© 2001 Wiley-Liss, Inc. genesis 29:188–195 (2001)
ARTICLE
Cell Populations and Morphogenetic Movements
Underlying Formation of the Avian Primitive Streak
and Organizer
Aaron Lawson and Gary C. Schoenwolf*
Department of Neurobiology and Anatomy, University of Utah School of Medicine, Salt Lake City, Utah
Received 28 November 2000; Accepted 25 January 2001
Published online 00 Month 2001
Summary: The cell populations and morphogenetic viewed from its dorsal or ventral surface, and its appear-
movements that contribute to the formation of the avian ance overtly defines the midline of the embryo and the
primitive streak and organizer—Hensen’s node—are rostrocaudal axis. Soon after its appearance, the primi-
poorly understood. We labeled selected groups of cells
with fluorescent dyes and then followed them over time tive streak undergoes progression, during which it rap-
during formation and progression of the primitive streak idly elongates rostrocaudally and assumes a linear pro-
and formation of Hensen’s node. We show that (1) the file. At about this time, the rostral end of the primitive
primitive streak arises from a localized population of streak becomes swollen, marking the location of the
epiblast cells spanning the caudal midline of Koller’s organizer (called Hensen’s node in avian embryos). Dur-
sickle, with the mid-dorsal cells of the primitive streak ing progression of the primitive streak, epiblast cells
arising from the midline of the epiblast overlying Koller’s programmed to form the mesoderm and endoderm be-
sickle and the deeper and more lateral primitive streak
cells arising more laterally within the epiblast overlying gin to ingress through the streak, thereby establishing
the sickle, from an arch subtending about 30°; (2) con- the germ layers. The final change that occurs in the
vergent extension movements of cells in the epiblast primitive streak is regression, a rostral-to-caudal shorten-
overlying Koller’s sickle contribute to formation of the ing of the streak that continues until Hensen’s node and
initial primitive streak; and (3) Hensen’s node is derived a short persisting remnant of the primitive streak is
from a mixture of cells originating both from the epiblast incorporated into the tail bud (Schoenwolf, 1979; for
just rostral to the incipient (stage 2) primitive streak and reviews of avian gastrulation, see Bellairs, 1986; Lemaire
later from the epiblast just rostral to the elongating
(stage 3a/b) primitive streak, as well as from the rostral and Kessel, 1997; Schoenwolf and Smith, 2000; Stern
tip of the progressing streak itself. Collectively, these and Canning, 1988).
results provide new information on the formation of the Several studies have focused on the induction of the
avian primitive streak and organizer, increasing our un- primitive streak and considerable evidence, based
derstanding of these important events of early develop- mainly on tissue transplantation, suggests that the induc-
ment of amniotes. genesis 29:188–195, 2001. tive signal is provided by the posterior (caudal) marginal
© 2001 Wiley-Liss, Inc. zone (PMZ; i.e., the posterior/caudal blastoderm ring
that forms an interface between the area pellucida and
Key words: cell behaviors; chick embryos; gastrulation; area opaca) (Bachvarova et al., 1998; Eyal-Giladi and
Hensen’s node Khaner, 1989; Eyal-Giladi et al., 1992; Khaner, 1998;
Khaner and Eyal-Giladi, 1986, 1989; Khaner et al., 1985).
Additional evidence implicating the PMZ in induction of
INTRODUCTION
the primitive streak comes from the demonstration that
secreted proteins, such as Vg1 (Shah et al., 1997), chor-
Gastrulation is a key event of early animal development. din (Streit et al., 1998), and Wnt-8c (Hume and Dodd,
It is characterized by extensive and highly coordinated 1993), are expressed in the PMZ, and some of these (i.e.,
morphogenetic movements that result in the formation
of the three primary germ layers: the ectoderm, meso-
derm, and endoderm. The hallmark of gastrulation in * Correspondence to: Gary C. Schoenwolf, Department of Neurobiology
higher vertebrates is the formation of the primitive and Anatomy, University of Utah School of Medicine, 50 N. Medical Drive,
streak, a thickening of the epiblast which in avian em- Salt Lake City, UT 84132.
bryos, appears suddenly at stage 2 (Hamburger and Ham- E-mail: Schoenwolf@med.utah.eduAaron Lawson’s Permanent address: Department of Anatomy, University
ilton, 1951) in the caudal third of the area pellucida. The of Ghana Medical School, P.O. Box 4236, Accra, Ghana, West Africa.
nascent primitive streak is triangular in shape when Contract grant sponsor: NIH, Contract grant number: NS 18112.
PRIMITIVE STREAK FORMATION 189
Vg1 and chordin) are capable of inducing primitive overlying the rostral part of Koller’s sickle. Additionally,
streaks when expressed ectopically. our results suggest that these precursor cells undergo a
Existing data on how the primitive streak forms are convergent-extension movement along Koller’s sickle
contradictory. Current ideas can be summarized with during their rostral displacement, thereby establishing
two extreme models. The first, based on the assumption the incipient (triangularly shaped) primitive streak. Sec-
that primitive streak precursor cells reside randomly ond, we labeled cells in areas “c” and “a” (i.e., at the
throughout the epiblast, proposes that formation of the rostral end of the primitive streak, and the epiblast just
primitive streak is achieved by a long-range caudomedial rostral to the streak, respectively; Garcia-Martinez et al.,
migration of precursor cells, followed by their aggrega- 1993) at stages 2 and 3a/b (Hamburger and Hamilton,
tion in the caudal midline (Stern and Canning, 1990). 1951; with stage 3 subdivided as described by Schoen-
The second, based on the assumption that the primitive wolf et al., 1992; also see Darnell et al., 1999), to deter-
streak precursor cells reside in the PMZ, proposes that mine their relative positions during progression of the
the formation of the primitive streak is achieved by the primitive streak, as well as their contributions to Hens-
short-range rostral displacement of streak precursor cells en’s node. Our results demonstrate that during progres-
from the midline of the PMZ (Eyal-Giladi et al., 1992; sion of the primitive streak, cells from area “c” move
Vakaet, 1970; Wei and Mikawa, 2000). Both of these rostrally to co-mingle with those from area “a,” and
models fail to account for the important fact that the together they form Hensen’s node. Thus, the term Hens-
initial primitive streak has a triangular profile. Moreover, en’s node should be used to designate the rostral end of
a recent descriptive study using scanning electron mi- the primitive streak only at stages near the end of prim-
croscopy suggested that the primitive streak does not itive streak progression (when these two populations of
arise from an extensive movement of cells from more precursor cells become intermixed) and later.
rostral or peripheral areas of the epiblast. Instead, epi-
blast cells that are already in the vicinity of the forming RESULTS
primitive streak at stage XIV (Eyal-Giladi and Kochav,
1976) seem to become transformed locally into primitive Mapping Primitive Streak Precursor Cells
streak cells by stage 2 (Lawson and Schoenwolf, 2001). To determine the origin of primitive streak precursor
However, the precise origin and identity of cells that cells, injections were made into the epiblast at the fol-
respond to the inductive influence of the PMZ, as well as lowing sites: overlying the midline of Koller’s sickle, at
the behaviors that such precursor cells exhibit to gener- the 3, 9, and 12 o’clock positions of the epiblast and at
ate the incipient primitive streak, remain largely un- equidistantly spaced intervals (unilaterally or bilaterally)
known. along the mediolateral extent of the epiblast overlying
Hensen’s node, the organizer of the avian embryo, first Koller’s sickle, either along its rostral or caudal borders
can be identified morphologically during the linear prim- (Fig. 1). Our results are summarized in Figure 8a–d.
itive streak stage. Fate mapping studies demonstrated Injections placed in the midline of the epiblast over-
that Hensen’s node at the fully elongated streak stage is lying Koller’s sickle (labeling essentially its entire rostro-
fated to form midline ectodermal, mesodermal, and caudal extent) labeled the midline dorsal surface cells of
endodermal derivatives (Garcia-Martinez et al., 1993; the primitive streak (for a description of these cells, see
Schoenwolf et al., 1992; Selleck and Stern, 1991). Re- Lawson and Schoenwolf, 2001), together with the un-
ports based largely on fate mapping and the expression derlying hypoblast and mesodermal cells adjacent to the
pattern of the homeobox gene, goosecoid, indicate that hypoblast (Fig. 2a–c). Additionally, such injections la-
Hensen’s node precursor cells exist as two populations beled the midline of the caudal area opaca. It is impor-
prior to the appearance of the primitive streak (Hatada tant to note that these so-called midline injections la-
and Stern, 1994; Izpisúa-Belmonte et al., 1993; Streit et beled clusters of cells that spanned the midline rather
al., 2000). Both populations seem to migrate to the than being restricted to the midline. It is also important
center of the embryo, where they come together to to note that by the stage illustrated in Figure 2b, labeled
colonize Hensen’s node (Streit et al., 2000). However, cells were already ingressing through the primitive
details of their positions and movement in relation to streak and moving laterally. Finally, not all cells in the
progression of the primitive streak are yet to be clarified. dorsoventral extent of the primitive streak are labeled
This study had two aims. First, we determined by fate (Fig. 2c); thus, more lateral cells moved into the primi-
mapping the origin of cells that contribute to the incip- tive streak during its formation, supporting the idea that
ient primitive streak. By labeling epiblast cells overlying formation of the primitive streak involves convergent
the mediolateral extent of Koller’s (Rauber’s) sickle (the extension (among other morphogenetic movements).
latter is defined as a sickle-shaped ventralward projec- Injections placed in the epiblast moved during the
tion from the epiblast located within the caudal area next few hours either centrifugally or centripetally de-
pellucida, just rostral to the posterior marginal zone; pending on their position: 12 o’clock injections moved
Khaner, 1993) with fluorescent dyes at stages XI–XIII, directly rostrally (centrifugally) with progression of the
and following their subsequent movement over time, we primitive streak, whereas 3 and 9 o’clock injections
show that the cells that contribute to the incipient prim- moved medially (centripetally) toward the progressing
itive streak reside in the middle two-thirds of the epiblast streak. Such epiblast injections never contributed la-
190 LAWSON AND SCHOENWOLF
FIG. 1. Scheme showing the locations of the injection sites in embryos at stages XI–XIII. The outer ring indicates the area opaca, the inner
disc, the area pellucida and the curvature at the bottom (i.e., caudal end) of the area pellucida, Koller’s sickle. The marginal zone is not
shown. Koller’s sickle lies within the area pellucida just rostral to the posterior (caudal) marginal zone. In one set of experiments on prestreak
embryos, injections were placed in the epiblast overlying the midline of Koller’s sickle (large white X), and in the epiblast of the area pellucida
at the 3, 9, and 12 o’clock positions (small black Xs). In another set of experiments on prestreak embryos, injections were placed in the
epiblast overlying the entire mediolateral extent of Koller’s sickle, either positioned along its rostral border or its caudal border. The oblique
lines indicate the angle (approximately 30°) subtended by the mediolateral extent of the portion of the epiblast overlying Koller’s sickle that
contributed to the formation of the primitive streak.
FIG. 2. Injections in the midline of the epiblast overlying Koller’s sickle and in the epiblast at 3, 9, and 12 o’clock positions in an embryo
at stage XIII. Whole-mounts (a, b) and a transverse paraffin section after peroxidase immunocytochemistry (c). (a) Locations of the injected
dye at time zero; the arrowhead marks the midline of the epiblast overlying Koller’s sickle. (b) Locations of the injected dye at 4 h of
reincubation. Labeled cells in the epiblast overlying the rostral border of the midline of Koller’s sickle moved rostrally during the 4 h of
reincubation to populate much of the rostrocaudal extent of the primitive streak. Labeled epiblast cells at the 3 and 9 o’clock positions
moved medially, whereas those at the 12o’clock position moved rostrally. (c) A transverse section through level c in b. The midline dorsal
surface cells of the primitive streak (arrowhead), the hypoblast and a few mesodermal cells adjacent to the hypoblast are labeled. Scale bar,
250 mm in a and b; 50 mm in c.
FIG. 3. Formation of the primitive streak in embryos in which the epiblast overlying the rostral border of Koller’s sickle was labeled with
DiI/CRSE at stage XII. Following reincubation, embryos were processed for peroxidase immunocytochemistry. (a) Within 6 h of reincubation,
labeled cells had migrated rostrally to form the triangularly-shaped, initial primitive streak characteristic of embryos at stage 2 (the initial
primitive streak stage). Inset: a similar embryo at time zero showing three injections in the epiblast overlying the rostral border of Koller’s
sickle. (b) Labeled cells from the epiblast overlying the rostral border of Koller’s sickle participated in the progression of the primitive streak.
Injections were placed only on one side of the midline in this embryo. (c, d) Transverse sections through levels c and d, respectively, in b.
Labeled cells were present on only the injected side of the midline (arrow marks the midline) throughout the rostrocaudal extent of the stage
3c primitive streak, except at its most rostral end (d) where Hensen’s node was just beginning to form. Scale bar, 250 mm in a and b; 50
mm in c and d.
PRIMITIVE STREAK FORMATION 191
beled cells to the primitive streak during its formation both within the midline and lateral portions of the epi-
and early progression, demonstrating conclusively that blast overlying the rostral border of Koller’s sickle.
the incipient primitive streak does not arise from a ran- In contrast to injections placed along the rostral bor-
dom delamination of epiblast cells followed by a wide- der of the epiblast overlying Koller’s sickle, injections
scale caudomedial migration to the forming streak. placed along its caudal border moved into area opaca,
Injections placed along the mediolateral extent of the without contributing to the primitive streak (Fig. 5a, b).
epiblast overlying Koller’s sickle were localized either to Sections revealed that such injections labeled the epi-
its rostral or caudal border. Those placed along its rostral blast in this area (data not shown).
border, labeled cells that populated the entire incipient
primitive streak (Fig. 3a, inset), with those placed uni- Mapping Hensen’s Node Precursor Cells
laterally remaining on that side of the midline as the To determine the origin and migratory routes of Hens-
primitive streak formed and underwent progression (Fig. en’s node precursor cells during primitive streak forma-
3b, c). Bilateral injections placed as far laterally in the tion and progression, injections were made at the tip of
epiblast overlying Koller’s sickle so as to subtend an the incipient primitive streak (area “c,” Fig. 7a), and also,
angle of about 30° (Fig. 1), contributed to the primitive in the epiblast immediately rostral to it (area “a,” Figs. 6a,
streak, but more lateral injections did not (rather, they 7a) in stage 2, 3a/b, and 3c embryos. Our results are
marked the caudal epiblast of the area pellucida). More- summarized in Figure 8c9–d9.
over, examination of embryos at stage 3c that contained Following about 4 h of reincubation, labeled cells in
injections placed along the mediolateral extent of the area “a” had moved rostrally, concomitant with progres-
epiblast overlying Koller’s sickle showed that the ex- sion of the primitive streak, but now these cells popu-
treme rostral end of the elongating primitive streak was lated the rostral end of the streak, constituting not only
not labeled (Fig. 3d), suggesting that this portion of the epiblast cells but essentially the entire population of area
primitive streak receives a contribution of cells from “c” cells at this stage (Fig 6b). Upon further incubation
outside of the region of the epiblast overlying Koller’s (15–24 h), such labeled cells colonized Hensen’s node
sickle (see following section). (Fig. 6c), from which they subsequently contributed to
Injecting the epiblast overlying Koller’s sickle of the the prechordal plate and rostral extraembryonic
same embryo with two types of markers that fluoresced endoderm and midline axial structures (Garcia-Martinez
at different wavelengths, revealed that the primitive et al., 1993). Because area “a” epiblast cells incorporated
streak precursor cells intermixed mediolaterally, provid- into the rostral end of the primitive streak between
ing direct evidence that they underwent a convergent- stages 2 and 3a/b, to determine whether additional epi-
extension movement during their rostral displacement blast cells contributed to Hensen’s node during subse-
(Fig. 4a, b, inset). The latter occurred in an orderly quent stages, we labeled cells in both areas “a” and “c” in
fashion with precursor cells located laterally along the same embryo at stage 3a/b with two types of markers
Koller’s sickle following an oblique course, contributing that fluoresced at different wavelengths. Such double
to outer and rostral portions of the incipient primitive labeling revealed conclusively that within 4 h of reincu-
streak, whereas those closer to the midline following a bation, labeled cells from area “c” had moved rostrally at
less oblique course, giving rise to inner and caudal por- a greater rate than area “a” cells, overlapping those from
tions of the incipient primitive streak. This orderly move- area “a” (Fig. 7b). These results support the view, sug-
ment is sufficient to account for the triangular shape of gested from previous descriptive studies using scanning
the incipient primitive streak. Thus, the incipient prim- electron microscopy (Lawson and Schoenwolf, 2001),
itive streak was formed from precursor cells located that late progression of the primitive streak and forma-
FIG. 4. A stage XI embryo with DiI/CRSE (red) and CFSE (green) injections in the epiblast overlying the rostral border of Koller’s sickle. The
oblique lines mark the approximate midline of the blastoderm and incipient primitive streak. (a) The numbers 1–7 indicate the locations of
the injected dye at time zero. (b) Within 4 h of reincubation, the new positions of the injected dyes (i.e., 19–79) show that labeled cells
underwent convergent extension to initiate formation of the primitive streak (note: some injection sites split into multiple sites). Inset: 8 h
of reincubation. The figure has been rotated so that the primitive streak (arrow) extends vertically down the midline. Scale bar, 400 mm in
a and b.
FIG. 5. A stage XIII embryo with DiI/CRSE injections in the epiblast overlying the caudal border of Koller’s sickle. (a) Injections at time zero
(because of the intense fluorescence, many individual injections appear fused as a single one). Arrows indicate the unlabeled epiblast
overlying the rostral border of Koller’s sickle. (b) After 4 h of reincubation, labeled cells had migrated caudally into the area opaca, without
contributing to the primitive streak (lateral borders of the primitive streak are indicated by vertical lines). Scale bar, 400 mm in a and b.
FIG. 6. DiI/CRSE injection of area “a” (epiblast just rostral to the incipient primitive streak) of a stage 2 embryo. White lines indicate the lateral
sides of primitive streak, which converge just caudal to the injection site. (a) Injection at time zero. (b) Following 4 h of reincubation, labeled
cells occupy the tip of the primitive streak. (c) By 24 h of reincubation (by stage 6), labeled cells had populated Hensen’s node (arrowhead)
and contributed to the prechordal plate and rostral extraembryonic endoderm (asterisk) and the midline axis. Scale bar, 400 mm in a–c.
FIG. 7. Double labeling of areas “a” (green) and “c” (red) at stage 3a/b. (a) Injection at time zero. (b) Cells from areas “a” and “c” migrated
rostrally, with those of area “c” overlapping those from level “a” (yellow) within 4 h of reincubation; arrowhead, incipient Hensen’s node. (c)
By 15 h of reincubation, both populations of precursor cells had populated Hensen’s node (arrow) and contributed to prechordal plate and
rostral extraembryonic endoderm and the midline axis (the red spot to the right of the green head process at about the 1 o’clock position
is artifactual i.e., dye that leaked between the epiblast and underlying vitelline membrane). Scale bar, 400 mm in a–c.
192 LAWSON AND SCHOENWOLF
that express Sonic hedgehog; further studies are under-
way to characterize in more detail the populations of
cells derived from areas “a” and “c”).
DISCUSSION
There are three principal findings of the present study to
discuss: (1) primitive streak precursor cells originate
from the epiblast overlying the rostral border of Koller’s
sickle, with the midline dorsal surface cells of the prim-
itive streak arising from the midline of the epiblast over-
lying Koller’s sickle and the deeper and more lateral
primitive streak cells arising more laterally within the
epiblast overlying the sickle, from an arch subtending
about 30°; (2) the incipient primitive streak is formed in
part through the convergent extension of its precursor
cells; and (3) during progression of the primitive streak,
Hensen’s node continuously receives contributions from
both epiblast cells located rostral to the primitive streak
(i.e., from area “a” as defined at stage 2 and at stage 3a/b)
and cells moving rostrally within the rostral end of the
elongating streak (i.e., from area “c”).
Origin of the Primitive Streak
Our labeling studies of the epiblast overlying Koller’s
sickle and the epiblast at the 3, 9, and 12 o’clock posi-
tions demonstrate that prior to the appearance of the
initial primitive streak, its precursor cells are localized
within the epiblast overlying the middle two-thirds
Koller’s sickle along its rostral border. Neither more
rostral epiblast cells, at the 3, 9, and 12 o’clock positions,
FIG. 8. Diagram summarizing the results of our labeling experi- nor more caudal epiblast cells, overlying the caudal bor-
ments. The outer ring indicates the area opaca, the inner disc, the der of Koller’s sickle, contribute to the formation of the
area pellucida, and the curvature at the bottom (i.e., caudal end) of
the area pellucida, Koller’s sickle. The marginal zone is not shown. initial primitive streak. Thus, our findings are in contrast
Koller’s sickle lies within the area pellucida just rostral to the caudal with the idea that the initial primitive streak arises from
(posterior) marginal zone. (a–d) Fate maps of the epiblast overlying the aggregation of cells originally distributed randomly
Koller’s sickle showing the origins of primitive-streak cells at stage throughout the epiblast (Stern 1992; Stern and Canning,
XII (a), and their locations at stages 2 (b), 3a (c), and 3c (d). With the 1990). Instead, they support the labeling studies of Bach-
exception of the epiblast overlying the midline of Koller’s sickle,
which contributes to the mid-dorsal surface cells of the primitive varova and co-workers (1998), which suggested a local-
streak, the epiblast overlying Koller’s sickle contributes cells to the ized origin of the primitive streak from the caudal mid-
entire dorsoventral thickness of the primitive streak. (c*–d*) Fate line of the area pellucida, as well as our previous labeling
maps of areas “a” (i.e., epiblast just rostral to the primitive streak) and descriptive studies (Lawson and Schoenwolf, 2001).
and “c” (i.e., rostral tip of the primitive streak) showing the origins of
Hensen’s node cells at stage 3a (c ) and stage 3c (d ) and their Moreover, our results do not support the view that the* *
locations at stages 3c and 3d/4, respectively. Note that area “c” primitive streak is formed exclusively from epiblast cells
cells migrate rostrally to intermix with area “a” cells, forming by localized in the caudal midline (Vakaet 1970; Wei and
stage 3d/4 Hensen’s node. Mikawa, 2000). Instead, we show that the epiblast over-
lying Koller’s sickle contributes to the primitive streak
along an arc that subtends an angle of about 30°—thus,
tion of Hensen’s node occurs both by the accretion of the middle two-thirds of the epiblast overlying Koller’s
newly delaminated epiblast cells to its rostral end and a sickle participates in primitive streak formation, not just
rostral movement of cells from its elongating tip. During its midline. Furthermore, epiblast cells located in the
subsequent development, cells in Hensen’s node moved exact midline overlying the rostral border of Koller’s
rostrally, contributing to midline cells. The red cells sickle prior to the appearance of the primitive streak,
contributed to the midline near the rostral end of the contribute to only a portion of the incipient streak,
neural plate, and the green cells contributed to the namely, its midline dorsal surface cells and its ventral
midline from about the mesencephalon level caudalward mesendoderm cells (see also Lawson and Schoenwolf,
(Fig. 7c). Although molecular labels were not used in 2001).
this study, based on their characteristic position, the In summary, our present findings provide a more ac-
green cells would be expected to form the midline cells curate fate map (Figure 8a–d) of the distribution of the
PRIMITIVE STREAK FORMATION 193
epiblast cells overlying Koller’s sickle that contribute to tive streak, rather than a triangular one (that is, the actual
the formation of the initial primitive streak, advancing shape of the incipient streak).
our understanding of the earliest events of formation of This study also demonstrates that subsequent to the
the primitive streak. establishment of the initial primitive streak, the precur-
sor cells also participate in its progression until the
Role of Convergent Extension in Formation of definitive (stage 4) primitive streak is formed. This un-
the Primitive Streak derscores the complexity of the process, which as sug-
By tagging the primitive streak precursor cells and gested in previous studies, involves the accretion of cells
following their movement over time, we present evi- to the growing rostral end of the primitive streak (Law-
dence that formation of the triangularly shaped, initial son and Schoenwolf, 2001), convergent-extension move-
(stage 2) primitive streak involves a convergent-exten- ments mediated by cell–cell intercalation (Lawson and
sion movement of its precursor cells. Convergent exten- Schoenwolf, 2001) and rostrocaudally oriented cell divi-
sion is defined as the narrowing of one dimension (e.g., sion (Wei and Mikawa, 2000).
the width or transverse extent) of a column of cells and The expression pattern of what we have described as
its concomitant elongation in a second dimension (e.g., group 1A gastrulation genes, namely Wnt 8c, Slug, Vg1,
the length or rostrocaudal extent) perpendicular to the and Nodal (Lawson et al., 2001), seems to correlate with
narrowing dimension. Over an 8-h period, labeled our current map of the origin of the primitive streak
groups of cells arranged mediolaterally at the outset of precursor cells and their rostral movement to establish
the experiment become interdigitated along the midline the incipient streak. We have defined such genes as
of the forming primitive streak (Figs. 4a, b, inset). This those that are expressed in the epiblast overlying
pattern of cell labeling is supported by studies using in Koller’s sickle and subsequently throughout the pro-
situ hybridization to detect gene expression in the form- gressing primitive streak, except at its extreme rostral
ing primitive streak (e.g., Lawson et al., 2001), providing end (Vg1 is an exception to this exception, as it is
further evidence of convergent extension during forma- expressed in the most rostral end of the progressing
tion of the primitive streak. We show here that cells primitive streak). Our results thus provide additional
from lateral portions of the precursor area of the epiblast evidence to support the suggestion that group 1A genes
overlying Koller’s sickle follow an oblique course to act in establishing the form and/or functional activities
contribute to more rostral portions of the streak, of the primitive streak (Lawson et al., 2001). An upregu-
whereas those closer to the midline contribute to more lation of the zinc-finger gene, Slug, along Koller’s sickle,
caudal portions of the streak. The movement occurring for example, may initiate the transformation of epiblast
during formation of the initial primitive streak is similar cells in the precursor region to primitive streak precur-
to that occurring during neurulation (particularly, shap- sor cells. Further evidence in support of the role of
ing of the neural plate; Smith and Schoenwolf, 1997) and group 1A genes in primitive streak formation comes
primitive streak progression (Lawson and Schoenwolf, from studies that show that an ectopic expression of
2001). In the latter case, measurements of the length and some of these genes, notably Vg1 (Shah et al., 1997) and
width of the essentially rectangular primitive streak dur- chordin (Streit et al., 1998), induces the formation of a
ing its progression show clearly that coordinated nar- primitive streak. The results of the this study provide an
rowing and lengthening occur. Such a relationship is important foundation upon which future studies on the
more difficult to quantify with an epiblast area that molecular mechanisms underlying primitive streak for-
begins as sickle-shaped and then transforms to triangu- mation can be built.
lar. Nevertheless, such shape changes are consistent
with a process of convergent extension. Thus, collec- Origin and Definition of Hensen’s Node, the
tively our results support the idea that formation of the Organizer of the Avian Embryo
primitive streak is driven at least in part by convergent It is well known that the rostral end of the avian
extension movements. primitive streak functions as an organizer when trans-
Although our results support the view of Eyal-Giladi planted to ectopic sites (Smith and Schoenwolf, 1998).
and co-workers (1992) on the origin of primitive streak Previous studies have shown that Hensen’s node, the
precursor cells, they are not in agreement with their spherical swelling present at the rostral end of the fully
model on how the precursor cells move rostrally to elongated (stage 3d/4) primitive streak, arises from two
establish the incipient streak. According to their model, populations of progenitor cells called central and poste-
the primitive streak forms in two stages. In the first stage rior cells (Streit et al., 2000). We have expanded these
of their model, cells located in the caudal midline move findings to show that the posterior progenitor cells arise
rostrally to occupy the most rostral portion of the prim- from only the midline of the epiblast overlying Koller’s
itive streak, with precursor cells located more laterally sickle, and not from more lateral sickle cells, and that the
moving to a midline position to replace the midline cells central progenitor cells arise continuously during pro-
that have moved rostrally. In the second stage of their gression of the primitive streak (i.e., during stages 2–3c)
model, the new midline cells move rostrally to populate from the epiblast just rostral to the elongating tip of the
the caudal portion of the streak. Their model, however, primitive streak (i.e., from the area we designate at each
would predict the formation of a linear incipient primi- stage as area “a”). Additionally, we have tracked the two
194 LAWSON AND SCHOENWOLF
populations of precursor cells throughout the process of Koller’s sickle between the DiI/CRSE injection sites (Fig.
primitive streak progression. Our results confirm the 4a). In six additional embryos, DiI/CRSE injections were
data of Streit and co-workers (2000) by directly demon- placed in the epiblast overlying the caudal border of
strating that the two populations of progenitor cells Koller’s sickle.
intermix to form the node during late progression, and
we agree with their suggestion that the term Hensen’s Mapping Hensen’s Node Precursor Cells
node should be reserved for the structure formed by For mapping Hensen’s node precursor cells, two sets
such intermixing. of experiments were done in embryos at stages 2–3c. A
total of 65 embryos was used for these experiments. In
MATERIALS AND METHODS the first set, a mixture of DiI/CRSE was injected at area
“a” (the epiblast just rostral to the primitive streak; Gar-
Fertilized chicken eggs were incubated at 38°C in hu- cia-Martinez et al., 1993), whereas in the second set, the
midified incubators to obtain embryos at stages XI–XIII DiI/CRSE was injected at area “c” (the rostral end of the
(Eyal-Giladi and Kochav, 1976) or stages 2–3a (Hamburg- primitive streak; Garcia-Martinez et al., 1993) and an
er and Hamilton, 1951; substaging of stage 3 embryos as additional injection of CSFE was placed at area “a.” Em-
described by Schoenwolf et al., 1992; also see Darnell et bryos were reincubated for periods ranging from 4 to
al., 1999). Culture dishes and embryos were prepared as 15 h prior to fixation.
described by Darnell and Schoenwolf (2000) for modi-
fied New (1955) culture. Embryos were injected with a Immunocytochemistry
mixture of 5-carboxytetramethylrhodamine, succinimi- All embryos labeled with the DiI/CRSE solution were
dyl ester (CRSE, Molecular Probes Inc., Eugene, OR, processed for immunocytochemistry as described by Pa-
USA) and 1,19-dioctadecyl-3, 3, 39-tetramethylindocarbo- tel and co-workers (1989), except that they were fixed
cyanine perchlorate (DiI, Molecular Probes Inc.), as de- with 4% paraformaldehyde in PBS and the peroxidase
scribed previously (Darnell et al., 2000). The dye solu- reaction was enhanced by the addition of 2% CoCl2 and
tion was injected with the aid of a Picrospritzer II Ni(NH4)2(SO4) per ml of DAB-PBT. The antibodies used
(General Valve Corp., Fairfield, NJ, USA) and a microma- were anti-rhodamine (rabbit IgG polyclonal, primary;
nipulator. All embryos were examined immediately after Molecular Probes Inc.) and horseradish peroxidase-con-
injection with fluorescence microscopy, and embryos in jugated goat anti-rabbit IgG (secondary antibody; Boehr-
which injections were too large or too small, or were inger Mannheim Corp., Indianapolis, IN, USA). The em-
misplaced, were discarded. bryos were first examined as whole-mounts; selected
Mapping Primitive Streak Precursor Cells ones were then processed for paraffin histology as de-
scribed in the following section.
For mapping primitive streak precursor cells, two sets
of experiments were done in embryos at stages XI–XIII. Paraffin Histology
The first set (involving 55 embryos) was designed to Embryos for histology were dehydrated with an as-
assess the contribution of cells located in the midline of cending graded series of ethanol up to 100% ethanol.
the epiblast overlying Koller’s sickle to the incipient They were taken through two 5-min changes of Histosol,
primitive streak, and also to test the hypothesis that cells infiltrated with Paraplast and embedded in fresh Para-
scattered randomly throughout the epiblast contribute plast. Sectioning was done at 10 mm.
to primitive streak formation. Small quantities of DiI/
CRSE were injected at four sites, namely, the midline of ACKNOWLEDGMENTS
the epiblast overlying Koller’s sickle, and the epiblast of
the area pellucida at the 3, 9, and 12 o’clock positions We thank Jean-Francois Colas and Raj Ladher for their
(Fig. 1). Embryos were then reincubated for 5 h. helpful comments on the manuscript.
The second set (involving 65 embryos) of experiments
was designed to determine whether the epiblast of LITERATURE CITED
Koller’s sickle contributes cells to the incipient primitive
streak. The epiblast overlying Koller’s sickle was divided Bachvarova RF, Skromne I, Stern CD. 1998. Induction of primitive
into rostral and caudal halves along its mediolateral ex- streak and Hensen’s node by the posterior marginal zone in the
tent. Three to eight injections of DiI/CRSE were placed early chick embryo. Development 125:3521–3534.
either along its rostral (60 embryos) or caudal (five em- Bellairs R. 1986. The primitive streak. Anat Embryol 174:1–14.
bryos) borders. In the former, the dye often extended Darnell DK, Schoenwolf GC. 2000. Culture of avian embryos. In: Tuan
RS, Lo CW, editors. Methods in molecular biology, Vol. 135:
slightly rostrally to include adjacent epiblast. Following Developmental biology protocols (Vol. 1). Totowa, NJ: Humana
injection, embryos were examined as described and then Press. p 31–38.
reincubated for periods ranging from 4 to 8 h. In a few Darnell DK, Stark MR, Schoenwolf GC. 1999. Timing and cell interac-
embryos injected in the epiblast overlying the rostral bor- tions underlying neural induction in the chick embryo. Develop-
der of Koller’s sickle, additional injections with 5-, 6-car- ment 126:2505–2514.Darnell DK, Garcia-Martinez V, Lopez-Sanchez C, Yuan S, Schoenwolf
boxyfluorescein diacetate, succinimidyl ester (CFSE, Molec- GC. 2000. Dynamic labeling techniques for fate mapping, testing
ular Probes Inc.) were placed in the epiblast overlying cell commitment, and following living cells in avian embryos. In:
PRIMITIVE STREAK FORMATION 195
Tuan RS, Lo CW, editors. Methods in molecular biology, Vol. 135: Lemaire L, Kessel M. 1997. Gastrulation and homeobox genes in chick
Developmental biology protocols (Vol. 1). Totowa, NJ: Humana embryos. Mech Dev 67:3–16.
Press. p 305–321. New DAT. 1955. A new technique for the cultivation of the chick
Eyal-Giladi H, Khaner O. 1989. The chick’s marginal zone and primitive embryo in vitro. J Embryol Exp Morphol 3:326–331.
streak formation. II. Quantification of the marginal zone’s poten- Patel NH, Martin-Blanco E, Coleman KG, Poole SJ, Ellis MC, Kornberg
cies—temporal and spatial aspects. Dev Biol 134:215–221. TB, Goodman CS. 1989. Expression of engrailed proteins in ar-
Eyal-Giladi H, Kochav S. 1976. From cleavage to primitive streak for- thropods, annelids, and chordates. Cell 58:955–968.
mation: A complementary normal table and a new look at the first Schoenwolf GC. 1979. Histological and ultrastructural observations of
stages of the development of the chick. I. General morphology. tail bud formation in the chick embryo. Anat Rec 193:131–148.
Dev Biol 49:321–337. Schoenwolf GC, Smith JL. 2000. Gastrulation and early mesodermal
Eyal-Giladi H, Debby A, Harel N. 1992. The posterior section of the
patterning in vertebrates. In: Tuan RS, Lo CW, editors. Methods in
chick’s area pellucida and its involvement in hypoblast and prim-
itive streak formation. Development 116:819–830. molecular biology, Vol. 135: Developmental biology protocols
Garcia-Martinez V, Alvarez IS, Schoenwolf GC. 1993. Locations of the (Vol. I). Totowa, NJ: Humana Press. p 113–125.
ectodermal and non-ectodermal subdivisions of the epiblast at Schoenwolf GC, Garcia-Martinez V, Dias MS. 1992. Mesoderm move-
stages 3 and 4 of avian gastrulation and neurulation. J Exp Zool ment and fate during avian gastrulation and neurulation. Dev
267:431–446. Dynamics 193:235–248.
Hamburger V, Hamilton HL. 1951. A series of normal stages in the Selleck MAJ, Stern CD. 1991. Fate mapping and cell lineage analysis of
development of the chick embryo. J Morph 88:49–92. Hensen’s node in the chick embryo. Development 112:615–626.
Hatada Y, Stern CD. 1994. A fate map of the epiblast of the early chick Shah SB, Skromne I, Hume CR, Kessler DS, Lee KJ, Stern CD, Dodd J.
embryo. Development 120:2879–2889. 1997. Misexpression of chick Vg1 in the marginal zone induces
Hume CR, Dodd J. 1993. Cwnt-8C: A novel Wnt gene with a potential primitive streak formation. Development 124:5127–5138.
role in primitive streak formation and hindbrain organization. Smith JL, Schoenwolf GC. 1997. Mechanisms of neurulation: Coming to
Development 119:1147–1160. closure. Trends Neurosci 20:510–517.
Izpisúa-Belmonte JC, Robertis EMD, Storey KG, Stern CD. 1993. The Smith JL, Schoenwolf GC. 1998. Getting organized. Curr Topics Dev
homeobox gene goosecoid and the origin of organizer cells in the Biol 40:79–110.
early chick blastoderm. Cell 74:645–659. Stern CD. 1992. Mesoderm induction and formation of the embryonic
Khaner O. 1993. Axis determination in the avian embryo. Curr Topics axis in the chick. Sem Dev Biol 3:65–71.
Dev Biol 28:155–180. Stern CD, Canning DR. 1988. Gastrulation in birds: A model system for
Khaner O. 1998. The ability to initiate an axis in the avian blastula is the study of animal morphogenesis. Experentia 44:651–657.
concentrated mainly at a posterior site. Dev Biol 194:257–266. Stern CD, Canning DR. 1990. Origin of cells giving rise to mesoderm
Khaner O, Eyal-Giladi H. 1986. The embryo-forming potential of the and endoderm in chick embryo. Nature 343:273–275.
posterior marginal zone in stages X through XII of the chick. Dev
Streit A, Lee KJ, Woo I, Roberts C, Jessell TM, Stern CD. 1998. Chordin
Biol 115:275–281.
Khaner O, Eyal-Giladi H. 1989. The chick’s marginal zone and primitive regulates primitive streak development and the stability of in-
streak formation. I. Coordinative effect of induction and inhibi- duced neural cells, but is not sufficient for neural induction in the
tion. Dev Biol 134:206–214. chick embryo. Development 125:507–519.
Khaner O, Mitrani E, Eyal-Giladi H. 1985. Developmental potencies of Streit A, Berliner AJ, Papanayotou C, Sirulnik A, Stern CD. 2000.
the area opaca and the marginal zone areas of the early chick Initiation of neural induction by FGF signalling before gastrula-
blastoderms. J Embryol Exp Morphol 89:235–241. tion. Nature 406:74–80.
Lawson A, Schoenwolf GC. 2001. New insights into critical events of Vakaet L. 1970. Cinephotomicrographic investigations of gastrulation
avian gastrulation. Anat Rec 262:238–252. in the chick blastoderm. Arch Biol 81:387–426.
Lawson A, Colas J-F, Schoenwolf GC. 2001. A classification scheme for Wei Y, Mikawa T. 2000. Formation of the avian primitive streak from
genes expressed during formation and progression of the avian spatially restricted blastoderm: Evidence for polarized cell division
primitive streak. Anat Rec 262:221–226. in the elongating streak. Development 127:87–96.