TD. 2014. Break-induced replication repair of damaged forks 669

induces genomic duplications in human cells. Science 343:88–91. 670

Couldrey C, Carlton MB, Nolan PM, Colledge WH, Evans MJ. 1999. A retroviral gene trap 671

insertion into the histone 3.3A gene causes partial neonatal lethality, stunted growth, 672

neuromuscular deficits and male sub-fertility in transgenic mice. Hum Mol Genet 8:2489–673

2495. 674

28

Crane E, Bian Q, McCord RP, Lajoie BR, Wheeler BS, Ralston EJ, Uzawa S, Dekker J, Meyer 675

BJ. 2015. Condensin-driven remodelling of X chromosome topology during dosage 676

compensation. Nature. doi:10.1038/nature14450 677

Deal RB, Henikoff JG, Henikoff S. 2010. Genome-wide kinetics of nucleosome turnover 678

determined by metabolic labeling of histones. Science 328:1161–1164. 679

Delaney K, Mailler J, Wenda JM, Gabus C, Steiner FA. 2018. Differential Expression of Histone 680

H3.3 Genes and Their Role in Modulating Temperature Stress Response in. Genetics 681

209:551–565. 682

Delaney K, Strobino M, Wenda JM, Pankowski A, Steiner FA. 2019. H3.3K27M-induced 683

chromatin changes drive ectopic replication through misregulation of the JNK pathway in C. 684

elegans. Nat Commun 10:1–15. 685

Dellino GI, Cittaro D, Piccioni R, Luzi L, Banfi S, Segalla S, Cesaroni M, Mendoza-Maldonado 686

R, Giacca M, Pelicci PG. 2013. Genome-wide mapping of human DNA-replication origins: 687

levels of transcription at ORC1 sites regulate origin selection and replication timing. 688

Genome Res 23:1–11. 689

Desprat R, Thierry-Mieg D, Lailler N, Lajugie J, Schildkraut C, Thierry-Mieg J, Bouhassira EE. 690

2009. Predictable dynamic program of timing of DNA replication in human cells. Genome 691

Res 19:2288–2299. 692

Dileep V, Gilbert DM. 2018. Single-cell replication profiling to measure stochastic variation in 693

mammalian replication timing. Nat Commun 9:427. 694

Eaton ML, Prinz JA, MacAlpine HK, Tretyakov G, Kharchenko PV, MacAlpine DM. 2011. 695

Chromatin signatures of the Drosophila replication program. Genome Res 21:164–174. 696

Ekundayo B, Bleichert F. 2019. Origins of DNA replication. PLoS Genet 15:e1008320. 697

Farkash-Amar S, Lipson D, Polten A, Goren A, Helmstetter C, Yakhini Z, Simon I. 2008. Global 698

organization of replication time zones of the mouse genome. Genome Res 18:1562–1570. 699

Feng H, Kipreos ET. 2003. Preventing DNA re-replication--divergent safeguards in yeast and 700

29

metazoa. Cell Cycle 2:431–434. 701

Fragkos M, Ganier O, Coulombe P, Méchali M. 2015. DNA replication origin activation in space 702

and time. Nat Rev Mol Cell Biol 16:360–374. 703

Frey A, Listovsky T, Guilbaud G, Sarkies P, Sale JE. 2014. Histone H3.3 is required to maintain 704

replication fork progression after UV damage. Curr Biol 24:2195–2201. 705

Groth A, Corpet A, Cook AJL, Roche D, Bartek J, Lukas J, Almouzni G. 2007. Regulation of 706

replication fork progression through histone supply and demand. Science 318:1928–1931. 707

Guo C, Kumagai A, Schlacher K, Shevchenko A, Shevchenko A, Dunphy WG. 2015. Interaction 708

of Chk1 with Treslin negatively regulates the initiation of chromosomal DNA replication. Mol 709

Cell 57:492–505. 710

Hansen RS, Thomas S, Sandstrom R, Canfield TK, Thurman RE, Weaver M, Dorschner MO, 711

Gartler SM, Stamatoyannopoulos JA. 2010. Sequencing newly replicated DNA reveals 712

widespread plasticity in human replication timing. Proc Natl Acad Sci U S A 107:139–144. 713

Henikoff S, Smith MM. 2015. Histone variants and epigenetics. Cold Spring Harb Perspect Biol 714

7:a019364. 715

Hiratani I, Ryba T, Itoh M, Rathjen J, Kulik M, Papp B, Fussner E, Bazett-Jones DP, Plath K, 716

Dalton S, Rathjen PD, Gilbert DM. 2010. Genome-wide dynamics of replication timing 717

revealed by in vitro models of mouse embryogenesis. Genome Res 20:155–169. 718

Hiratani I, Ryba T, Itoh M, Yokochi T, Schwaiger M, Chang C-W, Lyou Y, Townes TM, 719

Schübeler D, Gilbert DM. 2008. Global reorganization of replication domains during 720

embryonic stem cell differentiation. PLoS Biol 6:e245. 721

Hödl M, Basler K. 2009. Transcription in the Absence of Histone H3.3. Current Biology. 722

doi:10.1016/j.cub.2009.05.048 723

Holway AH, Kim S-H, La Volpe A, Michael WM. 2006. Checkpoint silencing during the DNA 724

damage response in Caenorhabditis elegans embryos. J Cell Biol 172:999–1008. 725

Huang W, Loganantharaj R, Schroeder B, Fargo D, Li L. 2013. PAVIS: a tool for Peak 726

30

Annotation and Visualization. Bioinformatics 29:3097–3099. 727

Hyun M, Choi S, Stevnsner T, Ahn B. 2016. The Caenorhabditis elegans Werner syndrome 728

protein participates in DNA damage checkpoint and DNA repair in response to CPT-729

induced double-strand breaks. Cell Signal 28:214–223. 730

Jang C-W, Shibata Y, Starmer J, Yee D, Magnuson T. 2015. Histone H3.3 maintains genome 731

integrity during mammalian development. Genes Dev 29:1377–1392. 732

Kamath RS, Fraser AG, Dong Y, Poulin G, Durbin R, Gotta M, Kanapin A, Le Bot N, Moreno S, 733

Sohrmann M, Welchman DP, Zipperlen P, Ahringer J. 2003. Systematic functional analysis 734

of the Caenorhabditis elegans genome using RNAi. Nature 421:231–237. 735

Kanemaki M, Labib K. 2006. Distinct roles for Sld3 and GINS during establishment and 736

progression of eukaryotic DNA replication forks. EMBO J 25:1753–1763. 737

Kim Y, Kipreos ET. 2007. The Caenorhabditis elegans replication licensing factor CDT-1 is 738

targeted for degradation by the CUL-4/DDB-1 complex. Mol Cell Biol 27:1394–1406. 739

Kramer M, Kranz A-L, Su A, Winterkorn LH, Albritton SE, Ercan S. 2015. Developmental 740

Dynamics of X-Chromosome Dosage Compensation by the DCC and H4K20me1 in C. 741

elegans. PLoS Genet 11:e1005698. 742

Kumagai A, Shevchenko A, Shevchenko A, Dunphy WG. 2010. Treslin collaborates with 743

TopBP1 in triggering the initiation of DNA replication. Cell 140:349–359. 744

Labib K, Diffley JFX. 2001. Is the MCM2–7 complex the eukaryotic DNA replication fork 745

helicase? Curr Opin Genet Dev 11:64–70. 746

Langley AR, Gräf S, Smith JC, Krude T. 2016. Genome-wide identification and characterisation 747

of human DNA replication origins by initiation site sequencing (ini-seq). Nucleic Acids Res 748

44:10230–10247. 749

Liu Q, Guntuku S, Cui XS, Matsuoka S, Cortez D, Tamai K, Luo G, Carattini-Rivera S, DeMayo 750

F, Bradley A, Donehower LA, Elledge SJ. 2000. Chk1 is an essential kinase that is 751

regulated by Atr and required for the G(2)/M DNA damage checkpoint. Genes Dev 752

31

14:1448–1459. 753

MacAlpine DM, Rodríguez HK, Bell SP. 2004. Coordination of replication and transcription along 754

a Drosophila chromosome. Genes Dev 18:3094–3105. 755

MacAlpine HK, Gordân R, Powell SK, Hartemink AJ, MacAlpine DM. 2010. Drosophila ORC 756

localizes to open chromatin and marks sites of cohesin complex loading. Genome Res 757

20:201–211. 758

Macheret M, Halazonetis TD. 2019. Monitoring early S-phase origin firing and replication fork 759

movement by sequencing nascent DNA from synchronized cells. Nat Protoc 14:51–67. 760

Macheret M, Halazonetis TD. 2018. Intragenic origins due to short G1 phases underlie 761

oncogene-induced DNA replication stress. Nature 555:112–116. 762

Macheret M, Halazonetis TD. 2015. DNA replication stress as a hallmark of cancer. Annu Rev 763

Pathol 10:425–448. 764

Mantiero D, Mackenzie A, Donaldson A, Zegerman P. 2011. Limiting replication initiation factors 765

execute the temporal programme of origin firing in budding yeast. EMBO J 30:4805–4814. 766

Marchal C, Sima J, Gilbert DM. 2019. Control of DNA replication timing in the 3D genome. Nat 767

Rev Mol Cell Biol. doi:10.1038/s41580-019-0162-y 768

Mesner LD, Valsakumar V, Cieslik M, Pickin R, Hamlin JL, Bekiranov S. 2013. Bubble-seq 769

analysis of the human genome reveals distinct chromatin-mediated mechanisms for 770

regulating early- and late-firing origins. Genome Res 23:1774–1788. 771

Michalet X, Ekong R, Fougerousse F, Rousseaux S, Schurra C, Hornigold N, van Slegtenhorst 772

M, Wolfe J, Povey S, Beckmann JS, Bensimon A. 1997. Dynamic molecular combing: 773

stretching the whole human genome for high-resolution studies. Science 277:1518–1523. 774

Miotto B, Ji Z, Struhl K. 2016. Selectivity of ORC binding sites and the relation to replication 775

timing, fragile sites, and deletions in cancers. Proc Natl Acad Sci U S A 113:E4810–9. 776

Moiseeva TN, Yin Y, Calderon MJ, Qian C, Schamus-Haynes S, Sugitani N, Osmanbeyoglu 777

HU, Rothenberg E, Watkins SC, Bakkenist CJ. 2019. An ATR and CHK1 kinase signaling 778

32

mechanism that limits origin firing during unperturbed DNA replication. Proc Natl Acad Sci 779

U S A 116:13374–13383. 780

Paranjape NP, Calvi BR. 2016. The Histone Variant H3.3 Is Enriched at Drosophila Amplicon 781

Origins but Does Not Mark Them for Activation. G3 6:1661–1671. 782

Petryk N, Kahli M, d’Aubenton-Carafa Y, Jaszczyszyn Y, Shen Y, Silvain M, Thermes C, Chen 783

C-L, Hyrien O. 2016. Replication landscape of the human genome. Nature 784

Communications. doi:10.1038/ncomms10208 785

Pourkarimi E, Bellush JM, Whitehouse I. 2016. Spatiotemporal coupling and decoupling of gene 786

transcription with DNA replication origins during embryogenesis in. Elife 5. 787

doi:10.7554/eLife.21728 788

Prioleau M-N, MacAlpine DM. 2016. DNA replication origins-where do we begin? Genes Dev 789

30:1683–1697. 790

Ramírez F, Ryan DP, Grüning B, Bhardwaj V, Kilpert F, Richter AS, Heyne S, Dündar F, Manke 791

T. 2016. deepTools2: a next generation web server for deep-sequencing data analysis. 792

Nucleic Acids Res 44:W160–5. 793

Rodríguez-Martínez M, Pinzón N, Ghommidh C, Beyne E, Seitz H, Cayrou C, Méchali M. 2017. 794

The gastrula transition reorganizes replication-origin selection in Caenorhabditis elegans. 795

Nat Struct Mol Biol 24:290–299. 796

Sakai A, Schwartz BE, Goldstein S, Ahmad K. 2009. Transcriptional and Developmental 797

Functions of the H3.3 Histone Variant in Drosophila. Current Biology. 798

doi:10.1016/j.cub.2009.09.021 799

Schindelin J, Arganda-Carreras I, Frise E, Kaynig V, Longair M, Pietzsch T, Preibisch S, 800

Rueden C, Saalfeld S, Schmid B, Tinevez J-Y, White DJ, Hartenstein V, Eliceiri K, 801

Tomancak P, Cardona A. 2012. Fiji: an open-source platform for biological-image analysis. 802

Nat Methods 9:676–682. 803

Schmid M, Durussel T, Laemmli UK. 2004. ChIC and ChEC; genomic mapping of chromatin 804

33

proteins. Mol Cell 16:147–157. 805

Schübeler D, Scalzo D, Kooperberg C, van Steensel B, Delrow J, Groudine M. 2002. Genome-806

wide DNA replication profile for Drosophila melanogaster: a link between transcription and 807

replication timing. Nat Genet 32:438–442. 808

Sonneville R, Querenet M, Craig A, Gartner A, Blow JJ. 2012. The dynamics of replication 809

licensing in live Caenorhabditis elegans embryos. J Cell Biol 196:233–246. 810

Stevens H, Williams AB, Matthew Michael W. 2016. Cell-Type Specific Responses to DNA 811

Replication Stress in Early C. elegans Embryos. PLOS ONE. 812

doi:10.1371/journal.pone.0164601 813

Stroud H, Otero S, Desvoyes B, Ramírez-Parra E, Jacobsen SE, Gutierrez C. 2012. Genome-814

wide analysis of histone H3.1 and H3.3 variants in Arabidopsis thaliana. Proc Natl Acad Sci 815

U S A 109:5370–5375. 816

Takahashi S, Miura H, Shibata T, Nagao K, Okumura K, Ogata M, Obuse C, Takebayashi S-I, 817

Hiratani I. 2019. Genome-wide stability of the DNA replication program in single 818

mammalian cells. Nat Genet 51:529–540. 819

Tanaka S, Nakato R, Katou Y, Shirahige K, Araki H. 2011. Origin association of Sld3, Sld7, and 820

Cdc45 proteins is a key step for determination of origin-firing timing. Curr Biol 21:2055–821

2063. 822

Tang MCW, Jacobs SA, Mattiske DM, Soh YM, Graham AN, Tran A, Lim SL, Hudson DF, 823

Kalitsis P, O’Bryan MK, Wong LH, Mann JR. 2015. Contribution of the two genes encoding 824

histone variant h3.3 to viability and fertility in mice. PLoS Genet 11:e1004964. 825

Técher H, Koundrioukoff S, Nicolas A, Debatisse M. 2017. The impact of replication stress on 826

replication dynamics and DNA damage in vertebrate cells. Nat Rev Genet 18:535–550. 827

Therizols P, Illingworth RS, Courilleau C, Boyle S, Wood AJ, Bickmore WA. 2014. Chromatin 828

decondensation is sufficient to alter nuclear organization in embryonic stem cells. Science 829

346:1238–1242. 830

34

White EJ, Emanuelsson O, Scalzo D, Royce T, Kosak S, Oakeley EJ, Weissman S, Gerstein M, 831

Groudine M, Snyder M, Schübeler D. 2004. DNA replication-timing analysis of human 832

chromosome 22 at high resolution and different developmental states. Proc Natl Acad Sci U 833

S A 101:17771–17776. 834

Wollmann H, Stroud H, Yelagandula R, Tarutani Y, Jiang D, Jing L, Jamge B, Takeuchi H, 835

Holec S, Nie X, Kakutani T, Jacobsen SE, Berger F. 2017. The histone H3 variant H3.3 836

regulates gene body DNA methylation in Arabidopsis thaliana. Genome Biol 18:94. 837

Woodfine K, Fiegler H, Beare DM, Collins JE, McCann OT, Young BD, Debernardi S, Mott R, 838

Dunham I, Carter NP. 2004. Replication timing of the human genome. Human Molecular 839

Genetics. doi:10.1093/hmg/ddh016 840

Woodward AM, Göhler T, Luciani MG, Oehlmann M, Ge X, Gartner A, Jackson DA, Blow JJ. 841

2006. Excess Mcm2-7 license dormant origins of replication that can be used under 842

conditions of replicative stress. J Cell Biol 173:673–683. 843

Zentner GE, Kasinathan S, Xin B, Rohs R, Henikoff S. 2015. ChEC-seq kinetics discriminates 844

transcription factor binding sites by DNA sequence and shape in vivo. Nat Commun 6:8733. 845

Zhong W, Feng H, Santiago FE, Kipreos ET. 2003. CUL-4 ubiquitin ligase maintains genome 846

stability by restraining DNA-replication licensing. Nature 423:885–889. 847

848

Figures and legends: 849

850

35

851

Figure 1. Replication timing is unchanged in absence of H3.3 (A) Schematic description of 852

Repli-seq. Embryonic cells were dissociated, exposed to an EdU pulse for 10 min and sorted 853

36

according to their DNA content. EdU-labelled DNA was sequenced and mapped to the genome. 854

(B) Representative genome browser views of Repli-seq and H3.3 ChIP-seq. Repli-seq signal is 855

shown for wild type (WT) and H3.3 null mutant (∆ H3.3) worms on regions of chromosomes I 856

and V. Early S phase is shown in black and late S phase in red. HIS-72 (H3.3) ChIP-seq signal 857

is shown in green for the same regions. ChIP-seq data from (Delaney et al., 2019). (C) Color-858

coded replication timing for each chromosome. Repli-seq signal from early (black), mid (orange) 859

and late (red) S phase for WT and ∆ H3.3. Data for each chromosome was sorted according to 860

the signal of early S phase in WT. (D) Gene expression levels for the genes present in domains 861

of early and late replication. RNA-seq data from (Kramer et al., 2015). (E) Genes sizes for the 862

genes present in domains of early and late replication. 863

864

37

865

Figure 2. Identification of replication origins in C. elegans embryos. (A) Schematic 866

description of the roles of CDT-1 and TRES-1 in replication origin firing. CDT-1 is required for 867

the licensing of all origins. TRES-1 is recruited only to origins that fire. (B) Localization of 868

38

HA::CDT-1 and FLAG::TRES-1 during the cell cycle. Immunofluorescence images using anti-HA 869

and anti-FLAG antibodies and DAPI are shown. Scale bar represents 5 µm. (C) Schematic 870

description of ChEC-seq. The protein of interest is fused with MNase. Upon activation with 871

calcium in purified nuclei, MNase cleaves and releases DNA fragments at the binding sites of 872

the fusion proteins. These small fragments are isolated and sequenced. (D) Representative 873

genome browser views of CDT-1 ChEC-seq (blue), TRES-1 ChEC-seq (violet) and EdU-seq 874

(pink) signal. Identified origins are highlighted in gray (licensed only) or dark and light green 875

(licensed and firing with and without EdU signal, respectively). Domains of late (red) and early 876

(gray) replication and positions of replication origins identified in previous studies are shown 877

below the genome browser. 878

879

39

880

Figure 3. Classification of replication origins in C. elegans embryos. (A) Heatmaps (top) 881

and average plots (bottom) of CDT-1 ChEC-seq (blue), TRES-1 ChEC-seq (violet) and EdU-seq 882

(pink) signal at replication origins. Origins were identified by peak calling using the CDT-1 883

ChEC-seq, TRES-1 ChEC-seq and EdU-seq datasets obtained from worms grown at 20°C and 884

separated into early, late and dormant origins through unsupervised clustering. Signal is shown 885

for a 50 kb window around each origin. (B) Heatmaps (top) and average plots (bottom) of HIS-886

72 (H3.3) ChIP-seq signal (Delaney et al., 2019) at replication origins, as in (A). 887

888

40

889

Figure 4. Replication origin dynamics is altered H3.3 null mutants. (A) Heatmaps (top) and 890

average plots (bottom) of TRES-1 ChEC-seq signal at replication origins, for wildtype (WT, 891

41

violet) and H3.3 null mutant (∆ H3.3, gray) worms grown at 25°C. Signal is shown for a 50 kb 892

window around each origin. (B, C) EdU-seq time course, for WT (pink) and ∆ H3.3 (gray) worms 893

grown at 25°C. (B) Genome browser views of EdU-seq signal at 0, 30 and 60 minutes, showing 894

fork progression at representative examples of early, late and dormant origins. (C) Heatmaps 895

(top) and average plots (bottom) of EdU-seq signal at 0, 30 and 60 minutes, at all origins. Signal 896

is shown for a 50 kb window around each origin. 897

898

42

899

Figure 5. Replication fork speed is not affected by loss of H3.3.. (A) Representative 900

examples of DNA combing images used to measure replication fork speed for wildtype (WT) 901

and H3.3 null mutant (∆ H3.3). IdU incorporation is shown in green and CldU incorporation in 902

magenta. Scale bar represents 10µm. (B) Fork speed determined by DNA combing at 20°C and 903

25°C for the WT and ∆ H3.3. 904

905

43

906

Figure 6. Loss of H3.3 results cell cycle delays and replication checkpoint activation. (A) 907

Determination of cell cycle length, from the onset of the cleavage furrow of the first cell division 908

(arrow) to nuclear envelope breakdown of the AB cell (star). Representative still images from a 909

movie of a developing H3.3 null mutant (∆ H3.3) embryo with and without depletion of chk-1 by 910

RNAi are shown. Scale bar represents 10µm. (B) Cell cycle timing for wild type (WT) and ∆ 911

H3.3 worms, with and without depletion of chk-1 by RNAi. (C) Embryonic lethality for WT and ∆ 912

H3.3, with and without depletion of chk-1 by RNAi. 913

914

44

915

Figure S1. Characteristics of early and late replicating domains. (A) Boundaries used to 916

sort embryonic cells according to their DNA content. Plot showing the number of cells counted 917

(cell counts) as a function of the amount of PI incorporation (PI-Area). Cells were sorted 918

according to their DNA content into early replication (gray), mid replication (yellow) and late 919

45

replication (red). (B) Color-coded replication timing for each chromosome. Repli-seq signal from 920

early (black), mid (orange) and late (red) S phase for wild type (WT) and H3.3 null mutant (∆ 921

H3.3) worms. Data is shown according to the chromosomal location. (C) Correlation plots 922

between Repli-seq data for early (left panel) or late (right panel) replication with HIS-72 ChIP-923

seq data (Delaney et al., 2019), in 10kb bins. The Spearman's rank correlation coefficient (r) is 924

shown. 925

926

46

927

Figure S2. Characteristics of replication origins. (A) FACS profiles from HU block 928

experiments for EdU-seq. Plots show distribution of EdU incorporation across cell populations in 929

absence of HU (left panel), upon HU block (center panel) and 30 minutes after release from the 930

HU block (right panel). (B) Classification of early, late and dormant origins found in this study 931

based on genomic location within (genic) or outside (non-genic) gene-coding regions. (C) Venn 932

diagrams representing the number of origins identified in this study that overlap with the origins 933

found in (Pourkarimi et al., 2016) (left) and in (Rodríguez-Martínez et al., 2017) (right). 934

935

47

936

Figure S3. Replication origin dynamics under non-stress conditions. EdU-seq time course, 937

for wild type (WT, pink) and H3.3 null mutant (∆ H3.3, gray) worms grown at 20°C. Heatmaps 938

(top) and average plots (bottom) of EdU-seq signal at 0, 30 and 60 minutes, at all origins. Signal 939

is shown for a 50 kb window around each origin. 940

941

48

942

Figure S4. Estimation of S-phase duration. Time course of EdU incorporation after HU 943

release at 20°C. Relative EdU incorporation is shown as a function of the time of incubation in 944

presence of EdU. The curve was obtained using a bell-shaped model with GraphPad Prism 945

version 8.0.0. 946

947

49

Table S1. Overview of C. elegans strains used in this study 948

949

Strain Genotype Source Description

N2 CGC Wild type

FAS43 his-69&his-70(uge44) III; his-72(tm2066) III; his-74(uge18) V; his-71(ok2289) X

Delaney et al., 2018

H3.3 null mutant

FAS50 cdt-1(uge34 [MNase::cdt-1]) I This study CDT-1 ChEC in WT background

FAS137 tres-1(uge102[MNase::tres-1]) I This study TRES-1 ChEC in WT background

FAS154 tres-1(uge104[MNase::tres-1]) I; his-69&his-70(uge44) III; his-72(tm2066) III; his-74(uge18) V; his-71(ok2289) X

This study TRES-1 ChEC in ∆H3.3 background

950