Pollen Tube Development in Petunia Hybrida Following Compatible and Incompatible Intraspecific Matings

J. Cell Sd. 47, 365-383 (1981) 365 Printed in Great Britain © Company of Biologists Limited ig8i

POLLEN TUBE DEVELOPMENT IN PETUNIA HYBRIDA FOLLOWING COMPATIBLE AND INCOMPATIBLE INTRASPECIFIC MATINGS

MARIA HERRERO* AND H. G. DICKINSON Department of Botany, Plant Science Laboratories, University of Reading, Whiteknights, Reading RG6 2AS, U.K.

SUMMARY Pollen tubes formed following compatible and incompatible intraspecific matings in Petunia have been examined with light and electron microscopes. Compatible and incompatible tubes develop in an identical fashion on the stigma but, on entry into the top 1 mm of the stylar transmitting tissue changes occur both to the cytology of the tubes and their rates of growth. The early cytological changes are common to tubes of both compatibilities but, although both types of tube accelerate on entry into the style, incompatible tubes grow more slowly than compatible. Cytological differences became apparent between compatible and incompatible tubes following a short period of growth in the style, the latter possessing thicker cell walls and a cytoplasm packed with both organelles and reserves. Incompatible tubes subsequently burst or simply cease growth and die. The characteristic image afforded by this cytoplasm resembles that of burst or dead compatible tubes, except in that proportions of the cell components may differ. These data are discussed in terms of current models proposed to explain pollen tube growth and the operation of the self-incompatibility response in Petunia.

INTRODUCTION As in many plants with gametophytic control of pollen compatibility with respect to the style, pollen tubes of Petunia formed following a self-mating fail to reach the ovary. Growing more slowly than tubes from a cross-pollination they eventually cease development in a region some two-thirds of the way down the style. Since much is known of the physiology of and structures involved in pollen tube growth, it is not unreasonable to expect that the differences in development between tubes of different compatibilities might easily be explained by changes in the ultrastructure and physiology of the tube cytoplasm. This has not proved to be the case. Striking ultrastructural differences between compatible and incompatible tubes were reported as early as 1966 by van der Pluijm & Linskens who described the walls of incompatible tubes as being far thicker than those derived from compatible crosses. In a comprehensive investigation into the cytology of pollen tube growth in Lyco- persicum (de Nettancourt et al. 1973), the arrest of incompatible pollen tubes was reported to result partly from the cessation of protein synthesis, itself caused by ribosome-coated endoplasmic reticulum forming into concentric whorls (de Nettan- court et al. 1974), and from the binding of an 'incompatibility protein' to wall-precursor containing bodies in the tube cytoplasm. Such results were not fully • Present address: INIA, Apartado 202, Zaragoza, Spain. 366 M. Herrero and H. G. Dickinson confirmed by work on Oenothera (Dickinson & Lawson, 19756) where growth of incompatible tubes appeared to be accompanied in the first instance by changes in the carbohydrate metabolism of the tube, with resultant changes in the cell wall. Data from this investigation did, however, reveal that considerable modification of the cytoplasmic components involved in cellulose biosynthesis is required to permit the very rapid rate of growth characteristic of these tube cells. Because the elongating pollen tubes are embedded in stylar tissue, physiological investigations have not been easy. Work by van der Donk (1974) has, however, shown differences between tubes of differing compatibilities to become evident very soon after pollination. So much so that Linskens (1975) has proposed the 'recognition' stage of the self-incompatibility response to operate in the stigma. Despite the considerable effort expended in these and other investigations, we are still in ignorance both as to the point in tube metabolism at which the self-incompatibility mechanism has its primary effect and, indeed, of the means by which pollen tube elongation is arrested. We report here an investigation into these events in Petunia, the results of which are discussed in terms of the accompanying changes that occur in the tissue of the pistil (Herrero & Dickinson, 1979).

MATERIALS AND METHODS Details of the plant material used in this study, together with the methods of preparation of tissue for light and electron microscopy are set out in Herrero & Dickinson (1979). The methods for staining resin-embedded material with Coomassie Brilliant Blue (CBB) are described by Fisher (1968).

RESULTS Pollen germination and tube growth at the stigma surface Within 30 min of pollination most grains germinate (Fig. 2) and develop a tube at least 30 jim in length. Conspicuous changes overcome the pollen during early stages in germination, firstly a fibrogranular matrix (Figs. 1, 3) that encased the mature grain

Fig. 1. Colpal region of uninucleate microspore (p), still retained in the anther. Fibrils (/) are visible in the sculptured exine, facing the loculus (/). x 8620. Fig. 2. Scanning electron micrograph of mature pollen grain (g) on the stigma (s). Note the absence of any deposition on the exine surface. The globules (/) are presumably derived from the stigmatic fluid. A germinal pore (p) is also visible, x 5640. Fig. 3. Low-power transmission micrograph of material shown in Fig. 6. The exine (e) is free from fibrils. The vegetative nucleus (n) and a germinal pore (/>) are also shown, x 2050. Fig. 4. Light-microscopic preparation of pollen grain germinating on the stigma, reacted to reveal acid phosphatase. Higher levels of the enzyme, which appears to be in packets, are contained in the tube (f) than the grain (g). x 730. Pig. 5. Tip of pollen tube growing on the stigmatic surface. Note the spherical vesicles (arrows). The cell wall (w) is fibrous and disorganized, and the ground cytoplasm very electron-opaque, x 16300. Pollen tube development in Petunia 367 368 M. Herrero and H. G. Dickinson in the anther becomes no longer detectable and, secondly, marked changes occur in the pollen cytoplasm. Here, instead of the electron-opaque protoplast characteristic of the grain prior to dispersal, a more electron-lucent cytoplasm is seen, containing increased amounts of rough endoplasmic reticulum, dictyosomes, and associated vesicles. The plasma membrane, previously smooth and featureless, now appears active, and is formed into numerous small projections associated with vesicles. Other components of this germinating protoplast are droplets of unsaturated lipid, diffuse fibrous masses, mitochondria, the vegetative nucleus and the generative cell. Although all the microbodies present in the cytoplasm appear identical under the electron microscope, light-microscopic histochemical tests indicate 2 classes of microbody to be present, one containing acid phosphatase (Fig. 4) and the other a peroxidase. The peroxidase-containing bodies survive for only some 60 min after germination, whereas those containing acid phosphatase may be found throughout development of the pollen tube. Clearly the cytoplasm of the young pollen tube on the stigma surface is contin- uous with that of the pollen grain, but since differences occur in the composition of the cytoplasm between the various regions of the tube, these areas are probably more helpfully described individually. At the tube tip, the cytoplasm is comparatively electron-opaque and the protoplast surface very irregular, such that the plasma membrane is undetectable (Fig. 5). At this surface are numerous vesicles, similar to those seen near dictyosomes deeper in the cytoplasm. The wall of the pollen tube tip consists solely of loosely woven fibrils showing no particular organization (Fig. 5). Back from the tip, the plasma membrane of the tube becomes better defined and the fibrils of the wall organized in particular directions (Fig. 6). Even at this point, the presence of callose is neither indicated by electron microscopy nor by cytochemical tests. While only small mitochondria may occasionally be discerned in the cytoplasm of the tube tip, here these organelles are far more frequent. Conspicuous also are large aggregates of fibrous material, similar to that of the tube wall, and masses of folded membrane reminiscent of the myelin figures of animal cells. Numerous vesicles still populate this cytoplasm, many of which appear associated either with the plasma membrane or with the fibrous masses (Fig. 6). Examination of the pollen tube wall close to the grain reveals it to consist of an outer, well organized fibrillar layer and an inner, electron-lucent layer adjacent to the plasma membrane (Fig. 7). Conspicuous in this cytoplasm are large numbers of

Fig. 6. Transverse section of pollen tube, as depicted in Fig. 5, but further back from the tip. Large fibrous masses are present (/), as are vesicles (arrows) and mitochondria (w). «>, wall, x 21400. Fig. 7. As Fig. 5, further back along the tube towards the grain, showing the presence of a thin electron-lucent layer (p) between the plasma membrane (arrows) and the fibrous wall (w). x 15200. Fig. 8. Transverse section of pollen tube, as depicted in Figs. 6, 7. Here the cytoplasm is more electron-lucent, containing elements of rough endoplasmic reticulum (e), mitochondria (m), and a paramural body (p). The lipidic stigmatic fluid (s) is also visible, x 14170. Pollen tube development in Petunia 369

. 8 370 M. Herrero and H. G. Dickinson membranous cisternae, normally identified with paramural bodies (Fig. 8). This cytoplasm is also rich in rough endoplasmic reticulum, mitochondria, plastids and microbodies (Figs. 7, 8). Light-microscopic tests indicate callose to be present in this wall (Fig. 10) and bodies containing acid phosphatase to populate the protoplast. By the time the tube has grown to a length of some 200 /tm the vegetative nucleus and the generative cell are also to be found in this region (Fig. 9). A large vacuole has developed in the pollen grain by this stage, sometimes extending into the tube itself. The cytoplasm remaining in the grain is rich in rough endoplasmic reticulum. It is striking that no differences exist in germination, cytoplasmic content and wall development between compatible and incompatible grains growing on the stigmatic surface.

Pollen tube growth in the transmitting tissue Between 2 and 3 h after pollination the first compatible and incompatible pollen tubes pass through the 'neck' of the style and into the transmitting tissue. Details of the organization of these cells are given elsewhere (Herrero & Dickinson, 1979). Coincident with the arrival of the pollen tubes a proportion of transmitting tissue cells burst, and the tubes continue their growth in a matrix formed partly from the residue of these ruptured cells (Fig. 11). Pollen tube growth in the style differs significantly from that on the stigma. This difference is reflected in the mode of wall synthesis, the composition of the tube cytoplasm and, of course, the further alterations to tube development that result from operation of the self-incompatibility system.

The cytology of compatible pollen tubes The changes to the compatible pollen tube as it enters the transmitting tissue are few. The small dictyosome vesicles characteristic of growth on the stigma appear to be replaced by larger vesicles, some 0-25 /tm in diameter. Despite the fact that dictyosomes are less frequent in the cytoplasm of the tube tip, these organelles seem

Fig. 9. Transverse section through a pollen tube growing in the stigmatic fluid (/). Both the vegetative nucleus (y) and the generative cell (g) are visible, x 10400. Fig. 10. Pollen grains developing in the stigma stained to reveal the presence of callose in the tube walls (arrows). Light micrograph, x 320. Fig. 11. Pollen tube (t) growing through a stylar matrix formed of cytoplasm (c) derived from ruptured cells. Note adjacent intact stylar cells (J). x 26000. Fig. 12. Transverse section through a pollen tube growing through the style, taken near to the tip. Vesicles (arrows) containing organized fibrillar content may be seen merging with the plasma membrane (p) and discharging their content into the wall (tv). x 18500. Fig. 13. Detail of pollen tube growing in the stylar tissue showing the protoplast periphery some distance from the tube tip. Here fibrillar (/) and electron-lucent (/) layers of the wall may be discerned, as may paramural membranous bodies (j>). x 38600. Fig. 14. Low-power light micrograph of a compatible cross-pollination stained to show the presence of callose (c) in the 'plugs' formed in the pollen tubes following the passage of the generative cell, x 90. Pollen tube development in Petunia

14 372 M. Herrero and H. G. Dickinson the most likely source of the larger vesicles, which, like the small vesicles, may be seen merging with the plasma membrane of the tube tip (Fig. 12). Unlike the smaller vesicles, however, these vesicles observed during growth in the style regularly contain an organized fibrillar content. The general electron opacity of the tube cytoplasm is also greater when in the style, and much of this electron-opacity can be shown, using enzymic digestion, to be sensitive to protease. Polyribosomes are also present in far greater numbers in these cells. In the tube cytoplasm further away from the tip, more paramural bodies may be seen (Fig. 13), and light-microscopic histochemistry reveals bodies containing esterase to be present in this region. Further back in the tube, callose commences to be formed at precisely the same distance from the tip as when the tube was growing on the stigma. The organelles associated with this synthesis are also similar to those found in earlier development of the tube. Once the generative cell has passed down the tube, some 6 h after pollination (Fig. 14), callosic plugs are formed. On the stigma, transverse sections of these' empty' tubes retain their original rounded appearance, while in the style they present an infolded or collapsed aspect (Fig. 15). The depth of the callosic wall varies during growth of the pollen tube, for, as it enters the lower style it becomes very thin, such that it is detectable in only few preparations. However, as the tube approaches the ovary (Fig. 16) callose once more becomes evident. In addition, profiles indicating the presence of burst pollen tubes are frequently conspicuous in this lower region of the style. The necrotic cytoplasm of these ruptured tubes regularly contains spherical bodies, apparently composed of radiating fibrils, and measuring about 0-25 /im in diameter. Enzymic digestion reveals the electron-opaque matrix of these spheroids to be insensitive to protease while the surrounding cytoplasm is digested by this enzyme. In many ways this discharged cytoplasm resembles that of some incompatible pollen tubes when examined either in the stigma or style 24 h after pollination (Fig. 17). Over the course of tube growth down the style, the generative cell continues to follow closely behind the tip. The generative cell (Fig. 18) which measures about 3 /tm long and 2 /im wide, is bounded by an electron-lucent space, contains mitochondria and a complex assembly of microtubules. No plastids appear to be present. As the generative cell passes through a particular length of tube, a callosic plug is formed behind it.

Fig. 15. Transverse section through a pollen tube (/) growing in the style, but following passage of the generative cell. Note how the wall is infolded (arrows), x 14400. Fig. 16. Light-microscope preparation of a compatible pollen tube entering the ovary (0), stained to reveal callose. Note the callosic plug (p) and the pollen tube (t) itself, x 460. Fig. 17. Necrotic cytoplasm from burst compatible pollen tubes at the base of the stylar transmitting tissue. Note the presence of the spherical fibrillar bodies (arrows) characteristic of degenerate pollen tube cytoplasm, x 6700. Fig. 18. Transverse section through a pollen tube showing the generative cell (g). The nucleus (n) is clearly visible, and small wefts of microtubules (arrows) may be discerned in the tube cytoplasm. The electron-lucent wall (w) between generative and vegetative cells is also evident, x 40050. Pollen tube development in Petunia

18 374 M. Herrero and H. G. Dickinson

Differences between the growth in the style of compatible and incompatible pollen tubes There appear to be no totally novel structures in the developing incompatible pollen tube, although the balance between cytoplasmic components may sometimes be altered considerably. The first indications of the incompatibility response may be seen soon after the pollen tube penetrates the transmitting tissue. These include an increased rate of apposition of the outer fibrillar wall at the tube tip (Fig. 19) and a densely ' packed' aspect of the cytoplasm, especially in the region of the generative cell (Fig. 20). Examination of incompatible tubes growing deep in the transmitting tissue reveals them to possess much thicker walls (Fig. 21). Both the fibrillar and the callosic electron-lucent walls are far more heavily deposited, and in regions distant from the tip, small crescent-shaped fibrillar inclusions may frequently be seen apparently immobilized inside the electron-lucent component of the wall (Fig. 21). The general cytoplasm of the tubes, both at the tips and elsewhere, is strikingly electron-opaque, but this electron opacity may be considerably reduced by treatment with protease (Fig. 22). The packing of cytoplasmic components is also increased, with greater numbers of almost all the inclusions previously described. Staining with Coomassie Brilliant Blue and treatment with PAS shows increased levels of protein and carbohydrate in these tubes (Figs. 23, 24). Within 24 h of growth in the transmitting tissue, a conspicuous change overcomes the cytoplasm of the majority of incompatible tubes. A marked increase in electron opacity occurs followed by a loss of identifiable form by most cytoplasmic inclusions (Fig. 25). Coincident with these events is the appearance of inclusions, between 0-2 and 0-3/tm in diameter, apparently composed of radiating fibrils (Fig. 26). As with the similar bodies seen in burst compatible tubes these inclusions are insensitive to protease, in sharp contrast to the cytoplasm investing them which is rapidly digested by this enzyme (Fig. 27). Also evident in the electron micrographs of this material are profiles indicating that a proportion of these incompatible tubes rupture (Fig. 28).

Fig. 19. Tangential section through the tip of an incompatible pollen tube (t) growing in the style. Note the increased deposition of fibrillar precursors (arrows) into the tube wall (w). The increased electron-opacity of the cytoplasm is also evident, x 17 500. Fig. 20. Sectioned incompatible pollen tube showing thickened fibrillar (/) and electron-lucent (/) walls, the generative cell (g), and electron-opaque cytoplasm, x 15340. Fig. 21. Detail of incompatible pollen tube deep in the transmitting tissue, showing fibrillar wall precursor elements (arrows) apparently embedded in the electron- lucent wall (/)• XII 580. Fig. 22. Material as shown in Fig. 21, but following treatment with protease. The cytoplasm (c) appears to be particularly sensitive to this enzyme, x 7370. Pollen tube development in Petunia 375 376 M. Herrero and H. G. Dickinson

28 Pollen tube development in Petunia 377

DISCUSSION Germination of the pollen grain and early tube growth on the stigma surface The first event to take place following the arrival of the pollen on the stigma is the disappearance of the fibrillar coating of the exine. Although tapetally-synthesized pollen grain coatings have been implicated in mechanisms of self-incompatibility (Dickinson & Lewis, 1973 a, b), this has only been shown in plants where pollen compatibility is sporophytically controlled. In Petunia where compatibility is controlled by the pollen genotype such an involvement would not be expected, a conclusion supported by the work of Fett, Paxton & Dickinson (1976) in Lilium, and Gilissen & Brantjes (1978) on Petunia, where washing of pollen was shown not to affect the self-incompatibility response. Activation of the pollen cytoplasm follows at approximately the same time as the disappearance of the fibrillar coating, with the anticipated increase in numbers of dictyosome vesicles, first reported by Larson (1965), and the increase in rough endoplasmic reticulum and polysome frequency described by Linskens, Schrauwen & Koning (1970). Once it has emerged from the grain, the manner of growth of the pollen tube differs little until it enters the transmitting tissue, and resembles closely the style of growth observed in vitro (Kroh, 1967). The disorganized fibrils at the tip of the tube presumably consist of polymerized and part-polymerized cellulose (O'Kelley & Carr, 1954), deposited from the vesicles present in the subjacent cytoplasm (Sassen, 1964). These vesicles have been shown to contain both cellulose and the enzymes necessary for its synthesis (Engels, 1974). Both the work here reported, and previous studies (Van de Woude, Morre & Bracker, 1971) indicate dictyosomes to be the source of these vesicles. How they are transferred to the growing tip is not clear. Work with inhibitors (Mascarenhas & Lafountain, 1972) has indicated that microtubules are unlikely to be involved in this process, but that cytochalasin-sensitive

Fig. 23 A. Incompatible pollen tubes (arrows) growing in the style reacted with PAS to reveal carbohydrate. Note the dense staining of the tube cytoplasm. Light micrograph, x 1800. B. As Fig. 23 A, but a compatible tube. The lack of staining of the tube cytoplasm (arrows) is evident. Light micrograph, x 1900. Fig. 24A. As Fig. 23 A, but stained with Coomassie Brilliant Blue. The tube contents (arrows) stain strikingly, x 1440. B. As Fig. 23 B, but stained with Coomassie Brilliant Blue. The tubes (arrows) appear to contain little protein, x 1600. Fig. 25. Necrosis in an incompatible pollen tube (t) some 24 h after entering the transmitting tissue. Note the thickened walls (iv), the presence of spherical fibrillar bodies (arrows), and the lack of identifiable organelles in the cytoplasm, x 4670. Fig. 26. Detail of cytoplasm of a tube such as that shown in Fig. 25. The spherical, fibrillar bodies (s) are clearly evident, together with some membrane (arrows), x 40910. Fig. 27. Ruptured incompatible pollen tube (i) in the transmitting tissue following digestion with protease. Note the sensitivity of this necrotic cytoplasm to the enzyme, x 337°- Fig. 28. As Fig. 27, but without enzymic digestion. Most of the features shown in Figs. 25 and 26 are also evident in this cytoplasm (c). x 2710. 378 M. Herrero and H. G. Dickinson elements, presumably microfilaments, organize the movement of these vesicles via cytoplasmic streaming. Since the material deposited at the tube-tip is eventually destined to form part of the cylindrical wall of the tube it must be flexible, a property no doubt conferrred upon it by the loose packing of the cellulosic microfibrils. Once finally deposited in the tube wall, these fibrils are then reinforced by a layer of callose. Little is known of the cytoplasmic structures involved in the synthesis of this glucan; Jensen & Fisher (1970) suggested that dictyosome vesicles were responsible for its deposition, whereas Cresti & van Went (1976) described callose synthesis in elements of' rough' endoplasmic reticulum. Most recently, however, both Cresti, Pacini, Ciampolini & Sarfatti (1977) and Ciampolini & Cresti (1977) have again indicated dictyosome vesicles as being the source of this polymer. Cell fractionation experiments by Helsper, Veerkamp & Sassen (1977) have revealed vesicles from the tube cytoplasm to contain /?-glucan synthetase, and to be capable of synthesizing alkali-insoluble 1-3 glucans. Elsewhere in plants where callose synthesis can be stimulated, for example in the stigmatic papillia of Raphanus (Dickinson & Lewis, 1973 a), the structures involved appear to be irregular-shaped vesicles, sometimes tube-like, emanating from elements of the endoplasmic reticulum. While nothing from the present investigation points directly to the identity of structures involved in callose synthesis, it is interesting that paramural bodies (Marchant & Robards, 1968), composed of irregular tubules and vesicles, are present without exception at the sites of callose deposition. The mechanism by which the generative cell moves into the tube is not evident from results obtained so far. However, since cytoplasmic streaming must of necessity be recirculatory in nature, it could hardly move this cell steadily down the tube. The assembly of microtubules around the generative cell is most striking, and it is not impossible that it is concerned with its mobility (Hoefert, 1971). Sanger & Jackson (1971), however, consider these microtubules to be responsible for maintaining the ellipsoid shape of the cell, rather than its movement, and more recently Cresti, van Went, Willemse & Pacini (1976) have identified fibrous masses in the pollen tube cytoplasm with the motility of the generative cell. The cytology of the stigma appears in no way to be disturbed by the passage of the pollen tube (Herrero & Dickinson, 1979), confirming the conclusion by Konar & Linskens (1966) that this tissue plays a passive role at this stage. What is taken up by the tube tip during this phase of growth is unknown, but it has tacitly been assumed that much of this development is supported by the reserves in the pollen grain (Linskens, 1964; Dickinson & Lawson 1975 a). The significance of the identical development of compatible and incompatible pollen tubes on this tissue is not easily evaluated. This similarity also extends to growth rate, tubes of both compatibilities growing at a rate of 150/im h"1 (Herrero & Dickinson, 1980). Certainly it may be that no factors are present in the stigma to affect the tube according to its compatibility, but it may equally be that the tube, growing with the aid of pollen-held reserves, may not yet be sensitive to the molecules responsible for the arrest of incompatible pollen tubes. Pollen tube development in Petunia 379

The changes in tube growth that occur on entry into the transmitting tissue The alterations to the cytology of the tube-tip that take place in the neck of the style may, in the main, be explained in the difference between the growth rates of the tubes; compatible tubes, for example, increase to a rate of some 520/im h"1 (Herrero & Dickinson, 1980). Certainly the cellulosic tube wall continues to be formed at the tube tip from vesicle-held precursors, but their source now appears to be cisternae of the endoplasmic reticulum. Fibrils often become apparent in these vesicles as they near the walls, and the vesicles themselves often fuse together before reaching the cell wall. In Oenothera there is evidence that this process becomes even further modified, with the phosphorylation stages necessary for the formation of the cellulose being carried out in the vesicles while appressed closely to mitochondria (Dickinson & Law- son, 1975 a). The increased electron-opacity of the cytoplasm presumably results from increased levels of protein in this region of the tube. It is in the neck of the style that a considerable difference in growth rate becomes evident between tubes of differing compatibilities, for incompatible tubes only increase in growth rate by 190 /im h"1 compared with the 370 /im h"1 of compatible tubes (Herrero & Dickinson, 1980). Ascher (1966) has suggested growth on the stigma and in the style to be controlled by 2 different operons, and that incompatible tubes fail to activate a high-velocity operon controlling stylar growth which enables it to utilize stylar materials for its subsequent development. If this is true, incompatible growth may thus be regarded as autotrophic. Such an inference has been supported by work of Rosen & Gawlick (1966) on the structure of tube tips, and also van der Donk (1975), who explained lack of growth of incompatible tubes in terms of an inability to activate the stylar genome, condemning the tube to grow only as long as its reserves lasted. Results from Herrero & Dickinson (1980), and those presented here do not wholly confirm these conclusions. Certainly, incompatible tubes do grow more slowly in the style than compatible ones, but both types accelerate on entry into this tissue. Incom- patible tubes, however, cease growth apparently while still containing plentiful reserves which, of course, may either remain from the pollen grain or be derived from the pistil. Equally, when growing in the transmitting tissue they resemble far more closely in structure and growth rate compatible tubes in the same region of the style, rather than tubes on the stigma. The incompatible tube, therefore, appears to undergo many of the changes experienced by compatible tubes on entry into the transmitting tissue, but suffers in addition some other effect resulting in the reduced growth rate. Incompatible tubes are capable of accumulating labelled precursors from the stigma to levels equivalent to those found in compatible tubes (Kroh, Miki-Hirosige, Rosen & Loewus, 1970), and appear to contain higher levels of detectable reserves, de Nettan- court et al. (1974) propose that protein synthesis is inhibited in incompatible tubes, a conclusion which receives support from Cresti et al. (1977) who showed, that gamma- irradiated compatible pollen tubes resembled in many ways incompatible tubes. On the other hand, our cytochemical data indicate that incompatible tubes contain high levels of proteins but, of course, we know nothing of the character of these polypep- tides. 380 JVf. Herrero and H. G. Dickinson Summarizing the evidence available from this present study, it seems most likely that both compatible and incompatible tubes undergo identical initial changes on entry into the transmitting tissue. These changes must involve an increase in uptake of materials from the pistil'and, in incompatible tubes, one or more of these materials must induce a metabolic imbalance that results in a decrease in tube growth rate. The bursting of many of the compatible tubes at the base of the transmitting tissue is perplexing. The microscopic data indicate that the tube walls are lacking the electron-lucent callosic layers here and it is not impossible that the absence of this rigid wall-layer is responsible for the rupture of the tubes. Why callose deposition should cease in this region is not clear. It is known that the oxygen tension is very low in this part of the style (Linskens& Schrauwen, 1966), and Tiipy (1959) has proposed that callose production is closely linked to respiration. This may be the case, but since cellulose and callose share much the same biosynthetic pathway, it is not easy to see how one carbohydrate should be formed, in the absence of, or at the expense of the other.

The differences in pollen tube growth following compatible and incompatible matings Since the compatibility of pollen tubes appears to become established as soon as they enter the top of the transmitting tissue, this neck region of the style probably plays the most important role in the self-incompatibility system in Petunia. Examina- tion of incompatible tubes in the stylar neck presents a confusing picture, but perhaps the most conspicuous features of such material 24 h after pollination are the small, spherical fibrillar vesicles. These bodies have been considered in most detail by de Nettancourt et al. (1973) who interpret them as resulting from the admixture of 'incompatibility protein' with pollen tube wall precursors, and point out the sirnilarity between these inclusions and the bodies liberated when pollen tubes encounter the synergid cells prior to fertilization (Jensen & Fischer, 1968; Diboll, 1968; van Went, 1970; Vazart, 1971). Dickinson & Lawson (1975&) concluded somewhat differently from work on Oenothera. They pointed out that these structures were.characteristic solely of burst or dead pollen tubes and did not constitute a primary product-of the self-incompatibility response. The chemical constitution of these bodies has yet to be fully elucidated, but Cresti & van Went (1976) conclude that they are fich in Callose, and form the basis of the callosic plugs characteristic of growing pollen tubes'. Since we here present evidence that these fibrillar bodies are also present in burst compatible tubes, the inference must be reinforced that they represent precursors of the cellulosic portion of the pollen tube wall which have, either as a result of tube bursting or necrosis, autopolymerized in situ. Although the fibrillar bodies seem not to be involved in the first events of the incompatibility response, the construction of the pollen tube wall becomes evidently different between tubes of different compatibilities only some 7 h after pollination. That increased amounts of callose were present in incompatible tubes was established as early as 1957 (Linskens & Esser, 1957), and shortly afterwards Tupy (1959) suggested that this did not necessarily represent an increased pace of callose synthesis, but could reflect a decreased rate of tube extension. When structural investigations of Pollen tube development in Petunia 381 these events began, it became clear that the components of the pollen tube wall differed not only in their quantity, but also in their disposition. Schlosser (1961) for example, reported thickening of the tube wall at the tip, while van der Pluijm & Linskens (1966) described thickening of the tube walls themselves, de Nettancourt et al. (1973) suggested that modification of the tube tip resulted in the demise of incompatible tubes, proposing that rupture of these cells followed a disappearance of the callosic wall and an expansion of the outer fibrillar wall into the intercellular spaces of the conducting tissue. Although many of the features described by de Nettancourt et al. (1973) are present in our material, we interpret them differently and suggest, as did Tupy (1959), that all the events observed may be explained in terms of a decreased rate of tube growth. Thus, we believe the thickened walls of the incompatible pollen tube to result from deposition of precursors at a rate equal to that in the compatible tubes which are, of course, growing faster. This continuity of 'normal' metabolism in tubes which are elongating slowly would not only explain the thick walls, but also the apparent continuing synthesis of cellulose once formation of the callose wall has commenced, for the 'change-over point' of wall precursor deposition would be 'focussed' in a far shorter length of tube. Further, most other cell products would also be accumulated at a rate appropriate to a compatible tube-elongation rate, thus explaining the high levels of reserves and other cytoplasmic constituents of the incompatible tubes. No unequivocal data point to the reason behind this slow growth. Rosen (1961) suggested that pollen tubes extend by hydrostatic pressure acting upon a 'weakened' tip. Extending this hypothesis to the situation in Petunia, the slow growth of incompatible tubes must result from increased wall pressure, or a decrease in turgor pressure. Heslop-Harrison (1978) explored the possibility that the tube wall becomes modified by a change in activity of cell-wall synthesizing enzymes, but it is equally pos- sible that incompatible tubes are unable to break down stylar materials into moieties which provide the necessary osmotic pressure to drive the tubes as fast as those resulting from compatible matings.

One of us (M.H.) would thank Reading University and the OECD for financial support. Thanks are also due to the Royal Society for the provision of photomicrographic apparatus, and to the Electron Microscope Unit of the Plant Science Laboratories, University of Reading.

REFERENCES ASCHER, P. D. (1966). A gene action model to explain gametophytic self-incompatibility. Euphytica 15, 179-183. CIAMPOLINI, F. & CRESTI, M. (1977). Main ultrastructural features of Petunia hybrida pollen tube. Caryologia 30 (4), 487-488. CRESTI, M., CIAMPOLINI, F., PACINI, E., SREE RAMULU, K., DEVREUX, M. & LANERI, U. (1977). Ultrastructural aspects of pollen tube growth inhibition after Gamma irradiation of Lyco- persicumperuvianum. Theoret. appl. Genet. 49, 297-303. CRESTI, M., PACINI, E., CIAMPOLINI, F. & SARFATTI, G. (1977). Germination and early tube development 'in vitro' of Lycopersicum peruvianum pollen; ultrastructural features. Planta 136 (3), 239-249. 382 M. Herrero and H. G. Dickinsoti CRESTI, M. & WENT, J. L. VAN (1976). Callose deposition and plug function in Petunia pollen tubes in situ. Planta 133, 35-40. CRESTI, M., WENT, J. L. VAN, WILLEMSE, M. T. M. & PACINI, E. (1976). Fibrous masses and cell and nucleus movement in the pollen tube of Petunia liybrida (Brief communication). Ada bot. need. 25(5), 381-383. DIBOLL, A. G. (1968). Fine structural development of the megagametophyte of Zea mays following fertilization. Am.jf. Bot. 55, 787-806. DICKINSON, H. G. & LAWSON, J. (1975 a). The growth of the pollen tube wall in Oenothera organensis.J. Cell Sci. 18, 519-532. DICKINSON, H. G. & LAWSON, J. (19756). Pollen-tube growth in the stigma of Oneothera organensis following compatible and incompatible intraspecific pollinations. Proc. R. Soc. B 188, 327-344- DICKINSON, H. G. & LEWIS, D. (1973a). Cytochemical and ultrastructural differences between intraspecific compatible and incompatible pollinations in Raphamts. Proc. R. Soc. B 183, 21-38. DICKINSON, H. G. & LEWIS, D. (19736). The formation of the tryphine coating the pollen grains of Raphanus and its properties relating to the self-incompatibility system. Proc. R. Soc. B 184, 149-165. DONK, J. A. W. M. VAN DER (1974). Synthesis of RNA and protein as a function of time and type of pollen tube style interaction in Petunia hybrida L. Molec. gen. Genet. 134, 93-98. DONK, J. A. W. M. VAN DER (1975). Recognition and gene expression during the incompatibility reaction in Petunia hybrida L. Molec. gen. Genet. 141, 305-317. ENGELS, F. M. (1974). Function of golgi vesicles in relation to cell wall synthesis in germinating Petunia pollen. IV. Identification of cellulose in pollen tube walls and golgi vesicles by X-ray diffraction. Acta bot. neerl. 23, 209-215. FETT, W. F., PAXTON, J. D. & DICKINSON, D. B. (1976). Studies in the self-incompatibility response in Lilium longiflontm. Am.jf. Bot. 63, 1104-1108. FISHER, D. B. (1968). Protein staining of ribboned epon sections for light microscopy. Histo- chemie 16, 92-96. GILISSEN, L. J. W. & BRANTJES, N. B. M. (1978). Function of the pollen coat in different stages of the fertilization process. Acta bot. neerl. 27, 205-212. HELSPER, J. P. F. G., VEERKAMP, J. H. & SASSEN, M. M. A. (1977). B-glucan synthetase activity in golgi vesicles of Petunia hybrida. Planta 133, 303-308. HERRERO, M. & DICKINSON, H. G. (1979). Pollen-pistil incompatibility in Petunia hybrida. Changes in the pistil following compatible and incompatible intraspecific crosses, jf. Cell Sci. 36, 1-18. HERRERO, M. & DICKINSON, H. G. (1980). Pollen tube growth following compatible and incompatible intraspecific pollinations in Petunia hybrida. Planta 148, 217-221. HESLOP-HARRISON, J. (1978). Recognition and response in the pollen-stigma interaction. In Cell-Cell Recognition, Symp. Soc. exp. Biol., xxxn, pp. 121-138. Cambridge University Press. HOEFERT, L. L. (1971). Pollen grain and sperm cell ultrastructure in Beta (abstract). In Pollen Development and Physiology (ed. J. Heslop-Harrison), pp. 177-185. London: Butterworth. JENSEN, W. A. & FISCHER, D. B. (1968). Cotton embryogenesis; the entrance and discharge of the pollen tube in the embryo sac. Planta 78, 158-183. JENSEN, W. A. & FISCHER, D. B. (1970). Cotton embryogenesis: the pollen tube in the stigma and style. Protoplasma 69, 215-235. KONAR, R. N. & LINSKENS, H. F. (1966). Physiology and biochemistry of the stigma fluid of Petunia hybrida. Planta 71, 372-387. KROH, M. (1967). Fine structure of Petunia pollen germinated 'in vivo'. Rev. Paleobot. Palyn. 3, 197-203. KROH, M., MIKI-HIROSIGE, H., ROSEN, W. & LOEWUS, F. (1970). Incorporation of label into pollen tube walls from myo-inositol labelled Lilium longiflorum pistils. PI. Physiol., Lancaster 45, 92-94- LARSON, D. A. (1965). Fine structural changes in the cytoplasm of germinating pollen. Am.jf. Bot. 52, 139-154. LINSKENS, H. F. (1964). Pollen physiology. A. Rev. PI. Physiol. 15, 255-270. LINSKENS, H. F. (1975). Incompatibility in Petunia. Proc. R. Soc. B 188, 299-311. Pollen tube development in Petunia 383 LINSKENS, H. F. & ESSER, K. L. (1957). Uber eine spezifische Anfarbung der Pollenschliiuche und die Zahl Kallosepropfen nach selbstung und fremdung. Natunvissenscliaften 44, 16. LINSKKNS, H. F. & SCHRAUWEN, J. (1966). Measurement of oxygen tension changes in the style during pollen tube growth. Planta 71, 98. LINSKENS, H. F., SCHRAUWEN, J. A. M. & KONING, R. N. N. (1970). Cell-free protein synthesis with polysomes from germinating Petunia pollen grains. Planta 90, 153-162. MARCHANT, R. & ROBARDS, A. W. (1968). Membrane systems associated with the plasmalemma of plant cells. Ann. Bot. 32, 457-471. MASCARENHAS, J. P. & LAFOUNTAIN, J. (1972). Protoplasmic streaming, cytochalasin B, and the growth of the pollen tube. Tissue & Cell 4, 11-14. NETTANCOURT, D. DE, DEVREUX, M., BOZZINI, A., CRESTI, M., PACINI, E. & SARFATTI, G. (J973)- Ultrastructural aspects of the self-incompatibility mechanism in Lycopersicum peruvianitm Mill. J. Cell Set. 12, 403-419. NETTANCOURT, D. DE, DEVREUX, M., LANERI, U., CRESTI, M., PACINI, E. & SARFATTI, G. (1974). Genetical and ultrastructural aspects of self- and cross-incompatibility in inter- specific hybrids between self-compatible Lycopersicum esculentum and self-incompatible L. peruvianum. Theoret. appl. Genet. 44, 278-288. O'KELLEY, J. C. & CARR, P. H. (1954). An electron micrographic study of the cell walls of elongating cotton fibres, root hairs, and pollen tubes. Am.jf. Bot. 41, 261-264. PLUIJM, J. VAN DER & LINSKENS, H. F. (1966). Feinstruktur der Pollenschlauche in Griffel von Petunia. Zilchter 36, 220-224. ROSEN, W. G. (1961). Studies on pollen-tube chemotropism. Am.jf. Bot. 48, 889-895. ROSEN, W. G. & GAWLICK, S. R. (1966). Relation of Lily pollen tube fine structure to pistil compatibility and mode of nutrition. In Electron Microscopy (ed. R. Vycda) (Proc. VI int. Congr. Electron Microsc, Kyoto), pp. 313-314. Tokyo: Maruzen. SANGER, J. M. & JACKSON, W. T. (1971). Fine structure study of pollen development in Haemanthus katherinae Baker. II. Microtubules and elongation of the generative cell.J. Cell Sci.S, 303-315. SASSEN, M. M. A. (1964). Fine structure of Petunia pollen grain and pollen tube. Ada bot. neerl. 13, 175-181. SCHLOSSER, K. (1961). Cytologische und cytochemische Untersuchungen uber das Pollen- schlauch-wachstum selbststeriler Petunia. Z. Bot. 49, 266-288. TUPY, J. (1959). Callose formation in pollen tubes and incompatibility. Biol. Plant 1, 192-198. VAN DE WOUDE, W. J., MORRE, D. J. & BRACKER, C. E. (1971). Isolation and characterization of secretory vesicles in germinated pollen of Li Hum longiflorum.J. Cell Set. 8, 331-351. VAZART, J. (1971). Degenerescence d'une synergide et penetration du tube pollinique dans le sac embryonnaire de Linum usitatissimum. Annh Univ. A.R.E.R.S., Reims 9, 89-97. WENT, J. L. VAN (1970). The ultrastructure of the fertilised embryo sac of Petunia. Ada bot. neerl. 19, 468-480.

(Received 17 May 1979 - Revised 12 August 1980)