Plant Cell Physiol. 41(12): 1365–1372 (2000)
JSPP © 2000
Abscission of Azolla Branches Induced by Ethylene and Sodium Azide
Eiji Uheda 1, 3 and Sumio Nakamura 2
1
2
Research Institute for Advanced Science and Technology, University of Osaka Prefecture, Gakuen-cho, Sakai, Osaka, 593 Japan
Biological Laboratory, Kanagawa Dental College, Inaoka-cho, Yokosuka, Japan
;
Treatment with ethylene accelerated the abscission of
branches of Azolla filiculoides plants. An Azolla plantlet
treated with ethylene at 10 ml liter–1 divided into 4–5 fragments after a lag period of 6–8 h. Ethylene-induced abscission was effectively inhibited by cycloheximide and was associated with an increase in the activities of cellulase and
polygalacturonase. At the fracture surface abscised after
treatment with ethylene, dissolution of the primary walls of
the abscission zone cells was apparent. However, the middle lamella between abscission zone cells was still present.
Immunoelectron microscopy using anti-unesterified pectin
(JIM5) and anti-methylesterified pectin (JIM7) monoclonal antibodies revealed the presence of both JIM5 and
JIM7 epitopes in the wall between abscission zone cells of
branches before abscission occurred. In the middle lamella
remaining after ethylene-induced abscission, only JIM7
epitopes were observed. The features of ethylene-induced
abscission described herein were different from those of the
rapid abscission induced by sodium azide, which implies
that they are mediated by different mechanisms. The possible mechanisms are discussed.
Key words: Azolla filiculoides — Cell wall — Ethylene —
Middle lamella — Rapid abscission — Sodium azide.
Introduction
Azolla is a small water fern with a widespread distribution in the world. The sporophyte of Azolla is 10–40 mm in
diameter (10–25 mm in the case of Azolla filiculoides) and
consists of multibranched stems (rhizomes) that bear deeply
bilobed leaves and determinant adventitious roots. The extensive branching pattern results in numerous stem apices and
various growth habits. Abscission of the branches or roots
allows fragmentation of the plants and facilitates vegetative
propagation (Rao 1935, Konar and Kapoor 1972, Addicott
1982, Peters and Calvert 1983). This plant has been used extensively and effectively in Asia as green manure in rice fields instead of chemical fertilizer, since it can utilize atmospheric N2
due to symbiosis with blue-green alga in its leaflet cavities.
The abscission of Azolla has some unique features. The
branches and roots of Azolla are rapidly abscised (within
20 min) in response to inhibitors of respiration, such as sodium
3
azide, 2,4-dinitrophenol and carbonyl cyanide m-chlorophenylhydrazone (Uheda and Kitoh 1994), and transient exposure to
high temperature stress (Uheda et al. 1999). The rapid abscission is assumed to participate in the propagation of this species
under various stresses. The process of the rapid abscission is
very similar to that of abscission in other plants with respect to
the separation of intact cells (Uheda and Kitoh 1994). The rapid abscission induced by sodium azide seems to involve an increase in the external medium pH. With respect to the speed of
the rapid abscission, we have proposed that, unlike abscission
in other plants, abscission in Azolla might be suspended at an
intermediate stage that precedes final separation (Uheda et al.
1994, Uheda et al. 1995b). Hydrolytic enzymes such as pectic
enzymes and cellulases might have already been synthesized,
but the activities might be suppressed. During the rapid abscission, these enzymes may be activated via a change in the external medium pH. However, our hypothesis remains to be
proven.
Ethylene has been reported to accelerate the abscission of
leaves, petals, flowers and fruit in a wide range of plant species
(Addicott 1982). In general, ethylene-induced abscission
usually requires for completion 10–48 h in fruits and leaves
and 2.5–8 h in flowers (Sexton and Roberts 1982). Inhibitors of
protein synthesis administered during the lag period prior to
abscission have been found to delay abscission markedly, with
cycloheximide being particularly potent in this regard (Sexton
and Roberts 1982). Upon exposure to ethylene, there is an
increase in the activity of enzymes responsible for degradation of primary wall and middle lamella in the abscission zone.
An increase in the activity of two enzymes, cellulase and polygalacturonase, is well documented to accompany the abscission (Osborne 1989, van Doorn and Stead 1997, González-Carranza et al. 1998).
If our hypothesis about the abscission induced by exogenous stimuli in Azolla is valid, ethylene-induced abscission
may be completed more rapidly in Azolla than in other plants.
Cycloheximide and actinomycin D may not affect ethylene-induced abscission. Furthermore, investigations on the effects of
ethylene on the abscission in Azolla might help us to understand abscission in this species including the rapid abscission.
However, most of the effects of ethylene on abscission in
Azolla remain unknown. Although the physiological and structural features of the sodium azide-induced abscission have been
described (Uheda et al. 1995b), here we compared the physiological and structural features of the ethylene-induced abscis-
Corresponding author: E-mail, [email protected]; Fax: +81-722-52-1163.
1365
1366
Abscission of Azolla branches
sion to those of the rapid abscission induced by sodium azide
simultaneously. We discuss the possible mechanisms for both
ethylene-induced and sodium azide-induced abscission. However, in order to compare both abscission systems in the same
publication, examinations were replicated.
Materials and Methods
Treatment with sodium azide and ethylene
Azolla filiculoides was grown in liquid culture (Uheda 1986) in a
growth cabinet under a 16-h light (26C) and 8-h dark (20C) cycle
(Uheda et al. 1995a). An exponentially growing plantlet of Azolla
(1515 mm) was rinsed with distilled water and transferred to an
Erlenmeyer flask (volume, 140 ml) that contained 40 ml of solution.
Each solution contained 10 mM MES [2-(N-morpholino)ethanesulfonic
acid] buffer at pH 5.7 and sodium azide at various concentrations.
Plants were treated with ethylene in an open-flow system. An Azolla
plantlet was placed in a flask and exposed to ethylene at various concentrations at a flow rate of 250 ml min–1. To prepare gases with various concentrations of ethylene, 200 l liter–1 ethylene in air (Takachiho
Co., Osaka, Japan) was diluted with air using gas flow meters (Model
RK-1150, Kofloc Co., Kyoto, Japan). The concentration of ethylene
was monitored with a gas chromatograph equipped with a Porapak N
column (diameter, 3 mm; length, 1 m; column temperature, 60C). The
plantlets were then incubated at 24C in light. Light was supplied by
fluorescent lamps (FLR 40, Toshiba Co., Tokyo, Japan) at an intensity
of 90 mol m–2 s–1 at plant level. At regular intervals, the branches
were enumerated. The location of each branch relative to the tip of the
plantlet was also determined under a binocular microscope. For examination of the effects of pretreatment with ethylene on the sodium
azide-induced abscission, an Azolla plantlet in a flask was treated with
10 l liter–1 ethylene as described above for various durations and then
sodium azide at concentrations of 0.1 and 0.5 mM was added to the
flask. At regular intervals, the number of branches was determined.
Treatment with cycloheximide and actinomycin D
For examination of the effects on the ethylene-induced abscission, plantlets were treated with 10 l liter–1 ethylene for 20 h as described above in the presence of cycloheximide or actinomycin D at
various concentrations. For examination of the effects on the sodium
azide-induced abscission, plantlets were floated on a solution of cycloheximide or actinomycin D at various concentrations. Then sodium
azide at 0.5 mM was added to each solution and incubated for 60 min
as described above.
Enzyme extraction
Azolla plantlets were treated with 0.5 mM sodium azide for 1 h,
or 10 l liter–1 ethylene for various periods. After treatment, proximal
bases about 2–3 mm in length from the fracture surface of an abscised
branch and from branches further than the fourth branch from the tip
that were still attached were cut and collected under a binocular microscope. When Azolla plantlets were abscised, many cells were released
from the abscission zone into the incubation medium. The released
cells were also collected by centrifugation for 5 min at 1,000g. As a
control, plants in flasks were incubated under air. Proximal bases of
branches (2–3 mm in length) further than the fourth branch from the
tip were cut from the plants. In order not to cause abscission during the
collection, the working-time was kept within 2 h. To collect an appreciable amount of abscission zone for enzyme assay, procedures were
repeated scores of times. The collected materials were stored at –80C
before use. The plant materials were homogenized in a small volume
of 20 mM MES [2-(N-morpholino)ethanesulfonic acid] buffer, pH 5.7,
that contained 3% polyvinylpolypyrrolidone (Sigma Chem. Co., St.
Louis, MO, U.S.A.) and 1 M sodium chloride. The homogenate was
centrifuged at 30,000g for 20 min, and the supernatant (extract) was
decanted and saved. The precipitated material was re-extracted in a
similar manner. The extracts were dialyzed against distilled water for
24 h, clarified by centrifugation at 30,000g for 15 min and then used
for enzymatic analysis. All of the above procedures were performed at
4C.
Enzyme assay
The activities of polygalacturonase and cellulase were determined by viscometric methods according to Bonghi et al. (1993) with
minor modifications. For the determination of cellulase activity, 3 ml
of an assay mixture that contained 0.075% (w/v) carboxymethylcellulose (Wako Chem. Co., Osaka, Japan) as a substrate, an appropriate
volume of plant extract and 20 mM sodium phosphate buffer, pH 6.5,
was incubated for 7 h at 37C. For the determination of polygalacturonase activity, 0.2% (w/v) sodium polypectate (Sigma) and 20 mM sodium acetate buffer, pH 5.0, were used as a substrate and buffer, respectively. The assay conditions were the same as those used for the
assay of cellulase. One unit of the activity of cellulase or of polygalacturonase was defined as the amount that caused a 1% decrease in the
substrate flow rate in the control experiments during a 1-h incubation
period. Control experiments were determined at zero time. Proteins
were quantitated as described by Bradford (1976), with bovine serum
albumin as the standard.
Light and transmission electron microscopy
Plantlets treated with 0.5 mM sodium azide for 60 min and 10 l
liter–1 ethylene for 20 h, and plantlets without treatment were separately fixed in 2.5% glutaraldehyde in 80 mM sodium phosphate buffer,
pH 7.2, for 3 h at 22C and post-fixed in 2% osmium tetroxide in the
same buffer for 2 h at 22C. They were dehydrated in a graded ethanol
series and then embedded in Spurr’s resin (Spurr 1969). Ultrathin sections (60–70 nm thick) were cut and stained as described by Roland
(1978), with the modifications described by Uheda et al. (1994), with
thiocarbohydrazide-silver proteinate, a stain specific for polysaccharides. For light microscopy, 1-m sections were placed on glass slides
and stained with 1% methylene blue for 20 s at 70C.
Immunoelectron microscopy
Plantlets were treated with a fixative (2% paraformaldehyde,
0.5% glutaraldehyde, 0.2% picric acid, 0.1% tannic acid and 0.1 M
cacodylate buffer, pH 7.2) for 2 h at room temperature. Plantlets were
then washed with 0.05 M cacodylate buffer, pH 7.2, dehydrated in a
graded ethanol series and embedded in LR White (London Resin. Co.,
London, U.K.). Ultrathin sections on a nickel grid were incubated with
normal goat serum (Cederlane Lab., Hornby, Ontario, Canada) diluted
1 : 30 in 0.05 M Tris-buffered saline, pH 7.2, to block nonspecific antibody binding for 10 min at room temperature. The sections on grids
were then treated with anti-unesterified pectin (JIM5) and anti-methylesterified pectin (JIM7) monoclonal antibodies (Knox et al. 1990)
diluted 1 : 100 in 0.05 M Tris-buffered saline, pH 7.2, for 16 h at 4C.
The grids were then rinsed in 0.05 M Tris-buffered saline which contained 0.2% (w/v) bovine serum albumin and incubated in a secondary
antibody solution containing goat anti-rat AuroProbe EM GARa G10
(Amersham, Arlington Heights, IL, U.S.A.) diluted 1 : 20 in 0.05 M
Tris-buffered saline containing 0.2% bovine serum albumin for 2 h at
room temperature. The grids were rinsed in distilled water and stained
with a saturated solution of uranyl acetate for 10 min and with basic
lead citrate for 5 min. As negative controls for JIM5 and JIM7, the incubation with the primary antibody was omitted. The sections were incubated with normal goat serum, rinsed, incubated in appropriate sec-
Abscission of Azolla branches
Fig. 1 Effects of ethylene on abscission of Azolla branches. Azolla
plants in flasks were treated with various concentrations of ethylene
for 20 h. Data represent means standard deviations (n=5).
1367
Fig. 2 Time course of ethylene-induced and sodium azide-induced
abscission. Azolla plantlets were incubated with ethylene at 10 ml
liter–1 (gray circles), with sodium azide at 0.5 mM (black circles) or
without treatment (black squares). Data represent means standard
deviations (n=6).
ondary antibody-gold conjugates, rinsed and then stained as described
above.
Results
The treatment with ethylene induced the abscission of
Azolla branches (Fig. 1, 2, 3). When Azolla plantlets were treated with ethylene at 10 l liter–1, each plantlet separated into
four to five fragments after a lag period of 6–8 h (Fig. 2). The
sites at which the branches abscised were further than the fifth
position from the tip (Fig. 3). Although an attempt was made to
Fig. 3 Dependence of ethylene-induced and sodium azide-induced
abscission on the location of branches relative to the tip. Ten Azolla
plantlets were treated with ethylene at 10 l liter–1 for 20 h (gray bars)
or 0.5 mM sodium azide for 60 min (black bars). Position no. 9 in the
sodium azide-induced abscission and position no. 10 in the ethyleneinduced abscission were furthest from the tip.
shorten the lag period or to increase the number of abscised
branches by increasing the concentration of ethylene to 100 l
liter–1, the increase in concentration did not enhance the effect
on abscission. Sodium azide as an accelerator of abscission was
more potent than ethylene (Fig. 2, 3). Abscission was completed within 30 min. Each frond was divided into about 10–12
fragments. The sites at which branches abscised were further
than the fourth position from the tip (Fig. 3). Ethylene-induced
abscission was effectively inhibited by cycloheximide, while
actinomycin D was relatively ineffective (Table 1). Increasing
Fig. 4 Activities of cellulase (white circles) and polygalacturonase
(black circles) during ethylene-induced abscission. Plants were cultured for desired times under 10 l liter–1 ethylene (solid lines) or air
(dotted line). Data represent means of two separate experiments.
1368
Abscission of Azolla branches
Table 1 Effects of cycloheximide and actinomycin D on
ethylene-induced and sodium azide-induced abscission
Treatment
Number of branches
None*
Ethylene (10 l liter–1)
+ cycloheximide (10 g ml–1)
+ actinomycin D (20 g ml–1)
Sodium azide (0.5 mM)
+ cycloheximide (10 g ml–1, 5 h)
+ actinomycin D (20 g ml–1, 5 h)
+ cycloheximide (10 g ml–1, 10 h)
+ actinomycin D (20 g ml–1, 10 h)
1.0
4.61.3
1.60.5
3.30.9
9.71.8
10.22.3
9.11.9
8.52.2
9.31.9
For examination of the effects on ethylene-induced abscission, plantlets
were treated with ethylene for 20 h in the presence of cycloheximide or
Actinomycin D. For examination of the effects on the sodium azideinduced abscission, plants were floated on solutions of cycloheximide
and actinomycin D for 5 or 10 h. Then after addition of 0.5 mM sodium
azide each solution was incubated for 60 min. Numbers are means standard deviations (n=6). * Plantlet was cultured without treatment for
20 h. Cycloheximide significantly inhibited the ethylene-induced
abscission (P<0.05).
Table 2 Activities of cellulase and polygalacturonase during
sodium azide-induced abscission
Treatment
Activity (unit (mg protein)–1)
Cellulase Polygalacturonase
None
0.9
Sodium azide-induced abscission 0.7
0.4
0.4
Plants were treated with or without 0.5 mM sodium azide for 60 min.
Numbers are means of two separate experiments.
the concentration of actinomycin D to 40 g ml–1 had no inhibitory effect (data not shown). The abscission seemed to be associated with an increase in the activities of cellulase and polygalacturonase (Fig. 4). By contrast, sodium azide-induced
abscission was not inhibited by either actinomycin D or cycloheximide (Table 1), nor was it associated with an increase in
cellulase and polygalacturonase activities during the abscission (Table 2). Pretreatment of Azolla plants with ethylene for
10 and 20 h followed by treatment with sodium azide had no
enhancing effect on sodium azide-induced abscission (Fig. 5).
Fig. 5 Effects of pretreatment with ethylene on the sodium azideinduced abscission. Azolla plantlets were pretreated with 10 l liter–1
ethylene for 0 (solid line), 10 (fine dotted line) and 20 h (thick dotted
line) and then treated with 0.1 (black, gray and white squares) and
0.5 mM (black, gray and white circles) sodium azide for various periods. Data represent means standard deviations (n=5).
The electron microscopic observation of sections of
branches without treatment revealed the middle lamella between the abscission zone cells weakly stained by thiocarbohydrazide silver proteinate (Fig. 6A, B, C). Individual abscission
zone cells adhered to adjacent cells via weakly stained middle
lamella. At the fracture surface abscised after treatment with
ethylene, dissolution and modification of the primary walls of
the abscission zone cells were apparent. However, the middle
lamella between abscission zone cells was still present in ethylene-induced abscission (Fig. 6D, E). In the rapid abscission induced by sodium azide, the middle lamella completely disappeared from the abscission zone cells (Fig. 6F).
Immunoelectron microscopy of sections of the untreated
plants revealed the presence of both JIM5 and JIM7 epitopes in
the middle lamella between the abscission zone cells (Fig. 7A,
B). The remaining middle lamella of the fracture surface abscised after treatment with ethylene reacted with only JIM7
(Fig. 7C, D). The fracture surface abscised after treatment with
sodium azide reacted neither with JIM5 nor with JIM7 (data
not shown). Control sections incubated only with secondary
antibody-colloidal gold were not labeled.
Fig. 6 Light (A) and electron (B–F) micrographs of Azolla plantlets before (A–C) and after treatment with ethylene at 10 l liter–1 for 20 h (D
and E) or with 0.5 mM sodium azide for 1 h (F). (A–C) The base of the sixth branch from the tip. Abscission zone cells are visible at the base of
the branch (arrowheads in A). The black star in A indicates that a branch is about to be abscised. Middle lamella between adjacent abscission zone
cells is faintly stained by thiocarbohydrazide-silver proteinate (double arrowhead in B and C). (D–E) The fracture surface of the sixth branch from
the tip after treatment with ethylene. Breakdown and modification of the primary walls of the cells at the fracture surface are apparent. However,
the middle lamella is still visible (arrowheads in D and E). (F) Fracture surface of the sixth branch from the tip after treatment with sodium azide.
Middle lamella has completely disappeared (arrowheads in F). Electron micrograph E is enlarged view of D. Bars represent 100 m in A and
1 m in B–F.
Abscission of Azolla branches
1369
1370
Abscission of Azolla branches
Fig. 7 Immunogold electron micrographs of Azolla plantlet labeled with JIM5 (A and C) or JIM7 (B and D). (A and B) The cell wall of the
abscission zone cells in the sixth branch before treatment. (C and D) The cell wall of the fracture surface of the sixth branch from the tip abscised
by treatment with ethylene at 10 ml liter–1 for 20 h. Bars represent 500 nm.
Abscission of Azolla branches
Discussion
Contrary to our expectations, the physiological and structural features of ethylene-induced abscission in Azolla, such as
time course, inhibitory effect of cycloheximide and breakdown
of primary wall, were similar to those in many other plants, although the middle lamella remained after abscission. The process of ethylene-induced abscission in Azolla appears to involve
a mechanism similar to that in many other abscission systems.
Our preliminary enzyme analysis suggested that, as in other abscission systems, both cellulase and polygalacturonase participate in the ethylene-induced abscission of Azolla. The reason
why actinomycin D was relatively ineffective is unknown. The
ineffectiveness of actinomycin D may be related to inadequate
transport to and/or uptake by abscission zone cells or, alternatively, the mRNA may be synthesized earlier (van Doorn and
Stead 1997). With respect to the remainder of the middle lamella between abscission zone cells after treatment with ethylene,
immunoelectron microscopic observations using anti-methyl
esterified pectin and anti-unesterified pectin monoclonal antibodies suggest that the phenomenon is closely associated with
the degree of methylation of the middle lamella pectin in the
abscission zone. Pectin with a high degree of methylation in
the middle lamella, which is resistant to hydrolysis by certain
endopolygalacturonases in vitro (Burns 1991), may also be resistant to polygalacturonase attack in vivo. Middle lamella remaining after ethylene-induced abscission has been reported in
plants other than Azolla (Webster 1973, Bonghi et al. 1993).
Bonghi et al. (1993) reported no increase in polygalacturonase
during the ethylene-induced abscission of Prunus leaf. Some
mechanisms depending on plant species might be involved in
the persistent presence of middle lamella pectin.
The physiological and structural features of sodium azideinduced abscission described herein were different from those
of ethylene-induced abscission. Furthermore, the pretreatment
with ethylene had no enhancing effects on sodium azide-induced abscission. Thus the sodium azide-induced abscission
and ethylene-induced abscission probably occurs via different
mechanisms. The lack of increase in cellulase or polygalacturonase activity after sodium azide-induced abscission and the absence of effects of cycloheximide and actinomycin D suggest
that, in the sodium azide-induced abscission, neither cellulase
nor polygalacturonase play a major role. In sodium azide-induced abscission, the dissolution of pectin may take place without activation of polygalacturonase. During the sodium azideinduced abscission, a rapid dissolution of both methylesterified and unesterified pectin is thought to occur. Even if activation of polygalacturonase, which is undetectable in our assay
conditions, takes place, the dissolution cannot be explained by
a simple activation of the same polygalacturonase which acts in
ethylene-induced abscission. In such a case, polygalacturonase
may hydrolyze the middle lamella in association with pectin
methylesterase, though pectin methylesterase is believed not to
1371
be involved in abscission (Osborne 1989, van Doorn and Stead
1997).
Azolla is not suitable for biochemical studies on abscission because the abscission zone is very small. The abscission
zone must be collected by dissection under a microscope. Such
work is laborious and time-consuming. Furthermore, many
cells and tissues other than those in the abscission zone contaminate the preparation, even if special care is taken for collection.
However, to clarify further details of Azolla abscission
(both ethylene-induced and sodium azide-induced abscission),
it is obvious that molecular and biochemical investigations of
the enzymes involved in abscission are necessary. We are currently investigating biochemical approaches to the above
issues.
Acknowledgments
We thank Dr. J. P. Knox (University of Leeds, U.K.) for providing monoclonal antibodies, JIM5 and JIM7, and Dr. T.Esaka (Osaka
Science Education Institute of Osaka Prefecture, Japan) for his collaboration in the observations by light and electron microscopy. This
work was supported in part by a Grant-in-Aid (no. 08640835) to E. U.
from the Ministry of Education, Science and Culture of Japan.
References
Addicott, F.T. (1982) Abscission. Berkeley, University of California Press, Berkeley.
Bonghi, Z.C., Casadoro, G., Ramina, A. and Rascio, N. (1993) Abscission in
leaf and fruit explants of Prunus persica (L.) Batsch. New Phytol. 123: 555–
565.
Bradford, M.M. (1976) A rapid and sensitive method for the quantitation of
microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72: 248–254.
Burns, J.K. (1991) The polygalacturonases and lyases. In The Chemistry and
Technology of Pectin. Edited by Walter, R.H. pp. 165–188. Academic Press,
San Diego.
González-Carranza, Z.H., Lozoya-Gloria, E. and Roberts, J.A. (1998) Recent
developments in abscission: shedding light on the shedding process. Trends
Plant Sci. 3: 10–14.
Knox, J.P., Linstead, P.J., King, J., Cooper, C. and Roberts, K. (1990) Pectin
esterification is spatially regulated both within cell walls and between developing tissues of root apicies. Planta 181: 512–521.
Konar, R.N. and Kapoor, R.K. (1972) Anatomical studies on Azolla pinnata.
Phytomorphology 22: 211–223.
Osborne, D.J. (1989) Abscission. Crit. Rev. Plant Sci. 8: 103–129.
Peters, G.A. and Calvert, H.E. (1983) Azolla-Anabaena azollae symbiosis. In
Algal Symbiosis. Edited by Goff, L.J. pp. 109–145. Cambridge University
Press, Cambridge.
Rao, H.S. (1935) The structure and life-history of Azolla pinnata R. Brown with
remarks on the fossil history of the Hydropterideae. Proc. Indian Acad. Sci.
Sect. B 2: 175–200.
Roland, J.C. (1978) General preparation and staining of thin sections. In Electron Microscopy and Cytochemistry of Plant Cells. Edited by Hall, J.L. pp. 2–
62. Elsevier/North Holland, Amsterdam.
Sexton, R. and Roberts, J.A. (1982) Cell biology of abscission. Annu. Rev. Plant
Physiol. 33: 133–162.
Spurr, A.R. (1969) A low-viscosity epoxy resin embedding medium for electron microscopy. J. Ultrastruct. Res. 26: 31–43.
Uheda, E. (1986) Isolation of empty packets from Anabaena-free Azolla. Plant
Cell Physiol. 27: 1187–1190.
Uheda, E. and Kitoh, S. (1994) Rapid shedding of roots from Azolla filiculoides
1372
Abscission of Azolla branches
plants in response to inhibitors of respiration. Plant Cell Physiol. 35: 37–43.
Uheda, E., Kitoh, S., Dohmaru, T. and Shiomi, N. (1995a) Isolation and analysis of gas bubbles in the cavities of Azolla leaves. Physiol. Plant. 93: 1–4.
Uheda, E., Kitoh, S. and Shiomi, N. (1999). Response of six Azolla species to
transient high-temperature stress. Aquat. Bot. 64: 87–92.
Uheda, E., Nakamura, S. and Kitoh, S. (1994) Events associated with the rapid
separation of cells from detached roots of Azolla filiculoides depend on pH. J.
Exp. Bot. 45: 1451–1457.
Uheda, E., Nakamura, S. and Kitoh, S. (1995b) Aspects of the very rapid abscission of Azolla branches: anatomy and possible mechanism. Int. J. Plant Sci.
156: 756–763.
van Doorn, W.G. and Stead, A.D. (1997) Abscission of flowers and floral parts.
J. Exp. Bot. 48: 821–837.
Webster, B.D. (1973) Ultrastructural studies of abscission in Phaseolus: ethylene effects on cell walls. Amer. J. Bot. 60: 436–447.
(Received May 18, 2000; Accepted September 28, 2000)