Introduction

Polyamines, including diamine putrescine, triamine spermidine and tetraamine spermine, are aliphatic nitrogen compounds distributed widely from bacteria to higher plants 1, 2 and have been implicated to play important roles in growth and development 3. At the cellular level, polyamines are organic cations, interacting with the macromolecules that possess anionic groups such as DNA, RNA, lipids and proteins, thereby influencing DNA conformation, gene expression and protein synthesis, and modulating enzyme activity. In higher plants, polyamines are able to affect membrane fluidity by binding to phospholipids in membrane 4 and mediate biotic and abiotic stress responses, such as pathogen infection, osmotic stress, potassium deficiency and wounding 5, 6, 7, 8, 9. Previous observations revealed that polyamines may be involved in a variety of plant developmental processes, such as cell division, root initiation, somatic embryogenesis, xylogenesis, flower development, fruit ripening and senescence 3. Recent studies have indicated that polyamines also affect the formation of plant architecture, such as internode elongation 10, 11, root branching 12 and shoot apical dominance 13.

The plant polyamine biosynthetic pathway is relatively simple 14. Putrescine is derived either from ornithine catalyzed by ornithine decarboxylase (ODC) or from arginine through several steps catalyzed by arginine decarboxylase (ADC), agmatine iminohydrolase and N-carbamoylputrescine amidohydrolase. Spermidine and spermine are synthesized from putrescine through spermidine and spermine synthases (SPDS and SPMS) from the donor of decarboxylated S-adenosylmethionine (dcSAM), which is produced from S-adenosylmethionine (SAM) by the action of S-adenosylmethionine decarboxylase (SAMDC).

Although application of inhibitors is very useful in studying the polyamine biosynthetic pathway and in exploring the functions of polyamines, isolation of mutants defective in polyamine biosynthesis can provide us a more powerful tool to understand their roles in plant growth and development, as the cases in yeast 15, 16, Leishmania donovani 17 and mouse embryonic cells 18, in which mutations in SAMDC cause an absolute requirement of spermidine and/or spermine in the media for survival. In Arabidopsis thaliana, mutants defective in ADC, SPMS and SPDS have been identified 11, 19, 20, 21, 22, 23, 24. However, no mutant defective in SAMDC has been reported yet, although SAMDC cDNAs were cloned and their transgenic plants have been generated 25, 26, 27, 28, 29. In this paper, we report the isolation and characterization of an Arabidopsis bushy and dwarf mutant, bud2. BUD2 encodes SAMDC4, a member of the SAMDC family. Our data indicate that SAMDCs are essential for plant embryogenesis and polyamines play an important role in plant growth and development.

Materials and methods

Plant growth, plasmid construction and plant transformation

Arabidopsis thaliana ecotype Columbia-0 (Col-0) wild type and mutant plants were grown on vermiculite saturated with 0.3 × B5 medium under continuous illumination (80-120 μE m−2 s−1) at 23 °C as described by Mou 30. For root observation, seeds were surface sterilized and vernalized at 4 ºC for 4 days and seedlings were grown vertically at 23 °C for 22 days.

For the complementation test, a 3.5-kb entire BUD2 gene (1.5-kb promoter, 1.044-kb coding sequence and 1.0-kb 3′ region) was amplified by PCR using the primers of PFB (5′ATG ATG AGGGATCC ATA A3′) and PRK (5′TAG GGTACC AGT TTC AAG TTA TGC3′), and the PCR product was cloned into the binary vector pCAMBIA1300 digested with BamHI and KpnI. For the antisense BUD2 construct, BUD2 cDNA was amplified with primers 5′CTC TTC TCT CTC TAC TCA CAC AC3′ and 5′GCG TGA CAT TGC CAC AAG TC3′, and ligated into the pGEM-T Easy vector (Promega, USA). The EcoRI-digested BUD2 cDNA was inserted into the pJL453 vector in an antisense orientation, and then cloned into pCAMBIA1300. The constructs were introduced into Agrobacterium tumefaciens strain GV3101 (pMP90RK) by electroporation and transformation was performed as described previously 31. Transformants were selected on 0.5 × MS agar plates supplemented with 40 mg L−1 hygromycin and 50 mg L−1 kanamycin. T2 plants were self-pollinated, and the T3 homozygous plants with a single insertion locus were used for detailed analyses.

DNA fiber FISH

Leaf nuclei were prepared according to the method described previously 32. A 2-mL suspension of nuclei was deposited at an end of a poly-L-lysine slide (Sigma, St. Louis, USA) and air dried for 5-10 min. Eight microliter of STE (0.5% (w/v) SDS, 5 mM EDTA, 100 mM Tris, pH 7.0) lysis buffer was pipetted on top of the nuclei, and slide was left at room temperature for 4 min. A clean coverslip was slowly dragged through samples. The slide was air dried, fixed in 3:1 ethanol: glacial acetic acid for 2 min and baked at 60 °C for 30 min. The BAC clone DNA was isolated, and labeled with either biotin-11-dUTP or digoxigenin-16-dUTP (Roche, Switzerland) by nick translation. The probe mixture (20 μL) was applied to the DNA fiber preparation, covered with a coverslip. The slide was placed in an 80 °C oven for 3 min and transferred to 37 °C for overnight incubation. Post-hybridization washing and signal detection were performed accordingly. Fiber FISH signal images were captured under the Olympus BX61 fluorescence microscope conjunct with a microCCD camera. Grayscale images were captured for each color channel, and then merged using the software of Image-Pro Plus.

RNA isolation, RACE-PCR and RNA gel blot

Total RNA was isolated with TRIzol reagent (Invitrogen, USA). The full-length BUD2 cDNA was amplified with a RACE Kit (Clontech, USA). The first strand cDNA was synthesized from 2.5 μg total RNA by using a cDNA synthesis kit (Promega, USA). RNA gel-blot analysis was carried out as described previously 33 with specific probes for each gene (Supplement 1).

Yeast complementation

BUD2 cDNA was digested with NotI, cloned into the yeast expression vector pREP5N 34 and introduced into yeast mutant spe2 cells 35. Yeast cells were cultured as described by Gupta 36. After a single clone was cultured in YPD medium for 24 h, the 10-μL cultured cells were added to 5-mL SD medium with or without supplement of spermidine and were incubated overnight at 30 °C. Cells were then diluted 500 times and A600 was measured at intervals of 4 h.

Measurement of cellulose, lignin and polyamines

Lignin and cellulose contents were assayed according to the methods described previously 37, 38. Analysis of polyamines was carried out according to the method described by Flores and Galston 39 with some modifications. Seedlings of 100 mg were homogenized in 1 mL of 5% (w/v) perchloric acid containing 10 nmol of 1,7-diaminoheptane·2HCl as the internal standard and cooled on ice for 1 h. The homogenate was centrifuged at 4 800 g for 20 min and the soluble polyamines in the supernatant were used for benzoylation. After drying under warm air, the benzoylated polyamines were redissolved in 100 μL of methanol and analyzed by HPLC (Model Agilent 1100, Agilent, USA). A 20-μL sample was applied to a reverse-phase C18 column (particle size 5 μm, 4.6 × 250 mm, Agilent), eluted with acetonitrile in water (a gradient of 36% to 41% for 8 min, followed by a 10-min elution of 41% and a gradient of 41% to 100% for 7 min) at a flow rate of 1 mL/min at room temperature, and monitored at 254 and 360 nm, respectively.

Microscopy and histology

Sections of inflorescences, roots, and petioles were prepared according to the method described previously 40 and stained with Safranin O and Fast Green (Fisher, USA). To observe the vascular patterns of Arabidopsis inflorescences and roots, the fresh hand-cut sections or microtome-sliced sections were observed directly under the UV light at 340-380 nm 41. For histochemical localization of lignin, freshly hand-cut sections (∼20 μm thick) of inflorescences or roots were incubated for 2 min in 2% phloroglucine solution (Sigma, USA), mounted in 50% HCl 42 and photographed with a Micro Color CCD camera (Apogee Instruments INC, USA).

Embryos were observed according to the method described previously 43. Seeds of 1-12 day-old were removed from the siliques and observed after transparency treatment under the Olympus BX61 differential interference contrast microscope conjunct with a CCD camera.

Results

Isolation and morphological characterization of bud2 mutant

Two mutants with obviously bushy and dwarf phenotype after flowering were isolated from a mutant collection generated with a sense/antisense RNA expression system (SARE) 44 and were designated as bud1 45 and bud2-1 (Figure 1), respectively.

Figure 1
figure 1

Morphology of bud2 and related mutants. (A) to (D) 26-day-old plants of wild type (A), bud2-1 (B), bud2-2 (Salk_007279) (C) and samdc1-1 (D). (E) to (H) 50-day-old plants of wild type (E), bud2-1 (F), bud2-2 (G) and samdc1-1 (H). (I) to (K) Roots of wild type (I), bud2-1 (J) and bud2-2 (K). Plants were grown on 0.5 × MS plates vertically for 22 days. (L) to (N) The primary and lateral roots of wild type (L), bud2-1 (M) and bud2-2 (N). Plants were grown on greenhouse for 50 days. Bars=1 cm.

Full size image

Compared with the wild type, bud2-1 plants displayed smaller and slightly curled leaves in seedlings, but severe morphological alterations were observed in plant height, apical dominance, length of internodes and petioles and the number of lateral roots (Figures 1B, 1F and 1J). Quantitative analysis revealed that the plant height and length of petioles in bud2-1 decreased by ∼57% and 37% compared to those of wild type, respectively. The hypocotyl length of the bud2-1 seedling grown in the dark was comparable with that of the wild type, but reduced to ∼56% when grown under light (data not shown). Moreover, the primary and lateral roots of bud2-1 were twisted and thicker than those of the wild-type plant (Figure 1L and 1M).

Altered vascularization in bud2 plants

To understand the alteration of the mutant plant at the cellular and tissue levels, we compared the anatomy of inflorescences, roots and petioles between bud2 and wild-type plants (Figure 2). In the mature inflorescences of wild-type plants, vascular bundles are mainly composed of phloem and xylem, where the phloem pole at the outer side of the bundle is juxtaposed with the xylem pole at the inner side of the bundle. Vascular bundles are arranged in a circle and connected by interfascicular fibers, forming a hollow cylinder that surrounds a central parenchymatous pith (Figure 2A and 2D) 46. Compared with wild type, the pattern of the vascular bundles and the hollow cylinder of bud2-1 inflorescences were largely affected (Figure 2B and 2E), due to a significant increase in the bundle size resulted from the increase in the number of tracheary vessels, although they were smaller than that of wild-type plants (Figure 2D and 2E).

Figure 2
figure 2

Comparison of vascular patterns of stems, roots and petioles between wild-type and bud2 plants. (A) to (C) Transverse sections of the first internodes of inflorescences of wild type (A), bud2-1 (B) and bud2-2 (C), observed under UV light (340-380 nm). (D) to (F) Transverse sections of the first internodes of inflorescences of wild type (D), bud2-1 (E) and bud2-2 (F), observed under light microscope. (G) to (I) Transverse sections of main roots of wild type (G), bud2-1 (H) and bud2-2 (I), observed under UV light (340-380 nm). (J) to (L) Transverse sections of main roots of wild type (J), bud2-1 (K) and bud2-2 (L), observed under light microscope. (M) to (O) Vascular systems in petioles of wild type (M), bud2-1 (N) and bud2-2 (O), observed under light microscope. (P) and (Q) Lignin and cellulose contents (milligrams per gram of total cell wall residues) of wild-type, bud2-1 and bud2-2 plants. (R) to (W) Wiesner's staining of the transverse sections of stems and longitudinal sections of lateral roots. (R) and (U) wild type; (S) and (V) bud2-1; (T) and (W) bud2-2. Bars in (A) to (O)=100 μm and in (R) to (W)=50 μm.

Full size image

Similar phenomenon was observed in the vascular system of roots (Figure 2G, 2H, 2J and 2K). Unlike the wild-type root, in which there are normally eight layers of cortical cells outside the vascular bundles and the xylem elements were separately arranged (Figure 2G and 2J), the vascular bundles of bud2-1 occupied nearly all the space of root sections with apparently increased size of the xylem elements, which were closely arranged in the bundle (Figure 2H and 2K). Similar vascularization was also observed in the bud2-1 petiole vascular system (Figure 2M and 2N).

Since the main compositions of the vascular system are lignin and cellulose, we therefore measured their contents in bud 2-1 and wild-type plants. The lignin content in bud2-1 increased significantly, ∼30% higher than that in wild-type plants (Figure 2P). However, the cellulose contents between bud2-1 and wild type had no apparent difference (Figure 2Q). The increase of lignin content in bud2-1 plants was further confirmed by the Wiesner's staining of the sections of stems (Figure 2R and 2S) and lateral roots (Figure 2U and 2V). These results indicate that vascular development and lignin biosynthesis are greatly affected in bud2-1.

Cloning and confirmation of BUD2

Because genetic analysis indicated that the phenotypic segregation in the F2 progeny fits the expectation for a single recessive nuclear mutation (wild-type: mutant =71: 25), it is very likely that the mutation in bud2-1 is due to a T-DNA insertion event. DNA gel-blot analysis of an F2 mapping population probed with the 32P-labelled 35S promoter fragment demonstrated that the mutant phenotypes were cosegregated with a 4-kb DNA fragment (data not shown). Tail-PCR analysis revealed that the T-DNA flanking sequence was in the BAC clone F17K4, chromosome 5. However, further analysis with PCR amplification and DNA gel blot demonstrated that a 75-kb genomic DNA fragment was substituted by the inserted T-DNA, which was confirmed by an extended DNA fiber FISH of two BAC clones, T16G12 and F17K4, situated on this region (Figure 3A to 3D). Sequence analysis indicated that the deleted region may contain up to 20 putative open reading frames (ORFs). Among all these ORFs, the SAMDC gene (At5g18930), which is involved in the polyamine biosynthetic pathway, appears to be our candidate gene BUD2, because the transgenic potato plants expressing an antisense cDNA of SAMDC displayed a stunted phenotype with short internodes and small leaves 26; the Arabidopsis acaulis5 (acl5) mutant defective in spermine biosynthesis 11, 23 and overexpression of ADC were also bushy and dwarf 10. Most importantly, another mutant collected from Salk T-DNA insertion stock (Salk_007279), in which T-DNA is inserted in the promoter region of At5g18930, also exhibited the same morphology as bud2-1 (Figure 1C, 1G, 1K and 1N). Further detailed histological examination revealed that the vascular morphology of Salk_007279 was highly similar to that of bud2-1 (Figure 2C, 2F, 2I, 2L, 2O, 2T and 2W).

Figure 3
figure 3

Chromosomal deletion in bud2-1 and complementation of mutants by BUD2. (A) Diagram of the 75-kb deletion in the bud2-1 genome. Red bar represents BAC T16G12 and the green represents BAC F17K4. (B) Florescence hybridization of the BAC T16G12 (red) and F17K4 (green) to extended DNA fibers from wild-type Col-0 plants. (C) Florescence hybridization of the BAC T16G12 (red) and F17K4 (green) to extended DNA fibers of bud2-1, showing a gap in the bud2-1 genome and the truncation in both T16G12 and F17K4. T-DNA substitution for the gap was confirmed in (D). (D) Florescence hybridization of BAC T16G12 (red), T-DNA (green, middle) and BAC F17K4 (green, right) to the bud2-1 extended DNA fibers. (E) Complementation of bud2 mutants by the BUD2 gene. 1, Wild type; 2, bud2-1; 3, bud2-2; 4, complemented bud2-1; 5, complemented bud2-2.

Full size image

To test whether BUD2 is the SAMDC gene, a genomic DNA fragment containing the entire At5g18930 with its 5′ and 3′ regions was introduced into both bud2-1 and Salk_007279 genomes, and the transgenic plants exhibited the wild-type morphology (Figure 3E, Supplement 2). We also generated the transgenic plants expressing an antisense BUD2 cDNA under the control of 35S cauliflower mosaic virus (CaMV) promoter and found that the transgenic plants displayed short internodes and reduced apical dominance as bud2-1 and Salk_007279 (data not shown). All these data demonstrate that it is the SAMDC (Accession number: DQ217414) that causes the phenotypic alterations in the bud2-1 and Salk_007279 plants. We therefore designated Salk_007279 as bud2-2.

BUD2 encodes an isoform of S-adenosylmethionine decarboxylase

To confirm that BUD2 has the enzyme activity of SAMDC, we carried out a functional complementation experiment by expressing BUD2 cDNA in the yeast mutant, spe2, in which the SAMDC gene is disrupted and therefore the supplement of spermidine is required in the medium for its growth 35. The spe2 cells transformed with BUD2 grew normally as did the wild-type cells in the medium without supplement of spermidine, whereas spe2 cells alone or transformed with an empty vector could not grow without the supplement of spermidine (Figure 4A and 4B), indicating that BUD2 does have the enzymatic activity of SAMDC.

Figure 4
figure 4

SAMDC activity of BUD2. (A) Functional complementation of yeast spe2 mutant cells by BUD2. The yeast cells were cultured in the SD medium supplemented with (+) or without (−) spermidine. 5NBUD2, spe2 cells transformed with BUD2; 5N, spe2 cells transformed with vector only. (B) Growth kinetics of SPE2 (triangle), spe2 (solid triangle) and spe2 transformed with BUD2 (5NBUD2, circle), or vector (5N, solid circle). spd, spermidine.

Full size image

In the Arabidopsis genome, SAMDC is a small gene family consisting of four highly homologous but distinguishable genes, SAMDC1 (At3g02470), SAMDC2 (At5g15950), SAMDC3 (At3g25570), and BUD2 or SAMDC4 (At5g18930) 29. Phylogenetic analysis indicated that SAMDC1 and SAMDC2 had the highest (78%) identity in amino acid sequences, whereas the SAMDC4 (BUD2) shared the lowest similarity to other members of SAMDC family (Figure 5A and 5B). Fortunately, another Salk T-DNA line, Salk_020185, was collected, in which the T-DNA was inserted in the 3′ UTR of SAMDC1. We thus named it samdc1-1. The samdc1-1 plant exhibited a similar but even severer phenotype compared to the bud2-1 or bud2-2 plants (Figure 1F to 1H).

Figure 5
figure 5

Arabidopsis SAMDCs and their expressions. (A) Homology analysis of SAMDCs in Arabidopsis. The identical amino-acid residues were shaded in the four SAMDCs. (B) Phylogenetic tree of SAMDCs in Arabidopsis. (C) Expressions of SAMDCs in various organs. 1 to 6 represents root, inflorescence stem, rosette leaf, cauline leaf, flower and silique, respectively. (D) Expressions of SAMDCs in wild-type, bud2-1, bud2-2 and samdc1-1 plants.

Full size image

To explore the roles of SAMDCs, we examined their organ-specific expressions with RNA gel blot. As shown in Figure 5C, SAMDC1 was strongly expressed in all the examined organs with a relatively higher level in siliques, but SAMDC2 was highly expressed in roots, leaves and flowers. SAMDC3 was weakly expressed in all organs. BUD2 (SAMDC4), however, accumulated ubiquitously in all examined organs at a low level (Figure 5C), but was undetectable in bud 2-1 and bud2-2 (Figure 5D), consistent with the phenotypic observations. In samdc1-1, the SAMDC1 mRNA level was apparently decreased, but the BUD2 transcripts were highly increased (Figure 5D). No significant differences at the expression levels of SAMDC2 and SAMDC3 were observed among wild-type, bud2 and samdc1 plants.

Disrupted homeostasis of polyamines in bud2 plants

Because SAMDCs are key enzymes in both spermidine and spermine biosynthesis, we therefore analyzed polyamine contents in wild type and bud2 plants. As shown in Table 1, compared to those in the wild type, the putrescine level in the bud2-1 plant increased by ∼25%, whereas the contents of spermidine and spermine decreased by ∼9% and 16%, respectively. Similar results were obtained from the bud2-2 mutant plant (Table 1). These results indicate that the disruption of BUD2 leads to an abnormal homeostasis of polyamines.

Table 1 Polyamine contents in wild type and bud2 plantsa
Full size table

SAMDCs are essential for embryogenesis

The fact that the complete deletion of SAMDC4 does not lead to the lethality of the bud2 mutant plant suggests that the SAMDC genes might be, at least in part, functionally redundant. To verify their redundancy, we carried out a careful genetic analysis of the progeny derived from the crosses between samdc1-1 and bud2-1 or bud2-2. Among the F2 progeny, the segregated plants with abnormal phenotypes were either SAMDC1samdc1/bud2bud2 or samdc1samdc1/BUD2bud2 genotype (Table 2), indicating that the double mutant samdc1samdc1/bud2bud2 is embryo lethal. We therefore examined the embryo developmental processes in the progeny derived from the self-pollinated SAMDC1samdc1/bud2bud2 or samdc1samdc1/BUD2bud2 plants, and found that about one-quarter of seeds were apparently different from others as early as six DAP (days after pollination). These seeds looked bleached and transparent under microscope and became eventually decayed (Figure 6A to 6C, Table 2). Microscopic examination revealed that the double mutant embryos began to show abnormalities at the globular stage and stopped further development afterwards (Figure 6D to 6M). This is consistent with the finding that the expression level of SAMDC1 is particularly high in siliques (Figure 5C). These results indicate that the SAMDC isogenes are essential for Arabidopsis embryogenesis.

Table 2 Genetic analysis of progeny derived from bud2 and samdc1
Full size table
Figure 6
figure 6

Embryo lethality of the bud2-1samdc1-1 double mutant. (A) to (C) Seed development of bud2-1samdc1-1 and wild-type plants at the stages of 8 days (8 DAP) of wild type (A) and bud2-1bud2-1/SAMDC1-1samdc1-1 (B), and 12 days (12 DAP) of bud2-1bud2-1/SAMDC1-1samdc1-1 (C). Arrows indicate aborted seeds. (D) to (H) Embryos of wild type at the triangle (D), heart (E), torpedo (F), bent-cotyledon (G) and mature (H) stages. (I) to (M) Embryos of bud2-1bud2-1samdc1-1samdc1-1 at the stages as given in (D) to (H). Note that embryo development was arrested at the global stage and the seeds decayed eventually. Bars in (D) to (M)=25 mm.

Full size image

Discussion

Redundancy and essentiality of the SAMDC family

In plants, the genes involved in the polyamine biosynthesis often belong to small families. So far, two members of ADC, SPDS and SPMS genes 6, 19, 24, 47, and four members of SAMDC genes have been identified in the Arabidopsis genome 29. Previous studies indicated that ADC and SPDS members are functionally redundant, because a mutation in either gene member does not cause significant phenotypic changes 20, 21, 24, 47. Nevertheless, the disruption of ACL5, one of SPMS genes, results in a severe dwarfism and an increased organ vascularization 11, 22, indicating that two SPMS genes are not completely redundant during developmental process.

The characterization of bud2 identified a mutation in one member of the SAMDC family that encodes enzymes necessary for the biosynthesis of both spermidine and spermine and is considered to control the rate-limiting step in the polyamine biosynthetic pathway. Bushiness, dwarfism and altered vascular development in bud2-1 and bud2-2 demonstrate that disruption of one of the four SAMDC genes, SAMDC4, apparently affects growth and development, but does not cause lethality, indicating that functions of SAMDCs are partially redundant. This is supported by SAMDC expressions in the samdc1-1 mutant, in which a decreased expression of SAMDC1 is accompanied by the elevated expression of SAMDC4. On the other hand, the four SAMDCs are more likely to play different roles in an organ-specific or responsive manner. Among them, SAMDC1 seems more important in affecting overall developmental processes, because knockdown mutant samdc1-1 not only displays a more severe phenotype but also elicits a compensatory expression of SAMDC4. Moreover, the different expression patterns of SAMDCs in organs also imply that these SAMDCs might be involved in different developmental or responsive processes. However, the embryo lethality of bud2samdc1 indicates that at least some SAMDCs are essential for plant survival. Therefore, further identification of loss of function mutants in SAMDC2 and SAMDC3 is necessary to fully understand their roles in plant growth and development.

Roles of plant polyamines

Previous studies on animal cells have indicated that polyamines play very important roles in cell division and growth. For example, a circadian change of polyamine levels is observed during cell cycle 48 and the translation initiation factor eIF5A's biochemical activity on cellular proliferation is tightly linked to cellular polyamine homeostasis, especially to spermidine modification 49. The observation on ODC knockout mouse suggests that putrescine plays a key role in cell growth 50. Moreover, the decreased level of spermine and spermidine is one of the important factors to induce cell apoptosis 18. In plants, polyamines are not only important for both stress responses and developmental processes 4, 51, 52 but also essential for plant survival, because the double mutants of ADCs, SPDSs and SAMDCs are all embryo lethal 24, 47.

Different polyamines appear to have different functions in multicellular systems. In higher plants, putrescine might be more related to root initiation 53, even though controversy exists 20, 54. Our findings that bud2 plants contain more putrescine and have the enhanced root growth appear to support the argument that putrescine has a promotive effect on root growth. On the other hand, spermidine seems to be more important during developmental process, because it is more abundant in embryo and young organs 55, and most importantly, disruption of two SPDSs in Arabidopsis leads to embryo lethal 24. Similarly, the embryo lethality of bud2samdc1 may result from deficiency in spermidine. Spermine appears to be unessential for plant survival, but it may play a role in the vascular development; because an increased organ vascularization of the double mutant acl5spms1 is observed, but its morphology is similar to that of acl5 19, 22. The enlarged vascular tissues and increased lignin content in bud2 might be attributable to the decreased spermine content. On the other hand, the bushy phenotype of bud2 and samdc1-1 suggests a possible involvement of auxin and/or cytokinin. Further analysis of bud2 and samdc1-1 responses to auxin and /or cytokinin might be helpful to understand the regulatory mechanism of polyamines in plant growth and development.