TGases (transglutaminases) are a class of calcium-dependent enzymes that catalyse the interactions between acyl acceptor glutamyl residues and amine donors, potentially making cross-links between proteins. To assess the activity of apple (Malus domestica) pollen TGase on the functional properties of actin and tubulin, TGase was prepared from apple pollen by hydrophobic- interaction chromatography and assayed on actin and tubulin purified from the same cell type. The enzyme catalysed the incorporation of putrescine into the cytoskeleton monomers. When tested on actin filaments, pollen TGase induced the formation of high-molecular-mass aggregates of actin. Use of fluorescein–cadaverine showed that the labelled polyamine was incorporated into actin by pollen TGase, similar to with guinea pig liver TGase. The pollen TGase also reduced the enzyme activity and the binding of myosin to TGase-treated actin filaments. Polymerization of tubulin in the presence of pollen TGase also yielded the formation of high-molecular-mass aggregates. Furthermore, the pollen TGase also affected the binding of kinesin to microtubules and reduced the motility of microtubules along kinesin-coated slides. These results indicate that the pollen TGase can control different properties of the pollen tube cytoskeleton (including the ability of actin and tubulin to assemble and their interaction with motor proteins) and consequently regulate the development of pollen tubes.
- actin filament
- plant transglutaminase
- pollen tube
The pollen is the male gametophyte of seed plants and its tube transports the male gametes toward the egg cell. The growth of pollen tubes is a complex process that needs the simultaneous activity of many different molecules working in the apical domain . The movement of the cytoplasmic content of pollen tubes is dependent on the pollen tube cytoskeleton , which consists essentially of microtubules, actin filaments and related proteins organized to facilitate and promote unidirectional cell expansion and the transport of sperm cells. Actin filaments are generally widespread in the pollen tube cytoplasm but are absent in the very tip; a consistent “actin fringe” is present immediately behind the tip and is thought to delimit the growth area . On the other hand, microtubules are preferentially organized as bundles in the cortical cytoplasm and are absent in the last 10–20 μm of the tube apex . Actin filaments are mainly involved in organelle movement and in tube growth , whereas microtubules may play a role in the transport of sperm cells , probably participate in the regulation of organelle movement  and might also guide pollen tube growth in vivo or have a role in directing the elongation activity .
Tubulin is post-translationally modified to perform its multitude of functions during pollen tube growth . On the other hand, post-translational modifications of actin are poorly known in eukaryotic cells . An important form of post-translational modification is the conjugation of polyamines, which is mediated by calcium-dependent TGases (transglutaminases; E.C. 18.104.22.168). TGases catalyse the interaction between an acyl acceptor glutamyl residue and amine donors, like lysyl residues, thus forming cross-links within the same or different proteins. Polyamines also act as physiological substrates of TGases: their terminal amino-group binds one or two glutamyl residues, giving rise to either mono-(γ-glutamyl)-polyamines or bis-(γ-glutamyl)-polyamines [10,11]. Consequently, the ‘bridge’ between two proteins can have different lengths: very short for the Lys–Gln binding, and gradually longer for polyamines with different backbone length. Moreover, polyamines exhibit positive charge and thus ‘cationise’ the partner protein. TGases form bridges between specific proteins, including cytoskeleton proteins or extracellular matrix proteins, and are probably involved in the regulation of cell growth and differentiation of animal and plant cells . TGases are also defined as “biological glues” for their capacity to join proteins . Previous results from animal cells suggested the involvement of TGase in the modification of cytoskeleton proteins. After early indications that actin is a substrate of TGase , actin in Physarum was recognized as such . Recently, the cytoskeleton substrates of TGase 2, one of the nine TGases identified in animals, were summarized by Robinson et al. . Polyamines are also involved in the rearrangement of actin and tubulin cytoskeleton as demonstrated by Pohjanpelto et al.  in CHO (Chinese hamster ovary) dividing cells.
The crude extract of apple (Malus domestica) pollen contained unpurified TGase activity and catalysed the incorporation of polyamines into proteins with molecular masses of 43 kDa and 52–58 kDa . These bands matched immunolabelled spots identified by mouse monoclonal antibodies to actin and α-tubulin. The supply of exogenous actin and tubulin in cell-free extracts of pollen enhanced the TGase activity. Moreover, actin of germinating pollen gave rise to cross-linked products of high molecular mass .
The present study was undertaken to analyse the involvement of purified pollen TGase in the post-translational modification of pollen actin and tubulin via Gln–Lys or polyamine bridges. Pollen from Malus domestica was chosen as the starting material because apple pollen was already shown to catalyse the incorporation of polyamines (therefore it presumably contains a TGase activity) . This work also assayed the polymerization/depolymerization properties of actin and tubulin after modification by TGase. This possibility was studied by supplying two different amines, FITC–cadaverine (for labelling the substrates), with only one amino group available for conjugation, and putrescine, with both the primary amino groups available for linkages and possible involvement in cross-bridge formation. We also assessed the TGase-induced modifications of actin/tubulin in connection with their related motor proteins, myosin and kinesin. Since the proper dynamic of actin filaments and microtubules, coupled to the activity of motor proteins, is responsible for organelle movement in pollen tubes , TGase may consequently be recognized to take part in the control of many active processes occurring during pollen tube growth.
MATERIALS AND METHODS
Mature pollen of M. domestica Borkh. cv. Red Chief was collected from experimental plots (Dipartimento di Colture Arboree, Bologna, Italy; Azienda Agricola Rondinini, Faenza, Italy). Handling and storage were performed as described by Bagni et al. .
Antibodies and chemicals
Rabbit antibody to plant actin was a generous gift of Dr. Richard Cyr (Department of Biology, Pennsylvania State University, PA, U.S.A.). Donkey polyclonal antibody to tubulin was purchased from Cytoskeleton (code number ATN02). Unless otherwise indicated, chemicals used in this work were purchased from Sigma–Aldrich. gplTGase (guinea pig liver TGase) was obtained from Sigma–Aldrich. The ID10 anti-TGase monoclonal antibody was raised against purified gplTGase; clones were screened out against gplTGase, ion exchange-purified extracts of 14-days old Pisum sativum leaves and against germinated Vicia faba cotyledons (used 14 days after imbibitions). ELISA-positive fractions were pooled, concentrated in a dialysis tube against PEG 8000 and stored at −20 °C. The specificity of ID10 for pollen TGase was tested against partially purified enzyme fractions and was compared with that of the well-known TGase antibody AbIII (Neomarker, Fremont, CA, U.S.A.) already assayed against plant  and pollen  TGases; the two antibodies showed the same immunorecognition profile.
Preparation of a pollen TGase-enriched fraction
Pollen was rehydrated at 30 °C with 100% relative humidity for 30 min. All purification steps were conducted at 4 °C. One gram of hydrated pollen was extracted using a Potter-Elvehjem tissue homogenizer with 100 ml of 25 mM Tris buffer, pH 8.0, 2 mM DTT (dithiothreitol), 0.1 mM PMSF and 1 mM EDTA. The extract was centrifuged at 20400 g for 10 min and the supernatant was dialysed overnight against 5 litres of dialysis buffer (5 mM Tris buffer, pH 8.0, and 1 mM 2-mercaptoethanol) with three changes of buffer. The dialysed sample was supplemented with 5 mM CaCl2 and centrifuged at 20400 g for 15 min to eliminate the calcium-induced protein precipitation. The supernatant was loaded on a Phenyl Sepharose HIC (hydrophobic-interaction chromatography) column (1.6×25 cm, Sigma–Aldrich), and equilibrated with 20 mM Tris buffer, pH 8.0, containing 5 mM CaCl2 and 1 mM 2-mercaptoethanol. Elution was performed at 2.5 ml/min with an isocratic flow of the same buffer containing EGTA instead of CaCl2, followed by two column washings with distilled water and ethylene glycol (50% in water) respectively. Samples were dialysed overnight against 1000 volumes of 5 mM Tris buffer, pH 8.0, and 1 mM 2-mercaptoethanol with three changes of buffer, then freeze-dried and stored at −80 °C. Samples were diluted in 50 mM Tris buffer, pH 8.0, containing 1.5 mM DTT and incubated for 5 min at room temperature (23 °C) before assaying the TGase activity.
Incorporation assay of biotin-labelled cadaverine by TGase
Dialysed fractions from the HIC and the pollen extract were tested by the biotin-cadaverine incorporation assay according to the method of Lilley et al.  with EGTA replacing 10 mM EDTA in negative controls.
Purification of actin from apple pollen
Purification of actin from apple pollen was performed as described in . Hydrated apple pollen (1 g) was germinated for 3 h in the growth medium (0.2 M sucrose, 0.33 mM H2BO3 and 1.3 mM Ca2NO3) at a concentration of 1 mg/ml. After germination, pollen was collected by centrifugation at 1000 g for 10 min. The pellet was washed with washing buffer (2 mM 2-mercaptoethanol, 0.005% NaN3, 0.2 mM CaCl2 and 2 mM Tris, pH 8.0) and the final pellet was suspended in 5 ml of the lysis buffer 1 (buffer A: 2 mM 2-mercaptoethanol, 0.2 mM ATP, 0.005% NaN3, 0.2 mM CaCl2, 2 mM Tris, pH 8.0, 1 mM PMSF and plant protease inhibitor cocktail from Sigma; code P9599). All purification steps were conducted at 4 °C unless otherwise stated. Cells were lysed with a Potter-Elvehjem Tissue Homogenizer. The homogenate was centrifuged at 10000 g for 20 min. The supernatant was then centrifuged at 100000 g for 30 min. The pellet was discarded, while the supernatant was precipitated with 2.2 M (NH4)2SO4 under agitation. After 10 min, the sample was centrifuged at 10000 g for 10 min. The resulting pellet was dissolved in 2 ml of buffer B (0.1 mM CaCl2, 2 mM β-mercaptoethanol, 10 mM imidazole pH 7.5, 0.5 mM ATP, 0.05% NaN3 and 1 mM PMSF) and dialysed overnight against buffer B. The sample was fractionated using a Mono-Q HR 5/5 column (GE HealthCare), equilibrated in buffer B over 20 column volumes (19.6 ml) with a linear gradient from 0 to 1 M KCl; elution was performed at 1 ml/min at room temperature. Fractions were analysed by spot-test with a plant actin antibody. Cross-reacting fractions were pooled and supplemented with 2 mM MgCl2. After a 10 h incubation at 4 °C to stimulate actin polymerization, the sample was centrifuged at 160000 g for 1 h 45 min. The pellet was dissolved in 300 μl of buffer C (2 mM 2-mercaptoethanol, 0.2 mM ATP, 0.005% NaN3, 0.2 mM CaCl2, 2 mM Tris, pH 8.0, and 1 mM PMSF) and dialysed overnight against buffer C using the 8 kDa Cut-Off Mini Dialysis kit (GE HealthCare). The sample was fractionated onto a Superdex 200 HR 10/30 column (GE HealthCare) and equilibrated with buffer C at room temperature with a flow rate of 0.8 ml/min. Chromatography steps were performed using an AKTA Purifier system (GE HealthCare). Fractions were assayed by spot-test with the actin antibody and positive fractions were pooled, frozen in liquid nitrogen and stored at −80 °C.
Purification of tubulin from apple pollen
Tubulin from apple pollen was purified as described in . Apple pollen (1 g) was germinated as described above and homogenized in 5 ml of lysis buffer 2 containing 50 mM Pipes, pH 6.8, 2 mM MgCl2, 2 mM EGTA, 1 mM DTT, plant protease inhibitor cocktail from Sigma and 0.1 mM GTP. All steps were carried out at 4 °C unless otherwise mentioned. After high-speed centrifugation, the supernatant was fractionated onto a Resource Q anion-exchange column (1×1 ml, GE HealthCare) equilibrated in buffer D (50 mM Pipes/KOH, pH 6.8, 2 mM MgCl2, 2 mM EGTA, 1 mM DTT, 1 mM PMSF and 0.1 mM GTP). The sample was eluted over 20 column volumes (20 ml) from 0 to 1 M KCl at room temperature with a flow rate of 1 ml/min. Fractions were analysed by spot-test with tubulin antibodies and positive fractions were pooled. The resulting 4 ml sample was loaded onto a HiPrep 16/60 Sephacryl S-300 column (GE HealthCare) equilibrated with buffer D at room temperature; elution was achieved at 1 ml/min. Fractions were again assayed by tubulin antibody. Cross-reacting fractions were pooled and applied to a Mono-Q HR 5/5 column equilibrated with buffer D. Protein separation was achieved over 20 column volumes (19.6 ml) from 0 to 1 M KCl at room temperature, with a flow rate of 1 ml/min. Fractions positive to the spot-test analysis were dialysed for several hours in Mini Dialysis kit 8 kDa cut-off against buffer D. Chromatography steps were performed using a AKTA Purifier system (GE HealthCare). The sample was frozen under liquid nitrogen and stored at −80 °C.
TGase radiometric assay
The in vitro reactivity of TGase with cytoskeleton proteins was performed as previously described by Serafini-Fracassini et al. , but modified as follows. The incubation mixture included 50 μl of pollen actin or tubulin (0.5 mg/ml), 15 μl of 500 mM Tris buffer, pH 8, 2 μl of TGase active fraction (2 mg/ml) and final concentrations of the following: 0.1 mM putrescine, 10 mM DTT and 2.5 mM CaCl2. As radioactive tracer, 444 kBq [1,4(n)-3H]-putrescine (0.55 TBq mmol−1; GE HealthCare) was added to the assay mixture to a total volume of 150 μl. After 120 min of incubation at 30 °C, the reaction was stopped and repeatedly pelleted with 5% (w/v) trichloroacetic acid as described  and the radioactivity was counted. Another aliquot was used for the γ-glutamyl polyamine derivatives analysis (see below). Enzyme reactions were also carried out in the presence of the commercial TGase inhibitor 1,3-dimethyl-2-[(2-oxopropyl)thio] imidazolium chloride (Zedira).
HPLC identification of γ-glutamyl polyamine derivatives
The incorporation of labelled putrescine into TGase substrate proteins was evaluated in the trichloroacetic acid-insoluble fractions as mono- and bis-γ-glutamyl polyamine derivatives. The final pellet was washed at least three times with anhydrous diethyl ether and proteolysed according to the method of Folk et al. . The digestion products were separated by ion-exchange chromatography using a Jasco HPLC system (4.5×90 mm column, packed with Ultropac 8 resin, Na+ form; Jasco Europe, Milan, Italy) and the five-buffer system previously described by Folk et al. . The identity of the polyamine derivatives was determined by comparison with the corresponding retention times of glutamyl polyamine standards. The identity of conjugated polyamines was determined after release of free polyamines by acid hydrolysis (6 M HCl) of the ion-exchange chromatographic fractions corresponding to the predicted retention times .
Assay of pollen TGase activity on actin filaments
In order to determine the effects of pollen TGase on cytoskeleton proteins, actin (Cytoskeleton; AKL-99) was diluted to 0.4 mg/ml in A-buffer (5 mM Tris/HCl, pH 8.0, 0.2 mM ATP and 0.2 mM CaCl2) and incubated for 1 h on ice. Actin was added to the polymerization buffer (0.5 M KCl, pH 8.0, 10 mM ATP and 20 mM MgCl2) and incubated for 1 h at room temperature. Actin and pollen TGase were mixed at a ratio of 10:1 in the presence of 1 mM FITC-labelled cadaverine and 7 mM CaCl2. All samples were incubated for 3 h at 37 °C. After incubation, 10 μl aliquots were perfused within a perfusion chamber and further incubated for 5 min. After brief staining with rhodamine–phalloidin buffer (10 μl of 10×polymerization buffer, 25 μl of 6.6 μM rhodamine–phalloidin stock from Invitrogen and 65 μl of water), samples were observed with a fluorescence microscope. Enzyme reaction of the remaining samples was stopped by the addition of 7 mM EGTA. Samples were centrifuged at 15000 g for 1 h at 25 °C and supernatants and pellets were processed for SDS/PAGE. Similar analysis was performed with rhodamine-labelled actin (Cytoskeleton, code number AR05); in this case, the rhodamine–phalloidin buffer was not applied.
Optical microscopy analysis of the incorporation of FITC–cadaverine into actin filaments
Samples were observed with a Zeiss Axiophot fluorescence microscope, equipped with a 63× objective. Images were captured with an MRc5 AxioCam video camera using AxioVision software. Samples were alternatively analysed with a Bio-Rad MicroRadiance Confocal microscope using a 63× objective.
Binding and enzyme assay of myosin with TGase-treated actin filaments
Actin filaments were prepared and treated with pollen TGase as described above. Myosin (Cytoskeleton, code MHO1) was diluted to 1 mg/ml in 4 mM Tris/HCl, pH 7.6, 2 mM MgCl2 and 100 mM KCl, and added to pre-treated actin filaments. After incubation for 30 min at 25 °C, samples were centrifuged at 15000 g for 1 h at 25 °C. Supernatants and pellets were processed for SDS/PAGE.
For the myosin enzyme assay, 10, 20 and 50 μl of actin filaments were added to myosin in the ATPase assay buffer (10 mM KCl, 1 mM EGTA, 2 mM MgCl2, 5 mM ATP and 20 mM Pipes, pH 7.0). After incubation at 25 °C for 5, 10, 20, 25 and 30 min, the amount of free inorganic phosphate was measured using the PhosFree kit from Cytoskeleton. Controls were performed with gplTGase (same amount of pollen TGase).
Polymerization/depolymerization assay of tubulin in the presence of pollen TGase
Tubulin aliquots (10 mg/ml, from Cytoskeleton, rhodamine-labelled, code TL331M, or unlabelled, code TL238) were diluted with buffer E (80 mM Hepes, pH 7.5, 1 mM EGTA, 1 mM MgCl2, 2 mM GTP and 20% glycerol) at 5 mg/ml. Tubulin samples were mixed with pollen TGase (10:1 ratio), 5 mM FITC–cadaverine and 7 mM CaCl2. Samples were incubated at 37 °C for 3 h, and then observed at the optical microscope [DIC (differential interference contrast) view, see below]. Enzyme reaction was blocked by adding 7 mM EGTA. Samples were centrifuged at 15000 g for 1 h at 25 °C, and supernatants were prepared for SDS/PAGE. Pellets were suspended in PEM buffer (80 mM Pipes, pH 6.8, 1 mM EGTA and 1 mM MgCl2); part of this was prepared for SDS/PAGE and the remainder was incubated for 30 min at 4 °C (depolymerization phase). Afterwards, samples were centrifuged at 15000 g for 1 h at 4 °C and supernatants and pellets were prepared for SDS/PAGE.
DIC microscopy of TGase-treated microtubules
Samples of treated and untreated microtubules were perfused into perfusion chambers and incubated for 5 min. Slides were washed with PEM buffer and observed at the optical microscope in DIC view to assess the effects of TGase-mediated incorporation of FITC–cadaverine on the microtubule structure. Samples were also observed in the rhodamine channel to prove that filamentous structures were definitely microtubules and in the FITC channel to visualize the incorporation of cadaverine.
Binding assay of kinesin to TGase-treated microtubules
Kinesin from Cytoskeleton (code number KR01, 0.1 mg/ml) was added to microtubules (treated and untreated with pollen TGase) prepared as described above. After incubation for 30 min at 25 °C, samples were centrifuged at 15000 g for 1 h at 25 °C. Supernatants were prepared for SDS/PAGE. Pellets were washed with ATP buffer (80 mM Pipes, pH 6.8, 7 mM EGTA, 1 mM MgCl2, 1 mM DTT, 20 μM taxol and 10 mM ATP); parts of the pellets were prepared for electrophoresis and the remaining part was incubated for 30 min at room temperature. Samples were centrifuged at 15000 g for 1 h at 25 °C and then prepared for electrophoresis.
in vitro motility assays
Kinesin solution (10 μl) from Cytoskeleton (0.1 mg/ml) was incubated for 5 min in perfusion chambers. Microtubules (treated or untreated with pollen TGase) were diluted 1:5 with motility buffer (80 mM Hepes, pH 7.5, 7 mM EGTA, 1 mM MgCl2, 1 mM DTT, 20 μM taxol and 5 mM ATP). Then, 10 μl of microtubule solution was incubated for 1 min in the perfusion chamber. Samples were observed at the optical microscope under DIC view. Motility of microtubules was recorded as described in .
Incorporation of putrescine into TGase-treated microtubules
Tubulin and microtubules were prepared as described above and adjusted to a concentration of 5 mg/ml. Pollen TGase was added to the ratio of 1:10 with tubulin supplemented with 7 mM CaCl2 and 0.1 or 1 mM putrescine. Samples were incubated at 37 °C for 3 h and then observed with DIC microscopy. The enzyme reaction was stopped by addition of 7 mM EGTA. Samples were centrifuged at 15000 g for 1 h at 25 °C. Supernatants and pellets were prepared for SDS/PAGE.
Electrophoresis and immunoblot
Electrophoresis of proteins was performed according to the method of Laemmli . Molecular-mass standards were from Sigma–Aldrich or from Bio-Rad (Precision series). Gels were stained with Bio-Safe Coomassie Blue (Bio-Rad). Immunoblot was performed according to Towbin et al. . Secondary antibodies against mouse and rabbit IgG were conjugated with horseradish peroxidase (GE HealthCare); secondary antibody to donkey IgG was conjugated with horseradish peroxidase (Cytoskeleton). Immunoreactive bands were detected with ECL® (enhanced chemiluminescence) Plus reagents from GE Healthcare.
The protein content was determined using a modified bicinchoninic acid method . BSA was used as the standard protein.
Isolation and characterization of a TGase-enriched fraction from apple pollen
A TGase activity-enriched fraction from apple pollen was eluted by HIC with 10 mM EGTA (Figure 1A) and contained a prominent band at 70 kDa (Figure 1B, lane 1), which catalysed transamidating reactions (see below). The EGTA-eluted fractions were enriched in TGase activity approx. 10-fold as shown by comparison with the HIC-unfractionated sample (Figure 1A). The 70 kDa polypeptide cross-reacted with two monoclonal TGase antibodies, ID10 and AbIII (Figure 1B, lanes 2 and 3). The cross-reactivity of ID10 was also evaluated on gplTGase and apple pollen extracts (lanes 4 and 5), while AbIII was tested on pollen extract (lane 6). Both antibodies cross-reacted with the 70 kDa band in pollen extracts.
Incorporation of polyamines in apple tubulin and actin
To assess the enzyme activity of pollen TGase on tubulin and actin, the latter two proteins were first isolated from apple pollen. Tubulin was purified by chromatography (Figure 2A, top panel), yielding about 150 μg of homogeneous tubulin (lane 4). After treatment with taxol, the protein was mostly found in the pellet (lane 6) and partly in the supernatant (lane 5), indicating that apple tubulin was competent at forming microtubules. The position of tubulin was monitored using a specific antibody (Figure 2A, lower panel).
Actin was isolated by precipitation with (NH4)2SO4 and by chromatography. Approx. 250 μg of actin was purified to homogeneity from apple pollen (Figure 2B, upper panel). After the Mono-Q step (lane 4), actin was still contaminated by additional proteins, whose level was reduced by the polymerization/depolymerization step (lane 5 and 6). The use of Superdex 200 column greatly enhanced the purity of actin (lane 7). Actin was monitored with a specific antibody (Figure 2B, lower panel). Figures 2(C) and 2(D) show the elution profiles of the digested trichloroacetic acid-insoluble fractions of apple tubulin (Figure 2C) and actin (Figure 2D) after incubation with the pollen TGase activity-enriched fraction and [3H]-putrescine. Incubation of tubulin and actin with 0.1 mM putrescine led to the recovery of bis-(γ-glutamyl)-putrescine and mono-(γ-glutamyl) derivatives, the free putrescine being the internal standard. The presence of the TGase inhibitor 1,3-dimethyl-2-[(2-oxopropyl)thio] imidazolium chloride in the assay drastically reduced the formation of γ-glutamyl-putrescine, whereas the absence of the TGase substrates (tubulin and actin) did not allow the production of derivatives.
Effects of pollen TGase on actin filaments
Because mammalian actin/tubulin reacted with pollen TGase comparably with pollen actin/tubulin and because of their availability in larger quantities, rabbit muscle actin and bovine brain tubulin were used in all subsequent assays. The effects of pollen TGase on actin filaments was first investigated by analysing the sedimentation properties of in vitro polymerized actin (Figure 3). Pollen TGase (Figure 3A, lane 1, arrow) and FITC–cadaverine were incubated with filamentous actin (Figure 3A, lane 3) polymerized from rabbit G-actin (Figure 3A, lane 2). As a control, F-actin was also examined in the absence of pollen TGase (Figure 3A, lane 4) but in the presence of FITC–cadaverine. The supernatant obtained in the presence of pollen TGase contained a small amount of actin (Figure 3A, lane 5), whereas the corresponding pellet showed the presence of large amounts of actin in addition to high-molecular-mass bands (Figure 3A, lane 6, arrowheads) showing molecular masses of 93, 133 and 155 kDa. The volumetric measure of those bands indicated that the ratio was 6:4:1. In the absence of pollen TGase, the supernatant (Figure 3A, lane 7) contained only a small amount of actin, and the corresponding pellet (Figure 3A, lane 8) did not reveal the presence of additional bands other than actin. An actin antibody (Figure 3B) reacted with actin at 43 kDa, but also with the high-molecular-mass polypeptides (Figure 3B, lane 6). When fractions were tested with the ID10 antibody in Figure 3(C), the pollen TGase was found in lane 1 (containing the enzyme only), in lane 3 (enzyme plus actin) and in the sample containing the actin-related high-molecular-mass aggregates (Figure 3C, lane 6). The pollen TGase was not found in the high molecular mass aggregates. As a control, the same assay performed with gplTGase and FITC–cadaverine yielded results similar to those obtained with pollen TGase (Supplementary Figure S1 at http://www.BiochemJ.org/bj/418/bj4180651add.htm).
To demonstrate that FITC–cadaverine bound to actin filaments, the pellet obtained in the presence/absence of pollen TGase was observed under fluorescence microscopy in the rhodamine channel (for visualizing actin) and in the FITC channel (for visualizing FITC–cadaverine). The sample in lane 6 of Figure 3(A) showed the presence of a conspicuous network of actin filaments (Figure 4A), which was also detected in the FITC-channel (Figure 4B). In the absence of pollen TGase (lane 8 of Figure 3A), we still observed the network of fluorescent actin filaments (Figure 4C), but no signal was detected in the FITC channel (Figure 4D), apart from random bright spots. This result indicated that FITC–cadaverine did not bind to actin in the absence of pollen TGase. The control in the presence of gplTGase was almost comparable with that of pollen TGase (Supplementary Figure S2 at http://www.BiochemJ.org/bj/418/bj4180651add.htm).
We also analysed the enzyme activity of myosin and its binding to actin filaments. First, the ATPase activity of rabbit myosin was tested in the presence of different concentrations of F-actin (10, 20 and 50 μg/ml), pre-treated (+TGase) and untreated with either pollen or gplTGase (Figure 5A). Actin alone showed no significant ATPase activity, neither did actin in combination with either pollen TGase or gplTGase. The ATPase activity of myosin alone was slightly detectable. In the absence of TGase, the ATPase activity of myosin increased with increasing concentrations of actin. An increasing trend was also observed with TGase-treated actin; however, the increase in ATPase activity was greatly reduced when compared to untreated actin.
To visualize the binding of myosin to actin, we used a sedimentation assay. In the absence of pollen TGase (Figure 5B), myosin bound almost completely to actin filaments (lane 2) and was not found in the supernatant (lane 1). When pollen TGase-treated actin was used, part of myosin (around 40%) was recovered in the supernatant (lane 3) and the remaining 60% in the pellet (lane 4). A comparable result was obtained with gplTGase (Supplementary Figure S3 at http://www.BiochemJ.org/bj/418/bj4180651add.htm).
Effects of pollen TGase on microtubules
We also analysed the effects of pollen TGase on the assembly/disassembly of microtubules at different temperatures. Pollen TGase (Figure 6A, lane 2) was mixed with rhodamine-labelled or unlabelled tubulin (lane 3), which was polymerized at 37 °C in the presence of FITC–cadaverine. In the absence of TGase, tubulin was partly found in the supernatant S1 (approx. 30%, lane 4) but more consistently in the pellet P1 (approx. 70%, lane 5). When Ca2+-free buffers were used (80 mM Pipes, pH 6.8, 1 mM MgCl2 and 1 mM EGTA), the percentage of tubulin in the pellet increased to 90% (results not shown). Around 55% of pollen TGase-treated tubulin was found in the supernatant S1 (lane 6), whereas the corresponding pellet P1 contained 45% of initial tubulin and additional high-molecular-mass bands (arrowheads) at 106, 121, 143 and 165 kDa (the last being the most abundant), plus a very heavy band in the gel top. When the pellets of lanes 5 and 7 were depolymerized, the TGase-untreated sample yielded a supernatant (S2, lane 8) enriched in tubulin, approx. 73% of total tubulin compared with the pellet P2 (lane 9). When TGase-treated tubulin was depolymerized, the supernatant S2 (lane 10) contained low levels of tubulin (32% of total tubulin) in comparison with the pellet P2 (lane 11); the pellet also contained the high-molecular-mass aggregates (arrowheads). The presence of tubulin in the high-molecular-mass aggregates was shown by immunoblot with the tubulin antibody (Figure 6B). Pollen TGase was detected by immunoblot in the tubulin pellet containing the high-molecular- mass aggregates (Figure 6B, lane 7). Although the pollen TGase remained in the P2 pellet after depolymerization (Figure 7B, lane 11), it was not found in association with the high-molecular-mass aggregates.
The microtubule pellets of lanes 5 and 7 in Figure 6(A) were observed with DIC microscopy. TGase-treated microtubules were observed as bundles from which single microtubules radiated (Figure 6D, left panel). The identity of microtubules was confirmed by fluorescence microscopy in the rhodamine channel (Figure 6D, central panel). The co-distribution of FITC–cadaverine with microtubules was evaluated in the FITC channel (Figure 6D, right panel). As a control, the pellet of lane 5 (TGase-untreated microtubules) was observed as single filaments, but not as bundles (Figure 6E). When unlabelled tubulin was used, fluorescence in the rhodamine channel was not detected, but it was detected in the FITC channel.
We investigated the effects of TGase and FITC–cadaverine on the binding of kinesin to microtubules. Tubulin (Figure 7A, lane 2) was polymerized into microtubules (Figure 7A, lane 4) in the presence of pollen TGase (Figure 7A, lane 3) and the mixture (Figure 7A, lane 5) was used for the binding assays. In controls, nearly all of kinesin pelleted with untreated microtubules (Figure 7A, lane 7) compared with the corresponding supernatant (Figure 7A, lane 6). When the pellet was washed with ATP (Figure 7A, release step), kinesin was found essentially in the supernatant (Figure 7A, lane 9) but not in the pellet (Figure 7A, lane 8). Kinesin bound only partially (approx. 60%) to TGase-treated-microtubules (Figure 7A, lane 11), compared with soluble kinesin (Figure 7A, lane 10, approx. 40%). After the ATP-releasing step, all kinesin was found in the supernatant (Figure 7A, lane 13) and not in the microtubule pellet (Figure 7A, lane 12). Pollen TGase was immunostained with ID10 (Figure 7B).
The effect of pollen TGase on the kinesin-mediated microtubule motility was analysed by an in vitro gliding assay. In the control (Figure 8A), untreated microtubules (arrows and arrowheads) moved with a speed of 0.4 to 0.6 μm/s on kinesin-coated glass slides. The mean velocity of TGase-treated microtubules was significantly lower (0.2–0.3 μm/s, Figure 8B) and we frequently observed presumptive microtubule bundles (black arrowheads) moving along kinesin-coated slides.
The ability of pollen TGase to form polyamine-mediated bridges between tubulin molecules was tested by mixing tubulin with pollen TGase and putrescine (Figure 9). Bovine brain tubulin (Figure 9A, lane 1) was polymerized in the presence of pollen TGase (Figure 9A, lane 2) and either 0.1 or 1 mM putrescine and centrifuged. In the absence of putrescine, the supernatant (Figure 9A, lane 3) contained small amounts of tubulin, whereas the pellet (Figure 9A, lane 4) showed high-molecular-mass aggregates (arrowheads) at 97, 106, 150 (the most abundant) and 258 kDa. In the presence of 1 mM putrescine, the supernatant contained only tubulin monomers (Figure 9A, lane 5), whereas the corresponding pellet showed larger amounts of the high-molecular-mass aggregates (Figure 9A, lane 6). When pollen TGase was omitted, the supernatant (Figure 9A, lane 7) showed essentially tubulin monomers, whereas the corresponding pellet (Figure 9A, lane 8) contained tubulin but not additional high-molecular-mass aggregates. Using DIC microscopy, the microtubule pellet of lane 4 (Figure 9A) showed the previously described bundles of microtubules (Figure 9B). Microtubules in the presence of pollen TGase and 1 mM putrescine (lane 6 of Figure 9A) appeared as amorphous structures without the typical filamentous aspect of microtubules (Figure 9C). When microtubules were incubated with putrescine but without TGase (lane 8 of Figure 9A), they were indistinguishable from untreated microtubules (results not shown). The same experiment was performed in the presence of 0.1 mM putrescine (Figure 9D). Arrows indicate the presence of lower amounts of high-molecular-mass aggregates. Using DIC microscopy, microtubules treated with TGase in the absence of putrescine (lane 6 in Figure 9D) appeared as bundles (Figure 9E). In contrast, the microtubule pellet after treatment with pollen TGase and 0.1 mM putrescine (lane 4 in Figure 9D) showed the presence of aberrant microtubule structures (Figure 9F).
Using the Ca2+-binding property of TGases, a 70 kDa polypeptide was identified in apple pollen extracts (Figure 1B), which showed molecular and immunological characteristics very similar to those of the well known TGase-2, which is widely expressed in many animal cells. The pollen protein shared typical functions of TGases, as shown by its capacity to catalyse the formation of glutamyl-polyamine derivatives, by its inhibition with a site-directed TGase inhibitor and by the comparative analysis of its products (Figures 2C and 2D). The pollen TGase recognized actin and tubulin, catalysing the formation of high-molecular-mass aggregates (Figures 3A and 6A). In addition, pollen TGase influenced the binding of myosin and kinesin to actin filaments (Figure 5B) and microtubules (Figure 7) respectively. Pollen TGase can also reduce the enzyme activity of myosin (Figure 5A) and the gliding activity of kinesin (Figure 8B). Taken together, these findings showed that pollen TGase potentially acts as regulator of the cytoskeleton activity in pollen tubes.
Pollen TGase induces the formation of high-molecular-mass aggregates of actin
Regulation of actin filament dynamics by post-translational modifications is poorly documented in plant cells. Evidence that actin is a substrate of the partially purified 70 kDa pollen TGase (Figures 2D and 3A) is in agreement with preliminary data showing that pollen actin was modified with polyamines after incubation with apple pollen crude extracts  to produce actin aggregates similar to those catalysed by TGase from bacteria , animals and Physarum polycephalum . In animals the formation of aggregates was attributed to intermolecular (and intramolecular) linkages by Gln–Lys cross-links. The intermolecular bond between Gln41 and Lys113 observed by linking ANP [N-(4-azido-2-nitrophenyl)-putrescine] makes G-actin more resistant towards proteolytic degradation and reduces the depolymerization of F-actin . In addition, the occurrence of intramolecular cross-linking between Gln41 and Lys50 in rabbit muscle actin was followed by stabilization of F-actin structure . Reports on animal or bacterial TGases indicated that Gln41 is the only glutamine residue on actin that reacts with primary derivatized amines such as ANP  and dansyl-ethylene-diammine  in the TGase-catalysed reaction  forming Gln–mono-derivatives. In the present paper, the FITC–cadaverine test (Figure 4) and the formation of radioactive mono-putrescine via unmodified radiolabelled polyamines (Figures 2C and 2D) suggested that the pollen TGase has identical activity. When unmodified polyamines were conjugated, a role in addition to cationization of the free polyamine amino group in F-actin stabilization can be discussed. F-actin-like filaments are reversibly formed when G-actin is incubated with spermine and spermidine at relatively low concentrations in the absence of TGase. This reversible polymerization was attributed to the formation of ionic linkages with polyanionic regions at the N-terminus of actin .
The formation of bis-putrescine derivatives by pollen TGase (Figures 2C and 2D) is a clear demonstration of cross-links between two glutamine residues. Assuming that Gln41 is the only residue modified by pollen TGase, the polyamine bis-derivatives should form intermolecular cross-links between Gln41 of different monomers or filaments. A single polyamine molecule can thus use its two terminal amino-groups to form either reversible ionic or irreversible covalent linkages with actin, the latter being mediated by TGase. Either excessive concentrations of polyamine or its absence disturbs the correct polymerization of actin in the presence of TGase; moreover, unordered bundles of actin appeared at high concentrations, suggesting that the actin binding sites are possibly saturated, giving rise to many mono-polyamine derivatives instead of forming the correct site-specific cross-links by bis-derivatives . The natural polyamine concentrations (estimated approx. 50 μM in hydrated pollen) should decrease the competition for the protein-binding sites, allowing bis-derivatives to form, as occurs with 100 μM putrescine (the optimal concentration used for the in vitro TGase assay reported here).
Pollen TGase affects the polymerization dynamic of microtubules
TGases were shown to catalyse the conjugation of putrescine to tubulin in animal embryos, generating high-molecular-mass cross-linkage products , as also occurred in cells committed to apoptosis  and in the human placenta . In plants, there are few studies on this topic; we previously showed that pollen tubulin was modified by polyamines after incubation with pollen crude extract, suggesting the occurrence of a catalysing TGase activity . Present results showed that pollen TGase changed the polymerization rate of tubulin in the presence of polyamines (Figure 6). Given that pollen TGase was never found in the actin and tubulin aggregates, its binding to aggregates should be transient or relatively weak to be broken by SDS. At 37 °C, TGase-treated tubulin assembled to a small extent into microtubules and, simultaneously, formed high-molecular-mass aggregates. Temperature-dependent depolymerization is decreased, suggesting that the conjugation of polyamines generated more stable microtubules (Figure 6). When the protein aggregates from the cadaverine (one binding site) and putrescine (two binding sites) assays were compared, some bands showed similar or identical molecular mass (i.e. 106 kDa); other aggregates appeared specific for the used polyamine, although the most abundant aggregates always showed a molecular mass of around 150 to 160 kDa (Figures 6 and 9). These aggregates are probably covalent dimers or trimers of tubulin produced by TGase activity. With DIC microscopy, the presence of relatively high concentrations (1 mM) of putrescine-generated amorphous structures with no filamentous appearance was observed (Figure 9). On the other hand, the presence of either the same concentration of FITC–cadaverine (having half the available binding sites) or TGase only yielded filamentous structures (presumably microtubule bundles). Putrescine at lower concentration (closer to the presumptive natural one) generated either filamentous or amorphous microtubule structures (Figure 6D). A reasonable hypothesis is that pollen TGase formed monomer aggregates at 1 mM putrescine concentration, which could slow down the polymerization rate of tubulin, thus explaining the low quantity of tubulin in the pellet. Rather than forming true microtubules, monomer aggregates could induce the formation of microtubule bundles, not necessarily formed only by covalent aggregates of tubulin but also by monomeric tubulin ‘entrapped’ into the aggregates.
Modification of actin and tubulin by pollen TGase influences the activity of myosin and kinesin
Actin filaments treated with pollen TGase bound less efficiently to myosin and enhanced the ATPase activity of myosin to a small extent, suggesting that pollen TGase had effects on the overall activity of the acto-myosin system (Figures 5A and 5B). Our findings are therefore in agreement with data showing that the cross-linking between Gln41 and Cys374 on adjacent monomers of actin filaments inhibits the acto-myosin motility and force generation [39,40]. In addition, the TGase-mediated intramolecular cross-linking between Gln41 and Lys50 disturbed the structure of G-actin but apparently had small effects on actin polymerization and on myosin interaction , or could even activate the ATPase activity of myosin subfragment 1 . Oriol-Audit  also reported that the polyamine-induced F-actin strongly triggered the Mg2+-ATPase activity of myosin.
In the present study, we showed that kinesin activity was reduced after polyamination of microtubules by pollen TGase. Specifically, kinesin bound with low affinity to TGase-modified microtubules but the ATP-releasing step was not affected (Figure 7A). The low binding affinity also affected the motor activity of kinesin; in vitro motility assays showed that TGase-modified microtubules moved with lower speed (at least 50% of reduction) compared with unmodified microtubules (Figures 8A and 8B). These results confirm that pollen TGase can also act as a ‘biological glue’ because of its capacity to cross-link proteins and to reduce their movement ability. These results are in agreement with the evidence that the motor activity of kinesin is also regulated by post-translational modification of tubulin. Members of the kinesin subfamily Kif5c moved predominantly along stable microtubules, which are mainly composed of detyrosinated tubulin . Other post-translational modifications that affected kinesin-based motility are glycylation, glutamylation and acetylation .
Hypothesis on the role of TGase during pollen tube growth
Results in the present paper suggest that the covalent cross-linking of polyamines by pollen TGase generates stable actin aggregates, which alter the dynamic properties of actin filaments and the interaction of myosin with actin. Consequently, pollen TGase may participate in actin regulation during pollen tube growth. Since the proper organization of the actin cytoskeleton is essential for tube growth and organelle movement , any disorganization of the motor apparatus generates critical changes in organelle motility and pollen tube growth. Given that Ca2+ is an important regulatory factor of TGase activity , local variation in Ca2+ concentration in the pollen tube may regulate the protein conformation and its enzymatic activity. We hypothesize that pollen TGase controls the transition between actin bundles and short filaments at the boundary between the apical and base domains of pollen tubes , favouring the assembly of more stable bundles. TGase activity may also be regulated by pH changes. The optimum pH value of TGase is alkaline  and the presence of an alkaline band in pollen tubes  suggests that local variation in pH may trigger TGase activity in these restricted areas. The polyamine concentration is also relevant by pushing toward the formation of mono- or bis-derivatives . Consequently, the actin–actin interactions, the actin–mono-polyamine linking and the possible ionic bonds between the free amino groups of bound polyamines with other pollen molecules are finely regulated by the in vivo conditions of the pollen tube.
The results of the present study also suggest that the TGase-mediated conjugation of polyamines is part of the mechanism that regulates microtubule dynamics and functioning. In eukaryotic cells, the dynamic of microtubules is dependent on MAPs (microtubule-associated proteins), which are poorly known in pollen tubes; the exception is represented by a couple of polypeptides that enhance microtubule polymerization and possibly mediate the interaction of microtubules with the plasma membrane . The functioning of pollen tube microtubules involves the control of vacuole positioning , the trafficking velocity of mitochondria  and the focusing of secretory vesicles in the tip domain (unpublished results from our laboratories and ). Results in the present manuscript, coupled to current literature, suggest that TGase-modified microtubules represent a subset of microtubules more resistant to depolymerization and along which the activity of motor proteins is partly inhibited.
In conclusion, these in vitro experiments show that pollen TGase affects the organization and functioning of the cytoskeleton; we propose that the activity of TGase might be modulated in vivo (under less extreme conditions) to exert a regulatory role on pollen tube growth.
This work was supported by the Italian Ministero dell'Università [grant numbers PRIN 2005 and PRIN 2007, http://prin.miur.it/] to the Dipartimento Biologia Evoluzionistica Sperimentale (DSF, University of Bologna) and by the Human Frontier Science Program (http://www.hfsp.org) to the Dipartimento Scienze Ambientali, University of Sienna.
Abbreviations: ANP, N-(4-azido-2-nitrophenyl)-putrescine; DIC, differential interference contrast; DTT, dithiothreitol; gplTGase, guinea pig liver transglutaminase; HIC, hydrophobic-interaction chromatography; TGase, transglutaminase
- © The Authors Journal compilation © 2009 Biochemical Society