J. Cell Sd. 82, 1-10 (1986)Printed in Great Britain © The Company of Biologists Limited 1986

STYLAR PEROXIDASE AND INCOMPATIBILITY

REACTIONS IN PETUNIA HYBRIDA

LUISA CARRARO, GIULIANA LOMBARDO AND F. M. GEROLADipartimento di Biologia, Sezione di Botanica Sistematica Universitd degli StudidiMilano, Via Celoria 26, 201'33 Milano, Italy

SUMMARYSelf-, cross- and non-pollinated styles of Petunia hybrida were treated with 3-3'-diamino-

benzidine for the ultrastructural localization of peroxidase activity. Wall peroxidases were absent incross-pollinated styles, but they were detectable as an osmiophilic deposit on the cell walls of theouter portion of the transmitting tissue in self- and non-pollinated styles. The cell layer showingperoxidase activity was thicker in the self-pollinated styles than in the non-pollinated ones. Inaccordance with current hypotheses on the mechanism involved in pollen incompatibility process,it is suggested: (1) that wall peroxidases present in the cells of the outer portion of the transmittingtissue are involved in the gametophytic self-incompatibility of Petunia; (2) that self-pollinationcauses an increase in the number of cells involved in the rejection process; (3) that non-pollinatedstyles, which are characterized by the presence of cell wall peroxidases in the outer portion of thetransmitting tissue, are 'prepared' to a certain extent for the rejection of incompatible pollen tubes.The removal of peroxidase activity thus seems to be an important step in the compatible pollinationprocess.

INTRODUCTION

The gametophytic self-incompatibility system of Petunia plants is characterizedby a pollen tube inhibition in incompatible combinations in the upper zone of thestylar transmitting tract (Heslop-Harrison, 1983). This stylar portion was called the'neck' by Herrero & Dickinson (1979), who stated that "the transmitting tissuecontains large spherical cells with characteristic ridges in their walls, which causethem to 'key' into an adjacent cell". Compatible and incompatible intraspecificpollinations induce cytological and metabolic changes in the stylar transmitting tissue(Herrero & Dickinson, 1979; Roggen, 1967); compatible and incompatible pollentubes growing in the stylar transmitting tissue of Petunia also show ultrastructuraldifferences (Cresti et al. 1979; Herrero & Dickinson, 1981). Regarding the incom-patibility reaction, Linskens (1975) and van der Donk (1975) have argued that therejection of pollen tubes is the reaction for which the style is prepared, whereasunaffected pollen tube growth is related to the synthesis of special enzymes that breakdown the incompatibility barrier.

In Nicotiana alata gametophytic self-incompatible flowers, many electrophoreticinvestigations have been carried out to detect the substances responsible for therecognition of the pollen tube along the style. Pandey (1967) and Bredemeijer

Key words: gametophytic self-incompatibility, Petunia hybrida, stylar transmitting tissue, wallperoxidases.

2 L. Carraro, G. Lombardo and F. M. Gerola

(Bredemeijer & Blaas, 1975) have reported the presence of peroxidase isoenzymes instyle extracts; however, the actual role of these enzymes in the gametophytic self-incompatibility mechanism is still an open question. Pandey (1967) has regardedperoxidases as a direct expression of the 5 gene; Bredemeijer (Bredemeijer & Blaas,1980) has reported an indirect non-specific action of peroxidases in the incompati-bility reactions. This author also showed that pollination and pollen tube growthcaused an increase in total peroxidase activity (Bredemeijer, 1974) and that there is apositive correlation between peroxidase isoenzyme 10 and intensity of the rejectionreaction (Bredemeijer & Blaas, 1975; Bredemeijer, 1976). In subsequent worksBredemeijer (1977, 1979) noted an even distribution of peroxidase 10 across thestyle, and he also proposed that only the fraction of peroxidase 10 present in theintercellular spaces and on the cell walls is responsible for pollen tube inhibition.

In a previous ultrastructural investigation on self-pollinated 'pin' flowers ofPrimula acaulis treated with 3-3'-diaminobenzidine (DAB) Carraro et al. (1985)reported the presence of peroxidases on the cell walls and intercellular spaces of thestylar transmitting tissue, whereas treatment with DAB did not provide evidence ofthe presence of peroxidases in cross-pollinated flowers.

To obtain further information supporting the hypothesis of the involvement ofperoxidases in the incompatibility reaction, we repeated the DAB treatment on self-,cross- and non-pollinated styles of Petunia hybrida.

MATERIALS AND METHODS

Seedlings of P. hybrida were raised from seeds of clones Ka3 and T2U (genotypes 22 and 33,respectively), known to be self-incompatible and cross-compatible (Herrero & Dickinson, 1979).Just before anthesis, the buds of T2U plants were detached and placed in small glass tubescontaining fresh water in a well-lit air-conditioned chamber, maintained at a constant temperatureof 25°C. Stamens and corolla were then removed and the pistils were divided into three groups.The first was self-pollinated by rubbing mature T2U anthers gently on the stigmas; the second wascross-pollinated with Ka3 anthers, the third was not pollinated. After 40 h the stigmas and upper2 mm of the style (the neck region) were cut away with a razor blade, prefixed in 3 % glutaraldehydein 0-05 M-cacodylate buffer, pH7-4 for 30min, at 4°C. Afterwards, a few pistils were postfixed incacodylate-buffered 1% osmium tetroxide, pH7-4, for 2h, at 4°C, dehydrated in ethanol andembedded in Araldite. Other style samples were used for the histochemical determination ofperoxidase. They were first rinsed in the same cacodilate buffer for 30min and then in 0-05 M-2-amino-2-methyl-l,3'-propandiol buffer, pH7-4, for 30min. Incubation with DAB (SigmaChemical Co., St Louis, MO) followed, for 60min, at 37°C. The incubation medium was alwaysprepared immediately before use: it contained lOmg DAB, 4-9ml of 0-05M-propandiol buffer,pH7-4, and 0-1 ml H2O2. The following procedures served as control to determine whether anyobserved precipitate could be reasonably ascribed to peroxidase: (1) preincubation for 30min at37°C with propandiol buffer, pH7-4, containing 0-02M-aminotriazole (3-amino-l,2,4-triazole,Aldrich Chemical Co., Inc., Milwaukee, WI), followed by incubation in DAB medium, alsocontaining 0-02M-aminotriazole (Rechcigl & Warren, 1963; Frederick & Newcomb, 1969); (2)preincubation for 30min at 37°C in propandiol buffer, pH7-4, containing O01 M-KCN, followedby incubation in DAB medium containing 0-01 M-KCN (Frederick & Newcomb, 1969; Burns &Little, 1949). After incubation, the segments were rinsed in 0-05 M-propandiol buffer, pH 7-4, andpostfixed in 1 % cacodylate-buffered osmium tetroxide, pH 7-4, for 2h at 4°C. The samples werethen dehydrated in ethanol and embedded in Araldite.

Ultrathin sections of the neck region were obtained with an LKB Ultrotome III and examined ina Philips TM 400 transmission electron microscope (TEM). Semithin sections were cut with a glass

Pervxidases and incompatibility in Petunia 3

knife and examined with a light microscope. The sections were not stained, to avoid any overlapwith reaction products.

RESULTS

Self-pollinated styles

By light microscopy of semithin cross-sections of the style neck, the transmittingtissue was seen to be characterized by spherical cells, whose size increased towardsthe periphery. After DAB treatment the cell walls of the outer portion of thetransmitting tissue appeared much darker than those of the central portion (Fig. 1).By electron microscopy of ultrathin sections, DAB treatment was seen to result inpronounced deposit of electron-opaque material only on the wall of the cells in theouter portion of the transmitting tissue (Figs 2, 4). No deposit was present in thecentral portion of the transmitting tissue (Fig. 3) or on the wall of pollen tubes(Fig. 4). The cells of the outer portion of the transmitting tissue (Fig. 2) werespherical and defined large intercellular spaces containing a granular electron-densematrix. The cells possessed a thin wall, evident nuclei, starch-containing plastids,small mitochondria and small vacuoles or a single large central vacuole. Pollen tubeswere observed in the outer portion of the transmitting tissue (Fig. 1); they appearedinfolded, close to the transmitting cells, and were characterized by degeneratedcytoplasm and a thick multilayered wall with fibrillar inclusions (Fig. 4). In thecentral portion of the transmitting tissue (Fig. 3) the cells were spherical and pos-sessed characteristic 'key junctions' as described by Herrero & Dickinson (1979).Chloroplasts containing large starch grains were particularly numerous.

When the DAB treatment was omitted, the deposition of electron-opaque materialon the cell walls of the transmitting tissue was absent and no difference was observedbetween the central and the outer portion of the transmitting tissue (data not shown).

When aminotriazole was added to the DAB medium, deposit of electron-opaquematerial on the cell walls of the outer portion of the transmitting tissue was stillpresent, even if less evident than in samples treated with DAB only (Fig. 5). Thiscontrol confirmed that the dark deposit was related to peroxidase activity, sinceaminotriazole is a specific inhibitor of catalase but not peroxidase (Carraro et al.1985).

When KCN was added to the DAB medium, the electron-opaque deposit on thecell walls of the outer transmitting tissue was completely absent, since KCN inhibitsboth catalase and peroxidase (Fig. 6) (Carraro et al. 1985).

Non-pollinated styles

Light microscopic observation of DAB-treated non-pollinated styles revealed thepresence of cell wall peroxidases in the outer portion of the transmitting tissue,However, the layer of cells characterized by the presence of peroxidase activityappeared to be reduced more than that in the self-pollinated styles (Fig. 7).

Electron microscopic examination of the outer transmitting tissue after DABtreatment revealed electron-opaque deposits on the cell walls similar to those

L. Carraro, G. Lombardo and F. M. Gerola

Figs 1-4. For legends see p. 7

Peroxidases and incompatibility in Petunia

cp

Figs 5-8. For legends see p. 7

L. Carraro, G. Lombardo and F. M. Gerola

Peroxidases and incompatibility in Petunia

Fig. 1. Semithin cross-section through the neck of a self-pollinated style, after DABtreatment; the outer portion of the transmitting tissue (op) is characterized by cell wallsmuch darker than those of the central portion (cp); incompatible pollen tubes aredetectable in the outer transmitting tissue (arrows). Light micrograph. X300.

Fig. 2. TEM of the outer portion of the transmitting tissue in a self-pollinated style,treated with DAB; a pronounced deposit of electron-opaque material is clearly visible onthe cell walls. The cells appear spherical, bound by a thin wall and defining largeintercellular spaces filled by an electron-dense matrix. The cells possess large nuclei (n),starch-containing plastids (/>), small mitochondria (m), small vacuoles or a single largecentral vacuole (v). X3600.

Fig. 3. TEM of the central transmitting portion of a self-pollinated style, treated withDAB. Electron-opaque material on the cell walls is absent. The cells have key junctionswith contiguous cells (arrows)./), starch-containing plastids. X4200.

Fig. 4. Detail of the outer portion of the transmitting tissue of a self-pollinated style,treated with DAB. An incompatible pollen tube (it) is clearly visible. Deposits ofelectron-opaque material are present on the transmitting cell walls (arrows). Theincompatible pollen tube possesses a thick multilayered wall (tv) with fibrillar inclusions(/) and lacks electron-opaque deposit. X7800.

Fig. 5. TEM of the outer transmitting portion of a self-pollinated style treated with DABin the presence of aminotriazole. A light deposit of electron-opaque material due toperoxidase activity is still visible on the cell walls of the transmitting tissue (arrows), it,incompatible pollen tube. X7800.

Fig. 6. TEM of the outer transmitting portion of a self-pollinated style treated with DABin the presence of KCN. Electron-opaque material on the cell walls of the transmittingtissue is absent, it, incompatible pollen tube; p, starch-containing plastids. XS000.

Fig. 7. Semithin cross-section through the neck of a non-pollinated style, treated withDAB. Only the outermost cell layer of the outer transmitting tissue (op) shows wallperoxidase activity. The remaining outer portion and the central portion (cp) of thetransmitting tissue contain cells with an almost indistinguishable cell wall devoid ofcontrast. Light micrograph. x400.

Fig. 8. TEM of the outer transmitting portion of a non-pollinated style after DABtreatment; only the outermost cell layer of this portion shows the electron-dense depositon the wall. The deposit disappears gradually in the inner cell layers, visible on the lowerleft side of the micrograph, p, starch-containing plastids; v, vacuoles. X3300.

Fig. 9. Semithin cross-section through the neck of a cross-pollinated style, after DABtreatment: central (cp) and outer portions (op) of the transmitting tissue possess cell wallsdevoid of contrast that did not react with the DAB medium. Light micrograph. X400.

Fig. 10. TEM of the outer transmitting portion of a cross-pollinated style, after DABtreatment. There is no electron-opaque deposit on the cell walls. The cells contain a densecytoplasm rich in organelles. The compatible pollen tube (ct) present among these cellsappears collapsed but not yet 'empty', n, nuclei; p, starch-containing plastids; m,mitochondria; u, vacuoles. X6000.

Fig. 11. TEM of the outer transmitting portion of a cross-pollinated style, after DABtreatment. Many compatible pollen tubes (ct) are detectable among the transmittingcells. The pollen tubes appear empty, flattened and collapsed. The transmitting cellspossess a large central vacuole and clear cell walls and lack any electron-opaque deposit.X330O.

8 L. Carraro, G. Lombardo and F. M. Gerola

observed in the self-pollinated samples (Fig. 8). The photograph was taken of theborder of the DAB-reacting cell layer; cells lacking the characteristic electron-opaquematerial in the wall can thus be observed.

Treatment with aminotriazole and KCN revealed that even in non-pollinatedstyles the electron-opaque deposit observed following DAB treatment was due toperoxidase activity (data not shown).

Cross-pollinated styles

In semithin cross-sections observed by light microscopy after DAB treatment, nodark deposit was observed on the cell walls of any portion of the transmitting tissue inthe neck region, resulting in an appearance devoid of contrast (Fig. 9). Whenobserved by electron microscopy, cell walls always appeared to be without electron-opaque deposit (Figs 10, 11). By comparison of these two figures the differentultrastructure of the transmitting cells is clearly evident. As previously reported(Herrero & Dickinson, 1979), probably owing to the elongation stage of the pollentube, the cells of the surrounding transmitting tissue appear to undergo largemodifications leading to cells characterized by large vacuoles and a thin peripherallayer of cytoplasm, and no longer rich in reserves (Fig. 11). At this level theintercellular spaces were filled with pollen tubes, which were almost empty; theyappeared flattened and collapsed and were characterized by a thin wall lacking anyfibrillar inclusion (Fig. 11). The absence of fibrillar inclusions in compatible pollentubes has also been previously reported (Herrero & Dickinson, 1981).

DISCUSSION

Whereas wall-peroxidase activity was detectable in both self- and non-pollinatedflowers, the 'peroxidase characterized' cell layer at the outer portion of the trans-mitting tissue was thicker in the self-pollinated than in the non-pollinated sample.Preliminary results might indicate a correlation between elongation of the incompat-ible pollen tubes and increase in thickness of the cell layer presenting peroxidaseactivity. These data are in agreement with previous observations (Pandey, 1967;Bredemeijer & Blaas, 1975; Bredemeijer, 1976) of an increase in peroxidase contentof the transmitting tissue following self-pollination. Our results indicating thepresence of peroxidases on the cell wall of non-pollinated and self-pollinated stylessupport the hypothesis of Bredemeijer (1977) that only the fraction of extracellularperoxidases is responsible for pollen tube growth inhibition. Also, our observation ofan even distribution of peroxidases between the central and the outer portion of thetransmitting tissue is reminiscent of the even distribution of peroxidase 10 observedby Bredemeijer (1979) between cortex and transmitting tissue of Nicotiana.

Of particular importance is the observation that compatible pollination does notcause the increase in the number of cells showing peroxidase activity, but does causethe disappearance of the peroxidase activity observed in non-pollinated samples.This finding is basically in agreement with the hypothesis of Linskens (1975) and van

Peroxidases and incompatibility in Petunia 9

der Donk (1975) that the style is prepared by nature for the rejection reaction of thepollen tube, whereas the unaffected growth of compatible pollen tubes is due to thebreak in the incompatibility barrier.

At this time we cannot distinguish between different mechanisms involved in thedisappearance of peroxidase activity in cross-pollinated styles, i.e. inhibition,sequestration or degradation of the enzyme. Further studies will be carried on thisaspect.

As previously reported by Herrero & Dickinson (1979), we also observed aconsiderable ultrastructural modification in the transmitting cells in the cross-pollinated samples, probably related to the elongation and growth of the pollen tube.However, in cross-pollinated styles the peroxidase activity was absent also in the wallof the cells that had not yet undergone extensive ultrastructural changes (seeFig. 10), indicating that the removal of peroxidase activity is a preceding process. Allthese results suggest that peroxidases are some of the components responsible forincompatibility reactions. To examine this hypothesis, the absence or presence ofperoxidases in very young buds of Petunia, which seem to lack the incompatibilitybarrier (Ascher, 1984), will be investigated.

The authors thank Professor Linskens, who kindly supplied the Petunia seeds. The authors areindebted to Dr P. D. Gerola for his critical reading and correction of the manuscript. This workwas also supported by M.P.I. (40%).

REFERENCES

ASCHER, P. D. (1984). Self-incompatibility. In Petunia (ed. K. C. Sink), pp. 92-110. Berlin,Heidelberg, New York, Tokyo: Springer-Verlag.

BREDEMEIJER, G. M. M. (1974). Peroxidase activity and peroxidase isoenzyme composition in self-pollinated, cross-pollinated and unpollinated styles of Nicotiana alata. Acta bot. neerl. 23,149-157.

BREDEMEIJER, G. M. M. (1976). Effect on bud-pollination and delayed self-pollination on theinduction of a possible rejection peroxidase in styles of Nicotiana alata. Acta bot. need. 25,107-116.

BREDEMEIJER, G. M. M. (1977). Peroxidase leakage and pollen tube growth inhibition in agedNicotiana alata styles. Acta bot. neerl. 26, 231-237.

BREDEMEIJER, G. M. M. (1979). The distribution of peroxidase isoenzymes, chlorogenic acidoxidase and glucose-6-phosphate dehydrogenase in transmitting tissue and cortex of Nicotianaalata styles. Acta bot. need. 28, 197-203.

BREDEMEIJER, G. M. M. & BLAAS, J. (1975). A possible role of a stylar peroxidase gradient in therejection of incompatible growing pollen tubes. Acta bot. need. 24, 37-48.

BREDEMEIJER, G. M. M. & BLAAS, J. (1980). Do S allele-specific peroxidase isoenzymes exist inself-incompatible Nicotiana alata? Theor. appl. Genet. 57, 119-123.

BURRIS, R. H. & LITTLE, H. N. (1949). Oxidases, peroxidases, and catalase. In RespiratoryEnzymes (ed. H. Lardy), pp. 170—173. Minneapolis: Burgess.

CARRARO, L., LOMBARDO, G. & GEROLA, F. M. (1985). Electron-cytochemical localization ofperoxidase in self- and cross-pollinated styles of Primula acaulis. Caryologia 38, 83-94.

CRESTI, M., CIAMPOUNI, F., PACINI, E., SARFATTI, G., VAN WENT, J. L. & WILLEMSE, M. T. M.

(1979). Ultrastructural differences between compatible and incompatible pollen tubes in thestylar transmitting tissue of Petunia hybrida.J. submicrosc. Cytol. 11, 209-219.

FREDERICK, S. E. & NEWCOMB, E. M. (1969). Cytochemical localization of catalase in leafmicrobodies (peroxisomes). J. Cell Biol. 43, 343-353.

10 L. Carraro, G. Lombardo and F. M. Gervla

HERRERO, M. & DICKINSON, H. G. (1979). Pollen-pistil incompatibility in Petunia hybrida:changes in the pistil following compatible and incompatible intraspecific crosses, J. Cell Sci. 36,1-18.

HERRERO, M. & DICKINSON, H. G. (1981). Pollen tube development in Petunia hybrida followingcompatible and incompatible intraspecific matings.,7. Cell Sci. 47, 365-383.

HESLOP-HARRISON, J. (1983). Self-incompatibility: phenomenology and physiology. Proc. R. Soc.B 218, 371-395.

LlNSKENS, H. F. (1975). Incompatibility mPetunia. Proc. R. Soc. B 188, 299-311.PANDEY, K. K. (1967). Origin of genetic variability: Combinations of peroxidase isozymes

determine multiple allelism of the 5 gene. Nature, Land. 18, 669-672.RECHCIGL, M. & WARREN, H. E. (1963). Role of catalase and peroxidase in the metabolism of

leucocytes. Nature, Land. 199, 1001-1002.ROGGEN, H. P. J. R. (1967). Changes in enzyme activities during the progame phase in Petunia

hybrida. Ada bot. neerl. 16, 1-31.VAN DER DONK, J. A. W. M. (1975). Recognition and gene expression during the incompatibility

reaction in Petunia hybrida L. Molec. gen. Genet. 141, 305-316.

(Received 8 July 1985 -Accepted, in revised form, 8 November 1985)