Planta

Planta 143, 161-179 (1978)

9 by Springer-Verlag 1978

Evidence for Initiation of Microtubules in Discrete Regions of the Cell Cortex in AzoUa Root-tip Cells, and an Hypothesis on the Development of Cortical Arrays of Microtubules B.E.S. Gunning, A.R. Hardham, and J.E. Hughes Department of Developmental Biology, Research School of Biological Sciences, Australian National University, P.O. Box 475, Canberra, A.C.T. 2601, Australia

Abstract. Complexes of microtubules, vesicles, and (to varying degrees) dense matrix material around the microtubules were seen along the edges of cells in root apices of A z o l l a p i n n a t a R.Br. (viewing the cells as polyhedra with faces, vertices and edges). They are best developed after cytokinesis has been completed, when the daughter cells are reinstating their interphase arrays of microtubules. They are not confined to edges made by the junction of new cell plates with parental walls, but occur also along older edges, Similar matrices and vesicles are seen amongst phragmoplast microtubules and where pre-prophase bands intersect the edges of cells. It is suggested that the complexes participate in the development of cortical arrays of microtubules. The observations are combined with others, made on pre-prophase bands and on the substructure of cortical arrays lying against the faces of cells, to develop an hypothesis on the development of cortical microtubules, summarised below: Microtubules are nucleated along the edges of cells, at first growing in unspecified orientations and then becoming bridged to the plasma membrane. Parallelism of microtubules in the arrays arises by inter-tubule cross-bridging. Lengths of microtubule are released from, or break off, the nucleating centres and are moved out onto the face of the cell by intertubule and tubule-membrane sliding, thus accounting for the presence there of short tubules with randomly placed terminations. The nucleating zones along cell edges might have vectorial properties, and thus be able to control the orientation of the microtubules on the different faces of the cell. Also, localised activation could generate localised arrays, especially pre-prophase bands, in specified sites and planes. Two possible reasons for the spatial restriction of nucleation to cell edges are considered. One is that the geometry of an edge is itself important; the other is that along most cell edges there is a persistent specialised zone, inherited at cytokinesis by the

daughter cells when the cell plate bisects the former pre-prophase-band zone.

Key words: A z o l l a - Microtubules - Microtubuleorganising centre - Pre-prophase band - Root (microtubules).

Introduction Microtubules were recognised as components of higher plant cells 15 years ago (Ledbetter and Porter, 1963). It was soon apparent that they might be important controlling agents in plant morphogenesis. Their orientation in the cell cortex in interphase was seen to mirror that of the cellulose microfibrils currently being deposited in the wall, a congruence now known to be widespread, and thought to be indicative of spatial interactions between microtubules and cellulose-synthetase complexes (see review by Hepler and Palevitz, 1974). Wall microfibrils influence the reaction of the wall to turgor forces and hence govern the shapes of cells. It follows that investigations of the control of cortical microtubule orientation are central to many problems of morphogenesis in plants. There are now numerous descriptions of spatial and temporal changes in microtubule abundance and orientation during interphase, and as the pre-prophase band, the spindle, and the phragmoplast, develop and break down in sequence, but little has been learned about the underlying regulatory systems that must exist in order that the sites, extent, and directionality of polymerisation of tubulin can be controlled in the plant cell. The concepts of microtubule-organising centres (MTOCs) and nucleating templates have been applied to great effect in organisms other than higher plants (Pickett-Heaps, 1969; Tucker, 1977) and a variety

0032-093 5/78/0143/0161/$03.80

B.E.S. Gunning et al. : Initiation of Cortical Microtubules in Azolla Root Cells

162 B,

: X,

40-44 i

O.

35-39

32 33

Fig. 1A-D. Diagrams illustrating positions and planes of section of subsequent micrographs of Azolla pinnata root cells. A Apical cell with three proximal faces labelled 1,2, 4; 1,2,3; and 4, 2, 3. Face 1,2, 3 is the most recent. Face 4, 5, 6 was formed at the previous division and face 7, 8, 9 two divisions previously. Part of face 1,2,3 is shown at a cell plate stage in Figures 3, 4 and after cytokinesis in Figures 5, 6; in both cases faces 5,6,7 and 4,5,6 also appear in the mierographs (dark rectangle represents field of view). B Junction of 1st tangential wall with acroscopic transverse inter-merophyte wall (Figs. 7-10), and junction of 1st tangential wall and sextant wall, sectioned slightly obliquely (Figs. 13 16). Shaded rectangles represent sections. C Diagram of the stele of an A.pinnata root. Black rectangle represents plane of section for Figures 40-44, at the edge of a xylem element (X). D Merophyte with tangential longitudinal pre-prophase bands in two sextant cells, sectioned as shown by the rectangle (Figs. 30-31). Part of the neighbouring merophyte in the helix is shown at the left, with another back-toback set of pre-prophase bands, as seen in Figures 35-39. The elevation to the right indicates another tangential longitudinal preprophase band, sectioned in the transverse plane of the root in Figures 32 33. Other labels: l=first tangential wall; 2=endodermis-pericyclewall; 3 = pericycle-inner stele wall

of in-vitro experiments has validated the physical reality of these foci for m i c r o t u b u l e i n i t i a t i o n (for a recent review, see Snyder a n d M c I n t o s h , 1976). As yet n o evidence has been f o r t h c o m i n g that m i c r o t u b u l e s in the cortex of h i g h e r - p l a n t cells derive f r o m localised centres. Indeed, t r a c k i n g of m i c r o t u b u l e s by serial

sectioning of flat or gently c u r v i n g faces of cells has shown that in these parts of the cell the cortical arrays consist of short, o v e r l a p p i n g c o m p o n e n t m i c r o t u b u les, a n d that a p a r t f r o m some clustering a n d linear a r r a n g e m e n t s , the m i c r o t u b u l e t e r m i n a t i o n s lie at rand o m ( H a r d h a m a n d G u n n i n g , 1977, 1978). A model of d e v e l o p m e n t based on self-propagation of the arrays by insertion of new m i c r o t u b u l e s alongside pre-existing ones was considered, b u t the p r o b l e m s of i n i t i a t i o n a n d c o n t r o l of directionality remained. I n the present work a t t e n t i o n is focused on the edges of cells (viewing the cells as p o l y h e d r a with faces, edges, a n d vertices), where tracking of microtubules by serial sectioning is impractical. Evidence is presented which d e m o n s t r a t e s the t r a n s i t o r y existence of specialised microtubule-vesicle complexes a l o n g cell edges at p a r t i c u l a r phases of the cell cycle. It is suggested that these structures represent microt u b u l e - n u c l e a t i n g sites. A c o n t i n u u m of events t h r o u g h the cell cycle can be envisaged, p r o v i d i n g a new c o n c e p t u a l f r a m e w o r k for e x p e r i m e n t a t i o n on the c o n t r o l of cortical arrays of microtubules. As in o u r w o r k o n p r e - p r o p h a s e b a n d s ( G u n n i n g et al., 1978 b) the i n t e r p r e t a t i o n s given here stem from a detailed knowledge of the sites of future cell divisions a n d of the relative ages of completed cell walls in the root tips of the water fern Azolla ( G u n n i n g et al., 1978a). A. pinnata was used t h r o u g h o u t , a n d the m e t h o d s of specimen p r e p a r a t i o n were as e m p l o y e d previously. Serial sectioning was f o u n d to be essential when trying to decide whether some of the structures to be described are m e a n i n g f u l or mere f o r t u i t o u s aggregations. A c c o r d i n g l y the illustrations

Fig. 2. Early cell plate with the central region (bracket) containing continuous plasma membrane, the remainder consisting of coated (small arrows) and non-coated vesicles, and numerous phragmoplast microtubules, some of which terminate in dense matrix material at the level of the ceil plate (arrow). Many of the vesicles closely resemble those seen amongst mierotubules in later micrographs. • 45,000. Scale marker = 1 gm Figs. 3 and 4. Periphery of phragmoplast in apical cell, oriented as shown in Figure 1A. In both sections (adjacent serials) note the vesicles (arrowheads) amongst the phragmoplast microtubules, and the peripheral microtubules appressed closely to the parental plasma membrane (arrows). Note also the abrupt bend in the cell plate (large arrowhead), suggestive of forces that had pulled the edge of the plate towards a site of fusion that was slightly out of alignment with respect to its remaining major part. Both • 34,000. Scale marker = 1 ~tm Figs. 5 and 6. The same wall junctions as in Figures 3, 4 but at

an older stage after some cell enlargement in the new merophyte. Clusters of microtubules are present where the section has cut through the cell edges (brackets), but microtubules also occur along the merophyte wall that is vertical in the micrographs. Section Nos. 1 and 3 in a ribbon, both x 34,000. Scale marker=l gm

B.E.S. Gunning et al. : Initiation of Cortical Microtubules in Azolla Root Cells

163

164

B.E.S. Gunning et al. : Initiation of Cortical Microtubnles in Azolla Root Cells

include several examples demonstrating repeated representation of specialised structures in successive sections.

Results Phragmoplast and Parental Wall In the Azolla root all of the metaphase plates that have been seen were sited and oriented at least approximately correctly. The marked anaphase reorientation exhibited by stomatal guard mother-cells (Palevitz and Hepler, 1974) has not been observed. The phragmoplasts grow centrifugally (Fig. 2) and contain microtubules, growing expanses of new plasma tembrane, and clouds of coated and noncoated vesicles. In later stages of development the phragmoplast microtubules are restricted to the peripheral zone (Figs. 3, 4), m a n y small vesicles lying amongst the microtubules, near where they terminate at the level of the cell plate. The vesicles are still present when the cell plate has grown sufficiently to have fused with the parental wall, so their role may be other than to contribute m e m b r a n e to the growing partition. M a n y of the microtubules at the extreme edge of the phragmoplast angle laterally towards the parental plasma membrane. Some pass along the inner face of the plasma m e m b r a n e away f r o m the site of fusion of cell plate and parental wall for a considerable distance (Figs. 3, 4; also see Fig. 1A for orientation). Pre-prophase bands in the apical cell are shown in Figures 8 and 9 of Gunning et al. (1978b). They have been m a p p e d in both transverse and longitudinal planes of section. So too have pre-prophase bands for other categories of division, and it seems to be general that cell plates fuse with the parental walls along the midline of the zones that were occupied earlier in the cell cycle by the pre-prophase bands. It also seems general that even where the cell plate is curved (as in the apical cell, Figs. 3, 4), it obeys Sachs' (1878) generalisation by joining the parental wall at right angles.

Post-cytokinesis As detailed in Gunning et al. (1978a), each successive merophyte along any one 120 ~ sector of the Azolla root was formed 3 cell cycles of the apical cell earlier than its acroscopic neighbour. The age of the first, i.e. most apical, representative of a given type of wall, whether transverse or longitudinal, is therefore equivalent to between 1 and 3 cycles of the apical cell. Amongst these first representatives some can be found

that are very recent products of completed mitoses. One criterion by which the age of a new cell can be categorised is the state of its cortical microtubules. The youngest have no, or incomplete, interphase cortical arrays. Other features of this category are described below. The next category consists of those cells possessing interphase arrays of predominantly transverse microtubules on the longitudinal walls and of microtubules showing much less parallelism on the transverse wails. If the cell in question is not about to differentiate terminally it will prepare for its next mitosis, and the third age category consists of cells with pre-prophase bands. As described in Gunning et al. (1978b), cells that contain interphase microtubules as well as those in the pre-prophase band, are taken to be younger than cells with preprophase-band microtubules together with microtubules lying between the band and the nucleus. Another criterion for categorising the stage of development of the first (most-acroscopic) representative of a particular type of cell is whether or not its precursor cell in the neighbouring merophyte in the apical direction along the root has reached its pre-prophase-band stage. The structures described below are best developed in cells that have recently divided, containing incomplete interphase arrays, and usually having neighbouring precursor cells that have not yet developed pre-prophase bands. Figures 5.-29 illustrate early stages of development (as defined above) of representatives of the three major categories of cell wall. Figures 5 and 6 deal with the transverse inter-merophyte walls that are laid down when the apical cell divides; Figures 7-20 are concerned with some of the longitudinal walls that are laid down in the formative zone; Figures 2l 29 show transverse walls formed during proliferative divisions in cell files.

Figs. 7-12. Intersection of the first representative of the first tangential wall (vertical in the micrographs, marked with large arrowhead) and the transverse inter-merophyte walls (horizontal) in merophyte No. 6 of an Azolla pinnata root. Figures %10 show the acroscopic end (see also Fig. 1B) and Figs. 11 and 12 the basiscopic end of the first tangential wall. Clusters of microtubule profiles (small arrows) are present where the two walls meet. In some sections vesicles accompany the microtubules (small arrowheads). Figures 7-9 are adjacent sections; Figure 10 is 16sections removed from Figure9; Figure 11 and 12 are separated by 5 sections. All x 38,000. Scale marker = 1 gm

Figs. 13-16. Adjacent sections at the junction of the first tangential wall (vertical), basiscopic transverse inter-merophyte wall (horizontal), and the first representative of the radial longitudinal sextant wall (grazing section, large arrowheads) in meropbyte No. 10 of an Azolla pinnata root (see also Fig. 1B). Microtubule-vesicle-matrix complexesare marked (brackets). All x 38,000. Scale marker = 1 gm in Figure 7

B.E.S. Gunning et al. : Initiation of Cortical Microtubules in Azolla Root Cells

165

166

B.E.S. G u n n i n g et al. : Initiation of Cortical Microtubules in Azolla Root Cells

In every case clusters of microtubule profiles, closely associated with vesicles 50-60 nm in diameter, are found in the angle of the recently formed wall and the parental wall. Often the clusters are partially or wholly embedded (except for emergent microtubules) in dense, amorphous matrix material that adds to the difficulty of resolving the structural details. It is by no means clear that the "vesicles" are bounded by a membrane. Some of them are angular (Figs. 19, 20, 31-33) and "particles" may in due course turn out to be a more correct term. The dense matrix and the vesicles are not present in every section; sometimes the microtubule clusters occur alone, and sometimes they are accompanied by vesicles, but no dense matrix. Figures 5 and 6 are two serial longitudinal sections of an apical cell (see Fig. 1A). Microtubule clusters are present in both daughter cells, i.e. in the apical cell and in the new merophyte, beside the line where the new wall fused with the old. In this case partial digestion of the Epoxy resin from an adjacent 1-ginthick section using alkaline ethanol (Lane and Europa, 1965), followed by staining with aniline blue and fluorescence microscopy, showed that the new cell wall was in the stage at which a callose-type reaction is given (Fulcher et al., 1976). The microtubules are not confined to the angle between the new wall and the old, but they are not clustered except at the wall junction. Four serial sections (the first 3 adjacent, the fourth 16 sections removed) of the acroscopic end of the first representative of the first tangential wall in a recent merophyte are shown in Figures 7-10. Figure 1B (left-hand side) gives the orientation of the plane of section. Figures 11 and 12 are separated from one another by 5 sections, and show the basiscopic end of the same wall. Microtubules and vesicles lie in the angles, on both sides of the new wall, but are not obvious in every section. Most examples do, however, contain microtubule profiles, taking up a range of orientations. No such structures were seen at the equivalent wall junctions in older merophytes. Figures 13-16 depict a later stage in the formative division sequence, where the sextant wall (product of "division 3 " in the numbering system of Gunning et al., 1978a) has just been laid down. As shown diagrammatically in Figure 1B, the micrographs show the intersection of the new sextant wall with an older stage of the first tangential wall than that shown in Figures 7-12. The plane of section is longitudinal, but slightly oblique, so that the cytoplasm on both sides of the line of intersection appears. Complexes of microtubules, vesicles, and dense matrix are traceable through the sequence of four sections, two corn-

plexes above the grazing section of sextant wall, and at least two below it. The wall shown in grazing section in Figures 17 and 18 was identified by reference to other walls in a long sequence of serial sections. It is the first representative, in the 18th merophyte from the apical cell, of the wall that separates the pericycle layer from the endodermal layer. Merophyte 15 (one merophyte closer to the apical cell in that sector of the root) contained a pre-prophase band for this wall. The wall that it is joining (right-hand side of Figs. 17 and 18) is the longitudinal inter-merophyte wall. The plane of section (approximately tangential-longitudinal) can be envisaged from Figure 1D, left-hand side, where pre-prophase bands for the division are depicted in highly diagrammatic form on both sides of an intermerophyte junction. Two adjacent sections are shown, and the complementary fit of the areas of cell wall may be used to show conclusively that some microtubule-vesicle-matrix complexes are on one side and some on the other side of the new wall. All are in the angle with the parental wall, and some are very close to the plasma membrane, if not attached to it. Microtubules fanning out from the complexes are clearly seen. Further details from the same sequence of sections (but not the sections used for Figs. 17 and 18) appear in Figures 19 and 20. By following the sequence of sections it became evident that the complexes were absent (a) from the face of the new wall, except near the angle with the parental wall, and (b) from the face of the parental wall, except near the angle with the new wall. They were much less conspicuous, but not completely lacking, at the same wall junction in the next-older merophyte (No. 21). The features described here for meroFigs. 17 and 18. Two adjacent sections of the junction of a radial longitudinal inter-merophyte wall (vertical, right-hand side of each micrograph) and the first representative of a tangential longitudinal wall between pericycle and endodermal layers (wail in grazing section) in merophyte No. 18 of an Azolla pinnata root. The portions of wall in the grazing sections fit together (e.g. dashed line). The microtubule-vesicle-matric complexes (brackets and stars) are both above and below the new wall. Those indicated by the star and the bracket in Figure 17 lie in one daughter cell, as does the cluster (bracket) at the lowest bracket in Figure 18 ; those above the dashed line in Figure 18 lie in the other daughter cell. Microtubules radiate away from certain complexes sited very close to the plasma membrane (e.g. stars), some of them (arrowheads) curving towards the predominant approximately transverse alignment. The plasmodesmata are not in rows, nor can any predominating orientation of microfibrils in the new wall be detected. Both x 35,000. Scale marker = 1 g m Figs, 19 and 20. Details of complexes from sections close to those used for Figures 17 and 18. Arrowheads indicate "vesicles " with distinctly angular outlines. Both x 56,000. Scale markers =0.1 g m

B.E.S. Gunning et al.: Initiation of Cortical Microtubules in Azolla Root Cells

167

168

B.E.S. Gunning et al. : Initiation of Cortical Microtubules in Azolla Root Cells

p h y t e s 15, 18, a n d 21 were also seen on the o t h e r side o f the r o o t in m e r o p h y t e s 16, 19, a n d 22. F i g u r e s 21-29 p r e s e n t c o m p a r a b l e o b s e r v a t i o n s on r e c e n t l y - f o r m e d transverse walls. The m o n t a g e o f Figure 21 r e p r e s e n t s an o u t e r cortex file in a m e r o p h y t e very s o o n after c o m p l e t i o n o f the s e c o n d r o u n d o f transverse division. The angles o f the two m o s t recent t r a n s v e r s e walls ( b u t n o t the o l d e r ones) with the side walls p r o v i d e f u r t h e r e x a m p l e s o f clusters o f mic r o t u b u l e profiles, s h o w n in m o r e detail in F i g u r e s 22 a n d 23. F i g u r e s 25 a n d 26 are serial to F i g u r e s 27 a n d 28, a n d s h o w d i a m e t r i c a l l y o p p o s e d corners m a d e b y the first transverse wall with the side walls in an e n d o d e r m a l file. F i g u r e 29 is similar, b u t refers to a transverse division in the pericycle. In all o f these cases the s a m e cell files in the next-younger m e r o p h y t e s h a d n o t u n d e r g o n e these divisions. The next-older m e r o p h y t e s s h o w e d no signs o f m i c r o t u bule c o m p l e x e s in the e q u i v a l e n t cell corners, b u t the cells d i d have n o r m a l , t r a n s v e r s e l y o r i e n t e d c o r t i c a l a r r a y s , u n l i k e those d e p i c t e d in the m i c r o g r a p h s . The a b o v e sets o f figures illustrate a n o t h e r phen o m e n o n , the generality o f which r e m a i n s to be established. It c o n c e r n s cell edges in the d a u g h t e r cells, o t h e r t h a n edges c r e a t e d b y the new cell wall. A s the first e x a m p l e , the serial sections o f which Figures 5 a n d 6 are r e p r e s e n t a t i v e s h o w m i c r o t u b u l e clusters in the angle m a d e by the m o s t recent a n d a p a r e n t a l wall, b u t also where two o l d e r walls (laid d o w n one a n d three cell cycles earlier t h a n the m o s t recent one) m e e t (see Fig. 1A). A s the s e c o n d e x a m ple, F i g u r e 24 illustrates the first r e p r e s e n t a t i v e o f a first t r a n s v e r s e division in an o u t e r c o r t e x initial cell. M i c r o t u b u l e c o m p l e x e s lie n o t o n l y close to the j u n c t i o n o f the new wall with the side walls, b u t also at the j u n c t i o n o f two l o n g i t u d i n a l walls, one o f which is seen in the m i c r o g r a p h in grazing section. Details a p p e a r in a series o f inserts, selected so as to highlight several features. First, the m i c r o t u b u l e s are o r i e n t e d in a v a r i e t y of directions, b u t often they are p a r a l l e l to one a n o t h e r within small groups. Second, within these g r o u p s p e r i o d i c c r o s s - b r i d g i n g is to be seen. Third, dense m a t e r i a l lies n e a r l y all a l o n g the p l a s m a m e m b r a n e , with m i c r o t u b u l e s f a n n i n g o u t f r o m this vicinity. T w o o t h e r general p o i n t s m a y be a d d e d before leaving the t o p i c o f m i c r o t u b u l e clusters a l o n g the edges o f recently d i v i d e d cells. E m p h a s i s has been p l a c e d (but n o t exclusively) on edges m a d e with a recently f o r m e d wall. W h a t has n o t been stressed is t h a t m i c r o t u b u l e - v e s i c l e - m a t r i x c o m p l e x e s have n o t been seen except in cell corners, w h e t h e r in y o u n g o r old cells. A l s o , the variety of o r i e n t a t i o n s o f the m i c r o t u b u l e s in a n d n e a r cell edges has received c o m ment. V a r i o u s stages o f d e v e l o p m e n t o f i n t e r p h a s e

cortical a r r a y a c c o m p a n y the c o n f i g u r a t i o n s in the edges. T h u s F i g u r e s 5 a n d 6 show n o r m a l transversely o r i e n t e d tubules a l o n g the face o f a wall in the new m e r o p h y t e , in a d d i t i o n to the clusters in the corners. A t the o t h e r e x t r e m e are wall faces t h a t lack m i c r o tubules, all m i c r o t u b u l e s being c o n f i n e d to the edges. In general, t r a n s v e r s e l y o r i e n t e d walls do n o t d e v e l o p a r r a y s o f p a r a l l e l m i c r o t u b u l e s , a n d y o u n g stages o f d e v e l o p m e n t o f the a r r a y s a l o n g such walls are n o t r e a d i l y d i s t i n g u i s h e d f r o m the m a t u r e state. A l o n g Figs. 21-29.

Azollapinnata root cells that have recently completed

transverse divisions Figs. 21-23. Figure 21 shows a file of outer cortex cells, sectioned longitudinally, with transverse inter-merophyte walls at top and bottom. The central transverse wall was laid down earlier than the two walls (large arrowheads) for the second round of transverse division, which were not present in the next merophyte in the apical direction. Two areas (brackets) are enlarged in Figure 22 and 23 to show the presence of microtubules (small arrows) and vesicles (arrowheads), in places embedded in a dense amorphous or finely fibrillar matrix (bracket). Figure21, x8,400; Figure22 and 23; x47,000. Scale markers1 lam Fig. 24. Part of an outer cortex ceil, with intermerophyte transverse walls at top and bottom, sectioned longitudinally, with a recent transverse wall laid down following the first round of divison in the initial cell (large arrowhead). The section is cut close to the intersection of two longitudinal walls, one seen in edge profile (left-hand edge, backing onto an intercellular space), and the other in grazing face view, showing plasmodesmata (many in short transverse rows). The lettered areas are shown at higher magnification in the insets A-D. The cell corner is occupied by many microtubules, in a variety of orientations. Many of them pass into dense material on or near the plasma membrane, especially where the section passes closer to the extreme edge near the bottom of the micrograph, in which area can be seen microtubule configurations as in Figures 17 and 18, fanning out from surface-located loci (bracket). Coated vesicles and their polygonal surface pattern are common, but smooth vesicles (arrowheads) are also present. Groups of parallel microtubules pass over the grazing section of the longitudinal wall, and within many of these groups periodic intertubule bridging is conspicuous (especially insets C and D). Figure 24, x 19,000; A, B, x36,400 (scale markers=l p,m); C, D, x 55,000 (scale markers=0.1 gin) Figs. 25 and 26. Diametrically opposite edges of a recently completed first transverse wall in the endodermis cell file, showing microtubule profiles (arrows) in the angles with the longitudinal walls. No microtubules were present elsewhere along the longitudinal walls in these sections, x 47,000. Scale marker =0.1 ~tm Figs. 27 and 28. As Figures 25 and 26. The bracket indicates a

cluster of microtubule profiles embedded in a dense matrix. x 47,000. Scale marker=0.1 gm Fig. 29. Angle between recently completed first transverse wall and longitudinal wall (left-hand side) in a pericycle file. Microtubules and dense matrix material are present (bracket). As in the endodermal examples in Figures25 28, no such structures were seen in older representatives of the same cell angles, x47,000. Scale marker = 0.1 gm

B.E.S. G u n n i n g et al. : Initiation of Cortical Microtubules in Azolla Root Cells

169

170

B.E.S. Gunning et al. : Initiation of Cortical Microtubules in Azolla Root Cells

longitudinal walls, however, where predominantly transverse tubules ultimately develop ( H a r d h a m and Gunning, 1977, 1978), it seems to be very c o m m o n for there to be a transitory stage with longitudinally 'directed microtubules, or a mixture of orientations in the plane of the wall. Face views of these walls show scattered plasmodesmata at this stage. Later stages, as seen in older merophytes with enlarged cells, generally have transverse microtubules and some of the plasmodesmata occur in short, transverse rows (Fig. 1 of Gunning et al., 1978b, shows the change in distribution of plasmodesmata. Examples of scattered plasmodesmata in new walls are seen here in Figs. 13-18, and of short rows in Fig. 24).

Pre-prophase Bands Figures 30-39 illustrate pre-prophase bands, focusing attention, as above, on the presence of vesicles and (less consistently) electron dense material at the edges of cells. Figures 4, 6, 8 and 11 in Gunning et al. (1978b) reinforce the pictures presented here. Back-to-back pre-prophase bands are shown in Figures 30 33. The spaces between and around the microtubules where the section intersects the cell edges are crowded with small vesicles. Serial sectioning showed that one of the bands in Figures 30 and 31 was relatively young, co-existing with interphase cortical microtubules; the other was relatively old, with numerous microtubules and bunches of microtubules between the band and the nucleus. Both stages of development had vesicles at the cell edges, not just in the sections shown, but also in other parts of the sequence. The dark-staining material around and between the microtubules in these sites is sometimes conspicuous (especially in Figs. 4 and 6 of Gunning et al., 1978b), and sometimes less so. It is not yet clear whether this variability reflects a developmental change, or inconsistent fixation and/or staining. In Figure 34 (for a low-magnification view, see Fig. 1 of Gunning et al., 1978b) dense sheaths surround short lengths of the microtubules, possibly at their terminations. Figures 32 and 33 also show dense material associated with m a n y of the microtubules. Figure 35 is a composite picture made by combining micrographs of two different sections from a ribbon of sections. Two pre-prophase bands are present in a staggered back-to-back arrangement as diagrammed in Figure 1D (left-hand side). As shown in more detail in Figures 36-39, all of the cell corners contain dense complexes of microtubules and vesicles. In one of the two cells the pre-prophase band was followed by serial sectioning to the two remaining corners, where the band passed from the transverse

walls onto the outer tangential face of the cell: every corner contained microtubules, vesicles and dense matrix material. Figures 8 and 9 of Gunning et al. (1978b) show a pre-prophase band in an apical cell. Where the band passes over a face of the cell there are no obvious specialisations (Fig. 9), but where it sweeps round the cell edge near the apex where the three proximal faces meet there is a pronounced dark-staining zone along the cell surface underlying the band of microtubules; there are also a few vesicles amongst the microtubules (Fig. 8). Reference to Figure 1A of the present paper shows that apical-cell pre-prophase bands pass round three edges. Two of these are formed at the previous division and one two divisions earlier. The zone referred to above as being darkly stained represents one of the set of two edges formed one cycle earlier.

Xylem 7'hickenings One aspect of the development of wall thickenings in xylem is relevant in the present context. The young xylem elements of A. pinnata roots are triangular in cross section (Gunning et al., 1978 a), and xylem-wall thickenings first appear at the cell edges. The appearance of the thickenings is preceded by the clustering of cortical microtubules from a previously more evenly dispersed type of array ( H a r d h a m and Gunning, 1978). Figures 4 0 4 4 illustrate part of a young xylem element, sectioned as shown in Figure 1C. A colchicine treatment had removed the majority of the microtubules, and was followed by a 36-h recovery period. It can be seen that in this recovering cell

Figs. 30 and 31. Adjacent sections showing portions of two longitudinal back-to-back, pre-prophase bands that anticipate the pericycle-endodermis wall in merophyte No. 15 of an Azolla pinnata root. The plane of section is indicated in Figure 1D. Serial sections showed that the band to the right gave way to transversely oriented cortical microtubules; the band to the left had a virtual gap in the centre of the wall, and numerous microtubules, some in parallel bunches, lay between the band and the cell nucleus (just out of the micrograph to the left). Accumulation of vesicles is seen at all cell edges-see especially Figure 31. Arrowheads in Figure 31 indicate "vesicles" with angular profiles. Figure 30, x 30,000; Figure 31, x 52,000. Scaie markers= 1 lxm Figs. 32 and 33. Azollapinnataroot. Two serial sections (separated by one section) of longitudinal back-to-back pre-prophase bands that anticipate the pericycle-inner stele division. The plane of section through the cell edges is indicated in Figure ID (right). Some microtubules can be tracked from one micrograph to the other (e.g. dotted line); others probably terminate between the sections (e.g. arrow). Numerous vesicles, some angular (arrowheads), and dense inter-tubular material are present, x 52,000. Scale marker = 1 gm

B.E.S. Gunning et al. : Initiation of Cortical Microtubules in Azolla Root Cells

171

172

B.E.S. Gunning et al. : Initiation of Cortical Microtubules in Azolla Root Cells

the bands of microtubules that precede xylem thickenings are f o r m i n g in the cell edge, and that the tubules are a c c o m p a n i e d by n u m e r o u s vesicles. The serial sections included here, and others in the sequence, show that this material was fixed at a stage in which the tubule-vesicle complexes did n o t extend far f r o m the edge that was being sectioned.

Discussion C o m p a n i o n studies o f the cortical microtubules in Azolla root-tip cells have dealt with reconstructions o f arrays lying against flat or gently curving faces o f cells ( H a r d h a m and Gunning, 1977, 1 9 7 8 ) a n d with the occurrence and localisation of pre-prophase bands ( G u n n i n g et al., 1978b). The present observations c o m p l e m e n t these investigations, utilising a detailed b a c k g r o u n d knowledge o f the Azolla r o o t ( G u n n i n g et al., 1978a) to examine identified regions o f cells at particular stages o f development. The main conclusion is that the edges o f cells can develop transitory specialisations in the f o r m of complex aggregates of microtubules, vesicles, and (to varying degrees) an electron-dense matrix surrounding parts o f the microtubules. Such complexes a p p e a r after cytokinesis and persist in a conspicuous f o r m for a period that c a n n o t yet be defined, but is considerably shorter than the period f r o m cytokinesis to the subsequent pre-prop h a s e - b a n d stage. Vesicles and matrix material also appear where pre-prophase bands pass a r o u n d cell edges, and vesicles occur a m o n g s t microtubules at the earliest stages o f development o f xylem-wall thickenings, again at cell edges. These observations m a y lead to new analyses o f a wide range of m o r p h o g e n e t i c processes in plants, and the following discussion will explore this potential significance.

Microtubule-nucleating Zones

The first step in the analysis is to suggest that the microtubule complexes are s y m p t o m a t i c of microtubule initiation. Electron-dense matrices such as seen in the complexes occur a r o u n d M T O C s (PickettHeaps, 1969, 1974, 1975) and microtubule-nucleatiug templates (Tucker, 1977) in a wide variety of organisms. The vesicles (which are sometimes angular, and m a y in fact be particles), are approximately the same size as 67 + 4 - n m - d i a m e t e r virus-like particles f o u n d in peri-centriolar regions o f ovary cells o f the Chinese hamster ( G o u l d and Borisy, 1977); material isolated f r o m these regions, and containing the particles, is capable o f nucleating microtubules in vitro.

The c o m p o n e n t s that a c c o m p a n y the microtubule clusters in the edges o f Azolla root-tip cells are therefore similar to those seen in other situations where microtubule arrays are k n o w n to originate. The disposition of the microtubules themselves adds further evidence. Some o f the cells illustrated here had microtubule complexes at cell edges but nowhere else, and had been fixed shortly after completion o f cytokinesis and before the interphase arrays had developed. Microtubule terminations are difficult to detect unequivocally where there are changes in orientation, as in cell corners, but it is clear that they are present in at least some o f the complexes (e.g. Figs. 11, 13, 17, 18, 22, 32, 33), unless there are a b r u p t right-angle bends in the microtubules in question at the level of the plasma membrane. The experiment on recovery f r o m colchicine treatment (Figs. 4 0 4 4 ) is again consistent with microtubule initiation at edges. Microtubule-vesicle-matrix complexes at cell edges have been illustrated here in planes o f section where the new wall is in edge-on profile as well as where it is seen in face view. The latter examples (Figs. 12-20) show why the complexes m a y be absent or inconspicuous when the f o r m e r plane o f section

Fig. 34. Part of a pre-prophase band for the tangential division that separates the pericycle from the innermost layer of the stele of an Azolla pinnata root (for low magnification view see Figure 1 of Gunning et al., 1978b). The wall in grazing view is a sextant and the lower part of the micrograph shows the edge between that sextant and transverse inter-merophyte wall (along base of picture). Dense matrices around microtubules, and numerous vesicles, are present in the edge of the cell, but are absent or less frequent further away from the edge. x 42,000. Scale marker = 1 gm Figs. 35-39. Composite micrograph made from views of two sections in a ribbon of serial sections of an Azolla pinnata root, showing the inter-merophyte longitudinal wall where it zig-zags between merophytes 15, 16, 18, and 19 of a root apex. The region below the dotted line derives from section No. 8 and that above it from section 13 of the ribbon. These two views were selected because the plane of section was slightly oblique and no single section passed through all parts of the pre-prophase bands that were present in merophytes 15 and 16. These bands were for the tangential perieycle-endodermis wall (see Fig. 1B, left-hand side, for diagrammatic representation). Merophytes 18 and 19 contain the walls (large arrowheads) that correspond to the bands seen in 15 and 16; both walls had microtubule-vesicle-matrix complexes where they met the longitudinal inter-merophyte wall (one example in Figure 36, bracket). Details are enlarged in Figures 36-39 (long arrows) ; Figure 36 is from section number 15 ; Figures 37 and 39 from section number 13; Figure 38 from section number 6 of the ribbon. White arrowheads in Figure 35 point to cell edges that are formed by bending a wall; they back onto edges that are formed by wall fusions. Figures 36-39 show the presence of many vesicles amongst pre-prophase band microtubules in all cell edges. Both pre-prophase bands appear to be losing bundles of microtubules (arrows in Figures 35 and 39) towards their respective nuclei, which lie out of the picture to right and left of Figure 35. Figure 35, x 13,200; Figures 36-39, • 35,000. Scale markers= 1 ~tm

B.E.S. Gunning et al. : Initiation of Cortical Microtubules in Azolla Root Cells

173

174

B.E.S. Gunning et aI. : Initiation of Cortical Microtubules in Azolla Root Cells

Figs. 40-44. Azollapinnataroot. Longitudinal adjacent serial sections of a young xylem element oriented as shown in Figure 1C. Fig. 40 is closest to the edge of the cell and the next four micrographs penetrate further into it. Only one band of microtubules out of a set that were spaced out along the cell preparatory to formation of xylem wall thickenings is included. Note vesicles amongst the tubules in Figures 40 43. By the level of Figure 44, further away from the cell edge, few microtubules and vesicles are present. From a root-tip fixed 36 h after a 2-h treatment with 5 x 10 3 M colchicine. All • 49,000. Scale marker= 1 gm

is used. They are of intermittent occurrence along the edge of the cell, so that sections m a y pass between them.

Development of Interphase Arrays The role of microtubule bridges. The microtubules that emerge from the complexes are not all as close to the plasma m e m b r a n e as they are in mature interphase arrays (Figs. 5, 9, 10, 23, 25-27). Some micrographs show tubules bending towards the plasma m e m b r a n e as they leave the corner (Figs. 10, 22, 23, 25, 26, 29). It may be conjectured that the assembly of an interphase array is a multi-step process, beginning with nucleation at a complex and then nondirected early growth, followed by attachment to the plasma m e m b r a n e by bridges which anchor the tubules into the plane of the cell cortex. F o r cell walls such as those lying transversely to the root axis this might complete the development of the array, producing non-specifically aligned cortical microtubules in the plane of the wall. F o r the longitudinal walls, where parallelism develops and the array ultimately is predominantly transverse to the root axis, a subsequent assembly step would be needed. Figures 17 and 18 illustrate what can be interpreted as an intermediate stage, in which microtubules fanning out from the microtubule complexes bend smoothly towards the predominant transverse orientation that is displayed further away from the edge of the cell, where the tubules are often in small, parallel bundles. Thus, parallelism could be achieved by cross-bridging

microtubules that have already been anchored in the plane of the cell surface. Cross-bridges to the plasma m e m b r a n e and between adjacent tubules have often been reported (see review by Hepler and Palevitz, 1974). Periodic intertubule bridges, sometimes in " h e r r i n g b o n e " patterns, occur in Azolla (Figures 24C, D and H a r d h a m and Gunning, 1978), as do microtubule-plasma membrane bridges, but before using these observations to add to the model of development of cortical arrays, it is necessary to refer to the detailed ultrastructure of the arrays, as established by tracking microtubules over the face of the cell by means of serial sectioning ( H a r d h a m and Gunning, 1977, 1978).

Microtubules against the faces of cells. It was shown that microtubule terminations do exist. The variation in microtubule length is marked, but the average is 2 4 gin, that is, a b o u t one eighth of the cell circumference. The probability of finding microtubules in the form of " h o o p s around the cell" is extremely low. With the exception of some clustering and occasional lines of terminations, the microtubules end in r a n d o m positions. The microtubules are not all strictly parallel. Some individuals, usually lying deeper in the cytoplasm than the majority, pass at large angles to the others. Some leave one bundle and angle across to become parallel to those in another bundle. Electrondense matrices were rarely seen at microtubule terminations, and in discussing the development of the arrays it was concluded that the flat faces of cells provided no compelling evidence for the existence of discrete MTOCs. Propagation of arrays by inser-

B.E.S. Gunning et al. : Initiation of Cortical Microtubules in Azolla Root Cells

tion of new microtubules alongside pre-existing ones was thought to be possible.

Developmental relationships between edges and faces. It is necessary to try to reconcile the above observations with the present results. If microtubules can develop individually at locations scattered over the faces of cells (a process which would be difficult to detect and impossible to rule out using the present methods), then the microtubule complexes at cell edges lack a raison d'atre. On the other hand it is entirely possible that some, if not all, of the microtubules that come to lie on the faces of the cells originate at the edges. Electron micrographs are static, but the microtubules, being short, might well be mobile (Hardham and Gunning, 1978). Thus acquisition of membrane-tubule and inter-tubule cross bridges near the microtubule complexes could not only bend the tubules towards a predominant orientation (Fig. 18), but also eventually bring about their release from the complexes, freeing them to slide out into the array on the face of the cell. Figure 45 is a diagrammatic representation showing the developing (A) and the mature (B) conditions. There is good evidence for

A

B

Fig. 45A and B. Diagram to illustrate the emergence of microtubules from complexes lying at intervals along a cell edge (vertical, in A), as compared with the more mature situation B. Mature edges probably have more microtubule terminations than shown here, and need further examination. The cell is polyhedral, unlike Figure 1 of Hardham and Gunning (1978) where, for purposes of making a mathematical derivation, the cell is shown diagrammatically as having a circular cross-section

175

inter-tubule sliding in other situations (Mclntosh, 1973; Gibbons, 1977; Sale and Satir, 1977), the force generation being mediated by mechanochemical effects seated in the arms and bridges of the microtubules. A form of feedback control can be envisaged if continued nucleation and growth of new microtubules is dependent on the existence of gaps in the array, into which new tubules can migrate. Without appropriate gaps the energetics of tubule migration could be so unfavourable that the removal of the product of the complexes would not occur, leading in turn to a diminution of microtubule growth and nucleation as the array matures. It has been noted here that the microtubule-vesicle-matrix complexes are less conspicuous after a certain period of time has elapsed since cytokinesis. It seems reasonable that the interpolation of new microtubules during cell expansion would be much less easily detectable than the initial regeneration of an array after mitosis. Virtually nothing is known about the polymerisation-depolymerisation reactions of tubulin from plant material. Accordingly, little can be said about other control mechanisms that might exist, or about the roles of the vesicles and the dense matrix. Possibilities range from provision of tubulin or of low-molecularweight cofactors such as guanosine triphosphate or high-molecular-weight microtubule-associated proteins, to regulation of the local ionic environment, as for instance by sequestering calcium ions. The above matters are at present virtually inaccessible to experimentation, but one facet of the system which can readily be examined concerns the establishment of arrays of microtubules with particular orientations. The major question is, do the postulated initiating zones along the edge angles of cells have vectorial properties? Hardham and Gunning (1978) found that the position of C-shaped microtubule terminations was patterned, the great majority of them pointing in the same direction in any given sequence of serial sections. Directionality of microtubule development emerged as a possibility. Most of the present evidence relates to newly formed edges of polyhedral cells, but some indication of microtubule complexes along older edges was also detected (Figs. 5, 6, 24). It was also noted that when microtubule arrays were forming along new cell walls in the longitudinal plane of the root there was a consistent tendency for development of a transitory nonoriented or longitudinal array (Figs. 7-16) prior to formation of the familiar transverse set of microtubules. Other examples of developmental changes in the predominant orientation of cortical microtubules are known (e.g. Robards and Kidwai, 1972; Palevitz and Hepler, 1976).

176

B.E.S. Gunninget al. : Initiationof Cortical Microtubulesin Azolla Root Cells

A potentially useful concept is that arrays with particular orientations might develop by the selective activation of microtubule-generating complexes at appropriate cell edges, each having specific and controllable vectorial properties. Longitudinal and transverse arrays could originate by growth of microtubules approximately at right angles to transverse and longitudinal edges, respectively. Interactions that maximise the extent of overlap of microtubules emanating from different longitudinal edges in the same cell would tend to generate and preserve a transverse orientation despite concomitant cell elongation. A trend towards satisfying all possible potential cross-bridging sites would be a suitable type of interaction, provided that the arrays form a continuum around the cell edges or else are anchored at each edge. It may not be coincidental that the average length of microtubules, as determined by serial sectioning, is about one eighth of a cell circumference (Hardham and Gunning, 1978), which can be approximated to half of a cell face, i.e. just enough for microtubules from opposite corners of a cell to meet and interact. With regard to other types of array, the possible origin of those on transverse walls has already been mentioned. Cross-bridging to generate parallelism, but without maximising overlap as envisaged for transverse systems, could produce arrays oriented in directions other than transverse. It would be particularly instructive to examine edges of cells that develop alternating crossed lamellate types of wall. In all of these situations there is scope for immediate investigation.

Development of Localised Arrays, Especially Pre-prophase Bands The above model of the development of interphase arrays of microtubules can be extended in relation to the development of spatially restricted systems, such as pre-prophase bands and those overlying xylem thickenings, by invoking localised activation of edge regions. The present evidence does not include images of the earliest stages of pre-prophase-band formation. Figure 34 could be such a stage, but extensive serial sectioning will be needed to make definitive identifications. However, in both early and late pre-prophase bands, vesicles resembling those already described for microtubule complexes are present where the bands pass around cell edges. At least some of the microtubules terminate at these edges. The types of influence, both internal, genetically controlled, and external, based on positional information and cellular interactions, that might specify sites of pre-prophase band formation are discussed elsewhere (Gunning et al., 1978b). Zones along the edges of cells, potentially

capable of initiating microtubules, could be the sensors for whatever gradients or other influences do in fact determine the polarity of the cell and the plane in which it is to divide. A new pre-prophase band could develop by microtubules interconnecting a set of co-ordinates consisting of locally activated 2-3-~tmwide portions of the edge zones. As with transverse interphase arrays, microtubule interactions and crossbridging would be necessary to establish and to maintain such interconnections. A similar sequence of events could underlie the early events in the development of xylem-wall thickening (Figs. 4044).

The Developmental Origin of Specialised Edge Zones Of equal importance to the potential functional significance of microtubule-organising complexes along cell edges is the possible cause of segregation of their activity to the edges of the cell. The edges of polyhedra are strategically located for influencing both of the faces that meet to make an edge of the cell, but at a deeper level of analysis, two possibilities are worth examining.

Geometrical considerations. The angle of curvature that can be assumed by microtubules appears to be limited. Their molecular architecture is not amenable to bending them around corners. Therefore, where mature arrays approach and sweep around the edges of the plant cell, they must leave the outermost zone of the cell cortex (e.g. Figs. 10, 23, 25, 27, 30-33, 35 39). One possible reason why edges become specialised rests upon this feature. For instance, it can be argued that the vesicles (particles) which occur along with microtubules in the complexes are produced somewhere else in the cell, and are propelled along microtubules in the manner envisaged for Golgi vesicles (Northcote, 1969), accumulating only where there is a suitable stagnant area, i.e. in the gap between the curving microtubules and the extreme edge. Certainly it is beyond doubt that the microtubule arrays in the edges contain far more vesicles than those against the faces. Acquisition of vesicles in this way would then confer specialised, but unknown, properties on the edges. This model is unsatisfactory in a number of respects, not least because it depends upon the prior existence of cortical microtubules, whereas the tubule-vesicle complexes can in fact have precedence, especially after cytokinesis. At this time, however, the complexes could originate by the microtubule organising material of the phragmoplast continuing the centrifugal migration that it displays during cell plate growth (Bajer and Bajer, 1972), until

B.E.S. Gunning et al. : Initiation of Cortical Microtubules in Azolla Root Cells it detaches into the edges that have been newly formed by fusion of the cell plate with the parental walls.

Bisection of former pre-prophase band sites. It cannot be excluded that the geometry of edges is in itself significant, but an alternative holds theoretical and conceptual advantages. One consistent feature of preprophase bands that has not hitherto been considered to be functionally significant is their width. It has been pointed out here, and by Gunning et al. (1978b), that in Azolla root tips, new cell walls are fused with parental walls precisely enough that the line of fusion can be shown to be along the approximate mid-line of the zone that had been occupied earlier in the cell cycle by the pre-prophase band. The former preprophase-band zone is approximately bisected and each daughter cell receives half. Each half-band passes round a daughter cell along new edges that circumscribe the new cell wall. With some exceptions, the older cell walls were also laid down by similar procedures, and they too are circumscribed by edges that carry half former pre-prophase band zones. The alternative hypothesis therefore is that localised special properties are conferred upon the cell cortex and plasma membrane at the time of formation of the pre-prophase band, and that, despite the loss of the pre-prophase microtubules, the former preprophase band zone retains an ability to participate in initiating (1) new interphase arrays after cytokinesis, and (2) new pre-prophase bands when the daughter cells prepare for their own mitotic cycle. The process would be self-perpetuating and would represent a form of cytoplasmic inheritance. A specialised zone of cell cortex is handed on to the next cell generation, which is capable of utilising the new and the earlier-formed zones to repeat the cycle. The emphasis that is placed upon multicellular tissues, where pre-prophase bands exist, underlines the need to examine situations where there are no pre-prophase bands, and where, if the hypothesis is correct, other mechanisms for generating microtubule arrays must exist. The only novel feature of the hypothesis is in the nature of the postulated microtubule-nucleating region and the possibility of localised spatial control. Bisection of the former pre-prophase band zone is analogous to fission of other types of MTOC such as the various types of spindle-pole body (reviewed by Hepler and Palevitz, 1974). Likewise, formation of new edge zones by generating new pre-prophase bands is analogous to de-novo formation of MTOCs, such as the blepharoplast of Marsilea (Hepler, 1976). Temporal control of MTOC activity is well known (see Pickett-Heaps, 1969, and Hepler and Palevitz, 1974).

177

Three features of the hypothesis need brief discussion: the possible mechanisms of ensuring equal partitioning and inheritance of the former pre-prophase band zone; the nature of the specialisation within this zone; and the type of signals that might activate it. Figures 3 and 4 show that, at late stages of development, the peripheral microtubules of the phragmoplast can lie closely appressed to the parental plasma membrane, spanning the zone where it is known that the pre-prophase band used to lie. Whilst other, coarse, orientation mechanisms must undoubtedly exist, the final stages of cell-plate alignment might be regulated by mechanochemical interactions between these peripheral microtubules and a specialised zone of cell cortex left behind after the disappearance of the pre-prophase band microtubules. The evidence of Figures 3 and 4 indicates that if the tubules that point in opposite directions from the extreme edge of the cell plate were to recognise and bridge onto the specialised zone, and then generate opposite forces by "sliding" (in opposite directions) over it, the edge of the plate would come to rest when the forces are equal and opposite, that is, when the plate is in the midline of the former pre-prophase band zone. Both sets of microtubules would by then have maximised their contact with the specialised zone, with the result that the edge of the cell plate bisects it. The pronounced kinks that are sometimes seen near the edges of cell plates (Figs. 3, 4) are consistent with fine adjustments to orientation in that vicinity. The suggestion also provides a mechanistic basis for the generalisation (Sachs, 1978) that the new wall following a cell division meets the old ones at an angle of 90 ~, even if the cell plate is curved. The hypothesis requires that a zone of cell cortex becomes specialised during that stage of the cell cycle when it is overlain by the pre-prophase band, and that the specialisation is retained. Nothing is known about such processes, but it may be relevant that tubulin can be present in membranes: Estridge (1977) provides evidence that tubulin-like polypeptides can be exposed on a membrane surface, and tubulin can be extracted from certain membranes and assembled into microtubules (Bhattacharyya and Wolff, 1976). One possible mechanism of final cell plate alignment therefore is that the peripheral phragmoplast microtubules interact with tubulin or tubulin-binding molecules that were set in place in the membrane or cortex during the pre-prophase band stage of the cell cycle.

Edges formed by bending the faces of cells. Only further, detailed observations will show whether microtubule initiation in cell corners happens because the geometry of a corner is favourable (as in the first

178

B.E.S. Gunning et al. : Initiation of Cortical Microtubules in Azolla Root Cells

hypothesis), or because corners carry a f o r m o f M T O C (as in the second hypothesis), or indeed for other, quite different reasons. Meanwhile the present observations include one example which at first sight is inconsistent with the " h a l f - p r e - p r o p h a s e - b a n d " model. Figures 35-39 show pre-prophase bands lying b a c k - t o - b a c k on b o t h sides o f a zig-zag longitudinal inter-merophyte wall (see Fig. 1B, left-hand side). Microtubules and vesicles are concentrated at every cell edge, but not every edge was f o r m e d by fusion o f a cell plate with parental walls. T w o o f those in Figure 35 were f o r m e d simply by bending the face o f a wall (left-hand side o f Fig. 37 and right-hand side o f Fig. 38), so that no half-pre-prophase bands were involved. However, backing o n t o the edges f o r m e d by bending are edges o f the neighbouring cells, with genuine f o r m e r p r e - p r o p h a s e - b a n d zones associated with microtubules and vesicles (right-hand side o f Fig. 37 and left-hand side o f Fig. 38). It is not to be ruled out that some f o r m o f intercellular, trans-wall signalling is operating, analogous to that postulated to occur in xylem, where thickenings often develop b a c k - t o - b a c k in adjacent cells (see discussion in T o r r e y et al., 1971). Further w o r k is needed to clear up these uncertainties, t h o u g h t it is difficult in meristematic tissue to find the type o f cell edge that arises by the bending o f a flat wall, and still m o r e difficult to find examples at the relevant stage o f the cell cycle and in the requisite plane o f section.

Conclusion The crucial observation in the present w o r k is that some, if not most, cortical microtubules m a y originate at restricted zones along the edges o f polyhedral cells. I f it is allowed that cortical microtubules function during interphase in relation to development o f the wall, and during pre-prophase as part o f the apparatus determining the site and plane o f division, then any putative organising regions for such microtubules assume special significance as potential sites o f m o r phogenetic control in the cell. Recognition o f these sites has only been b r o u g h t a b o u t after intensive investigation of an object possessed of unusually precise developmental p r o g r a m m e s , and it remains to be seen whether a similar analysis can be applied to angiosperm tissues. Perhaps the main p r o b l e m will be to find situations in which the plane of division is sufficiently predictable to show whether the former prep r o p h a s e b a n d zone is bisected at cytokinesis. Elements o f this discussion have been taken far b e y o n d the immediate interpretation o f the observations, in order to generate testable hypotheses concerning the regulation o f cortical microtubule

arrays in plant cells. It is n o w 15 years since these microtubules were discovered (Ledbetter and Porter, 1963), and in that time operational analyses of the systems that must be present, controlling microtubule f o r m a t i o n and disposition, and hence functions, have lagged far behind straightforward p h e n o m e n o l o g i c a l descriptions. It is h o p e d that the micrographs and suggestions m a d e here provide a new impetus and conceptual f r a m e w o r k for investigating unsolved problems o f plant-cell morphogenesis in which microtubules are involved,

References Bajer, A.S., Mol6-Bajer, J.: Spindle dynamics and chromosome movements. Int. Rev. Cytol., Suppl. No. 3. New York-London: Academic Press 1972 Bhattacharyya, B., Wolff, J. : Polymerisation of membrane tubulin. Nature 264, 576-577 (1976) Estridge, M.: Polypeptides similar to the e and /? subunits of tubulin are exposed on the neuronal surface. Nature 268, 60-63 (1977) Fulcher, R.G., McCully, M.E., Setterfield, G., Sutherland, J.: /?1,3-Glucans may be associated with cell plate formation during cytokinesis. Can. J. Bot. 54, 539-542 (1976) Gibbons, I.R.: Structure and function of flagellar microtubules. In: International cell biology 1976 1977, pp. 348-357, Brinkley, B.R., Porter, K.R., eds. New York: Rockefeller University Press 1977 Gould, R.R., Borisy, G.G. : The pericentriolar material in chinese hamster ovary cells nucleates microtubule formation. J. Cell Biol. 73, 601