INTRODUCTION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

It is currently accepted that fruit set depends on the coordinated action of positive growth signals as a result of pollination and fertilization (Gillaspy et al., 1993). Parthenocarpy, the formation of seedless fruit, occurs naturally in many species (George et al., 1984) and can be artificially induced by hormone application, mainly auxins and gibberellins (GAs; Goodwin, 1978; Schwabe and Mills, 1981; García-Martínez and Hedden, 1997). In natural parthenocarpy, it has been suggested that the expression of parthenocarpic genes can affect the pattern of hormone production, transport, and/or metabolism leading to hormone levels in the ovary capable of promoting growth even in the absence of pollination and fertilization (Nitsch, 1970; Gillaspy et al., 1993).

Natural parthenocarpy has been widely studied in tomato (Lycopersicon esculentum Mill.) because of its potential use to improve fruit set under unfavorable environmental conditions. Different tomato genotypes carrying gene(s) for parthenocarpy have been discovered or selected, such as pat, pat-2, and pat-3/pat-4 (Philouze, 1983; George et al., 1984; Lukyanenko, 1991). In these three genetic systems, the mechanism necessary for fruit set and development seems to be switched on before pollination and fertilization (Fos and Nuez, 1996, 1997; Mazzucato et al., 1998). In addition, recent evidence supports the hypothesis that parthenocarpic fruit set and growth induced by pat-2 (Fos et al., 2000) and pat-3/pat-4 (Fos et al., 2001) depend on GAs.

There are no reports in the literature showing parthenocarpic fruit set after polyamine (PA) application. However, several lines of evidence suggest that PAs may have a role in early fruit development in different species (Costa and Bagni, 1983; Evans and Malmberg, 1989; Egea-Cortines and Mizrahi, 1991). In tomato, inhibitors of Orn decarboxylase (ODC; EC 4.1.1.17) prevent the development of pollinated tomato ovaries, and this inhibition was counteracted by putrescine application in the presence of the Arg decarboxylase (ADC; EC 4.1.1.19) pathway (Cohen et al., 1982). High levels of free PAs at anthesis (AN) and during the initial stages of fruit development have been reported (Cohen et al., 1982; Teitel et al., 1985; Egea-Cortines et al., 1993). A rapid and transient increase in the amount of free PAs and a decrease of conjugated PAs after pollination and during early parthenocarpic fruit development induced by auxin (Mizrahi and Heimer, 1982; Alabadí et al., 1996) and GAs (Alabadí et al., 1996) has been observed. The activity of putrescine biosynthetic enzymes ODC and ADC is also elevated during the early stages of fruit development in tomato (Heimer et al., 1979; Cohen et al., 1982). Also, a transient increase of ADC and ODC activities takes place during early parthenocarpic fruit induced by auxin and GAs, with a maximum activity reached after 5 d for 2,4-dichlorophenoxyacetic acid (2,4-D) and after 8 d for GA₃ treatment (Alabadí et al., 1996). In addition, expression of the ODC and spermidine synthase (SPDS) genes is up-regulated during early fruit development after pollination and 2,4-D and GA₃ application (Alabadí and Carbonell, 1998, 1999). In the case of GA-induction, the expression pattern of both genes was similar to that induced by pollination (Alabadí and Carbonell, 1998, 1999). Finally, a correlation between maximal ODC transcript levels and maximal ODC activity has been observed (Alabadí et al., 1996). These results support the idea that ODC is the primary enzyme in the regulation of putrescine biosynthesis during early fruit development in tomato, which takes place mainly by cell division until about 10 d after AN (Gillaspy et al., 1993). In addition, it suggests an interaction between GAs and PAs during early tomato fruit development.

In this work, the involvement of PAs in tomato parthenocarpic fruit development controlled by pat-2 has been investigated. Our results show that PA application induces partial parthenocarpic fruit set and growth of wild-type (wt) ovaries and that PAs are necessary for parthenocarpic growth of pat-2 ovaries. Supporting this idea, we have found that the pat-2 mutation activates PA biosynthesis through the ODC pathway, leading to increased free spermine content in ovaries before pollination compared with wt ovaries. This increase may be attributable to the higher GA levels of unpollinated pat-2 ovaries.

	RESULTS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Effect of PA Application on Parthenocarpic Growth of Tomato cv Madrigal (MA)/wt and MA/pat-2 Ovaries

Unpollinated MA/wt ovaries did not grow (20 d after treatment their average fresh weight was 22 mg versus 16 mg at AN) compared with pollinated ovaries (Table I, Experiment I). Application of GA₃ induced fruit set of unpollinated MA/wt ovaries, although the weight of parthenocarpic fruits was significantly lower than those developed from pollinated ovaries (Table I, Experiment I). This is in agreement with previous results reported for the MA cultivar (Fos et al., 2000). Application of putrescine, spermidine, or spermine induced partial fruit set of unpollinated MA/wt ovaries (three to five ovaries developed out of 13 treated ovaries; Fig. 1). The weight of parthenocarpic fruits was in all cases significantly lower than that of seeded fruits (Table I, Experiment I). Weights of parthenocarpic fruits induced by GA₃ and spermidine were not significantly different (Table I, Experiment I).

View larger version (79K): [in this window] [in a new window]   Figure 1.   Effect of PAs on parthenocarpic fruit growth of MA/wt unpollinated ovaries. Fruits collected 20 d after hand pollination or application of 1 mM putrescine, 1 mM spermidine, and 1 mM spermine to ovaries from emasculated flowers.

Unpollinated MA/pat-2 ovaries developed similarly to hand-pollinated MA/pat-2 ovaries (Table I, Experiment II). Application of paclobutrazol, an inhibitor of the GA biosynthetic enzyme ent-kaurene oxidase, prevented growth of unpollinated MA/pat-2 ovaries (only one ovary developed, although significantly less than untreated ovaries; Table I, Experiment II). It is already known that paclobutrazol inhibition is fully reverted by GA₃ (Fos et al., 2000). Application of spermidine, the most efficient PA in inducing parthenocarpy in MA/wt ovaries (Table I, Experiment I), reverted the inhibition of paclobutrazol (Table I, Experiment II). The weight of developed fruits was not significantly different to hand-pollinated fruits or to fruits treated only with spermidine (Table I, Experiment II). Application of spermidine did not alter the growth of unpollinated MA/pat-2 ovaries (Table I, Experiment II).

Effect of Inhibitors of Putrescine Biosynthesis on Fruit Set and Development of Unpollinated MA/pat-2 Ovaries

As previously shown, MA/pat-2 ovaries developed parthenocarpically in the absence of pollination (Table I, Experiment II). Application of -difluoromethyl-Orn (DFMO) and -difluoromethyl-Arg (DFMA), which are irreversible inhibitors of the putrescine biosynthetic enzymes ODC and ADC respectively, prevented fruit growth from unpollinated pat-2 ovaries (Table II). Application of putrescine fully reverted the inhibition of DFMA but not that of DFMO (Table II, Experiment I). The inhibition of DFMO could not be reverted either by the application of spermidine (Table II, Experiment II). However, the inhibitory effect of both DFMO and DFMA, single or combined, on fruit set of unpollinated pat-2 ovaries was negated by application of GA₃ (Table II). The final weight of DFMA + GA₃-treated fruits was similar to that of control fruits. However, fruits of significantly smaller weight than control were obtained in the cases of DFMO + GA₃ and DFMO + DFMA + GA₃ treatments (Table II). The final weight of fruits treated with DFMO + DFMA + putrescine + GA₃ was significantly higher than those treated with DFMO + GA₃ and DFMO + DFMA + GA₃, although significantly smaller than developed fruits where ODC activity was not inhibited (control, DFMA + putrescine, and DFMA + GA₃; Table II).

PA Content of MA/wt and MA/pat-2 Ovaries during Flower Development

Contents of free putrescine, spermidine, and spermine in flower bud (FB), preanthesis (PR), and AN ovaries did not vary significantly during flower development in MA/wt and in MA/pat-2 (data not shown). Therefore, mean value levels (micrograms per gram fresh weight) of PAs from ovaries of all developmental stages were compared between MA/wt and MA/pat-2 (Fig. 2). In both lines, levels of putrescine (200-250 µg g¹ fresh weight) were higher than those of spermidine (100-140 µg g¹ fresh weight), and these were higher than those of spermine (20-40 µg g¹ fresh weight; Fig. 2). In MA/wt ovaries, the average content of putrescine was about 25% higher than in MA/pat-2 ovaries, whereas spermidine and spermine contents were about 25% and 80% higher, respectively, in MA/pat-2 than in MA/wt ovaries. The difference was only significant in the case of free spermine content (Fig. 2).

View larger version (17K): [in this window] [in a new window]   Figure 2.   Content of free putrescine, spermidine, and spermine in non-parthenocarpic (wt) and parthenocarpic (pat-2) tomato MA ovaries. Values are mean contents of FB, PR, and AN ovaries ± SE, n = 12. Means of spermidine content were significantly different (P < 0.05).

ADC and ODC Activities in MA/wt and MA/pat-2 Ovaries

We investigated whether the differences in PA content (Fig. 2) observed between MA/wt and MA/pat-2 ovaries are associated to differences in the activity of putrescine biosynthetic enzymes. ADC and ODC activities did not vary significantly in the two lines during flower development (data not shown), so mean values from ovaries of all developmental stages of MA/wt and MA/pat-2 were compared. The levels of ADC were not different between MA/wt and MA/pat-2 ovaries, whereas ODC activity was significantly higher (about 50%) in MA/pat-2 than in MA/wt ovaries (Fig. 3).

View larger version (20K): [in this window] [in a new window]   Figure 3.   ADC and ODC activities in non-parthenocarpic (wt) and parthenocarpic (pat-2) tomato MA ovaries. Values are mean contents of FB, PR, and AN ovaries ± SE, n = 6. Means of ODC activity were significantly different (P < 0.05).

Transcript Levels of PA Biosynthesis Genes in MA/wt and MA/pat-2 Ovaries during Flower Development

To determine whether the differences in PA content (Fig. 2) and in ODC activity (Fig. 3) are primarily regulated at the level of gene expression, we investigated transcript levels of genes encoding ADC, ODC, and SPDS. Transcript levels of the ADC gene were similar in MA/wt and MA/pat-2 ovaries, and no differences were observed along flower development (Fig. 4). ODC transcript levels in MA/wt ovaries did not vary during flower development, whereas in MA/pat-2 ovaries they were higher at FB and PR than at AN (Fig. 4). Importantly, ODC transcripts were more abundant in pat-2 ovaries than in wt ovaries at FB and PR (Fig. 4). The temporal pattern of SPDS transcript levels was similar to that found for the ODC gene: They did not change in MA/wt ovaries and were slightly more abundant in MA/pat-2 than in MA/wt ovaries at FB and PR stages (Fig. 4).

View larger version (102K): [in this window] [in a new window]   Figure 4.   Transcript levels of PA biosynthesis genes (ADC, ODC, and SPDS) in non-parthenocarpic (wt) and parthenocarpic (pat-2) tomato MA ovaries. FB, Flower bud; PR, preanthesis; AN, anthesis.

	DISCUSSION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Application of putrescine, spermidine, and spermine to unpollinated wt ovaries induced partial parthenocarpic growth in tomato (Table I; Fig. 1). To the best of our knowledge, this is the first report on the role of PAs as parthenocarpy inducers and supports the idea of the involvement of PAs in early tomato fruit growth. This is in agreement with the observation that inhibitors of putrescine biosynthesis prevented growth of pollinated tomato ovaries and that putrescine application could negate this inhibition (Cohen et al., 1982). It is of interest that spermidine was capable of producing fruits of similar size to those obtained with GA₃ and larger than those obtained with putrescine and spermine (Table I). It is not possible to ascertain, however, whether GAs and PAs act independently on inducing fruit set in tomato, although the lower efficiency of PAs compared with GA₃ suggests that PAs may not act as a primary signal.

It has been previously reported that paclobutrazol prevents growth of unpollinated MA/pat-2 ovaries and that this inhibition is fully reverted by application of GA₃ (Fos et al., 2000). We show now that paclobutrazol inhibition is also negated by spermidine application (Table I, Experiment II). These results suggest a role of PAs in inducing parthenocarpic fruit development in tomato. Moreover, the higher efficiency of spermidine to revert paclobutrazol inhibition in MA/pat-2 ovaries (Table I, Experiment II) than to induce parthenocarpic fruit set and development in MA/wt ovaries (Table I, Experiment I) suggests that pat-2 enhances the response capability to PAs.

The parthenocarpic fruit growth induced by pat-2 mutation depends on GAs (Table I, Experiment II), in agreement with previous report (Fos et al., 2000). We show now that it also depends on PAs and that both putrescine biosynthetic pathways are required for parthenocarpic growth of pat-2 ovaries (Table II). The DFMA effect could be reverted by putrescine-treatment, whereas that of DFMO could not be reverted by neither putrescine nor spermidine. The absence of reversion of the DFMO effect by PAs, in contrast to that of DFMA, is not easy to explain. It has been reported that putrescine counteracts the effect of ODC inhibitors on development of pollinated tomato ovaries (Cohen et al., 1982). The discrepancy with our results may be attributable to the different genotypes employed in both studies and/or to the absence of pollination in pat-2 ovaries.

The presence of a protein complex involving enzymes of spermidine and spermine biosynthesis in Arabidopsis has recently been proposed (Panicot et al., 2002). It is possible that the tomato ODC is part of a similar protein complex, in which the putrescine synthesized is directly transferred to the next enzyme in the pathway, not necessarily SPDS. Because of steric constraints, no free putrescine would be accessible to the complex, explaining why exogenous putrescine does not revert the DFMO effect. In the case of ADC, there may be no constraints to make the free, applied putrescine accessible to the next enzyme, explaining therefore why putrescine reverts the DFMA effect. A second hypothesis could be a differential subcellular compartmentalization of ODC and ADC activities, as previously suggested (Galston et al., 1997; Tiburcio et al., 1997; Walden et al., 1997). ODC and ADC activities have been localized in the nucleus and vacuole (Walker et al., 1987; Slocum, 1991) and in chloroplast and mitochondria (Walker et al., 1987; Borrell et al., 1995), respectively. A differential subcellular compartmentalization of the two enzymes might produce different accessibility of the enzymes to applied putrescine.

On the other hand, GA₃ could revert the effect of DFMO and DFMA, separated and combined (Table II), indicating that the lack of reversibility of DFMO inhibition by putrescine was not attributable to a toxic effect of the inhibitor. Because both ODC and ADC activities are necessary for parthenocarpic growth, the GA reversion of DFMO and DFMA inhibition could be explained by GA-induced expression of ODC and ADC genes. A transient increase of both ODC transcript and activity levels in GA₃-treated tomato ovaries, with maximum levels 8 d after treatment has been reported. This ODC regulation pattern is similar to that observed after pollination (Alabadí et al., 1996; Alabadí and Carbonell, 1998).

The content of free spermine was significantly higher in MA/pat-2 than in MA/wt unpollinated ovaries (Fig. 2). This is consistent with: (a) the capacity of putrescine, spermidine, and spermine applications to induce fruit set in MA/wt ovaries; (b) the high efficiency of spermidine to revert paclobutrazol inhibition of MA/pat-2 ovaries growth; and (c) the parthenocarpic capability of unpollinated MA/pat-2 ovaries. A rapid and transient increase in the amount of free PAs during early parthenocarpic tomato fruit development induced by 2,4-D and GA₃ applied on day before AN has been previously found (Alabadí et al., 1996). The elevated free spermine levels found in MA/pat-2 unpollinated ovaries before AN may be, therefore, a consequence of the higher GA content in unpollinated pat-2 ovaries (Fos et al., 2000).

ADC and ODC pathways of putrescine biosynthesis are operative in unpollinated tomato ovaries before pollination (Figs. 3 and 4). In the case of pat-2, both pathways of putrescine biosynthesis seem to be necessary for parthenocarpic growth (Table II). ADC activity was similar in MA/wt and MA/pat-2 unpollinated ovaries, whereas ODC activity was significantly higher in MA/pat-2 than in MA/wt unpollinated ovaries (Fig. 3). ADC and ODC activities were found to increase after auxin or GA treatment (Alabadí et al., 1996). ODC, SPDS (Alabadí and Carbonell, 1998, 1999), and ADC (C. Acosta, M.S. Agüero, and J. Carbonell, unpublished data) transcripts have also been detected in developing fruits of tomato. No differences in ADC transcript levels were observed between the two lines (Fig. 4). However, ODC transcript levels were much higher in MA/pat-2 than in MA/wt ovaries. Differences in SPDS transcript level between MA/wt and MA/pat-2 were smaller than those found for the ODC gene (Fig. 4). A transient increase of ODC and SPDS transcripts in 2,4-D- and GA₃-treated tomato ovaries has been observed previously, with maximum levels between 5 and 8 d after stimulation (Alabadí and Carbonell, 1998, 1999). Also, maximal ODC transcript content correlates with maximal ODC activity in ovaries post-AN (Alabadí et al., 1996). These results support the idea that parthenocarpy induced by pat-2 is associated to an increase of ODC activity and transcript levels (Fig. 5). The elevated spermine content and SPDS transcript level suggest that pat-2 regulate several steps of the PA biosynthesis pathway, probably as a result of higher GA levels in unpollinated pat-2 ovaries (Fig. 5.).

View larger version (10K): [in this window] [in a new window]   Figure 5.   Proposed regulation of PA biosynthesis by pat-2 in unpollinated tomato ovaries. pat-2 activates the ODC pathway of putrescine biosynthesis and the spermidine and spermine biosynthesis genes through an increase of GA content (see upwards arrow) in unpollinated ovaries. Question mark indicates that SPMS gene expression has not been analyzed yet.

In conclusion, all results presented show that: (a) PAs induce partial parthenocarpy in tomato, (b) parthenocarpic growth of pat-2 depends on PAs and requires both ADC and ODC putrescine biosynthetic pathways, and (c) pat-2 activates PA biosynthesis through the ODC pathway leading to elevated free spermine content in unpollinated ovaries, probably as a result of a higher concentration of active GAs in pat-2 ovaries.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Plant Material

A non-parthenocarpic tomato (Lycopersicon esculentum Mill. cv Madrigal) line (referred to as MA/wt) and its corresponding near-isogenic parthenocarpic line carrying the pat-2 gene (Par54-11, referred to as MA/pat-2) were used in the experiments. The parthenocarpic line MA/pat-2 was derived from the cross of MA/wt and Severianin (Nuez et al., 1985) and backcrossed to the non-parthenocarpic line MA/wt four times (Fos and Nuez, 1996).

Plants were grown in an air-conditioned greenhouse set at an average temperature of 15°C in 25-L pots containing a peat:soil (1:1, v/v) mixture, irrigated with nutrient solution under natural light conditions. The temperatures fluctuated according to the environment, but extremes were never higher than 25°C (day) or lower than 6°C (night).

Ovaries from flowers at FB, PR, and AN stages were collected from the second to sixth clusters, as previously described (Fos and Nuez, 1996), for PA analysis. Ovaries were frozen immediately in liquid N₂ and stored at 80°C until extraction.

Plant Hormone Application

MA/wt and MA/pat-2 flowers from the second to fourth cluster were emasculated 1 d before AN to prevent self-pollination and on the same or next day ovaries were either hand-pollinated or treated with GA₃ (0.29 mM; Sigma-Aldrich Química SA, Alcobendas, Spain) or diverse doses of putrescine, spermidine, or spermine (Sigma-Aldrich). Paclobutrazol (0.85 mM; Duchefa, Haarlem, The Netherlands) and DFMO or DFMA (1 mM; Marion Merrell Dow Inc., Cincinnati) were applied to unpollinated MA/pat-2 ovaries alone or combined with other substances. All treatments were applied to the ovary in 20 µL of 10% to 20% (v/v) methanol solution containing 0.1% (v/v) Tween 80 (Sigma-Aldrich). Control ovaries were treated with the same volume of solvent solution. Each treatment was carried out on nine to 18 ovaries (three to four plants, one to two clusters per plant, and two to three ovaries per cluster).

Quantification of PAs

Extraction and quantification of free PAs was based on the method described by Smith and Davies (1985). Ovaries (about 200 mg fresh weight) were homogenized in 5 volumes of 0.2 M cold HClO₄ with 100 µg mL¹ 1,6-diaminohexane (Aldrich, Buchs, Switzerland) as an internal standard. Homogenates were centrifuged at 15,000g for 15 min at 4°C. Aliquots (100 µL) of the supernatants were mixed with 200 µL of saturated Na₂CO₃ and 400 µL of dansyl chloride (Sigma-Aldrich) in acetone (10 µg mL¹) in a tapered reaction vial. Mixtures were incubated at 60°C for 1 h. A 100 µL aliquot of a solution of L-Pro (100 µg mL¹) was added. After 30 min, dansylated PAs were extracted twice with 500 and 300 µL of toluene (HPLC grade). The mixture of toluene fractions was dried out under N₂ flow. The residue was dissolved in 200 µL of acetonitrile (HPLC grade), filtered through HV-4 filters (Millipore, Molsherim, France), and analyzed immediately by HPLC. Aliquots (2-5 µL) of samples were injected onto a reverse-phase column (200 mm long, 4.6 mm i.d.) of Hypersil ODS (5 µm) and eluted with a programmed water-acetonitrile solvent gradient from 60% to 90% in 30 min at a flow rate of 1.5 mL min¹ at 35°C. An attached fluorescence spectrophotometer (420 AC, Waters Associates, Milford, MA) equipped with an integrator LCI-100 (PerkinElmer Instruments, Norwalk, CT) was used to quantify the dansylated samples. Excitation and emission wavelengths were 365 and 425 nm (band pass), respectively.

Extraction and Assay for ADC and ODC

Frozen material (200-250 mg fresh weight), previously ground in a mortar with liquid N₂, was homogenized in 5 volumes of 100 mM potassium phosphate buffer, pH 7.5, with 20 mM -mercaptoethanol, 10 µM leupeptin, and 100 µM pyridoxal. The homogenate was centrifuged at 15,000g for 15 min at 4°C, and the supernatant was used for ADC and ODC assays. ADC and ODC activities were assayed according to Birecka et al. (1985) by measuring the ¹⁴CO₂ liberated from L-[1-¹⁴C]Arg (55 mCi mmol¹; American Radiolabeled Chemicals, St. Louis) and L-[1-¹⁴C]Orn (55 mCi mmol¹; American Radiolabeled Chemicals), respectively. An aliquot of 50 µL of enzymatic extract were mixed with 100 µL of substrate (including 0.3 µCi of labeled compound) to give a final concentration of 5 mM Arg or 10 mM Orn. The amount of protein was 100 to 250 µg per assay. The reaction was carried out at 30°C for 1 h in tapered vials bearing a No. 1 paper disc (Whatman, Clifton, NJ) impregnated with 20 µL of 4 N KOH. The reaction was stopped by injecting 200 µL of cold 15% (w/v) trichloroacetic acid. After 30 min at 30°C, the vials were opened, filters were air dried, and radioactivity was determined in a liquid scintillation counter (1409, PerkinElmer Wallac, Gaithersburg, MD). All assays were run per duplicate. Controls were carried out using homogenization buffer instead of enzymatic extracts. Protein concentration in the enzymatic extracts was determined according to Bradford (1976) using bovine serum albumin fraction V (Sigma-Aldrich) as standard.

RNA Extraction and Northern Analysis

Total RNA was isolated from homogenized, frozen tissues (about 150-200 mg fresh weight) as described by Bartels and Thompson (1983). Total RNA (20 µg) was fractionated in a 1.5% (w/v) agarose gel containing 6.7% (v/v) formaldehyde, transferred in 20× SSC to nylon membranes (Nytran, Schleicher & Schuell, Dassel, Germany), and cross-linked with a UV Stratalinker 800 (Stratagene, La Jolla, CA). The filters were prehybridized for 1 h at 42°C in a solution containing 50% (v/v) formamide, 5× SSC, 50 mM Na₂HPO₄, pH 6.3, 1× Denhardt's solution, 0.1% (w/v) SDS, 0.1 mg mL¹ denatured salmon sperm DNA, and 10 µg mL¹ poly(A⁺) RNA. Hybridization probes were prepared from full-length cDNA clones of ODC (pD12; Alabadí and Carbonell, 1998), ADC (Leargdeca; Rastogi et al., 1993), and SPDS (pSPDtom.5; Alabadí and Carbonell, 1999). Probes were labeled with ³²P using the Ready-To-Go DNA labeling kit (Amersham Biosciences AB, Uppsala). Hybridizations were performed at 42°C for at least 16 h. Filters were washed three times for 5 min at room temperature in 1× SSC, 0.1% (w/v) SDS and three times for 15 min at 65°C in 0.1× SSC, 0.1% (w/v) SDS.

Statistical Methods

Statistical treatments of the data were made by ANOVA and using the Fisher's LSD procedure to discriminate among the means (Tables I and II) and the Student's t test for comparison of means (Figs. 2 and 3; Statgraphics Plus program, v3.1 for Windows, Statistical Graphics, Rockville, MD).