Differential activity of the promoters of ADC1 and ADC2
The existence of two ADC genes in Arabidopsis (Galloway et al., 1998) is in agreement with the duplicated status of its genome (Arabidopsis Genome Initiative, 2000). The stabilization of genome organization has provided the opportunity for functional divergence of the two paralogues. Although the protein sequences of ADC1 and ADC2 show a high degree of homology, the possibility of divergence in terms of substrate specificity and enzymatic regulation cannot be discounted. Biochemical analysis of mammalian ADC has shown that it could use both ornithine and arginine, although the specific ODC inhibitor, difluoromethylornithine, had no effect on enzyme activity with either substrate (Regunathan & Reis, 2000). Previous studies have emphasized the hypothesis that Arabidopsis ADC1 and ADC2 may be targeted to different subcellular localizations (Hanfrey et al., 2001). This implies action on distinct metabolic pools and thus strongly suggests important functional differences. In addition to possible variation of protein and enzyme functions between ADC1 and ADC2, important developmental and stress-response modifications of mRNA expression between ADC1 and ADC2 in Arabidopsis have been reported recently (Soyka & Heyer, 1999; Perez-Amador et al., 2002; Piotrowski et al., 2003).
Many studies have insisted on the importance of post-transcriptional and post-translational regulation of ADC accumulation and activity (Malmberg & Cellino, 1994; Borrell et al., 1996; Watson & Malmberg, 1996). Thus in Arabidopsis plantlets stressed by potassium deficiency, ADC mRNA did not correlate with increase of ADC enzyme activity (Watson & Malmberg, 1996). However, in other situations, such as response to acid and salt stresses, ADC mRNA levels were found to correlate with ADC activity (Perez-Amador et al., 1995; Chattopadhyay et al., 1997). However, analysis of ADC1 and ADC2 promoters showed striking differences, such as various cis-acting regulatory elements, and the presence of a transposable element specifically in the promoter of ADC1 (El Amrani et al., 2002). In the present study analysis of homozygotic pADC::GUS transgenic lines confirmed that in imbibed seeds, seedlings, roots, stems, leaves and flowers under temperate conditions, pADC1 activity was significantly more development- and tissue-limited than pADC2, which had a more general pattern of activity. The more restricted pattern of activity for pADC1 may be related to the presence of the transposable element through disruption of cis-acting regulatory elements, as has recently been shown for the Zea mays a1 gene (Pooma et al., 2002), or through epigenetic regulation. Introduction of the pADC1::GUS transgene resulted in over 40% of silenced lines, whereas pADC2::GUS lines all showed reporter gene activity. The ADC1 promoter is likely to be more sensitive to epigenetic regulation as its sequence contains a transposable element that is present in nearly 2000 copies in the Arabidopsis genome (El Amrani et al., 2002). In plants and animals DNA methylation is involved in heritability and flexibility of epigenetic states, and it has been shown that transposable elements are the primary targets of genomic DNA methylation (Okamoto & Hirochika, 2001). As insertion of transposable elements may influence expression of neighbouring genes via DNA methylation, epigenetic regulation may be one of the mechanisms controlling ADC1 expression.
However, situations of highly contrasting promoter activity were shown to correlate with corresponding variation of mRNA levels. Thus Piotrowski et al. (2003) have reported ADC2 mRNA to be at much higher levels than ADC1 mRNA in roots. This is, at least partially, consistent with the differential activities of ADC1 and ADC2 promoters in roots of plants grown under temperate conditions (Fig. 3b,d). By contrast, in young leaves and rosette leaves both pADC1 and pADC2 showed significant activity (Fig. 3a,d), and both ADC1 and ADC2 mRNA have been shown to be present (Perez-Amador et al., 2002; Piotrowski et al., 2003). Detection of pADC1 and pADC2 activities in inflorescences was also associated with the detection of ADC1 and ADC2 transcripts in flowers and siliques (Soyka & Heyer, 1999). Thus these various relationships between ADC1 and ADC2 promoter activities and mRNA levels strongly suggested that both promoter sequences were functional. Moreover, the reporter gene strategy had previously shown that pADC1, but not pADC2, was highly active in pollen grains (El Amrani et al., 2002). This was partially correlated with the presence of the sequence AAATGA, which has been described as capable of driving pollen-specific expression independently of orientation (Weterings et al., 1995). This sequence was present six times in both orientations in the ADC1 promoter, and only once in the ADC2 promoter. This provided strong evidence that evolution had selected divergent cis-acting regulatory elements in the promoter sequences of the two ADC paralogues, and that these divergent elements were functional. Moreover, this strongly indicated that the presence of the transposable element did not interfere negatively with this pollen-specific expression, and may even contribute to pollen-specific expression through the presence of an additional AAATGA sequence (Fig. 1). Finally, the striking differences of tissue-specific promoter activity clearly showed that ADC genes were not transcribed at a basic level in all tissues, with subsequent regulation at post-transcriptional and post-translational levels, and correlatively suggested that transcriptional control may be important for ADC gene expression.
High ADC2 gene expression during seed germination and seedling growth under temperate conditions
Seed maturation is followed by embryo developmental arrest during seed dehydration; on germination, embryo arrest is lifted and cell division resumes (Raz et al., 2001). These developmental stages are highly regulated and are of great importance in the life cycle of monocarpic species such as Arabidopsis. The ADC paralogues were found to show contrasting promoter activities during these developmental stages. None of these genes showed activity during seed maturation and dehydration. During germination, which starts with uptake of water by the quiescent dry seed and terminates with emergence of the embryonic axis (Bewley, 1997), no ADC1 promoter activity was observed, whereas ADC2 promoter activity was high as early as 10 h after seed imbibition, and before primary root emergence (Fig. 2b). This correlated with the detection of significant levels of ADC2 mRNA in imbibed seeds (Fig. 6a). It has been shown that germination of barley seed was promoted by addition of exogenous polyamines (putrescine, spermidine and spermine), and it is suggested that endogenous polyamines may play a growth-promoting role complementary to ethylene in the normal course of barley germination (Locke et al., 2000). At the end of the germination process, ADC2 promoter activity was high in the emerging radicle. Involvement of ADC in germination and radicle extrusion is in agreement with the general idea that ADC is active in elongating cells (Nam et al., 1997). The specific reason why the ADC2 gene is upregulated during this process remains to be elucidated.
Under temperate conditions, pADC2 was highly active in seedling and plantlet roots, which is in accordance with the higher level of ADC2 transcripts in roots (Piotrowski et al., 2003). A high level of ADC transcripts in roots is also in general agreement with high ADC activity in roots (Hanfrey et al., 2001) and a high level of putrescine in roots (Watson et al., 1998; Piotrowski et al., 2003). Moreover, the pattern of pADC2::GUS activity reveals a tight relationship with rhizogenesis processes. Lateral root formation involves the stimulation of pericycle cells to proliferate and create a new root meristem. Formation of the lateral root primordium may be divided into different stages that can be characterized by histology, cell division patterns, and gene expression (Malamy & Benfey, 1997). Activation of pADC2 was found to be induced at stage I of primordium development, before the first periclinal division occurs. At later stages of development, strong GUS staining was observed in the whole primordium. Arabidopsis spe mutants, which are characterized by low levels of ADC activity, are deeply affected in root development, with enhanced lateral root formation in single mutants and a compact root system in double mutants (Watson et al., 1998). A number of previous studies have associated ADC activity with root growth and root branching (Biondi et al., 1993; Watson et al., 1998; Hummel et al., 2002). The present study shows that this association is correlated with activation of the ADC2 promoter in the early stages of lateral root formation. However, to our knowledge no study has yet addressed the mechanisms of polyamine action on root development.
Activation of both ADC1 and ADC2 promoters was found in the aerial parts of plantlets, with pADC2 showing much higher activity than pADC1. ADC1 and ADC2 mRNA levels have been reported to be equivalent in plantlet leaves (Piotrowski et al., 2003). Similarly, ADC1 and ADC2 mRNA levels in whole plantlets showed much less contrast (Fig. 6b) than the differences in promoter activity (Fig. 4i–n). Thus important post-transcriptional regulation probably acts on ADC1 and ADC2 transcripts. Expression of ADC2 gene in stems is in accordance with increased ADC mRNA levels and increased ADC activity in hypocotyls of germinating soybean (Nam et al., 1997). Expression of ADC1 and ADC2 genes in leaves is in accordance with the physiological importance of polyamines in leaves for photosynthetic activity (Chang et al., 2000). At least one of the ADC proteins is likely to be targeted to the chloroplast (Perez-Amador et al., 2002), and the oat ADC has been shown to be localized in the chloroplast (Borrell et al., 1995). Moreover, light was found to be a strong inducer for ADC1 and ADC2 promoter activity, with no or little promoter activity during dark growth (Fig. 4a–n). This strong induction by light was consistent with the presence of many putative light-responsive cis-acting regulatory elements in the promoters of both ADC1 and ADC2. Similarly, the promoter region of the carnation ADC gene is rich in light-responsive, cis-acting regulatory elements, and ADC transcripts increase tenfold after light exposure (Chang et al., 2000). In the case of ADC2, promoter activity was also increased in the dark in the presence of exogenous sucrose, especially in roots (Fig. 4a–d). This was consistent with the presence of numerous sucrose-responsive, cis-acting regulatory elements in the promoter of ADC2. It is thus tempting to speculate that this induction by sucrose, resulting in high pADC2 activity in roots, may be part of shoot–root relationships during light growth of plants through the transport of sucrose from shoot to root.
The pattern of pADC2 activity during seedling development was found to be affected in dominant ethylene-mutant backgrounds, even in the presence of light and sucrose, which normally resulted in high pADC2 activity (Fig. 3). Gallardo et al. (2002) have recently highlighted the importance of S-adenosylmethionine synthesis for seedling development of Arabidopsis. This is consistent with an essential role of endogenous ethylene during seedling development, which has also been described in other species (Petruzzelli et al., 2000). Thus ADC2 would be part of an array of genes activated by ethylene during seedling development.
High ADC1 gene expression in response to chilling
Chilling, which has been shown to increase ADC activity and polyamine levels (Lee 1997; Shen et al., 2000; He et al., 2002) had a strong effect on ADC1 and ADC2 promoter activity. Strikingly, growth under chilling conditions thoroughly modified the respective patterns of pADC1 and pADC2 activity, with pADC1 and pADC2 becoming, respectively, highly active and poorly active in roots (Fig. 4i–p). This strong effect appeared to be a specific response to chilling rather than a general response to stress, as salt treatment did not change the patterns of ADC1 and ADC2 promoter activity. The chilling effect was correlated to changes in mRNA level, and consistent with the specific presence of two copies of a low-temperature response element in the promoter of ADC1 and with the potential impact of the transposable element on gene expression, as a copy of this low-temperature-response element is part of the ADC1 transposable element (Fig. 1). Moreover, changes in temperature may also affect epigenetic control of ADC1 promoter activity. Further work should therefore determine whether this ADC1/ADC2 functional divergence derives directly from the presence of the transposable element in the ADC1 promoter.
The functional role of ADC activity in response to chilling is highlighted by previously reported correlations between agmatine accumulation and frost resistance in wheat (Racz et al., 1996), and by the involvement of polyamines and increased ADC activity in chilling tolerance of cucumber (Shen et al., 2000). In Arabidopsis, D-Arg inhibition of ADC drastically reduced plantlet development at low temperature and caused symptoms of chilling injury (data not shown). Thus in Arabidopsis the polyamine response to chilling is shown to correlate with transcriptional activation of the ADC1 promoter. This switching effect of chilling on ADC1 and ADC2 relative expression therefore provides a good experimental model for understanding the specific roles of ADC1 and ADC2.