Spermidine-Binding Proteins. Purification and Expression Analysis in Maize

Abstract

Polyamine-binding proteins have been identified in a wide range of organisms, including mammals, yeasts, and bacteria. In this work, we have investigated specific spermidine binding to plant membrane proteins purified from microsomes of etiolated maize (Zea mays) coleoptiles. In the final purification step, specific spermidine-binding activity (K_d 6.02 10⁻⁷ m) was eluted from a HiTrapQ fast-protein liquid chromatography column at about 0.25 m NaCl, and sodium dodecyl sulfate-polyacrylamide gel electrophoresis of the most active fraction showed a major polypeptide of about 60 kD and another copurifying 18-kD protein. Competition experiments, performed on HiTrapQ active fractions, confirmed the specificity of the binding. Upon Sephadex G-100 gel filtration, spermidine binding was associated almost exclusively with the 18-kD protein. On the basis of the N-terminal sequences, degenerate oligonucleotide probes were designed and used to isolate, by reverse transcriptase-polymerase chain reaction and polymerase chain reaction, cDNA fragments of about 1 kb for the 60-kD protein, and 0.9 kb for the 18-kD protein. Northern-blot analysis performed on etiolated coleoptiles and different tissues from 10-d-old maize plants indicated the presence of two different mRNAs of 1.7 and 0.7 kb. Southern-blot analysis indicated that the genes encoding the 60- and 18-kD proteins are probably derived from differential processing of the same precursor mRNA. Using rabbit polyclonal antibodies raised against these proteins, affinity purification and dot-blot experiments detected analogous membrane proteins in monocot and dicot plants.

Aliphatic polyamines are ubiquitous polycations implicated in a wide range of organisms in several aspects of growth and development, although their specific mechanism of action is not well understood. In plants, they are involved in processes such as embryogenesis, senescence, and flowering (Bagni, 1989).

In animal and plant systems, it was postulated that polyamines may play a role in the post-translational modifications of proteins (Äschlimann and Paulsson, 1994; Serafini-Fracassini et al., 1995), as well as in the modulation of many enzyme activities such as protein kinase, phosphatase, ATPase (Datta et al., 1986; Friedman, 1986; Reggiani et al., 1992), and 1,3-β-glucan synthase (Kauss and Jeblick, 1985). Despite the large amount of information about the covalent interactions between polyamines and proteins (Serafini-Fracassini et al., 1995), little is known on the noncovalent binding of polyamines to membrane proteins, although this may represent one of the first steps of their action at cellular level. The interaction between polyamines and membranes is suggested to be an intermediate in important cellular events such as membrane fusion (Schuber, 1989) and transmission of receptor-mediated signals (Koenig et al., 1983).

In Escherichia coli, two periplasmic polyamine-binding proteins (PotD and PotF), which are part of two membrane transport systems (pPT104 and pPT79) and which bind spermidine and putrescine, respectively, have been isolated and crystallized (Kashiwagi et al., 1993; Pistocchi et al., 1993; Sugiyama et al., 1996; Vassylyev et al., 1998). Structural resolution of PotD led to the elucidation of the spermidine recognition mechanism and of the main chain folding at the binding site (Sugiyama et al., 1996). In Bacillus subtilis, Woolridge et al. (1997) reported that the excretion of spermidine can be catalyzed by the multidrug transporter Blt. From sequence similarity between Blt and proteins encoded by the yeast Saccharomyces cerevisiae, four proteins were identified, one of which is a vacuolar membrane protein that catalyzes a proton gradient-dependent polyamine transport (Tomitori et al., 1999). This protein has 12 putative transmembrane segments, and the Glu residues, which may interact with polyamines, are located in positions similar to those in the PotE protein, which catalyzes uptake and excretion of putrescine in the transporter pPT71 of E. coli (Kashiwagi et al., 2000). In animal systems, photoaffinity labeling has been used in attempts to identify polyamine-binding sites. Leroy et al. (1995) characterized polyamine-binding sites on the β-subunit of the protein kinase CK2, a Ser/Thr kinase of Drosophila melanogaster, and different photoprobes were used to identify putative polyamine-binding proteins that could be part of the polyamine transporter in tumor cells (Felschow et al., 1995, 1997). A 16-kD binding protein for S-adenosyl-Met, a precursor for spermidine synthesis structurally related with spermidine, was recently isolated and characterized from the green alga Chlamydomonas reinhardtii (Nakano et al., 2000).

In vascular plants, previous work described the main characteristics of spermidine binding to purified zucchini (Cucurbita pepo) plasma membrane vesicles (Tassoni et al., 1996), including structural specificity. Total proteins were subsequently solubilized from plasma membrane vesicles, and the specific spermidine binding activity characterized (dissociation constant [K_d] of 5 μm). A partial purification of two proteins (44 and 66 kD) associated with specific polyamine-binding activity was performed (Tassoni et al., 1998). The present work reports the purification, N-terminal sequencing, and expression analysis of specific spermidine-binding proteins in maize (Zea mays).

RESULTS

Spermidine-Binding Protein Purification

Solubilization of spermidine-binding activity from maize coleoptile microsomes was tested by using a selection of detergents at different concentrations and by an acetone powder method, which gave the best recovery of activity (Table I).

Table I.

Solubilization of specific [¹⁴C]spermidine-binding activity from maize coleoptiles

Detergent	Concentration	Specific Activity
		pmol g fresh wt−1
Octylglucoside	1.0% (w/v)	0.79 ± 0.17
Triton X-100	1.0% (v/v)	1.16 ± 0.30
	0.5% (v/v)	1.47 ± 0.40
Nonidet P-40	1.0% (v/v)	0
	0.5% (v/v)	0.80 ± 0.30
Acetone powder	–	1.84 ± 0.61

Detergents were added to a 2-g fresh wt mL⁻¹ microsome resuspension. Values represent the mean ± se of triplicate samples.

The purification protocol first involves diethylaminoethyl (DEAE)-BioGel anion-exchange chromatography, where the major part of spermidine-binding activity was eluted between 0.1 and 0.2 m NaCl in binding buffer (Fig. 1A). Some binding activity was subsequently eluted at 0.5 m NaCl (0.40 ± 0.02 pmol g fresh weight⁻¹, about 25% of the total activity), suggesting the presence of other spermidine-binding proteins in maize coleoptiles, but this fraction was not considered further. No additional activity was eluted at 1 m NaCl. The second purification step utilized an octyl-agarose column from which spermidine-binding activity was recovered by elution with binding buffer (Fig. 1B). Final purification involves an HiTrapQ fast-protein liquid chromatography (FPLC) column and elution with a 0 to 0.5 m NaCl linear gradient in binding buffer (Fig. 1C). Spermidine-binding activity was eluted at about 0.25 m NaCl, and the most active fraction (no. 5) showed a major polypeptide of about 60 kD after SDS-PAGE and blotting, together with a minor 18-kD protein (Fig. 2). The binding activity of fractions collected from the different purification steps was confirmed by ultracentrifugation assay with Amicon (Beverly, MA) filters. Relative to the initial acetone powder extract, the specific spermidine-binding activity (expressed as picomoles per milligrams of protein) was enriched in HiTrapQ FPLC fraction 5 by about 1,500-fold (Table II).

Figure 1.

Purification of spermidine-binding proteins from maize coleoptiles. A, DEAE-Bio-Gel anion-exchange column. Elution was performed with binding buffer supplemented with increasing concentrations of NaCl. Bars, [¹⁴C]Spermidine-specific binding activity; dotted line, protein elution pattern (A₂₈₀). B, Octyl-agarose hydrophobic interaction column. Elution was performed by binding buffer with decreasing concentrations of ammonium sulfate. Bars, [¹⁴C]Spermidine-specific binding activity; dotted line, protein elution pattern (A₂₈₀). C, HiTrapQ FPLC column. Protein elution was performed with a 0 to 0.5 m NaCl linear gradient in binding buffer at a flow rate of 1 mL min⁻¹. ▴, [¹⁴C]Spermidine-specific binding activity, dotted line, protein elution pattern (A₂₈₀); solid line, NaCl gradient.

Figure 2.

SDS-PAGE of HiTrapQ-FPLC fractions. Proteins were separated by 12% (w/v) SDS-PAGE, blotted onto nitrocellulose, and stained (Kumar et al., 1985). M, M_r markers.

Table II.

Purification of specific [¹⁴C]spd-binding activity from maize coleoptiles (260 g fresh wt)

Purification Step	Protein	Specific [14C]spd Binding	Specific Activity	Recovery
	mg	pmol	pmol mg protein−1	%
Acetone powders (crude extract)	120	475	4.0	100
DEAE (0.2 m NaCl)	83.2	71.6	0.9	15.0
Octyl-agarose (buffer fraction)	3.4	64.2	18.9	13.5
HiTrapQ (fraction 5)	0.005	32.6	6270	6.9

The binding activity was measured by glass fiber filter assay. Values are the mean of eight different experiments.

The concentration dependence of spermidine binding to purified protein (Fig. 3) showed that saturation was approached at about 5 μm polyamine concentration, with an apparent K_d of 6.02 10⁻⁷ m. Using purified fraction 5, competition experiments were carried out between [¹⁴C]spermidine ([¹⁴C]spd) and unlabeled spermidine or other polyamines at a 20 μm concentration (50-fold higher than the [¹⁴C]spd). Unlabeled spermidine inhibited [¹⁴C]spd-specific binding by 95%, spermine (a tetraamine) inhibited it by 85%, and norspermidine (a triamine that lacks a CH₂ group with respect to spermidine) inhibited it by 78%, whereas the diamine putrescine only inhibited specific spermidine binding by about 20%.

Figure 3.

Concentration dependence of specific [¹⁴C]spermidine binding to purified HiTrapQ fraction 5 proteins. Values represent the mean ± se of triplicate samples.

To determine which of the two polypeptides was mainly responsible for spermidine binding, HiTrapQ fraction 5 was incubated for 5 min on ice with 90 kBq of [¹⁴C]spd and it was then loaded on a gel-filtration Sephadex G-100 column. The elution profile of [¹⁴C]spd showed the presence of a radioactivity peak at an apparent molecular mass of 20 kD (Fig. 4). SDS-PAGE analysis confirmed the presence of an 18-kD protein (Fig. 4, inset) in the active fractions.

Figure 4.

Sephadex G-100 gel-filtration elution profile. HiTrapQ FPLC fraction 5 was applied to a Sephadex G-100 column after incubation with [¹⁴C]spd. Proteins were eluted with binding buffer plus 0.1 m NaCl, and 0.7-mL fractions were collected. ▴, Radioactivity (dpm) elution profile; ○, A₂₈₀ protein elution profile. Inset, Silver-stained 12% (w/v) SDS-PAGE of the active fractions.

N-Terminal Sequencing and Database Searching

Copurification of the 60- and 18-kD polypeptides was confirmed using other purification protocols (data not shown), suggesting that the 60-kD protein may also have a role in the spermidine-binding process. The N-terminal sequences (Fig. 5) showed no similarity between the two proteins. Database searching (FASTA and BLASTP programs) identified expressed sequence tags (ESTs) for the 18-kD protein only. Figure 6 shows the alignment of the 18-kD protein with the deduced partial amino acid sequence of the ESTs. The best similarity, over 20 amino acids of overlap, was found with two ESTs from maize (about 85%), maize 1 (clone BE051852 from maize glume cDNAs library) and maize 2 (clone AW330986 from maize mixed adult tissues cDNA library). Lower similarity was found with sorghum (78.9%; clone AW677515, cDNA from dark-grown sorghum) and rice (65%; clone AU033236, cDNA from rice etiolated shoot). All of these clones are of unidentified function. Database searches (FASTA3 and BLASTP2 programs) performed using the deduced amino acid sequences of maize 1 and maize 2 ESTs gave similarity with some proteins of molecular mass ranging from 20.1 to 19.5 kD, the sequences of which are also aligned in Figure 6. Arabidopsis proteins Arabidopsis 1 (accession no. AL022605) and Arabidopsis 2 (accession no. AC007168) are of unknown function, whereas oilseed rape protein (accession no. AF314810) is a dehydration stress-induced protein. Two other proteins of about 19 kD, one from peppers (Capsicum annuum; accession no. AAF63515, induced during hypersensitive response to tobacco mosaic virus) and another from tobacco (Nicotiana tabacum; accession no. BAB13708, elicitor inducible protein), show less similarity (data not shown).

Figure 5.

N-terminal sequences of 60- and 18-kD proteins and corresponding degenerate nucleotide probes. Mixtures of nucleotides are indicated by the following one-letter code: N: a, g, c, t; H: a, t, c; R: a, g; S: g, c; Y: c, t; and M: a, c. Amino acid sequences are shown in the single-letter code below the oligonucleotide sequence. X, Undetermined amino acid. G(T), Possible alternative amino acids. The protein sequence data reported in this paper will appear in the SWISS-PROT Data Bank under the accession numbers P82867 and P82868.

Figure 6.

Alignment of the N-terminal 18-kD protein sequence with deduced amino acid sequences of ESTs and database sequences. Aligned sequences, Maize 1 (BE051852), maize 2 (AW330986), sorghum (Sorghum bicolor; AW677515), rice (Oryza sativa; AU033236), Arabidopsis 1 (AL022605), oilseed rape (Brassica napus; AF314810), and Arabidopsis 2 (AC007168). Black shading represents conserved amino acids in at least four of the aligned sequences; dark-gray shading represents blocks of similarity; light-gray shading represents weakly similar amino acids. The alignment was performed with the PILEUP program (University of Wisconsin, Genetics Computer Group package).

To give insight into the function of these proteins, N-terminal sequences, maize 1 and maize 2 EST sequences, were used to perform additional searches against motif databases (Pfam, Blocks, and Emotif), but no significant results were obtained.

Northern- and Southern-Blot Analysis

On the basis of the N-terminal sequences, two degenerate oligonucleotide probes were designed for each protein (Fig. 5) and by reverse transcription (RT)-PCR and PCR, cDNA fragments of about 1 kb for the 60-kD protein, and 0.9 kb for the 18-kD protein were amplified.

These fragments were used as probes for northern-blot analysis performed on total RNA extracted from etiolated maize coleoptiles and from primary leaves and stems collected from 10-d-old light-grown maize plants. Both probes gave the same pattern of expression detecting two main mRNAs of 1.7 and 0.7 kb (Fig. 7, A and B, lanes 1–3). In both hybridizations, the expression of the 1.7-kb transcript did not show any significant variation, whereas the 0.7-kb transcript seemed to be differentially expressed among the different tissues (about 70% higher in 10-d-old leaves and stems compared with coleoptiles, based on densitometric analysis). Equal RNA loading was confirmed by ethidium bromide staining and by hybridization with an 18S rRNA probe (data not shown).

Figure 7.

Expression analysis in maize. A and B, Northern-blot analysis of the total RNA extracted from etiolated coleoptiles (lane 1), 10-d-old primary leaves (lane 2), and stems (lane 3). The membrane was hybridized with the cDNA fragment corresponding to the 60-kD protein (A, 3-d exposure) or to the 18-kD protein (B, 16-h exposure). C and D, Southern-blot analysis: Genomic DNA was digested with NotI (lane 1), KpnI (lane 2), EcoRI (lane 3), and HindIII (lane 4) restriction enzymes. The membrane was hybridized with the cDNA probe related to the 60-kD protein (C, 5-h exposure) and to the 18-kD protein (D, 3-h exposure). E, Cross hybridization. The cDNA fragment corresponding to the 60-kD protein (about 100 ng of cDNA) was hybridized with the [α-³²P]dCTP-labeled cDNA fragment related to the 18-kD protein (15-min exposure).

For Southern-blot analysis, the genomic DNA, isolated from etiolated maize coleoptiles, was digested with NotI, KpnI, EcoRI, and HindIII restriction enzymes. Hybridization performed with the two cDNA fragments gave the same hybridization pattern (Fig. 7, C and D, lanes 1–4). Figure 7E shows that the two cDNA fragments cross hybridize.

Immunochemical Analysis

Using HiTrapQ fraction 5 in ELISA as coating antigen (0.05 μg protein per well), the polyclonal antiserum titer was measured to be 1:8,000. ELISA performed with fractions from the final HiTrapQ purification column confirmed maximum reaction at fraction 5, the peak spermidine-binding fraction (data not shown). After SDS-PAGE and western blotting of the fractions, the antiserum did not interact with the 60-kD protein, whereas it poorly recognized the 18-kD band (data not shown). Hence, the antibodies recognize efficiently only the native form of the proteins.

An immunoaffinity resin was prepared by crosslinking of the rabbit IgG to a HiTrap-Protein A column (Amersham Biosciences, Piscataway, NJ) and loaded with a partially purified maize fraction (0.05–0.2 m NaCl DEAE fraction). Two polypeptides of 60 and 18 kD were eluted from the resin with Gly-HCl buffer (Fig. 8, lane 1). Similar results were obtained (Fig. 8, lane 2) when the column was loaded with partially purified fraction from zucchini hypocotyls (0.2 m NaCl DEAE fraction).

Figure 8.

Protein A-antibody affinity column purification. Lane 1, Purification from maize coleoptiles; lane 2, purification from zucchini hypocotyls. Proteins were separated by 12% (w/v) SDS-PAGE, blotted onto nitrocellulose, and stained (Kumar et al., 1985). M, M_r markers.

A positive signal was detected in dot-blot assays performed with 0.3 μg of total proteins solubilized from microsomes of different species and tissues, namely maize coleoptiles and mung bean (Vigna radiata) and zucchini hypocotyls (Fig. 9). Purified proteins from HiTrapQ fraction 5 were used as control. A poor signal was found with oat (Avena sativa) coleoptile proteins. Preimmune serum gave no signal (data not shown).

Figure 9.

Dot-blot assay. Total proteins (0.3 μg) solubilized from oat (O) and etiolated maize coleoptiles (MC), or from mung bean (MB) and zucchini hypocotyls (Z) were spotted on nitrocellulose. A positive control was performed with an equal amount of HiTrapQ fraction 5-purified proteins (PP). Antiserum dilution: 1:500.

DISCUSSION

We have purified, from microsome membranes of etiolated maize coleoptiles, an 18-kD polypeptide that seems directly involved in specific spermidine binding (Fig. 4) and another copurifying 60-kD polypeptide probably related to the process (Fig. 2). The apparent K_d was calculated to be 6.02 10⁻⁷ m, an affinity about one and two orders of magnitude greater than those found with zucchini hypocotyls plasma membrane total solubilized proteins (Tassoni et al., 1998), and vesicles (Tassoni et al., 1996), respectively.

Competition experiments confirmed the specificity of the binding activity. Unlabeled spermidine exerted the highest inhibitory effect, followed by the tetraamine spermine and norspermidine. The diamine putrescine competed poorly. These results are similar to those previously observed in zucchini hypocotyls (Tassoni et al., 1996, 1998), where it was proposed that the spermidine-binding site should contain at least three negatively charged groups whose distance from one another resembles that of the positively charged amino groups of spermidine. This kind of conformation of the spermidine-binding site was already found in E. coli PotD protein (Sugiyama et al., 1996) and in human deoxyhypusine synthase (Liao et al., 1998).

Previous work on zucchini plasma membrane proteins (Tassoni et al., 1998) suggested the existence of two proteins (66 and 44 kD) associated with spermidine binding. Affinity purification using antibodies raised against purified maize spermidine-binding proteins yielded specifically the 60- and 18-kD maize polypeptides (Fig. 8, lane 1). With zucchini protein extracts, polypeptides of similar size were obtained (Fig. 8, lane 2). Moreover, the antibodies crossreacted in dot-blot assays with microsomal extracts from zucchini and mung bean, and to a lesser extent, with oat extracts (Fig. 9). The presence of polyamine-binding proteins in other plants might be anticipated from database searches of ESTs from different plants (maize, sorghum, and rice) that show sequence similarity with the 18-kD protein and by the presence of other proteins of uncertain function but of molecular mass of about 19 to 20 kD (Fig. 6). Unfortunately, no sequence similarities were found for the 60-kD protein.

Southern-blot analysis (Fig. 7, C and D) suggests that the genes encoding the 60- and 18-kD proteins are tightly linked and that both present in single copy. The cross hybridization between the cDNA fragments corresponding to the two proteins (Fig. 7E) suggests differential processing of the same precursor RNA. This process could be regulated in a tissue-specific manner (Fig. 7) or in response to different stimuli (Nishiyama et al., 1999; Bassett et al., 2000). However, it is not possible to rule out the presence of alternative polyadenylation or transcription initiation sites.

Further research is in progress to clarify the relation between the two proteins and to elucidate the functionality of the spermidine-binding process.

MATERIALS AND METHODS

Plant Material

Maize (Zea mays hybrid Cecilia, Pioneer Hi-Bred S.p.a., Parma, Italy) seeds were grown in moist vermiculite at 24°C ± 1°C in total darkness for 6 d. Etiolated coleoptiles were excised and directly used for microsome isolation and Southern-blot analysis. For dot-blot analysis, mung bean (Vigna radiata), zucchini (Cucurbita pepo), and oat (Avena sativa) seeds were grown for about 5 to 7 d in the same conditions, and coleoptiles or hypocotyls were excised and directly used for microsome isolation.

Maize plants for northern-blot analysis were grown for 10 d in daylight on clay beads. Primary leaves and stems were then separated with a razor blade before being frozen in liquid nitrogen and stored at −80°C.

Microsome Isolation and Protein Solubilization

All procedures were carried out at 4°C under room light. The plant material was homogenized (Sorvall Omni-Mixer 17106; DuPont Instruments, Wilmington, DE) using 1 volume of grinding buffer (50 mm Tris-HCl, pH 8.0, 250 mm Suc, 2 mm EDTA, 0.1 mm 1, 4-dithiothreitol, and 0.2 mm phenylmethylsulphonyl fluoride). The homogenate was filtered through nylon mesh and was centrifuged at 4,400g for 20 min (2-HS centrifuge, JA-20 rotor; Beckman Coulter, Fullerton, CA). The supernatant was recentrifuged at 48,000g for 45 min (2-HS centrifuge, JA-20 rotor; Beckman-Coulter) to yield a microsomal pellet.

For detergent solubilization, the microsome pellet was resuspended in the binding buffer (50 mm Tris-HCl, pH 8.0, 50 mm Suc, and 2 mm EDTA) at a ratio of 2 g fresh weight mL⁻¹ supplemented with octylglucoside (Calbiochem, La Jolla, CA), Triton X-100 (Pierce, Rockford, IL) or Nonidet P-40 (Pierce). After gentle stirring for 30 min at 4°C, the extracts were centrifuged at 48,000g for 45 min and the supernatant was used for binding assays. Acetone powder protein solubilization was performed according to Venis (1977) and the powders, resuspended in binding buffer at a ratio of 5 g fresh weight mL⁻¹, were centrifuged at 48,000g for 30 min before being used for binding assays or protein purification.

Binding Assays

The standard binding assay (adjusted to a 1-mL final volume with binding buffer) contained protein fraction equivalent to 2 to 3 g fresh weight and 1.85 kBq of [¹⁴C]spermidine (specific activity of 4.07 GBq mm⁻¹; Amersham Biosciences, Piscataway, NJ). Assays were carried out in the absence or presence of 0.1 mm unlabeled spermidine to give total and nonspecific binding (samples A and B, respectively). Samples were incubated on ice for 5 min and were then filtered through glass-fiber filters (Whatman GF/B; Whatman, Clifton, NJ) previously soaked for about 2 h in 0.3% (v/v) polyethylenimine (Sigma, St. Louis; Bruns et al., 1983), and they were then placed in a vacuum filtration unit (model 1225; Millipore, Bedford, MA) connected to a vacuum pump. After rapid filtration, filters were rinsed with 5 mL of binding buffer and were then placed overnight in 5 mL of scintillation cocktail (Ultima Gold; Packard BioScience, Groningen, The Netherlands) before determining radioactivity in a scintillation counter (LS 6500; Beckman-Coulter). Specific binding was obtained by subtracting the radioactivity of sample B from that of sample A. All binding experiments were repeated at least twice with triplicate samples. EBDA program (version 3.0; GA McPherson, Elsevier-Biosoft, Cambridge, UK) was used to calculate the apparent K_d. To confirm polyethylenimine-soaked filter results, an ultracentrifugation binding assay through micropartition filters (Micropartition kit MPS-1; Amicon, Beverly, MA) was used. The assay samples were as described above, except that the final volume was adjusted to 0.7 mL with binding buffer. The binding mixture was incubated for 5 min at 4°C and was then centrifuged at 2,000g for 40 min at 15°C (24 × 14-mL rotor; Europa Bioproducts, Cambridge, UK). The dry filters were placed in the scintillation cocktail, and radioactivity in sample B was subtracted from sample A to determine the specific binding.

Spermidine-Binding Proteins Purification

Acetone powders from about 300 g fresh weight of etiolated coleoptiles were resuspended in the binding buffer at a ratio of 5 g fresh weight mL⁻¹, centrifuged at 48,000g for 30 min at 4°C, and applied to a DEAE-BioGel (Bio-Rad, Hercules, CA) anion-exchange column (about a 10-mL column volume) previously equilibrated in the same buffer. Elution was monitored at A₂₈₀ (Monitor UV-1; Amersham Biosciences), and fractions within each salt step were collected. After an initial elution with binding buffer plus 0.05 m NaCl to eliminate inert proteins, the binding activity was eluted with the same buffer plus 0.2 m NaCl. This fraction was precipitated overnight at 70% saturation of ammonium sulfate, centrifuged for 30 min at 48,000g at 4°C, resuspended in about 8 to 10 mL of binding buffer plus 1 m ammonium sulfate, and loaded onto a 2-mL octyl-agarose (Sigma) hydrophobic column previously equilibrated with binding buffer plus 1 m ammonium sulfate. Low-activity proteins were eluted in decreasing salt steps (0.5 and 0.25 m ammonium sulfate), and the main spermidine-binding activity recovered by elution with binding buffer. This fraction was precipitated overnight at 70% saturation of ammonium sulfate, taken up to 0.5 mL in binding buffer, and injected onto an HiTrapQ FPLC 1-mL column (Amersham Biosciences). Elution was performed at a flow rate of 1 mL min⁻¹ with a 0 to 0.5 m NaCl linear gradient in binding buffer, and 10 fractions of 2 mL each were collected. Protein fractions were separated by 12% (w/v) SDS-PAGE (Laemmli, 1970), the gel was blotted onto a nitrocellulose membrane, and it was stained by the starch-iodide method of Kumar et al. (1985).

For gel-filtration experiments, HiTrapQ fraction 5 (about 25 μg of protein) was incubated for 5 min on ice with about 90 kBq of [¹⁴C]spermidine and was then loaded onto a column (20 × 1.6 cm) of Sephadex G-100 equilibrated and eluted with binding buffer plus 0.1 m NaCl. Protein elution was monitored at A₂₈₀ (Monitor UV-1; Amersham Biosciences), and fractions (0.7 mL) were collected and analyzed for radioactivity in a scintillation counter (LS 6500; Beckman-Coulter). The column was equilibrated with a set of molecular mass markers. SDS-PAGE (12%, w/v) and silver staining (Sammons et al., 1981) were performed on fractions that showed spermidine-binding activity.

In all experiments, protein content was determined following Lowry et al. (1951) with bovine serum albumin as standard.

N-Terminal Protein Sequencing and Database Searching

Active fraction from the HiTrapQ column (fraction 5) was separated by 12% (w/v) SDS-PAGE with Tris-tricine running buffer (upper buffer: 0.2 m Tris-HCl, pH 8.9; tank buffer, 0.1 m Tris-HCl, 0.1 m tricine, and 0.1% [w/v] SDS, pH 8.25), after having pre-run the gel for 30 min with the same running buffer plus 0.2 mm thioglycolic acid, to prevent N-terminal blocking. The proteins were then transferred onto a Immobilon-P (Millipore) polyvinylidene difluoride membrane and stained with Coomassie Blue. The N-terminal sequences were determined for the 60-kD protein at the Babraham Institute (Microchemical Facilities, Cambridge, UK) and for the 18-kD protein at the Alta Bioscience Laboratories (Biochemical School, University of Birmingham, Birmingham, UK).

Protein similarity searching was performed by FASTA and BLASTP programs (The Arabidopsis Information Resource database, Pearson and Lipman, 1988) or FASTA3 and BLASTP2 programs (EMBL, Heidelberg). Protein motif searching was carried out in Pfam (Washington University, St. Louis), Blocks (Fred Hutchinson Cancer Research Center, Seattle), and Emotif (Stanford University, Stanford, CA) databases. Sequence alignment was performed by PILEUP program (University of Wisconsin, Genetics Computer Group package).

Partial cDNA Isolation: Northern- and Southern-Blot Analysis

Total RNA was extracted from 6-d-old etiolated maize coleoptiles as described by Michael et al. (1996), with minor modifications.

On the basis of the translated N-terminal sequences, two different degenerate oligonucleotide probes were designed for 60- and 18-kD proteins (Fig. 5). Primers AL100 or AL102, in combination with NotI(dT)18 [5′-AACTGGAAGAATTCGCGGCCGCAGGAAT(18)-3′] (25 pm each) and 0.5 μg of total extracted RNA, were utilized for RT-PCR with a Ready-To-Go RT-PCR Beads kit (Amersham Biosciences) according to the manufacturer's instructions. Amplification was carried out in a Gene Amp PCR System 2400 (PerkinElmer Instruments, Norwalk, CT) with the following temperature parameters: 30 min at 42°C, 5 min at 95°C, followed by 35 cycles of 1 min at 95°C, 2 min at 57°C[AL100 and NotI(dT)18 primer combination] or 50°C [AL102 and NotI(dT)18 primer combination], 2 min at 72°C, with final 10-min extension at 72°C. Aliquots (0.5–1 μL) of these reaction mixes were used as template for semi-nested PCR amplifications (Ready-To-Go PCR Beads kit) with AL101 or AL103 in combination with NotI(dT)18 primer (15 pm each) with the following temperature parameters: 5 min at 95°C, followed by 35 cycles of 1 min at 95°C, 2 min at 58.5°C [AL101 and NotI(dT)18 primer combination] or 57°C [AL103 and NotI(dT)18 primer combination], 2 min at 72°C, with a final 10-min extension at 72°C.

cDNA fragments were extracted using a QIAquick gel extraction kit (QIAGEN, Hilden, Germany), α[³²P]dCTP labeled by oligo-labeling (Ready-To-Go DNA Labeling Beads; Amersham Biosciences), and used as probes for northern- and Southern-blot analysis.

Total RNA and genomic DNA for northern and Southern blots were extracted from approximately 1.5 g of tissue, as described by Michael et al. (1996) with minor modifications. For Southern-blot analysis, about 30 μg of genomic DNA extracted from etiolated maize coleoptiles was digested overnight at 37°C with about 80 to 100 units of different restriction enzymes: NotI (Boehringer Mannheim, Summerville, NJ), KpnI (Promega, Madison, WI), EcoRI (Boehringer), and HindIII (Promega).

Northern- and Southern-blot analysis was performed by standard procedures (Sambrook et al., 1989) using Hybond-N nylon membranes (Amersham Biosciences). For northern blots, equal loading was checked by ethidium bromide staining and by hybridization with an 18S rRNA probe (kindly provided by Dr. Marianne van Buuren, Agricultural Faculty, University of Bologna, Bologna, Italy). Prehybridization (2 h) and hybridization (16 h) were carried out at 42°C following Sambrook et al. (1989). Membranes were washed at high stringency for 2 h at 65°C and hybridized membranes were exposed to Biomax film (Eastman-Kodak, Rochester, NY) at −80°C for 15 min to 72 h. Densitometric analysis of exposed films was performed by image analysis software (Phoretix International, Newcastle-upon-Tyne, UK).

For cross hybridization, a blot with about 100 ng of the cDNA fragment related to the protein of 60 kD was performed, and the membrane was hybridized with the cDNA fragment related to the 18-kD protein.

Polyclonal Antiserum Production and Evaluation

Active fractions from HiTrapQ FPLC column chromatography were dialyzed overnight against 32-fold-diluted phosphate-buffered saline (10 mm Na₂HPO₄, 1.8 mm KH₂PO₄, 0.13 m NaCl, and 3 mm KCl, pH 7.4 with HCl), freeze dried, and resuspended in water to give the 1× phosphate-buffered saline concentration. Several preparations of about 300 g fresh weight were utilized to give about 170 μg of total purified protein. A single rabbit was immunized with 70 μg of protein, and three successive boosts of about 35 μg were given at 2-week intervals. The preimmune serum bleed was collected 6 d prior to the primary immunization.

To determine antibody titer, ELISA plates (Nunc Immuno II plates; Nalge Nunc International, Naperville, IL) were coated with 0.05 μg of purified protein (fraction 5 HiTrapQ column), and different dilutions of immune and preimmune sera were used as primary antibody. To test the reactivity of the FPLC fractions, the plates were coated with 50 μL of fraction 1 to 10 from the HiTrapQ column, and a serum dilution of 1:900 was used. For both assays, a 1:1,000 dilution of the second antibody goat anti-rabbit horseradish peroxidase-conjugated IgG (Sigma) was used. After development with 3,3′-5,5′-tetramethylbenzidine (Harlow and Lane, 1988), A₄₅₀ was determined using an ELISA reader (2550 EIA; Bio-Rad).

Three milliliters of the rabbit antiserum was adsorbed to a 1-mL HiTrap Protein A column (Amersham Biosciences), and IgG were crosslinked to the matrix by 20 mm dimethylpimelimidate in 0.2 m sodium tetraborate (pH 9.0) buffer (Harlow and Lane, 1988). Non-crosslinked antibodies were eluted from the column with 0.1 m Gly-HCl (pH 2.5). For affinity purification, acetone powders of maize coleoptiles (about 80 g fresh weight) or zucchini hypocotyls (about 150 g fresh weight) were partially purified on a DEAE column: 0.05 to 0.2 m NaCl elution for maize and 0.2 m NaCl elution for zucchini, both in binding buffer. The DEAE fractions were precipitated overnight at 70% saturation of ammonium sulfate, resuspended to a volume of 20 mL with TBSTw buffer plus 1% (v/v) fish gelatin (TBSTw: 25 mm Tris-HCl, 0.13 m NaCl, 3 mm KCl, and 0.1% [v/v] Tween 20, pH 7.4), and loaded onto the Protein A-antibody affinity column previously equilibrated with the same buffer. Two washes with TBSTw, supplemented respectively with 0.2 m and 0.5 m NaCl, were performed before the final elution with 0.1 m Gly-HCl (pH 2.5). Maize and zucchini Gly eluates were analyzed by 12% (w/v) SDS-PAGE.

For dot-blot assays, total proteins were solubilized from different plants by the acetone powder method (Venis, 1977), and about 0.3 μg of protein was dot blotted onto nitrocellulose. Immune and preimmune sera were used on separate membranes at 1:500 dilution. The membranes were developed by biotin/avidin-horseradish-peroxidase reaction (1:1,000 dilution of goat anti-rabbit IgG biotin conjugate and avidin-peroxidase conjugate antibodies; Sigma) as stated by Bauly et al. (2000).

Upon request, all novel materials described in this publication will be made available in a timely manner for noncommercial research purposes.