Biol Cell (1990) 68, 119-127 119

© Elsevier, Paris

Original article

Mitochondrial development during Drosophila oogenesis:

distribution, density and in situ RNA hybridizations
Sylvette Tourmente *~, Pierre Lecher 2 Fabienne Degroote 3, Michel Renaud 4
t Laboratoire de Biochimie M6dicale, Facultd de Mddecine, Universit~ Clermont I, place Henri-Dunant, F-63000
Clermont-Ferrand ; 2Laboratoire de Biochimie ; SLaboratoire de Gdndtique, Universitd Blaise Pascal, URA CNRS 360,
Aubi~re; 4Centre O R S A N de recherche en Biotechnologie, Paris, France
(Received 16 October 1989; accepted 17 December 1989)

Summary - The changes in distribution and density of mitochondria and the level of mitochondrial RNA during Drosophila oogenesis
were studied simultaneously in the 3 cell types ie follicle cells, nurse cells and oocyte, making up the egg chamber. Up to stage 6,
mitochondrial density (mitochondrial and cellular areas ratio) was elevated and increased similarly in both follicle and nurse cells.
Thereafter the mitochondrial density of follicle cells continued to increase and that of the nurse cells declined markedly while the
nurse cell mitochondria assembled in dense groups and decreased in size. This can be related to a transfer of nurse cell cytoplasm,
including mitochondria, to the oocyte. In the oocyte from stage 4 to stage 7 we observed a significant decrease of the mitochondriai
density due to the absence of mitochondrial biogenesis. Then the cytoplasm transfer caused mitochondrial density to increase up
to the level found in the nurse cells at the end of oogenesis. The mature oocyte contains enough mitochondria to supply 15 000 somatic
cells. Our results strongly suggest that the variations in size, distribution and density of mitochondria relate to the particular energetic
requirements of the different cell types during the first half of oogenesis. Later they relate to the developmental requirements of
the nurse cells and the oocyte, in particular the storage of mitochondria in the oocyte. The level of mitochondrial RNA was studied
through in situ hybridization. Throughout oogenesis the follicle and nurse cell RNA evolved similarly. Up to stage 9, there was no
change in RNA densities in these cells, suggesting a correlation with the cell volume and/or the nuclear DNA content. Thereafter
the cellular RNA concentration declined rapidly. In the oocyte the RNA concentration evolved differently especially from stage 10
to the end, the RNA density being stabilized. This can be related to the injection of nurse cell mitochondria, followed by their assign-
ment to reserve status. Our results suggest that the mt RNA density is under extramitochondrial control mechanisms.

Drosophila oogenesis / mitochondria / mt DNA expression I in situ hybridization

Introduction coded by nuclear genes whereas the RNAs are mitochon-

Meroistic oogenesis is a rapid process of oocyte develop-
ment characteristic o f Diptera and certain other
holometabolous insects.
The egg chamber is the functional unit of oogenesis in
Drosophila and consists of 3 cell types. A single oocyte
is connected through cytoplasmic bridges to 15 sister nurse
ceils. Follicle cells surround this 16-cell complex. During
oogenesis nurse cells and follicle cells reach a high degree
of D N A polyploidization. Then only a little over 3 days
are required for an egg chamber to develop from stage 1
to stage 14, the final stage of oogenesis.
The oocyte mass increases around 100 000 times during
this period [19]. This increase is essentially due to the ac-
cumulation o f storage proteins and RNA, and requires a
large amount of energy produced by mitochondria via the
respiratory chain. The A T P synthetase and 3 multienzyme
complexes a m o n g 4 of the respiratory chains consist of
both nuclear and mitochondrial encoded polypeptides.
Likewise, the mitochondrial translation enzymes are en-
* Correspondence and reprints
Abbreviations: mt DNA, mitochondrial DNA; mt RNA, mitochondrial
RNA; rRNA, ribosomal RNA
drially encoded [for a review, see 30].
This dual origin involves coordinated interactions bet-
ween the cellular energy requirement, the mitochondrial
activity and the expression of the two genomes. Moreover
in oogenesis these coordinated interactions must be con-
nected with a developmental process since mitochondria
are also stored in the egg.
Considering the molecular analysis of m t D N A structure
and expression in higher organisms [for a review, see 30],
considerable progress has been made during the past
decade. However the cellular control mechanisms of the
m i t o c h o n d r i a l b i o s y n t h e t i c activities (replication,
transcription, protein synthesis) are still poorly under-
stood. This is due in part to the lack o f detailed informa-
tion available at the cellular level. Most of the data are
related to mitochondrial D N A replication and are in agree-
ment with an extramitochondrial control mechanism [8,
11, 20, 37]. However, the existence of different replicative
mitochondrial subpopulations [37] in the Xenopus oocyte
suggeste a developmental control of mitochondriogenesis.
Results on mitochondrial R N A and protein synthesis ap-
pear fragmentary and often contradictory. This is probably
due to the various biological materials and conditions us-
ed. Some of the results obtained with damaged cells or cells
under stress-like conditions suggest a mitochondrial gene
specific expression [30] while others suggest a global
qualitative control o f the entire genome [36].
120 s Tourmente et a!
Little or only f r a g m e n t a r y data are available on the
development o f the mitochondrial mass and the genomic
expression in connection with oogenesis or any develop-
mental process. The m o r e refined analysis concerns the
X e n o p u s oogenesis [10, 29, 37] for which heterogeneous
mitochondrial distribution, replication activities and R N A
contents were described. This suggests the existence o f dif-
ferent mitochondrial subpopulations in a single cell, which
further complicates the situation.
The Drosophila oogenesis with its 3 well defined cell
types seems to us a particularly interesting model for an
extensive study o f the mitochondrial development and the
accumulation o f mitochondrial R N A s .
In this paper we present a comparative study o f the
developmental changes, mitochondrial shape, cellular
distribution and density, and the accumulation o f
mitochondrial R N A s in the 3 cell types during the m a j o r
part o f oogenesis. We show i m p o r t a n t variations firstly
o f the mitochondrial cellular distribution and o f their den-
sity, secondly o f the m t R N A levels, both depending on the
cell type and the developmental stage.
Materials and Methods
Biological material
Drosophila melanogaster females (Canton-S strain) bred at 18°C,
approximately 5 days old were utilized.
Ultrastructure
The ovaries of Canton-S females were dissected, f'Lxedfor 15 min
in pH 7.2 Ringer solution with 4% paraformaldehyde and 0.1 °70
glutaraldehyde, then post-ffLxed for 2 h in a Ringer solution with
2°7o osmium tetroxide. The ovaries were then dehydrated with
increasing concentrations of ethanol and propylene oxide [23],
then embedded in Epon resin. Ultra-thin sections were contrasted
with 2% uranyl acetate and lead citrate [31]. This method was
chosen because it produces good contrast for mitochondrial
observation.
To facilitate the identification of different stages, toluidine-
blue stained semi-thin sections were prepared along with the
ultrathin sections. The stages were identified according to King's
criteria [lff].
The ulfra-thin sections were examined with a Jeol electron
microscope.
In situ RNA hybridization on frozen sections
The probe used represented an equimolecular mixture of 4 recom-
binant plasmids, pDmt EB, pDmt EC, pDmt ED, pDmt HCD
covering the entire mt DNA coding sequence except the small
ribosomal RIqA (rRNA) [4]. The 3H-labelled probes were
prepared as described previously [5, 16].
Frozen sections of ovaries were prepared as follows: 50-100
ovaries rapidly dissected in ice-cold Ringer solution were transfer-
red to a drop of OCT embedding medium (Miles Laboratory)
mounted on a cryostat specimen holder and then frozen by im-
mersing in liquid nitrogen. 8-/zm sections were placed on gelatin-
coated slides [14], incubated 2 min at 50"C and then dried at
room temperature (2 h). The sections were fbced in a 4%0 parafor-
maldehyde, 0.I M phosphate buffer, pH 7.0 (20 min, room
temperature) and dehydrated in a graded series of ethanol con-
centrations. The dried sections can be stored at room temperature
for a long time before hybridization.
In situ hybridization, exposure and development were carried
out according to [16]. Slides were examined by light microscopy.
Determination of grain densities
Silver grains were counted manually on micrographs of sections
using a Digicont counter. Cellular surface areas were measured
with transparent millimeter paper.
The cellular density of grains was calculated by dividing the
number of grains by the whole cellular surface area (including
cytoplasm and nuclei).
The mean values were calculated with their 95°7o confidence
x+t.s.
intervals ~ ; where s is the standard deviation, n the number
of observations, and t represents Student's t.
Identification of developmental stage
Developmental stages were identified according to King's
parameters [19], taking into account the type of preparation. In
fact egg chambers of frozen sections were comparatively larger
than chambers of semithin sections in the resin.
Results
Cellular distribution and density o f the mitochondria
Whatever the stage and the cell type, the m i t o c h o n d r i a
have m a n y cristae. We determined the width o f the
m i t o c h o n d r i a ; it did not differ significantly between cell
types or developmental stages. F o r this reason we give in
table I the average size o f the m i t o c h o n d r i a o f the 3 cell
types for all stages combined. H o w e v e r the average length
o f the m i t o c h o n d r i a differed. The oocyte m i t o c h o n d r i a
were always the shortest and their surface area was ap-
proximately 50°7o smaller t h a n that o f the other cell types
except in nurse cells f r o m stage 9 and later stages.
In the follicle cells and in the oocyte no differences in
mitochondrial length were observed between the different
stages. H o w e v e r the maximal length o f the m i t o c h o n d r i a
o f the nurse cells increased f r o m stage 3 to 8 ( f r o m 1 to
1.4 Fzm, respectively), and decreased a r o u n d stage 9 (to
0.9/~m, a length approximately equal to that o f the oocyte
mitochondria).
Table I. Mean values of the maximal length, minimal width and
area of mitochondria in the 3 types of ovarian cells. The mean
values (X) are given according to the equation:
x+_t.__
S
S
where t is the Student t-distribution value at P = 0 . 0 5 and ~nn
is the SEM (standard error mean deviation). ~Assuming an
~LW
ellipse A = • 2Mean value of the maximal length of nurse
4
cell mitochondria calculated from the values at different stages
(see text): Stage 3 : 1 /zm; stage 4, 5, 6 : 1 . 2 5 /~m; stage 7:
1.45 tzm ; stage 8 : 1.35/~m ; stage 10 : 0.9/~m.
Oocyte Nurse cells Follicle cells
(n = 360) (n = 2080) (n = 330)
Length (L)/~m 0.82+0.04 1.21 +0.182 1.15 +0.02
Width (W)/zm 0.2 +_0.03 0.25 +_0.02 0.24+_0.02
Area I (A) b t m 2 0.13+_0.03 0.24+_0.07 0.22_+0.03
 Mitochondrial development during Drosophila oogenesis 121

Follicle cells mitochondria are distributed over the entire cytoplasm.

The follicle cells surrounding the egg chamber have a dense These ceils disappear before stage 10.
cytoplasm rich in ribosomes. During the entire oogenesis,
the concentration o f the mitochondria is at its highest in Nurse cells
these cells until they form the chorion at stages 12 and 13. The 15 nurse cells, like the follicle cells, also contain a
Figure 1B shows that the number of mitochondria in dense cytoplasm loaded with ribosomes. The vesicles of
the follicle cells as well as the size o f the ceils themselves the endoplasmic reticulum are smaller than those o f the
increase exponentially. From stages 3 to 8 growth procedes follicle cells and form a delicate " c r o w n " around the cell
in one single phase. The slope is slightly steeper for the nucleus.
mitochondrial increase than for the increase in cellular sur- Although the total surface area of the nurse cells exceeds
face area. that o f the follicle cells, their respective increase in sur-
Due to the high number of long mitochondria which face area over time is very similar, at least until stage 10
constantly increases (200-fold from stage 2 to 10) the total (fig 1B and 1C).
mitochondrial surface area in cross-section increases In contrast to the follicle cells, the increase in the number
2.5-fold relative to the follicle cellular surface area (fig 2). o f mitochondria is divided into several phases (fig 1C).
This relative increase is continuous from stage 3 to 8 and The initial exponential phase continues until stage 6 and
then reaches a plateau (12%). Maximally 1.2 mitochon- resembles the exponential phase of the follicle cells. Then,
dria/t~m 2 were found for stage 8 and by extrapolation 0.5 in nurse cells, the increase of the mitochondrial/cellular
mitochondria//~m 2 for stage 2. These values are in the surface area ratio (fig 2) is due to the enlargement o f the
same range as those published in [18] for stages 3 - 6 . size of mitochondria (table I). This increase in the ratio
Because the morphology of the follicle cells has already is identical in follicle cells and nurse cells up to stage 6
been described in numerous previous studies [19, 25], we (fig 2). We measured 0.2 and 0.9 mitochondria//~m 2 of
describe here only our observations on the distribution of nurse cellular surface area in stages 3 and 6 respectively,
the mitochondria. Up to stage 6 follicle cells are cubic and corresponding approximately to the concentrations
their mitochondria are randomly distributed around the previously found for stages 3 - 6 [18].
cell nucleus (figs 3 and 4). From stage 7 on 2 types of folli- During a second phase, at stages 6 and 7, the number
cle cells appear. In the follicle cells adjacent to the oocyte of mitochondria stays at the same level and then increases
that become elongated and narrow, the mitochondria are from stage 8 on (fig 1C). The ratio of the mitochon-
found at the two poles longitudinally from the cell nucleus. drial/cellular surfaces area declines rapidly during the sec-
This mitochondrial distribution remains unchanged until ond phase (fig 2). Finally the decline appears to level off
stage 10. In the follicle cells that become flattened from at stage 10, when the area surface ratio reaches a minimal
stage 8 on and are adjacent to the nurse cells, the value similar to that found in the oocyte (fig 2).

1C
A
, i , , , , == ~ i , i i a ,
Time (hours)
10, 20, 50, 40 50 ~0 70
Os
1~ ,..i =
,.J ,~, 14
.F.i"" . =.}g 13 T

"t
12i
=* a,1 11
('o
=,, 10
m "J IO=
9

® 8
I I :~13 I 4 N S 17 IIII0]10H;
20 40 60
I I ~e I 3 I 4 N S I 7 IgllellOllJ
20 40
II~1314ROl~Jelo~
.....~),~e
0 7
Itage
a) 6
t_

Fig 1. Time course of the development of mitochondrial numbers 5

and cell surface during oogenesis. A, Oocyte: cell surface area
(o), total number of mitochondria (e), total volume of the egg ._m 4
chamber (dotted line); B: follicle cells: cell surface area (A), total
number of mitochondria (v); C: nurse cells: cell surface area
(o), total number of mitochondria (=). The surface areas of the 2 \\x +II
different cell types in figures IA, B and C have been estimated
from our own electron micrographs and from data previously
published [19]. The egg chamber volumes in figure I A were
$
calculated assuming an ellipsoid shape until stage 12 and then
a cylindrical shape for stages 13 and 14 [19]. Our egg chamber
I! 13 I, 116 17 18191,ollll
11 I'2""13~r4
volume curve differs from King's curve [19]. It can be explained Stage
from a different representation of the developmental stages. King
[19] assigns the same abscissa interval to each stage while we take Fig 2. Cell type and stage dependent variations of the mito-
account to the development time. Only standard deviations from chondrial surface area with respect to the cellular surface area.
more than 3 independent observations are shown on the curve Oocyte ( . ) ; follicle cell (A); nurse cell (O); standard deviation
(see caption under table I). (see table I) is given as in figure 1.
122 s Tourmente et al

It is of interest to compare these results with the mor- obtain information on localization and density of total
phological development of the cells and the location of the mtRNA. The probes used were a mixture revealing all
mitochondria. transcripts except the small rRNA, which probably
During stages 3 and 4, the mitochondria are more or develops like the large r R N A as found in Drosophila em-
less regularly arranged around the cell nucleus and form bryos [4]. One can also assume that our results essentially
a crown that becomes more marked in some stage-4 ceils reflect the behaviour of the m a j o r RNAs ie r R N A and to
(fig 3). During stage 5 the mitochondria form a relatively a lesser extent the cytochrome oxidase RNA, together
prominent crown in some nurse cells, while they become representing about 70% of the total m t R N A , according
aggregated into packages in others. At stage 6 (fig 4) all to [4] and to our own observations.
mitochondria are grouped in packages that increase in size In situ hybridization on frozen sections was chosen over
during stage 7. hybridization on semi-thin sections, due to faster ex-
During the last phase at stage 8 as well as stage 9 the perimental processing and a relatively higher labelling
mitochondrial packages become elongated and there is an intensity [5]. The methodology differed from [16] in
increase in the relative number of mitochondria that have 2 ways : first, Drosophila ovaries were not prefixed before
a shape which suggests that they are in the process of freezing and second, the mean probe length was shorter
dividing. Finally at stage 10, the mitochondria distribute (30-120 bp). In order to quantitate the results, tritiated
themselves over the entire cytoplasm of the nurse cells. rather than biotinylated probes were chosen.
Two control experiments were carried out, RNAse treat-
Oocyte ment (2 mg/ml, 1 h, 37°C) of the section before hybridiza-
The oocyte, the only cell that survives at the end of tion and hybridization with a pBR 322 vector probe.
oogenesis, also has a cytoplasm rich in ribosomes. In the Labelling was insignificant in both cases (results not
first part of oogenesis the oocyte cytoplasm is easily reco- shown).
gnized through its richness of endoplasmic reticular vesicles. Figure 5 illustrates typical results obtained on a series
The growth characteristics of the oocyte, already of sections hybridized with the same probe and exposed
described [19], are shown in figure 1A. Its growth is divid- for the same time. We observed a significant signal for
ed into 2 exponential phases. every cell type and stage examined. The similar cellular
The first one lasts until stages 7 - 8 , with an approximate localization of the labelling compared to that of mitochon-
10-h doubling time, similar to that of the follicle and nurse
cells.
The more rapid second phase has a 2.5-h doubling time
and lasts until stage 12. Thereafter the oocyte growth slows
down significantly.
We found that the oocyte was the only cell type where
mitochondria did not increase in number between stages
4 and 7, during and up to the end of the first growth phase.
Consequently, the mitochondrial/cellular surface area
ratio decreases sharply (5-fold) during this period (fig 2).
From stage 7 on, the number of mitochondria increases
exponentially at the same rate as the cellular surface area.
Thus, the mitochondrial/ceUular surface area ratio (fig 2),
after a rapid increase, reaches a plateau value equal to that
of the nurse cells. One can notice that this increase takes
place just after the beginning of the decrease in nurse cells
but is not associated with a change in the distribution.
Throughout the oocyte development, the mitochondria are
evenly ~stributed in the cytoplasm. Contrary to [27] we
did not observe a specific association between mitochon-
dria and yolk platelets.
Because we have considered only those sections where
the surface area of the oocyte is large enough to be
representative and the relative scarcity of oocyte sections
before stage 8, fewer observations were made on this cell
type than on the others. Nevertheless, this disadvantage
is compensated for by the more homogeneous distribution
of the mitochondria throughout the developmental stages.
Figure 2 gives the ratio, for the 3 cell types, between
the mitochondrial area and the entire cellular surface area
including the nucleus. It was of interest to consider the
effect of the relative variation of the size of the nucleus
on the density of the mitochondria in the cytoplasm. This
was tested and no qualitative differences were observed
(results not shown).

In situ hybridization o f mtRNA on frozen sections

Fig 3. Electron micrograph of stage 4 egg chamber (x 4 000);
The level of the mitochondrial RNAs was examined by in (n) : nurse cell ; (f) : follicle cell ; (ER) : endoplasmic reticulum ;
situ hybridization techniques [5, 6] which allowed us to (m) : mitochondria.
 Mitochondrial development during Drosophila oogenesis
dria and the lack of, or only very weak nuclear labelling
strongly support the mitochondrial specificity of hybridiza-
tion and suggest that the mitochondrial population is
homogeneous with regard to the mtRNA.
Figure 5 clearly visualizes variations of grain distribu-
tion and density with cell type and developmental stage.
From stage 6 to 8, we observed a heterogeneous distribu-
tion of the grains in the nurse cells. From stage 10 on, the
grain distribution became homogeneous in the nurse cells
and diminished. In the oocyte, the grain distribution re-
mained uniform over all stages.
Because of the small size of mitochondria, only the
cellular surface area can be estimated directly in light
microscopy. The quantitative study of the number of
grain//zm 2 of cellular surface area given in figure 6 shows
two clearly distinct phases in follicle and nurse cells; a first
plateau phase lasting until stage 9 followed by a declining
Fig 4. Electron micrograph of stage 6 egg chamber ( x 4 500) ; (o) : oocyte; (n) : nurse cell; (f) : follicle cell ; (er) : endoplasmic reticulum;
(m) : mitochondria.
123
phase from stage l0 on. During the declining phase, the
grain densities dropped around 6 and 3 times in follicle
and nurse cells respectively.
For the oocyte, data from the first half of oogenesis
were insufficient to verify a plateau. From stage 5 on,
up to stage 10, we observed a slow but steady decrease in
grain density. Contrary to follicle and nurse cells, this
decrease stops at this point and reaches a final plateau
around 2 times lower than that of the 2 other plateau
values.
The different plateau values found in the 3 cell types
deserve special consideration. Grain density was always
highest in follicle cells, intermediate in the oocyte and
lowest in the nurse cells.
It must be noted that using the same probe we found
identical qualitative and quantitative results with another
Drosophila strain, Charolles (results not shown).
124 S Tourmente et al

Discussion and Conclusion
Drosophila oogenesis represents a well documented system
consisting of 3 cooperating cell types, nurse cells, oocyte
and follicle cells, developing over 78 h to build a mature
oocyte 100000 times bigger than at the beginning of
oogenesis. Each cell type has particular functions and
energetic requirements. The energy requirements for all
of these processes are assured by mitochondria.
Despite some relatively well known molecular aspects,
there is a lack of information about the mitochondrial
biogenesis and its cellular control mechanism. In order to
obtain such information we wanted to closely examine and
compare mitochondrial changes between the 3 cell types
throughout oogenesis.
The main conclusions we can draw are:
First, the dynamic evolution of the mitochondrial
population in the three cell types throughout oogenesis as

Fig 5. In situ RNA hybridization on frozen sections of Drosophila ovaries, using a ~H-labelled mitochondrial probe covering the
entire coding sequence except the small rRNA (see Materials and Methods). Bright and dark field micrographs are shown for each
frozen section (x 125). Stage 14 corresponds to the mature oocyte. The characteristic grain packages are particularly visible in stages
6, 7 and 8 on the upper dark field micrograph.
 Mitochondrial developmentduring Drosophila oogenesis
stated by the important and specific changes of the
number, shape, distribution and density of mitochondria
and of the mtRNA level.
According to the criteria, presented in [3] the mitochon-
dria we observed appeared to be active and mobile, judg-
ing from their aspect and rapid shift in position between
the different stages of oogenesis particularly in nurse cells.
This was confirmed by experiments using rhodamine 123
[32], where we observed important respiratory activities
in all cell types and stages, including the cells and the stages
where the mitochondria have a smaller size (not shown).
Second, the good correlation between these changes, the
biosynthetic activities and the energetic and developmen-
tal requirements, suggesting the existence of several sim-
ple control mechanisms of these mitochondrial parameters
and that these controls are important for the cell.
These conclusions and hypotheses are supported by the
comparison of the biological functions of the 3 cell types
with their mitochondrial/cellular surface area ratio con-
sidered as an indicator of the cell respiratory capacity. The
published data demonstrate significant biosynthetic ac-
tivities during the entire oogenesis in follicle cells. In suc-
cession, they first divide rapidly 4 times between stages 2
and 6 [19], then they enlarge and their nuclear DNA
polyploidizes significantly ( 4 5 - 5 0 C) [26]. From stage 8
on, the cells contribute to the production of vitellogenins
[6] which represent 22°7o of all proteins of the mature
oocyte. In stages 9 and 10 they synthesize the vitelline
membrane [12] and finally, over a very short time inter-
val (5 h), they synthesize the chorion [35]. These data
strongly suggest that these cells have important and con-
stant energy requirements. This is in agreement with the
high and increasing density of mitochondria.
In the nurse cells up to stage 7, a similar conclusion can
be drawn. Indeed, as in the follicle cells, the nuclear DNA
polyploidizes but it reaches much higher levels, ie
512-1024 C [19]. The nurse cells synthesize the majority
of the RNA stored in the mature oocyte [19] and the syn-
thesis of most of the RNA begins with the start of
oogenesis. This transcription is accompanied by a transla-
tion which is accelerated from stage 7. The rapid and
characteristic decline of this ratio and our own observa-
tions concerning the aggregation of mitochondria in groups
between stages 6 and 9, associated with the more intense
mitochondrial replication activity (not shown) and likewise
the brisk increase of the oocyte mitochondria content are
totally in agreement with the literature data relating to the
nurse cell cytoplasm transfer into the oocyte through
cytoplasmic channels [1, 19]. However the subsequent
3-fold reduction in the mitochondrial/cellular surface area
ratio (fig 2) seems to be contradictory to the literature data
about the stimulation of transcriptional activity of several
genes during this phase [1, 2, 22, 33, 34] suggesting an
increased rather than a reduce energy demand. Never-
theless, it has been calculated [17] that owing to the strong
polyploidization, the production of the 6000 different
RNAs of the nurse cells requires only a relatively slow rate
of transcription. Thus it may not be a limiting factor at
this level.
On the other hand the aggregation of mitochondria has
also been observed in oocytes o f other organisms, eg fish
[9], amphibia [37] and sea anemones [21] which, contrary
to Diptera, do not have nurse or follicle cells. This ag-
gregation could be related to a particular activity rather
than to the cytoplasmic transfer.
Concerning the oocyte, the smallest-sized mitochondria,
the lack of mitochondrial biogenesis during stages 4 - 7
125
(while the oocyte size increases 5-fold) leading to the
smallest ratio of the 3 cell types suggest modest energetic
requirements of this cell. This is also in good agreement
with the literature data showing a low protein synthesis
activity [40], the absence of DNA replication [19], and little
transcriptional activity [15, 24]. One can hypothesize that
according to the normal mitochondrial/cellular surface
area ratio found in stage 4, the nuclear coded "house
keeping" genes needed for the mitochondrial assembly
would be transcribed at the beginning of oogenesis, at a
time when the nurse cells and oocyte are nearly equivalent.
These transcripts would be completely consumed around
stage 4 explaining the lack of biogenesis from this stage
on. Then the sudden increase in the number of mitochon-
dria would be entirely due to the transfer of nurse cell
cytoplasm which is precisely localized from stage 7 on.
The value of 4°/0 that we found for the mitochondrial/
cellular surface area ratio is close to that calculated from
the literature data [28, 38] assuming 1.5 × 107 mtDNA
molecules/oocyte, and a single mtDNA molecule/
mitochondrion. This low value seems to be a priori con-
tradictory to the common idea of a high mtDNA content
in the mature oocyte. In fact, despite this and owing to
its large cell volume the mature oocyte contains enough
mitochondria to supply some 15 000 somatic cells. This is
sufficient to assure embryogenesis without any external
nutrient requirement except oxygen.
Contrary to the cellular distribution and density of
mitochondria and to the mtRNA distribution which can
be easily correlated to the biological functions of the 3 cell
types throughout oogenesis, the mtRNA density shows the
more unexpected results : a constant and parallel mtRNA
density until stage 9 and then an abrupt, significant and
simultaneous decrease in the mtRNA concentration in the
follicle and nurse cells at stage 10.
Until stage 9 two hypotheses could be drawn concern-
ing the constancy of mtRNA density:
First the mtRNA level directly or not correlates to the
cellular surface area, and more probably to the cell volume,
according to a mechanism which remains to be determined.
Second the mtRNA level correlates to the nuclear DNA
content, since some data [19, 26] indicate a parallel varia-
40 eo .~. Time [ hours
, .'9.~.
20 ..~.~ ..~
A B C
1.4
c 1;
o
~o_ I.( • .~°
c_ a o - °
.6
"~ .5
"1o
O
.3
-~ .2
d
,1 21 3 I" ~q6171819Jql] ' 121314 N 61718191~1 , 12 [ 3] 4 ~isl ?lslel'olll
Stage
Fig 6. In situ mitochondrial RNA hybridization on Drosophila
ovary sections : cellular grain density of follicular cell (A) ; nurse
cell (B); oocyte (C). The grain density values are calculated as
described in Materials and Methods. Only standard deviations
of means from more than 3 independent observations are shown.
126 S Tourmente et al
tion of D N A content and cell surface area both in nurse
and follicle cell during this period.
Some authors [1] have already described some examples
o f nuclear coded R N A whose level varies in parallel with
the D N A content during this period. Concerning the I ele-
ment [7], the data also show similar behaviour (Bucheton
and Tourmente, personal communication).
On the contrary the brisk decrease at stage 10 seems to
be characteristic for the mtRNA, since an increase has been
shown for most of the cellular R N A around stage 9 [1].
For the I element a constant density of the R N A has been
observed (Bucheton and Tourmente, personal com-
munication).
This decrease could reflect either a negative and specific
control of mt R N A level or a decrease of cell D N A con-
centration. Our results do not allow us to discriminate
between these hypotheses. Nevertheless the higher level of
m t R N A in the follicle ceils compared to the nurse cells cor-
responding to their higher nuclear D N A concentration and
the end of D N A accumulation even though the cellular sur-
face area constantly increases [19] are in favor of a nuclear
D N A content correlation in the 2 cell types.
In the oocyte the limited number of results do not allow
us to determine clearly whether a plateau exists before stage
7. From stage 7 up to stage 10, contrary to nurse and folli-
cle cells (and confirmed by other experiments with a
ribosomal D N A specific probe, results not shown), the
m t R N A level decreases. Presumably this can be related to
the injection into the oocyte of the nurse cell cytoplasm
including mitochondria.
In any case considering the low nuclear D N A content
of oocyte (4c [19]), its higher m t R N A level compared to
that of nurse cells does not agree with our last hypothesis
except if the oocyte m t R N A turnover is completely dif-
ferent from what it is in follicle and nurse cells. This could
be related both to the lack of mitochondrial biogenesis and
the stability of the RNA density f r o m stage 10 up to the
end. This is in agreement with a relative quiescence of the
oocyte.
Such control mechanisms of the m t R N A level could be
enforced at least by 2 types of nuclear coded factors the
first one a mitochondrial R N A polymerase or a transcrip-
tional factor [13] whose expression would be unspecifica]Jy
correlated with the cell volume, and more probably nuclear
D N A co~tent. The second one would control m t R N A
stability. Its expression would be specifically related to the
cell type and the developmental stage, explaining the dif-
ferent behaviour between the oocyte and the other two cell
types.
Although numerous intriguing questions remain to be
elucidated, especially at the quantitative and molecular
levels, our results agree with the idea that mitochondria
are an integral part of the developmental process.
Acknowledgments
The authors thank R Durand (Laboratoire Biochimie, Clermont-
Ferrand), WJ Gehring (Biozentrum, Basel) for helpful discus-
sions and critical reading of the manuscript. They thank U Kloter
(Biozentrum, Basel), the laboratoire of Zoologic et Protistologie
(Clermont-Ferrand) for laboratory facilities and A Marie
(Laboratoire Biochimie, Clermont-Ferrand) for helpful technical
assistance. They also thank MJ Goncalves for typing the
manuscript. This work was supported by grants from the CNRS,
INSERM. S Tourmente was the recipient of a French MRT
predoctoral scholarship and a Ligue Regionale Contre le Cancer
grant.
References
1 Ambrosio L, Schedl P (1984) Gene expression during
Drosophila melanogaster oogenesis: analysis by in situ
hybridization to tissue sections. Dev Biol 105, 80-92
2 Ambrosio L, Schedl P (1985) Two discrete modes of histone
gene expression during oogenesis in Drosophila melano-
gaster. Dev Biol 111,220-231
3 Bereiter-Hahn J, V6th M (1983) Metabolic control of
shape and structure of mitochondria in situ. Biol Cell 47,
309-322
4 Berthier F, Renaud M, Alziari S, Durand R (1986) RNA
mapping on Drosophila mitochondrial DNA: precursors and
template strands. Nucleic Acids Res 14, 4519-4533
5 Binder M, Tourmente S, Roth J, Renaud M, Gehring WJ
(1986). In situ hybridization at the electron microscope level:
localization of transcripts on ultrathin sections of Lowicryl
K4M embedded tissue using biotinylated probes and pro-
tein A-gold complexes. J Cell Biol 102, 1646-1653
6 Brennan MD, Weiner A J, Goralski T J, Mahowald AP
(1982) The follicle cells are a major site of vitellogenin syn-
thesis in Drosophila melanogaster. Dev Biol 89, 225-236
7 Bucheton A, Paro R, Sang HM, Pelisson A, Finnegam DJ
(1984) The molecular basis of I-R hybrid dysgenesis in
Drosophila melanogaster: identification, cloning and prop-
erties of the I factor. Cell 38, 153-163
8 Clayton DA (1982) Replication of animal mitochondrial
DNA. Cell 28, 693-705
9 Clerot JC (1976) Les groupements mitochondriaux des
cellules germinales des poissons t~16ost~ens cyprinid~s. I.
Etude ultrastructurale. J UItrastruct Res 54, 461-475
10 El Meziane A, Callen JC, Mounolou JC (1989) Mitochon-
drial gene expression during Xenopus laevis development:
a molecular study. EMBO J 6, 1649-1655
11 Enster L, Schatz G (1981) Mitochondria: a historical review.
J Cell Biol 91, 225-227
12 Fargnoli J, Waring GL (1982) Identification of vitelline
membrane proteins in Drosophila melanogaster. Dev Biol
92, 306-314
13 Fisher RP, Topper JN, Clayton DA (1987) Promoter selec-
tion in human mitochondria involves binding of a transcrip-
tion factor to orientation-independent upstream regulatory
elements. Cell 50, 247-248
14 Gall JG, Pardue ML (1971) Nucleic acid hybridization in
cytological preparations. Methods Enzymol 21,470-480
15 Haelin M, Roos C, Cassab A, Mohier E (1987) Oocyte-
specific transcription of fs (1) K 10: a Drosophila gene af-
fecting dorsal-ventral developmental polarity. Embo J 6,
801-807
16 Hafen E, Levine M, Garber RL, Gehring WJ (1983) An im-
proved in situ hybridization method for the detection of
cellular RNAs in Drosophila tissue sections, and its applica-
tion for localizing transcripts of the homeotic Antennapedia
gene complex. EMBO J 2, 617-623
17 Jacobs-Lorena M, Hough-Evans BR, Britten R J, Davidson
EH (1980) Complexity of RNA in developing ooctyes of
Drosophila melanogaster. Dev Biol 76, 509-513
18 King RC, Mills RP (1962) Oogenesis in adult Drosophila XI :
studies of some organelles of the nutrient stream in egg
chambers of Drosophila melanogaster and Drosophila
willistoni. Growth 26, 235-253
19 King RC (1970) Ovarian development in Drosophila
melanogaster. Academic Press, New York
20 Kuroiwa T (1982) Mitochondrial nuclei. Int Rev Cytol 75,
1-59
21 Larkman AU (1984) The fine structure of mitochondria and
the mitochondrial cloud during oogenesis in the sea anenome
actinia. Tissue & Cell 16, 393-404
22 Loyd JE, Raft EC, Raft RA (1981) Site and timing of syn-
thesis of tubulin and other proteins during oogenesis in
Drosophila melanogaster. Dev Biol 86, 272-284
23 Luft JH (1961) Improvements in Epoxy resin embedding
methods. J Biophys Biochem Cytol 9, 404-414
24 Mahowald AP, Tiefert M (1970) Fine structural changes in
 Mitochondrial development during Drosophila oogenesis
the Drosophila oocyte nucleus during a short period of RNA
synthesis. Wilhelm Roux's Arch 165, 8-25
25 Mahowald AP (1972) Ultrastructural observations on
oogenesis in Drosophila. J Morphol 137, 29-48
26 Mahowald AP, Caulton JH, Edwards MK, Floyd AD
(1979). Loss of centrioles and polyploidization in follicle
cells of Drosophila melanogaster. Exp Cell Res 118,
404-410
27 Marco R, Batuecas B, Garesse R (1983). Mitochondrial
storage in Drosophila melanogaster yolk granules. EDRC,
Cambridge
28 Merten SH, Pardue ML (1981) Mitochondrial DNA in
Drosophila. An analysis of genome organization and
transcription in Drosophila melanogaster and Drosophila
virilis. J Mol Biol 153, 1-21
29 Mignotte F, Tourte M, Mounolou JC (1987) Segregation of
mitochondria in the cytoplasm of Xenopus vitellogenic
oocytes. Biol Cell 60, 97-102
30 Quagliariello E, Slater EC, Palmieri F, Saccone C, Kroon
AM (1985) Achievements and Perspectives of Mitochondria
Research, vol 1. Elsevier, Amsterdam
31 Reynolds ES (1963) The use of lead citrate at high pHs, an
electron-opaque stain in electron microscopy. J Cell Biol 17,
208-212
32 Ronot X, Benel L, Adolphe M, Mounolou JC (1986).
127
Mitochondrial analysis in living cells : the use of rhodamine
123 and flow cytometry. Biol Cell 57, 1-8
33 Ruddell A, Jacobs-Lorena M (1984) Preferential expression
of actin gene during oogenesis of Drosophila. Dev Bio1105,
115-120
34 Ruddell A, Jacobs-Lorena M (1985) Biphasic pattern of
histone genes expression during Drosophila oogenesis. Proc
Natl Acad Sci USA 82, 3316-3319
35 Spradling AC, Mahowald AP (1979) Identification and
genetic localization of mRNAs from ovarian follicle cells of
Drosophila melanogaster. Cell 16, 589-598
36 Stepien G, Renaud M, Savre I, Durand R (1988). Juvenile
hormone increases mitochondrial protein synthesis and
cytochrome oxidase activity in Drosophila cells, lnsect
Biochem 18, 313-321
37 Tourte M, Mignotte F, Mounolou JC (1984) Heterogeneous
distribution and replication activity of mitochondria in
Xenopus laevis oocytes. Eur J Cell Biol 34, 171-178
38 Travaglini EC, Petrovic J, Schultz J (1968). Two satellite
cytoplasmic DNAs in Drosophila. J Cell Biol 39, 136a
39 Wernette CM, Kaguni LS (1986) A mitochondrial DNA
polymerase from embryos of Drosophila melanogaster. J
Biol Chem 261, 14764-14770
40 Zalokar M (1959) Sites of ribonucleic acid and protein syn-
thesis. Exp Cell Res 19, 184-186