Angiosperm stigmas have long been known to exhibit high levels of peroxidase activity when they are mature and most receptive to pollen but the biological function of stigma peroxidases is not known. A novel stigma-specific class III peroxidase gene, SSP (stigma-specific peroxidase) expressed exclusively in the stigmas of Senecio squalidus L. (Asteraceae) has recently been identified. Expression of SSP is confined to the specialized secretory cells (papillae) that compose the stigma epidermis. The literature on stigma peroxidases and hypotheses on their function(s) is reviewed here before further characterization of SSP and an attempt to determine its function are described. It is shown that SSP is localized to cytoplasmic regions of stigmatic papillae and also to the surface of these cells, possibly as a component of the pellicle, a thin layer of condensed protein typical of ‘dry’ stigmas. Enzyme assays on recombinant SSP showed it to be a peroxidase with a preference for diphenolic substrates (ABTS and TMB) and a pH optimum of ∼4.5. In such assays the peroxidase activity of SSP was low when compared with horseradish peroxidase. To explore the function of SSP and other stigmatic peroxidases, levels of reactive oxygen species (ROS) in stigmas of S. squalidus were investigated. Relatively large amounts of ROS, principally H2O2, were detected in S. squalidus stigmas where most ROS/H2O2 was localized to the stigmatic papillae, the location of SSP. These observations are discussed in the context of possible functions for SSP, other peroxidases, and ROS in the stigmas of angiosperms.
Pollination is one of the most critical stages in the life cycle of a flowering plant, involving a complex series of cell–cell interactions that constitute the pollen–pistil interaction (Heslop-Harrison, 1978). This critical cellular ‘courtship’ between the haploid male gametophyte (pollen) and the diploid female cells (pistil) of the sporophyte determines whether fertilization will occur and is a paradigm for the study of cell–cell recognition in plants.
As a prelude to fertilization, pollen must first establish molecular congruity/compatibility with the stigma and then germinate to produce a pollen tube that penetrates the stigma and grows through the transmitting tissue of the style to locate an ovule within the ovary. Initiation and successful completion of this sequence of events depends upon the stigma and style providing the exact requirements for pollen germination and sustained growth and guidance of the pollen tube through the pistil and ovary (Heslop-Harrison, 2000; Herrero, 2003; Swanson et al., 2004). The pollen must therefore be programmed to respond appropriately at every step of this interaction. The timing of pollination is critical, because the stigmatic surface of the pistil is only receptive to pollen for a relatively short period. Pollination outside this window of female receptivity will result in reduced seed set or no seed set (Herrero, 2003).
At maturity, when stigmas are ‘ripe’ for pollination they are characterized by high levels of peroxidase activity (Dupuis and Dumas, 1990; Dafni and Motte Maues, 1998; Fig. 1) and also esterase activity (Hiscock et al., 1994, 2002a). Surprisingly, despite concerted efforts to characterize stigma and pollen molecules that regulate self-recognition during the self-incompatibility (SI) response, and also pollen tube guidance (Franklin-Tong, 2002; Johnson and Preuss, 2002; Hiscock and McInnis, 2003; Swanson et al., 2004), virtually nothing is known about the function of these abundant stigmatic enzymes. Interestingly, while esterase activity shows no appreciable change during stigma development (Hiscock, 1993, 2004), peroxidase activity increases dramatically as pistils mature, reaching a peak when the stigma is most receptive to pollen (Dupuis and Dumas, 1990; Seymour and Blaylock, 2000; Stpiczynska, 2003). Indeed, the tests most widely used to determine pistil receptivity involve measuring stigma peroxidase activity using various peroxidase tests (Dafni and Motte Maues, 1998).
Despite their morphological diversity, angiosperm stigmas can be divided into two basic types: ‘wet’ and ‘dry’ (Heslop-Harrison and Shivanna, 1977; Hiscock et al., 2002b; Hiscock, 2004; Edlund et al., 2004). ‘Wet’ stigmas produce a copious surface secretion that may be either predominantly aqueous (e.g. Liliaceae) or predominantly lipidic (e.g. Solanaceae). By contrast, ‘dry’ stigmas bear no surface secretion; instead the cuticle is overlaid by a condensed surface layer of protein, the proteinaceous pellicle (Heslop-Harrison et al., 1975; Hiscock, 2004). Interestingly, despite these fundamental differences, both stigma types show similar high levels of esterase and peroxidase activity (Fig. 1), suggesting that these enzymes are fundamentally important for stigma function.
The function of stigmatic peroxidases is unknown
Peroxidases generally catalyse the breakdown of hydrogen peroxide to yield highly oxidizing intermediates, which oxidize a variety of organic and inorganic reducing substrates. Understanding peroxidase function in plants has proved problematic because of their broad range of substrate preferences and lack of tissue specificity. Most plant peroxidases belong to a large superfamily of secreted isozymes collectively called class III peroxidases (Welinder, 1992). Most of the characterized class III plant peroxidases are expressed in a range of plant tissue types and play diverse roles in plant metabolism and physiological processes such as oxidative stress responses, lignification, suberization, auxin metabolism, cross-linking of cell wall components, defence against pathogen attack, and salt tolerance (Welinder, 1992; Penel et al., 2003). Recently H2O2 and other reactive oxygen species (ROS) have been shown to be involved in cell signalling in plants, where they regulate diverse aspects of plant metabolism and cell growth (Neill et al., 2002; Foreman et al., 2003; Rentel and Knight, 2004). H2O2 has specifically been shown to be involved in stomatal closure (Neill et al., 2002) and the hypersensitive response (Grant and Loake, 2000). Given that peroxidases can generate H2O2 as well as consume it (Bolwell, 1996; Blee et al., 2001; Bolwell et al., 2002), peroxidases should now be considered as potentially important components of signal transduction pathways in plants.
Ever since peroxidases were identified as a major enzyme component of angiosperm stigmas (Pandey, 1967) there has been much speculation about the function of these enzymes in stigmas. Early isozyme studies in Nicotiana suggested a potential role in pollination processes, particularly SI (Pandey, 1967; Bredemeijer and Blaas, 1975). These studies showed that stylar peroxidase activity was higher after incompatible pollinations compared with compatible pollinations. Unpollinated pistils, however, showed stylar peroxidase activities comparable to those of selfed (incompatible) pistils, so Bredemeijer (1974) concluded that in the compatible style peroxidase activity was somehow inactivated. Closer inspection indicated that only the peroxidase activity associated with the extracellular secretion of the stylar transmitting tissues was reduced following compatible pollinations (Bredemeijer and Blass, 1975). Later Carraro et al. (1986) went so far as to suggest that peroxidases were responsible for the inhibition of incompatible pollen in Petunia and Nicotiana. Today this is known to be incorrect because SI in the Solanaceae is mediated by stylar S-RNases (McCubbin and Kao, 2000). However, in Nicotiana, Lycopersicon, and Petunia a peroxidase gene is linked to each of their respective S loci (ten Hoopen et al., 1998), but these S-linked peroxidase loci have never been characterized. The many accounts of altered peroxidase activity in pistils of Nicotiana and Petunia after pollination (Pandey, 1967; Bredemeijer, 1974, 1977, 1979; Bredemeijer and Blaas, 1975, 1980, 1981; Carraro et al., 1986) cannot be ignored because they suggest a potential role for peroxidases in pollen–pistil interactions that demands further investigation.
Even though stigmas, particularly wet stigmas, provide an ideal environment for bacteria and fungi, they rarely show signs of infection, suggesting that stigmas are protected from pathogen attack. It is well known that pistils express a number of pathogenesis-related (PR) proteins, some of which are peroxidases (Curtis et al., 1997) so, as an alternative role to that proposed in pollen–pistil interactions, it is possible that peroxidases may provide enhanced protection against pathogen attack when the stigma is ‘primed’ to receive pollen.
SSP: a stigma-specific peroxidase from Senecio squalidus
The first plant peroxidase gene expressed exclusively in stigmas (McInnis et al., 2005) was recently identified. Importantly, this peroxidase, SSP (stigma-specific peroxidase), is expressed specifically in the specialized secretory cells (papillae) of the stigma epidermis and is developmentally regulated. Expression of SSP is undetectable in small flower buds, but increases during flower development, to reach a maximum in newly opened flowers when stigmas are most receptive to pollen. These properties make SSP the first characterized plant peroxidase with such a precise cell-specific and developmentally regulated expression profile. Interestingly, SSP is also polymorphic. So far, six SSP alleles have been isolated from S. squalidus that share an overall amino acid identity of 95–99% (McInnis et al., 2005). Indeed, SSP was initially identified because two alleles, SSP1 and SSP2, appeared to be linked to the S1 and S2 self-incompatibility alleles, respectively (Hiscock et al., 2003). Characterization of additional SSP alleles, however, revealed that SSP was not located at the S locus of S. squalidus (McInnis et al., 2005), although like the peroxidase loci identified in Solanaceous SI species, SSP does appear to be linked to the Senecio S locus. What then is the function of SSP? Extensive searches of the public databases have yet to identify a peroxidase with close sequence similarity to SSP (McInnis et al., 2005) and among the 73 peroxidase genes identified in Arabidopsis (Valério et al., 2004) those with closest sequence similarity to SSP are not expressed exclusively in stigmas of Arabidopsis (S McInnis and S Hiscock unpublished results). Thus a reverse genetic analysis of an Arabidopsis orthologue of SSP is not currently possible. The continued characterization of SSP and further attempts to obtain information about the potential function(s) of SSP and stigma peroxidases in S. squalidus are described here.
Materials and methods
Senecio squalidus plants were grown in a greenhouse under a 16/8 h day/night photoperiod according to Hiscock (2000).
Unless specified, all chemicals were obtained from Sigma.
Production of SSP antibody
Polyclonal antiserum was raised in a rabbit against a synthetic peptide (CSVGRPNQLTFSKK) specific to a predicted external loop/helix region of SSP. This region was chosen to avoid potential cross-reaction with other peroxidases. The C-terminal Cys residue (not present in the SSP protein sequence) was included in the peptide to allow conjugation of the peptide to a carrier protein for efficient antibody production. Peptide synthesis and antibody production and purification were carried out by a commercial company (Affiniti Research Products Ltd., Exeter, UK).
Electrophoresis and western blotting
Total stigma protein was extracted and quantified according to McInnis et al. (2005). SDS-PAGE and electroblotting were carried out using the Bio-Rad Mini-Protean II system according to Hiscock et al. (1994). To visualize total protein, gels were stained with Coomasie brilliant blue (Hiscock et al., 1994). Following blotting, membranes were placed in blocking buffer (1× PBS, 0.1% v/v Tween-20, and 3% w/v dried non-fat milk [Marvel]) for 1 h with gentle agitation. Primary antibody was then diluted 1:1000 in blocking buffer and applied to the blot for 1 h at RT. The blot was then washed with wash buffer (1× PBS, 0.1% w/v Tween-20, 0.5% w/v BSA, and 1% w/v Marvel) for three 10 min washes. The membrane was then placed in secondary antibody (HRP-conjugated goat anti-rabbit) diluted 1:1000 in wash buffer for 1 h. This was then washed, as above, and bands visualized using enhanced chemiluminesence (ECL) reagents (Amersham) according to the manufacturer's instructions. Isoelectric focusing (IEF) was carried out using a Multiphor II system (Amersham Life-Sciences) according to McInnis et al. (2005). Peroxidase activity of protein bands was determined using a commercially available peroxidase indicator reagent (Sigma) again according to McInnis et al. (2005).
Immunolocalization of SSP
SSP protein was localized on ultra-thin sections of Senecio stigmas fixed and embedded in LR-White resin according to Armstrong et al. (2002). Sections were mounted on Pioloform-coated nickel grids and then reacted with Anti-SSP serum largely following the method of Armstrong et al. (2002): Anti-SSP serum and preimmune serum (negative control) were diluted 1:20 in buffer (1× PBS, 1% BSA pH 7.2); goat anti-rabbit IgG conjugated to 15 nm gold particles (Sigma) was diluted 1:20 in the same buffer. Sections were viewed with a JEOL 2000EX transmission electron microscope at 80 kV.
Structural modelling of SSP
A three-dimensional model of the structure of SSP was obtained based on the amino acid sequence of the mature protein using a computer simulation of protein folding (Anthony Clarke and Richard Sessions, Department of Biochemistry, University of Bristol) and X-ray diffraction data from the Arabidopsis thaliana neutral peroxidase AtP N (PDB: 1QGJ) (Mirza et al., 2000) as a reference.
Production of recombinant SSP
A PCR product containing the mature SSP cDNA coding sequence, flanked by NcoI (5′) and XhoI (3′) restriction sites was generated using the following primers: CATGCCATGGGTTGCAAAGTTGGTTTCTACCAGGC (5′) and CCGCTCGAGTCATTAATTAATTCTGTTACAAACCCTAC (3′). This was inserted into the expression vector pET28a and expressed in E. coli BL21 (DE3). Insoluble protein was purified, solublized in 8 M urea, 20 mM TRIS–HCl (pH 8.5) and subjected to chromatography using a 5 ml HiTrap Q HP column (Amersham). Purified protein was then refolded in the presence of haem, according to Neilson et al. (2001) and subjected to a further round of anionic chromatography in 20 mM TRIS–HCl (pH 8.5). The proportion of correctly folded recombinant SSP was assessed spectrophotometrically (A403/A275) and judged to be sufficiently high purity (RZ=1.9) for enzyme assay studies.
The absorbance spectra of recombinant SSP (native and compound II) were measured in the visible range (from 275 nm to 700 nm) using a Unicam 500 dual wavelength spectrophotometer. One millilitre of rSSP or soybean peroxidase (SBP) solution (both ∼0.6 mg ml−1 in 50 mM HEPES, pH 7.0) was placed in a cuvette and absorbance scanned at 0.5 nm bandwidths from 275–700 nm. Compound II was then generated by adding 4 mg sodium dithonite (Fluka) to the cuvette and vortexing vigorously before rescanning (Dunford, 1999).
Peroxidase activity was initially assayed spectrophotometrically at 415 nm with 1 mM ABTS, 25 mM H2O2 in 50 mM sodium-acetate buffer, pH 3.8–5.6 (Ruzin, 1999). A range of potential substrates (each 1 mM in sodium-acetate buffer, pH 4.51 with 90 mM H2O2) were tested for change in absorbance with SSP (∼6 μg) at the specified wavelengths: ABTS (415 nm), guaiacol (470 nm), phenol (400 nm), ascorbic acid (265 nm), catechol (295 nm), TMB (400 nm), and chloronaphthol (400 nm). The effect of pH on substrate reduction activity was tested using 1 mM ABTS, 90 mM H2O2 in 50 mM glycine/HCl (pH 2.2 and 3.0), 50 mM sodium acetate (pH 3.5–5.5), 50 mM phosphate-citrate buffer (pH 4.4–4.8), 50 mM MES (pH 6.0), 100 mM sodium phosphate (pH 5.5–7.5), 50 mM HEPES (pH 7.0–8.0), 50 mM TRIS–HCl (pH 8.0), and 100 mM glycine-NaOH (pH 9.0–13.0) (Ruzin, 1999). The effect of calcium concentration on enzyme activity (0.5–20 mM CaCl2 in 50 mM sodium acetate buffer, pH 4.51) was also tested (as in Medda et al., 2003). IAA oxidase activity was tested according to the method of Ferrer et al. (1992).
Peroxidase activity was localized by immersing excised pistils in a solution containing 0.1 M guaiacol, 0.1 M H2O2, in 20 mM phosphate buffer pH 4.5, until an orange/red colour was observed (approximately 1–3 min). ROS/H2O2 localization was performed by immersing excised pistils in a solution containing the ROS indicator dye TMB (3,3′,5,5′-tetramethylbenzidine-HCl, 0.1 mg ml−1 in TRIS-acetate, pH 5.0) until a blue colour was observed (Barceló et al., 2002). Stained pistils were observed using a Nikon dissecting microscope and photographed with a Nikon Coolpix digital camera.
ROS/H2O2 measurement was performed using the fluorescent ROS indicator dye DCFH2-DA (dichlorodihydrofluoroscein diacetate, Calbiochem). Pistils of S. squalidus were excised from 1 d post-anthesis flowers and immediately immersed in 5 ml of 50 μM DCFH2-DA (2′,7′-dichlorodihydrofluorescein diacetate, Calbiochem) in MES-KCl buffer (5 mM KCl, 10 mM MES, 50 μM CaCl2, pH 6.15) for 10 min followed by a wash step in fresh buffer for 15 min. Other pistils were treated with 1 M sodium pyruvate (Sigma) in MES–KCl buffer for 30 min, followed by a wash step, before finally treating with DCFH2-DA as above. Negative controls were treated with buffer only. Imaging with confocal microscopy was performed on a Nikon PCM2000 (excitation 488 nm, emission 515–560 nm) and data analysed with SCION IMAGE software (Scion, Frederick, MD). Relative fluorescence intensity values were calculated as average intensities from several pistils analysed in different experiments.
Immunolocalization of SSP in Senecio stigmas
Western blotting confirmed the specificity of the polyclonal SSP antiserum. A single cross-reacting band of ∼35 kDa, the predicted molecular weight of SSP, was obtained in the lane containing total Senecio stigma proteins (Fig. 2). Preimmune serum showed no cross reactivity with any proteins in the stigma extract (data not shown). A strong signal at ∼35 kDa was also obtained in the lane containing recombinant SSP, thereby confirming the specificity of the anti-SSP serum. The multiple cross-reacting bands in the recombinant SSP lane indicate the presence of background amounts of misfolded forms of SSP.
To determine the cellular location of SSP in the Senecio stigma, immunolocalization of SSP was performed on ultra-thin sections of mature unpollinated stigmas (Fig. 3). Gold particles were observed predominantly in the cytoplasm of stigmatic papillae and also on the cell surface (Fig. 3). Control sections treated with preimmune serum showed no significant cross-reaction between the gold-conjugated secondary antibody and stigmatic papillae (data not shown). Interestingly, very few gold particles were observed bound to the cell wall or the apoplastic spaces between cell walls of papillae. Instead the majority of gold particles associated with the cell wall were bound to its surface where they are most likely associated with the proteinaceous pellicle, a thin layer of condensed protein material that overlays the cuticle of dry or semi-dry stigmas (Fig. 3A–C). Indeed, in Fig. 3B and C the pellicle can be seen lifting from the papilla cuticle with gold particles clearly bound to it. Within the cytoplasm, gold particles occurred mainly in clusters, suggesting that SSP may be localized to specific compartments, organelles or vesicles, although it was difficult to discern the structural nature of these compartments. (Fig. 3A, B). Nevertheless, in some preparations gold particles could be seen distinctly associated with the Golgi apparatus (Fig. 3D).
A structural model of SSP
A three-dimensional model for the structure of SSP was obtained based on the amino acid sequence of the mature SSP protein and the X-ray diffraction pattern of a similar class III peroxidase from Aradidopsis (AtP N [PDB:1QGJ]; Mirza et al., 2000) using a computer simulation (Fig. 4). The structure of SSP conformed to that of a typical class III plant peroxidase with a well-defined substrate channel leading to an active site containing the haem prosthetic group. The region of the protein composed of the amino acid sequence to which the anti-SSP antibody was made was confirmed to consist of an extended loop and helix (Fig. 4) as predicted from analysis of the linear amino acid sequence. Polymorphic residues previously described in McInnis et al. (2005), mapped mainly to surface regions on alpha helices or exposed loops and are therefore unlikely to impact on the overall structural conformation or biochemistry of the different SSP allozymes.
Peroxidase activity of native and recombinant SSP
An in-gel peroxidase assay of total Senecio stigma proteins separated by IEF identified at least five peroxidases with pIs of approximately 9, 8, 7, 6, and 4.5 (Fig. 5A). Two of these peroxidases with pIs of ∼8 and ∼7 corresponded to SSP allozymes (alleles), SSP1 and SSP2, respectively (McInnis et al., 2005). The strong but diffuse peroxidase staining at pI 9 and pI 4.5 is probably associated with more than one peroxidase band. Peroxidase activity of the two SSP bands was far lower than the activity of these two regions suggesting that the majority of the peroxidase activity detected in the Senecio stigma (Figs 1, 8) is associated with the pI 4.5 and pI 9 peroxidases.
Recombinant SSP (rSSP) was folded in the presence of haem to obtain a mature protein which demonstrated peroxidase activity in assays with ABTS (Fig. 5B). The peroxidase activity of rSSP was low compared with horseradish peroxidase (HRP) and was optimal at concentrations of H2O2 (100 mM) well in excess of those concentrations of H2O2 (10 mM–25 mM) that gave optimal HRP activity (Fig. 5b). Changes in Ca2+ concentration had no effect on the peroxidase activity of rSSP with ABTS unlike the calcium-sensitive Euphorbia peroxidase characterized by Medda et al. (2003).
Absorbance spectra of rSSP in the Soret region were measured alongside those of soybean peroxidase (SBP) and found to be comparable (Fig. 6). Both rSSP and SBP showed a shift in maximal absorbance from the ‘native’ form to compound II when reduced with sodium dithionite. This change in absorbance is characteristic of peroxidases and is associated with the iron in the haem ligand [prosthetic group] changing from the Fe3+ state to the Fe4+=O…H+* state. The Soret peak shift for rSSP was from 411 nm (‘native’ rSSP) to 424 nm for compound II, compared with a peak shift of 403 nm to 433 nm for native versus compound II SBP. The RZ (purity number=A360/A280) calculated for this preparation of rSSP was 1.9.
Peroxidase activity of SSP was also assayed against a range of substrates to identify potential biochemical pathways in which SSP might participate. Recombinant SSP showed activity with diphenols, having relatively high activity with ABTS (Fig. 5B) and TMB, but lower activity with chloronaphthol (data not shown). In every case the activity of recombinant SSP was low compared to HRP (see Fig. 5 for ABTS assay). Interestingly, recombinant SSP showed no activity with the monophenols catechol, phenol, and guaiacol (data not shown). Like the majority of class III peroxidases, SSP was unable to use ascorbic acid as a substrate. SSP did not have IAA oxidase activity so is unlikely to be involved in auxin catabolism within the stigma. Assays with ABTS revealed the pH optimum for rSSP to be between pH 4.4 and pH 4.5 in both sodium acetate and phosphate-citrate buffers (Fig. 7).
Stigmas of Senecio generate hydrogen peroxide constitutively
Treatment of Senecio stigmas with TMB resulted in a blue colour deposit in cells of the stigma (Fig. 8A) indicating constitutive production of ROS, most probably H2O2. A similar pattern of staining was observed in stigmas treated with guaiacol to visualize peroxidase activity (Fig. 8B). Because rSSP was unable to react with guaiacol in in vitro assays it is unlikely that this in vivo assay reflects activity of SSP, but rather reflects the activity of one or more of the other three or more stigmatic peroxidases identified by in gel assays (Fig. 5A). Interestingly, TMB did not stain the pseudo-papillae situated at the ends of the stigma lobes (Fig. 8A). These cells, which act as ‘pollen-presenters’, also showed very low levels of peroxidase activity relative to the rest of the stigma (Fig. 8B). Confocal microscopy, following treatment of Senecio pistils with the ROS-sensitive fluorescent probe DCFH2-DA, confirmed high levels of constitutive ROS/H2O2-production by Senecio stigmas and showed that ROS/H2O2 production is confined principally to the surface papillae (Fig. 8C, D). Treating pistils with sodium pyruvate (a scavenger of H2O2) prior to treatment with DCFH2-DA resulted in reduced fluorescence in the papillae (Fig. 8E, F). This reduction in fluorescence increased as the concentration of sodium pyruvate was increased from 10 mM to 1 mM indicating that the ROS produced by papillae consists predominantly of H2O2.
Peroxidases of unknown function are a ubiquitous feature of mature angiosperm stigmas making peroxidase assays a useful means of assessing stigma receptivity/viability (Daphne and Motte Maues, 1998). The biological function of stigmatic peroxidases is, however, not known. At least five peroxidases were identified in the stigma of Senecio squalidus, one of which corresponded to the protein product of a peroxidase gene, SSP, which had previously been characterized and shown to be expressed exclusively in stigmas (McInnis et al., 2005). Immunolocalization confirmed that the SSP protein is present only in stigmatic papillae. In the papillae SSP was associated predominantly with the cytoplasm and the surface of the papilla cell wall, where SSP was localized to the proteinaceous pellicle (Fig. 3). This cellular distribution was unexpected because class III plant peroxidases are most frequently localized within the cell wall (Bestwick et al., 1997, 1998; Do et al., 2003). Enzyme assays confirmed that native SSP and recombinant SSP have peroxidase activity, but this enzyme activity appeared low compared with other class III peroxidases, such as HRP, and some other peroxidases identified in stigma extracts of Senecio. Unlike ELP, the Euphorbia peroxidase characterized by Medda et al. (2003), this low level of enzymatic activity was not due to a suboptimal concentration of Ca2+. Interestingly, SSP showed its highest peroxidase activity at concentrations of H2O2 far in excess of that producing optimal enzyme activity for HRP. This observation may prove significant for the function of SSP in vivo given the relatively high levels of constitutive H2O2 production observed in Senecio stigmas. The pH optimum for SSP (pH 4.5) equated well with that of other class III plant peroxidases, which are mainly active at acidic pH (Welinder, 1992). It is interesting that rSSP did not show enzyme activity with the mono-phenolic substrates phenol, catechol, and guaiacol. This suggests that SSP is unlikely to be involved in cross-linking cell wall phenolic compounds, a conclusion supported indirectly by the absence of significant amounts of SSP in the cell walls of stigmatic papilla as revealed in the immunolocalization studies. The lack of any detectable IAA-oxidase activity indicates that SSP is not directly involved in auxin regulation within the stigma. These findings suggest that SSP has a unique, as yet undefined, biological function within the stigma.
Searches for potential orthologues of SSP in the public databases have yet to identify any convincing candidates (McInnis et al., 2005; S McInnis, S Hiscock, unpublished results). Phylogenetic analysis of peroxidases with the highest sequence similarity to SSP revealed that SSP belongs to a distinct clade of class III plant peroxidases that possess two introns, instead of the more usual three introns (McInnis et al., 2005). This smaller clade of peroxidases appears to have evolved out of the three-intron group by the loss of an intron, an event that must have predated the divergence of eudicots (McInnis et al., 2005). Within the two-intron clade the two peroxidase genes with greatest sequence similarity to SSP that are reported to be expressed in flowers are the tomato cationic peroxidase, CEVI16 (Gadea et al., 1996), and the cotton cationic peroxidase POD2 (Delannoy et al., 2003). It is not known whether these genes are expressed in stigmas, but both are expressed in a variety of other tissues, so unlike SSP they are not stigma-specific, or even flower-specific.
The two-intron peroxidase clade also contains four Arabidopsis peroxidases with sequence similarity to SSP, but none are expressed specifically in flowers (Justesen et al., 1998; Welinder et al., 2002; Valério et al., 2004), so are unlikely to have functional identity with SSP. Interestingly, not one of the 73 Arabidopsis peroxidase genes appears to be expressed exclusively in flowers (Valerio et al., 2004), so at present it is not possible to investigate the function of an Arabidopsis SSP orthologue using gene knock-outs. Given the large numbers of peroxidase genes identified in the sequenced genomes of Arabidopsis (73) and rice (138) (Passardi et al., 2004) a high level of redundancy may be expected among class III plant peroxidases, so it remains possible that the function of SSP in Senecio stigmas may be carried out by an unrelated peroxidase or peroxidases in other species, perhaps even by a peroxidase that is not found exclusively in stigmas.
What then could be the function of SSP and other stigma-expressed peroxidases? This study's finding that the stigmas of Senecio generate significant quantities of ROS/H2O2 constitutively may be important in the context of the function of peroxidases in stigmas, especially because the ROS/H2O2 is produced principally in the stigmatic papillae where SSP is localized. Production of ROS/H2O2 by stigmas is not confined to Senecio and may be quite widespread among angiosperms: recent work has shown that stigmas from 20 different species across a range of different angiosperm groups (basal monocolpates, monocots, and eudicots) all produce ROS/H2O2 constitutively (S McInnis et al., unpublished data). Given that the stigmas of these species also have high levels of peroxidase activity, it is tempting to speculate upon a functional relationship between ROS/H2O2 production in stigmas and peroxidase activity. Perhaps SSP and other stigmatic peroxidases are important for regulating levels of H2O2 in stigmas or perhaps stigmatic peroxidases produce H2O2. In most plant cells H2O2 is generated by one or more NADPH oxidases (NOX) (Neill et al., 2002; Foreman et al., 2003), but some plant peroxidases have also been shown to generate H2O2 (Bolwell, 1996; Blee et al., 2001). Given that it is only possible to obtain minimal levels of peroxidase activity from native and recombinant SSP, an important next step is to determine whether rSSP can generate H2O2 as well as consume it.
So far it appears that SSP is unique among class III plant peroxidases, so it is important now to characterize other peroxidases expressed in stigmas from other species, especially Arabidopsis to determine whether SSP is unique in its stigma-specificity and/or its stigma-functionality. Most plant peroxidase genes are expressed in a variety of tissue types and are frequently induced by particular stresses and other environmental factors (Hiraga et al., 2001; Valério et al., 2004). However, the precise papilla cell-specific expression and localization of SSP suggests that it may have a key role in stigma function. One obvious role for SSP, and other stigma peroxidases, could be in some aspect of the pollen–stigma interaction, for instance in loosening stigma cell wall components to allow penetration and growth of pollen tubes within the stigma. Alternatively, acting indirectly through H2O2 metabolism, SSP and other stigma peroxidases may form the components of signalling systems activated during the pollen–stigma interaction that could mediate, for instance, species-specific pollen recognition. Another possibility is that stigmatic peroxidases may be involved in defence against pathogen attack, as increased expression of peroxidase genes is known to be associated with the hypersensitive response and stress (Mohan et al., 1993; Harrison et al., 1995; Gadea et al., 1996; Hiraga et al., 2001; Cheong et al., 2002; Delannoy et al., 2003; Do et al., 2003). It has been shown that pistils express a number of pathogenesis-related (PR) proteins constitutively, some of which are peroxidases (Curtis et al., 1997), so it is possible that the high constitutive levels of peroxidase activity in stigmas (and style) may contribute to enhanced protection against pathogen attack when the pistil is ‘primed’ to receive pollen.
Interestingly there is growing evidence that ROS and H2O2 are important in protecting nectar from pathogen attack (Carter and Thornburg, 2004). ROS and H2O2 have been shown to accumulate to very high levels in nectar and a superoxide dismutase, Nectarin 1, contained in nectar, is thought to be responsible for generating the H2O2 (Carter and Thornburg, 2000). Like nectar, the nutrient-rich secretions of stigmas are a potentially attractive environment for microbes, but like nectar, the receptive surfaces of stigmas rarely, if ever, experience microbe attack. Perhaps by analogy with nectar therefore, stigma secretions are protected from microbial infection by ROS and H2O2. Stigma peroxidases, such as SSP, may well be important elements of such a ROS/H2O2-based system of defence. If such a defence system is responsible for protecting stigmas from microbial attack, compatible pollen must either be immune to the toxic effects of high levels of ROS/H2O2 or able to down-regulate its production in some way. At present, there can only be speculation about a possible functional link between stigmatic peroxidases and high levels of stigmatic ROS/H2O2, but these findings in Senecio do highlight some important new avenues of research on the role of ROS and ROS-related enzymes in angiosperm reproduction, specifically in pollen–stigma interactions and defence by stigmas against pathogens.
Thanks are due to Anthony Clarke and Richard Sessions, Department of Biochemistry, University of Bristol for producing the structural model of SSP, Keith Stobart and Tom Fraser for help with the enzyme assays, Tim Colborn for help preparing the figures, and Mike Ames for maintaining the plants. This work was supported by the BBSRC and a pump-priming grant from the Faculty of Science, University of Bristol which funded SMM for six months.