Evidence for Autoregulation of Cystathionine γ-Synthase mRNA Stability in Arabidopsis

See allHide authors and affiliations

Science  12 Nov 1999:
Vol. 286, Issue 5443, pp. 1371-1374
DOI: 10.1126/science.286.5443.1371


Control of messenger RNA (mRNA) stability serves as an important mechanism for regulating gene expression. Analysis ofArabidopsis mutants that overaccumulate soluble methionine (Met) revealed that the gene for cystathionine γ-synthase (CGS), the key enzyme in Met biosynthesis, is regulated at the level of mRNA stability. Transfection experiments with wild-type and mutant forms of the CGS gene suggest that an amino acid sequence encoded by the first exon of CGS acts in cis to destabilize its own mRNA in a process that is activated by Met or one of its metabolites.

Genetic studies of metabolic pathways in bacteria and yeast have revealed important regulatory mechanisms. For example, studies of amino acid biosynthesis operons in bacteria led to an understanding of mRNA attenuation (1), and the histidine biosynthesis pathway of yeast led to an understanding of the complex interplay between general and pathway-specific controls (2). With the exception of tryptophan biosynthesis inArabidopsis (3), genetic methods have not been extensively used to analyze amino acid biosynthesis in plants. To study the molecular mechanisms for regulation of methionine biosynthesis in plants, we used Arabidopsis mutants, termed mto1, that overaccumulate soluble Met (4).

Met, a sulfur-containing amino acid, functions not only as a protein component but also as a precursor of S-adenosylmethionine, the primary methyl donor in many transmethylation reactions and, in plants, a precursor of the phytohormone ethylene (5). Met is an essential dietary amino acid for mammals. Studies with the aquatic plant Lemna have shown that the cellular concentration of soluble Met remains unchanged over a 3000-fold range in sulfur availability (6), indicating that Met biosynthesis is tightly regulated in plants. Cystathionine γ-synthase (CGS) catalyzes the first committed step in Met biosynthesis, and it has been suggested to be a key regulatory site of the pathway (5). Indeed, CGS activity in Lemna and barley is regulated positively and negatively in response to the availability of Met (7).

Analyses of CGS expression in mto1-1 mutant plants revealed that the steady-state levels of CGS mRNA, protein, and enzyme activity are three- to fivefold higher than in wild-type plants (Fig. 1, A to C). Application of Met to wild-type plants reduced the amount of mRNA for CGS, whereas no such effect was observed in the mto1-1 mutant (Fig. 1D). This suggests that wild-type plants down-regulate the level of CGS mRNA in response to exogenous Met, and this regulation is impaired in themto1 mutant.

Figure 1

CGS expression in wild-type (WT) and mto1-1 mutant (mto1). (A to C) Levels of CGS mRNA (A), protein (B), and enzyme activity (C) in leaves of 3-week-old plants (27). Total RNA (10 μg) was analyzed by RNA blot hybridization (21) with CGS cDNA used as a probe (28). The membrane was rehybridized (26) with a control ubiquitin (UBQ) cDNA (28) (A). Protein extracts (2 μg) were subjected to immunoblot analysis (21) by using rabbit antiserum to CGS. CGS protein migrated at about 50 kD. Asterisk indicates a band that was also detected with the control serum (B). CGS activities (29) relative to wild-type activity (5.28 milliunits/mg) are shown. Error bars indicate standard deviations of three independent samples (C). (D and E) Effect of Met on CGS mRNA accumulation. Various concentrations of Met were applied to plants (27) (D) or callus culture (8) (E) for 3 days and total RNA (10 μg) was analyzed as in (A). The 3′ and 5′ regions of CGS cDNA (28) were also used as probes in (E). (F and G) Time course of Met effect in callus culture (8). Samples were withdrawn at time points as indicated after Met (0.3 mM) alone (F) or Met (0 or 0.3 mM) and ActD (100 μg/ml) (11) treatments (G), and total RNA (10 μg) was analyzed as in (A). Control in (G) is ethidium bromide staining of 26S ribosomal RNA. Arrowheads indicate positions of truncated mRNA (E to G). Representative results of at least triplicate experiments are shown (A to G).

A liquid callus culture system (8) was used for further studies. As with whole plants, the steady-state level of CGS mRNA was reduced by feeding calli from wild-type plants with Met (Fig. 1E), although the response was less significant than in whole plants (9). In addition to an overall reduction in CGS mRNA, a minor band about 500 bases shorter than the full-length transcript was evident in Met-treated wild-type calli (9). Two lines of evidence indicate that this minor mRNA species is a form truncated at the 5′ end of the transcript. The minor band is observed after polyadenylate selection (10) and a probe covering the 3′ untranslated region hybridizes to it, whereas one covering the 5′ region does not (Fig. 1E). This truncated transcript may be an intermediate in the degradation of CGS mRNA (see below).

Time course studies showed that the level of CGS mRNA in wild-type calli, but not in mto1-1 mutant calli, is reduced within 2 hours after Met treatment and, simultaneously, the truncated transcript appeared (Fig. 1F). To determine whether the reduced accumulation of CGS mRNA in wild type is subject to transcriptional or posttranscriptional regulation, we studied mRNA turnover after blocking transcription by treating calli with actinomycin D (ActD) (11) (Fig. 1G). In the absence of applied Met, turnover of CGS mRNA was faster in wild type than in the mto1-1 mutant. Met treatment accelerated the turnover in wild type but not in themto1-1 mutant (12), which indicates that the regulation involves a posttranscriptional event. The truncated transcript was also observed in wild-type calli after Met treatment and declined as the main band decayed.

The mto1 mutation and the CGS gene mapped close to each other on chromosome 3 (13). Sequence analyses of fivemto1 mutants (14) revealed single base changes in the CGS coding region, giving rise to alterations in the amino acid sequence (Table 1). The mutations were clustered in a small region of eight amino acids located about 80 residues from the NH2-terminus, with two of the independent mutations being identical. Hereafter, the wild-type amino acid sequence defined by the mto1 mutations is referred to as the MTO1 region and the corresponding nucleotide sequence isMTO1.

Table 1

Nucleotide and amino acid changes in mto1and silent mutations. The mto1-3 and mto1-5mutants are independent of each other because they were isolated from different batches of mutagenized population.

View this table:

The role of the MTO1 region was studied by transient expression experiments. The coding region from exon 1 (amino acids 1 to 183) of CGS (15), with or without mto1 mutations, was fused in-frame to the 5′ end of the Escherichia coliβ-glucuronidase (GUS) reporter gene. The constructs were placed under the control of a cauliflower mosaic virus (CaMV) 35S RNA promoter and used in transfection experiments. Reporter activity was lower for the construct carrying wild-type exon 1 than for those carrying mto1 mutations and reporter activity was repressed by incubation with Met (Fig. 2A). In contrast, reporter activity was insensitive or less sensitive to Met treatment for the constructs carrying mto1 mutations (16) or for a construct carrying only the first four amino acids of exon 1 (Δ5–183) (17). Thus, wild-type exon 1 contains a sequence that is both necessary and sufficient for down-regulation of reporter gene activity in response to applied Met.

Figure 2

Role of CGS exon 1. We used cells from liquid callus cultures (8) for these experiments. (A) Effect of wild-type and mutant exon 1 on reporter activity in a transient expression system. We transfected wild-type protoplasts (30) with 10 μg of plasmid carrying the GUS reporter. The plasmids carried CGS exon 1 from either wild type (WT) or mutants as indicated (31). Transfected protoplasts were incubated for 48 hours with (filled box) or without (hatched box) 0.1 mM Met (9). A control plasmid (10 μg) carrying the LUC reporter (31) was cotransfected as an internal standard, and GUS activities were normalized with LUC activities. The GUS/LUC values relative to the wild-type exon 1 construct in the absence of Met are shown. Error bars indicate standard deviations of at least five experiments. (B) Cotransfection experiments. Wild-type protoplasts were transfected with 10 μg each of two plasmids carrying wild-type (WT) ormto1-1 mutant (mto1) CGS exon 1 fused to GUS or LUC reporters (31). Cotransfection was carried out in all four combinations. After 48 hours of incubation, GUS and LUC activities were determined and normalized with protein content. Reporter activities relative to those of wild-type exon 1 combination are shown. Error bars indicate standard deviations of three to five experiments.

To test whether the nucleotide or amino acid sequence is important for this regulation, we mutated the sequence of exon 1 by introducing base changes into the MTO1 region that do not alter the amino acid sequence (silent mutations) (Table 1). Transfection experiments showed that the silent mutants behaved as did the wild type (Fig. 2A), which suggests it is the amino acid sequence that plays a role in regulation.

We examined whether the MTO1 region functions in cis or in trans by cotransfecting plasmid constructs carrying different reporter genes and exon 1 of CGS from wild type or the mto1-1 mutant. Neither the wild-type nor mutant exon 1 affected the reporter activity of the other (Fig. 2B), which indicates that the mutation acts in cis.

Comparison of the CGS amino acid sequence from four plant species showed that the region encompassing exon 1 of ArabidopsisCGS has a low overall homology and variable length (Fig. 3A). A notable exception was a stretch of 38 amino acids that includes the MTO1 region (Fig. 3B). Conservation of the MTO1 region among widely different plant species suggests that it plays a functional role (18).

Figure 3

Alignment of CGS amino acid sequence. (A) We aligned the deduced amino acid sequences of the complete coding region for CGS from Arabidopsis (GenBank database accession no. AB010888), soybean (Glycine max; accession no. AF141602), common ice plant (Mesembryanthemum crystallinum; accession no. AF069317), and maize (Zea mays; accession no. AF007786) by using the CLUSTALW program (32). The alignment was scanned with a window of 11 amino acids, and the identical (filled box) or similar (shaded box) amino acids among the four plant species were counted. (B) Alignment of amino acid sequence for the highly conserved region covering the MTO1 region. Those amino acids that are identical (reversed) or similar (shaded) among the four plant CGS are marked. At,Arabidopsis; Gm, soybean; Mc, common ice plant; Zm, maize. Positions of mto1 mutations are marked with asterisks.

These results suggest that the exon 1 polypeptide of CGS acts in cis to down-regulate its own mRNA stability in response to excess Met. Although it is unusual for a polypeptide to act in cis, a plausible explanation is that the regulation occurs during translation when the nascent polypeptide and its mRNA are in close proximity. A model for such a regulation mechanism predicts a role for the exon 1 polypeptide of CGS in destabilizing its own mRNA during translation and an activation of this process by Met or one of its metabolites (19). Alterations in the amino acid sequence of the MTO1 region abolish either of the reactions. The mechanism may generate an intermediate mRNA species with a truncated 5′ end (Fig. 1E).

The β-tubulin gene in animals (20) has a similar mechanism in that an amino acid sequence acts in cis to regulate its own mRNA stability, although the responsible amino acid sequences are placed differently between the two systems. The stability of β-tubulin mRNA is down-regulated by the unassembled β-tubulin subunits, with the NH2-terminal tetrapeptide of nascent β-tubulin being responsible for the regulation (20). The presence of another system reported here, together with the genetics ofArabidopsis and identification of a possible degradation intermediate, will help in defining the molecular mechanisms of this mode of mRNA regulation.

Much is known about the feedback regulation at the level of enzyme activity in metabolic pathways in plants (5), but very little is known about feedback regulation at the level of gene expression. The system reported here adds another mechanism to the repertoire of metabolic controls (1, 2, 5).

  • * Present address: Fukujuen CHA Research Center, Kyoto 619-0223, Japan.

  • Present address: Laboratory of Molecular Phylogenetics, Institute of Molecular and Cellular Biosciences, The University of Tokyo, Tokyo 113-0032, Japan.

  • To whom correspondence should be addressed. E-mail: naito{at}


View Abstract

Navigate This Article