Planta
Planta (Berl.)128, 4 9 - 5 8 (1976)
9 by Springer-Verlag 1976
Some Ultrastructural Observations on the Nature of Foliar Abscission in Impatiens sultani Roy Sexton Department of Biology, University of Stirling, Stirling FK9 4LA, U.K.
Summary. Both scanning and transmission electron microscopes have been used to study the anatomy of the abscission zone of Impatiens sultani Hook. Evidence is presented to show that the fracture line follows the middle lamella in all the living cells of the abscission zone including those in the vascular traces. The separation of these cells is preceded by a breakdown of the middle lamellar region of the wall. The characteristics of this process vary in different cell types. Accompanying this breakdown is an enlargement of inner cortex cells mainly in a direction parallel to the axis of the petiole. It is suggested that this expansion of cells is necessary to produce the tensions which rupture the cuticle and xylem vessels prior to separation. The occurrence of transfer cells and tyloses in the abscission zone is also described and the physiological implications of their presence discussed.
Introduction
There is now a considerable literature concerned with the morphological changes which occur during the process of leaf abscission (see review Webster, 1973 a). Most of these observations have been made by direct light microscope examination, sometimes used in conjunction with histochemical procedures. The limited resolution of the light microscope has imposed considerable restraint on these investigations. As a result, observations made at the limits of the instrument's resolution have been controversial, particularly in respect to such features as the state of membranes and organelles, and the substructure of the wall. In spite of the problems posed by the fixation and embedding of this tissue, those observations which have been made with the electron microscope have done much to clarify the situation (Bornman, 1967a; Jensen and Valdo-
vinos, 1967, 1968; Sexton and Hall, 1974; Valdovinos and Jensen, 1968; Webster, 1968, 1973b). In this investigation both the scanning and transmission electron microscopes have been used to discover the nature and extent of the hydrolytic processes which lead to the separation of abscission zone cells. Unlike the earlier electron microscope studies the investigation is not limited to a description of parenchyma cells alone but describes the changes which occur in all the tissue types found in the separation zone. The developmental implications of these results are discussed and proposals concerning the mechanism of fracture put forward.
Materials and Methods Plants of Impatiens sultani were propagated in a greenhouse maintained above a minimum temperature of 18~ C and in which natural daylight was supplemented by incandescent lamps. When the plants had reached 40 cm in height foliar abscission was induced by delaminating the petiole 20 mm above the point of leaf insertion. Complete abscission of the petiole occurred 36 h after induction. Just prior to fracture the incipient separation zone could be easily recognized. The zone was excised and fixed for 3 h in 2% glutaraldehyde in 25 mM sodium cacodylate buffer pH 7.2. Following thorough washing in buffer the tissue was post fixed in 2% OsO4 in 100 mM cacodylate buffer for 3 h. Dehydration was carried out in an ethanol series and the material was finally embedded in Epon-Araldite. Transverse sections of the separation zone were cut with a Reichert OM U3 ultratome. The sections were stained in uranyl acetate and lead citrate as described by Reynolds (1963). The material was examined using an A.E.I. Corinth 275 microscope though some of the low power photographs were taken with a JEM 100C. Material for the scanning electron microscope was prepared by placing the piece of material on a drop of 1% gelatin on a specially insulated t cm stub (E.M. Ventions, Maryland). The stub was placed in a bath of liquid nitrogen which both freezes the tissue and fixes it firmly to the stub base. The material was quickly transferred to a Cambridge Stereoscan and the instrument evacuated as quickly as possible. Using this technique the specimen could be viewed in the instrument for periods of up to 30 min. This simple method gave better results than either freeze or critical point drying.
50 Determinationsof the waterpotentialof the sap expressedfrom the abscissionzonecellsweremadeusingthe depressionof freezing point methoddescribedby Cohenand Atsmon(1972). Observations and Discussion
After removal of the leaf blade the remaining petiole abscises after approximately 36 h. Break strength measurements indicate that the zone begins to weaken 18 h after induction. This initial weakening is followed by a very rapid decline in the strength of the zone. Sections through the zone at various stages in the process indicate that separation is first visible in the cortex cells on the adaxial side of the stele. The fracture spreads rapidly throughout the zone until the vascular traces, outer collenchymatous cells and epidermes are all that remain intact. A break then occurs through the upper epidermis and a small mechanical force is all that is necessary to break the stele. Abscission in Impatiens is apparently not preceded by cell division in the abscission layer (Gawadi and Avery, 1950). Examination of sections through the fracture surfaces showed them to be covered with swollen rounded cells. Counts of sections through ten abscission zones indicated that 81% of the parenchyma cells on the fracture surfaces were intact. This observation is similar to that of Lloyd (1914a) and indicates that the fracture line has followed the middle of the walls (schizolysis, Correns, 1899, quoted by Lloyd, 1914 b) rather than passing directly through the cells of the zone (rhexolysis), which would have resulted in broken cells on the fracture surfaces. It could still be argued that the small proportion of ruptured cells present on the surface represent evidence for rhexolysis; however Abeles (1973) and Sexton and Hall (1974) have pointed out that due to the delicate nature of these cells it is very easy to damage them when preparing the material. It was found that the scanning electron microscope could be used to examine the fracture surfaces without having to handle or interfere with the cells in any way, thus decreasing the risk of such artefacts. Photographs of both proximal and distal fracture surfaces are shown in Figs. 1 and 2 and both appear to be covered in rounded, apparently intact cells (Fig. 3). Nearly all the cortical parenchyma cells are intact; however, a greater proportion of cells in the epidermis and stele do appear ruptured. The increased depth of focus of the SEM can be used with advantage to study the gross morphology of the fracture surfaces. As can be seen in Fig. 1 and 2, a ring of swollen parenchyma cells surrounding the central ' U ' shaped stelar complex stand out from the rest of the cortex cells on both proximal and distal surfaces. Since these two groups of cells would be juxtaposed in the intact tissue it seems likely that their
R. Sexton: UltrastructuralObservationson Foliar Abscission longitudinal expansion would result in the stretching of the stele producing the tension which ultimately causes its fracture. Careful examination of the xylem vessels after separation reveals several features consistent with this proposal. They have clearly been mechanically ruptured and also the spirals of thickening appear to have been pulled apart in the process. A similar expansion of the cortex cells has been described in many other abscission zones (see Wright and Osborne, 1974). Examination of the separation zone just prior to fracture reveals that the cells of the mid-cortex have separated along the line of the middle lamella (Figs. 9, 10). The middle lamella is virtually lost and the spaces produced become partially filled with a dispersion of microfibrillar material. Wall breakdown of this type has been found in other cases where parenchyma cells of the abscission zone have been examined (Bornman, 1967; Davenport and Marinos, 1971; Sexton and Hall, 1974; Valdovinos and Jensen, 1968). This observation is consistent with both the histochemical (Webster, 1973 a) and biochemical data (Morre, 1968) which indicate that pectin is the main component to be lost from the walls. The layer of cells in which the middle of the wall is removed is three to four cells in width (Fig. 10). The extensive nature of the wall hydrolysis suggests that these cells actively secrete the wall degrading enzymes. On either side of the secretory cells is a further layer where degradation is limited to the walls parallel to the axis of the petiole. This is probably the result of enzyme diffusion from the secretory layer along the intercellular spaces which run predominantly in this direction. The fine structural features of these cells will be the subject of a more extensive report but those micrographs which have been included give some important information. Firstly, the cells on both the proximal or distal side of the final fracture line have intact plasmalemma and tonoplast membranes (Figs. 9, 10). This is contrary to the observations made by Bornman (1967) which led to the hypothesis that while the integrity of these membranes is maintained, the free movement of hydrolytic enzymes into the wall is restricted. In Impatiens this can not be the case since the walls are at an advanced stage of breakdown while the membranes still appear intact. Secondly, the cytoplasm appears to be fairly active, there being present large numbers oforganelles (mitochondria, chloroplasts and dictyosomes) as well as considerable areas of polysome-covered rough endoplasmic reticulum. The expansion of the cells of the inner cortex is not so apparent in fixed and embedded material (Fig. 14). The dehydration procedures unfortunately result in the shrinkage and separation of the cells (Fig. 10). The simplest explanation for the enlargement
R. Sexton: Ultrastructural Observations on Foliar Abscission
51
Figs. 1 and 2. Scanning electron micrographs of the proximal (Fig. 1) and distal fracture surfaces (Fig. 2). Note the absence of broken ceils and the ring of enlarged parenchyma cells (arrows) which is raised above the general cortex. The plane of fracture in the stele (St) is displaced. Fig. 1 x 104, Fig. 2 x 130 Fig. 3. A scanning electron micrograph showing details of the cells on the proximal fracture surface. Note the intact rounded nature of most cells and the apparent gelatinous consistency of the material between them (arrow) x 800
52
R. Sexton : Ultrastructural Observations on Foliar Abscission
Fig. 4. A low power electron micrograph of a transverse section through the distal fracture surface on the adaxial side of the petiole. Epi and Epii are two epidermal cells joined by a strip of cuticle (Cu). Coi, Coii and Coiii are underlying collenchymatous cells. The probable points of attachment of Epi and Col to the fracture surface are indicated by the triangular symbols. The regions outlined on Epii are shown magnified in Figs. 5 and 6. Note the intact nature of the cells and the integrity of tonoplast and plasma-membrane x 1,100 Fig. 5. An enlarged view from a serial section of the side wall of the epidermal cell Epii in Fig. 4. Note how the central region of the wall has been broken down. The line of fracture is shown by double headed arrows x 7,000
53
Fig. 6. An electron micrograph of a group of collenchymatous cells on the distal fracture surface. The line of fracture is shown by the double headed arrows. Note the enlarged cell walls, fenestrated at the corners (F). Dispersed microfibrillar material (DM) fills the region between the cells x 1,600 Fig. 7. An enlarged area of the exposed wall of the epidermal cell Epii in Fig. 4. Cell wall degradation can be seen to be occurring just below the cuticle (Cu) • 20,000 Fig. 8. An electron micrograph of two separating xylem transfer cells. Note that the middle lamellar region of the wall has disappeared, the primary wall (PW) is extensively degraded and that the secondary wall ingrowths (Sg0 have also been hydrolysed x 900
Fig. 9. A view of the fracture line (double headed arrows) between two parenchyma cells. Note the absence of the middle lameIla region of the wall x 15,000 Fig. 10. A montage of the separation zone in the region of the cortical parenchyma. The direction of fracture is indicated by the arrow. Note how the cells have become completely separated from one another. All the cells in the micrograph were examined and shown to have an intact tonoplast and plasmamembrane x 800 Fig. 11. A high power view of the sieve element shown in Fig. 12. Note how the fracture has passed through .the centre of the sieve plate (Sp) and the callose deposits on the surface of the sieve area Fig. 12. A low pov~ei electron micrograph of the vascular region on the proximal fracture face. The line of separation can be seen to run through the end walls of a sieve tube (Se), a phloem transfer cell (Pt) and some cambial initials (Ca) • 2,000
55
Fig. 13. A low power electron micrograph of a longitudinal section through the xylem region proximal to the fracture line. Note the large numbers of xylem transfer cells (Tc) and closely adpressed tyloses (Ty). The tylose Tyi is derived from an adjacent transfer cell x 2,000 Fig. 14. A low power electron micrograph of a swollen cortical cell stelar cells (St) will give some indication of the cell's size x 850
(Ec) on the distal fracture surface. Comparison with the adjacent
Fig. 15. A view of the xylem in the separation zone just prior to fracture. Wall breakdown between parenchyma cells (P) and transfer cells (Tc) is shown by arrows x 1,800
56
of these cells is that the wall degrading enzymes weaken the wall allowing it to expand as a result of the turgor pressure within. It could be argued that decreasing the solute potential inside the cell should have a similar effect. In order to investigate this possibility we compared the water potential of the sap of cells which were beginning to expand with those from control tissue. To obtain cells in this state, fracture was induced prematurely at the stage when 50 g was required to break the zone. The central cortex cells on the two exposed fracture surfaces were removed with a razor blade and quickly frozen in liquid nitrogen. The sap was extracted by the centrifugation technique of Cohen and Atsmon (1970) and the osmotic potential of the sap determined by the depression of freezing point method. The value obtained for the control tissue was-5.70 bars whilst that for proximal and distal fracture surfaces was-6.30 bars a n d - 6 . 2 0 bars respectively. These results indicate that the osmotic potential of the sap had decreased to some extent prior to expansion of the cells. Thus it is probable that the enlargement of the abscission zone cells is due to both wall weakening and an increase in the concentration of the sap. Fracture of the upper epidermis is also the result of wall breakdown rather than mechanical rupturing of cells. After separation up to three or four detached epidermal cells can often be seen floating free of the fracture surface. An example is the epidermal cell Epi in Fig. 4. Closer examination of serial sections of the epidermal cell next to it (Epii) shows that breakdown of the middle lamella occurs in the basal and side walls (Fig. 5), and that this fracture runs along the exposed surface of the cells just below the cuticle (Fig. 7). There are several reports that there is a pectin band in this position under the cuticle, though it could not be distinguished in the control tissue. The cuticle itself and some cells in the lower epidermis are mechanically ruptured and this appears to be the last event which occurs before separation of the petiole. Just beneath the epidermis, particularly in the axil of the leaf, are several rows of thick walled collenchymatous cells (Figs. 4, 6). The changes which occur in these cells before separation are rather different to those in the rest of the cortex. The middle region of the wall first becomes swollen and electron dense (Figs. 4, 6) and then fenestrations proliferate at the corners of the cell (Fig. 6). The central region of the wall becomes even more dispersed (DM Fig. 6) before separation ultimately occurs. This dispersed microfibrillar material is probably what light microscopists have described as a gelantinized wall. The thin layer of cytoplasm which lines both the epidermal and these outer cortex cells appears quite normal and, as in the other cortical cells, both tonoplast and plasmalemma
R. Sexton : U l t r a s t r u c t u r a l O b s e r v a t i o n s on F o l i a r A b s c i s s i o n
are intact. Some limited enlargement of these cells occurs prior to fracture particularly in the epidermis. The localized stresses generated by this expansion probably provide the force necessary to separate the weakened walls. The anatomical changes in the stele which accompany abscission are more complex. The first visible sign of change is the formation of tyloses in xylem vessels. These are produced as outgrowths from the xylem parenchyma and transfer cells which surround the vessels (Fig. 13), and occur mainly on the proximal side of the fracture. In the period just prior to fracture the xylem becomes almost totally blocked by tyloses, which become tightly adpressed, producing a resistance to water movement (Fig. 13). The functional significance of the tyloses in the abscission zone is not clear. Bornman (1967 b) suggests that tylose formation in cotton is not causally involved in abscission since after a variety of treatments there was an inverse relationship between the rate of abscission and the number of tyloses. Scott et al. (1967) obtained the opposite result in bean. The blocking of the xylem, no matter how complete, must contribute to the water stress in distal tissues. The loss of turgor and shrinkage of the pulvinus cells distal to the separation zone has been recorded during abscission of bean. This results in a decrease in pulvinar diameter (Leopold, 1967; Morre, 1968), and the development of tensions and stresses across the separation zone which lead ultimately to its fracture. In Impatiens no differential changes in the diameter of the petiole on proximal and distal sides of the zone occurs. Thus in this case the role of tyloses in producing stresses across the zone would seem less important. One of the initial steps in the formation of a tylose is the degradation of the pit membrane in the vessel wall. If tylose formation occurs with sufficient frequency it must have a significant effect on the strength of the vessel walls, Thus it is possible that tylosis serves mainly to weaken the stele particularly in abscission zones such as those in Impatiens where there are large numbers of tyloses. In addition the blockage of the xylem may well be important after fracture both by preventing the breakage of the water columns in the xylem, hence maintaining a water supply to the fracture zone, and by restricting the entry of pathogens. Another feature of the stele in the abscission region is the large numbers of transfer cells (Fig. 13). The physiological significance of the presence of these cells is not clear. Gunning et al. (1970) have described transfer cells in association with departing leaf traces and Pate et al. (1970) consider that their role is to supply nutrients to the developing meristematic zones in the node. The transfer cells in the abscission zone may be part of this same nutritional apparatus though
R. Sexton : Ultrastructural Observations on Foliar Abscission
it is possible that they could be involved specifically in supplying nutrients to the separation zone cells. The fracture of the xylem elements is purely mechanical as described above, though the living cells of the stele separate as a result of the degradation of the central regions of their walls. In the xylem parenchyma and cambial initials (Figs. 12, 15) breakdown is similar to that described in the cortical parenchyma, though rather less extensive. Most of the degradation is limited to the wall along which fracture will eventually pass and is not so obvious in the walls at right angles to the plane of fracture. This sort of polarized breakdown of the wall would be expected if two rows of adjacent cells were producing the wall degrading enzymes. The highest concentration of enzyme would be found in the common wall. In the xylem transfer cells breakdown is not limited to the middle lamella but involves considerable dispersion and breakdown of the primary wall and secondary wall ingrowths (Fig. 8). In all these cells the cytoplasm seems normal (Figs. 8, 12, 15) and similar to that described in other cells of the separation layer. Ruptured cells are more frequent in the stele particularly near the xylem vessels. It is thought that this damage results from the sudden release of tensions which build up in the xylem prior to fracture. The separation of the phloem is interesting since cell wall swelling and breakdown not only occur in companion and transfer cells (Fig. 12) but also in the sieve elements (Figs. 11, 12). In the latter the fracture line passes right through the centre of the sieve plate (Fig. 11). This was not expected since it is known that some wall degrading enzymes are synthesised ' de novo' (Lewis and Varner, 1970) and sieve elements appear to have neither the necessary synthetic machinery for the production of these enzymes nor the dictyosomal apparatus which is necessary for their secretion (Cronshaw, 1974; Hall and Sexton, 1974). Wall breakdown in the sieve tubes could be the result of enzyme secretion by neighbouring cells but unfortunately it is difficult to determine if this has taken place. Callose is present around the sieve plate areas both at the fracture surface (Fig. 11) and several cells proximal to it. In view of the rapidity with which callose forms when the phloem is damage (Escherich, 1970) particularly in Impatiens it would seem probable that this is a wound reaction. It was hoped on the basis of the observations reported above to determine which cells in the separation zone were involved in the production of cell wall degrading enzymes. It is possible to state with some conviction that parenchyma cells, xylem transfer cells and the collenchymatous cells in the axil of the leaf are capable of producing these enzymes. This can be deduced because large enough tracts of these cells exist
57
for the possible diffusion of the enzyme from neighbouring tissues to be disregarded. Unfortunately most of the other cell types in the abscission zone are in close proximity to other tissues and thus it is impossible to identify with certainty the secretory cells. The problem is well illustrated by the situation in the epidermis, where two features point to the conclusion that these cells produce their own wall degrading enzymes. Firstly, the epidermis is often more extensively degraded than adjacent tissues and secondly there is a lack of significant gradients of wall breakdown from one side of the epidermis to the other. Neither condition would be expected if enzyme diffusion were responsible, though this explanation can not be totally excluded. Clarification of this issue must await the characterization of cytoplasmic features which allow the identification of secreting cells. I would like to acknowledge the patience and skill of Mr. G. Jamieson in helping prepare and section this material. My thanks are also due to Mr. T. Forrest, Mr. J. Goodall and Dr. A.W. Robards for advice concerning both transmission and scanning electron microscopy, and to Professor H. Meidner for help with the preparation of the manuscript.
References Abeles, F.B. : Ethylene in plant biology. London-New York: Academic Press 1973 Bornman, C.H. : Some ultrastructural aspects of abscission in Coleus and Gossypiurn. S. Afr. J. Sci. 63, 325 330 (1967a) Bornman, C.H.: The relationship between tylosis and abscission in cotton (Gossypium hirsutum L.). S. Aft. J. Agri. Sci. 10, 143154 (1967b) Cohen, D., Atsmon, D. : Measurments of water potential and hormone transport associated with growth of cucumber hypocotyls. In: Proceedings 7th International Conference of Plant Growth Substances. pp. 23-30. Ed. : Carr, D.J. Berlin-Heidelberg-New York: Springer 1972 Cronshaw, J. : Phloem differentiation and development. In : Dynamic aspects of plant ultrastructure, pp. 391413. Ed.: Robards, A.W. London: McGraw Hill 1974 Davenport, T.I., Marinos, N.G. : Cell separation in isolated abscission zones. Aust. J. Biol. Sci. 24, 709 715 (1971) Escherich, W. : Biochemistry and fine structure of phloem in relation to transport. Ann. Rev. Plant Physiol. 21, 193-214 (1970) Gawadi, A.G., Avery, G.S. : Leaf abscission and the so called "abscission layer". Amer. J. Bot. 37, 172-180 (1950) Gunning, B.E.S., Pate, J.S., Green, L.W. : Transfer cells in the vascular system of stems. Taxonomy, association with nodes and structure. Protoplasma (Wien)71, 147-171 (1970) Hall, J.L., Sexton, R.: Fine structure and cytochemistry of the abscission zone of Phaseolus leaves. II. Localization of peroxidase and acid phosphatase in separation zone cells. Ann. Bot. 38, 855 858 (1974) Jensen, T.E., Valdovinos, J.G. : Fine structure of abscission zones. I. Abscission zones in the pedicels of tobacco and tomato flowers. Planta (Berl.) 77, 298 318 (1967) Leopold, A.C.: The mechanism of foliar abscission. Symp. Soc. exp. Biol. 21,507 516 (1967) Lewis, L.N., Varner, J.E. : Synthesis of cellulase during abscission
58 of Phaseolus vulgaris leaf explants. Plant Physiol. 46, 194-199 (1970) Lloyd, F.E. : Injury and abscission in Impatiens sultani. 6th Annual report of the Quebec Society for the Protection of Plants from Insect and Fungal Diseases. 72 79 (1914a) Lloyd, F.E.: Abscission. Ottawa Nat. 28, 61 75 (1914b) Morre, D.J.: Cell wall dissolution and enzyme secretion during leaf abscission. Plant Physiol. 43, 1545-1559 (1968) Pate, J.S., Gunning, B.E.S., Milliken, F.F.: Function of transfer cells in nodal regions of stems particularly in relation to nutrition in young seedlings. Protoplasma (Wien) 71,313 334 (1970) Reynolds; E.S. : The use of lead citrate at high pH as an electron opaque stain in electron microscopy. J. Cell Biol. 17, 208~12 (1963) Sexton, R., Hall, J.L.: Fine structure and cytochemistry of the abscission zone cells of Phaseolus leaves. I. Ultrastructural changes during abscission. Ann. Bot. 38, 849-854 (1974) Scott, F.M., Miller, L.W., Webster, B.D., Leopold, A.C. : Structural
R. Sexton : Ultrastructural Observations on Foliar Abscission changes during bean leaf abscission. Amer. J. Bot. 54, 730-734 (1967) Valdovinos, J.G., Jensen, T.E. : Fine structure of abscission zones. II. Cell wall changes in abscising pedicels of tobacco and tomato flowers. Planta (Berl.) 83, 295-302 (1968) Webster, B.D. : Anatomical aspects of abscission. Plant Physiol. 43, 1512-1544 (1968) Webster, B.D. : Anatomical and histochemical changes in leaf abscission. In: Shedding of plant parts, pp. 45 83. Ed.: Kozlowski, T.T. New York-London : Academic Press 1973 a Webster, B.D. : Ultrastructural studies of abscission in Phaseolus: Ethylene effects on Cell walls. Amer. J. Bot. 60, 436447 (1973b) Wright, M., Osborne, D.J. : Abscission in Phaseolus vulgaris. The positional differentiation of ethylene induced expansion growth of specialized cells. Planta (Bed.) 120, 163 171 (1974)
Received 2 July; accepted 8 September 1975
Planta (Berl.)128, 4 9 - 5 8 (1976)
9 by Springer-Verlag 1976
Some Ultrastructural Observations on the Nature of Foliar Abscission in Impatiens sultani Roy Sexton Department of Biology, University of Stirling, Stirling FK9 4LA, U.K.
Summary. Both scanning and transmission electron microscopes have been used to study the anatomy of the abscission zone of Impatiens sultani Hook. Evidence is presented to show that the fracture line follows the middle lamella in all the living cells of the abscission zone including those in the vascular traces. The separation of these cells is preceded by a breakdown of the middle lamellar region of the wall. The characteristics of this process vary in different cell types. Accompanying this breakdown is an enlargement of inner cortex cells mainly in a direction parallel to the axis of the petiole. It is suggested that this expansion of cells is necessary to produce the tensions which rupture the cuticle and xylem vessels prior to separation. The occurrence of transfer cells and tyloses in the abscission zone is also described and the physiological implications of their presence discussed.
Introduction
There is now a considerable literature concerned with the morphological changes which occur during the process of leaf abscission (see review Webster, 1973 a). Most of these observations have been made by direct light microscope examination, sometimes used in conjunction with histochemical procedures. The limited resolution of the light microscope has imposed considerable restraint on these investigations. As a result, observations made at the limits of the instrument's resolution have been controversial, particularly in respect to such features as the state of membranes and organelles, and the substructure of the wall. In spite of the problems posed by the fixation and embedding of this tissue, those observations which have been made with the electron microscope have done much to clarify the situation (Bornman, 1967a; Jensen and Valdo-
vinos, 1967, 1968; Sexton and Hall, 1974; Valdovinos and Jensen, 1968; Webster, 1968, 1973b). In this investigation both the scanning and transmission electron microscopes have been used to discover the nature and extent of the hydrolytic processes which lead to the separation of abscission zone cells. Unlike the earlier electron microscope studies the investigation is not limited to a description of parenchyma cells alone but describes the changes which occur in all the tissue types found in the separation zone. The developmental implications of these results are discussed and proposals concerning the mechanism of fracture put forward.
Materials and Methods Plants of Impatiens sultani were propagated in a greenhouse maintained above a minimum temperature of 18~ C and in which natural daylight was supplemented by incandescent lamps. When the plants had reached 40 cm in height foliar abscission was induced by delaminating the petiole 20 mm above the point of leaf insertion. Complete abscission of the petiole occurred 36 h after induction. Just prior to fracture the incipient separation zone could be easily recognized. The zone was excised and fixed for 3 h in 2% glutaraldehyde in 25 mM sodium cacodylate buffer pH 7.2. Following thorough washing in buffer the tissue was post fixed in 2% OsO4 in 100 mM cacodylate buffer for 3 h. Dehydration was carried out in an ethanol series and the material was finally embedded in Epon-Araldite. Transverse sections of the separation zone were cut with a Reichert OM U3 ultratome. The sections were stained in uranyl acetate and lead citrate as described by Reynolds (1963). The material was examined using an A.E.I. Corinth 275 microscope though some of the low power photographs were taken with a JEM 100C. Material for the scanning electron microscope was prepared by placing the piece of material on a drop of 1% gelatin on a specially insulated t cm stub (E.M. Ventions, Maryland). The stub was placed in a bath of liquid nitrogen which both freezes the tissue and fixes it firmly to the stub base. The material was quickly transferred to a Cambridge Stereoscan and the instrument evacuated as quickly as possible. Using this technique the specimen could be viewed in the instrument for periods of up to 30 min. This simple method gave better results than either freeze or critical point drying.
50 Determinationsof the waterpotentialof the sap expressedfrom the abscissionzonecellsweremadeusingthe depressionof freezing point methoddescribedby Cohenand Atsmon(1972). Observations and Discussion
After removal of the leaf blade the remaining petiole abscises after approximately 36 h. Break strength measurements indicate that the zone begins to weaken 18 h after induction. This initial weakening is followed by a very rapid decline in the strength of the zone. Sections through the zone at various stages in the process indicate that separation is first visible in the cortex cells on the adaxial side of the stele. The fracture spreads rapidly throughout the zone until the vascular traces, outer collenchymatous cells and epidermes are all that remain intact. A break then occurs through the upper epidermis and a small mechanical force is all that is necessary to break the stele. Abscission in Impatiens is apparently not preceded by cell division in the abscission layer (Gawadi and Avery, 1950). Examination of sections through the fracture surfaces showed them to be covered with swollen rounded cells. Counts of sections through ten abscission zones indicated that 81% of the parenchyma cells on the fracture surfaces were intact. This observation is similar to that of Lloyd (1914a) and indicates that the fracture line has followed the middle of the walls (schizolysis, Correns, 1899, quoted by Lloyd, 1914 b) rather than passing directly through the cells of the zone (rhexolysis), which would have resulted in broken cells on the fracture surfaces. It could still be argued that the small proportion of ruptured cells present on the surface represent evidence for rhexolysis; however Abeles (1973) and Sexton and Hall (1974) have pointed out that due to the delicate nature of these cells it is very easy to damage them when preparing the material. It was found that the scanning electron microscope could be used to examine the fracture surfaces without having to handle or interfere with the cells in any way, thus decreasing the risk of such artefacts. Photographs of both proximal and distal fracture surfaces are shown in Figs. 1 and 2 and both appear to be covered in rounded, apparently intact cells (Fig. 3). Nearly all the cortical parenchyma cells are intact; however, a greater proportion of cells in the epidermis and stele do appear ruptured. The increased depth of focus of the SEM can be used with advantage to study the gross morphology of the fracture surfaces. As can be seen in Fig. 1 and 2, a ring of swollen parenchyma cells surrounding the central ' U ' shaped stelar complex stand out from the rest of the cortex cells on both proximal and distal surfaces. Since these two groups of cells would be juxtaposed in the intact tissue it seems likely that their
R. Sexton: UltrastructuralObservationson Foliar Abscission longitudinal expansion would result in the stretching of the stele producing the tension which ultimately causes its fracture. Careful examination of the xylem vessels after separation reveals several features consistent with this proposal. They have clearly been mechanically ruptured and also the spirals of thickening appear to have been pulled apart in the process. A similar expansion of the cortex cells has been described in many other abscission zones (see Wright and Osborne, 1974). Examination of the separation zone just prior to fracture reveals that the cells of the mid-cortex have separated along the line of the middle lamella (Figs. 9, 10). The middle lamella is virtually lost and the spaces produced become partially filled with a dispersion of microfibrillar material. Wall breakdown of this type has been found in other cases where parenchyma cells of the abscission zone have been examined (Bornman, 1967; Davenport and Marinos, 1971; Sexton and Hall, 1974; Valdovinos and Jensen, 1968). This observation is consistent with both the histochemical (Webster, 1973 a) and biochemical data (Morre, 1968) which indicate that pectin is the main component to be lost from the walls. The layer of cells in which the middle of the wall is removed is three to four cells in width (Fig. 10). The extensive nature of the wall hydrolysis suggests that these cells actively secrete the wall degrading enzymes. On either side of the secretory cells is a further layer where degradation is limited to the walls parallel to the axis of the petiole. This is probably the result of enzyme diffusion from the secretory layer along the intercellular spaces which run predominantly in this direction. The fine structural features of these cells will be the subject of a more extensive report but those micrographs which have been included give some important information. Firstly, the cells on both the proximal or distal side of the final fracture line have intact plasmalemma and tonoplast membranes (Figs. 9, 10). This is contrary to the observations made by Bornman (1967) which led to the hypothesis that while the integrity of these membranes is maintained, the free movement of hydrolytic enzymes into the wall is restricted. In Impatiens this can not be the case since the walls are at an advanced stage of breakdown while the membranes still appear intact. Secondly, the cytoplasm appears to be fairly active, there being present large numbers oforganelles (mitochondria, chloroplasts and dictyosomes) as well as considerable areas of polysome-covered rough endoplasmic reticulum. The expansion of the cells of the inner cortex is not so apparent in fixed and embedded material (Fig. 14). The dehydration procedures unfortunately result in the shrinkage and separation of the cells (Fig. 10). The simplest explanation for the enlargement
R. Sexton: Ultrastructural Observations on Foliar Abscission
51
Figs. 1 and 2. Scanning electron micrographs of the proximal (Fig. 1) and distal fracture surfaces (Fig. 2). Note the absence of broken ceils and the ring of enlarged parenchyma cells (arrows) which is raised above the general cortex. The plane of fracture in the stele (St) is displaced. Fig. 1 x 104, Fig. 2 x 130 Fig. 3. A scanning electron micrograph showing details of the cells on the proximal fracture surface. Note the intact rounded nature of most cells and the apparent gelatinous consistency of the material between them (arrow) x 800
52
R. Sexton : Ultrastructural Observations on Foliar Abscission
Fig. 4. A low power electron micrograph of a transverse section through the distal fracture surface on the adaxial side of the petiole. Epi and Epii are two epidermal cells joined by a strip of cuticle (Cu). Coi, Coii and Coiii are underlying collenchymatous cells. The probable points of attachment of Epi and Col to the fracture surface are indicated by the triangular symbols. The regions outlined on Epii are shown magnified in Figs. 5 and 6. Note the intact nature of the cells and the integrity of tonoplast and plasma-membrane x 1,100 Fig. 5. An enlarged view from a serial section of the side wall of the epidermal cell Epii in Fig. 4. Note how the central region of the wall has been broken down. The line of fracture is shown by double headed arrows x 7,000
53
Fig. 6. An electron micrograph of a group of collenchymatous cells on the distal fracture surface. The line of fracture is shown by the double headed arrows. Note the enlarged cell walls, fenestrated at the corners (F). Dispersed microfibrillar material (DM) fills the region between the cells x 1,600 Fig. 7. An enlarged area of the exposed wall of the epidermal cell Epii in Fig. 4. Cell wall degradation can be seen to be occurring just below the cuticle (Cu) • 20,000 Fig. 8. An electron micrograph of two separating xylem transfer cells. Note that the middle lamellar region of the wall has disappeared, the primary wall (PW) is extensively degraded and that the secondary wall ingrowths (Sg0 have also been hydrolysed x 900
Fig. 9. A view of the fracture line (double headed arrows) between two parenchyma cells. Note the absence of the middle lameIla region of the wall x 15,000 Fig. 10. A montage of the separation zone in the region of the cortical parenchyma. The direction of fracture is indicated by the arrow. Note how the cells have become completely separated from one another. All the cells in the micrograph were examined and shown to have an intact tonoplast and plasmamembrane x 800 Fig. 11. A high power view of the sieve element shown in Fig. 12. Note how the fracture has passed through .the centre of the sieve plate (Sp) and the callose deposits on the surface of the sieve area Fig. 12. A low pov~ei electron micrograph of the vascular region on the proximal fracture face. The line of separation can be seen to run through the end walls of a sieve tube (Se), a phloem transfer cell (Pt) and some cambial initials (Ca) • 2,000
55
Fig. 13. A low power electron micrograph of a longitudinal section through the xylem region proximal to the fracture line. Note the large numbers of xylem transfer cells (Tc) and closely adpressed tyloses (Ty). The tylose Tyi is derived from an adjacent transfer cell x 2,000 Fig. 14. A low power electron micrograph of a swollen cortical cell stelar cells (St) will give some indication of the cell's size x 850
(Ec) on the distal fracture surface. Comparison with the adjacent
Fig. 15. A view of the xylem in the separation zone just prior to fracture. Wall breakdown between parenchyma cells (P) and transfer cells (Tc) is shown by arrows x 1,800
56
of these cells is that the wall degrading enzymes weaken the wall allowing it to expand as a result of the turgor pressure within. It could be argued that decreasing the solute potential inside the cell should have a similar effect. In order to investigate this possibility we compared the water potential of the sap of cells which were beginning to expand with those from control tissue. To obtain cells in this state, fracture was induced prematurely at the stage when 50 g was required to break the zone. The central cortex cells on the two exposed fracture surfaces were removed with a razor blade and quickly frozen in liquid nitrogen. The sap was extracted by the centrifugation technique of Cohen and Atsmon (1970) and the osmotic potential of the sap determined by the depression of freezing point method. The value obtained for the control tissue was-5.70 bars whilst that for proximal and distal fracture surfaces was-6.30 bars a n d - 6 . 2 0 bars respectively. These results indicate that the osmotic potential of the sap had decreased to some extent prior to expansion of the cells. Thus it is probable that the enlargement of the abscission zone cells is due to both wall weakening and an increase in the concentration of the sap. Fracture of the upper epidermis is also the result of wall breakdown rather than mechanical rupturing of cells. After separation up to three or four detached epidermal cells can often be seen floating free of the fracture surface. An example is the epidermal cell Epi in Fig. 4. Closer examination of serial sections of the epidermal cell next to it (Epii) shows that breakdown of the middle lamella occurs in the basal and side walls (Fig. 5), and that this fracture runs along the exposed surface of the cells just below the cuticle (Fig. 7). There are several reports that there is a pectin band in this position under the cuticle, though it could not be distinguished in the control tissue. The cuticle itself and some cells in the lower epidermis are mechanically ruptured and this appears to be the last event which occurs before separation of the petiole. Just beneath the epidermis, particularly in the axil of the leaf, are several rows of thick walled collenchymatous cells (Figs. 4, 6). The changes which occur in these cells before separation are rather different to those in the rest of the cortex. The middle region of the wall first becomes swollen and electron dense (Figs. 4, 6) and then fenestrations proliferate at the corners of the cell (Fig. 6). The central region of the wall becomes even more dispersed (DM Fig. 6) before separation ultimately occurs. This dispersed microfibrillar material is probably what light microscopists have described as a gelantinized wall. The thin layer of cytoplasm which lines both the epidermal and these outer cortex cells appears quite normal and, as in the other cortical cells, both tonoplast and plasmalemma
R. Sexton : U l t r a s t r u c t u r a l O b s e r v a t i o n s on F o l i a r A b s c i s s i o n
are intact. Some limited enlargement of these cells occurs prior to fracture particularly in the epidermis. The localized stresses generated by this expansion probably provide the force necessary to separate the weakened walls. The anatomical changes in the stele which accompany abscission are more complex. The first visible sign of change is the formation of tyloses in xylem vessels. These are produced as outgrowths from the xylem parenchyma and transfer cells which surround the vessels (Fig. 13), and occur mainly on the proximal side of the fracture. In the period just prior to fracture the xylem becomes almost totally blocked by tyloses, which become tightly adpressed, producing a resistance to water movement (Fig. 13). The functional significance of the tyloses in the abscission zone is not clear. Bornman (1967 b) suggests that tylose formation in cotton is not causally involved in abscission since after a variety of treatments there was an inverse relationship between the rate of abscission and the number of tyloses. Scott et al. (1967) obtained the opposite result in bean. The blocking of the xylem, no matter how complete, must contribute to the water stress in distal tissues. The loss of turgor and shrinkage of the pulvinus cells distal to the separation zone has been recorded during abscission of bean. This results in a decrease in pulvinar diameter (Leopold, 1967; Morre, 1968), and the development of tensions and stresses across the separation zone which lead ultimately to its fracture. In Impatiens no differential changes in the diameter of the petiole on proximal and distal sides of the zone occurs. Thus in this case the role of tyloses in producing stresses across the zone would seem less important. One of the initial steps in the formation of a tylose is the degradation of the pit membrane in the vessel wall. If tylose formation occurs with sufficient frequency it must have a significant effect on the strength of the vessel walls, Thus it is possible that tylosis serves mainly to weaken the stele particularly in abscission zones such as those in Impatiens where there are large numbers of tyloses. In addition the blockage of the xylem may well be important after fracture both by preventing the breakage of the water columns in the xylem, hence maintaining a water supply to the fracture zone, and by restricting the entry of pathogens. Another feature of the stele in the abscission region is the large numbers of transfer cells (Fig. 13). The physiological significance of the presence of these cells is not clear. Gunning et al. (1970) have described transfer cells in association with departing leaf traces and Pate et al. (1970) consider that their role is to supply nutrients to the developing meristematic zones in the node. The transfer cells in the abscission zone may be part of this same nutritional apparatus though
R. Sexton : Ultrastructural Observations on Foliar Abscission
it is possible that they could be involved specifically in supplying nutrients to the separation zone cells. The fracture of the xylem elements is purely mechanical as described above, though the living cells of the stele separate as a result of the degradation of the central regions of their walls. In the xylem parenchyma and cambial initials (Figs. 12, 15) breakdown is similar to that described in the cortical parenchyma, though rather less extensive. Most of the degradation is limited to the wall along which fracture will eventually pass and is not so obvious in the walls at right angles to the plane of fracture. This sort of polarized breakdown of the wall would be expected if two rows of adjacent cells were producing the wall degrading enzymes. The highest concentration of enzyme would be found in the common wall. In the xylem transfer cells breakdown is not limited to the middle lamella but involves considerable dispersion and breakdown of the primary wall and secondary wall ingrowths (Fig. 8). In all these cells the cytoplasm seems normal (Figs. 8, 12, 15) and similar to that described in other cells of the separation layer. Ruptured cells are more frequent in the stele particularly near the xylem vessels. It is thought that this damage results from the sudden release of tensions which build up in the xylem prior to fracture. The separation of the phloem is interesting since cell wall swelling and breakdown not only occur in companion and transfer cells (Fig. 12) but also in the sieve elements (Figs. 11, 12). In the latter the fracture line passes right through the centre of the sieve plate (Fig. 11). This was not expected since it is known that some wall degrading enzymes are synthesised ' de novo' (Lewis and Varner, 1970) and sieve elements appear to have neither the necessary synthetic machinery for the production of these enzymes nor the dictyosomal apparatus which is necessary for their secretion (Cronshaw, 1974; Hall and Sexton, 1974). Wall breakdown in the sieve tubes could be the result of enzyme secretion by neighbouring cells but unfortunately it is difficult to determine if this has taken place. Callose is present around the sieve plate areas both at the fracture surface (Fig. 11) and several cells proximal to it. In view of the rapidity with which callose forms when the phloem is damage (Escherich, 1970) particularly in Impatiens it would seem probable that this is a wound reaction. It was hoped on the basis of the observations reported above to determine which cells in the separation zone were involved in the production of cell wall degrading enzymes. It is possible to state with some conviction that parenchyma cells, xylem transfer cells and the collenchymatous cells in the axil of the leaf are capable of producing these enzymes. This can be deduced because large enough tracts of these cells exist
57
for the possible diffusion of the enzyme from neighbouring tissues to be disregarded. Unfortunately most of the other cell types in the abscission zone are in close proximity to other tissues and thus it is impossible to identify with certainty the secretory cells. The problem is well illustrated by the situation in the epidermis, where two features point to the conclusion that these cells produce their own wall degrading enzymes. Firstly, the epidermis is often more extensively degraded than adjacent tissues and secondly there is a lack of significant gradients of wall breakdown from one side of the epidermis to the other. Neither condition would be expected if enzyme diffusion were responsible, though this explanation can not be totally excluded. Clarification of this issue must await the characterization of cytoplasmic features which allow the identification of secreting cells. I would like to acknowledge the patience and skill of Mr. G. Jamieson in helping prepare and section this material. My thanks are also due to Mr. T. Forrest, Mr. J. Goodall and Dr. A.W. Robards for advice concerning both transmission and scanning electron microscopy, and to Professor H. Meidner for help with the preparation of the manuscript.
References Abeles, F.B. : Ethylene in plant biology. London-New York: Academic Press 1973 Bornman, C.H. : Some ultrastructural aspects of abscission in Coleus and Gossypiurn. S. Afr. J. Sci. 63, 325 330 (1967a) Bornman, C.H.: The relationship between tylosis and abscission in cotton (Gossypium hirsutum L.). S. Aft. J. Agri. Sci. 10, 143154 (1967b) Cohen, D., Atsmon, D. : Measurments of water potential and hormone transport associated with growth of cucumber hypocotyls. In: Proceedings 7th International Conference of Plant Growth Substances. pp. 23-30. Ed. : Carr, D.J. Berlin-Heidelberg-New York: Springer 1972 Cronshaw, J. : Phloem differentiation and development. In : Dynamic aspects of plant ultrastructure, pp. 391413. Ed.: Robards, A.W. London: McGraw Hill 1974 Davenport, T.I., Marinos, N.G. : Cell separation in isolated abscission zones. Aust. J. Biol. Sci. 24, 709 715 (1971) Escherich, W. : Biochemistry and fine structure of phloem in relation to transport. Ann. Rev. Plant Physiol. 21, 193-214 (1970) Gawadi, A.G., Avery, G.S. : Leaf abscission and the so called "abscission layer". Amer. J. Bot. 37, 172-180 (1950) Gunning, B.E.S., Pate, J.S., Green, L.W. : Transfer cells in the vascular system of stems. Taxonomy, association with nodes and structure. Protoplasma (Wien)71, 147-171 (1970) Hall, J.L., Sexton, R.: Fine structure and cytochemistry of the abscission zone of Phaseolus leaves. II. Localization of peroxidase and acid phosphatase in separation zone cells. Ann. Bot. 38, 855 858 (1974) Jensen, T.E., Valdovinos, J.G. : Fine structure of abscission zones. I. Abscission zones in the pedicels of tobacco and tomato flowers. Planta (Berl.) 77, 298 318 (1967) Leopold, A.C.: The mechanism of foliar abscission. Symp. Soc. exp. Biol. 21,507 516 (1967) Lewis, L.N., Varner, J.E. : Synthesis of cellulase during abscission
58 of Phaseolus vulgaris leaf explants. Plant Physiol. 46, 194-199 (1970) Lloyd, F.E. : Injury and abscission in Impatiens sultani. 6th Annual report of the Quebec Society for the Protection of Plants from Insect and Fungal Diseases. 72 79 (1914a) Lloyd, F.E.: Abscission. Ottawa Nat. 28, 61 75 (1914b) Morre, D.J.: Cell wall dissolution and enzyme secretion during leaf abscission. Plant Physiol. 43, 1545-1559 (1968) Pate, J.S., Gunning, B.E.S., Milliken, F.F.: Function of transfer cells in nodal regions of stems particularly in relation to nutrition in young seedlings. Protoplasma (Wien) 71,313 334 (1970) Reynolds; E.S. : The use of lead citrate at high pH as an electron opaque stain in electron microscopy. J. Cell Biol. 17, 208~12 (1963) Sexton, R., Hall, J.L.: Fine structure and cytochemistry of the abscission zone cells of Phaseolus leaves. I. Ultrastructural changes during abscission. Ann. Bot. 38, 849-854 (1974) Scott, F.M., Miller, L.W., Webster, B.D., Leopold, A.C. : Structural
R. Sexton : Ultrastructural Observations on Foliar Abscission changes during bean leaf abscission. Amer. J. Bot. 54, 730-734 (1967) Valdovinos, J.G., Jensen, T.E. : Fine structure of abscission zones. II. Cell wall changes in abscising pedicels of tobacco and tomato flowers. Planta (Berl.) 83, 295-302 (1968) Webster, B.D. : Anatomical aspects of abscission. Plant Physiol. 43, 1512-1544 (1968) Webster, B.D. : Anatomical and histochemical changes in leaf abscission. In: Shedding of plant parts, pp. 45 83. Ed.: Kozlowski, T.T. New York-London : Academic Press 1973 a Webster, B.D. : Ultrastructural studies of abscission in Phaseolus: Ethylene effects on Cell walls. Amer. J. Bot. 60, 436447 (1973b) Wright, M., Osborne, D.J. : Abscission in Phaseolus vulgaris. The positional differentiation of ethylene induced expansion growth of specialized cells. Planta (Bed.) 120, 163 171 (1974)
Received 2 July; accepted 8 September 1975
