/ . Embryol exp. Morph. Vol. 69, pp. 115-126, 1982
Printed in Great Britain © Company of Biologists Limited 1982
\ \ 5
Cell membrane regionalization in
early mouse embryos as demonstrated by
5 -nucleotidase activity
By L. IZQUIERDO 1 AND C. EBENSPERGER 1
From the Department of Biology, University of Chile
SUMMARY
The distribution of 5'-nucleotidase activity in pre-implantation mouse embryos is studied
by means of a cytochemical method adapted from Uusitalo & Karnovsky (1977). The enzyme
activity is detected, from the 4-cell stage up to the morula stage, on discrete patches of the
cell membane between blastomeres. Appropriate controls show that this distribution is not a
localization artifact due to selective retention of the enzyme reaction product in the narrow
interblastomeric spaces. In early blastocysts, as the blastocoel expands the enzyme activity
on its lining disappears. The external surface of the trophectoderm in early blastocysts lacks
any enzyme activity, whereas in late blastocysts a strong enzyme activity is detected at the
embryonic trophectoderm, decreasing in intensity towards the opposite pole of the embryo.
These results are compared to previous observations by other authors and the differences
are mainly ascribed to differences in the cytochemical procedure employed.
We conclude that during cleavage a gradual cell membrane regionalization unfolds, revealing a pattern that may be related to morphogenesis; in particular, to the localization of
zonular tight junctions around the peripheral blastomeres of the morula (Izquierdo, 1977;
Izquierdo, Lopez & Marticorena, 1980).
INTRODUCTION
The inside-outside model originally proposed by Tarkowski & Wroblewska
(1967) still offers the most coherent explanation for the differentiation of inner
cell mass and trophectoderm during the transition from morula to blastocyst.
The model aptly reduces the problem to a particular case of differential gene
activity elicited by diverse cellular microenvironments. Nevertheless, positional
information required for blastocyst morphogenesis (Izquierdo, 1977; Johnson,
Pratt & Handyside, 1981) may depend, not on which genes are transcribed in
different cells but rather on where a gene product localizes within a cell. A case
in point is the proposed causal relationship between peripheral sealing of
interblastomeric spaces by zonular tight junctions and the development of the
blastocoel (Ducibella & Anderson, 1975; Izquierdo, Fernandez & Lopez,
1976; Ducibella, 1977; Fernandez & Izquierdo, 1980). If tight junctions were
exclusively found around peripheral blastomeres, one might assume that their
1
Authors' address: Department of Biology, University of Chile, Casilla 653, Santiago,
Chile.
116
L. IZQUIERDO AND C. EBENSPERGER
5'-nucleotidase in early mouse embryos
117
localization could be explained by differential gene activity leading to the
synthesis of specific tight-junction components. However, incipient focal
junctions, admittedly not easy to classify, can be seen between inner blastomeres of the morula (Ducibella, Albertini, Anderson & Biggers, 1975; Izquierdo
et al. 1976). Furthermore, when the number of blastomeres at blastulation is
diminished by different procedures, peripheral sealing is not prevented even
though proper inside and outside cells may not exist (Smith & McLaren, 1977;
Fernandez & Izquierdo, 1980; Izquierdo & Becker, 1982); this situation would
preclude the emergence of an inside-outside morphogenetic pattern, if based on
the interaction of cells that express different genes. These and other observations
(see Johnson, Pratt & Handyside, 1980) led us to assume that the search for
positional information demands, not only an analysis of cellular regionalization
within the embryo but also a description of the regionalization of cellular components within embryonic cells.
We described in a previous work the regionalization of the cell membrane in
preimplantation mouse embryos as revealed by the distribution of alkaline
phosphatase activity (Izquierdo et al, 1980). Results were interpreted according
to a model which proposes that a morphogenetic pattern emerges during
cleavage as a result of localized formation of discrete patches of 'new' Cell
membrane (Izquierdo, 1977). In the present work we have recourse as a marker
to the cytochemical demonstration of 5'-nucleotidase activity. Two reports have
already dealt with the onset and distribution of 5'-nucleotidase activity in
preimplantation embryos (Vorbrodt, Konwinski, Solter & Koprowski, 1977;
Nizeyimana-Rugina & Mulnard, 1979); however, this matter requires further
investigation for several reasons: (1) former observations were not intended to
prove or disprove a morphogenetic model, (2) they disagree in several points
and (3) cytochemical methods for the demonstration of 5'-nucleotidase have
recently been revised and improved (Uusitalo & Karnovsky, 1977; Yamashina
& Kawai, 1979; Klaushofer & von Mayersbach, 1979).
Figs. 1-6. Whole mounts of embryos in different stages of development. Bars
represent 20 /*m.
Fig. 1. Late 4-cell embryo (58 h of development). A patch of 5'-nucleotidase activity
is seen between the two blastomeres in focus.
Fig. 2. Early morula (58 h of development). The enzyme activity is detected in interblastomeric spaces whereas no activity is detected on the surface of the embryo.
Fig. 3. Late morula (80 h of development). Observations similar to Fig. 2.
Fig. 4. Early blastocyst (84 h of development). The enzyme activity is detected between cells of the inner mass and in some places of the blastocoel (b) lining. No
activity is detected on the outer aspect of the trophectoderm.
Fig. 5. Late blastocyst (104 h of development). Enzyme activity is detected between
cells of the inner mass and on the outer surface of the trophectoderm at the embryonal pole.
Fig. 6. A 2-cell embryo (later cleaved into 3 cells) paired with an early morula and
cultured for 5 h. The enzyme activity is detected between blastomeres of the morula.
No activity is detected in the 3-cell embryo nor on the surface between both embryos.
118
L. IZQUIERDO AND C. EBENSPERGER
', .A
5'-nucleotidase in early mouse embryos
119
MATERIALS AND METHODS
About 720 embryos from spontaneously ovulating Swiss-Rockefeller mice
were used. The age of the embryos was reckoned starting at 0 h of the day in
which the vaginal plug was detected. Oviducts were flushed with Biggers'
medium (Biggers, Whitten & Whittingham, 1971) and the embryos were processed for the cytochemical demonstration of 5'-nucleotidase activity according
to a modified version of Uusitalo & Karnovsky's method (1977). Modifications
refer to the use of a different fixative and incubation medium. In preliminary
trials we compared the effect of different fixatives containing tris-maleate or
cacodylate buffers at pH values ranging from 6 to 7-5; we also compared the
effect of different incubation media containing Pb(NO3)2 at concentrations
ranging between 0-25 and 6 mM. Non-specific lead deposits were more frequent
when high lead concentrations were used; however, results were not strictly
proportionate, and even low lead concentrations occasionally produced unspecific deposits. In these preliminary trials 120 embryos were used.
Best results regarding sensitivity and specificity (see controls, below) were
obtained with the following procedure. Embryos were fixed in 1 % (w/v)
glutaraldehyde in 0-1 M-tris-maleate buffer pH 6 during 60min at room temperature. After three rinses in 0-1 M-tris-maleate buffer pH 7-2, embryos were
incubated during 50 min at 37 °C in a medium containing 0-1 M-tris-maleate
buffer pH 7-2, 1-5 mM 5'AMP (Sigma), 1-5 mM Pb(NO3)2 and 10 mM-MgSQ4.
The fixative, the buffers used for rinsing and the incubation medium contained
5% (w/v) glucose (Sundstrom & Mornstad, 1975; E. Rodriguez, personal
communication).
For light microscopy, after incubation the embryos were rinsed three times
Fig. 7. Whole mounts of morulae. From left to right: normal control embryo;
embryo incubated with 1-5 mw sodium /?-glycsrophosphate instead of 5'AMP;
embryo incubated without Mg2+; embryo incubated with 005 M-NaF.
Figs. 8-11. Electron micrographs of embryos in different stages of development.
Bars represent 1 fim.
Fig. 8. Early morula (60 h of development). The enzyme reaction product is confined
to the interblastomeric surface and does not extend beyond the site where tight
junctions are localized (arrow). The cell surface facing the zona pellucida (zp)
lacks reaction product. Unspecific lead deposits are observed in the perivitelline
space.
Fig. 9. Late morula (80 h of development). The wide interblastomeric spaces are
lined by the enzyme reaction product.
Fig. 10. Nascent blastocyst (80 h of development). A cytoplasmic 'vesicle' (v)
emptying its contents into the growing blastocoel (b). The blastocoel is still lined
by the enzyme reaction product whereas no reaction product is seen on the vesicle
wall.
Fig. 11. Nascent blastocyst (80 h of development). The enzyme reaction product
lining the expanding blastocoel (b) is discontinuous.
120
L. IZQUIERDO AND C. EBENSPERGER
5'-nucleotidase in early mouse embryo
121
in tris-maleate buffer pH 7-2 and the enzyme reaction product was developed
with 0-1 % (w/v) ammonium sulphide. The embryos were then whole-mounted
in glycerol.
For electron microscopy, after incubation the embryos were rinsed three
times in 0-1 M-cacodylate buffer pH 7-2 and then post-fixed in 3 % (w/v)
glutaraldehyde in cacodylate buffer during 60 min at room temperature. After
rinsing in buffer, post-fixation continued for 60 min at room temperature With
OSO4 (w/v) in 0-1 M-cacodylate buffer. The embryos were rinsed again in buffer,
dehydrated in graded acetone and embedded in a low-viscosity resin (Spurr,
Poly science). Thin sections were obtained in a Porter-Blum ultramicrotome, the
grids were stained during 2 min in 4 % (w/v) uranyl acetate in methanol and
observed in a Philips EM 300 electron microscope.
Two kinds of control were used. In order to determine the specificity of the
enzyme reaction we tested (a) incubation in substrate-free medium; (b) incubation with sodium /?-glycerophosphate or ATP as analogous substrates; (c) incubation without the activator; (d) incubation at 0 °C; (e) incubation after heating
the embryos for 30 min at 95 °C; (/) incubation in a medium with 0-05 M-NaF.
In order to discard a localization artifact that might be produced by retention
of the enzyme reaction product in the interblastomeric spaces, we fused early
morulae with 2-cell embryos, according to a method described elsewhere
(Izquierdo et al. 1980). These chimaeras were cultured for 5 h before being
processed for 5'nucleotidase demonstration.
RESULTS
In embryos with less than four blastomeres, 5'-nucleotidase activity is absent,
but it is consistently present in embryos with more than four blastomeres.
Embryos with 4 cells, recovered before 48 h of development, intermingled with
2- and 3-cell embryos, show a negative reaction, while the reaction is positive
in 4-cell embryos recovered after 58 h of development, intermingled with early
morulae. That is, the onset of enzyme activity, as recognized by this cytochemical method, coincides with the late 4-cell stage.
Light microscopy shows .that the enzyme reaction product appears between
blastomeres as small patches that become larger as cleavage proceeds. In morulae
and early blastocysts the external surface of the embryo lacks any enzyme
activity. In late blastocysts, which have more than 106 h of development,
Figs. 12-15. Late blastocyst (104 h of development). Strong enzyme activity is
detected on the external surface of the trophectoderm at the embryonal pole
(Fig. 12), gradually decreasing towards the opposite pole (Figs. 13, 14) where no
activity is detected (Fig. 15)! Bar represents 1 fim.
Fig. 16-19. Another late blastocyst (104 h of development) showing a similar distribution of the enzyme reaction product; this diminishes from the embryonal pole
(Fig. 16) towards the abembryonal pole (Fig. 19). Bar represents 1 fim.
122
L. I Z Q U I E R D O AND C. EBENSPERGER
Fig. 20. Diagram representing 5'-nucleotidase activity in different developmental
stages. The enzyme reaction product is indicated by bold lines.
enzyme activity can be detected on the external surface of the polar trophectoderm, gradually decreasing towards the opposite pole of the embryo (Figs. 1-5).
Light microscope observations, excluding control series, were performed on
198 embryos.
The controls on the localization of the enzyme reaction product, using
aggregated embryos, showed that each embryo presented the characteristic
pattern of its stage and that no activity was found in the artificial cleft between
them (Fig. 6). Twenty-three embryo pairs were observed. As to controls designed
to test the specificity of the enzyme activity, all gave negative results (Fig. 7).
One hundred and seventy-six embryos were observed with the light or the
electron microscope during these tests.
Although there is evidence that the fixation procedure breaks down the
permeability barrier to phosphatase substrates (Uusitalo & Karnovsky, 1977)
electron microscopy shows that the enzyme reaction product is confined to the
surface of the plasma membrane. At the 4-cell stage the enzyme reaction product
appears in discrete regions of the cell surface of adjoining blastomeres. These
regions grow as development continues, but always remain restricted to interblastomeric surfaces. Therefore a distinct regionalization of the plasma membrane can be recognized in peripheral blastomeres: the membrane facing the
zona pellucida is devoid of enzyme reaction product, whereas the membrane
facing the blastomeres is full of it. The precise boundary of these regions
coincides with the site where zonular tight junctions become established
(Figs. 8 and 9).
During blastulation, the cell membranes surrounding the nascent blastocoel
reveal 5'-nucleotidase activity, but as the cavity expands its lining of enzyme
reaction product becomes discontinuous and soon disappears. Electron micrographs suggest that the wall of the cytoplasmic vesicles which empty their
contents into the blastocoel may contribute to its enlarging surface; however, a
unit membrane encircling these vesicles is not clearly seen (Figs. 10 and 11).
Fourth-day blastocysts show enzyme reaction product between cells of the
5'-nucleotidase in early mouse embryo
123
inner mass and in few patches around the blastocoel surface. Enzyme activity
is rarely detected at the small contact surface between trophectoderm cells or at
the external surface of the embryo. In fifth-day blastocysts, enzyme activity
appears on the outer surface of the trophectoderm at the embryonal pole,
decreasing in intensity towards the abembryonal pole (Figs. 12-19). Electron
microscope observations, excluding control series, were performed on approximately 180 embryos. A summary of our results is presented in fig. 20.
DISCUSSION
Our observations are in keeping with those by Vorbrodt et al. (1977) except
in that they report 5'-nucleotidase activity on the external surface of the embryo
beginning at the advanced morula stage, while we do not observe it until the
late blastocyst. The results reported by Nizeyimana-Rugina & Mulnard (1979)
barely show any resemblance to ours, as these authors describe enzyme activity
on the embryo surface at all stages and also within cells, in the form of cytoplasmic and nuclear clusters. The discrepancies, particularly those between
our observations and those of Nizeyimana-Rugina & Mulnard, can be at least
partly ascribed to differences in the method employed for demonstrating the
enzyme activity; especially to differences in fixative composition and in lead
content of the incubation medium.
Parallel biochemical and cytochemical experiments on mouse lymphocytes
have shown that the standard fixation with glutaraldehyde in cacodylate buffer
pH 7-2-7-4 causes an important inhibition of 5'-nucleotidase activity which c&n
be avoided by using a slightly acidic fixative buffered with tris-malaete (Uusitalo
& Karnovsky, 1977). These resits have been confirmed in acinar cells of rat
salivary glands and in liver tissue (Yamashina & Kawai, 1979; Klaushofer & von
Mayersbach, 1980). Since similar results were obtained in our preliminary trials,
the composition of the fixative we used thereafter differs considerably from
that used by Vorbrodt's group or by Nizeyimana-Rugina & Mulnard.
Better preservation of the enzyme activity by proper fixation allows for a
reduction of lead concentration in the incubation medium, thus avoiding the
formation of non-specific lead deposits. Uusitalo & Karnovsky (1977) used 1-8
mM-Pb(NO3) instead of 3-6 mM, which is the concentration in the standard
Wachstein-Meisel medium (Wachstein & Meisel, 1957). Klaushofer & von
Mayersbach (1979) found that in liver rat tissue, prepared by w-butanol freeze
substitution, best results were obtained with only 0-6 mM, and our preliminary
trials also showed the advantage of low concentration. We used 1-5 mMPb(NO3)2 in the routine procedure, Vorbrodt et al. (1977) used 3-6 mM and
Nizeyimana-Rugina & Mulnard (1979) used 60 mM.
Our controls support the specificity of the cytochemical method employed,
and most probably discard a localization artifact caused by selective retention
of the reaction product between blastomeres. The interblastomeric spaces
124
L. IZQUIERDO AND C. EBENSPERGER
experimentally produced by fusing embryos retain no reaction product and
later, when enzyme activity appears on the external surface of advanced blastocyst, the reaction product showed no sign of diffusion. We then conclude that
our procedure is a reliable method for the cytochemical demonstration of
5'-nucleotidase activity, and hence that the cell membrane regionalization it
reveals deserves an interpretation.
Observations reported here fit in with a model which proposes that localized
'new' cell membrane is formed during cleavage, thus developing a pattern which
might provide every blastomere with information about its age and position
(Izquierdo, 1977). According to this model, we have formerly interpreted the
cell membrane regionalization revealed by alkaline phosphatase activity
(Izquierdo et al. 1980), and the same rationale would fit the observations
reported here. Briefly: since no 5'-nucleotidase activity is detected prior to the
4-cell stage and a large quantity of cell membrane is required for cleavage,
after the 4-cell stage one might recognize ' old' cell membrane, which is devoid
of 5'-nucleotidase activity, from 'new' cell membrane, which presents activity.
If 'new' cell membrane is formed at the cleavage furrows, leaving a patch
between sister blastomeres, the outer surface of peripheral blastomeres would
remain covered solely by 'old' cell membrane. This interpretation accounts
quite well for our results, not so well for Vorbrodt's results and not at all for
results obtained by Nizeyimana-Rugina & Mulnard.
Our interpretation may also account for the appearance of 5'-nucleotidase
activity on the outer surface of advanced blastocysts and for the stronger
reaction on the polar trophectoderm as compared to the mural trophectoderm.
It has been shown that trophectoderm cells close to the inner cell mass divide
faster than those farther away (Copp, 1978); therefore, trophectoderm cells at
the abembryonal pole, which have gone through fewer divisions, according to
our model should have less 'new 'membrane, that is, less 5-'nucleotidase
activity.
In its most extreme form, the interpretation we offer assumes that discrete
patches of cell membrane are formed at the leading edge of cleavage furrows,
and indeed, evidence has already been advanced in other materials which would
support such an idea (review in Izquierdo, 1977). However, several objections
can be raised to this scheme as applied to preimplantation embryos. First, both
5'-nucleotidase and alkaline phosphatase (Ishiyama & Izquierdo, 1977; Izquierdo
et al. 1980) show a negative reaction in early 4-cell stage and a positive
reaction in late 4-cell stage, without intervening cell divisions; secondly, the
cytochemical demonstrations of 5'-nucleotidase and alkaline phosphatase have
not revealed any cytoplasmic component that might be regarded as a membrane
precursor, near to the place where the membrane is supposedly assembled.
Even a less extreme interpretation, if based on the formation of patches of'new'
cell membrane, would imply that these enzymes are prevented from diffusing
on the plane of the membrane; however, up to now we have seen no ultra-
5'-nucleotidase in early mouse embryo
125
structural evidence of cytoplasmic attachments in appropriate positions
(see Izquierdo et al. 1980) and our observations on the effect of cytoskeletal
inhibitors are still indecisive.
In this and former reports we have considered 5'-nucleotidase and alkaline
phosphatase activities exclusively as markers of cell membrane regionalization
without any regard for their physiological role in preimplantation embryos.
This remains to be elucidated. In other better known and comparable systems,
such as lymphoblast differentiation and lymphocyte proliferation, the rol$ of
5'-nucleotidase seems to be independent of cell membrane formation (Dornard,
Bonnafous, Gavach & Mani, 1979; Edwards et al. 1979). It is worth mentioning
though, that by analogy with the formation of cell wall in bacteria it has been
suggested that phosphatases, such as 5'-nucleotidase and alkaline phosphatase,
might be involved in the synthesis of cell membrane (Fishman, 1974).
We have as yet found no direct relationship between 5'-nucleotidase or
alkaline phosphatase and the establishment of zonular tight junctions; however,
the regionalization revealed by these enzymes offers a clue to the study of
spatial signals that may be required, among other things, for the positioning of
cell junctions and the morphogenesis of the blastocyst.
This research was supported by grants from the Ford Foundation, PNUD/UNESCO
and the University of Chile.
REFERENCES
J. D., WHITTEN, W. K. & WHITTINGHAM, D. G. (1971). The culture of mouse
embryos in vitro. In Methods in Mammalian Embryology (ed. J. D. Daniel), pp. 86-115.
COPP, A. J. (1979). Interaction between inner cell mass and trophectoderm of the mouse
blastocyst. II. The fate of the polar trophectoderm. /. Embryol. exp. Morph. 51, 109-120.
DORNARD, J. BONNAFOUS, J. C. GAVACH, C. & MANI, J. C. (1979). 5'nucleotidase-facilitfcted
adenosine transport by mouse lymphocytes. Biochemie 61, 973-977.
DUCIBELLA, T., ALBERTINI, D. F., ANDERSON, E. & BIGGERS, J. D. (1975). The preimplantation
mammalian embryo; characterization of intercellular junctions and their appearance
during development. Devi Biol. 45, 231-250.
DUCIBELLA, T. & ANDERSON, E. (1975). Cell shape and membrane changes in the eight*cell
mouse embryo: pre-requisites for morphogenesis of the blastocyst. Devi Biol. 47, 45«-58.
DUCIBELLA, T. (1977). Surface changes of the developing trophoblast cell. In Development
in Mammals, vol. 1, (ed. M. H. Johnson), pp. 5-30.
EDWARDS, L. N., GELFAND, E. W., BURK, L., DOSCH, H. M. & Fox, I. H. (1979). Distribution
of 5'nucleotidase in human lymphoid tissues. Proc. natn. Acad. Sci. U.S.A. 76, 3474-3476.
FERNANDEZ, M. S. & IZQUIERDO, L. (1980). Blastocoel formation in half and double mouse
embryos. Anat. Embryol. 160, 77-81.
FISHMAN, W. H. (1974). Perspectives on alkaline phosphatase isoenzymes. Am. J. Med. 56,
617-646.
ISHIYAMA, V. & IZQUIERDO, L. (1977). The onset of phosphatase activity in early mammalian
embryos. /. Embryol. exp. Morph. 42, 305-308
IZQUIERDO, L., FERNANDEZ, M. S. & LOPEZ, T. (1976). Cell membrane and cell junctions in
differentiation of preimplanted embryos. Archs Biol. Med. exp. 10, 130-134.
IZQUIERDO, L. (1977). Cleavage and differentiation. In Development in Mammals, vol. 2,
(ed. M. H. Johnson), pp. 99-118.
IZQUIERDO, L., LOPEZ, T. & MARTINCORENA, P. (1980). Cell membrane regions in preimplantation mouse embryos. /. Embryol. exp. Morph. 59, 89-102.
BIGGERS,
5
EMB 69
126
L. IZQUIERDO AND C. EBENSPERGER
L. & BECKER, M. I. (1982). Effect of Li+ on preimplantation mouse embryos
J. Embryol. exp. Morph. 67, 151-159.
JOHNSON, M. H., PRATT, H. P. M. & HANDYSIDE, A. H. (1981). The generation and recognition of positional information in the preimplantation mouse embryo. In Cellular and
Molecular Aspects of Implantation (ed. S. R. Glasser & D. W. Bullock), pp. 55-74.
KLAUSHOFER, S. M. & VON MAYERSBACH, H. (1979). Freeze substituted tissue in 5'nucleotidase histochemistry. Comparative histochemical and biochemical investigations. /.
Histochem. Cytochem. 27, 1583-1587.
NIZEYIMANA-RUGINA, E. & MULNARD, J. (1979). Ultrastructural localization of 5'nucleotidase in preimplantation mouse embryos. Archs. Biol. 90, 131-140.
SMITH, R. & MCLAREN, A. (1977). Factors affecting the time of formation of the mouse
blastocoele. /. Embryol. exp. Morph. 41, 79-92.
SUNDSTROM, B. & MONRSTAD, H. (1975). Lead citrate-containing media for use at alkaline
pH: their stabilization with glucose and increased buffer strength. Stain Technol. 50, 287288.
TARKOWSKI, A. K. & WROBLEWSKA, J. (1967). Development of blastomeres of mouse eggs
isolated at the 4- and 8-cell stage. /. Embryol. exp. Morph. 18, 155-180.
UUSITALO, R. J. & KARNOVSKY, M. J. (1977). Surface localization of 5'nucleotidase on the
mouse lymphocyte. /. Histochem. Cytochem. 25, 87-96.
VORBRODT, A., KONWINSKI, M., SOLTER, D. & KOPROWSKI, H. (1977). Ultrastructural cytochemistry of membrane-bound phosphatases in preimplantation mouse embryos. Devi
Biol. 55, 117-134.
WACHSTEIN, M. & MEISEL, E. (1975). Histochemistry of hepatic phosphatases at a physiologic pH. Am. J. Clin. Pathol. 27, 13-23.
YAMASHINA, S. & KAWAI, K. (1979). Cytochemical studies on the localization of 5'nucleotidase in the acinar cells of the rat salivary glands. Histochem. 60, 255-263.
IZQUIERDO,
{Received 26 June 1981, revised 2 December 1981)