Upload
others
View
1
Download
0
Embed Size (px)
Citation preview
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 ([email protected]), Guillaume Balavoine ([email protected]) 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