1

1

2

3

4

Cell lineage and cell cycling analyses of the 4d micromere using live imaging in the marine 5

annelid Platynereis dumerilii 6

7

8

9

B. Duygu Özpolat1,3,*, Mette Handberg-Thorsager2, Michel Vervoort1, 10

and Guillaume Balavoine1,* 11

12

AFFILIATIONS 13

1 – Institut Jacques Monod, Paris, France 14

2 – Max Planck Institute of Molecular Cell Biology and Genetics, Dresden, Germany 15

3 – Currently at the Marine Biological Laboratory, Woods Hole, MA, USA 16

17

*Correspondence to: 18

B. Duygu Özpolat (dozpolat@mbl.edu), Guillaume Balavoine (guillaume.balavoine@ijm.fr) 19

20

21

22

2

ABSTRACT 23

24

Cell lineage, cell cycle, and cell fate are tightly associated in developmental processes, but in 25

vivo studies at single-cell resolution showing the intricacies of these associations are rare due to 26

technical limitations. In this study on the marine annelid Platynereis dumerilii, we investigated 27

the lineage of the 4d micromere, using high-resolution long-term live imaging complemented 28

with a live-cell cycle reporter. 4d is the origin of mesodermal lineages and the germline in many 29

spiralians. We traced lineages at single-cell resolution within 4d and demonstrate that 30

embryonic segmental mesoderm forms via teloblastic divisions, as in clitellate annelids. We also 31

identified the precise cellular origins of the larval mesodermal posterior growth zone. We found 32

that differentially-fated progeny of 4d (germline, segmental mesoderm, growth zone) display 33

significantly different cell cycling. This work has evolutionary implications, sets up the 34

foundation for functional studies in annelid stem cells, and presents newly-established 35

techniques for live-imaging marine embryos. 36

37

Keywords: live imaging, lineage tracing, cell cycle, teloblasts, primordial germ cells, mesoderm, 38

segmentation, polychaete, annelid 39

40

41

42

43

3

INTRODUCTION 44

Development of a multicellular organism requires precise regulation of the cell cycle, which has crucial 45

roles in cell lineage establishment, cell fate decisions, and maintenance of pluripotency. Many embryonic 46

and post-embryonic developmental processes involve stem cells that repeatedly give rise to tissue 47

founder cells while also self-renewing at each round of division. Cell cycle regulation defines the correct 48

timing and pacing of divisions for generating the progenitor cells, as well as maintaining the potency of 49

stem cells themselves (Ables and Drummond-Barbosa, 2013; Barker, 2014; Yasugi and Nishimura, 2016). 50

Understanding the cell cycle characteristics of stem cells and the implications of cell cycle regulation 51

requires a combined lineage tracing and live-cell cycle analysis approach at single-cell resolution. 52

However, such high-resolution lineage tracing has been challenging in many traditional and emerging 53

animal model systems, due to a wide range of practical limitations that spans from the inaccessibility of 54

embryos or tissues of interest, to the unavailability of tools and techniques (reviewed in Kretzschmar and 55

Watt, 2012). Therefore, in order to understand cycling behavior of stem cells and their progeny in vivo, 56

studies are needed in organisms where continuous observations are feasible in intact individuals and 57

tissues. 58

Spiralians are a group of Protostomes including segmented worms (Annelida), mollusks, ribbon 59

worms, and flatworms, many of which undergo a stereotyped program of early cell divisions known as 60

spiral cleavage (Conklin, 1897; Henry, 2014; Lyons et al., 2012; Seaver, 2014; Wilson, 1892). Blastomeres 61

that arise from spiral cleavage show determinate cell fates and a strict correlation exists between cell 62

division timing and cell fate determination. The 4d micromere (also called M for Mesoblast) is one of the 63

micromeres created during the 4th spiral cleavage, and it is an evolutionarily conserved blastomere across 64

Spiralia (Lambert, 2008). In most spiralians, 4d gives rise to mesoderm, endoderm, and primordial germ 65

cells (PGCs) (Ackermann et al., 2005; Gline et al., 2011; Kang et al., 2002; Lyons et al., 2012; Meyer et al., 66

4

2010; Rebscher, 2014; Shimizu and Nakamoto, 2014; Swartz et al., 2008). Within a given species, 4d 67

follows stereotypical division patterns. Differences across species in the 4d lineage division program have 68

been proposed as a mechanism for obtaining the diverse body plans present across spiralians (Lyons et 69

al., 2012). Yet, despite the immense diversity within spiralians, only a few studies have looked into the 4d 70

lineage in detail, and some of these studies could not employ high-resolution lineage tracing due to 71

techniques used (Fischer and Arendt, 2013; Gline et al., 2011, 2009; Goto et al., 1999b; Lyons et al., 72

2012). In addition, very limited data is available detailing the relationship between the 4d lineage cell 73

cycle characteristics and the resulting differences in cell fates (Bissen, 1995; Bissen and Weisblat, 1989; 74

Smith and Weisblat, 1994). Thus, the spiralian 4d lineage provides an exciting embryonic stem cell model 75

system for linking cell fate potency, cell lineage, cell cycle, and morphogenetic processes. 76

77

Annelids (segmented worms) are a large but understudied group of spiralians containing many 78

species with broadcast-spawning, giving large numbers of relatively small, yolk-poor and translucent 79

embryos and larvae, which are amenable to functional developmental studies. Annelids are 80

characterized by repeated (metameric) body parts called segments (Balavoine, 2014). In clitellate 81

annelids, a taxon including the leeches and earth worms (Zrzavý et al., 2009), founder cells of segmental 82

tissues are generated during embryonic development by stem cells called teloblasts (Anderson, 1973a; 83

Devries, 1973, 1972; Goto et al., 1999b; Penners, 1924; Weisblat and Shankland, 1985; Zackson, 1982). 84

Specific teloblasts make specific tissue types. The 4d micromere is the originator of the M teloblasts 85

(mesoteloblasts) that make the trunk mesoderm: the first division of 4d generates two bilaterally 86

symmetric stem cells named Mesoteloblast-Left (ML) and Mesoteloblast-Right (MR). ML and MR then 87

make the left and right mesodermal bands, respectively, via a well-described division program (Weisblat 88

and Shankland, 1985; Zackson, 1982): the teloblasts repeatedly divide asymmetrically to self-renew the 89

ML/MR stem cells and to give rise to tissue precursor cells (primary blast cells) through iterated divisions. 90

5

Each primary blast cell (much smaller in size compared to the teloblasts they have split from) follows a 91

stereotyped program of cell divisions with fixed fate, generating clonal regions of tissues in adjoining 92

segments. Micromere 4d and its daughters ML and MR are evolutionarily conserved embryonic stem 93

cells across spiralians (Lambert, 2008; Lyons et al., 2012). Their teloblastic nature in non-clitellate 94

annelids has been suggested before (Anderson, 1973b; Fischer and Arendt, 2013), but direct evidence for 95

teloblasts outside of clitellate annelids is still missing. 96

97

Platynereis dumerilii (P. dumerilii) is a marine annelid (Errantia, Nereididae) suitable to address 98

the questions outlined above. ásà a à E a tà Pol haete ,à P. dumerilii is phylogenetically distant from 99

clitellates (Struck et al., 2011; Weigert and Bleidorn, 2016) and presumably much closer in anatomy to 100

the last common ancestor of annelids (Balavoine, 2014). Based on comparative genome analyses, P. 101

dumerilii has also been suggested to belong to a slow-evolving lineage, thus potentially bearing genomic 102

ancestral features of annelids (Raible et al., 2005; Raible and Arendt, 2004). In addition, P. dumerilii has 103

externally-fertilized, relatively fast-developing, transparent embryos which can be injected for lineage 104

tracing and can be cultured at the lab for the full life cycle (Ackermann et al., 2005; Backfisch et al., 105

2014). Embryos develop into free-swimming planktonic larvae in about 24 hours-post-fertilization (hpf). 106

By 48 hpf, segmental organization starts to become apparent, mostly evident by the repetition of paired 107

bilateral bristle bundles (chaetae) on each segment (Fischer et al., 2010). At this stage, a mesodermal 108

posterior growth zone (MPGZ) has formed anterior to the presumptive pygidium (the posterior-most 109

non-segmental region), juxtaposed with the four putative PGCs (pPGCs). The MPGZ and pPGCs, as a cell 110

cluster, sit at the converging point of the left and right mesodermal bands, and both express Vasa mRNA 111

and protein (Rebscher et al., 2012, 2007). The first two divisions of ML and MR in P. dumerilii give rise to 112

the pPGCs (Fischer and Arendt, 2013). However, how the MPGZ and pPGCs end up next to each other, 113

and the exact embryonic origin of the MPGZ within the 4d lineage are not yet known. Previous studies 114

6

show that the mesodermal bands, and eventually the segmental mesoderm also originate from the 4d 115

micromere in P. dumerilii (Ackermann et al., 2005; Fischer and Arendt, 2013), but whether the segmental 116

mesoderm forms via stereotyped teloblastic divisions of primary blast cells (à la clitellate) is also 117

unknown. 118

Here, using high-resolution live imaging techniques complemented with a live-cell cycle reporter 119

we developed, we report an extensive analysis for the 4d lineage at single-cell resolution, and an 120

investigation of cell cycling patterns of several lineages that originate from the 4d micromere. We have 121

developed imaging techniques for both embryos and larvae that are easy to implement and can be 122

applied to other annelids and spiralians, as well as other metazoans with ciliated larvae. We show that a 123

pair of mesoteloblasts (ML and MR), similar to what has been observed in clitellate annelids, are active 124

during P. dumerilii embryogenesis and that they give rise to the mesodermal derivatives and pPGCs via 125

asymmetric cell divisions. A series of four contiguous primary blast cells produced on each side of the 126

larva proliferate to produce mesodermal blocks that each correspond to a distinct larval hemisegment. 127

We show that M cells, after having produced the four larval segments, undergo an abrupt transition in 128

their cycling behavior and start dividing much more slowly and symmetrically. These final divisions of the 129

mesoteloblasts give rise to cells that form the MPGZ in the early larvae. This MPGZ cells remain in 130

contact with the pPGCs, which are produced earlier and arrested in G0/G1. Differences we observed in 131

cell cycling patterns in differentially-fated lineages that originate from a single cell (4d) provide 132

foundational information to start delineating the relationship between cell cycle regulation, cell lineage, 133

and generation of different cell fates. 134

135

136

137

7

RESULTS 138

Establishment of a work flow for long time-lapse live imaging and cell lineage tracing in marine 139

embryos and larvae 140

Live imaging at single-cell resolution over extended time periods, and analysis of these 4D image 141

datasets for determining cell lineages require overcoming technical barriers. In this work, we used a 142

standard scanning confocal microscope, which allowed us to trace cells in a specific body region of the 143

embryo. To this end, we overcame a number of specific difficulties, piecing together a complete 144

workflow that could be applied to other marine embryos and larvae. For the immobilization of embryos 145

and larvae, we mounted samples in a thin layer of low-melting sea-water agarose using glass-bottom 146

dishes. After injection and a careful selection of normal developing embryos at the time of mounting, we 147

observed normal development of all embryos and larvae in agarose. For embedding swimming stages, 148

we developed a new deciliation protocol, allowing for permanent immobilization, even after cilia regrew 149

(see Materials and Methods). 150

Intense exposure to laser lights in a scanning confocal microscope caused phototoxicity 151

problems, resulting in embryonic deaths. To solve this problem, each image stack (at a resolution of 512 152

x 512 pixels) was acquired in roughly two minutes, and was followed by a recovery time of at least five 153

minutes. This time seemed to allow for the natural elimination of the toxic free radicals created by the 154

laser light, before reaching deleterious concentrations for the cells. In these conditions, we observed 155

development of the embryos and larvae that are morphologically indistinguishable from control animals. 156

This time resolution (about 7 minutes, or longer depending on the stack thickness imaged), 157

which helped limiting exposure of samples to the laser light, in turn limited the possibility of using 158

automated lineage tracing. In addition, for datasets that have low time- or image-resolution, or embryos 159

which have significantly variable cell sizes and nuclei, segmentation and lineage tracing algorithms 160

8

available via different platforms are currently not able to carry out an automated lineage analysis with 161

high accuracy (Ulman et al., 2017). Consequently, lineage tracing in such samples still largely rely on 162

manual curation. For determining cell lineages at single-cell resolution in Platynereis dumerilii (P. 163

dumerilii), we used a combination of labelling techniques. In addition to the conventional nuclear and 164

membrane labelling, we also used a live-cell cycle reporter. This added the necessary information to 165

trace lineages accurately, making it easier to predict when a cell will divide and follow daughter cells 166

after a division, without the need to increase time resolution and thus light exposure. 167

168

Construction of a cell cycle reporter and analysis of its cycling patterns 169

In order to visualize cell cycle progression in live P. dumerilii embryos and larvae, we first 170

constructed a fluorescent cell cycle reporter. Live-cell cycle reporters rely on fusion of a truncated cell 171

cycle protein containing a degron motif (also called destruction box) to a fluorescent protein. As a result, 172

the fluorescent protein becomes a visible reporter of the corresponding cell cycle phase, and gets 173

degraded when the endogenous cell cycle protein normally gets degraded. These cell cycle reporters are 174

also called FUCCI (fluorescent ubiquitination-based cell cycle indicator) (Sakaue-Sawano et al., 2008; 175

Zielke and Edgar, 2015). For developing a live-cell cycle reporter in P. dumerilii, we first identified in the 176

P. dumerilii genome and transcriptomes the cdt1 gene (Figure 1A), on which specific G1 phase reporters 177

are based in human and zebrafish (Sakaue-Sawano et al., 2008; Sugiyama et al., 2009). Metazoan Cdt1 178

proteins contain a well-characterized degron motif called PIP box (Q-x-x-[I/L/M/V]-T-D-[F/Y]-[F/Y]-x-x-x-179

[R/K]) (Havens and Walter, 2009), which interacts with the ubiquitin ligase complex CRL4 pathway for 180

degradation. The amino acid sequence of the degron in human Cdt1 (hCdt1) is QRRVTDFFARRR and it is 181

at position aa3-14. We found a conserved degron QTSVTNFFASRK at the same location in P. dumerilii 182

amino acid sequence (Figure 1-figure supplement 1). We also searched for a second degradation 183

9

peptide, the Cy motif (aa68-70 in hCdt1), that interacts with SCF E3 ligase for degradation (Fujita, 2006; 184

Nishitani et al., 2006). This second degron, however, cannot be identified unambiguously in metazoan 185

proteins (including P. dumerilii) outside of mammals. Assuming that the PIP box is the most important for 186

P. dumerilii Cdt1 degradation, we then proceeded by fusing a large truncated Pdu-Cdt1 sequence (aa1-187

147) containing the PIP box degron, but excluding the other functional interaction or DNA-binding 188

domains, to different fluorescent proteins (mVenus, mCherry, mKO2, mAG) (Figure 1A). The fused 189

constructs were then cloned into pCS2+ expression vector, transcribed in vitro, and injected as mRNA 190

into embryos at 1-cell stage (see Methods for details). Among all constructs tested, only mVenus-191

Cdt1(aa1-147) showed fluorescence located to the nucleus and clear cycling without producing 192

phenotypic effects (Figure 1B). The other constructs which had different fluorescent proteins but the 193

same Cdt1 peptide did not produce any fluorescent signal (a problem also encountered by other 194

researchers while establishing cell cycle reporters in other systems). 195

To understand the temporal characteristics of cycling patterns of mVenus-Cdt1(aa1-147), we 196

used live imaging and 5-ethynyl- ’-deoxyuridine (EdU) labeling assay. We first carried out a detailed 197

time-point analysis for individual cells using live imaging. HistoneH2A-mCherry mRNA was co-injected 198

with the mVenus-cdt1(aa1-147) mRNA for continuous nuclear labeling. The imaging was done in several 199

embryos (n=10, 4 replicates) but here we report results from a representative embryo (Videos 1.1 and 200

1.2). We found that mVenus-Cdt1(aa1-147) in P. dumerilii was present during G1 phase, and was then 201

degraded, probably during S phase. In contrast to the widely-used hCdt1 reporter, which is specific to 202

G1, the P. dumerilii Cdt1 reporter started accumulating again shortly before mitosis probably in late G2 203

(Figure B,àseeà ellà a àa dàthe à ellà à efo eà itosis .àCo se ue tl ,à Ve usàsig alà asàalsoàp ese tà204

during mitosis (can be seen in Videos 1.1, 1.2). Fluorescence was suddenly dispersed in the cytoplasm 205

when the nuclear envelope broke down during mitosis (because the truncated chimeric protein, unlike 206

the native Cdt1, does not bind to chromatin), and then quickly concentrated in the daughter nuclei when 207

10

nuclear envelope reformed at the end of telophase. We observed contrasting cycling patterns in sister 208

cells. As an example, nucleus a à etai edàtheà Ve usàsig alà fo à o eàtha àa àhou à Figure 1B, upper 209

row), while nucleus à kept signal for only about 8 minutes (Figure 1B, lower o .à Theà ellà à210

eventuall à di idedà atà à i utes à hileà ellà a à asà stillà i àG /“àphase.à Similar cycling patterns were 211

observed for other cycling cells in several videos we have obtained, thus establishing mVenus-Cdt1(aa1-212

147) as an efficient cycling marker. 213

Next, in order to determine more precisely when mVenus-Cdt1(aa1-147) was degraded, we used 214

S-phase-labeling with EdU. Cdt1 protein is part of the DNA-replication complex, normally present at the 215

beginning of S phase, but targeted for degradation during S phase in order to prevent re-replication 216

(Arias and Walter, 2006; Fujita, 2006; Liu et al., 2004). Thus, we tested whether mVenus-Cdt1(aa1-147) 217

signal overlapped with the EdU signal. Embryos injected with mVenus-cdt1(aa1-147) and HistoneH2A-218

mCherry mRNAs were incubated until 12 hours-post-fertilization (hpf), then treated with a very short 219

pulse of EdU (3 minutes) and immediately fixed after quick rinsing (as EdU detection cannot be carried 220

out live, specimens needed to be fixed). The EdU treatment was long enough to mark the S phase but 221

too short for most of the cells to progress much further in the cell cycle. As a result, the vast majority of 222

EdU(+) cells were expected to be in S phase at the time of sample fixation. We analyzed cell nuclei from 6 223

embryos (from 2 independent experiments) treated this way (Figure 1C-D’’’). We observed that 15.6% 224

(n=51) of cells were positive for mVenus, mCherry and EdU, demonstrating that mVenus-Cdt1(aa1-147) 225

was present during at least part of the S phase (Figure 1D). Only a few cells (2.1%, n=7) were positive for 226

mCherry and EdU, but not mVenus, suggesting that mVenus-Cdt1(aa1-147) was degraded at some point 227

during the S phase (Figure D’ . Taking into account that the vast majority but not all Edu(+) cells are also 228

mVenus(+), this strongly suggests that mVenus-Cdt1(aa1-147) is degraded towards the end of the S 229

phase and is no longer present as the cells enter the G2 phase. The majority of cells (75.3%, n=247) were 230

found to be positive for mVenus and mCherry, but lacked EdU signal (Figure D’’). These cells were either 231

11

in G1 or late G2 phase. Finally, cells that were only mCherry(+) marking the early G2 phase made up a 232

smaller percentage (7%, n=23) of all cells analyzed (Figure D’’’). We further confirmed that the cell 233

cycle reporter is working properly by analyzing the ciliated cells that are known to be terminally 234

differentiated in the annelid swimming larvae. These cells exit the cell cycle, and thus are in G0, no 235

longer degrading Cdt1. As expected, we observed that they accumulated mVenus-Cdt1(aa1-147), once 236

they stopped cycling. The mVenus signal reached much higher levels than in any cycling cell and 237

remained at high levels in these fully differentiated multiciliated cells (Figure 1-figure supplement 2, 238

Videos 2.1 and 2.2). 239

In P. dumerilii, mVenus-Cdt1(aa1-147) reporter is present during a larger portion of the cell cycle 240

(Figure 1E). It is being degraded during the course of the S phase, similar to the vertebrate reporter, and 241

it starts to accumulate again during the late G2 phase. This pattern reflects the behavior of endogenous 242

Cdt1 degradation reported in other organisms: In human cells, Cdt1 starts getting degraded in the S 243

phase, thus it is expected to be present at least partially during this phase (Shiomi et al., 2014). In 244

embryonic stem cells, Cdt1 starts accumulating at the end of G2 phase, which is promoted by its 245

interaction with Geminin (Ballabeni et al., 2004). Even though our mVenus-Cdt1(aa1-147) construct is 246

not specific to the G1 phase, its cycling behavior is very pronounced as the protein is at barely detectable 247

levels in late S and early G2 phases and therefore, together with HistoneH2A-mCherry, allows to follow 248

cell cycle progression in living embryos. We call the two constructs (HistoneH2A-mCherry and mVenus-249

Cdt1(aa1- àtogethe à Pdu-FUCCI ài àtheà estàofàthis manuscript. Next, we show our results of 4d 250

micromere lineage analysis at single-cell resolution, via live imaging combining Pdu-FUCCI with 251

membrane labeling. The layer of information into cell cycle progression provided by mVenus-Cdt1(aa1-252

147) was crucial in accurate manual curation of lineages in the experiments reported below. Because we 253

could not increase the time resolution due to phototoxicity, cell cycle progression was used for 254

predicting cell division, and differentiating the daughter cells from the surrounding cells. 255

12

256

257

258

259

Live imaging the cell cycling and migration behavior of putative primordial germ cells 260

In P. dumerilii,àtheà esode alà dà i o e eà alsoà alledà M àgi esà iseàtoàt oàs et i all à261

positioned sister cells ML (Meso lastà Left à a dà MRà Meso lastà Right by symmetric cell division 262

(Fischer and Arendt, 2013; Rebscher et al., 2012, 2007). The following two highly asymmetric divisions of 263

ML and MR give rise to four small cells (1ml/r and 2ml/r), which have been suggested to form the 264

germline (putative primordial germ cells, pPGCs). Thus, the pPGCs are the earliest lineage to segregate 265

from 4d in P. dumerilii. pPGCs become mitotically quiescent right after birth. As a result, they can be 266

labeled with a short pulse of EdU around their time of birth (between 3-8 hpf), and be detected even 267

days after EdU incorporation as they retain undiluted EdU while the neighboring somatic cells have 268

largely diluted EdU through multiple mitoses (Rebscher et al., 2012). We confirmed that when embryos 269

were treated between 5-7 hpf or 6-8 hpf, and raised to 24 hpf or 48 hpf all four pPGCs were labeled with 270

EdU under our culturing conditions (5-7 hpf treatment shown in Figure 2A for samples raised until and 271

fixed at 24 hpf, in Figure 2B for samples raised until and fixed at 48 hpf; 6-8 hpf data not shown). When 272

embryos were treated with EdU between 7-9 hpf, four out of six larvae had only two EdU positive pPGCs 273

at 72 hpf (data not shown), indicating that pPGCs stop incorporating EdU around this time frame. In 274

addition, to test whether pPGCs synthesized DNA after they were born in the next two days of 275

development, we carried out EdU incorporation assays after pPGCs were born, between 11-24 hpf 276

(Figure 2C) and 24-48 hpf (Figure 2D). We did not detect any pPGCs positive for EdU in the larvae (n>16) 277

13

in either of these treatments, confirming DNA synthesis does not take place during this time frame in the 278

pPGCs. 279

We then analyzed the pPGCs for their cell cycle state using Pdu-FUCCI injection and live imaging. 280

In this and following experiments, the injection cocktail included Pdu-FUCCI and an additional EGFP 281

construct with a cell membrane-localization signal (EGFP-caax) for visualization of individual cell shapes. 282

First, we injected 1-cell stage embryos with Pdu-FUCCI and EGFP-caax, and imaged the larvae live at 24 283

hpf and 48 hpf. We found that pPGCs were mVenus-Cdt1(aa1-147) positive at 24 hpf and 48 hpf, 284

suggesting that they are arrested in G0/G1 (Figure 2E, 2F). Next, in samples injected with the same 285

cocktail but only in the D quadrant at 4-cell stage, we carried out high time-resolution (every 10 minutes) 286

time-lapse imaging to determine if Pdu-FUCCI could allow tracking the pPGCs (Figure 3, Figure 3-figure 287

supplement 1, Video 3.1). We chose to inject the D quadrant only, as it eventually gives rise to the 4d 288

lineage from which pPGCs originate. In addition, injecting only the D quadrant facilitated the observation 289

of its descending lineages, because the rest of the embryo remained unlabeled. Starting from 8 hpf, 290

injected embryos were imaged for about 12 hours, with 10-minute intervals. Temperature conditions 291

during live imaging were higher than 18°C (which is the usual culture temperature for P. dumerilii) 292

(Fischer et al., 2010). Thus, development was accelerated during imaging allowing us to record a broader 293

developmental period, while the sample had normal development (Video 3.2). This imaging period 294

corresponds to the stages from stereoblastula to mid-late trochophore (roughly 8 hpf to 38 hpf at 18˚C) 295

(See Table 1 for time-points and corresponding stages at different temperatures). Using the high-296

resolution time-lapse video, we were able to track the pPGCs, and observed that mVenus signal persisted 297

in pPGCs after they were born (Figure 3A-H, Figure 3-figure supplement 1 for close-ups). Thus, in 298

addition to differentiated cells such as ciliated cells, Pdu-FUCCI is a useful tool to trace the cells that are 299

undifferentiated but arrested in G0/G1, such as the pPGCs. 300

14

We then investigated the migration behavior of pPGCs in the same imaging dataset from the 301

sample injected in the D quadrant with Pdu-FUCCI and EGFP-caax mix. When live imaging was started, 4d 302

cell had already divided once giving rise to the left and right mesoblasts (ML and MR), and ML and MR 303

had divided twice giving rise to two pPGCs each (Figure 3A-á’’, Figure 3-figure supplement 2 shows the 304

first two divisions of ML and MR in another sample). pPGCs at this stage (stereoblastula) were located on 305

the external surface of the embryo (Figure á’-á’’,àVideos 4.1, 4.2). However, soon after they were born 306

pPGCs started moving towards each other, formed a cluster, and then moved towards the vegetal pole 307

through epiboly (Figure B’-D’’). The movement of pPGCs continued on the surface until 20 hpf (or about 308

5 hours of imaging –Table 1) (Figure D’-D’’ . Then they started to position internally (Figure E’-G’’ , 309

eventually assuming a postero-dorsal position forming a bridge between the left and right mesodermal 310

bands which converge at the posterior midline (Figure 3H-H’’,à 2E, 2F). This analysis establishes a 311

framework for locating pPGCs through these developmental stages and allow for tracking these cells in 312

experimental manipulation conditions such as drug treatment or cell ablation in future studies. 313

314

Each mesodermal hemisegment is the progeny of individual blast cells produced in successive divisions 315

In addition to giving rise to pPGCs, 4d (M) blastomere is the origin of the trunk mesoderm in P. 316

dumerilii (Ackermann et al., 2005). The trunk somatic mesoderm of P. dumerili larvae is organized in 317

three largely similar segmental blocks (Brunet et al., 2015). Each mesodermal segment is composed of 318

identical left and right halves called mesodermal hemisegments. In clitellate annelids, ML and MR divide 319

asymmetrically giving rise to an anterior to posterior linear series of smaller segmental founder cells 320

calledà p i a à lastà ells à PBCs .àThere are as many PBCs on each side as there are segments. Each PBC 321

generates a metamerically-iterated clonal mesodermal domain distributed over three consecutive 322

hemisegments. However, each PBC makes o eàhe iseg e t- o th àofà esode (Gline et al., 2011; 323

15

Weisblat and Kuo, 2014). This stereotypic, anterior to posterior-oriented, asymmetric division process is 324

termed teloblastic growth. In P. dumerilii, to determine whether the 4d lineage (after the divisions that 325

generate pPGCs) makes the mesodermal hemisegments in the larva via the teloblastic process as in the 326

clitellates, we followed divisions of this lineage and analyzed the origin of mesodermal cell clusters (4d 327

lineage is summarized in Figure 4A). If mesoderm formation in P. dumerilii follows a teloblastic process, a 328

key expectation is the presence of clonally homogeneous mesodermal regions, each originating from one 329

specific primary blast cell as in the clitellate annelids. To do this, we manually traced cell lineages 330

originating from 4d blastomere in live imaging datasets using Imaris software (for details, see Materials 331

and Methods). Here we only report data for the left mesoblast (ML) (Sample A), but confirmed in other 332

samples (Samples B and C) that MR behaves like ML (Figure 5-figure supplement 1, Videos 5.1, 5.2). 333

Imaris creates a temporal series of fully re-orientable 3D reconstructions of the embryo, allowing 334

the manual tracking of cell divisions. Spots corresponding to the nucleus of each cell are generated and 335

cell lineages can be connected via these spots across time (Video 6.1). In addition to nuclear divisions, 336

the cell cycle reporter helped us predict when cell division was approaching, and the cell membrane 337

rounding (visualized via the EGFP-caax construct) helped us detect without ambiguity each mitosis and 338

the resulting daughter cells. Once a cell lineage tree is available, subsets of lineages can be highlighted 339

to visualize where the progeny is located in the 3D reconstruction. Using this feature, we highlighted the 340

progeny of each blast cell that splits from ML (Figure 4). We looked at the origin and clonality of cells 341

that formed each segmental mesoderm. Here we adopt a similar 3-character nomenclature (rank, fate, 342

and side) (Fischer and Arendt, 2013; Lyons et al., 2012). For example, 5ml is the small daughter, of the 343

fifth asymmetric division of the left M teloblast whereas its sister cell, 5ML is the self-renewed M 344

teloblast. 345

16

P. dumerilii has three true larval segments with bundles of chaetae on each side and one cryptic 346

anterior segment (alsoà alledà seg e tà à (Saudemont et al., 2008),à o à seg e tà I à (Steinmetz et al., 347

2011)). The cryptic segment does not bear chaetae and is incorporated into the head of the larva (Brunet 348

et al., 2015; Fischer et al., 2010; Steinmetz et al., 2011). We adopt the following order from anterior to 349

posterior progression: cryptic segment, segment 1, 2, and 3. The numbered segments are already clearly 350

identifiable in the final stacks of our videos corresponding to mid-trochophore stage, because two 351

ectodermal pockets corresponding to chaetal sacs are starting to form in each hemisegment, and 352

mesodermal tissues that will give the segmental musculature are starting to surround them (Video 3.1, 353

last frame). We identified four clear mesodermal sub-lineages that give distinct clonal domains of 354

segmental mesodermal precursor cells arranged in bilateral anterior to posterior series (Figure 4D-G, 355

Figure 4-figure supplement 1 .à Weà fou dà that,à p oge à ofà ea hà à lastà ellà f o à là toà à là356

contributes mostly to a specific larval segment: 6ml contributes to the larval hemisegment 1 (Figure 4E-357

E’’’’,àVideos 6.8, 6.9); 7ml to the larval hemisegment 2 (Figure 4F-F’’’’,àVideos 6.10, 6.11); and 8ml to the 358

larval hemisegment 3 (Figure 4G-G’’’’,àVideos 6.12, 6.13). These clonal blocks were easily discernible at 359

earlier stages of development as well (Figure 4-figure supplement 1). In addition, 5ml proliferates to give 360

a mesodermal block located just anterior to the three true segmental blocks originating from 6ml-8ml 361

and will possibly give rise to the mesoderm of the cryptic larval hemisegment (Figure 4D-D’’’’,àVideos 6.6, 362

6.7). These lineages are summarized in Video 7. 363

364

To confirm that the clonal blocks of cells generated by each blast cell corresponded to segmental 365

anlagen, we bring forward additional observations. First, we further analyzed the dorso-ventral position 366

of the clonal mesodermal clusters at the final time point (t=72, or about 37,5 hpf of development (Table 367

1) of the time lapse dataset. At this time point, chaetal sacs (ectodermal structures used as anatomical 368

17

markers of larval segment formation) had already started forming with a few elongating bristles evident, 369

and each clonal mesoderm cluster from a single PBC overlapped with a pair of chaetal sacs present in 370

each hemisegment (Figure 4-figure supplement 2A-B’, .àThen we analyzed the cross sections of the 3D 371

stacks with the Imaris lineage spots. We found that spots marking the mesodermal lineage lie just 372

beneath each pair of chaetal sacs within a larval hemisegment (Figure 4-figure supplement 2C-D’’ ,à373

confirming that these were mesodermal cells forming the layer beneath the elongating chaetal sacs at 374

this stage. Second, we analyzed another time-lapse movie showing the development of the trochophore 375

larva up to a stage where segmentation becomes thoroughly apparent (26 – 54 hpf) (Figure 4-figure 376

supplement 3, Video 8). Mesodermal blocks were initially beneath the elongating and deepening chaetal 377

sacs (Figure 4-figure supplement 3A-B, Video 8), but they eventually formed mesodermal pockets 378

surrounding the chaetal sacs like sheaths (Figure 4-figure supplement 3C-D, Video 8). Interestingly, 379

before the chaetal sacs started to elongate and sink into the mesodermal layer, clear mesodermal 380

segment boundaries were briefly visible, in the form of dorso-ventrally-stretched cells (Figure 4-figure 381

supplement 3C, arrows). In other models, these boundaries have been shown to correspond to acto-382

myosin cables that act as active compartmental limits preventing cell mixing (Monier et al., 2011). 383

Furthermore, in P. dumerilii, previous works have shown that segmental anlagen become distinguishable 384

at the transcriptional level already by 34 hpf, evident by the striped pattern of gene expression in the 385

pre-segmental mesoderm and ectoderm (Saudemont et al., 2008; Steinmetz et al., 2011). Thus, even 386

though segments are not anatomically distinct yet, at the molecular level mesodermal clusters start to 387

show evidence of segmentation. Together, these observations support early determination of segmental 388

anlage, and suggest that there is minimal (if any) mixing of mesodermal cells across segments at the 389

stages we carried out the mesodermal cell lineage analysis and after. 390

We also looked into the fate of blast cells segregated prior to these segmental lineages. 1ml 391

and 2ml (1mr and 2mr for the right side) corresponded to the bilateral pairs of pPGCs as described above 392

18

(Figure 4A-B). These cells did not divide further and stayed in the immediate vicinity of the M teloblastic 393

cells near the vegetal pole of the embryo. 3ml and 4ml blast cells, unlike 5ml to 8ml, underwent a limited 394

number of symmetric mitoses. These small 3ml and 4ml lineages drifted towards the animal pole. Due to 395

the orientation of the sample during imaging, it was not possible to trace the lineages 3ml and 4ml 396

confidently after a certain time point. Our confocal stacks cover only about 60 µm thickness (the 397

embryos are 160 µm) from the posterior end, and the progeny of 3ml and 4ml move out of the imaged 398

area. However, the limited information we obtained suggests they may be giving rise to anterior-most 399

non-segmental mesodermal tissues (Videos 6.2, 6.3, 6.4, 6.5). Overall, the data support that ML shows 400

teloblastic behavior and the progeny of mesodermal primary blast cells contribute to distinct larval 401

segments similar to what is observed in clitellate annelids. 402

403

Mesodermal posterior growth zone originates from 8ML and 8MR 404

At the 8th division of ML and MR, cells 8ML/R and 8ml/r are generated. As described above, 405

8ml/r are the founder cells of the 3rd segmental mesoderm (Figure 4G, Figure 4-figure supplement 1, 406

Video 7). We next investigated the progeny of 8ML/MR (Figure 5). We compared three separate 407

individuals (Samples A, B, C) for the 8ML and 8MR divisions and found them to be similar in all samples 408

(Figure 5-figure supplement 1, Videos 5.1, 5.2). At the time they were born, 8ML/MR were positioned 409

immediately next to the pPGCs (Figure 6U-U’;àFigure 4-figure supplement 1; Video 6.14, 6.15). By mid-410

trochophore stage (around time points 55-60) all individuals had three cells on the left and right sides 411

near pPGCs (Figure 5-figure supplement 1A-C’ .àT oàofàtheseà ellsàdi idedàe e tuall à a ou dàtheàsa eà412

time in the left and right sides, but not perfectly simultaneously) (Figure á’-á’’’;à Figure 5-figure 413

supplement 1á’’-C’’ àe pa di gàtheàp oge àofà MLàa dà MRàtoà à ellsào àea hàsideà àtotal .àThus,à à414

the last time point of our datasets (corresponding to the end of mid-trochophore stage), 8ML and 8MR 415

19

had given rise to five cells each that were surrounding the four pPGCsàa dàfo edàaàposte io à idge à416

between the right and left mesodermal bands. They were positioned slightly posterior to the pPGC 417

cluster (Figure 5B-B’’ ,à i à theà egion which has been previously described as the mesodermal posterior 418

growth zone (Rebscher et al., 2007). We thus suggest that these cells represent the founding lineage ring 419

of mesodermal stem cells that will be active later in juvenile development in the segment addition zone 420

(Gazave et al., 2013). 421

422

Both the asymmetry and orientation of embryonic mesoteloblast divisions gradually change 423

We analyzed the 4d lineage divisions at single-cell resolution for size asymmetry. The division 424

as et à asedào à ellà size à fo à d’sà t oàdaughte à ellsàMLàa dàMRàhasà ee àp e iousl àdes i edà425

using bright-field microscopy only until the 7th division (7ML/R and 7ml/r) conclusively (Fischer and 426

Arendt, 2013), and the divisions after this were not possible to observe due to the limitations of this 427

imaging technique. We took advantage of high-resolution confocal imaging with injection of the 428

fluorescent markers Pdu-FUCCI (nuclear) and EGFP-caax (cell membrane, making the cell size very easy 429

to observe). Here, we show the imaging dataset of an embryo (Sample A, same dataset as reported 430

above) injected into its D quadrant and live-imaged from 8 hpf until larval stages when segmental 431

mesoderm morphology started becoming obvious (Table 1, Video 3.1). In the confocal stacks, we were 432

able to investigate division of 4d lineage further than previously-reported, and we analyzed the size 433

asymmetry as well as cell division angle for this lineage. 434

The first division of 4d is equal, creating two sister mesoblast cells positioned bilaterally 435

symmetrically (ML and MR) (Figure 6A-B, Figure 3-figure supplement 2). However, most of the rest of the 436

divisions were unequal, as expected of teloblasts. The first two divisions of ML and MR, are the most 437

unequal. They gave rise to large daughter cells (1ML, 1MR, 2ML, 2MR) and very small ones (1ml, 1mr, 438

20

2ml, 2mr). 1ml/r and 2ml/r are the pPGCs as explained above (Figure 4A-C; Figure 3-figure supplement 439

2). For the next rounds of divisions, we report the data for ML in detail, but we confirmed MR behaves 440

similarly in two more samples investigated (Figure 5-figure supplement 1, Videos 5.1, 5.2). After the 441

initial divisions that gave rise to the pPGCs, the mesoblast (now named 2ML) continued to divide, 442

budding off small cells: division of 2ML, 3ML, 4ML, 5ML were all unequal (Figure 6E, 6G, 6J, 6L, 443

respectively). However, the size difference between the two daughter cells progressively decreased. This 444

seemed to be due to the fact that 3ml and 4ml were small precursors compared to the later segmental 445

precursors, 5ml, 6ml (compare Figure 6E, 6G to 6J, 6L) but also due to the gradual decrease in size of the 446

remaining M teloblast. Following this tendency, division of 6ML into 7ML and 7ml was roughly equal in 447

size (Figure 6O). We observed that 7ML then divided unequally again, giving rise to 8ML and 8ml (Figure 448

6S). But, surprisingly this time, the size difference was inverted: 8ml, precursor of the mesoderm of the 449

3rd segment was the large cell and 8ML, the remaining teloblast cell, was now very small. The further 450

divisions of this remaining teloblast 8ML are essentially equal and give a cluster of quite small cells. Thus, 451

we confirm that 4d lineage shows asymmetric divisions similar to clitellate mesoteloblasts. However, 452

unlike in the leech, in P. dumerilii the mesodermal teloblasts decreased in size very rapidly through this 453

8-division sequence, and we observed a reversed asymmetry at the end of the sequence. The 454

mesodermal teloblastic lineage (derived from 8ML/MR) that will presumably produce the segments of 455

the juvenile thus starts from a cluster of very small cells. 456

As observed in previous work (Fischer and Arendt, 2013), the orientation of divisions also 457

sequentially shifted in a clock-wise fashion for the left mesoteloblast divisions (Figure 6E, 6G, 6J, 6L, 6O, 458

6S, summarized in Figure 6V), when looking from the posterior pole. Consequently, whereas the first 459

series of precursors (pPGCs, 3ml, 4ml) were budded off roughly in the direction of the future blastopore, 460

the later segmental precursors (5ml to 8ml) were budded off progressively more towards the anterior of 461

21

the embryo, also reflecting the gradual change of size of the teloblast cell and its shifting position 462

towards the vegetal pole. 463

464

Divisions generating segmental mesoderm precursors happen in quick succession, then cell cycling 465

slows down significantly 466

We assessed whether the divisions giving rise to segmental precursors follow some kind of 467

seg e tatio à lo k ,à o pa a leà toà theà a à so itesà a eà p odu edà atà aà egula à ti eà i te alà i à468

vertebrates (Cooke and Zeeman, 1976). This is not the case in P. dumerilii as the time interval increases 469

gradually between divisions. Precisely, at the experimental temperatures 4ML, 5ML, 6ML, 7ML divisions 470

happened in 30 mins, 30 mins, 40 mins, 50 mins after the preceding division, respectively. Thus, there is 471

oà seg e tatio à lo k à utàaà seg e tatio àos illato àofàde easi gàf e ue ài àtheàe o.à 472

We next used the live cell cycle reporter Pdu-FUCCI to compare the cell cycling patterns of some 473

of the different lineages that arise from the 4d micromere. To do this, mean fluorescence intensity 474

information was extracted for each spot from the cell lineage datasets using statistics module in Imaris 475

and these were plotted as graphs (Figure 7A, Figure 7-source data 1). Each spot was manually placed 476

inside each nucleus (Figure á’), and mean fluorescence intensity of pixels within each spot was 477

calculated for subsets of lineages (see Materials and Methods for details). The measurement reflects the 478

abundance of mVenus-Cdt1(aa1-147) protein, providing quantitation of the cycling of the construct. 479

pPGCs stopped cycling after they were born (Figures 2, 3; Figure 3-figure supplement 1) and as expected, 480

a roughly linear plot, without significant valleys and peaks was obtained for these cells (Figure 7B). The 481

mesoblast ML, in contrast, had clear cycling with prominent valleys and peaks (Figure 7B, Figure 7-figure 482

supplement 1A). Next, we selected a sub-lineage within primary blast lineages 7ml, and 8ml, as well as a 483

sub-lineage within 8ML for analyzing cell cycling patterns. In segmental mesoderm lineages within 7ml 484

22

a dà l,à eàsele tedà e ui ale t àsu -lineages based on the location of daughter cells as the primary 485

blast cell divided (see Materials and Methods for details of the lineage selection criteria). The 486

comparison of 7ml sub-lineage to 8ml sub-lineage revealed that initially these two are out of phase 487

(valleys and peaks do not overlap) even though cycling periodicity is similar. However, by time point 46 488

they become synchronized as shown by the overlapping graphs (Figure 7C). On the other hand, 8ml and 489

8ML sub-lineages differ from each other drastically: 8ML has much fewer valleys and peaks (Figure 7D, 490

Figure 7-figure supplement 2A-E, Figure 7-figure supplement 1C), indicating slower cell cycling compared 491

to 8ml (Figure 7D, Figure 7-figure supplement 2F-J, Figure 7-figure supplement 1B). We have also 492

confirmed that the cycling patterns are similar in different samples and in the right mesoblast (Figure 7-493

figure supplement 2). Overall, these data show that live cell cycle reporter Pdu-FUCCI is a useful tool for 494

revealing changing cycling signatures within lineages and across lineages. This tool can be used for 495

similar purposes in other developmental processes and stages. 496

497

DISCUSSION 498

Here we report techniques for long-term live imaging of P. dumerilii embryos and larvae, and 499

demonstrate the feasibility of a FUCCI live-cell cycle reporter in this species. Using these techniques and 500

tools, we reveal previously unknown characteristics of the development of this marine annelid, in 501

particular the fact that the daughters of the blastomere 4d (ML and MR) produce the trunk mesoderm by 502

teloblastic divisions, as in clitellate annelids. Starting with the 5th division, each division of this pair of 503

cells gives rise to the mesoderm of one segment. ML and MR also give rise, by their two first divisions, to 504

the putative Primordial Germ Cells (pPGCs), and starting with the 8th round of division, to the 505

mesodermal posterior growth zone (MPGZ), which is part of the segment addition zone (SAZ). 506

23

Interestingly, the different cell types that originate from 4d (segmental mesoderm cells, posterior growth 507

zone cells, and germ cells) show distinct cell cycle characteristics. 508

509

510

511

Live imaging the 4d (M) lineage in P. dumerilii with a live-cell cycle reporter 512

Long-term imaging at single-cell resolution for lineage analyses has been technically challenging 513

in most spiralians including annelids. The micromere lineages (such as 4d) are small and very difficult to 514

label individually in non-clitellate annelids (i.e. non-clitellate Sedentaria, Errantia, and basal annelids) 515

(Irvine and Seaver, 2006), as opposed to the leech and the sludge worm Tubifex (both clitellates), which 516

have larger and accessible embryos that enable injecting single blastomeres to follow specific lineages 517

(Gline et al., 2011, 2009, Goto et al., 1999a, 1999b; Storey, 1989; Weisblat and Shankland, 1985). Despite 518

the difficulty of injections, cell lineage analyses have been carried out nevertheless. Some of the studies 519

used DiI injections for labeling specific blastomeres and analyzing the domains they gave rise to in fixed 520

samples, in which case there is no continuous observation by time-lapse imaging, thus no resolution at 521

individual cell-level (Ackermann et al., 2005; Irvine and Seaver, 2006). Other studies used bright-field 522

microscopy for live observations, but only earlier stages of development could be analyzed as cell sizes 523

become too small at later stages to confidently follow without any labeling, and/or the larvae start 524

moving due to ciliary band formation (Fischer and Arendt, 2013; Wilson, 1892 in another nereidid). As a 525

result, behavior of specific lineages with detailed cell cycling characteristics at single-cell resolution could 526

not be observed, and whether the lineages showed teloblastic activity in non-clitellate annelids could not 527

be directly observed. Taking advantage of injecting mRNAs coding for fluorescently-labeled proteins, 528

24

confocal imaging, and specimen immobilization, we were able to live-image the 4d lineage until late 529

trochophore stages (summarized in Figure 8) and analyzed its cell cycling patterns. Overcoming the long-530

standing challenge of immobilization of the trochophore larvae was particularly significant, because the 531

larvae start spinning from 15 hpf and later actively swim via several ciliary bands. Thus, live imaging for 532

extended time periods to follow cell lineages has not been possible. Our DMSO treatment method 533

(Materials and Methods) for cilia elimination can be potentially applied to other marine ciliated larvae as 534

well as freshwater ciliates (Balavoine, Özpolat, Duharcourt, unpublished results). 535

We generated a live-cell cycle reporter in P. dumerilii, using the gene Pdu-cdt1 (Figure 1). Cdt1 536

cell cycle component in mammal and zebrafish FUCCI (fluorescent ubiquitination-based cell cycle 537

indicator) is specific to the G1 phase (Sakaue-Sawano et al., 2008; Sugiyama et al., 2009). Our reporter 538

mVenus-Cdt1(aa1-147) is not specific to G1, but it was sufficient to observe cell cycling patterns, 539

therefore is a useful tool to study cell lineages and cell cycle dynamics during development of P. 540

dumerilii. To our knowledge, this is the first live-cell cycle reporter available in a spiralian model system. 541

Studies are in progress to develop reporters for other phases of cell cycle in P. dumerilii, as well as 542

transgenic lines containing one or more of the live cell cycle constructs. 543

Cell cycle patterns of the 4d lineage 544

4d is a conserved blastomere across many spiralian organisms with diverse body plans (Lambert, 545

2008; Lyons et al., 2012). However, only limited data are available on detailed 4d lineage/fate maps, cell 546

cycle ha a te isti s,à a dà ho à ellà leà egulatio à isà elatedà toà diffe e tà ellà t pesà dà p og a à547

generates (Bissen, 1997, 1995; Bissen and Weisblat, 1989; Chen and Bissen, 1997). In P. dumerilii, 4d 548

generates differentially-fated cell populations, which then display different cycling characteristics. pPGCs 549

(the first lineages to segregate from 4d) are mitotically quiescent and presumably do not divide until 550

later in development. However, lineages that arise right after pPGCs, the precursors of mesoderm, show 551

25

diverse cell cycling patterns (Figure 7). For example, we found early cycles of ML/R to be very short in P. 552

dumerilii (Figure 7-figure supplement 1A). Then cell cycle becomes slightly longer within mesodermal 553

lineages with relatively extended G2 phase. These findings in P. dumerilii are reminiscent of the leech, in 554

which, early ML/R divisions have short cell cycle durations due to absence of G1 and a very short G2 555

phase. Later cycles in the leech teloblasts become longer due to an increase in the G2 phase length, and 556

appearance of G1 phase (Bissen and Weisblat, 1989). 557

558

In stark contrast to the fast cycling lineages of the trunk mesoderm (3ml/r-8ml/r), precursors of 559

the MPGZ (8ML/R) are a slow cycling lineage which has relatively short G1 and long G2 phases (Figure 7-560

figure supplement 1C). These cells resemble stem cells in other organisms (such as hydra, mammals, 561

Drosophila, and C. elegans) because of their slow cycling as well as having short G1 and longer G2 phase 562

(Ables and Drummond-Barbosa, 2013; Buzgariu et al., 2014; Hsu et al., 2008; Roccio et al., 2013; Seidel 563

and Kimble, 2015) How will the cycling characteristics of the growth zone cells change with slow or fast 564

growth (for example, based on nutrient availability), or in response to injury (for example, during 565

segment addition in the regenerated tail, which grows faster) remains to be determined. 566

567

The first lineage segregated from 4d: putative primordial germ cells 568

In organisms where germline specification takes place early during development, the PGCs are 569

segregated before gastrulation (Extavour and Akam, 2003). This also appears to be the case in P. 570

dumerilii. We confirmed earlier observations (Fischer and Arendt, 2013; Rebscher et al., 2012, 2007) that 571

suggested that four putative PGCs (pPGCs; 1ml/r and 2ml/ à ells àa iseàf o àtheàfi stàt oàdi isio sàofà d’sà572

daughters, ML/R. Our and previous observations suggest these cells make only pPGCs (Ackermann et al., 573

26

2005; Rebscher et al., 2012, 2007), thus pPGCs are made separately from the segmental mesoderm 574

lineages which arise during the subsequent divisions. Here, we refer to these cells asà putati e àPGCsà575

because a continuous lineage tracing (or genetic lineage tracing) by following these progenitors into 576

adulthood when they would form the gametes has not been carried out. However, these cells fit the PGC 577

criteria used in other model systems: they express germline/multipotency program (GMP) genes (Juliano 578

et al., 2010) and Vasa protein, at day 4 of development they migrate anteriorly (Rebscher et al., 2012, 579

2007), they are mitotically quiescent (Figures 2, 3) but after migration they start proliferating to make 580

future gametes (Rebscher, 2014). However, considering that some annelids can regenerate their 581

germline (Herlant-Meewis, 1946; Özpolat et al., 2016; Özpolat and Bely, 2016; Stéphan-Dubois, 1964), 582

these PGCs set aside early in development in P. dumerilii may not be the only source of the germ cells in 583

this species. In particular, the ectodermal and mesodermal cells that constitute the SAZ of the juvenile 584

worm persistently express the GMP genes such as vasa, piwi or nanos (Gazave et al., 2013; Rebscher et 585

al., 2007). It is currently not known whether the GMP is shared between these apparently different sets 586

of cells because the program is crucial for multipotency, or whether the GMP expression in SAZ, and 587

growth zone cells in general, reflects a persistent ability of these cells to produce germline cells. A similar 588

situation has been described in planarians, where both the neoblast cells (the planarian stem cells) and 589

the germ cells express piwi (Handberg-Thorsager, 2008; Nakagawa et al., 2012). In planarians, the 590

neoblasts give rise to the germ cells (Morgan, 1902; Newmark et al., 2008; Solana, 2013). Only further 591

investigation involving genetic cell tracing or cell ablation will establish whether P. dumerilii can produce 592

(or regenerate) germline cells independently of the pPGCs discussed in this work. 593

4d lineage gives rise to the PGCs in other annelid species as well (Rebscher, 2014), but with some 594

differences: in contrast to P. dumerilii, which segregates its pPGCs during the first two divisions of 4d 595

(Figure 8B), PGCs arise much later from within the segmental mesoderm lineage in clitellates studied so 596

far, as opposed to being set aside during the earlier divisions of ML/R. For example, in the leech, blast 597

27

cells 16ml/r and 18 to 23ml/r (or sm10 and sm12-22 in the leech nomenclature) give rise to both 598

segmental mesoderm and the PGCs within the associated segments (Cho et al., 2014; Kang et al., 2002) 599

(Figure 8C .àCellsàp odu edà àtheàfi stàsi àdi isio sàofà d’sàdaughte sà o t i uteà ostl àtoàgutàe dode ,à600

to head and pharynx musculature, but not to the germline. Likewise, in Tubifex, PGCs originate from 601

within the progeny of 10ml/r and 11ml/r (Kitamura and Shimizu, 2000; Niwa et al., 2013; Oyama and 602

Shimizu, 2007). In other spiralians such as mollusks, pPGCs are segregated from the 4d lineage during 603

relatively early divisions (Rebscher, 2014), for example 2ml/r in the slipper snail (Lyons et al., 2012). 604

However, the lineage that generates the pPGCs in the slipper snail also gives rise to mesodermal cell 605

types. At the metazoan-wide level, it has been suggested that late epigenetic specification inside trunk 606

mesoderm is the ancestral process and that early specification, ofte àli kedàtoàtheàe iste eàofàaà ge à607

plas , has evolved independently multiple times in metazoans (Extavour and Akam, 2003). What role 608

these different segregation and specification patterns play, and whether they have implications in the 609

ability to regenerate germ cells at later stages of development and in adulthood in a given species need 610

to be further investigated. 611

612

Formation of segmental and growth zone mesoderm in P. dumerilii 613

We have established, for the first time, the early pattern of segmental mesoderm formation in P. 614

dumerilii (Figure 4, and Figure 8 for summary). We found that during embryogenesis, segmental 615

precursor cell segregation is similar to that in the leech or other clitellate embryogeneses. One important 616

diffe e eà isà that,à o l à fou à seg e tsà o eà a te io à pti à seg e tà a dà th eeà istle-bearing larval 617

trunk segments) are formed during P. dumerilii embryonic/larval development. This is in sharp contrast 618

to the 32 segments formed during leech embryogenesis and to the numerous segments formed in the 619

other non-leech clitellates (Anderson, 1973a; Balavoine, 2014; Goto et al., 1999b; Weisblat and 620

28

Shankland, 1985; Zackson, 1982). Nevertheless, as in clitellates, pairs of mesodermal precursors (called 621

primary blast cells (PBCs)) are sequentially produced in an anterior to posterior progression, and 622

correspond to a mesodermal segment of the larva. The segregation of the series of segmental PBCs in P. 623

dumerilii (from 5ml/r to 8ml/r) is preceded by two pairs of precursors (3ml/r and 4ml/r) that divided only 624

a few times during our analysis and probably contribute to the head mesoderm. These two pairs of 625

anterior head mesoderm precursors may be related to the six pairs of em precursors known in the leech 626

that contribute to endoderm and pharynx musculature (Figure 8C) (Gline et al., 2011). 627

We were not able to determine the precise tissue-type contribution of each mesodermal PBC 628

because our lineage analysis ends at the beginning of late trochophore stage, before cell differentiation 629

takes place. Additional imaging that extends further in development suggests that segmental blocks of 630

mesodermal cells contributed mostly to a single hemisegment (Figure 4-figure supplement 3, Video 8). 631

However, we cannot exclude the possibility that each clonal block will eventually contribute 632

complementary sets of different tissue types to two or more contiguous segments in the juveniles, as 633

they do in the leech. Future studies will also determine whether these precursors give rise to a few 634

neuronal cells (as in the clitellates), as well as other tissues such as the gut endoderm, and possibly 635

pygidium (the non-segmental posterior end) as it has been previously suggested (Starunov et al., 2015). 636

Another important difference between P. dumerilii and clitellate annelids resides in the size of 637

the mesodermal teloblasts and their evolution through embryogenesis. In the leech and other clitellates, 638

M teloblasts are huge cells that undergo a long series of very asymmetric divisions, budding off a series 639

ofà u hà s alle à esode alà PBCs,à efo eà the à e o eà e hausted à i à te sà ofà toplas i à olu eà640

after they produce the most posterior segments (Shimizu, 1982; Weisblat and Kuo, 2014). In P. dumerilii, 641

correlating with the small number of segments produced during embryogenesis, M teloblasts are born 642

ithàaà ediu àsizeàa dàappea àtoàgetà ui kl à e hausted àth oughàaàsho tàse iesàofàas et i àdi isio s 643

29

(Figure 6). It will be interesting to find out in the future, whether such M teloblasts exists in non-clitellate 644

annelids that form more segments during the embryonic/larval stage, such as Capitella which forms 13 645

segments (Balavoine, 2014; Seaver, 2016), and whether teloblast size evolved in proportion to the 646

number of larval segments (Figure 9, also see discussion below). 647

We also found that after the progenitors of the four mesodermal segments have been produced, 648

the remaining M teloblast cells in P. dumerilii gave rise to the precursors of mesodermal posterior 649

growth zone (MPGZ), which is the likely source of the mesodermal component of the Segment Addition 650

Zone (SAZ). Most annelids (but not leeches) continue growing from their posterior end by adding new 651

segments throughout their lives (Özpolat and Bely, 2016) (Figure 9). The new tissues that form new 652

segments originate from the SAZ which is a specific ring of small cells (presumably teloblast-like) within 653

the posterior growth zone, and has mesodermal and ectodermal components (Balavoine, 2014). The SAZ 654

expresses many Germline/Multipotency markers (Gazave et al., 2013; Juliano et al., 2010; Özpolat and 655

Bely, 2016, 2015). Much is still unknown about the exact embryonic origins of the SAZ in annelids. 656

Studies using bright-field microscopy, and DiI injections suggested that the mesodermal SAZ originates 657

from the 4d blastomeres (Ackermann et al., 2005; Anderson, 1973b; Rebscher et al., 2007), but which 658

cells within the 4d lineage make up the growth zone has not been identified to date. Our work 659

successfully identifies 8ML/MR as the origin of the mesodermal SAZ in the larva. A slowly cycling small 660

cluster of cells are made by 8ML/MR, located immediately anterior to the pPGCs, consistent with the 661

previous studies that described this region asàtheàVasa + à esode alàposte io àg o thàzo e à MPG)) 662

(Rebscher et al., 2012, 2007). Our imaging dataset is only until mid-/late-trochophore stage, thus we do 663

not know whether all or a subset of these cells will produce the mesodermal SAZ in the later larval stages 664

and juveniles. Thus, how these precursors go onto form the ring-like mesodermal SAZ and how they 665

contribute to new tissues in juvenile worms remain to be determined. 666

30

This study in P. dumerilii helps putting different types of development in the annelids into a 667

phylogenetic perspective (Figure 9). In the leeches, the totality of a species-specific number of segments 668

are made during embryogenesis through the activity of large embryonic teloblasts. In contrast, in most 669

marine annelids, the majority of segments are made during post-embryonic juvenile development, 670

through the activity of inconspicuous posterior stem cells whose characteristics are still largely unknown. 671

Earlier works in P. dumerilii (de Rosa et al, 2005; Gazave et al, 2013) have suggested that posterior stem 672

cells in the juvenile may have a teloblast-like activity. We establish in this article that the mesodermal 673

teloblasts which are very similar to the leech are indeed at play in the P. dumerilii embryo, contributing 674

to the formation of the three larval segments. We also show that these embryonic teloblasts are most 675

likely at the origin of the teloblast-like posterior stem cells of the juvenile (post-embryonic teloblasts), 676

thus possibly establishing a cellular continuity in segment formation in P. dumerilii. In addition, it will be 677

useful to determine if embryonic ectodermal teloblasts exist in P. dumerilii and whether they are at the 678

origin of the ectodermal posterior stem cells as well. It is very likely that the ancestral life cycle in 679

annelids involved these two phases of segment formation (embryonic/larval and post-embryonic) (Figure 680

9). The large teloblasts of Clitellate embryos would have evolved in the context of an acceleration of 681

development, most or all segment formation happening in the embryonic stage (Balavoine 2014). 682

Meanwhile, parallels among all spiralians remain to be determined. It is legitimate to wonder whether 683

embryonic teloblasts are present in mollusks having a form of mesodermal segmentation such as chitons 684

or aplacophorans (Scherholz et al., 2015), or whether post-embryonic teloblasts are present in other 685

posteriorly growing phyla, such as nemerteans. Future studies using high-resolution live imaging and 686

genetic lineage tracing methods on these exciting but understudied spiralian groups, as well as additional 687

studies in the annelids will help answer the open questions surrounding teloblasts and their involvement 688

in the evolution of different body plans. 689

690

31

691

692

693

MATERIALS AND METHODS 694

Key Resources Table 695

Reagent type (species) or

resource Designation

Source or reference

Identifiers Additional information

gene (Platynereis dumerilii)

Pdu-cdt1 NA GenBank No: MF614951

From the full 2337-bp coding sequence, a 1946-bp part including the start site, and ending before the stop codon was amplified

strain, strain background (P. dumerilii)

Wild Type Institut Jacques Monod cultures

recombinant DNA reagent

pCS2+mVenus-cdt1(aa1-147)

this paper GenBank No: MF614950

recombinant DNA reagent

pEXPTol2-H2A-mCherry

Other

Plasmid was not specifically produced for this paper, and has been published before. The source plasmids used for the constructions of this plasmid was provided by Caren Norden Lab (MPI-CBG, Dresden, Germany).

recombinant DNA reagent

pEXPTol2-EGFP-CAAX

Other

Plasmid was not specifically produced for this paper, and has been published before. The source plasmids used for the constructions of this plasmid was provided by Caren Norden Lab (MPI-CBG, Dresden, Germany).

antibody Acetylated-tubulin

Sigma T7451, RRID:AB_609894

(1:500)

sequence-based reagent

Pdu-cdt1- Forward Primer

this paper TTGTTTTCTTGTTGAGGTGGGATG

sequence-based reagent

Pdu-cdt1- Reverse Primer

this paper GGAGATGACGAGACAGGCAG

32

sequence-based reagent

pCS2+ vector backbone – Forward Primer

this paper CTAGAACTATAGTGTGTTGTATTACGT

sequence-based reagent

pCS2+ vector backbone – Reverse Primer

this paper AGAGGCCTTGAATTCGAATCG

commercial assay or kit

Gibson assembly master mix

NEB France E2611S

commercial assay or kit

Click-iT EdU Alexa Fluor 647 Imaging Kit

Life Technologies C10340

commercial assay or kit

SP6 mMessage kit

Ambion, ThermoFisher

AM1340

commercial assay or kit

MEGAclear kit

Ambion, ThermoFisher

AM1908 RNA elution option 2 with the modifications - see methods for details

commercial assay or kit

Proteinase K Ambion, ThermoFisher

AM2548 20 µg/µl final concentration

software, algorithm Bitplane Imaris (Version 8.2)

software, algorithm Onionizer

https://figshare.c

om/s/1c0dcd120

d13deb888ba

696

Animal Cultures and Staging 697

Platynereis dumerilii embryos were obtained from lab cultures at the Institut Jacques Monod. Cultures 698

are reared based on previous protocols (Dorresteijn et al., 1993) (available in English by Fischer and 699

Dorresteijn on www.platynereis.de). Embryos and larvae are staged after Fischer et al 2010. All staging is 700

made at 18°C unless otherwise stated. For live time-lapse imaging, the imaging temperatures were 701

typically higher (estimated 26-28°C). We provide a table with calculated developmental times and stages 702

corresponding to each time point (Table 1). 703

704

705

Gene Cloning 706

33

Pdu-cdt1 coding sequence was obtained from transcriptome databases by the Jékely and Arendt Labs 707

(jekely-lab.tuebingen.mpg.de/blast; 4dx.embl.de/platy). An amino acid alignment against several species 708

was carried out to confirm the sequence (MUSCLE alignment in GENEIOUS 6.1.8). From the full 2337-bp 709

coding sequence, a 1946-bp part including the start site, and ending before the stop codon was amplified 710

(Pdu-cdt1- Forward Primer: TTGTTTTCTTGTTGAGGTGGGATG; Reverse Primer: 711

GGAGATGACGAGACAGGCAG). The PCR fragment was gel purified, cloned into TOPO-TA pCRII plasmid, 712

and sequenced (GenBank No: MF614951). 713

714

Construction of Nuclear, Membrane, and Cell Cycle Constructs 715

Nuclear and membrane constructs: We created the expression constructs pEXPTol2-EGFP-CAAX and 716

pEXPTol2-H2A- Che à à LRà Re o i atio à Rea tio à a o di gà toà theà a ufa tu e ’sà i st u tio sà717

(Gateway Technology, Invitrogen). The individual plasmids used were the entry plasmids pENTR- p5E-718

CMV-SP6, pENTR-pME- EGFP-CAAX and pENTR-p3E-polyA or pENTR- p5E-CMV-SP6, pENTR-pME-H2A-719

mCherry, and pENTR-p3E-polyA and the destination plasmid pDestTol2pA2 (Kwan et al., 2007). 720

Cell cycle construct: Using an alignment of P. dumerilii Cdt1 (Pdu-cdt1) with vertebrate Cdt1 amino acid 721

sequences, we sub-cloned a C-terminal truncation of Cdt1 that is comparable to FUCCI constructs made 722

by others previously (Sakaue-Sawano et al., 2008; Sugiyama et al., 2009), including the PIP degron, 723

excluding Cdt1-Geminin interaction site and Cdt1-MCM2-7 binding site (Figure 1A). The truncated Pdu-724

cdt1 sequence was pla edà atà theà ’à e dà ofà Ve usà fluo es e tà p otei .à To ensure the stability of 725

expressed or injected mRNA for the fused mVenus-cdt1(aa1-147) construct, we added a KOZAK 726

se ue eài ediatel àupst ea àtoàtheàsta tà odo àofà Ve us.àTheà o se edà CáCC àKO)áKàsequence 727

from H. sapiens was used, because we could not determine a consensus sequence for P. dumerilii. 728

Fi all ,àse e alà“TOPà odo àse ue esà e eàaddedàtoàtheà ’àe dàofàtheà o st u t.àThisàKO)áK-mVenus-729

34

cdt1(aa1-147) sequence was synthetized by a provider (IDTDNA gBlocks, Belgium). The synthetized 730

fragment (200 ng) was resuspended in 20 µl nuclease-free water to a final concentration of 10 ng/µl, and 731

it was cloned using Gibson reaction into pCS2+ expression vector which was amplified via PCR (pCS2+ 732

vector backbone – Forward Primer: CTAGAACTATAGTGTGTTGTATTACGT; Reverse Primer: 733

AGAGGCCTTGAATTCGAATCG) (GenBank No: MF614950). For the Gibson reaction, Gibson assembly 734

master mix (NEB, France, E2611S) and NEB 5-alpha Competent E. coli (NEB C2987H) were used, following 735

theà a ufa tu e ’sàp oto ol.à 736

737

In vitro Transcription of mRNA For Microinjections 738

To express constructs in P. dumerilii embryos, we injected in vitro-transcribed mRNA into fertilized 739

oocytes. To prepare the mRNA, each vector was linearized, gel-purified, and used for the in vitro 740

transcription reaction (HistoneH2A-mCherry: BglII / SP6; EGFP-caax: ClaI / SP6; mVenus-cdt1(aa1-147): 741

Not-I / SP6). For in vitro transcription, SP6 mMessage kit from Ambion (AM1340) was used, and the 742

a ufa tu e ’sà p oto olà asà follo edà u tilà theà e dà ofà DNaseà stepà à µlà DNaseà à i sà atà ˚C .à Fo à743

purification of mRNA, MEGAclear kit from Ambion (AM1908) was used, following RNA elution option 2 744

with the following modifications: elution uffe à asà heatedà toà ˚C,à thisà a edà elutio à uffe à asà745

appliedà toà filte à a t idgeà o tai i gà RNá,à theà tu esà e eà keptà atà ˚Cà heatedà plateà fo à à i utesà746

before centrifuging for elution. 747

748

Embryo Preparation and Microinjections 749

For preparing the embryos for microinjections, fertilized embryos were dejellied and their cuticle was 750

softened via enzymatic degradation. To do this, fertilized embryos were transferred to a nylon 80µm-751

mesh at about 1 hour-post-fertilization (hpf), washed 10 times using filtered natural sea water to remove 752

35

the jelly secretion, treated with 20 µg/µl Proteinase K (Ambion, AM2548) in 30 mL seawater for 1 min, 753

then washed again 10 times, and transferred to a 6-well plate. 754

755

For microinjections, an agarose injection platform (made of 1.5% agarose in filtered natural sea water) 756

was prepared (Lauri et al., 2014) (Figure 10A-á’’’ .àI je tio àsolutio sà e eàp epa edàusi gàtheàfollo i gà757

final concentrations: mVenus-cdt1(aa1-147) – 25 ng/µl, HistoneH2A-mCherry – 75ng/µl, EGFP-caax – 758

75ng/µl, 0.07% phenol red (Sigma, P0290, 0.5% stock). Microinjections were carried out using an 759

inverted Zeiss Axiovert 135 scope, Eppendorf Femtojet 4i, and Eppendorf Transferman Micromanipulator 760

(Figure 10A). Micro injection needle (Sterile Eppendorf Femtotips II (930000043) was angled at about 15-761

20 degrees (Figure 10á’), and was connected to an Eppendorf universal capillary holder (920-00-739-2) 762

with adapter (no: 5) for Femtotips. Femtojet pressure settings varied depending on the needle opening 763

size, approximately in the following ranges: pi[hPa]=200-300 psi; pc[hPa]=30-300 psi; ti= 1-4 seconds (pi: 764

injection pressure, pc: back pressure, ti: injection time). Embryos were injected at 1-cell stage, 2-cell 765

stage (into CD half), and 4-cell stage (into D quadrant). Injected embryos were transferred back to the 6-766

ellàplateàa dà e eàkeptàatà ˚Cài u ato àu tilà ou ti g.à 767

768

Mounting Samples for Imaging 769

For all mounts, we used glass-bottom dishes (MatTek Corporation 35 mm dish, P35G-0-10-C), as this 770

method enabled keeping embryos and larvae in sea water which could be easily renewed if desired. As a 771

result, samples were healthy after mounting during extended periods of imaging, and even after imaging 772

they continued to grow in the agarose. Mounting was carried out on a warm plate to keep low-melting 773

agarose from solidifying while allowing enough time for rotating the samples to the desired position 774

using an eyelash tool (Figure 10-figure supplement 1). Low melting agarose (Invitrogen, 16520050) was 775

prepared as 1% solution in filtered natural sea water by microwaving, and 500 µl aliquots were kept in -776

36

˚Càu tilàuse.àTheseàtu esà e eàtha edài àaà ate à athàatà ˚Cà ightà efo eà ou ti g,àa dàkeptài àd à777

heati gà lo kàatà ˚Cào eàtha ed.à 778

779

Injected embryos and larvae can be checked for normal development based on a number of criteria 780

observed at different stages of development. For injected embryos to be imaged early (around 7-8 hpf at 781

18°C developmental time), we identified those injected samples without any visible deformations, clear 782

animal-vegetal polarization, and four oil droplets with typical appearance at this stage (Fischer et al., 783

2010). For samples to be imaged after ciliary band development (15 hpf or later, at 18°C developmental 784

time), we looked for normal ciliary bands, normal swimming behavior, chaetal sacs with developing 785

bristles, and larval eyes. We picked healthy samples based on these criteria, as well as the samples with 786

optimal fluorescence signal strength (which varies based on the amount of mRNA injected), and 787

appropriate angle for mesoderm imaging after mounting. 788

789

Depending on the developmental stage to be imaged, samples had to be mounted before or after the 790

ciliary band formation and the start of swimming movements. Mounting embryos or larvae with 791

functional cilia will result in the specimens rotating in their agarose spherical lodge. For stages after the 792

cilia form, immobilization requires deciliation at the time of mounting. We thus carried out imaging for 793

both pre-hatching (embryonic and pre-larval) and post-hatching (larval) stages using two different 794

techniques: 795

796

Mounting technique 1 - Mounting before ciliary band formation (Figure 10B): Cilia do not appear before 797

15 hpf (18°C). We observed that specimens mounted in low-melting agarose before 15 hpf (usually 798

around 5-6 hpf) usually will not move even after the ciliary bands develop, because the growing cilia are 799

37

impaired by the agarose covering. These specimens were simply placed onto the glass coverslip, excess 800

sea water was removed, and low-melting agarose was added to cover the sample. 801

802

Mounting technique 2 – Mounting after ciliary bands are formed (Figure 10C): If specimens are to be 803

imaged after the ciliary band formation, we found that incubating the larvae for 10 minutes in a small 804

volume of 10% DMSO / 90% seawater, followed by sudden flushing with fresh seawater caused clumping 805

and breakage of the cilia, thus immobilizing the specimens temporarily. Embryos and larvae can then be 806

mounted in low-melting agarose in the next ten minutes before the cilia start to grow back. These 807

DMSO-treated larvae recover completely and grow without abnormalities. Mounting injected samples 808

during larval stages may be desirable when a specific and precise orientation of the sample is required 809

starting at a specific stage. 810

811

EdU Cell Proliferation Analysis and Immunohistochemistry 812

EdU assay on injected samples: Embryos were injected at 1-, 2-, or 4-cell stage with HistoneH2A-mCherry 813

and mVenus-cdt1(aa1-147) solution mix, and were raised until 12 hpf. At 12 hpf, they were treated with 814

5 µM EdU in natural sea water for 3 minutes, rinsed quickly and fixed in 4% PFA for 1 hour. After fixation, 815

they were washed in 1X PBS a few times and were processed for the EdU reaction next day, using the 816

Click-iTà EdUà ále aà Fluo à à I agi gà Kità Lifeà Te h ologies,à C à kità follo i gà theà a ufa tu e ’sà817

protocol, except EdU reaction was carried out for 45 mins. 818

819

EdU assay and Immunohistochemistry: These experiments were done according to previously-published 820

protocols (Asadulina et al., 2012; Gazave et al., 2013). For EdU assays, embryos or larvae were incubated 821

in EdU (incubation conditions same as above) for the desired time period. Then EdU solution was washed 822

by transferring the samples several times through sea water. The samples were incubated to grow until 823

38

the desired stage and fixed in 4% PFA for 2 hours. The fixed samples were either processed just for EdU 824

reaction, or additional immunohistochemistry was carried out (primary antibody: Acetylated-tubulin 825

(1:500) (Sigma, T7451), secondary antibody: anti-mouse Alexa Fluor 488 (Molecular Probes, 44408S), and 826

DAPI (1:000 from stock concentration of 1mg/ml, overnight). EdU reaction was developed on the final 827

day of immunohistochemistry protocol. 828

829

Imaging 830

Immunohistochemistry and EdU reaction samples were mounted in 33% TDE (Asadulina et al., 2012) 831

using standard slides and coverslips, and were imaged using a Zeiss LSM710 confocal scope. For the 832

injected samples treated with EdU, endogenous mCherry and mVenus signals (expressed from the 833

i je tedà RNá à a eà dete ta leà i à ui k-fi ed à spe i e sà seeà a o e ,à thusà p e ludi gà theà eedà fo à834

antibody detection. 835

836

For live imaging, a Zeiss LSM780 confocal scope was used. In the samples coinjected with both nuclear 837

and membrane constructs, we took advantage of mVenus (YFP) being nuclear and EGFP being strictly 838

localized to the cell membrane, and imaged the two fluorophores together, using single laser (514) for 839

excitation. This allowed quick simultaneous acquisition of all three fluorophores decreasing phototoxicity 840

due to laser exposure. Fluorescent protein production from the injected mRNA started becoming 841

detectable around 5-6 hours after injections, thus imaging could not be carried out earlier. For the 842

samples imaged for cell lineage analysis (Samples A, B, C in Figure 5-figure supplement 1), 1.13 um step 843

size and 60-70 um total stack thickness was used, representing the proximal embryonic hemisphere 844

roughly. The 4D Z-stack for Sample A has been uploaded online as a dataset and can be accessed at 845

https://doi.org/10.5281/zenodo.1063531. 846

847

39

Image Analysis and Processing 848

EdU assays, and immunostaining stacks were analyzed and exported using Fiji release of ImageJ 849

(Schindelin et al., 2012). Nuclear signal counts were done manually in Fiji for Edu assays on injected 850

samples. For live imaging datasets, 3D-projected stacks were exported as videos also using Fiji, except 851

cell lineage tracking (see below). Time frame, and frequency of imaging is indicated for each video 852

specifically. Annotated videos were edited using Adobe Premiere Pro CC 2017, and video encoding was 853

adjusted to web-viewing using Hanbrake. 854

To better represent graphically an embryo that is spherical in shape, or parts of the embryo that are laid 855

out in a curved manner,à eà oteàaàFijiàs iptà alledà O io ize .à(The onionized images appear in Figure 856

1-figure supplement 2C, and Figure 4-figure supplement 3.) This is essentially a tool that allows in silico 857

flattening of the embryo at various tissue depths. Concretely, the script extracts the intersection of the 858

3D stacks and a series of near spherical surfaces located at increasing depth from the tissue surface. The 859

i te se tio sà a eà the à Dà p oje tedà asà ifà su essi eà o io à la e s à ofà theà sa pleà e eà flatte ed.à Theà860

sta kà ;à ;àz à isà epla edà àaà ;à ;àol àsta kà e eà ol à ep ese tào io à la e sàofà i easi gàdepth.àThisà861

allows flattening of ectodermalà tissuesàfo à lo à ol à aluesàa dà esode alà fo àhighe à ol values. The 862

FIJI script is available on FigShare.com (https://figshare.com/s/1c0dcd120d13deb888ba). 863

All the cell lineage and cell cycle analyses were done using Bitplane Imaris (Version 8.2). 3 different 864

samples from 2 independent injection batches were used for lineage analysis, and these are letter coded 865

throughout the manuscript: Sample A, Sample B, Sample C. In the Figures 3-7, data from Sample A is 866

shown (except in Figure 5B-B’’à“a pleàBàisàsho ,àa dàdataàf o à“a plesàBàa dàCàa eàp ese tedài àtheà867

figure supplements. 868

869

40

Imaris – Cell Lineage: Zeiss confocal stacks were uploaded to Imaris, and cell lineage tracing was done 870

manually using spots feature, volume and slice views in Imaris (as opposed to automated tracing, which 871

does not work well when cell and nuclei sizes change drastically, as is the case in P. dumerilii embryos). 872

Based on the manual curation and connection of spots across time frames, a cell lineage tree is 873

automatically created by Imaris. Different color codes were assigned to different lineages (such as blue, 874

orange, and pink). Descendants of particular blastomeres were selected using a different spot color 875

(yellow) for highlighting the clonal regions, and time-lapse videos including these highlighted spots were 876

exported using Imaris (Dataset DOI: 10.6084/m9.figshare.5613019). 877

878

Imaris – Cell Cycle: Spots (each sphere marking a cell in Imaris dataset) can be used for measuring 879

fluorescence intensity for the pixelsàthatàfallài toàthatàspot’sà olu e. We used this feature for obtaining a 880

quantitation of the cell cycle construct. To do this, spot size and position were manually adjusted so that 881

each spot was slightly smaller than the nucleus signal, and was positioned to be completely inside the 882

nucleus. Fluorescence intensity mean calculation for the mVenus channel was extracted for different cell 883

lineages. During mitosis (when cell nucleus disappeared) the spot was positioned roughly to contain 884

chromosomes. Criteria for choosing lineages to compare was as follows: we picked always the larger of 885

the two sister cells after a division, and if the division was equal, we picked the one more 886

posterior/closer to the Mesoblast. Cell cycle graphs were drawn in Excel and were modified in Adobe 887

Illustrator CC 2017. 888

889

ACKNOWLEDGEMENTS 890

We thank the Balavoine Lab Members, Vervoort Lab Members, and Ryan Null for helpful discussions and 891

feedback on the manuscript. We thank Solène “o gà fo à helpà ithà theà o io izatio à ode. We 892

acknowledge the Company of Biologists for the travel award to BDO, which enabled a trip to Florian 893

41

Rai le’sàla o ato à MFPL,àVie a,àáust ia àfo àlea i gài je tio s.àWeàtha kàKa i àVadi alaà Rai leàLa à894

for his assistance during this training period. We acknowledge the ImagoSeine facility (member of the 895

France BioImaging infrastructure supported by the French National Research Agency, ANR-10-INSB-04, 896

Investments of the future ) and its experienced staff for their assistance in imaging and image analysis. 897

We also thank Caren Norden Lab (MPI-CBG, Dresden, Germany) for providing the plasmids for the 898

construction of EGFP-caax and HistoneH2A-mCherry vectors. Finally, we thank the editors and reviewers 899

for their comments that helped improve this manuscript. 900

901

AUTHOR ORCIDS 902

B. Duygu Özpolat: orcid.org/0000-0002-1900-965X 903

Mette Handberg-Thorsager: orcid.org/0000-0002-3908-7233 904

Guillaume Balavoine: orcid.org/0000-0003-0880-1331 905

906

COMPETING INTERESTS 907

Authors declare no financial competing interests. 908

909

REFERENCES 910

911

Ables, E.T., Drummond-Barbosa, D., 2013. Cyclin E controls Drosophila female germline stem cell 912

maintenance independently of its role in proliferation by modulating responsiveness to niche 913

signals. Development 140, 530–40. doi:10.1242/dev.088583 914

Ackermann, C., Dorresteijn, A., Fischer, A., 2005. Clonal domains in postlarval Platynereis dumerilii 915

(Annelida: Polychaeta). J. Morphol. 266, 258–280. doi:10.1002/jmor.10375 916

Anderson, D.T., 1973a. Chapter 3: Oligochaetes and Leeches, in: Kerkut, G.A. (Ed.), Embryology and 917

Phylogeny in Annelids and Arthropods. Pergamon Press, New York, pp. 51–92. doi:10.1016/B978-0-918

08-017069-5.50008-1 919

42

Anderson, D.T., 1973b. Chapter2: Polychaetes, in: Kerkut, G.A. (Ed.), Embryology and Phylogeny in 920

Annelids and Arthropods. Pergamon Press, New York, pp. 7–50. doi:10.1016/B978-0-08-017069-921

5.50007-X 922

Arias, E.E., Walter, J.C., 2006. PCNA functions as a molecular platform to trigger Cdt1 destruction and 923

prevent re-replication. Nat. Cell Biol. 8, 84–90. doi:10.1038/ncb1346 924

Asadulina, A., Panzera, A., Verasztó, C., Liebig, C., Jékely, G., 2012. Whole-body gene expression pattern 925

registration in Platynereis larvae. Evodevo 3, 27. doi:10.1186/2041-9139-3-27 926

Backfisch, B., Kozin, V. V, Kirchmaier, S., Tessmar-Raible, K., Raible, F., 2014. Tools for gene-regulatory 927

analyses in the marine annelid Platynereis dumerilii. PLoS One 9, e93076. 928

doi:10.1371/journal.pone.0093076 929

Balavoine, G., 2014. Segment formation in Annelids: patterns, processes and evolution. Int. J. Dev. Biol. 930

58, 469–83. doi:10.1387/ijdb.140148gb 931

Ballabeni, A., Melixetian, M., Zamponi, R., Masiero, L., Marinoni, F., Helin, K., 2004. Human Geminin 932

promotes pre-RC formation and DNA replication by stabilizing CDT1 in mitosis. EMBO J. 23, 3122–933

3132. doi:10.1038/sj.emboj.7600314 934

Barker, N., 2014. Adult intestinal stem cells: critical drivers of epithelial homeostasis and regeneration. 935

Nat. Rev. Mol. Cell Biol. 15, 19–33. doi:10.1038/nrm3721 936

Bissen, S.T., 1997. Developmental control of cell division in leech embryos. BioEssays 19, 201–207. 937

doi:10.1002/bies.950190305 938

Bissen, S.T., 1995. Expression of the cell cycle control gene, cdc25, is constitutive in the segmental 939

founder cells but is cell-cycle-regulated in the micromeres of leech embryos. Development 121, 940

3035–3043. 941

Bissen, S.T., Weisblat, D.A., 1989. The durations and compositions of cell cycles in embryos of the leech, 942

Helobdella triserialis. Development 106, 105–118. 943

Brunet, T., Lauri, A., Arendt, D., 2015. Did the notochord evolve from an ancient axial muscle? The 944

axochord hypothesis. BioEssays 37, 836–850. doi:10.1002/bies.201500027 945

Buzgariu, W., Crescenzi, M., Galliot, B., 2014. Robust G2 pausing of adult stem cells in Hydra. 946

Differentiation 87, 83–99. doi:10.1016/j.diff.2014.03.001 947

Chen, Y., Bissen, S.T., 1997. Regulation of cyclin A mRNA in leech embryonic stem cells. Dev. Genes Evol. 948

206, 407–415. doi:10.1007/s004270050070 949

Cho, S.-J., Vallès, Y., Weisblat, D.A., 2014. Differential expression of conserved germ line markers and 950

delayed segregation of male and female primordial germ cells in a hermaphrodite, the leech 951

helobdella. Mol. Biol. Evol. 31, 341–54. doi:10.1093/molbev/mst201 952

Conklin, E.G., 1897. The embryology of crepidula,A contribution to the cell lineage and early 953

development of some marine gasteropods. J. Morphol. 13, 1–226. doi:10.1002/jmor.1050130102 954

Cooke, J., Zeeman, E.C., 1976. A clock and wavefront model for control of the number of repeated 955

structures during animal morphogenesis. J. Theor. Biol. 58, 455–76. 956

Devries, J., 1973. La formation et la destinee des feuillets embryonnaires chez le lombricien Eisenia 957

foetida. Arch. Anat. Microsc. 15–38. 958

Devries, J., 1972. Etude descriptive et experimentale du developpement embryonaire chez le Lombricien 959

Eisenia foetida. Universite de Bordeaux I. 960

Do esteij ,àá.W.C.,àO’G ad ,àB.,àFis he ,àá.,àPo het-Henneré, E., Boilly-Marer, Y., 1993. Molecular 961

spe ifi atio àofà ellàli esài àtheàe oàofàPlat e eisà á elida .àRou ’sàá h.àDe .àBiol.à ,à –962

269. doi:10.1007/BF00363215 963

Extavour, C.G., Akam, M., 2003. Mechanisms of germ cell specification across the metazoans: epigenesis 964

and preformation. Development 130, 5869–84. doi:10.1242/dev.00804 965

Fischer, A.H., Henrich, T., Arendt, D., 2010. The normal development of Platynereis dumerilii (Nereididae, 966

Annelida). Front. Zool. 7, 31. doi:10.1186/1742-9994-7-31 967

43

Fischer, A.H.L., Arendt, D., 2013. Mesoteloblast-like mesodermal stem cells in the polychaete annelid 968

Platynereis dumerilii (Nereididae). J. Exp. Zool. B. Mol. Dev. Evol. 320, 94–104. 969

doi:10.1002/jez.b.22486 970

Fujita, M., 2006. Cdt1 revisited: complex and tight regulation during the cell cycle and consequences of 971

deregulation in mammalian cells. Cell Div. 1, 22. doi:10.1186/1747-1028-1-22 972

Gazave, E., Béhague, J., Laplane, L., Guillou, A., Préau, L., Demilly, A., Balavoine, G., Vervoort, M., 2013. 973

Posterior elongation in the annelid Platynereis dumerilii involves stem cells molecularly related to 974

primordial germ cells. Dev. Biol. 382, 246–67. doi:10.1016/j.ydbio.2013.07.013 975

Gline, S.E., Kuo, D.-H., Stolfi, A., Weisblat, D.A., 2009. High resolution cell lineage tracing reveals 976

developmental variability in leech. Dev. Dyn. 238, 3139–51. doi:10.1002/dvdy.22158 977

Gline, S.E., Nakamoto, A., Cho, S.-J., Chi, C., Weisblat, D.A., 2011. Lineage analysis of micromere 4d, a 978

super-phylotypic cell for Lophotrochozoa, in the leech Helobdella and the sludgeworm Tubifex. 979

Dev. Biol. 353, 120–133. doi:10.1016/j.ydbio.2011.01.031 980

Goto, A., Kitamura, K., Arai, A., Shimizu, T., 1999a. Cell fate analysis of teloblasts in the Tubifex embryo 981

by intracellular injection of HRP. Dev. Growth Differ. 41, 703–13. 982

Goto, A., Kitamura, K., Shimizu, T., 1999b. Cell lineage analysis of pattern formation in the Tubifex 983

embryo. I. Segmentation in the mesoderm. Int J Dev Biol, 317–327. 984

Handberg-Thorsager, M., 2008. Neoblast and Germ Cell Dynamics in the Planarian Schmidtea 985

mediterranea: The role of the nanos and piwi gene families (Dinámica celular de los neoblastos y la 986

línea germinal en la planaria Schmidtea mediterranea: El papel de las familias génicas nanos y p. 987

Universitat de Barcelona. 988

Havens, C.G., Walter, J.C., 2009. Docking of a Specialized PIP Box onto Chromatin-Bound PCNA Creates a 989

Degron for the Ubiquitin Ligase CRL4Cdt2. Mol. Cell 35, 93–104. doi:10.1016/j.molcel.2009.05.012 990

Henry, J.Q., 2014. Spiralian model systems. Int. J. Dev. Biol. 58, 389–401. doi:10.1387/ijdb.140127jh 991

Herlant-Meewis, H., 1946. Co t i utio à àl’EtudeàdeàlaàR g atio à hezàlesàOligo h tesà- 992

Reconstitution du germen chez Lumbricillus lineatus (Enchytraeides) (deuxième partie). Arch. Biol. 993

Paris 57, 197–306. 994

Hsu, H.-J., LaFever, L., Drummond-Barbosa, D., 2008. Diet controls normal and tumorous germline stem 995

cells via insulin-dependent and -independent mechanisms in Drosophila. Dev. Biol. 313, 700–712. 996

doi:10.1016/j.ydbio.2007.11.006 997

Irvine, S.Q., Seaver, E.C., 2006. Early annelid development, a molecular perspective., in: Rouse, G.W., 998

Pleijel, F. (Eds.), Reproductive Biology and Phylogeny of Annelida. Science Publishers, Enfield, NH, p. 999

93–140 (page 99). 1000

Juliano, C.E., Swartz, S.Z., Wessel, G.M., 2010. A conserved germline multipotency program. 1001

Development 137, 4113–26. doi:10.1242/dev.047969 1002

Kang, D., Pilon, M., Weisblat, D. a, 2002. Maternal and zygotic expression of a nanos-class gene in the 1003

leech Helobdella robusta: primordial germ cells arise from segmental mesoderm. Dev. Biol. 245, 1004

28–41. doi:10.1006/dbio.2002.0615 1005

Kitamura, K., Shimizu, T., 2000. Analyses of Segment-Specific Expression of Alkaline Phosphatase Activity 1006

in the Mesoderm of the Oligochaete Annelid Tubifex: Implications for Specification of Segmental 1007

Identity. Dev. Biol. 219, 214–223. doi:10.1006/dbio.1999.9602 1008

Kretzschmar, K., Watt, F., 2012. Lineage Tracing. Cell 148, 33–45. doi:10.1016/j.cell.2012.01.002 1009

Kwan, K.M., Fujimoto, E., Grabher, C., Mangum, B.D., Hardy, M.E., Campbell, D.S., Parant, J.M., Yost, H.J., 1010

Kanki, J.P., Chien, C.-B., 2007. The Tol2kit: A multisite gateway-based construction kit forTol2 1011

transposon transgenesis constructs. Dev. Dyn. 236, 3088–3099. doi:10.1002/dvdy.21343 1012

Lambert, J.D., 2008. Mesoderm in spiralians: the organizer and the 4d cell. J. Exp. Zool. Part B Mol. Dev. 1013

Evol. 310B, 15–23. doi:10.1002/jez.b.21176 1014

Lauri, a., Brunet, T., Handberg-Thorsager, M., Fischer, a. H.L., Simakov, O., Steinmetz, P.R.H., Tomer, R., 1015

44

Keller, P.J., Arendt, D., 2014. Development of the annelid axochord: Insights into notochord 1016

evolution. Science (80-. ). doi:10.1126/science.1253396 1017

Liu, E., Li, X., Yan, F., Zhao, Q., Wu, X., 2004. Cyclin-dependent Kinases Phosphorylate Human Cdt1 and 1018

Induce Its Degradation. J. Biol. Chem. 279, 17283–17288. doi:10.1074/jbc.C300549200 1019

Lyons, D.C., Perry, K.J., Lesoway, M.P., Henry, J.Q., 2012. Cleavage pattern and fate map of the 1020

mesentoblast, 4d, in the gastropod Crepidula: a hallmark of spiralian development. Evodevo 3, 21. 1021

doi:10.1186/2041-9139-3-21 1022

Meyer, N.P., Boyle, M.J., Martindale, M.Q., Seaver, E.C., 2010. A comprehensive fate map by intracellular 1023

injection of identified blastomeres in the marine polychaete Capitella teleta. Evodevo 1, 8. 1024

doi:10.1186/2041-9139-1-8 1025

Monier, B., Pélissier-Monier, A., Sanson, B., 2011. Establishment and maintenance of compartmental 1026

boundaries: role of contractile actomyosin barriers. Cell. Mol. Life Sci. 68, 1897–1910. 1027

doi:10.1007/s00018-011-0668-8 1028

Morgan, T.H., 1902. Growth and regeneration in Planaria lugubris. Arch. Entw. Mech. Org. 13, 179–212. 1029

Nakagawa, H., Ishizu, H., Hasegawa, R., Kobayashi, K., Matsumoto, M., 2012. Drpiwi-1 is essential for 1030

germline cell formation during sexualization of the planarian Dugesia ryukyuensis. Dev. Biol. 361, 1031

167–76. doi:10.1016/j.ydbio.2011.10.014 1032

Newmark, P.A., Wang, Y., Chong, T., 2008. Germ cell specification and regeneration in planarians. Cold 1033

Spring Harb. Symp. Quant. Biol. 73, 573–81. doi:10.1101/sqb.2008.73.022 1034

Nishitani, H., Sugimoto, N., Roukos, V., Nakanishi, Y., Saijo, M., Obuse, C., Tsurimoto, T., Nakayama, K.I., 1035

Nakayama, K., Fujita, M., Lygerou, Z., Nishimoto, T., 2006. Two E3 ubiquitin ligases, SCF-Skp2 and 1036

DDB1-Cul4, target human Cdt1 for proteolysis. EMBO J. 25, 1126–36. 1037

doi:10.1038/sj.emboj.7601002 1038

Niwa, N., Akimoto-Kato, A., Sakuma, M., Kuraku, S., Hayashi, S., 2013. Homeogenetic inductive 1039

mechanism of segmentation in polychaete tail regeneration. Dev. Biol. 381, 460–470. 1040

doi:10.1016/j.ydbio.2013.04.010 1041

Oyama, A., Shimizu, T., 2007. Transient occurrence of vasa-expressing cells in nongenital segments 1042

during embryonic development in the oligochaete annelid Tubifex tubifex. Dev. Genes Evol. 217, 1043

675–90. doi:10.1007/s00427-007-0180-1 1044

Özpolat, B.D., Bely, A.E., 2016. Developmental and molecular biology of annelid regeneration: a 1045

comparative review of recent studies. Curr. Opin. Genet. Dev. doi:10.1016/j.gde.2016.07.010 1046

Özpolat, B.D., Bely, A.E., 2015. Gonad establishment during asexual reproduction in the annelid Pristina 1047

leidyi. Dev. Biol. 405, 123–36. doi:10.1016/j.ydbio.2015.06.001 1048

Özpolat, B.D., Sloane, E.S., Zattara, E.E., Bely, A.E., 2016. Plasticity and regeneration of gonads in the 1049

annelid Pristina leidyi. Evodevo 7, 22. doi:10.1186/s13227-016-0059-1 1050

Penners, A., 1924. Die Entwicklung des Keimstreifs und die Organbildung bei Tubifex rivulorum Lam. 1051

Zool. Jb. Abt. Anat. Ontog. 251–308. 1052

Raible, F., Arendt, D., 2004. Metazoan Evolution: Some Animals Are More Equal than Others. Curr. Biol. 1053

14, 106–108. doi:10.1016/S0960-9822(04)00030-2 1054

Raible, F., Tessmar-Raible, K., Osoegawa, K., Wincker, P., Jubin, C., Balavoine, G., Ferrier, D., Benes, V., de 1055

Jong, P., Weissenbach, J., Bork, P., Arendt, D., 2005. Vertebrate-Type Intron-Rich Genes in the 1056

Marine Annelid Platynereis dumerilii. Science (80-. ). 310, 1325–1326. doi:10.1126/science.1119089 1057

Rebscher, N., 2014. Establishing the germline in spiralian embyos. Int. J. Dev. Biol. 58, 403–411. 1058

doi:10.1387/ijdb.140125nr 1059

Rebscher, N., Lidke, A.K., Ackermann, C.F., 2012. Hidden in the crowd: primordial germ cells and somatic 1060

stem cells in the mesodermal posterior growth zone of the polychaete Platynereis dumerillii are 1061

two distinct cell populations. Evodevo 3, 9. doi:10.1186/2041-9139-3-9 1062

Rebscher, N., Zelada-gonzález, F., Banisch, T.U., Raible, F., Arendt, D., 2007. Vasa unveils a common 1063

45

origin of germ cells and of somatic stem cells from the posterior growth zone in the polychaete 1064

Platynereis dumerilii. Dev. Biol. 306, 599–611. doi:10.1016/j.ydbio.2007.03.521 1065

Roccio, M., Schmitter, D., Knobloch, M., Okawa, Y., Sage, D., Lutolf, M.P., 2013. Predicting stem cell fate 1066

changes by differential cell cycle progression patterns. Development 140, 459–70. 1067

doi:10.1242/dev.086215 1068

Sakaue-Sawano, A., Kurokawa, H., Morimura, T., Hanyu, A., Hama, H., Osawa, H., Kashiwagi, S., Fukami, 1069

K., Miyata, T., Miyoshi, H., Imamura, T., Ogawa, M., Masai, H., Miyawaki, A., 2008. Visualizing 1070

spatiotemporal dynamics of multicellular cell-cycle progression. Cell 132, 487–98. 1071

doi:10.1016/j.cell.2007.12.033 1072

Saudemont, A., Dray, N., Hudry, B., Le Gouar, M., Vervoort, M., Balavoine, G., 2008. Complementary 1073

striped expression patterns of NK homeobox genes during segment formation in the annelid 1074

Platynereis. Dev. Biol. 317, 430–43. doi:10.1016/j.ydbio.2008.02.013 1075

Scherholz, M., Redl, E., Wollesen, T., Todt, C., Wanninger, A., 2015. From complex to simple: myogenesis 1076

in an aplacophoran mollusk reveals key traits in aculiferan evolution. BMC Evol. Biol. 15, 201. 1077

doi:10.1186/s12862-015-0467-1 1078

Schindelin, J., Arganda-Carreras, I., Frise, E., Kaynig, V., Longair, M., Pietzsch, T., Preibisch, S., Rueden, C., 1079

Saalfeld, S., Schmid, B., Tinevez, J.-Y., White, D.J., Hartenstein, V., Eliceiri, K., Tomancak, P., 1080

Cardona, A., 2012. Fiji: an open-source platform for biological-image analysis. Nat. Methods 9, 676–1081

682. doi:10.1038/nmeth.2019 1082

Seaver, E.C., 2016. Annelid models I: Capitella teleta. Curr. Opin. Genet. Dev. 39, 35–41. 1083

doi:10.1016/j.gde.2016.05.025 1084

Seaver, E.C., 2014. Variation in spiralian development: insights from polychaetes. Int. J. Dev. Biol. 58, 1085

457–467. doi:10.1387/ijdb.140154es 1086

Seaver, E.C., Paulson, D.A., Irvine, S.Q., Martindale, M.Q., 2001. The spatial and temporal expression of 1087

Ch-en, the engrailed gene in the polychaete Chaetopterus, does not support a role in body axis 1088

segmentation. Dev. Biol. 236, 195–209. doi:10.1006/dbio.2001.0309 1089

Seidel, H.S., Kimble, J., 2015. Cell-cycle quiescence maintains Caenorhabditis elegans germline stem cells 1090

independent of GLP-1/Notch. Elife 4. doi:10.7554/eLife.10832 1091

Shimizu, T., 1982. Development in the freshwater oligochaete Tubifex. Dev. Biol. Freshw. Invertebr. 286–1092

316. 1093

Shimizu, T., Nakamoto, A., 2014. Developmental significance of D quadrant micromeres 2d and 4d in the 1094

oligochaete annelid Tubifex tubifex. Int. J. Dev. Biol. 58, 445–456. doi:10.1387/ijdb.140102tsh 1095

Shiomi, Y., Suenaga, N., Tanaka, M., Hayashi, A., Nishitani, H., 2014. Imaging Analysis of Cell Cycle-1096

Dependent Degradation of Cdt1 in Mammalian Cells, in: Methods in Molecular Biology (Clifton, 1097

N.J.). pp. 357–365. doi:10.1007/978-1-4939-0888-2_18 1098

Smith, C.M., Weisblat, D.A., 1994. Micromere fate maps in leech embryos: lineage-specific differences in 1099

rates of cell proliferation. Development 120. 1100

Solana, J., 2013. Closing the circle of germline and stem cells: the Primordial Stem Cell hypothesis. 1101

Evodevo 4, 2. doi:10.1186/2041-9139-4-2 1102

Starunov, V. V, Dray, N., Belikova, E. V, Kerner, P., Vervoort, M., Balavoine, G., 2015. A metameric origin 1103

for the annelid pygidium? BMC Evol. Biol. 15, 25. doi:10.1186/s12862-015-0299-z 1104

Steinmetz, P.R.H., Kostyuchenko, R.P., Fischer, A., Arendt, D., 2011. The segmental pattern of otx, gbx, 1105

and Hox genes in the annelid Platynereis dumerilii. Evol. Dev. 13, 72–9. doi:10.1111/j.1525-1106

142X.2010.00457.x 1107

Stéphan-Du ois,àF.,à .àLaàLi g eàge i aleàdesàTu ella i sàetàdesàá lidesàda sàl’ olutio à1108

o aleàetàlaà g e atio ,ài :àWolff,àE.à Ed. ,àL’O igi eàdeàLaàLi g eàGe i aleàChezàLesàVe t sà1109

et Chez QuelquesàG oupsàd’I e t s.àHermann, Paris, pp. 115–136. 1110

Storey, K., 1989. The effects of ectoteloblast ablation in the earthworm. Development 107, 533–545. 1111

46

Struck, T.H., Paul, C., Hill, N., Hartmann, S., Hösel, C., Kube, M., Lieb, B., Meyer, A., Tiedemann, R., 1112

Purschke, G., Bleidorn, C., 2011. Phylogenomic analyses unravel annelid evolution. Nature 471, 95–1113

8. doi:10.1038/nature09864 1114

Sugiyama, M., Sakaue-Sawano, A., Iimura, T., Fukami, K., Kitaguchi, T., Kawakami, K., Okamoto, H., 1115

Higashijima, S., Miyawaki, A., 2009. Illuminating cell-cycle progression in the developing zebrafish 1116

embryo. Proc. Natl. Acad. Sci. U. S. A. 106, 20812–7. doi:10.1073/pnas.0906464106 1117

Swartz, S.Z., Chan, X.Y., Lambert, J.D., 2008. Localization of Vasa mRNA during early cleavage of the snail 1118

Ilyanassa. Dev. Genes Evol. 218, 107–13. doi:10.1007/s00427-008-0203-6 1119

Thamm, K., Seaver, E.C., 2008. Notch signaling during larval and juvenile development in the polychaete 1120

annelid Capitella sp. I. Dev. Biol. 320, 304–318. doi:10.1016/j.ydbio.2008.04.015 1121

Ul a ,àV.,àMaška,àM.,àMag usso ,àK.E.G.,àRo e e ge ,àO.,àHau old,àC.,àHa de ,àN.,àMatula,àP.,àMatula,à1122

P., Svoboda, D., Radojevic, M., Smal, I., Rohr, K., Jaldén, J., Blau, H.M., Dzyubachyk, O., Lelieveldt, 1123

B., Xiao, P., Li, Y., Cho, S.-Y., Dufour, A.C., Olivo-Marin, J.-C., Reyes-Aldasoro, C.C., Solis-Lemus, J.A., 1124

Bensch, R., Brox, T., Stegmaier, J., Mikut, R., Wolf, S., Hamprecht, F.A., Esteves, T., Quelhas, P., 1125

Demirel, Ö., Malmström, L., Jug, F., Tomancak, P., Meijering, E., Muñoz-Barrutia, A., Kozubek, M., 1126

Ortiz-de-Solorzano, C., 2017. An objective comparison of cell-tracking algorithms. Nat. Methods. 1127

doi:10.1038/nmeth.4473 1128

Weigert, A., Bleidorn, C., 2016. Current status of annelid phylogeny. Org. Divers. Evol. 1129

doi:10.1007/s13127-016-0265-7 1130

Weisblat, D.A., Kuo, D.-H., 2014. Developmental biology of the leech Helobdella. Int. J. Dev. Biol. 58, 1131

429–443. doi:10.1387/ijdb.140132dw 1132

Weisblat, D.A., Shankland, M., 1985. Cell lineage and segmentation in the leech. Philos. Trans. R. Soc. 1133

Lond. B. Biol. Sci. 312, 39–56. 1134

Wilson, E.B., 1892. The cell-lineage of Nereis.A contribution to the cytogeny of the annelid body. J. 1135

Morphol. 6, 361–480. doi:10.1002/jmor.1050060301 1136

Yasugi, T., Nishimura, T., 2016. Temporal regulation of the generation of neuronal diversity in Drosophila. 1137

Dev. Growth Differ. 58, 73–87. doi:10.1111/dgd.12245 1138

Zackson, S.L., 1982. Cell clones and segmentation in leech development. Cell 31, 761–770. 1139

doi:10.1016/0092-8674(82)90330-0 1140

Zielke, N., Edgar, B. a., 2015. FUCCI sensors: powerful new tools for analysis of cell proliferation. Wiley 1141

Interdiscip. Rev. Dev. Biol. n/a-n/a. doi:10.1002/wdev.189 1142

) za ý,àJ.,àŘíha,àP.,àPi lek,àL.,àJa ouško e ,àJ.,à .àPh logeny of Annelida (Lophotrochozoa): total-1143

evidence analysis of morphology and six genes. BMC Evol. Biol. 9, 189. doi:10.1186/1471-2148-9-1144

189 1145

1146

1147

1148

FIGURE LEGENDS 1149

Figure 1 - FUCCI cell cycle reporter for Platynereis dumerilii. A) P. dumerilii Cdt1 complete amino acid 1150

sequence above, showing the domains including PIP destruction box (N terminal). Below: a 1151

representation of the fused cell cycle construct containing mVenus fluorescent protein and a truncated 1152

47

part of Pdu-Cdt1. mVenus is a yellow fluorescent protein but is shown in green color for convenience 1153

throughout the manuscript. B) Frames (every 4 minutes) from time-lapse (Video 1.1 and Video 1.2) 1154

sho i gàdi isio àa dà ellà li gàofàa àe o i à ellà la eledàasà a àa dàitsàdaughte sà a àa dà .à1155

Fluorescent channels are shown separately for HistoneH2A-mCherry (red - nuclear) and mVenus-1156

Cdt1(aa1-147) (green - cycling nuclear). The sister cells show significantly different cycling patterns, as 1157

indicated by bars marking the different cell cycle phases (M/G1/S/G2). C) Example of an embryo which 1158

was injected with mRNA at 2-cell stage, EdU-incubated for 3 minutes at 12hpf, and fixed immediately. 1159

The part below the dashed line is the injected side. All channels are merged. Red: Histone2A-mCherry; 1160

Green: mVenus-Cdt1(aa1-147); Magenta: EdU. D-D’’’ à E a plesà ofà ellsà thatà displa edà diffe e tà olo à1161

combinations, and counts and percentages of cells based on data from 6 embryos (2 independent 1162

experiments). E) Cartoon summarizing the Pdu-FUCCI fluorescence patterns based on the time-lapse 1163

observations in B and EdU observations in D-D’’’. 1164

1165

Figure 1 – Figure Supplement 1 - Cdt1 amino acid alignments of destruction box sequences from 1166

different species. Cdt1 amino acid alignment showing the region containing the PIP BOX (degron motif). 1167

The conserved motif for this degron in the human sequence is Q-x-x-[I/L/M/V]-T-D-[F/Y]-[F/Y]-x-x-x-1168

[R/K]. We found a highly-conserved region in the P. dumerilii sequence (aa3-14) corresponding to the PIP 1169

degron. For the cell cycle construct, a truncated (aa1-147) piece of Pdu-cdt1 sequence containing the PIP 1170

Box was used. 1171

Figure 1 – Figure Supplement 2 – Multiciliated cells that exit the cell cycle can be visualized using Pdu-1172

FUCCI. A) Cartoon of 48 hpf larva (ventral view), anterior up, slightly rotated up from the posterior end. 1173

B-C’’ à Ve us-Cdt1(aa1-147) (nuclear) and EGFP-caax (membrane) live-imaged together via single green 1174

channel. The orientation of sample in B is similar to cartoon in A. Strong mVenus signal persists in the 1175

48

nuclei of prototroch (arrowheads) (see Video 2.1 for earlier time points of the prototroch cells). C) 1176

Ectoderm reconstruction of a larva injected with mVenus-Cdt1(aa1-147), HistoneH2A-mCherry, and 1177

EGFP-caax (see Methods and Video . à fo àtheà o io izatio àp o ess .àTheàstageà o espo dsàtoà late-1178

t o hopho eàstage.àPa at o hà ellsàsho àst o gà Ve usàsig alà a o headsài àC’àa dàC’’ .àD àCa too àofà1179

larva at 72 hpf, ventral view. Larvae were treated with EdU from 36 hpf until 72 hpf, and processed for 1180

immunohistochemistry (Acetylated-tubulin for cilia) and EdU reaction. Prototroch cells (E-E’’ ,à and 1181

pa at o hà ellsà F,àG’ àdoà otài o po ateàEdUàafte àthe àa eà o à o fi i gàthe àha eàexited cell cycle 1182

and do not proceed to S phase. Stars indicate the troch cell nuclei. 1183

Figure 2 – pPGCs do not incorporate EdU after they are born and retain mVenus signal. pPGCs 1184

(arrowheads) incorporate and retain the EdU signal (red) when treated during the time they are born (5-1185

7 hpf), and can be easily detected at 24 hpf (A) or 48 hpf (B) due to retention of EdU. However, they do 1186

not incorporate EdU, if treated after they are born (C and D). Live samples that were injected with Pdu-1187

FUCCI (HistoneH2A-mCherry, mVenus-cdt1(aa1-147)) and EGFP-caax and imaged at 24 hpf (E) and 48 hpf 1188

(F) show that pPGCs have the nuclear mVenus signal (green), also suggesting that they have stopped 1189

cycling (see also Figure 3, Figure 3-figure supplement 1, and Figure 7B). Green bars in the timeline 1190

schemas show the time period in which pPGCs are born. Red arrows in A-D show the period of EdU 1191

treatment. At least n=6 samples were imaged for each EdU assay, and a representative sample is shown. 1192

1193

Figure 3 – Time-lapse series of pPGC migration. Embryo (Sample A) was injected with Pdu-FUCCI and 1194

EGFP-caax in the D quadrant and live-imaged with high time resolution (every 10 minutes). Time-lapse 1195

imaging started at 8 hpf and continued for 12 hours. Selected snapshots (elapsed time indicated on the 1196

lower-left corner) shown in the figure. A-H) Z-projections from subsets of stacks showing the sample 1197

from a postero-late alào ie tatio ,àdo salàup,à e t alàdo .àpPGCsàa eài di atedà ithàa o heads.àá’-H’ à1198

49

Same dataset analyzed in Imaris as Z-projection of full stacks, and rotated to a perfect posterior view in 1199

theàsoft a e’sà Dà ie e .àállàfluo es e tàsig alàisàsho ài àg e à olo à setàa tifi iall ài àI a is àa dàpPGCsà1200

(pink spots) were traced using the spots feature (Video . .à á’’-H’’ à Lateral view of the same Imaris 1201

dataset, posterior to the right, dorsal up (Video 4.2). Yellow arrow shows the approximate path of pPGCs 1202

asàthe à ig ate.àNoteàthatàthe àt a elào àtheàsu fa eà u tilàE’’ ,àthe àsta tàassu i gàa ài te alàpositio à1203

(indicated with the dashed yellow line) with the onset of epiboly. For the original video showing all time 1204

points and full-projection of stacks see Video 3.1. For close-ups see Figure 3-figure supplement 1. (In the 1205

rest of the manuscript, this sample is analyzed for other developmental processes and is referred to as 1206

Sample A. Information from additional samples (Samples B and C) is provided in figure supplements.) 1207

1208

Figure 3 – Figure Supplement 1 – Close-ups of time-lapse video of pPGCs. Close-ups of pPGCs for the 1209

frames from Video 3.1 (Sample A) shown in Figure 3A-H are shown with merged fluorescent channels 1210

and green fluorescent channel separately. Note that pPGCs (arrowheads) retain the nuclear mVenus 1211

signal throughout. Towards the end of the time-lapse, signal in pPGCs becomes weak and harder to see 1212

in these 3D-projections, however measurements of nuclear signal show mVenus in pPGCs does not cycle 1213

(Figure 7B). 1214

1215

Figure 3 – Figure Supplement 2 – The first two divisions of ML and MR. Frames showing the first two 1216

divisions of ML and MR, from a different sample injected with Pdu-FUCCI and EGFP-caax. Imaging started 1217

at 7:20hpf, done every 6 minutes, allowing to observe the first two divisions. These divisions give rise to 1218

the four pPGCs (dashed circles). pPGCs have both HistoneH2A-mCherry (red) and mVenus-Cdt1(aa1-147) 1219

(green) right after cell divisions are completed, indicated by the orange-green nucleus color. 1220

50

1221

Figure 4 – Primary blast cells give rise to clonal bocks of mesoderm. The same sample as shown in 1222

Figure 3 (Sample A) was analyzed for mesodermal lineages in Bitplane Imaris. 4-cell stage embryo 1223

injected with Pdu-FUCCI (HistoneH2A-mCherry, mVenus-cdt1(aa1-147)) and EGFP-caax into D quadrant 1224

was live-imaged starting at 8 hpf every 10 minutes for 12 hours, from a posterior-lateral angle which 1225

enabled the tracing and visualization of the left mesoblast (ML=4d1) lineage. Schema in A shows 4d 1226

lineage: ML (blue), MR (orange), pPGCs (pink), and highlighted subsets (yellow) are all color-coded 1227

(summarized in C) throughout this and following figures. Note that only a limited number of cells were 1228

traced for MR. Panel B (zoomed-i àf o àB’) shows the starting frame of the time-lapse, and the position 1229

of ML and MR, with all three fluorescent constructs (see Video 3.1 for the original time-lapse with 1230

annotations, and Video 6.1 with spots for lineage tracing of ML). I àtheàpa elsàD’- G’’,àtheàlastàti eàpoi tà1231

from this time-lapse (Sample A) is shown with EGFP-caax and mVenus-Cdt1(aa1-147) only (both in 1232

green), but HistoneH2A-mCherry (red) was removed for easier observation of the colored spots. 5ml 1233

lineage makes the mesoderm of the left half of the anterior-most cryptic hemisegment (D-D’’’ ,à là1234

makes the 1st larval hemisegment (E-E’’’ ,à làthe 2nd larval hemisegment (F-F’’’ ,àa dà làtheà dàla alà1235

hemisegment (G-G’’’ .à Pa elsàD-Gàa eà theà sa eào ie tatio àasàD’-G’à utào l à sho i gà theà spots.àD’-G’à1236

sho àposte io ào ie tatio ,àa dàD’’-G’’àsho àlate alào ie tatio . The lineages highlighted in yellow in D-1237

G’’àa eàalsoàsho ài à ello ài àtheàli eageàt eesà D’’’-G’’’ .à“eeàVideo 7 for an animated version of these 1238

lineages highlighted. Videos 6.2-6.13 show the full time-lapse of each lineage from posterior and lateral 1239

orientations, as well as 360-degree rotation. 1240

1241

1242

51

Figure 4 – Figure Supplement 1 – Progeny of each blast cell at an intermediate time point. For a 1243

comparison to the final time point (12h 00min) shown in Figure 4. Clonal domains (in yellow) of cells 1244

each primary blast cell gives rise to are already obvious at time point 8h 10min. Videos 6.1 through 6.15 1245

show all time points, each highlighting the progeny of a primary blast lineage separately. Blue: ML 1246

lineage, Yellow: subset highlighted within ML lineage, Pink: pPGCs, Orange: subset of MR lineage for 1247

reference. 1248

1249

Figure 4 – Figure Supplement 2 – Mesodermal clonal clusters align with chaetal sacs specific to each 1250

hemisegment. Chaetal sacs (ectodermal) are evident by the final time point (t=72, or 12h 0min) of the 1251

time-lapse (A-á’ ,à hi hà o espo dsàtoàa outà : àhpfàde elop e talàti eà Ta leà .àTheà haetalàsa sà1252

are often used as anatomical landmarks for determining larval segments, as there are 2 sacs per 1253

hemiseg e tà á’ .à Theà lo alà esode alà luste sà I a isà spots à alig à ithà theà haetalà sa sà ofà theà1254

relevant segment, an example for the 2nd seg e tàisàsho ài àB,àB’,àhighlightedà ithà ello àspotà olo .à1255

The spots are not ectodermal and do not overlap with the chaetal sacs, but instead lie beneath them at 1256

this stage (C-D’’ .àFo ào tai i gà ossàse tio sài àC’-D’’,àthe Imaris 3D stack with lineage spots at t=72 was 1257

exported, and resliced in FIJI for visualizing cross sections. Orange lines in C and D indicate the section 1258

pla esàsho ài àC’,C’’àa dàD’,D’’.àRedàdashedàli eà a ksàtheàapp o i ateà o de à et ee àthe ectoderm 1259

(the chaetal sacs) and the mesoderm. Note that the spots marking the mesoderm are positioned mostly 1260

elo àtheà haetalàsa sà C’’,D’’ , and those spots that overlap in these projections are located in-between 1261

the sacs. 1262

1263

Figure 4 – Figure Supplement 3 – Live-imaging of the formation of mesodermal segmental blocks and 1264

the corresponding chaetal sacs. Development of the right mesodermal band tissues, between the early 1265

52

trochophore and the late trochophore stages (from 26 hpf to 54 hpf) was time-lapse imaged (Video 8). 1266

The embryo was injected in the D quadrant with HistoneH2A-mCherry (red) and EGFP-caax (green) 1267

mRNA, and was raised until swimming larva stage and imaged at a time resolution of every 67 minutes 1268

(18°C equivalent). Two visualizations of the same time point are shown in each panel: i) an onion layer 1269

projection deep enough to exclude all surface ectoderm (A-D), and ii) a curved section along the Z axis 1270

that follows the formation of the ventral row of chaetal sacs (á’-D’). An interpretation of the 1271

morphogenetic events is given with color code for each type of tissue. The position of section curves is 1272

indicated by white dashed lines (D, D’ . At 28 hpf (á,àá’), the mesodermal band is still a bean-shaped, 1273

rapidly-proliferating block of tissue. At 31 hpfà B,à B’), while the mesodermal band spreads dorso-1274

ventrally, chaetal sacs start to form as pockets of ectodermal tissue above the mesodermal band. At 36 1275

hpf C,àC’ , the mesodermal band shows transient but very clear intersegmental constrictions possibly 1276

corresponding to segmental anlage. Note that this is about the same developmental point as shown in 1277

the last time point of the lineage-tracing dataset (Sample A) in the previous figures. Each segmental 1278

mesodermal block corresponds to one of the clonal blocks identified in the lineage tracing. Chaetal sacs 1279

elongate and each mesodermal block starts to surrou dàaàpai àofà haetalàsa sà C’ .àátà51 hpf D,àD’ , the 1280

chaetal sacs are fully elongated and the mesoderm surrounds each one of them like a sheath. Note that 1281

approximate 18°C equivalents of developmental time in hours-post-fertilization (hpf) are shown in the 1282

panels. 1283

Figure 5 – 8ML/R give rise to the mesodermal posterior growth zone. Continuation of lineage analysis 1284

(in Figure 4) of Sample A shows that the 8ML cell created during the 8th division gives rise to a few cells 1285

near the pPGCs á’-á’’ ,à he eà theà esode alà poste io à g o thà zo eà isà lo ated.à Theà MLà li eageà isà1286

highlightedà i à theàt eeà á’’’ .àá othe àsa pleà “a pleàB à ithàdiffe e tào ie tatio à isàsho à i àB-B’’.àáà1287

subset of 3D-projected focal planes show how the 8ML/8MR lineages are located immediately posterior 1288

toàtheàpPGCsà B’,àB’’ .àTheàsa eà olo à odi gàasài àFigure 4. is used: Blue is ML lineage, orange is MR, pink 1289

53

is pPGCs, and yellow 8ML. For a comparison of 3 different samples with 8ML/MR lineages traced see 1290

Figure 5-figure supplement 1 and Videos 6.3 and 6.4. See Videos 6.14 and 6.15 for full time-lapse with 1291

spotsàsho i gàli eages,àf o àposte io àa dàlate alào ie tatio s,àasà ellàasà ˚à otatio à Video 6.14). All 1292

signal from fluorescent constructs is shown in green (EGFP-caax and mVenus-Cdt1(aa1- ,ài àá’-á’’’ ào à1293

in grayscale (Histone2A-mCherry, EGFP-caax, and mVenus-Cdt1(aa1-147)) to highlight general 1294

morphology. 1295

1296

Figure 5 – Figure Supplement 1 – Cell lineages from additional samples compared with the sample 1297

shown in the main text. For easier comparison, only some cell lineages in ML and MR are highlighted in 1298

each sample. Sample A is the one shown in the main text (A). Samples B and C were partially traced for 1299

confirming lineages (B, C). In all three samples, 7ml gives rise to mesodermal cells located on the left side 1300

ofàtheà dàla alàseg e tà á’, B’, C’ .à à asào se edàtoà akeàaàs et i à egio àtoà lào àtheà ightà1301

o l à a al zedà i à theà thi dà sa pleà C,à C’ .à “i ila l ,à MLà a dà MRà ga eà iseà toà li eagesà lo atedà ea 1302

pPGCs (in pink), and lineage trees of all three sa plesàfo àtheseà ellsàlookà e àsi ila .àBla kàli esài àá’’,à1303

B’’,à C’’à i di ateà theà ti eà poi tà sho à i à s apshotsà i à á,à B,à C.à Theseà ti eà poi tsà e eà hose à fo à1304

comparison of each sample at an equivalent time point where both 8ML and 8MR has given rise to 3 1305

cells each, and some of these cells are about to divide. All fluorescence signal from Pdu-FUCCI 1306

(HistoneH2A-mCherry, mVenus-Cdt1(aa1-147)) and EGFP-caax is shown in grayscale. 1307

1308

Figure 6 – Cell divisions of 4d lineage giving rise to the mesodermal primary blast cells. Same dataset 1309

shown in Figures 4 and 5 (Sample A) was analyzed for cell division size asymmetry and orientation. In A, 1310

schema shows the 4d lineage divisions, ML in blue color. The spot diagrams next to the images show the 1311

division by which each primary blast cell is created. The direction of division shown in spot diagrams 1312

54

represents the actual division axis. B is a zoomed-i àpa elàf o àB’àsho i gàtheà eso lastsàafte à à ou dsà1313

of division (2ML=4d111, and MR= d .àTheào ie tatio àofài agi gàisàdepi tedài àB’.àI àC,àfluo es e eà1314

key shows the color and location of each construct. Throughout D-Uà e gedà ha els àa dàD’-U’à g ee à1315

ha elào l às apshotsàofàML’sàdi isio sàf o àtheàlive imaging dataset are shown, outlined with dashed 1316

lines. Note that the initial divisions are highly asymmetric in size (E, G, J, L, S) while the division that 1317

creates 7ml and 7ML appears symmetric (O). In V, the division axes for all primary blast cells are shown 1318

together. See Video 3.1 for the time-lapse dataset the snapshots are derived from (Sample A). 1319

á o headsài àU,àU’:àpPGCs.àTi eàisàsho àasàelapsedài agi gàti eà Ta leà à 1320

1321

Figure 7 – Examples of cell cycle patterns of different cell lineages within 4d. A) The lineages selected 1322

for cell cycle signature analysis are indicated in color (matching the colors used in the subsequent 1323

g aphs .àá’ àE a pleàofàaàspotàpla edài àtheà iddleàofàtheà u lea àsig alà o l ào eàfo alàpla eàf o àtheà1324

stack shown here), and used for o tai i gà Ve usà fluo es e eà i te sit à ea à easu e e tsà fo à1325

each nucleus per time point in Imaris. B) Comparison of mVenus intensity across 3 cell lineages. 8ml 1326

(light blue) is plotted as a continuation of the mesoblast ML (dark blue). Star indicates the time point 1327

7ML divides into 8ml and 8ML in the lineage tree (A) and the cell cycle graph (B). C) Two lineages from 1328

7ml and 8ml show similar cycling patterns. Despite being slightly out of phase at the beginning, they 1329

become synchronized around time point 45. D) Same 8ml lineage is compared this time with a lineage 1330

from 8ML, which is the lineage that gives rise to the mesodermal posterior growth zone cells, and is 1331

much slower in its cycling. In B and D, the dashed boxes show the excerpts from the graphs analyzed in 1332

Figure 7-figure supplement 1. Time axis shows time points (Table 1). Fluorescence measurement is in 1333

arbitrary units (Figure 7-source data 1). Fluo es e tà p otei à olo sà a eà i di atedà i à theà Fluo es e eà1334

Ke à o .à 1335

55

1336

Figure 7 – Figure Supplement 1 – Details from cell cycle graphs. Excerpts from the cell cycle graphs 1337

presented in Figure 7 are shown here with snapshots of the cell nuclei corresponding to the time points 1338

on the graphs (indicated as matching numbers on the graphs and above the images). The cartoons above 1339

cell snapshots show simplified versions of those cells, with nuclei color matching the fluorescent signal. 1340

Red: mCherry only; Yellow: mCherry + mVenus; Orange: mCherry + lower mVenus. Dots in C indicate 1341

time points not shown here. M: Mitosis, G1: Gap phase 1, S: Synthesis, G2: Gap phase 2. 1342

1343

Figure 7 – Figure Supplement 2 - Cell cycle graphs compared across different live-imaged samples. 1344

Sample A is the sample shown in the main figures in the manuscript. Cell cycle graphs obtained from the 1345

same cell lineages from two additional time-lapse samples (Samples B and C) are compared with Sample 1346

A. The volume of mRNA injected in each sample slightly differs leading to differences in the measured 1347

intensity of fluorescence. For easier comparison, data was normalized by finding the maximum for each 1348

series, dividing all time points by the maximum, and multiplying by 100. The same letter codes are used 1349

for the samples as in Figure 5-figure supplement 1. Cell lineages show very similar cycling patterns in all 1350

samples (A-E show 8ML and 8MR from different samples; F-J show ML+8ml from different samples) 1351

(Figure 7-source data 1). 1352

1353

Figure 8 – ML lineage tree indicating known and novel sub-lineages in P. dumerilii, and comparison of 1354

P. dumerilii with the leech. Lineage tree of cells that arise from the left mesoblast (ML) traced in Imaris 1355

is plotted in A. Colors show the extent of lineage analyses earlier researchers accomplished. The same 1356

li eageàt eeà isà sho à i àá’à ithà olo -coded sub-lineages for each primary blast cell, pPGCs, and 8ML. 1357

56

The same color codes are used in the Platynereis larva cartoon in B. On the left in B, 4d lineage diagram 1358

is shown with primary blast cells labeled with the mesodermal region they contribute to. Note how the 1359

cells from the first two divisions of ML (pPGCs) end up next to the cells in the mesodermal posterior 1360

g o thàzo eà MPG) ,à hi hàa eà o à u hàlate .àFo à o pa iso ,à dà DM’’ àli eageàdiag a àa dàp i a à1361

blast cell lineages (again color-coded) are shown for the leech (a clitellate annelid) (Cho et al., 2014; 1362

Gline et al., 2011; Rebscher, 2014). The first six divisions of ML make cells that contribute to non-1363

segmental mesoderm in anterior regions and the gut (both in dark gray). The following divisions make 1364

segmental primary blast cells, each or which contributes to the mesoderm of two consecutive 1365

hemisegments (for example, sm1 contributes to part of 1st segment and part of 2nd segment 1366

mesoderm, in total making 1 segment-worth of mesoderm). Note that in the leech, PGCs (pink) arise in 1367

later cell divisions, as a subset within the segmental mesoderm population that originate from a primary 1368

blast cell (sm10, and sm12 through 22). 1369

1370

Figure 9 – Phylogenetic distribution of embryonic and post-embryonic teloblasts in some annelids, and 1371

the relation to the number of segments made. A simplified annelid phylogeny (Weigert and Bleidorn, 1372

2016) summarizes the main annelid species investigated for segment formation and teloblasts. The 1373

number of segments appearing during the pelagic larval phase (but patterned during embryogenesis) 1374

and the number of segments added posteriorly after the start of benthic life can vary considerably 1375

(Balavoine, 2014). In leeches, a fixed number of segments is made during direct embryonic development 1376

and none added after hatching. In non-leech clitellates (such as Tubifex), the embryos make a few tens of 1377

segments during embryogenesis and add many more after hatching. In non-clitellate annelids (Errantia, 1378

Sedentaria, and early branching), the number of larval segments is variable, and many more are added in 1379

post-larval development: In Nereidids (Errantia) only 3 larval segments develop, while in Capitella 1380

57

(Sedentaria), 13 larval segments (Thamm and Seaver, 2008) and in Chaetopterus (an early-branching 1381

annelid) 15 larval segments develop (Seaver et al., 2001). Embryonic and post-embryonic teloblasts differ 1382

in the way they have been evidenced. Embryonic teloblasts in the Clitellates (Weisblat and Kuo, 2014; 1383

Goto et al, 1999; Nakamoto et al, 2000) and in Platynereis (this work) have been directly observed in live 1384

specimens. By contrast, post-embryonic teloblasts are inconspicuous cells that are identified only so far 1385

by their molecular stem cell signature (Gazave et al, 2013; Dill and Seaver, 2008; Özpolat and Bely, 2016). 1386

No direct observation of post-embryonic teloblast patterns of division is available so far. No direct 1387

evidence for embryonic teloblasts giving birth to post-embryonic teloblasts exists in any species so far to 1388

our knowledge. The ancestor of annelids presumably had both larval and post-larval segments, and thus 1389

it raises the question of the presence of both embryonic and post-embryonic teloblasts in this ancestral 1390

annelid, and even more broadly in other spiralians such as mollusks. Structures shown in the figure are 1391

olo à odedàa dàe plai edài àtheà s olàke .à 1392

1393

Figure 10 – Injection setup and general experimental outline for imaging samples. A) A Zeiss inverted 1394

scope with a gliding stage, coupled with the Eppendorf Femtojet microinjector and Eppendorf 1395

Transferman micromanipulation system were used for injections. Femtotip ready-to-use injection 1396

needles were back-filled with injection solution á’ .à á à aga oseà platfo à á’’ à ithà aà g oo eà la geà1397

enough to contain embryos was prepared using a plastic custom-made mold. Under a dissecting scope, 1398

s allài isio sà e eà adeàtoàtheà ightàsideàofàtheàg oo eà á’’’ ,àa dàtheseài isio sà e eàusedàtoà e o e 1399

embryos from the needle by sliding the needle through them. The agarose platform was placed into a 1400

small petri dish lid and covered with filtered sea water before the embryos were transferred into the 1401

dish. In B and C, the steps of general experimental outline for live imaging samples at different stages are 1402

listed. 1403

58

1404

Figure 10 - Figure Supplement 1 – Heating plate setup for mounting samples in glass-bottom dishes. 1405

One of the heating blocks was turned upside down to use its flat surface, and was painted with black 1406

marker to provide contrasting background for mounting samples. The heating plate was placed under a 1407

ste eos ope.àTe pe atu eà asàsetàtoà à˚C.àGlass-bottom dishes were pre-warmed on this surface. One 1408

injected sample was placed on the dish with as little sea water as possible. Next, a drop of 100 µl of 1% 1409

low-melting agarose was added on, and before the agarose started solidifying, the sample was rotated to 1410

the desired position with the help of an eyelash tool. If needed, more agarose was slowly added with the 1411

help of a micropipette to cover the glass surface completely. Then the plate was transferred on a colder 1412

surface to let agarose solidify. Once agarose solidified (in about 3-5 minutes), filtered natural sea water 1413

was added on top to cover the aga ose,à a dà theà dishesà e eà o e edà ithà theà lidà a dà keptà atà à ˚Cà1414

incubator, in a wet chamber, until imaging. 1415

1416

1417

1418

FIGURES 1419

FIGURE 1 1420

59

1421

FIGURE 1 – Figure Supplement 1 1422

1423

1424

FIGURE 1 – Figure Supplement 2 1425

60

1426

1427

1428

1429

1430

1431

1432

1433

FIGURE 2 1434

61

1435

1436

1437

1438

1439

FIGURE 3 1440

62

1441

1442

1443

1444

1445

1446

1447

1448

FIGURE 3 – Figure Supplement 1 1449

63

1450

1451

1452

1453

1454

1455

FIGURE 3 – Figure Supplement 2 1456

64

1457

1458

1459

1460

1461

1462

FIGURE 4 1463

65

1464

1465

1466

1467

1468

FIGURE 4 – Figure Supplement 1 1469

66

1470

1471

1472

1473

1474

1475

FIGURE 4 – Figure Supplement 2 1476

67

1477

1478

1479

1480

1481

1482

1483

1484

1485

1486

FIGURE 4 – Figure Supplement 3 1487

68

1488

1489

1490

1491

1492

1493

FIGURE 5 1494

69

1495

1496

1497

1498

1499

1500

FIGURE 5 – Figure Supplement 1 1501

70

1502

1503

1504

1505

1506

1507

1508

1509

1510

FIGURE 6 1511

71

1512

1513

1514

1515

1516

1517

1518

1519

FIGURE 7 1520

72

1521

1522

1523

1524

1525

1526

FIGURE 7 – Figure Supplement 1 1527

73

1528

1529

1530

1531

1532

1533

1534

1535

1536

1537

FIGURE 7 – Figure Supplement 2 1538

74

1539

1540

1541

1542

1543

1544

1545

FIGURE 8 1546

75

1547

FIGURE 9 1548

76

1549

1550

1551

1552

1553

1554

1555

1556

1557

1558

FIGURE 10 1559

77

1560

1561

1562

1563

1564

1565

1566

1567

1568

FIGURE 10 – Figure Supplement 1 1569

78

1570

1571

1572

1573

1574

1575

1576

1577

VIDEO LEGENDS 1578

79

Videos for Figure 1 1579

Video 1.1 – Cell cycle reporter Pdu-FUCCI time-lapse in the early embryo. These videos show an embryo 1580

injected with Histone2A-mCherry and mVenus-cdt1(aa1-147) mRNAs at 1-cell stage, imaged every 4 1581

minutes, starting around 6:30 hpf for about 2 hours. The cells marked by arrowheads in the videos (ab, a, 1582

b) are shown in detail in Figure 1. Red: mCherry, Green: mVenus 1583

Video 1.2 – Cell cycle reporter Pdu-FUCCI time-lapse in the early embryo. Same Video 1.1 is shown with 1584

mVenus (green) channel only. 1585

Video 2.1. – Time-lapse of prototroch cells being born and exiting cell cycle. Embryo was injected with 1586

Pdu-FUCCI (Histone2A-mCherry and mVenus-cdt1(aa1-147) and EGFP-caax mRNA at 1-cell stage, and 1587

imaged every 10 minutes for about 6 hours. mVenus and EGFP signals are imaged from a single channel. 1588

Video 2.2 - Explanation of the onionization process, and display of all the onion layers. Video is made 1589

from stack same as the sample shown in Figure 1-figure supplement 1C. Embryo was injected with Pdu-1590

FUCCI and EGFP-caax mRNA at 1-cell stage. 1591

Videos for Figure 3 1592

Video 3.1 – Original 3D-projected time-lapse dataset with annotations (for Sample A). Embryo was 1593

injected with Pdu-FUCCI and EGFP-caax mRNA into the D quadrant at 4-cell stage, and imaged every 10 1594

minutes. Note that uninjected side of the embryo appears dark. Also see Videos 5.1 and 5.2 for 1595

additional samples imaged from a different batch of fertilization and injections. 1596

Video 3.2 – Video of stack showing the same individual from Video 3.1 imaged next day. The sample 1597

displays normal development after injection and imaging. 1598

Video 4.1 - pPGC migration traced in Imaris. Same confocal time-lapse dataset (Sample A) from Video 1599

3.1 was used for this analysis. Pink spots mark the pPGC nuclei. Posterior view. 1600

Video 4.2 - pPGC migration traced in Imaris. Same confocal time-lapse dataset (Sample A) from Video 1601

3.1 was used for this analysis. Pink spots mark the pPGC nuclei. Lateral view 1602

Videos for additional Samples 1603

Video 5.1 – Time-lapse of Sample B. Video shows time-lapse 3D projections of Sample B, imaged every 1604

12 mins for 13 hours. Embryo was injected with Pdu-FUCCI and EGFP-caax mRNA into the D quadrant at 1605

4-cell stage. 1606

Video 5.2 – Time-lapse of Sample C. Video shows time-lapse 3D projections of Sample C, imaged every 1607

12 mins for 13 hours. Embryo was injected with Pdu-FUCCI and EGFP-caax mRNA into the D quadrant at 1608

4-cell stage. 1609

Videos for Figure 4 1610

80

Video 6.1 – Time-lapse of the ML lineage, and 360 rotation of the last time point. Video shows spots 1611

from Imaris lineage tracing for extensively-traced ML lineage (blue), some tracing for MR lineage 1612

(orange), and pPGCs from both MR and ML (pink). Data set is the same as shown in Video 3.1 (Sample A). 1613

Video 6.2 – 3ml, posterior view. 3ml sub-lineage is highlighted (yellow) within the ML lineage (blue) 1614

time-lapse video. 1615

Video 6.3 – 3ml, lateral view. 1616

Video 6.4 – 4ml, posterior view. 4ml sub-lineage is highlighted (yellow) within the ML lineage (blue) 1617

time-lapse video. 1618

Video 6.5 – 4ml, lateral view. 1619

Video 6.6 – 5ml, posterior view. 5ml sub-lineage is highlighted (yellow) within the ML lineage (blue) 1620

time-lapse video. 1621

Video 6.7 – 5ml, lateral view. 1622

Video 6.8 – 6ml, posterior view. 6ml sub-lineage is highlighted (yellow) within the ML lineage (blue) 1623

time-lapse video. 1624

Video 6.9 – 6ml, lateral view. 1625

Video 6.10 – 7ml, posterior view. 7ml sub-lineage is highlighted (yellow) within the ML lineage (blue) 1626

time-lapse video. 1627

Video 6.11 – 7ml, lateral view. 1628

Video 6.12 – 8ml, posterior view. 8ml sub-lineage is highlighted (yellow) within the ML lineage (blue) 1629

time-lapse video. 1630

Video 6.13 – 8ml, lateral view. 1631

Video 6.14 – 8ML, posterior view. 8ML sub-lineage is highlighted (yellow) within the ML lineage (blue) 1632

time-lapse video. 1633

Video 6.15 – 8ML, lateral view. 1634

Video 7 – Animation of the segmental mesoderm clonal regions. Posterior and lateral views together. 1635

Video for Figure 4-figure supplement 3 1636

Video 8 – Segment formation in the P. dumerilii larva. Embryo was injected with HistoneH2A-mCherry 1637

and EGFP-caax mRNA, raised until 24 hpf swimming larval stage, immobilized with DMSO treatment and 1638

mounted in low-melting agarose for time-lapse live-imaging. The resulting 4D dataset was modified in 1639

81

FIJI with the Onionizer (see Methods) to reveal the curved ectodermal and mesodermal larval segments 1640

forming. Upper row is the lateral view, middle row is the cross-section, and lower row is the label key 1641

with corresponding colors. The timer shows the calculated developmental time. See Figure 4-figure 1642

supplement 3 for details. 1643

TABLES 1644

Table 1 – Calculated Developmental Time and Corresponding Stage For Each Time Point at Imaging 1645

Temperatures For the video (Sample A) from which we show most of the lineage tracing data in this 1646

manuscript, we calculated what each time point roughly corresponds to in the normal staging by Fischer 1647

et al 2010. This calculation was done based on the starting stage, end stage, and how long this normally 1648

takes under 18˚ C conditions. For example, Video 3.1 (used in Figures 3, 4, 5 and 6) starts at 8 hpf 1649

(stereoblastula stage), and in the final time point the sample appears towards the end of mid 1650

trochophore stage. Even though elapsed time is 12 hours of imaging (72 time points), under 18˚ C 1651

incubation conditions, about 35-40 hours is required to reach this stage. Thus, we calculated each time 1652

point in the live-image dataset corresponding roughly to: T = hou sà+à i sà àtp ,à he eà tp àisàti eà1653

point, and T is the calculated developmental time. According to this, the end calculated developmental 1654

time for Video 3.1 is: T = 8h + (25 mins x 72) = 37 hours 35 minutes. 1655

1656

82

Table 1 – Calculated Developmental Time and Corresponding Stage For Each Time Point at Imaging 1657

Temperatures 1658

time

point (tp)

Elapsed

Imaging Time

(h:m:s)

Calculated

Developmental Time

(T) at 18˚àC (h:m:s)

Corresponding Developmental

Stage at 18˚àC (Fischer et al.,

2010)

1 0:00:00 8:00:00 stereoblastula

3 0:20:00 8:50:00 stereoblastula

5 0:40:00 9:40:00 stereoblastula

7 1:00:00 10:30:00 stereoblastula

9 1:20:00 11:20:00 stereoblastula

11 1:40:00 12:10:00 stereoblastula

13 2:00:00 13:00:00 protrochophore

15 2:20:00 13:50:00 protrochophore

17 2:40:00 14:40:00 protrochophore

19 3:00:00 15:30:00 protrochophore

21 3:20:00 16:20:00 protrochophore

23 3:40:00 17:10:00 protrochophore

25 4:00:00 18:00:00 protrochophore

27 4:20:00 18:50:00 protrochophore

29 4:40:00 19:40:00 protrochophore

31 5:00:00 20:30:00 protrochophore

33 5:20:00 21:20:00 protrochophore

35 5:40:00 22:10:00 protrochophore

37 6:00:00 23:00:00 protrochophore

39 6:20:00 23:50:00 protrochophore

41 6:40:00 24:40:00 early trochophore

43 7:00:00 25:30:00 early trochophore

45 7:20:00 26:20:00 mid-trochophore

47 7:40:00 27:10:00 mid-trochophore

49 8:00:00 28:00:00 mid-trochophore

51 8:20:00 28:50:00 mid-trochophore

53 8:40:00 29:40:00 mid-trochophore

55 9:00:00 30:30:00 mid-trochophore

57 9:20:00 31:20:00 mid-trochophore

59 9:40:00 32:10:00 mid-trochophore

61 10:00:00 33:00:00 mid-trochophore

63 10:20:00 33:50:00 mid-trochophore

65 10:40:00 34:40:00 mid-trochophore

67 11:00:00 35:30:00 mid-trochophore

69 11:20:00 36:20:00 mid-trochophore

71 11:40:00 37:10:00 mid-trochophore

72 11:50:00 37:35:00 mid-trochophore

1659

G2

G2 M

G1

MM G1 S

M

mVenus cdt1

Pdu-cdt1

Geminin interaction MCM2-7 interaction

PIP

1 778

1-147

EdUEdU+mVe+mCh mVe mCh

51 15.6

247 75.3

7 2.1

23 7

(n) %

64 mins0 min

0 min 52 mins

a b

a

b

aababab b

S

G2

G1

M

50 μm 12 hpf

HistoneH2A-mCherry

+ mVenus-cdt1(aa1-147)

HistoneH2A-mCherry

mVenus-cdt1(aa1-147)

=Pdu-FUCCI

A

B

E

C D

D’

D’’

D’’’

Caenorhabditis elegans cdt1

Drosophila melanogaster cdt1

Bombyx mori cdt1

Hydra vulgaris cdt1

Lottia gigantea cdt1

Ciona intestinalis cdt1

Xenopus laevis cdt1

Rattus norvegicus cdt1

Mus musculus cdt1

Capitella teleta cdt1

Platynereis dumerilii cdt1

Homo sapiens cdt1

PIP BOX

EdU 36 to 72 hpf

EdU 36 to 72 hpf

EdU 36 to 72 hpf

15 μm

*

**

***

**

**

L

Anterior

Posterior

Anterior

Posterior

R

50 μm

25 μm

*

*

* **

*

*

*

* **

*

Ed

UE

dU

+D

AP

IA

cet-

tub

. +

EdU

+D

AP

I

Acet-

tub.

+

EdU

+D

AP

I

Acet-

tub

. +

EdU

+D

AP

I

Acet-

tub

.

+ E

dU

Acet-

tub.

+ E

dU

late-trochophore late-trochophore

56hpf

A

DE

E

F

F F’

G

G

G’

E’

E’’

B C C’

C’’

C’’C’

mVenus-cdt1(aa1-147)

EGFP-caaxmVenus-cdt1(aa1-147)

EGFP-caax HistoneH2A-mCherry

EdU

5-7

hpf

EdU

5-7

hpf

EdU

11-2

4 h

pf

FU

CC

I in

jecte

d

FU

CC

I in

jecte

dE

dU

24-4

8 h

pf

E F

B

D

A

C

EdUDAPI

EdUDAPI

EdUDAPI

EdUDAPI

0-hpf 18-hpf6-hpf 12-hpf 24-hpf

mRNA injection at 1-2 hpf live snapshot live snapshotmRNA injection at 1-2 hpf

EdU treat 11-24 hpf Fix

0-hpf 18-hpf6-hpf 12-hpf 24-hpf

EdU treat 24-48 hpf Fix

0-hpf 48-hpf6-hpf 24-hpf

EdU treat 5-7 hpf Fix

0-hpf 18-hpf6-hpf 12-hpf 24-hpf

EdU treat 5-7 hpf Fix

0-hpf 6-hpf 12-hpf 48-hpf24-hpf

0-hpf 6-hpf 12-hpf 48-hpf24-hpf

48 hpf

48 hpf

48 hpf24 hpf

24 hpf

24 hpf

mVenus-cdt1(aa1-147)

EGFP-caax

HistoneH2A-mCherry mVenus-cdt1(aa1-147)

EGFP-caax

HistoneH2A-mCherry

50 μm

D

6h 30min

8h 10min

9h 50min

11h 30min

6h 30min

8h 10min

9h 50min

11h 30min

6h 30min

8h 10min

9h 50min

11h 30min

0h 00min

1h 30min

3h 10min

4h 50min

0h 00min

1h 30min

3h 10min

4h 50min

0h 00min

1h 30min

3h 10min

4h 50min

postero-lateral posterior lateral postero-lateral posterior lateral

A A’ A’’ E E’ E’’

F F’ F’’

G G’ G’’

H H’ H’’

B B’ B’’

C C’ C’’

D D’ D’’

8 hpf

mVenus-cdt1(aa1-147)

EGFP-caax

HistoneH2A-mCherry mVenus-cdt1(aa1-147)

EGFP-caax

HistoneH2A-mCherry mVenus-cdt1(aa1-147)

EGFP-caax

HistoneH2A-mCherry mVenus-cdt1(aa1-147)

EGFP-caax

HistoneH2A-mCherry mVenus-cdt1(aa1-147)

EGFP-caax

HistoneH2A-mCherry mVenus-cdt1(aa1-147)

EGFP-caax

HistoneH2A-mCherry

0h 00min 6h 30min

8h 10min

9h 50min

11h 30min

6h 30min

8h 10min

9h 50min

11h 30min

1h 30min

3h 10min

4h 50min

0h 00min

1h 30min

3h 10min

4h 50min

A E

F

G

H

E’

F’

G’

H’

B

C

D

A’

B’

C’

D’

mVenus-cdt1(aa1-147)

EGFP-caaxmVenus-cdt1(aa1-147)

EGFP-caaxmVenus-cdt1(aa1-147)

EGFP-caaxHistoneH2A-mCherry mVenus-cdt1(aa1-147)

EGFP-caax

HistoneH2A-mCherry

7:29 hpf

7:47 hpf

7:59 hpf

8:17 hpf

7:20 hpf

ML MR

mVenus-cdt1(aa1-147)

EGFP-caax

HistoneH2A-mCherry

Right M blastomere (MR = 4d2) and its progeny

4D

4d111

=2ML

4d211

=2MR

0h 00min = 8 hpf

Left M blastomere (ML = 4d1) and its progeny

COLOR KEY

Highlighted subset within ML (blue) lineage

4 putative Primordial Germ Cells (pPGCs)

4d 4D

3D

ML

1ml

2ml

3ml

4ml

2mr

1mr1ML

2ML

3ML

6ml

5ml

7ml

8ml

1MR

MR

12h 0min

12h 0min

12h 0min

12h 0min

lateralposteriorposterior

12h 0min

12h 0min

12h 0min

12h 0min

12h 0min

Progeny of 5ml

Progeny of 7ml

Progeny of 8ml

12h 0min 12h 0min12h 0min

Progeny of 6ml

cryp.seg.

seg.1

seg.2

seg.3

ML

3ml

PGCs

4ml5ml 6ml

7ml8ml 8ML

ML

3ml

PGCs

4ml5ml 6ml 7ml

8ml 8ML

ML

3ml

PGCs

4ml5ml 6ml 7ml

8ml 8ML

ML

3ml

PGCs

4ml5ml 6ml

7ml 8ml8ML

A B

B

D D’ D’’ D’’’

E E’ E’’E’’’

F F’ F’’F’’’

G G’ G’’G’’’

CB’

mVenus-cdt1(aa1-147)

EGFP-caax

HistoneH2A-mCherry

mVenus-cdt1(aa1-147)

EGFP-caax

5ml

8:10

Posterior6ml

7ml

8ML

8ml

mVenus-cdt1(aa1-147)EGFP-caax

posterior

CR

OS

S-S

EC

TIO

N

lateral

posterior posterior

seg.2

seg.2

seg.2seg.1

seg.3

12h 0min 12h 0min12h 0min 12h 0min

12h 0min 12h 0min

mesoderm

ectoderm

mesoderm

ectoderm

Progeny of 7ml: mesoderm of seg.2

A C DA’

B B’

C’

C’’

D’

D’’

mVenus-cdt1(aa1-147)

EGFP-caax

mVenus-cdt1(aa1-147)

EGFP-caax

mVenus-cdt1(aa1-147)

EGFP-caax

28 hpf 31 hpf

36 hpf 51 hpf

A

A’ B’

ant.

post. post.post. post.

post. post.post. post.

seg.1

seg.2seg.3

ant. B

C

C’ D’

D

E

ant. ant.

ant. ant.

ant. ant.

20 µm

20 µm

8ML

8MR

pPGCs

L

Anterior

Posterior

R

12h 0min12h 0min

posterior lateral lateral

8ML

Progeny of 8MLA

B B’

B’’

A’ A’’ A’’’

A’’’’

8ml8ML

mVenus-cdt1(aa1-147)EGFP-caax

mVenus-Cdt1(aa1-147)EGFP-caax

HistoneH2A-mCherry

50 µm

A

A’

A’’

B

B’

B’’

C

C’

C’’8ML

8ML

pPGCs pPGCs pPGCs

7ml

8MR

8MR 8ML7ml

8MR

7mr

8ML

7ml8MR

8ML 8MR 8ML 8MR

Nuclear:

HistoneH2A-mCherry

Nuclear:

mVenus-cdt1 (aa1-147)

Membrane: EGFP-Caax

Fluorescence key Division axes of ML

4D

4d111

=2ML

4d211

=2MR

0h 00min = 8 hpf

P

Dorsal

Ventral

A

0h 00min

0h 10min

0h 20min

0h 40min

0h 50min

1h 10min

1h 20min

1h 40min

1h 50min

2h 00min

2h 20min

2h 30min

2h 50min

3h 00min

3h 10min

3h 20min

3h 40min

4d

3D

ML

1ml

2ml

1ML

2ML

2ML

5ML

6ML

7ML

8ML

5ml

5ml 4ml3ml

6ml

6ml

7ml

7ml

8ml

8ml

MR

3ml

3ML

4ml

4ML

AB

B

B’ C V

D D’ J J’ O O’

P P’

Q Q’

R R’

S S’

T T’

U U’

K K’

L L’

M M’

N N’

E E’

F F’

G G’

I I’

mV

en

us

inte

nsi

ty m

ea

n

Figure

Sup.1A

Figure

Sup.1B

Figure Sup. 1C

*

8ml

ML

pPGC

mV

en

us

inte

nsi

ty m

ea

nm

Ve

nu

sin

ten

sity

me

an

B

C

D

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72

8ml7ml

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65

8ml8ML

0

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72

8ML8ml

pPGC

MLA

*

A’

7ml

0 5 10 15 20 25 30 35 40 45 50 55 60 65 70�me

10

20

30

40

50

60

70

80

0 5 10 15 20 25 30 35 40 45 50 55 60 65 70�me

0 5 10 15 20 25 30 35 40 45 50 55 60 65 70�me

10

20

30

40

50

60

20

40

60

80

100

120

140

A

B

C

5

5

6

67

7

8

8

19 20 21 2322 24

22

23

24

19

20

21

20

20

22

22

25

25

29

29

34

34

39

39

40

40

43

43

G2

M M

M MG1 S

G2M MG1 G1S

S(?)

. . . . .. . . . . . . . . . .

mVenus-cdt1(aa1-147)

EGFP-caax

Fluorescence Key

HistoneH2A-mCherry

mV

en

us

inte

nsi

ty m

ea

n

SampleC-8MR

SampleC-8ML

SampleB-8MR

SampleB-8ML

SampleA-8ML

0

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58

0 5 10 15 20 25 30 35 40 45 50�me

20

40

60

80

100

120

mV

en

us

inte

nsi

ty m

ea

n

0

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58

0 5 10 15 20 25 30 35 40 45 50�me

20

40

60

80

100

120

mV

en

us

inte

nsi

ty m

ea

n

0

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58

0 5 10 15 20 25 30 35 40 45 50�me

20

40

60

80

100

120

mV

en

us

inte

nsi

ty m

ea

n

0

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58

0 5 10 15 20 25 30 35 40 45 50�me

20

40

60

80

100

120

mV

en

us

inte

nsi

ty m

ea

nA

B

C

D

E

0

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58

0 5 10 15 20 25 30 35 40 45 50�me

20

40

60

80

100

120

mV

en

us

inte

nsi

ty m

ea

n

SampleC-MR+8mr

SampleC-ML+8ml

SampleB-MR+8mr

SampleB-ML+8ml

SampleA-ML+8ml

0

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58

0 5 10 15 20 25 30 35 40 45 50�me

20

40

60

80

100

120

mV

en

us

inte

nsi

ty m

ea

n

0

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58

0 5 10 15 20 25 30 35 40 45 50�me

20

40

60

80

100

120m

Ve

nu

sin

ten

sity

me

an

0

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58

0 5 10 15 20 25 30 35 40 45 50�me

20

40

60

80

100

120

mV

en

us

inte

nsi

ty m

ea

n

0

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58

0 5 10 15 20 25 30 35 40 45 50�me

20

40

60

80

100

120

mV

en

us

inte

nsi

ty m

ea

nF

G

H

I

J

0

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58

0 5 10 15 20 25 30 35 40 45 50 55�me

20

40

60

80

100

120

55

55

55

55

AML

Wilson, 1892

Fischer and Arendt, 2013

This study

ML

3ml

pPGCs

4ml5ml 6ml

7ml8ml 8ML

A’

4d

ML

1ml

2ml

3ml

4ml

1ML

2ML

3ML

5ml

4ML

6ml

5ML

7ml

6ML

8ml

7ML

8ML

MR

?

?

larval

segment 1

larval

segment 2

larval

segment 3

MPGZ

cryptic

segment

pPGCs

Platynereis Helobdella (leech)

segmental

mesoderm 9

segmentalmesoderm 1-8

segmental

mesoderm 10

segmental

mesoderm 11

segmental

mesoderm 12

segmental

mesoderm 13

DM’’

ML

em1

em2

em3

em4

em5

em6

sm9

sm1-8

sm10

sm11

MR

non-segmental

mesoderm

(head and

digestive tract)

sm12

sm13

sm14

sm15

segmental

mesoderm 23

segmental

mesoderm

(14-22)sm16

sm23

sm22

B C

anterior

anterior

seg.1-4

seg.5

seg.6

seg.7

seg.8

seg.9

seg.10

seg.11

seg.12

seg.13

seg.14

seg.15

seg.16

seg.17

Chaetopterus

Clite

llata

AN

NE

LID

AErra

ntia

Sedenta

ria

juvenile/adultembryo/larva

Leeches

?

teloblasts teloblasts unknown

segmental mesoderm progenitors

made during embryogenesis

made in the juvenile/adult

post-embryonic teloblasts

in segment addition zone

post-embryonic teloblasts

not investigated

SYMBOL KEY

Capitella

Non-leech

clitellatesTubifex

Helobdella

Errant

“polychaete”

Sedentary

“polychaete”

Other segmented

spiralians

Early-branch

annelids

ancestral annelid with embryonic

and post-embryonic teloblasts?

Platynereis

?

??

?

?

?

?

segmental mesoderm

Chitons

For Imaging embryonic and

pre-larval stages (7-24 hpf)

Inject at 1-, 2-, or 4-cell stage

Select healthy embryos via

external morphology

(4 oil droplets)

Mount on glass-bottom dish,

on the 37 C heating plate

(See figure supplement)

(See figure supplement)

Start imaging

Start imaging

Mount on glass-bottom dish,

on the 37 C heating plate

Treat with 10% DMSO for

10 mins to eliminate cilia.

Select healthy embryos via

external morphology, and

behavior (larval eyes,

normal swimming)

Inject at 1-, 2-, or 4-cell stage

For imaging larval stages

(24-60 hpf)

(0-2 hpf)

(5-6 hpf)

(5-6 hpf)

(7-8 hpf)

(0-2 hpf)

(24 hpf)

(24 hpf)

(24.5 hpf)

(25 hpf)

CD D CD D

Wash (filtered natural sea water)

A’

C

A’’

A’’’

B

A

Glass-bottom dish is placed

on heating plate at 37 °C

Low-melt agarose

Low-melt agarose

Objective

Filtered natural sea water

is added after agarose solidified

Glass coverslip

Eyelash tool

to position the embryo

Injected embryo

A B