Antisense RNA techniques were used to repress the expression of GSH2encoding GSH synthetase
[16]. Transgenic Arabidopsisplants expressing the GSH2complementary DNA (cDNA) in the antisense
orientation under the regulation of CaMV 35S promoter were generated; confirmed by genomic DNA,
RNA, and protein gel blot analyses; and analyzed for thiol content. Among the herbicide-resistant T2 seg-
regates containing the anti-GSH2andbarselectable marker gene, the GSH level dropped to about 60%
of the wild-type level. These plants had an average 90-fold increase in -EC levels and a 25% increase in
total thiols compared with the wild type. The herbicide-sensitive segregates, on the other hand, showed a
thiol composition similar to that of wild-type plants [36]. This observation is consistent with GSH con-
trolling its own synthesis by feedback control on -EC synthetase. The lowered GSH levels would relax
the feedback control, resulting in the observed elevation in -EC concentrations as well as total thiols.
These results demonstrate that the same feedback mechanism observed in vitro also works in vivo in
higher plants.
- Substrate Availability
The control by substrate availability and -EC synthetase activity was demonstrated in poplars by Foyer
and colleagues [4,35,37–42]. They showed that the amount of -EC synthetase but not of GSH synthetase
controls the GSH level in these transgenic plants. Glycine, the substrate for GSH synthetase, was shown
to be limiting in the dark. The change of glycine pool size diurnally correlates well with the change of the
GSH pool size. In plants overexpressing -EC synthetase, the amount of GSH can be increased substan-
tially by feeding cysteine, suggesting that under these conditions the availability of this amino acid lim-
its synthesis of GSH.
As a key substrate, cysteine can be an important limiting factor of GSH synthesis. The amino acid
cysteine occupies a central position in sulfur metabolism in higher plants. It serves as a hub linking sul-
fur assimilation and utilization. Accumulating evidence suggests that there is coordination between these
pathways (see later).
D. Temporal and Spatial Regulation
- Expression Pattern of Arabidopsis GSH1 Gene and GSH Synthesis
To appreciate fully the complexity of multilevel regulation, the spatial control of GSH1expression must
be investigated. The expression pattern of GSH1should provide clues to where GSH is synthesized. To
address this problem, the expression pattern of the GSH1promoter was examined in the same Arabidop-
sistransgenic lines used for translational control analysis. The expression of GSH1is developmentally
regulated and tissue specific. In germinating seeds and young seedlings, the GSH1promoter activity is
primarily localized in roots, indicating early onset of GSH synthesis in this organ. As seedlings develop
into the rosette stage, the activity appears in leaves and remains high in roots. In mature plants, the activ-
ity is localized in rosette leaves, flower buds, flowers, and young siliques. Strong GUS staining is always
observed in the root system of young seedlings and in mature roots and is consistent with the -ECS pro-
tein level. In flower buds, only stigma and immature anthers show GUS activity. In flowers, GUS activ-
ity is exclusively localized in anthers, stigma, and receptacle. GUS activity is localized in the tip and the
base of young siliques. The spatial pattern of GUS activity in flowering plants is rather intriguing and may
reflect the complexity of GSH synthesis at the whole plant level.
This expression pattern is also similar to that of the gene coding for the bZIP transcriptional factor
TGA6 that specifically binds as-1–type elements in vitro [43]. Because to the nature of the translational
reporter fusion, the GUS activity should report the localization of -ECS. The spatial expression pattern
of -ECS during development indicates that the expression of -ECS is subjected to mutilevel control and
GSH homeostasis in a whole plant body is complex. These data also indicate that GSH transport between
cells and among organs must be functioning in order to maintain an effective GSH level in each cell.
RNA gel blot analysis of steady-state transcript levels for GSH1andGSH2in different tissues pri-
marily agrees with the GUS expression pattern. The transcripts for GSH1as well as GSH2were de-
tected in all tissues examined with relatively higher levels in siliques, roots, and inflorescence stems.
-ECS protein is present in all tissues analyzed, although the protein levels vary from tissue to tissue.
The disjunction between transcription and translation is obvious in these data; notably, the protein level
for -ECS is highest in roots while the GSH1transcript level is lower there than in siliques and inflo-
MULTILEVEL REGULATION OF GLUTATHIONE HOMEOSTASIS 543