/ . Embryo/, exp. Morph. Vol. 37, pp. 13-31, 1977
Printed in Great Britain
{3
An autoradiographic analysis of nucleic
acid synthesis in the presumptive primordial germ
cells of Xenopus laevis
BY MARIE DZIADEK 1 AND K. E. DIXON 1
From the School of Biological Sciences, The Flinders University of
South Australia
SUMMARY
Microinjection of [3H]thymidine into Xenopus laevis embryos between late blastula
(stage 10) and early tadpole (stage 44) showed that the presumptive primordial germ cells
synthesise DNA between stages 10-33. The percentage of labelled cells was highest between
stages 10 and 16, declined sharply between stages 22 and 26 and rose again between stages
26 and 33. The fluctuations in the labelling patterns together with increase in the number of
presumptive primordial germ cells and direct observation of germ cells in mitosis suggested
that the germ cells divide three times between stages 10 and 44. The first divisions probably
take place during gastrulation (stages 10-12), the second relatively synchronously at about
stages 22-24 and the third series again relatively synchronously about stages 37-39. This
period of proliferative activity is distinguishable on the one hand from the cleavage divisions
in which the number of germ cells does not increase and on the other hand from the next
proliferative phase by a period of mitotic inactivity. Microinjection of [3H]uridine showed
that the presumptive primordial germ cells synthesize RNA only in mid-gastrula to early
tail-bud-stage embryos. There is no obvious simple causal relationship between RNA
synthesis and the movement of the germ plasm to the nucleus, or with division of the germ
cells or with their migration out of the endoderm.
INTRODUCTION
The synthesis of nucleic acids in early amphibian embryos has been studied
because a knowledge of the time of onset and course of genetic activity during
early development is important for our understanding of the processes which
ultimately lead to the formation of differentiated cells. In almost all of these
studies, changes in the rates of synthesis of DNA and of the major species of
RNA have been determined using whole embryos (reviewed in Gurdon, 1968,
1974). In Xenopus laevis embryos, synthesis of DNA is rapid during early
cleavage but the rate declines at the end of cleavage (Graham & Morgan, 1966),
whereas nuclear RNA synthesis in general cannot be detected until late in
cleavage (but see Decroly, Cape & Brachet, 1964; Gurdon & Woodland, 1969).
The synthesis of low molecular weight RNA probably including tRNA begins
1
Authors' address: School of Biological Sciences, The Flinders University of South
Australia, Bedford Park, South Australia 5042, Australia.
2
EMB 37
14
MARIE DZIADEK AND K. E. DIXON
about stage 8 (Gurdon & Woodland, 1969) but ribosomal RNA synthesis
begins later, at the gastrula and early neurula stages (Brown & Littna, 1966;
Woodland & Gurdon, 1968; Gurdon & Woodland, 1969). However, the
assumption underlying most of these studies, that all cells of the embryo are
equivalent, is clearly not correct; for example, the endoderm cells divide less
frequently than ectoderm cells (Gurdon, 1968). Nevertheless, there have been
relatively few autoradiographic investigations which permit study of nucleic
acid synthesis in particular types of cells or even in cells from particular regions
of the embryos. In one such study Graham & Morgan (1966) investigated DNA
synthesis in the endoderm cells of X. laevis embryos from stages 7-28 and
calculated the cell cycle times. Bieliavsky & Tencer (1960) and Denis (1964)
studied RNA synthesis in Xenopus embryos and reported that it begins at the
gastrula stage and cytoplasmic incorporation becomes evident by at least late
gastrula.
Bachvarova & Davidson (1966) have also studied RNA synthesis in X. laevis
embryos and concluded that presumptive endoderm and mesoderm cells show
a sudden increase of at least 20-fold just prior to the beginning of gastrulation
and shortly after, presumptive ectoderm cells become activated at the midgastrula stage.
The germ-cell lineage in amphibians provides an almost unique opportunity
to study changes in nucleic acid synthesis as cells begin to differentiate, since
the germ cells can be recognized, even in early cleavage stages, by the presence
of a readily visualized cytoplasmic marker termed germ plasm (Bounoure, 1934,
1964; Blackler, 1970; Whitington & Dixon, 1975). In a previous report from
this laboratory, Whitington & Dixon (1975) have described changes in size,
number and position of the presumptive primordial germ cells (i.e. germ cells
which have not yet migrated into the gonads and hence still lie in the endoderm).
The results suggested that the germ cells divide 2-3 times during the endodermal
phase between blastula and stage 41. Subsequently, we have confirmed, using
autoradiography with [3H]thymidine that, contrary to earlier reports (Blackler,
1970), the presumptive primordial germ cells synthesize DNA and divide during
this period (Dziadek & Dixon, 1975).
The experiments reported here were designed to ascertain whether the divisions
of the presumptive primordial germ cells are fixed in number and synchronous.
The synthesis of RNA was also studied to determine the time of onset and the
course of genetic activity in the germ cells, relative to other cells in the embryo.
The time when RNA synthesis begins in the germ cells might represent the point
when a separate and distinct germ cell lineage is established. The subsequent
pattern of RNA synthesis might be correlated with other events in the lineage
such as movement of the germ plasm to the nucleus, division of the germ cells
and their migration to the genital ridges.
DNA and RNA synthesis in Xenopus germ cells
15
MATERIALS AND METHODS
Embryos
Ovulation was induced in mature female Xenopus laevis by injection of
300-500 units of chorionic gonadotrophin (Chorulon: Organon). Eggs were
stripped manually and fertilized with a macerated testis (after Wolf & Hedrick,
1971). Embryos developed in aged tap water and were staged according to
Nieuwkoop & Faber (1967).
Microinjection apparatus and procedure
The injection apparatus consisted of a fine glass micropipette mounted on
a micromanipulator and attached to a microsyringe by plastic tubing. Micropipettes were drawn from pyrex capillary tubing of 0-3 mm internal diameter
and 0-5 mm external diameter. The syringe, tubing and micropipette were filled
with paraffin oil. The flow of fluid was controlled by a micrometer attached to
the microsyringe and the level of the solution in the micropipette was shown by
the meniscus at the paraffin-solution interface which was viewed through a stereo
microscope. The micropipette was graduated by means of an externally attached
scale to deliver volumes of approximately 0-08 /tl or multiples thereof, calculated
from the dimensions of the micropipette.
Embryos for injection were supported on moistened filter paper and excess
water was removed. Even hatched embryos were successfully immobilized by
this treatment and hence an anaesthetic was not required. Injections of blastula
and gastrula stages were made into the animal hemisphere, stages 12-41 into the
endoderm and stages after 42 into the posterior body cavity. To aid penetration
of the micropipette, prehatching stages were irradiated with u.v. light (250 nm;
ca. 8000-12000 ergs/mm2) immediately prior to injection.
Labelled substrates
DNA synthesis: [3H]methyl thymidine (1 mCi/ml, specific activity 56 Ci/mM)
was used in all experiments. The standard dose of 0-08 /A represented 0-08 /*Ci
per embryo.
RNA synthesis: [3H]uridine (1 mCi/ml, 44 Ci/mM) was used at the standard
dose of 0-08 /i\.
Histology and autoradiography
Embryos were fixed in Smith's fixative (Jones, 1954) for 12-24 h, embedded
in paraffin and serially sectioned at 5 /tm. Mitotic indices were determined in
sections stained with the Feulgen reaction (after Deitch, Wagner & Richart,
1968). The incorporation of [3H]thymidine into DNA and [3H]uridine into
RNA was visualized by autoradiography of sections coated with Uford K 2
emulsion (after Rogers, 1973), exposed for 1 week and 4 weeks respectively,
developed with Kodak D 19 and stained through the emulsion (Volkonsky,
1928).
16
MARIE DZIADEK AND K. E. DIXON
Measurement of radioactivity
The amount of [3H]thymidine injected was determined from embryos homogenized ultrasonically in distilled water, by counting aliquots for tritium in
Triton X-114/Xylene-based scintillation fluid (Anderson & McClure, 1973) at
an average calculated efficiency of 63 %. To measure the amount of radioactivity remaining unincorporated into DNA, individual embryos were
homogenized ultrasonically in 2 ml of 10 % trichloracetic acid (TCA) at 4 °C.
The precipitate containing DNA was removed by centrifugation, and the TCA
in the supernatant extracted with three ether washes. Aliquots of the extracted
supernatant were counted for tritium in the scintillation counter. The proportion of this radioactivity which was available for incorporation into DNA
was determined by chromatography. The TCA-free supernatant was concentrated by evaporation, unlabelled thymidine (10 /4, 20 mM) was added and
the solution then chromatographed for 16 h with thymine, thymidine and
thymidine triphosphate markers using 75% isopropanol: 2 5 % ammonium
hydroxide. After air drying, the spots originating from the extract corresponding
to thymidine, thymine and thymidine triphosphate were separately eluted in
distilled water, freeze dried and counted for tritium in the scintillation counter.
RESULTS
Preliminary experiments were performed with [3H]thymidine injections in
order to test the following technical aspects : (a) the reproducibility of the
volume delivered by the microinjection apparatus; (b) whether the labelled
substrate was available to all the cells of the embryo; (c) whether subsequent
development of the embryos was affected; and (d) the time of availability of
[3H]thymidine for incorporation into DNA. Similar experiments were not
carried out with [3H]uridine (with the exception of (b) - see later) either because
the results of the thymidine experiments were relevant (as in (a)) or because
questions relating to subsequent development and time of availability were not
applicable because of the short time between injection and fixation.
(a) Reproducibility of the dose
The reproducibility of the volume delivered by the apparatus was measured
by injecting five replicates of each of 0-08, 0-12, 0-25 and 0-5 /A. of [3H]thymidine
into individual vials containing 4 ml of Dioxan scintillation fluid which were
then counted for tritium. The variance between replicates of a given volume
(Table 1) indicates that the smaller the volume, the greater the variability of the
dose. However, even at the lowest dose, the reproducibility was sufficient for
comparative autoradiographic studies.
The degree of leakage from injected embryos was estimated by injecting 0-08,
0-12, 0-25 and 0-50 /A of [3H]thymidine into individual stage-9 blastulas which
DNA and RNA synthesis in Xenopus germ cells
17
Table J. Variability between doses of[3H]thymidine delivered by
microinjection
Approximate
volume
injected (//.I)
No. of
samples
Variance (%)
0-5
0-25
012
008
5
5
5
5
0-34
1-4
9-2
12-2
Table 2. Leakage of [zH]thymidine from X. laevis blastulas after
injection of different volumes
Approximate
volume
injected (fo\)
°/( 3 cpm in embryo
M±SEMof
5 replicates
% cpm in solution
M ± SEM of
5 replicates
0-25
012
008
55-7 ±10-2
940 ± 6 1
94-7 ±3-7
43-4 ±7-5
6-0±l-4
5-3 + 0-7
were then placed separately into 2 ml of 10 % Holtfreter's solution for 30 min.
The amounts of radioactivity in each of five embryos at each dose and in the
corresponding solutions were then measured in the scintillation counter. The
results (Table 2) show, as expected, that leakage decreased when the volume
injected was reduced. Only 5 % leakage occurred with doses of 0-08 /d. There
have been no previous studies in which the reproducibility of the dose and the
amount of leakage have been investigated.
(b) Availability of thymidine to all cells of the embryo
Embryos at 17 stages between 9-44 of development were injected with
0-08 /tCi of [3H]thymidine, fixed 8 h later and prepared for autoradiographic
analysis of the distribution of labelled nuclei within the embryo. Three embryos
were analysed at each stage. Labelled nuclei were observed in all regions of the
embryos. However, the proportion of labelled cells and the density of silver
grains over individual nuclei were noticeably different in the various regions and
tissues of the embryos. These differences are probably due in part to variations
in the rate of cycling between cells of different parts of the embryo (Graham &
Morgan, 1966). We conclude that [3Hlthymidine was available to all cells of the
embryos. This result confirms and extends an earlier finding of Graham &
Morgan (1966) that injection of 0-04/iCi [3H]thymidine into X. laevis blastulas
was sufficient to label all nuclei within 10 min.
18
MARIE DZIADEK AND K. E. DIXON
(c) The effect of injections on development of the embryos
The injection of [3H]thymidine may adversely affect embryos in two ways:
either by excess thymidine blocking DNA synthesis and therefore cell division,
an effect measurably shortly after injection, or by the action of radioactive
disintegrations on the genetic material with consequent abnormalities later in
development. We determined the extent of these effects by comparing the
mitotic index and rate and normality of embryogenesis in uninjected and
injected mid-blastula embryos, a stage at which the rate of cell division is high
(Graham & Morgan, 1966). Embryos were either injected with 008 /id
[3H]thymidine after u.v. irradiation, irradiated with u.v. without injection, or
untreated. One hour later, five embryos from each treatment were fixed, then
serially sectioned and stained with the Feulgen reaction. The mitotic index of
the endoderm cells was determined in ten randomly selected sections of each
embryo. Approximately 200 nuclei were counted in each embryo. Endoderm
cells were selected because of our interest in the presumptive primordial germ
cells which are located in the endoderm.
The mitotic indices were the same in all three groups (0-29 ± 0-01), indicating
that cell proliferation was not immediately affected by the injected thymidine.
The injected embryos also developed at the same rate as controls, as judged by
external criteria, showing that there were no long term effects due to genetic
damage. We conclude therefore that a dose of 0-08 fid [3H]thymidine, corresponding to approximately 40 ^Ci/g n a d no effect, a finding which agrees
with the only other study using X. laevis embryos, in which doses of approximately 20 jLiCi/g had no adverse effects (Graham & Morgan, 1966).
(c) Time of availability of label
The time after injection for which the radioactive thymidine was available
for incorporation into DNA was measured in stage-9 blastulas and stage-29
tail-buds injected with 0-08 jnCi [3H]thymidine. Fifteen minutes after injection
and at 2 h intervals up to 12 h after injection, the total amount of radioactivity
and the amount of radioactivity not incorporated into DNA were determined.
Two separate batches of stage-9 blastulas and one batch of stage-29 embryos
were injected and at each time interval, five embryos from each batch were
examined.
The results (Fig. 1) show that thymidine is rapidly incorporated into DNA
during the first 4 h and then more slowly up to 12 h after injection. The rate of
incorporation taken over 8 h was similar in both batches of blastulas and in the
stage-29 embryos. Chromatographic analysis of the unincorporated radioactivity at 8 h showed that approximately 70 % (equivalent to 9000 cpm) was in
thymidine and TTP, and hence available for incorporation into DNA.
To determine whether this amount of radioactivity could be detected autoradiographically if incorporated into DNA, blastulas were injected with a dose
DNA and RNA synthesis in Xenopus germ cells
2
4
6
8
Time after injection (h)
10
19
12
Fig. 1. Time of availability of label after injection of 008 /*Ci [3H]thymidine into
two batches of blastulas (O, • ) and stage-29 embryos (A), expressed as the
percentage of the original injected label remaining unincorporated into DNA.
of [3H]thymidine corresponding to the amount of radioactivity still available for
incorporation into DNA. One hour later they were fixed and prepared for
autoradiography. After an exposure of 1 week, distinct nuclear labelling was
observed throughout the embryo. We concluded therefore that the amount of
[3H]thymidine injected in our experiments (0-08 fiCi) ensured that DNA
synthesis could be studied any time up to 8 h after injection.
In other embryonic systems, the time of availability varies widely. For
example, after injection of 0-5/*Ci/g [3H]thymidine into larval Pleurodeles
wait Hi, the label was available for only 1 h whereas it was available for 4 h in
young metamorphosed specimens (Brugal, 1971; Chibon & Brugal, 1973). At
the other extreme, [3H]thymidine was still available one week after injection into
larval and pupal Cecropia (Bowers & Williams, 1964). It is therefore obviously
necessary to determine the time of availability of [3H]thymidine in each system.
This had not previously been done in X. laevis embryos.
DNA synthesis
A large number (500-1000) eggs were fertilized simultaneously, providing
a batch of embryos developing at the same rate. At late blastula (stage 9-10),
20 embryos were injected with 0-08 ju,Ci [3H]thymidine and fixed 2 h later
(stage 11). At this time, a further 20 embryos from the stock batch were injected
with the same amount of [3H]thymidine and fixed 8 h later. This sequence of
20
MARIE DZIADEK AND K. E. DIXON
Table 3. DNA synthesis in presumptive primordial germ cells in X. laevis embryos,
expressed as percentage of germ cells labelled after 8-h pulses with 0-08 f.iCi [ 3 i/]thymidine.
The developmental intervals shown indicate the times of injection and fixation
expressed as stages of development.
Interval
(stages)
10-11
11-16
16-22
22-26
26-31
31-33
33-34/35
34/35-36
36-37/38
37/38-39
39-40
40-41
41^2
42-43
43-44
No. of
embryos
examined
No. of
germ cells
identified
3
2
3
3
2
2
2
3
3
3
3
3
3
3
32
10
17
13
18
16
11
22
42
46
61
83
71
81
9
280
Labelled
germ cells
(%)
56-3
600
35-3
7-7
33-3
250
0
0
0
0
0
0
0
0
0
a
Fig. 2. Autoradiograph of presumptive primordial germ cell in late gastrula embryo
injected with [3H]thymidine and fixed 1 h later. The nucleus is heavily labelled (only
a proportion of the grains are in focus). Pigment granules are visible around the
periphery of the germ plasm, x 1300.
Fig. 3. Presumptive primordial germ cell in mitosis in stage 22 embryo. The germ
plasm lies at either pole of the spindle, x 1750.
DNA and RNA synthesis in Xenopus germ cells
21
Table 4. DNA synthesis in presumptive primordial germ cells in X. laevis embryos,
expressed as percentage of germ cells labelled after 2-h pulses between stages
16-28 and 8-h pulses between stages 28 and 45 with 0-08 fid [3H]thymidine
Interval
(stages)
No. of
embryos
examined
No. of
germ cells
identified
Labelled
germ cells
(%)
16-17
17-19
19-21
21-23
23-24
24-25
25-23
28-32
32-33
33-35
35-36/37
36/37-33
38-40
40-41
41-42
42-43
43-44
44-45
3
3
3
2
3
3
3
2
2
3
3
3
3
3
3
3
2
2
26
15
25
26
2?
42
16
16
16
21
20
36
55
45
66
66
40
47
69-2
73-3
480
26-9
20-7
42-9
68-8
250
18-8
9-5
0
0
0
0
0
0
0
0
injections followed by fixation 8 h later was repeated until the embryos reached
stage 44, by which time the germ cells had left the endoderm. In this way,
development between early gastrula and stage 44 was completely covered by
14 pulses each of 8-h duration and one pulse of 2 h. The fixed embryos were
prepared for autoradiography and the percentages of labelled germ cells and
endoderm cells determined.
[3H]thymidine incorporation by presumptive primordial germ cells
Labelled germ cells were observed in all embryos from stages 10-33 (Fig. 2)
but in all 32 embryos between stages 33-44, none of the 697 germ cells examined
was labelled (Table 3). In all cases, endoderm cells immediately adjacent to the
germ cells were labelled, indicating that [3H]thymidine was available to the
germ cells. Between stages 10—33, the percentage of labelled germ cells was
highest in the earliest stages (10-16), declined sharply between stages 22-26 and
then rose again between stages 26-33.
A second experiment was performed to confirm that the germ cells did not
synthesize DNA after stage 33 and also to investigate more closely the changes
in the patterns of DNA synthesis in the germ cells with development. Embryos
were injected and fixed so that the period of development between stages 16-28
was covered by seven pulses each of 2-h duration, and stages 28-45 by 11 pulses
each of 8-h duration. The results are summarized in Table 4 and show that after
22
MARIE DZIADEK AND K. E. DIXON
Table 5. DNA synthesis in presumptive primordial germ cells in X. laevis embryos,
expressed as percentage of germ cells labelled after 2-h pulses with 0-08 /id
[3H]-thymidine
Interval
(stages)
No. of
embryos
examined
No. of
germ cells
identified
Labelled
germ cells
'%)
20-21
21-22
22-23
23-24
3
6
2
2
18
48
13
16
5-6
31 3
53-9
87-5
stage 35, none of the 375 germ cells examined were labelled. Furthermore, the
percentage of labelled germ cells gradually decreased from 73 % at stages 17-19
to 20 % at stage 23, 24 and then increased to 69 % again at stages 26-28. In
a third experiment covering development between stages 21-24, embryos at
stage 21 contained only 5-6 % labelled germ cells compared with 87 % at
stages 23-24 (Table 5). These results therefore confirm that the presumptive
primordial germ cells do not synthetize DNA at about stage 22 or after about
stage 33.
Germ cells in mitosis were observed at only two stages in development. Three
mitotic germ cells were recorded in stage-12 embryos and four in stage-22
embryos (Fig. 3).
These results suggest the following interpretation. The germ cells synthesize
DNA and divide as the embryo develops from blastula to stage 44, i.e. while
they are in the endoderm. Thus our earlier results (Whitington & Dixon, 1975;
Dziadek & Dixon, 1975) are confirmed. However, DNA is not synthesized
during the whole of this time but only between blastula and about stage 33. The
reduction in the percentage of labelled germ cells at stages 21-24 divides the
period when DNA is synthesized into two relatively discrete phases. The germ
cells apparently go through two cycles, with mitoses at about stage 21-24 and
again after stage 33. The identification of dividing germ cells at stage 22 and the
increase in number of germ cells at about stages 36-39 lend support to this
suggestion. The presence of dividing germ cells in stage-12 embryos similarly
suggests the possibility that a round of mitotic activity may also occur about
this time. Many determinations of germ cell numbers in this laboratory indicate
that the presumptive primordial germ cells divide, on the average, three times
(more rarely twice in some batches of embryos) between blastula and stage 44.
If three divisions take place, then according to the results presented here, the
first may occur early in this period, at about stage 12, the second towards the
middle of the period, at about stage 22 and the third towards the end, at about
stage 36.
DNA and RNA synthesis in Xenopus germ cells
23
50-
40-
JJ 30
20 -
10-
10
11
16
22
26
31
33
35
36
38
39
40
42
43
44
Stages of development
Fig. 4. Percentage of labelled endoderm cells after 8-h pulses of [3H]thymidine in
embryos at stages 10-44 of development. Each interval, e.g. stages 16-22, gives the
stage injected with [3H]thymidine and the stage when fixed, and represents 8-h
development time.
[3H]thymidine incorporation by endoderm cells
At all stages of development from early gastrula to stage 44, at least 25 % of
the endoderm cells were synthesizing DNA. The percentage of labelled cells
fluctuated, however, with peaks of DNA synthesis at approximately 48-h
intervals, at stages 16, 36 and 44 (Fig. 4). These results suggest that endoderm
cells pass through a number of more or less synchronous cell cycles between
early gastrula and stage 44. The patterns of DNA synthesis in the endoderm
24
MARIE DZIADEK AND K. E. DIXON
cells of Xenopus embryos from blastula to stage 18 were studied by Graham &
Morgan (1966). The labelling indices reported by them were higher for comparable incubation times than those recorded in this study but the discrepancy
may lie in the criteria used to classify cells as labelled.
RNA synthesis
A large number of synchronously developing embryos was obtained as
previously described. Beginning at late blastula, 12 embryos were injected with
0-08 /*Ci [3H]uridine and fixed 1 h later. At this time another 12 embryos were
injected and subsequently fixed 1 h later. Between late blastula and late
gastrula, five such intervals of 1 h were used to determine the time when RNA
synthesis was initiated in the presumptive primordial germ cells and in the
endoderm cells. Other groups of embryos were progressively injected and fixed
1 h later at ten different stages between neurula and stage 45 (see Table 6). Three
embryos from each injected group were sectioned and autoradiographed and the
proportions of labelled germ cells and endoderm cells determined.
In blastula to late gastrula embryos, only some of the cells were synthesizing
RNA. In blastula-stage embryos, a small proportion of the cells in the ectoderm
over the blastocoel was labelled (Fig. 5) and, in early gastrula stages, all the
ectoderm cells and the endoderm cells next to the blastocoel were labelled
(Fig. 6). As gastrulation proceeded, endoderm cells progressively further away
from the blastocoel began to incorporate [3H]uridine into RNA until in the
completed gastrula, only cells in the yolk plug remained unlabelled. After
gastrula, labelled cells were distributed throughout the embryo.
This pattern of labelling could indicate that not all regions of the embryo are
accessible to the labelled substrate. However, the following observations do not
support this contention:
(i) the pattern of labelling was constant for each developmental stage;
(ii) although the embryos were always injected in the animal hemisphere, the
depth of injection could not be controlled precisely, and hence the distribution
of labelled cells would have varied between embryos if rapid diffusion through
the embryo was not possible;
(iii) the distribution of labelled cells changed progressively with development
although the amount of label injected and the incubation time remained
constant;
(iv) at later stages of development, all cells incorporated [3H]uridine even
though the dimensions of the embryo had changed, more cells were present and
the cells were probably bound more tightly together as histogenesis began;
(v) injections of [3H]thymidine produced uniform labelling over the whole
embryo within 10 min (Graham & Morgan, 1966).
In order to establish conclusively that the patterns of RNA synthesis in
gastrulating embryos were due to the progressive initiation of RNA synthesis
along the animal-vegetal axis and not to variable diffusion rates, a batch of
DNA and RNA synthesis in Xenopus germ cells
Fig. 5. Autoradiograph of late blastula embryo after 1-h pulse of [3H]uridine. The
nuclei of presumptive ectoderm cells lining the blastocoel are heavily labelled
whereas endodermal cells and cells towards the surface of the embryo have not begun
to synthesize RNA. x 110.
Fig. 6. Autoradiograph of early gastrula embryo after 1-h pulse of [3H]uridine.
Endoderm nuclei (insert a, b) near blastocoel are labelled whereas germ cell deeper
in endoderm (inset c) is not. Presumptive ectoderm and mesoderm cells adjacent to
the blastocoel are labelled whereas those further away are not. x 110.
Fig. 7. Autoradiograph of stage-39 embryos after 1-h pulse of [3H]uridine. Cells in
the mesoderm and ectoderm (arrows) are heavily labelled, but nuclei in the yolky
endoderm are only lightly labelled, x 200.
Fig. 8. Autoradiograph of stage-16 embryos after 1-h pulse of [3H]uridine. The
nucleus is lightly labelled but there are no grains over either the germ plasm or the
general cytoplasm. Pigment granules occur between the yolk platelets, x 1100.
25
26
MARIE DZIADEK AND K. E. DIXON
stage-9 embryos was injected with 0-08/*Ci [3H]uridine. One third of the
injected embryos were incubated at 20 °C for 1 h then fixed and prepared for
autoradiographic examination. The remaining embryos were incubated at 8 °C
for either 1 or 2 h to allow diffusion of the substrate from the point of injection
without incorporation of the label into RNA, then incubated for a further 1 h
at 20 °C after which they were fixed and prepared for autoradiographic
examination. Three embryos in each group were examined. The patterns of
labelling were identical in control and experimental embryos and similar to
those we had previously observed in embryos fixed 1 h after injection. Therefore, we concluded that RNA synthesis is initiated progressively along the polar
axis between blastula and gastrula. Bieliavsky & Tencer (1960) have previously
reported the existence of a gradient in RNA synthesis in gastrula-stage embryos.
Our results differ from those of Bachvarova & Davidson (1966) in two major
respects. Firstly, in our experiments RNA synthesis was initiated at a slightly
later stage than that reported by Bachvarova & Davidson (1966) but this
difference is small enough to be dismissed. Secondly, we have noted the existence
of a polar gradient in the initiation of RNA synthesis whereas Bachvarova and
Davidson reported that the presumptive endoderm and mesoderm cells become
activated first, followed shortly after by the presumptive ectoderm cells. There
is no obvious way to account for the differences in these results except as consequences of the different procedures used in exposing the cells of the embryo
to the labelled substrate. In our experiments [3H]uridine was injected into the
intact embryo and we have attempted to show experimentally that the patterns
of labelling are not due to any deficiencies resulting from the use of the injection
procedure. On the other hand Bachvarova & Davidson (1966) immersed embryos
cut into halves in medium containing [3H]uridine. If this procedure disrupts the
polar gradients usually assumed to be present in the egg, then the observed
pattern of RNA synthesis would differ from that in the intact embryo.
In later embryos, mesoderm and ectoderm cells were always more heavily
labelled than endoderm cells (Fig. 7), indicating they were synthesizing RNA
at a greater rate. Denis (1964) has earlier reported making a similar observation.
[3H]Uridine incorporation by presumptive primordial germ cells
The percentages of germ cells synthesizing RNA at the different stages of
development examined are shown in Table 6.
Between blastula and gastrula, the germ cells conformed to the general
patterns of RNA synthesis in the embryo. That is, at the blastula stage, no
germ cells were incorporating [3H]uridine but as RNA synthesis was gradually
initiated through the embryo, so the germ cells also began to synthesize RNA.
The time of initiation of RNA synthesis in germ cells was therefore dependent
on their position within the embryo, which varied both between embryos and
between different germ cells in one embryo (see also Whitington & Dixon, 1975)
Germ cells close to the blastocoel started RNA synthesis at early gastrula, while
DNA and RNA synthesis in Xenopus germ cells
27
Table 6. RNA synthesis in presumptive primordial germ cells in X. laevis embryos,
expressed as percentage of germ cells labelled after 1-h pulses with 0-08 juCi
[zH]uridine (sum of three experiments)
Interval
(stages)
9-10
12
\2\
22/23
24
26
29
30
33
36
39
41
45
No. of
embryos
examined
No. of
germ cells
identified
Labelled
germ cells
(%)
7
11
2
2
5
3
3
2
2
1
4
3
3
27
61
7
7
23
22
10
8
15
17
40
32
35
0
45-9
100
100
60-9
13-6
0
0
0
0
0
0
0
those at the yolk plug began last of all, at late gastrula. However, when RNA
synthesis began in the germ cells it was always in concert with neighbouring
endoderm cells. Thus the initiation of RNA synthesis appears not to be directly
related to the movement of the germ plasm to the nucleus which takes place in
all germ cells by early gastrula (stage 10) (see also Whitington & Dixon, 1975),
but many germ cells were not labelled until later stages. It seems likely therefore
that the initiation of RNA synthesis is controlled by some external factor which
does not distinguish between germ cells and other endoderm cells. The pattern
of labelling observed suggests that the blastocoel may have some role in this
process. All germ cells had started RNA synthesis by the beginning of neurulation (stage 12^), but by stage 24 the percentage of germ cells incorporating
[3H]uridine had decreased markedly. In all 18 embryos between stages 29-45,
none of the 157 germ cells examined were labelled, although all of the endoderm
cells continued to synthesize RNA.
An additional experiment was performed in an attempt to determine whether
the RNA synthesized in the nuclei of the germ cells between stages 10-26 was
eventually transported to the cytoplasm. A batch of embryos was injected with
0-08 /.id [3H]uridine at stages 10| and 21/22 and incubated for 4, 6 and 8 h
before fixation and preparation for autoradiographic analysis. Seventy-eight
germ cells were identified in 13 embryos and in no case could we conclude that
the cytoplasm was labelled. This observation should be treated with caution,
however, because the high proportion of the cytoplasm occupied by the germ
plasm (which was not labelled) and by the yolk granules means that any labelled
RNA in the cytoplasm would be distributed widely and therefore would be
more difficult to visualize autoradiographically. Denis (1964) has reported that
28
MARIE DZIADEK AND K. E. DIXON
endodermal cells exhibit solely nuclear RNA synthesis until stage 27, a finding
which is in agreement with our observations. Small nucleoli were visible in the
germ cells in the earliest of these embryos, and therefore, if transport of RNA to
the cytoplasm does not take place, it cannot be accounted for by the absence of
nucleoli.
DISCUSSION
Our preliminary experiments showed that 0-08 /*Ci [3H]thymidine can be
injected into early embryos of X. laevis with an acceptable degree of reproducibility, with only slight leakage, and without affecting cellular proliferation.
After injection, the thymidine was available for incorporation into DNA by all
cells of the embryo for at least 8 h, and therefore we based our study of DNA
synthesis in presumptive primordial germ cells on pulses of 8 h or less.
The results of this study have confirmed the suggestion previously made by
Whitington & Dixon (1975) that the presumptive primordial germ cells divide
as the embryo develops between blastula and stage 44 (see also Dziadek &
Dixon, 1975). After comparing the numbers of germ cells at these different
stages, Whitington & Dixon (1975) proposed that two or three divisions of each
cell took place. The results reported here, and other unpublished studies from
our laboratory, indicate that three divisions are more usual. The first division
probably takes place during gastrulation, the second relatively synchronously
at about stages 22-24 and the third series probably again relatively synchronously, considering the short period of time available before migration out of the
endoderm, at about stages 36-39. The germs cells then remain mitotically
quiescent until after they have entered the genital ridges, when a second
proliferative phase begins about stages 48-52 (Kalt & Gall, 1974; Ziist & Dixon,
in preparation), about 5-12days later according to the 'Normal Table of
Development' (Nieuwkoop & Faber, 1967). In other animals in which the
proliferative activity of presumptive primordial germ cells has been studied,
either cell division is apparently more or less continuous as in mammals and
birds (see Hardisty, 1967) or there are periods of mitotic activity separated by
periods of quiescence, e.g. Drosophila (Sonnenblick, 1950), fish, amphibians
and reptiles (Hardisty, 1967). However, the synchrony of the divisions has not
previously been investigated.
The germ cells begin to synthesize RNA during gastrulation, at the same time
as adjacent cells in the embryo, apparently in response to a polar gradient.
Germ cells in different regions of the endoderm begin synthesis at different
times but always in concert with neighbouring endoderm cells. Thus, RNA
synthesis in the germ cells is not required for the movement of the germ plasm
to the nucleus, a shift which occurs relatively synchronously so that by early
gastrula (stage 10), germ plasm surrounds the nucleus in all germ cells. However,
the possibility that the germ plasm influences RNA synthesis quantitatively or
qualitatively cannot yet be dismissed. A causal relationship between RNA
DNA and RNA synthesis in Xenopus germ cells
29
synthesis and division of the germ cells also seems unlikely, since the first
mitosis probably precedes the onset of [3H]uridine incorporation in the germ
cells. If particular division patterns can be initiated without the release of
information, i.e. without RNA synthesis, the information specifying these
patterns may be encoded in the egg. Furthermore, if our tentative findings that
the incorporated uridine is confined to the nucleus can be confirmed, the
likelihood that RNA synthesis is required for migration of the germ cells out
of the endoderm is diminished since intranuclear RNA synthesis presumably
cannot affect cytoplasmic functions. However, if these possibilities are eliminated,
other explanations for the function of the RNA synthesized are not obvious.
Although it is now well established that most of the RNA synthesized in the
nucleus never reaches the cytoplasm (Scherrer et al. 1970) the failure of any
newly synthesized RNA to be transported to the cytoplasm would be difficult
to interpret.
The three divisions of the presumptive primordial germ cells in the endoderm
represent a discrete period of proliferative activity and have been termed
cloning divisions (Whitington & Dixon, 1975), implying that the number of
germ cells increases but the determined state of the cells does not alter. If the
RNA synthesized in the nucleus of the germ cells during this phase does not
reach the cytoplasm, change in the determined state as a result of new nucleocytoplasmic interactions seems less likely and the suggestion of Whitington &
Dixon (1975) gains some substance. This compartment of the germ cell lineage
is distinguishable on the one hand from the first compartment of the lineage in
which the initial clone of cells is segregated by unequal divisions and in which
RNA is not synthesized and on the other hand separated by a period of mitotic
inactivity from the next compartment in which a second proliferative phase and
sexual differentiation take place (Ziist & Dixon, 1976).
The point at which a separate and distinct germ-cell lineage is established is
difficult to determine. The differences in patterns of nucleic acid synthesis in
germ cells and endoderm cells indicate that germ cells are differentiated from
endoderm cells at least by stage 24 when RNA synthesis ceases in germ cells.
However, differences in thymidine-labelling indices of germ cells and endoderm
cells immediately after gastrulation may suggest that germ cells have already
differentiated from neighbouring endoderm cells. If the germ plasm is involved
in the establishment of the germ cell lineage, it may act shortly after moving to
its perinuclear position by exerting some influence on synthetic activities
within the nucleus. However, its mode of action, as with other ooplasmic
localizations, remains obscure. Our results at least eliminate the possibilities
that, at these stages in development, RNA is synthesized in the germ plasm or
that the presence of the germ plasm around the nucleus inactivates the nucleus.
EMB 37
30
MARIE D Z I A D E K AND K. E. DIXON
REFERENCES
L. E. & MCCLURE, W. O. (1973). An improved scintillation cocktail of highsolubilizing power. Analyt. Biochem. 51, 173-179.
APELGOT, S. (1966). Actions letales comparees des rayons x et des beta du tritium sur
E. coli. Int. J. Radiat. Biol. 10, 495-508.
BACHVAROVA, R. & DAVIDSON, E. H. (1966). Nuclear activation at the onset of amphibian
gastrulation. /. exp. Zool. 163, 285-296.
BIELIAVSKY, N. & TENCER, R. (1960). Etude de l'incorporation de l'uridine tritiee dans les
oeufs d'amphibiens. Expl Cell Res. 21, 279-285.
BLACKLER, A. W. (1970). The integrity of the reproductive cell line in the Amphibia. Curr.
Top. Devi Biol. 5, 71-87.
BOUNOURE, L. (1934). Recherches sur la lignee germinale chez la grenouille rousse aux
premiers stades du developpement. Annls Sci. nat. lOe ser. 17, 67-248.
BOUNOURE, L. (1964). La lignee germinale chez les batraciens anoures. In VOrigine de la
Lignee germinale (ed. E. Wolff), pp. 207-234. Paris: Hermann.
BOWERS, B. & WILLIAMS, C. M. (1964). Physiology of insect diapause. XIII. DNA synthesis
during the metamorphosis of the Cecropia silkworm. Biol. Bull. mar. biol. Lab., Woods
Hole 126, 205-219.
BROWN, D. D. & LITTNA, E. (1966). Synthesis and accumulation of DNA-like RNA during
embryogenesis of Xenopus laevis. J. molec. Biol. 20, 81-94.
BRUGAL, G. (1971). Relations entre la proliferation et la differentiation cellulaires: etude
autoradiographique chez les embryons et les jeunes larves de Pleurodeles waltlii Michah
(Amphibien, Urodele). Devi Biol. 24, 301-321.
CHIBON, P. & BRUGAL, G. (1973). Duree de disponsibilite de la thymidine exogene chez la
larve et le jeune du triton Pleurodeles waltlii Michah. Ann. Embryol. Morph. 6, 81-92.
DECROLY, M., CAPE, M. & BRACHET, J. (1964). Studies on the synthesis of ribonucleic acids
in embryonic stages of Xenopus laevis. Biochim. biophys. Ada 87, 34-39.
DEITCH, A. D., WAGNER, R. & RICHART, R. M. (1968). Conditions influencing the intensity
of the Feulgen reaction. J. Histochem. Cytochem. 16, 371-379.
DENIS, H. (1964). Effets de l'actinomycine sur le developpement embryonaire. Etude autoradiographique influence de l'actinomycine sur la synthese des acides nucleiques. Devi
Biol. 9, 458-472.
DZIADEK, M. & DIXON, K. E. (1975). Mitosis in presumptive primordial germ cells in postblastula embryos of Xenopus laevis. J. exp. Zool. 192, 285-291.
GRAHAM, C. F. & MORGAN, R. W. (1966). Changes in the cell cycle during early amphibian
development. Devi Biol. 14, 439-460.
GURDON, J. B. (1968). Nucleic acid synthesis in embryos and its bearing on cell differentiation.
Essays in Biochemistry 4, 26-68.
GURDON, J. B. (1974). The Control of Gene Expression in Animal Development. Oxford:
Clarendon Press.
GURDON, J. B. & WOODLAND, H. R. (1969). The influence of the cytoplasm on the nucleus
during cell differentiation, with special reference to RNA synthesis during amphibian
cleavage. Proc. R. Soc. B 173, 99-111.
HARDISTY, M. W. (1967). The number of vertebrate primordial germ cells. Biol. Rev. 42,
265-287.
JONES, R. MCC. (1954). Handbook of Microscopical Technique, 3rd ed. New York: Hafner.
KALT, M. R. & GALL, J. G. (1974). Observations on early germ cell development and
premeiotic ribosomal DNA amplification in Xenopus laevis. J. Cell Biol. 62, 460-472.
NIEUWKOOP, P. D. & FABER, J. (1967). Normal Table of Xenopus laevis (Daudin), 2nd ed.
Amsterdam: North-Holland.
ROGERS, A. W. (1973). Techniques of Autoradiography, 2nd ed. Amsterdam: Elsevier.
ANDERSON,
SCHERRER, K., SPOHR, G., GRANBOULAN, N., MOREL, C , GROSCLAUDE, J. & CHEZZI, C.
(1970). Nuclear and cytoplasmic messenger-like RNA and their relation to the active
messenger RNA in polysomes of HeLa cells. Cold Spring Harb. Symp. quant. Biol. 35,
539-554.
DNA and RNA synthesis in Xenopus germ cells
31
B. P. (1950). The early embryology of Drosophila melanogaster. In Biology of
Drosophila (ed. M. Demerec), pp. 62-167. New York: Hafner.
VOLKONSKY, M. (1928). Sur une nouvelle modification de la technique d'Altmann. Bull.
Histol. appl. Physiol. Path. 5, 220-222.
WHJTINGTON, P. M C D . & DIXON, K. E. (1975). Quantitative studies of germ plasm and germ
cells during early embryogenesis of Xenopus laevis. J. Embryol. exp. Morph. 33, 57-74.
WOLF, D. P. & HEDRICK, J. L. (1971). A molecular approach to fertilization. II. Viability and
artificial fertilization of Xenopus laevis gametes. Devi Biol. 25, 348-359.
WOODLAND, H. R. & GURDON, J. B. (1968). The relative rates of synthesis of DNA, sRNA
and rRNA in the endodermal region and other parts of Xenopus laevis embryos. /. Embryol.
exp. Morph. 19, 363-385.
SONNENBLICK,
{Received 19 December 1975, revised 26 July 1976)
3-2