Download McClay et al.

5113
Development 127, 5113-5122 (2000)
Printed in Great Britain © The Company of Biologists Limited 2000
DEV5407
A micromere induction signal is activated by β-catenin and acts through
Notch to initiate specification of secondary mesenchyme cells in the sea
urchin embryo
David R. McClay*, Robert E. Peterson, Ryan C. Range, Anne M. Winter-Vann and Michael J. Ferkowicz
Department of Biology, DCMB Group, Box 91000, Duke University, Durham, NC 27708, USA
*Author for correspondence (e-mail: [email protected])
Accepted 25 September; published on WWW 2 November 2000
SUMMARY
At fourth cleavage of sea urchin embryos four micromeres
at the vegetal pole separate from four macromeres just
above them in an unequal cleavage. The micromeres have
the capacity to induce a second axis if transplanted to the
animal pole and the absence of micromeres at the vegetal
pole results in the failure of macromere progeny to specify
secondary mesenchyme cells (SMCs). This suggests that
micromeres have the capacity to induce SMCs.
We demonstrate that micromeres require nuclear βcatenin to exhibit SMC induction activity. Transplantation
studies show that much of the vegetal hemisphere is
competent to receive the induction signal. The micromeres
induce SMCs, most likely through direct contact with
macromere progeny, or at most a cell diameter away. The
induction is quantitative in that more SMCs are induced by
four micromeres than by one. Temporal studies show that
the induction signal is passed from the micromeres to
macromere progeny between the eighth and tenth cleavage.
If micromeres are removed from hosts at the fourth
cleavage, SMC induction in hosts is rescued if they later
receive transplanted micromeres between the eighth and
tenth cleavage. After the tenth cleavage addition of
induction-competent micromeres to micromereless
embryos fails to specify SMCs.
For macromere progeny to be competent to receive the
micromere induction signal, β-catenin must enter
macromere nuclei. The macromere progeny receive the
micromere induction signal through the Notch receptor.
Signaling-competent micromeres fail to induce SMCs if
macromeres express dominant-negative Notch. Expression
of an activated Notch construct in macromeres rescues
SMC specification in the absence of induction-competent
micromeres.
These data are consistent with a model whereby βcatenin enters the nuclei of micromeres and, as a
consequence, the micromeres produce an inductive ligand.
Between the eighth and tenth cleavage micromeres induce
SMCs through Notch. In order to be receptive to the
micromere inductive signal the macromeres first must
transport β-catenin to their nuclei, and as one consequence
the Notch pathway becomes competent to receive the
micromere induction signal, and to transduce that signal.
As Notch is maternally expressed in macromeres,
additional components must be downstream of nuclear βcatenin in macromeres for these cells to receive and
transduce the micromere induction signal.
INTRODUCTION
Balzer, 1967, pp. 111-112; based upon Boveri, 1901). Boveri,
and later Horstadius (Horstadius, 1928), isolated fragments of
eggs using the subequatorial pigmented band on Paracentrotus
eggs as a landmark. They found that fertilized fragments of
eggs containing the vegetal-most cytoplasm were capable of
forming skeleton, but no other part of the egg had that
capability.
Boveri and Horstadius further studied the ‘regulatory
activity’ of micromeres. Micromere regulatory activity was
demonstrated in a classic experiment by Horstadius.
Micromeres recombined with animal halves of 8-cell stage
embryos induced the animal halves to produce endoderm. In
the absence of micromeres the animal halves developed into
dauer blastulae with no mesoderm or endodermal structures
The earliest known specification event in sea urchin embryos
occurs in micromeres. As a result of an unequal fourth
cleavage, micromeres arise at the vegetal pole below their large
sister macromeres. Experiments at the turn of the last century
demonstrated that cytoplasmic determinants localized to the
vegetal pole were the earliest potential cause of this
specification event. In 1901 Boveri recognized that “the area
nearest the vegetative pole possesses the greatest potentiality
to bring development completely to the pluteus stage. It is the
‘priority region’ where differentiation begins. And when
differentiation has begun, ‘from this center all other regions
are determined in their role by a regulatory action” (cited in
Key words: Notch, β-catenin, Induction, Specification, Sea urchin
5114 D. R. McClay and others
(Horstadius, 1939). Those experiments were interpreted to
suggest that one of the early functions of micromeres was to
provide an inductive signal to the macromere cells above them.
Horstadius further showed that a secondary axis could be
induced if he transplanted micromeres to the animal pole. This
experiment dramatically demonstrated the inductive capacity
of micromeres. For more than 75 years since that experiment
there has been interest in the mechanism of micromere
induction.
Recent experiments returned to the micromere inductive
capabilities at a molecular level. Micromeres were transplanted
to the animal pole, as Horstadius had done, and the induced,
ectopic, secondary gut expressed lineage markers just like
those expressed in the endogenous gut (Ransick and Davidson,
1993). Furthermore, Ransick also found that micromeres were
necessary for normal gut specification, at least for the early
expression of endoderm markers (Ransick and Davidson,
1995). In other recent experiments absence of micromeres led
to the absence of secondary mesenchyme cells (SMCs) in host
embryos (Sweet et al., 1999). This led Sweet et al. to propose
that an endogenous micromere signal induces SMCs. The
mechanism of that induction was not clear, though from
experiments by others (Sherwood and McClay, 1999),
signaling through the Notch pathway was suggested by Sweet
et al. as an obvious possibility.
Maternal Notch is distributed on every blastomere in early
cleavage. When SMCs are specified Notch disappears from
cells of the presumptive SMC lineage (Sherwood and McClay,
1997). This loss of Notch is the earliest known lineage marker
for the SMCs. Manipulation of Notch alters the number of
SMCs: if embryos express constitutively active Notch, extra
SMCs develop; conversely if one expresses dominant-negative
Notch in the embryo, SMCs are reduced in number or
eliminated altogether (Sherwood and McClay, 1999). These
data suggest that Notch signaling is involved in the SMC
specification.
Another early molecular event in vegetal specification
involves components of the Wnt pathway. Late in the fourth
cleavage β-catenin enters the nuclei of micromeres (Logan
et al., 1999). If nuclear entry of β-catenin is prevented, the
micromeres fail to become specified toward their normal
skeletogenic fate (Logan et al., 1999) (Wikramanayake et al.,
1998) (Emily-Fenouil et al., 1998), and in the absence of
nuclear β-catenin, markers show that micromeres become
specified as ectoderm (Logan et al., 1999). Micromeres without
nuclear β-catenin do not have the ability to induce a secondary
axis when transplanted to the animal pole of host embryos
(Logan et al., 1999). Thus β-catenin appears to be required for
micromere signaling competence. Furthermore, macromeres
lacking nuclear β-catenin fail to specify endoderm and
macromeres expressing excess β-catenin are vegetalized.
Given that both β-catenin and Notch appear to be involved
in early specification of vegetal pole tissues, and that the
micromere induction signal is required for endoderm and
mesoderm specification, we designed a series of experiments
to determine the nature of the micromere induction signal
pathway. The results of that effort show that β-catenin is
upstream of a micromere induction signal, and β-catenin is also
upstream of the macromere’s response to that induction signal.
We demonstrate that the micromere induction signal induces
SMCs at the eighth to tenth cleavage through activation of
Notch. If micromeres are eliminated, SMC specification is
rescued by expression of activated Notch signal in
macromeres. The possible existence of an earlier endoderminducing activity of micromeres is discussed.
MATERIALS AND METHODS
Animals
Adult Lytechinus variegatus were obtained from Jennifer Keller
(Beaufort, NC, USA) and from Susan Decker (Hollywood, FL, USA).
Gametes were harvested and fertilized in artificial sea water (ASW)
as described (Hardin et al., 1992). Embryos were cultured at 21-23°C.
mRNA preparation and injection into zygotes
All LvNotch, β-catenin and cadherin DNA constructs used in this
study have been previously described (Miller and McClay, 1997;
Sherwood and McClay, 1999) and were used as templates to generate
in vitro transcribed 5′ capped mRNAs using the T3 mMessage
mMachine kit (Ambion). The concentrations of mRNAs were
determined and the mRNAs were mixed with glycerol (40% v/v,
containing FITC-dextran 40 kD, 10%) for injection as described
(Sherwood and McClay, 1999). Eggs were passed (3-5×) through 102
µm Nitex cloth to remove the jelly coat, and were aligned on 60 mm
Petri dish covers treated with 1% protamine sulfate, fertilized and
injected as described (Sherwood and McClay, 1999). Embryos to be
dissected later were fertilized in 5 mM p-aminobenzoic acid (PABA)
to prevent fertilization membrane hardening.
Micromere transplants
Embryos to be dissected were transferred by mouth pipet to calciumfree SW. They were then inserted into Kiehart chambers (Kiehart,
1982) that were inverted relative to the original protocol. A joystick
micromanipulator was used to manipulate a glass needle as the
dissection tool. The needle was broken to make a blunt tip with an
inside diameter just smaller than the diameter of the cells to be
transplanted. Micromeres were removed by suction using a Gilmont
2 ml microsyringe containing silicon oil attached to the needle to
provide a suction action. For micromere transplantations, the four
micromeres were first removed from host embryos then donor
micromeres were harvested from either red (RITC-stained, as in
Logan et al., 1999) or green embryos (FITC coinjected with RNA)
and transplanted into position on the hosts. The color of the PMCs
later in development unambiguously identified the origin of the
transplanted micromere (thus preventing misinterpretation due to
inadvertent failure to remove all micromeres from the host embryo).
The embryo was slightly compressed between two coverslips for
several minutes to hold the donor cell(s) in place. The chamber was
then flooded gradually with ASW and the embryos were then
transferred to 96-well plates for culture in ASW.
Immunolocalization and image analysis
Late mesenchyme blastula embryos were fixed in glass depression
slides with cold methanol for 10 minutes. After fixation, the embryos
were washed 2× with PBS, blocked for 10 minutes in PBS/5% normal
goat serum (NGS) (Gibco/BRL), and incubated in primary antibody
for 2-3 hours at room temperature or overnight at 4°C, again with the
antibody diluted in PBS/5% NGS. Embryos were then washed 3×
with PBS/5% NGS, incubated in secondary antibody (Cy5-, Cy2- or
CY3-conjugated; Jackson Immunoresearch Laboratories) for 2 hours,
washed 3× with PBS, and mounted in 7:3 (v/v) glycerol:PBS. βcatenin was localized with a guinea pig α-β-catenin pAb as
previously described (Logan et al., 1999; Miller and McClay, 1997).
Notch was identified with an antibody that stains the extracellular
EGF-repeat region of the molecule (Sherwood and McClay, 1997),
and SMCs were stained with SMC-1, a marker that specifically stains
Micromere induction signal 5115
this lineage shortly after it is first specified (Sweet et al., 1999).
Images were obtained by sequential confocal sections of embryos
using a Zeiss 410 laser-scanning microscope. All images were
collected at 2 µm intervals. To calculate the area of SMC territory
the length of SMC specified tissue was measured in each section of
a complete stack. Given that each image was 2 µm apart the area was
then calculated. In the images below, blue was the assigned
pseudocolor of cells labeled with fluorescein dextran or stained with
Cy2 secondary antibody; red was the assigned pseudocolor for cells
labeled with rhodamine or stained with Cy3 secondary antibody;
green was the assigned pseudocolor for cells stained with Cy5
secondary antibody.
RESULTS
Micromeres are necessary for induction of SMCs
If one removes micromeres no SMCs are specified (Sweet et
al., 1999). Fig. 1 confirms this and shows further that if one
removes micromeres and then transplants donor micromeres
from another embryo, the ability to make SMCs is rescued.
Donor micromeres were labeled with rhodamine
isothiocyanate (Logan and McClay, 1997). Host embryos were
unlabeled. Host micromeres were removed at the 16-cell stage
and replaced with red donor micromeres (Fig. 1C), or not
replaced (Fig. 1B). Fig. 1A shows a control embryo with a
normal area of SMC specification at the vegetal plate. Notch
disappears from presumptive SMCs between the eighth and
tenth cleavage, making it a convenient early marker of SMC
specification (Sherwood and McClay, 1997; Sherwood and
McClay, 1999) (Fig. 1A,C). The Notch disappearance involves
endocytosis of the extracellular region, presumably as part of
the SMC specification process (Sherwood and McClay, 1999).
Apical Notch staining at this time is observed lateral to the
SMCs in a ring around the vegetal plate. Lineage marking
shows these cells to be presumptive endoderm (Sherwood and
McClay, 1997; Sherwood and McClay, 1999). If micromeres
are removed at the 16-cell stage, the embryo fails to activate
Notch signaling, no SMCs are specified at the vegetal pole, and
instead the cells express apical Notch at that location,
indicating that endoderm has been specified in the absence of
SMC induction (Fig. 1B). The requirement for micromeres is
indicated in Figs 1A,C in that if present, or returned to
micromereless embryos, (red donor cells in Fig. 1C), SMCs are
specified. Thus micromeres are necessary for induction of
SMCs.
Activation of β-catenin and the presence of
micromeres induce SMCs
Loss of nuclear β-catenin (Logan et al., 1999) or a block in
Notch signaling with a dominant-negative form of Notch,
results in few, if any, SMCs (Sherwood and McClay, 1999).
We concluded that both β-catenin and Notch are necessary for
SMC specification. These earlier experiments did not indicate
which cells require β-catenin signaling and which cells receive
the Notch signal, since the constructs are expressed
ubiquitously. To sort out the requirements for β-catenin and
Notch we undertook transplant experiments where a donor and
host combination permitted us to learn where β-catenin and
Notch were required for the induction sequence.
In the first set of experiments, if one vegetalizes embryos by
ubiquitous expression of the stable form of β-catenin, embryos
specify excess endoderm, and in the absence of micromeres,
no SMCs (Fig. 2A). If micromeres are present, excess
expression of β-catenin results in overproduction of endoderm
plus excess numbers of SMCs (Fig. 2B). In these embryos the
entire surface of the embryo expresses apical Notch except for
the presumptive SMC region. Micromeres induce SMCs no
matter where they are placed in such embryos. For example,
in Fig. 2C, red control micromeres were placed at the animal
pole of activated β-catenin-expressing hosts and a second site
of SMC specification was set up there. Absence of β-catenin
in the nucleus has the reciprocal phenotype since no mesoderm
or endoderm are specified in such cases (Logan et al., 1999).
We conclude that β-catenin is necessary for specification of
vegetal tissues and micromeres are necessary for specification
of SMCs in those vegetal tissues.
The difficulty with these data is that they reveal little about
how β-catenin might be involved in SMC and endoderm
specification other than being required. To establish details
about how β-catenin and Notch work in SMC and endoderm
specification we undertook a series of experiments designed to
learn where and when these molecules are required in the
induction process. The following set of experiments
sequentially examines the induction sequence and the roles of
β-catenin and Notch.
Micromeres require nuclear β-catenin to attain
competence as inducing cells
The animal pole transplant experiment is the classic way to
demonstrate inducing activity of micromeres (Ransick and
Davidson, 1993; Horstadius, 1939). Earlier we demonstrated
that nuclear β-catenin is necessary for that inductive capacity
(Logan et al., 1999). Unfortunately that experiment does not
allow one to conclude that the activity exhibited at the animal
pole is the same activity normally used by micromeres at the
vegetal pole. Accordingly we asked whether β-catenin is
required by micromeres for their normal inductive activity at
the vegetal pole.
Fig. 3 shows the result of a micromere swap experiment
to test the hypothesis that β-catenin is required for
endogenous micromere inductive competence. Because βcatenin is present in all cells and used by other cells in the
embryo for signaling (see below), it was important to alter
β-catenin just in the micromeres. This was possible with
the swap experiment. Two populations of embryos were
cultured. Control micromere donor embryos were labeled
red at the 8-cell stage and micromeres removed from these
embryos at the 16-cell stage. Micromeres from embryos
injected with truncated cadherin are labeled blue in Fig. 3.
The expression of truncated cadherin in the plasma
membrane eliminates the nuclear signaling capacity of
endogenous β-catenin by preventing it from entering the
nucleus (Logan et al., 1999). Host embryos had their
micromeres removed at the 16-cell stage. Immediately
thereafter they received either red control donor micromeres,
or blue micromeres that were unable to signal through
nuclear β-catenin. The red control micromeres induced
SMCs as expected (Fig. 3A) (n=79/79). Blue micromeres
lacking nuclear β-catenin failed to induce SMCs (Fig. 3B)
(n=29/31) and also failed to ingress into the blastocoel, since
without nuclear β-catenin they were unable to follow the
PMC lineage. Thus transplantation of micromeres lacking
5116 D. R. McClay and others
Fig. 1. Micromeres are necessary for SMC
specification. (A) A control embryo is stained with
Notch (green) and a marker for PMCs (blue). At the
vegetal pole the central region is unstained due to an
earlier consumption of Notch, delineating SMC
specification (between the arrowheads). Lateral to the
vegetal pole Notch is strongly expressed apically on
cells that are fated to be endoderm, thereby providing
a marker of endodermal specification.
(B) Micromeres were removed at the 16-cell stage
(confirmed by the absence of PMCs in the
blastocoel). The entire vegetal plate stains apically,
indicating endoderm specification and no SMC
specification. (C) Red donor micromeres were transplanted to the vegetal pole of micromereless embryos at the 16-cell stage. SMC
specification occurred, as revealed by the unstained region in the vegetal plate (arrowheads).
Fig. 2. Augmentation of β-catenin signaling expands
the endoderm and accommodates expanded SMC
induction. Zygotes were injected with 0.03 pg of
RNA encoding stabilized β-catenin. (A) Apical
Notch surrounds a micromereless embryo, indicating
a strongly vegetalized phenotype, yet no SMCs are
specified. (B) The same amount of β-catenin was
injected, and this time micromeres were present. The
resulting embryo was vegetalized and had a vastly
expanded area of SMC specification in addition
(arrowheads). (C) Red micromeres were added to
the animal pole of vegetalized embryos inducing an
SMC territory at the animal pole. Blue endogenous PMCs induced the vegetal SMC territory (lower arrowheads) and the red PMCs were
responsible for inducing the animal SMC territory (upper arrowheads).
nuclear β-catenin mimics the phenotype seen in the
micromereless embryos as in Fig. 1. These data support the
hypothesis that β-catenin is necessary in micromeres for
production of the micromere inductive signal.
The micromere induction signal is short-range and
quantitative
Now that we had an assay for micromere inductive activity a
number of experiments could be performed to characterize the
properties of the signal. First, we asked whether the signal acts
at a short range. Normally the vegetalmost veg 2 cells are
specified as SMCs while other veg 2 cells are specified as
endoderm. The cells specified as SMCs normally are in close
proximity to micromere progeny. From experiments with
activated Notch we knew that the region of SMC specification
could be expanded to include most, if not all, veg 2 progeny
(Sherwood and McClay, 1999). First, we designed experiments
to ask whether micromere position dictated the induction, or if
induction always occurred in the vegetalmost veg 2 cells.
The first hypothesis tested was the prediction that SMC
induction occurs through short-range interactions between
micromeres and responding macromere progeny. To test this
hypothesis we needed to distinguish the location of the added
micromeres relative to the true vegetal pole. All micromeres
were removed from hosts at the 16-cell stage. A single red
signaling-incompetent micromere was added at the true vegetal
pole to provide a spatial point of reference. As shown above,
a micromere expressing the cadherin cytoplasmic tail is
incapable of signaling and inducing SMCs. To confirm that,
identical control cadherin-injected embryos (including the
embryo from which the micromere was taken) were cultured
and in all cases these controls developed as ectoderm-only
dauer blastulae. This gave us a point of reference to locate the
original vegetal pole later in development. Next, we placed a
single blue signaling-competent micromere at a position lateral
to the vegetal pole. To label the blue signaling micromere,
FITC was introduced into that donor embryo shortly after
fertilization (actually the ‘blue’ is the assigned pseudo-color of
Fig. 3. β-catenin is required in micromeres for the micromeres to
induce SMCs. (A) Control red micromeres were transplanted to the
vegetal pole of micromereless embryos at the 16-cell stage. Those
cells became PMCs (red) and induced the formation of SMCs at the
vegetal pole. (B) Blue micromeres from embryos expressing
truncated cadherin were transplanted to the vegetal pole of control
micromereless embryos. No SMCs were induced as shown by the
expression of Notch throughout the vegetal pole region, and the blue
micromeres remained in the blastoderm rather than ingress as PMCs.
Micromere induction signal 5117
Fig. 4. The micromere inductive signal is short-range. A red nonsignaling micromere was placed at the vegetal pole to provide a point
of reference. Blue signaling-competent micromeres were
transplanted at several distances from the vegetal pole. (A,B) Two
examples where SMCs were induced lateral to the vegetal pole. In
(A) the inducing micromere was placed closer to the vegetal pole
than in (B). Arrowheads in A and B mark the vegetal pole and the
center of the induced SMC territory. The endoderm is delineated by
the apical pattern of Notch expression in green.
the FITC fluorescence obtained with confocal imaging). At the
mesenchyme blastula stage the embryos were fixed and the
position of SMC specification was determined relative to the
position of the inserted red signal-incompetent micromere. As
shown in Fig. 4, the reference red cells failed to induce SMCs
and were surrounded by endoderm, as stained with apical
Notch. In every case the site of SMC induction occurred at the
position of the inserted blue signaling micromere (we looked
only at transplants in the vegetal hemisphere). We conclude
that the SMC inductive signal is indeed short-range and may
in fact require direct contact with micromeres. Further, we
conclude that cells lateral to the true vegetal pole are perfectly
capable of being induced to be SMCs, though normally they
become endoderm in the absence of the short-range micromere
signal.
We next asked whether there is a quantitative aspect to
micromere induction activity, that is, do more micromeres
induce larger numbers of SMCs? This experiment was
performed in two ways. First, different numbers of micromeres
were transplanted to the vegetal pole of micromereless hosts.
After induction the area of SMC specification was measured.
Fig. 5 shows the results of that experiment. The area of SMC
specification at the beginning of gastrulation shows an
incremental increase in embryos with 0, 1, 2 or 4 micromeres
(0, 644, 1408 and 2335 µm2, respectively, of SMC induction
surface area (n=6+ embryos measured with each combination).
We conclude that there is a quantitative aspect to the ability of
micromeres to induce SMCs.
Another quantitative question is whether the induction
response is amplified if vegetalized micromeres are the
inducing cells. Ideally this question will be addressed by direct
overexpression of the inductive ligand by micromeres. Since
the inducing ligand has yet to be identified our only choice was
to ask whether vegetalized micromeres convey an extra signal.
Donor micromeres were taken from embryos expressing the
stabilized form of β-catenin at a level known to vegetalize
Fig. 5. The number of micromeres correlates with the size of the
SMC induction field. 0 (A), 4 (B), 1 remaining (C), or one red
micromere added to a micromereless embryo (D) are compared using
apical Notch to delineate the extent of endoderm surrounding the
SMC-specified territory. The results show the sections with the
widest expanse induced to specify SMCs (between arrows). The
actual areas of the territories measured are reported in the text.
whole embryos (Logan et al., 1999). When untreated host
micromereless embryos received one vegetalized micromere
there was a slight increase in induced SMCs relative to controls
but the difference was not statistically different. Response
to the induction, however, is affected by vegetalization. If a
single control micromere is tranplanted to a vegetalized
micromereless host, the area of SMC induction is twice that of
the control response (one control micromere transplanted to the
vegetal pole of a control micromereless host). We cannot know
at present from this experiment whether or not vegetalized
micromeres produce more inductive signal, but even if they do,
the inductive influence of a vegetalized micromere is not
significantly greater than control induction. The receptive
area of sensitivity to the inductive signal is affected by
vegetalization though the mechanism for that increased
sensitivity is not known. From these experiments we conclude
that the area of SMC induction depends on the number of
micromeres present at the vegetal pole. The inductive response
does not appear to be amplified significantly when the
micromeres are vegetalized.
The β-catenin signal in macromeres is necessary
before these cells are competent to receive the
micromere induction signal
Macromeres and their progeny import β-catenin into their
nuclei at the fifth cleavage (Logan et al., 1999). From then until
about the tenth cleavage β-catenin continues to be localized in
nuclei of veg 2 cells and their progeny. Veg 1 cells lose nuclear
β-catenin at the sixth cleavage but veg 1 progeny fated to
become endoderm again localize β-catenin to their nuclei at
around the tenth to eleventh cleavage (Logan et al., 1999;
Sherwood and McClay, 1999). In earlier studies we showed
5118 D. R. McClay and others
Fig. 6. A β-catenin signal in macromeres is
necessary for SMC specification. (A) Control
embryos at the 16-cell stage had their
micromeres removed and replaced with control
Signaling-competent micromeres. The embryo
in A has PMCs that ingressed normally, and
SMCs that were specified at the vegetal plate as
normal (arrow points to cells stained with SMC
1 antibody). (B) RNA expressing truncated
cadherin makes an embryo specify neither
endoderm nor SMCs (no SMC-1 staining), and
fails to specify the micromere lineage (no PMCs
in the blastocoel). (C) Control micromeres were
added to a micromereless embryo that had been
injected with RNA to truncated cadherin. Those
micromeres became PMCs, had inductive competence, ingressed normally and expressed the PMC marker, but the macromeres, without
nuclear β-catenin, were unable to respond to the control micromere induction signal, and failed to specify either endoderm or SMCs.
Fig. 7. SMCs are induced through
the Notch signal. (A) One red
micromere was transplanted onto
the vegetal plate of control
micromereless embryos. As before,
SMCs were specified (between
arrowheads). (B) Fertilized eggs
received RNA expressing activated
Notch. At the 16-cell stage
micromeres were removed from
these embryos. In the absence of
micromeres, but in the presence of
activated Notch, SMCs were
specified in these embryos
(between arrowheads). (C) Red signaling-competent micromeres were added to micromereless embryos that had been injected with RNA
expressing dominant-negative Notch RNA. These embryos failed to specify SMCs despite having a competent induction signal. (D) The dominantnegative Notch was expressed only in micromeres. These cells still induced SMCs when transplanted to control micromereless hosts (between
arrowheads). Thus, Notch is unnecessary in micromeres for the induction signal, but necessary in macromeres for reception of the signal.
that if β-catenin is prevented from entering the nuclei of
macromeres these cells develop as ectoderm and fail to specify
either endoderm or SMCs (Logan et al., 1999). In that
experiment β-catenin also failed to enter micromere nuclei.
Thus, it was possible that entry of β-catenin into macromere
nuclei required the micromere signal and/or endoderm
specification required β-catenin in micromeres. To address
these questions we asked if signaling-competent micromeres
could rescue endoderm and/or SMCs in macromeres lacking
nuclear β-catenin.
Fig. 6 shows the results of a swap experiment in which
normal (red) micromeres were transplanted to nuclear βcatenin-minus macromeres. As shown in Fig. 6C, when
signaling-competent micromeres are provided the β-cateninminus macromeres fail to specify SMCs (nor do they specify
endoderm). Embryos in Fig. 6 are stained with two markers,
both of which are mouse IgMs. The SMC marker (arrow, Fig.
6A), stains a population of SMCs (Sweet et al., 1999), and the
other marker stains PMCs (the same color in this figure). The
failure to express the SMC marker in macromeres lacking
nuclear β-catenin (Fig. 6C) indicates that β-catenin is
necessary in macromeres to receive the micromere induction
signal. In Fig. 6C the control micromeres ingress to become
PMCs but the remainder of the embryo remains ectoderm-like.
Control micromereless embryos failed to specify SMCs as in
Fig. 1 (Fig. 6B). We conclude that β-catenin is required in
macromeres for these cells to become receptive to the
micromere induction signal.
From earlier data we suspected the micromere induction
signal to act through Notch (McClay and Sherwood, 1999;
Sweet et al., 1999). This prompted an examination of the
relationship between β-catenin and Notch in macromeres.
Notch in macromere progeny is maternal and is expressed on
macromeres even in the absence of nuclear β-catenin. Thus,
β-catenin is necessary either to augment maternal Notch with
zygotic copies of the Notch protein, or to establish other
components in the complete Notch signal transduction
pathway. Double injections of RNA encoding activated
Notch and the truncated cadherin fail to rescue β-catenininhibited macromeres (data not shown). Since activated
Notch fails to rescue the β-catenin-deficient macromeres we
conclude that β-catenin is independently necessary in both
micromeres and macromeres. In micromeres β-catenin is
necessary for specification of the PMC lineage and for the
pathway leading to production of the micromere inductive
signal. In macromeres β-catenin is required for vegetal
specification and for competence of the Notch signaling
pathway.
The micromere induction signal acts through Notch
to specify SMCs
The experiments of Sweet et al. (1999), our earlier expression
Micromere induction signal 5119
data (Sherwood and McClay, 1999), and the experiments above
suggested that Notch might be the receptor for the micromere
induction signal. If that were true several predictions were
testable. First, in the absence of the micromere induction
signal, provision of an activated Notch signal to macromere
progeny should rescue SMCs. To test this hypothesis
micromeres were removed from embryos expressing the
constitutively activated form of Notch. Fig. 7B shows that a
territory of SMC specification appears in these embryos. As
shown earlier, uninjected micromereless embryos fail to
specify SMCs. Thus, activated Notch rescues SMCs in the
absence of the micromere inductive signal.
A second prediction is that provision of signaling-competent
micromeres should fail to induce SMCs if dominant-negative
Notch is supplied to macromeres. Results in Fig. 7C show that
prediction to be supported as well. Signaling-competent
micromeres do indeed fail to induce macromeres expressing
the dominant-negative form of Notch.
Third, if Notch is required in macromeres only, micromeres
expressing the dominant-negative form of Notch should be
competent to signal and induce SMCs (Fig. 7D). This
prediction is correct. SMCs are specified by micromeres
expressing dominant-negative Notch. These data are consistent
with the model that Notch, in macromere progeny, is the
receptor for the micromere inductive signal.
The rescue experiment with activated Notch demonstrates a
curious property of the induction response. Earlier it was
shown that the signal from micromeres provides spatial
localization of SMC specification. Fig. 7 shows that in the
absence of micromeres the SMCs are specified at the vegetal
pole. All cells of the micromereless embryo in Fig. 7B express
activated Notch but only vegetal pole cells become specified
as SMCs. The number of cells that become specified as SMCs
varies with the concentration of RNA to the activated Notch.
We conclude that the position of SMC specification is decided
both by proximity to micromere-released signal, and by an
intrinsic bias in the vegetalmost veg 2 cells. The data support
a model whereby the most sensitive area for reception of the
micromere induction signal is at the vegetal pole. Lateral to
that, induction will occur if the micromeres are placed
ectopically over vegetally specified cells, or if activated Notch
signal is provided the territory is expanded.
The micromere induction signal acts between the
eighth and tenth cleavage division, or approximately
at the time when the Notch signal is active in veg 2
cells
The disappearance of Notch from presumptive SMCs occurs at
around 7 hours after fertilization or roughly between the eighth
to tenth cleavages (Sherwood and McClay, 1997; Sherwood
and McClay, 1999). Based on the phenotypes seen with either
activated or dominant-negative Notch, it was proposed that
Notch is removed from the plasma membrane of SMCs as it
signals or shortly thereafter, thereby creating the vegetal region
devoid of Notch. Knowing that Notch leaves the vegetal plate
during the eighth to tenth cleavages we asked when, relative to
those cleavage divisions, does the micromere induction signal
occur.
Micromeres were removed from the embryos at the fourth
cleavage. The micromereless embryos were then cultured for
varying periods of time before receiving blue, signaling-
competent micromeres from donor embryos. In the first
experiments we transferred micromeres from 16-cell stage
donors to older host micromereless embryos. Hosts received
the donor micromeres at the fifth-seventh and ninth cleavages,
or after being without micromeres for 1-3 or 5 cleavages. After
as many as 3 (or 4 cleavages), readdition of micromeres
allowed SMCs to be specified in the host embryos (Fig. 8C).
However, if the micromereless embryos were missing
micromeres for 5 cleavages we saw no SMC territory specified
and the embryos displayed apical Notch throughout the vegetal
hemisphere as is typical of control micromereless embryos
(Fig. 8D). We conclude that micromeres are unnecessary for
SMC induction for several cleavages after their appearance and
that the inductive signal is necessary at or sometime shortly
after the eighth cleavage.
It was possible that failure to induce SMCs at the ninth
cleavage was because the young micromeres simply were not
capable of immediately inducing SMCs when transferred at the
ninth cleavage. If this were the case we predicted that older
donor micromeres should be able to immediately induce if
transplanted to ninth cleavage hosts (assuming this is actually
when the induction occurs). Accordingly micromeres were
labeled red with RITC and transferred initially to surrogate
unlabeled host embryos for varying intervals. At the same time
the experimental hosts had their micromeres removed. When
the micromereless embryos reached the ninth cleavage the red
micromeres were harvested from the surrogate hosts and
inserted into the vegetal plate of the micromereless hosts. The
older micromeres induced SMCs (Fig. 8E). We conclude that
the micromere induction signal is unnecessary for SMC
induction prior to the eighth or ninth cleavage. We were unable
to rescue induction when micromeres were returned to
embryos at tenth cleavage or thereafter. These data further
support the model that an immediate consequence of the
micromere induction signal is the surface disappearance of
Notch in the cells that have just received the induction signal
through Notch.
DISCUSSION
Micromeres are shown to induce SMCs through Notch
signaling. The induction occurs between the eighth and tenth
cleavages and it relies on close or direct contact between the
micromeres and the targeted cells of the veg 2 layer. As a
consequence maternal Notch disappears from the cells that will
become SMCs during or following reception of the micromere
signal. The dynamics of Notch disappearance are incompletely
understood but appear to be similar to those seen in other
systems using the Notch signaling pathway. By
immunofluorescence, the extracellular region of Notch
relocates to intracellular vesicles at the eighth to tenth
cleavages; coincident with the time that SMCs receive the
micromere induction signal (Sherwood and McClay, 1997).
Whether endocytosis occurs during or following the signal
reception is not known. Either way, Notch is lost from the
newly specified SMC population, leaving that population
negative for Notch. Of the many uncertainties that remain, the
precise time that micromeres become competent to signal is
not known since the micromere ligand that activates Notch has
not been identified. Based on the timing experiments it is
5120 D. R. McClay and others
Fig. 8. Micromeres induce at the eighth to tenth cleavages. Micromeres were removed from host embryos at the 16-cell stage. In (A), a control
micromereless embryo expresses Notch throughout the vegetal region. Blue donor micromeres were returned to similarly treated hosts either
immediately, one cleavage later (fifth) (B), three cleavages later (seventh) (C), or 5 cleavages later (ninth) (D). The blue micromeres were in
each case from fourth cleavage donors. In (E) ninth cleavage micromereless embryos received red donor micromeres that were also in the 9th
cleavage at the time of donation. The embryos were stained with Notch (green) to indicate apical endoderm and the area of SMC specification
(cells missing apical Notch at the vegetal pole, between arrowheads).
predicted that the induction signal is present on the micromeres
as early as the sixth cleavage.
The micromere induction signal described herein appears
to be a single event that activates Notch. However, these data
do not mean that the Notch induction signal is the only signal
from the micromeres. Earlier studies (Ransick and Davidson,
1995), suggest that an additional signal from micromeres may
be involved in specification of endoderm. They showed that
expression of Endo 16, an early endoderm gene, depends
upon the presence of micromeres for its full expression at the
vegetal plate. Endo 16 appears in macromere progeny in the
absence of micromeres, but presence of micromeres between
the fourth and sixth cleavage increases the level of Endo 16
expression at the mesenchyme blastula stage significantly
(Ransick and Davidson, 1995). Between the fifth and eighth
cleavage β-catenin is in the nuclei first of macromeres and
then of the veg 2 layer (the layer immediately above
micromeres). Our experiments have shown that β-catenin
nuclear entry is essential for macromere endodermal
specification. It occurs cell-autonomously, even in the
absence of micromeres (Logan et al., 1999). The
Ransick/Davidson data point to an additional micromereinduced requirement between the fourth and sixth cleavages
that somehow amplifies the initial endoderm specification.
Curiously, the endoderm induction appears to be independent
of SMC specification because, as shown here, SMC
specification does not require micromeres to be present
during the fourth to seventh cleavages.
Removal of micromeres delays invagination of the
archenteron. In part that delay could be due to the failure to
fully specify endoderm in the absence of micromeres, but
another possible explanation is due to the absence of SMCs.
Normally, invagination begins at the vegetal pole with bottle
cells initiating the inward bending primary phase of
archenteron formation (Nakajima and Burke, 1996; Kimberly
and Hardin, 1998). Neither of these groups examined the
lineage origins of the bottle cells but by location they are likely
to be SMCs. When the entire vegetal pole is endoderm, there
may be a reduced capacity to initiate archenteron formation.
Invagination of the archenteron eventually occurs, but during
the delay the embryo also regulates and replaces the SMCs
(Sweet et al., 1999).
If β-catenin is depleted from cells so that it cannot enter
the nucleus, there is no specification of micromeres
(Wikramanayake et al., 1998; Emily-Fenouil et al., 1998;
Logan et al., 1999). The asymmetric division that forms the
micromeres still occurs, but those micromeres are incapable of
inducing a second axis at the animal pole (Logan et al., 1999),
and they are incapable of inducing SMCs (Fig. 3). Thus, one
of the pathways initiated by nuclear β-catenin in micromeres
is the synthesis or activation of the ligand that binds to Notch.
The Delta or Serrate homologues are the obvious candidates
for the micromere signal but these have yet to be identified and
tested for this ligand function in the sea urchin. From the data
we have so far, fourth cleavage micromeres are incapable of
immediately inducing through Notch. By about the sixth to
seventh cleavage micromere progeny have the ability to induce
eighth to tenth cleavage veg 2 progeny to specify SMCs
through Notch. Thus we predict that when the ligand is
identified it will be produced zygotically, or activated within
one or two cleavage divisions of nuclear β-catenin in
micromeres, and within one or two cleavages after ligand
appearance, the SMC induction occurs through the Notch
signal.
Micromeres begin moving β-catenin into their nucleus
during the 20 minutes between the fourth and fifth cleavages.
In a population of embryos fixed early between the fourth and
fifth cleavages we see only a fraction of the embryos with
nuclear β-catenin. That fraction increases if embryos are fixed
later in the 20 minute period between cleavages, and by fifth
cleavage all micromeres have nuclear β-catenin. These data
suggest that β-catenin accumulates in micromere nuclei
during the fourth interphase. Is there a Wnt signal to activate
β-catenin to go to the nucleus? So far the evidence is negative
and points to a cell-autonomous mechanism. If we separate
blastomeres at each cleavage and constantly culture the
isolated cells separately, β-catenin still enters micromere
nuclei at the correct time (Logan et al., 1999). This suggests
that a cell-autonomous mechanism governs nuclear entry.
Further, in separate experiments with wnt 8 we find no
Micromere induction signal 5121
evidence of a wnt signal being involved in directing β-catenin
to micromere nuclei (A. H. Wikramanayake et al.,
unpublished). β-catenin is maternally expressed and appears
to be expressed equally by each cell. In situ data also indicate
that β-catenin RNA is equally distributed throughout the
embryo. By the 16-cell stage the protein is associated with
adherens junctions all around the embryo (Miller and
McClay, 1997), but then the protein goes into the nuclei only
in micromeres (and in the lower tier of macromeres a
cleavage later). Inhibition of endogenous GSK-3 enhances
movement of β-catenin to the nuclei (Emily-Fenouil et al.,
1998), but this does not offer any proof of how nuclear entry
normally occurs. Thus, the earliest known molecular event of
consequence for mesoderm and endoderm specification
occurs at the 16-cell stage, though as yet it is not known what
triggers the localized nuclear entry of β-catenin.
The induction signal is not immediately present on
micromeres but appears shortly thereafter in a βcatenin-dependent process
When micromeres from the fourth cleavage are transplanted to
the vegetal pole of micromereless embryos at the seventh
cleavage, the micromeres induce SMCs during the eighth to
tenth cleavage. Fourth cleavage micromeres fail to induce if
transplanted to ninth cleavage micromereless hosts. If βcatenin is prevented from entering the nucleus of micromeres
the cells are unable to induce SMCs. Together the timing
experiments suggest that the micromeres are competent to
signal within about 2-3 cleavages of the fourth, or after about
80 minutes of elapsed time. This is plenty of time for several
steps in a pathway prior to the production or activation of the
Notch ligand, though the synthesis and insertion of the Notch
ligand is likely to be slow relative to the production of any
transcriptional regulatory factors upstream of that ligand
synthesis. Whether the β-catenin-to-ligand production is a
direct or an indirect pathway remains to be determined.
The micromere induction signal acts at short range
No matter where the induction occurs in the vegetal
hemisphere, if micromeres are present, SMC induction occurs
by direct contact or within at most one cell diameter away from
the micromere. Several years ago we showed that micromeres
and their PMC progeny extend thin filopodia that extend the
apparent range of contact of those cells (Miller et al., 1995).
Thus even if the cell body is two or more diameters away from
an induced cell it is hard to rule out the possibility that the
inducing cell makes direct contact with the induced cell via
thin filopodia. These data support the hypothesis that induction
requires direct contact or very short distances between the
inducing and the receiving cell. Possible exceptions to a direct
contact requirement might include cases where the embryo is
vegetalized or when the Notch signal is activated at a high
level. Even then direct contact might be required in that as a
consequence of vegetalization the expansion of the signaling
source might be expanded.
The inductive signal acts through Notch
Notch disappears from the presumptive SMCs at the vegetal
pole (Sherwood and McClay, 1999), and our experiments
support this disappearance as being a part of the SMC
specification sequence. The Notch disappearance appears to be
directly related to the Notch signal. Several curious features of
this signaling are worth mentioning. When activated Notch
specifies SMCs in the absence of micromeres, endogenous
Notch disappears from the SMCs as well. Thus one
consequence of Notch signaling is removal of Notch from the
cell surface, even if that Notch was not involved in the
signaling. In any Notch-related induction it is likely that only
a small portion of the surface Notch actually is used for the
signal, but as shown by the subsequent disappearance, an early
consequence of Notch signaling is removal of all remaining
Notch receptor from the cell surface (at least in this case). This
biology is useful but also complicates future experiments. We
are fortunate in these experiments because removal of all
remaining Notch provides a convenient marker to indicate that
SMC specification has been initiated. The complication will
occur in future experiments if only a small amount of Notch
actually performs the signaling and the remaining Notch is
removed by endocytosis. Western analysis of cells allows one
to visualize the fragments of Notch. Unfortunately, the
endogenous post-signaling disappearance of Notch makes it
difficult to distinguish between fragmentation due to Notch
signal transduction, and fragmentation due to post-signaling
degradation.
A start has been made toward characterization of the
micromere signal. There are many gaps to fill, however. We do
not know what triggers the β-catenin-induced signaling. We
know very little about the pathway between β-catenin and the
appearance of the inducing signal. We know little about how
Notch transduces the signal and even less about downstream
events. Nevertheless, from this study the biological framework
is in place for those detailed analyses.
REFERENCES
Balzer, T. B. (1967). Life and Work of a Great Biologist. Berkeley, CA:
University of California Press.
Boveri, T. (1901). Uber die Polaritat des Seeigel-Eies. Verh. Phys. Med. Ges.
Wurzburg 34, 145-176.
Emily-Fenouil, F., Ghiglione, C., Lhomond, G., Lepage, T. and Gache, C.
(1998). GSK3beta/shaggy mediates patterning along the animal-vegetal axis
of the sea urchin embryo. Development 125, 2489-2498.
Hardin, J., Coffman, J. A., Black, S. D. and McClay, D. R. (1992).
Commitment along the dorsoventral axis of the sea urchin embryo is altered
in response to NiCl2. Development 116, 671-685.
Horstadius, S. (1928). Uber die determination des keimes bei echinodermen.
Acta Zoologica 9, 1-191.
Horstadius, S. (1939). The mechanics of sea urchin development as studied
by operative methods. Biol. Rev. 14, 132-179.
Kiehart, D. P. (1982). Microinjection of echinoderm eggs: apparatus and
procedures. Meth. Cell Biol. 25, 13-31.
Kimberly, E. L. and Hardin, J. (1998). Bottle cells are required for the
initiation of primary invagination in the sea urchin embryo. Dev. Biol. 204,
235-250.
Logan, C. Y. and McClay, D. R. (1997). The allocation of early blastomeres
to the ectoderm and endoderm is variable in the sea urchin embryo.
Development 124, 2213-2223.
Logan, C. Y., Miller, J. R., Ferkowicz, M. J. and McClay, D. R. (1999).
Nuclear beta-catenin is required to specify vegetal cell fates in the sea urchin
embryo. Development 126, 345-357.
McClay, D. R. and Logan, C. Y. (1996). Regulative capacity of the
archenteron during gastrulation in the sea urchin. Development 122, 607616.
Miller, J., Fraser, S. E. and McClay, D. (1995). Dynamics of thin filopodia
during sea urchin gastrulation. Development 121, 2501-2511.
Miller, J. and McClay, D. R. (1997). Changes in the pattern of adherens
5122 D. R. McClay and others
junction-associated β-catenin accompany morphogenesis in the sea urchin
embryo. Dev. Biol. 192, 310-322
Nakajima, Y. and Burke, R. D. (1996). The initial phase of gastrulation in
sea urchins is accompanied by the formation of bottle cells. Dev. Biol. 179,
436-446.
Ransick, A. and Davidson, E. H. (1993). A complete second gut induced by
transplanted micromeres in the sea urchin embryo. Science 259, 1134-1138.
Ransick, A. and Davidson, E. H. (1995). Micromeres are required for normal
vegetal plate specification in sea urchin embryos. Development 121, 32153222.
Sherwood, D. R. and McClay, D. R. (1997). Identification and localization
of a sea urchin Notch homologue: insights into vegetal plate regionalization
and Notch receptor regulation. Development 124, 3363-3374.
Sherwood, D. R. and McClay, D. R. (1999). LvNotch signaling mediates
secondary mesenchyme specification in the sea urchin embryo.
Development 126, 1703-1713.
Sweet, H. C., Hodor, P. G. and Ettensohn, C. A. (1999). The role of micromere
signaling in notch activation and mesoderm specification during sea urchin
embryogenesis [In Process Citation]. Development 126, 5255-5265.
Wikramanayake, A. H., Huang, L. and Klein, W. H. (1998). β-Catenin is
essential for patterning the maternally specified animal-vegetal axis in the
sea urchin embryo. Proc. Natl. Acad. Sci. USA 95, 9343-9348.