INTRODUCTION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Nitrogen is a crucial plant macronutrient that exists in the environment in several inorganic forms. Plants acquire their nitrogen from two principal sources: the soil in the form of nitrate (or nitrite), which is converted to ammonia by the sequential reductive action of nitrate and nitrite reductases; and in legumes from the atmosphere through symbiotic nitrogen fixation (Lea and Ireland, 1999). Nitrate and N₂ are reduced to NH₃, which in turn is assimilated via the joint action of Gln synthetase (GS; EC 6.3.1.2) and Glu synthase (GOGAT; Lam et al., 1996; Ireland and Lea, 1999). GS catalyzes the ATP-dependent condensation of NH₃ with Glu to yield Gln. GOGAT transfers the amido group of Gln to -ketoglutarate to subsequently produce Glu (Temple et al., 1998b; Ireland and Lea, 1999).

Higher plant GS is an octameric enzyme of 320 to 380 kD (Stewart et al., 1980) that occurs as a number of isoenzymes, the subunits of which are encoded by members of a small multigene family (Bennett et al., 1989; Peterman and Goodman, 1991; Roche et al., 1993; Temple et al., 1995; Dubois et al., 1996). These GS isoforms are located in the cytosol (GS₁) or chloroplast/plastid (GS₂) and assimilate ammonia produced by different physiological processes in different plant organs. In the roots, NH₃ is taken up directly by the roots or is produced by the reduction of NO₃ (Ireland and Lea, 1999), in the cotyledons, NH₃ is produced by the breakdown of nitrogenous compounds, whereas the NH₃ in nodules is produced by the fixation of atmospheric N₂ produced by the symbiont (Atkins, 1987). The major GS isoform in the leaves is GS₂ and it is located in the mesophyll cells and its major role is to assimilate NH₃ resulting from the reduction of nitrate and to re-assimilate NH₃ released during photorespiration (Lam et al., 1995). GS₁ in leaves and stem is localized primarily in the phloem elements (Brears et al., 1991; Kamachi et al., 1992; Dubois et al., 1996; Sakurai et al., 1996) and it is postulated that it functions to generate Gln for transport.

The GS₁ genes in all plants studied are members of small gene families and the different members are differentially regulated (Bennett et al., 1989; Peterman and Goodman, 1991; Roche et al., 1993; Temple et al., 1995; Dubois et al., 1996). Little is known about the regulatory mechanism underlying the regulation of the GS₁ genes. However, there is evidence accumulating that suggests the involvement of metabolites in the expression of GS₁ genes in plants (Hayakawa et al., 1990; Kozaki et al., 1991; Miao et al., 1991; Sukanya et al., 1994; Sakakibara et al., 1996; Temple et al., 1996; Oliveira and Coruzzi, 1999). There is evidence that suggests that the ratio of cellular Gln to Glu could be one of the potential regulatory parameters for the expression of the GS₁ genes in radish (Watanabe et al., 1997) and Arabidopsis (Oliveira and Coruzzi, 1999). All these studies emphasize regulation of the GS₁ genes at the transcriptional level. GS₂ genes are regulated by light (Edwards and Coruzzi, 1989; Edwards et al., 1990; Oliveira and Coruzzi, 1999), and photorespiratory ammonia has also been shown to play a direct role in the induction of GS₂ gene in pea (Edwards and Coruzzi, 1989), whereas in bean the effect of photorespiration is indirect (Cock et al., 1990, 1991).

In bacteria, different regulatory mechanisms, which include transcriptional, post-transcriptional, and post-translational modifications, control the GS enzyme to ensure optimal utilization of nitrogen substrates (Reitzer and Magasanik, 1987), and some recent reports suggest that similar regulatory mechanisms may be occurring in higher plants (Temple et al., 1996, 1997, 1998a; Ortega et al., 1999). In this context, it is interesting to point out that a PII (a component of the nitrogen regulatory system in Escherichia coli) homolog has recently been isolated from Arabidopsis and castor bean and has been shown to play a role in signaling the status of carbon and nitrogen in plants (Hsieh et al., 1998). In bacteria, PII acts as an allosteric effector that indirectly regulates GS via other components of the nitrogen regulatory system (Ntr) at the transcriptional and post-translational level in response to nitrogen availability (Merrick and Edwards, 1995; Ninfa et al., 1995; Reitzer, 1996). It has been demonstrated that the intracellular concentrations of -ketoglutarate and Gln are effectors that play a key role in controlling the GS adenylation state and the transcription of GS genes in bacteria (Jiang et al., 1998).

Another level of regulation that has been well demonstrated for bacterial GS is at the level of holoprotein turnover. Although normally stable, GS is turned over when cells are starved for nitrogen (Fulks and Stadtman, 1985), suggesting that the intracellular level of GS in bacterial cells is also regulated by proteolysis. The degradation of GS in E. coli and Klebsiella aerogenes appears to involve two steps: First, the enzyme is inactivated by oxidative modification of a single His residue per subunit (Levine, 1983); second, the altered enzyme is degraded by endogenous proteases that are capable of degrading the oxidized enzyme, but exhibit little activity on the native GS (Rivett and Levine, 1990). A similar phenomenon has been demonstrated for GS from soybean (Glycine max) roots (Ortega et al., 1999), suggesting that the mechanism of GS turnover in plants is the same as in bacteria. Furthermore, it has been shown that the GS enzyme in plants is protected against oxidative turnover by its substrates (Ortega et al., 1999; Suganuma et al., 1999).

The focus of this paper is to determine if GS₁ genes in alfalfa (Medicago sativa) are regulated at steps other than transcription. As a first approach we have introduced into alfalfa a GS₁ gene driven by the constitutive cauliflower mosaic virus (CaMV) 35S promoter to ensure expression of the GS₁ gene in all cell types and to bypass the transcriptional component of regulation. Transgenic tobacco plants with the CaMV 35S promoter driving an alfalfa GS₁ gene had previously been shown to have increased GS activity in the leaves (Eckes et al., 1989; Temple et al., 1993). Similar increase in GS activity was detected in the leaves of transgenic Lotus japonicus and in birdsfoot trefoil transformed with the CaMV 35S promoter driving an alfalfa GS₁ gene and a soybean GS₁ gene, respectively (Temple et al., 1994; Hirel et al., 1997; Vincent et al., 1997). Our analysis of the transgenic alfalfa plants with the CaMV 35S-GS₁ gene construct has included measurement of steady-state levels of the GS transcripts and the GS protein in the different organs of the plant grown under different nitrogen treatments. Our results suggest that GS₁ genes in alfalfa, besides being regulated at the transcriptional level, may also be regulated at the level of transcript stability and holoprotein turnover and that these steps may be under metabolic control.

	RESULTS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Constitutive Overexpression of a Soybean GS₁ (Gmgln₁) Gene in Alfalfa Showed the Accumulation of the Gmgln₁ Transcript in the Leaves, But Not in the Nodules

A gene construct consisting of the CaMV 35S promoter driving a soybean GS₁ gene and containing the 3'-nopaline synthetase (NOS) terminator (pGS20Q; Fig. 1A) was introduced into alfalfa and the presence of the transgene in the genome of the putative transformants was tested by PCR using specific primers for the Gmgln₁ gene and the CaMV 35S promoter (data not shown). This soybean GS₁ gene is similar, but not identical to the previously identified ammonia-inducible form reported by Miao et al. (1991) and probably represents an allelic variant (accession no. AF301590). We will refer to this particular soybean GS₁ gene as Gmgln₁. These transformants showed no visible phenotypic difference when compared with control plants. The transformed and control plants were inoculated with Sinorhizobium meliloti and 30 d after inoculation the leaves and nodules were harvested. The RNA isolated from these tissues was then subjected to RNA-blot analysis using the Gmgln₁ gene-specific probe. The blot was also hybridized with a probe for rRNA to verify equal loading of total RNA into each lane. The lanes containing the RNA from the leaves of five representative transformants showed a hybridizing band with the Gmgln₁-specific probe (Gmgln₁- 3'-untranslated region [UTR]), whereas the lanes with the RNA from the control leaves did not show any hybridization (Fig. 1B), suggesting that the Gmgln₁ gene was transcribed and the corresponding transcript accumulated in the leaves of the alfalfa transformants. However, the nodule RNA from the same transformants showed no trace of hybridization with the Gmgln₁-specific probe even after prolonged exposure of the filters to the x-ray film (Fig. 1B). Similar results were obtained when alfalfa transformants containing the CaMV 35S-MsGS100 (alfalfa GS₁ gene; Temple et al., 1993) were analyzed for the presence of the MsGS 100 transcripts in the leaves and nodules (data not shown).

View larger version (61K): [in this window] [in a new window]   Figure 1.   Analysis of Gmgln1 gene transcript in pGS20Q (CaMV 35S-Gmgln1) transformed alfalfa plants. A, Map of the CaMV 35S-Gmgln1-Nos 3' gene construct. B, Total RNA (20 µg per lane) isolated from the leaves and nodules of non-transformed control (C1, C2, and C3) and transgenic (T12, T22, T26, T35, and T52) alfalfa plants were fractionated on formaldehyde-agarose gels. The gel was blotted onto nitrocellulose and was hybridized with a 32P-labeled 3'-UTR of the Gmgln1 soybean cDNA (Gmgln1, 3'). The same blot was hybridized to a ribosomal RNA probe (28S rRNA) as a control for RNA loadings.

Nitrogen Feeding Reduces the Level of GS₁ Transcript in Leaves of Alfalfa

To test if the nitrogen status of a cell may have a role in affecting transcript accumulation of the GS₁ transgene in the nodules of alfalfa, vegetatively propagated control and transgenic alfalfa plants were grown in the presence or absence of nitrate and the level of the transcript corresponding to the transgene was measured. Because the plants in the two treatments were clonal, any difference in the level of the transcript corresponding to the transgene can be attributed directly to the treatment. Nitrate is taken up by the roots and is reduced to ammonia in the leaves, which is then assimilated into Gln via the action of GS₂ and GOGAT (Lea and Ireland, 1999). Measurement of nitrate levels in the leaves shows that nitrate is transported into the leaves and the levels are significantly higher in the nitrate-fed plants compared with the non-nitrate-fed plants (data not shown). The feeding experiment was repeated three times and similar results were obtained each time. Only the results of a representative experiment are shown in Figure 2. RNA isolated from the leaves and roots of transgenic (two representative plants) and control plants (same age as the transformants) grown with 10 mM KCl or 10 mM KNO₃ for 10 d was subjected to RNA-blot analysis using different GS₁ and GS₂ sequences as probes. The blot was also probed with an actin cDNA probe as a representative of a general housekeeping gene and with an rRNA gene probe to evaluate equal loading of total RNA. The slightly lower signal for rRNA in the leaf samples is attributed to the fact that total leaf RNA samples also contain the chloroplast rRNA. However, within each group the loads appear uniform.

View larger version (53K): [in this window] [in a new window]   Figure 2.   Effect of nitrogen fertilization on the steady-state levels of transcripts corresponding to the transgene and the endogenous GS1 gene in alfalfa plants transformed with the pGS20Q (CaMV 35S-Gmgln1) gene construct. A, Total RNA (20 µg per lane) isolated from the roots and leaves of control (C1) and transgenic plants (T12 and T35) fed for 10 d with 10 mM KCl or 10 mM KNO3 was fractionated on formaldehyde-agarose gels. The gels were blotted onto nitrocellulose and sequentially hybridized after stripping to a 32P-labeled full-length cDNA clone Gmgln1 (Gmgln1, coding), the 3'-UTR of the soybean gln1 cDNA (Gmgln1, 3'), the 3'-UTR of the "constitutive" alfalfa GS1 isoform (GS100, 3'), a full-length alfalfa cDNA clone for GS2, actin cDNA from soybean (Actin), and a soybean 28S ribosomal RNA gene fragment (28S rRNA). B, Band intensity in each case was quantified and the GS and actin band intensities were standardized against the intensity of the 28S rRNA hybridization signal and plotted.

The hybridization signals were subjected to quantitation using the BioImage Intelligent quantifier (Genomic Solutions, Ann Arbor, MI) and the values for hybridization signals with the different GS gene probes and the actin probe were standardized against the hybridization signal with the rRNA gene probe and the values were plotted (Fig. 2B). The Gmgln₁ coding region and the 3'-UTR probes showed a low level of hybridization signal with the root RNA samples from the transformants and a much stronger signal for the leaf RNA from the transformants. Although in the nitrate fed and non-nitrate-fed plants the hybridization signal for the root RNA with the Gmgln₁ probe remained the same, the signal for the leaf RNA from the non-nitrate-fed plants was >3-fold higher than the signal obtained for leaf RNA from nitrate-fed plants. The pMsGS 100 3'-UTR hybridized strongly to the root RNA when compared with leaf RNA and no significant difference was observed in the signal intensity between the roots of nitrate and non-nitrate-fed control plants. However, the roots of the transformans showed a slight drop in the hybridization signal with the MsGS100 3'-UTR probe when treated with nitrate. With the leaf RNA, the hybridization signal with the MsGS100 3'-UTR probe showed a 2.5-fold drop in intensity in the nitrate-fed plants compared with the non-nitrate-fed plants. The actin probe showed no difference in the hybridization signals between the control and transgenic samples, though the roots showed a slightly higher signal than the leaves. Moreover, the samples from nitrate-fed plants showed slightly higher hybridization signals with the actin probe for the roots and the leaves than the KCl-fed plants. These results indicate that in the nitrate-fed plants, there is a specific reduction in the level of GS₁ transcripts in the leaves for the transgene and the endogenous GS₁ genes.

To determine how constitutive overexpression of GS₁ transgene affects the expression of GS₂ gene(s), the RNA blots described above were also probed with an alfalfa GS₂ cDNA (Zozaya-Garza and Sengupta-Gopalan, 1999). The GS₂ probe showed no hybridization with the root RNA from the non-nitrate-fed plants, but showed detectable levels of hybridization with the root RNA samples from the nitrate-fed plants. As expected, the GS₂ gene probe hybridized strongly to leaf RNA, the signal being approximately 2-fold higher for the nitrate-fed plants compared with the non-nitrate-fed plants. It is noteworthy that the leaves from the transformants showed a lower level of GS₂ transcripts compared with the control in the KNO₃- and KCl-fed plants.

Alfalfa Transformants Containing the CaMV 35S--glucuronidase (GUS) Gene Construct Showed Accumulation of the Corresponding Transcripts in the Nodules, and Nitrate Feeding Has No Effect on the CaMV 35S Promoter

To check if the absence of transcripts for the transgenes (Gmgln₁ and MsGS100) in the nodules was due to the non-functionality of the CaMV 35S promoter in the nodules of alfalfa, a gene construct consisting of the CaMV 35S-GUS-NOS 3' (Fig. 3A) was introduced into alfalfa and the leaves, stem, roots, and nodules were stained for GUS activity. GUS activity was seen in all organs, including the nodules (Fig. 3B). There appeared to be a wide range in the intensity of GUS staining in the nodules from different alfalfa transformants, but all nodules tested showed some level of GUS staining. In the roots, GUS activity was mostly associated with the vasculature (Fig. 3B, panels 2 and3), whereas in the leaves (Fig. 3B, panel 1) and nodules (Fig. 3, panels 3 and 4), GUS staining appeared more homogeneous. The results indicate that the CaMV 35S promoter is active in the alfalfa nodules and thus the absence of the Gmgln₁ transcripts in the nodules is possibly due to RNA turnover.

View larger version (36K): [in this window] [in a new window]   Figure 3.   Analysis of expression of CaMV 35S promoter-GUS gene in transgenic alfalfa. A, Map of the gene construct, consisting of the udiA gene that encodes for GUS in between the CaMV 35S promoter and the NOS 3' terminator. The positions of the relevant restriction sites are indicated. B, Histochemical analysis of GUS activity in leaf (1), roots (2, 3) and nodules (3, 4), of the alfalfa plants transformed with the CaMV 35S-GUS-NOS 3' gene construct. Blue precipitate indicates the location of GUS activity. C, Total RNA (20 µg per lane) isolated from the leaves of control (C1) and plants transformed with a GUS gene behind the CaMV 35S promoter (T1) fed for 10 d with 10 mM KCl or 10 mM KNO3 was fractionated on formaldehyde-agarose gels. The gel was blotted onto nitrocellulose and hybridized to a 32P-labeled BglII-EcoRI DNA fragment (A). The RNA in the gel was visualized after ethidium bromide staining to check for RNA loadings. The hybridization intensity and the ethidium bromide staining intensity were quantified. Intensity of hybridization with the GUS probe was standardized against the intensity of the 28S rRNA band and plotted.

The drop in the level of the transcript for the Gmgln₁ in alfalfa transformants fed with nitrate can be attributed to repression of transcription of the CaMV 35S promoter by nitrate or the destabilization and turnover of the GS₁ transcripts. To address this issue, non-nodulated alfalfa plants containing the CaMV 35S-GUS construct were grown in the presence of KNO₃ or KCl for 10 d, and the leaf RNA was analyzed for GUS transcript. The hybridization signal was quantified and the band intensities were standardized against the ethidium bromide staining intensity of the 28S rRNA bands and plotted. As seen in Figure 3C, nitrate feeding had no significant effect on the GUS transcript level. The results suggest that the decreased level of GS₁ transcript in the KNO₃-fed transgenic plants is probably due to increased GS₁ transcript turnover.

No Increase in GS₁ Polypeptide Level Is Detected in the Leaves of Alfalfa Transformants Expressing the Gmgln₁ Gene

Protein extracts from roots and leaves of control and transformed alfalfa plants (same as those used for RNA analysis, Fig. 2) were subjected to GS activity measurements (Fig. 4A) and immunoblot analysis using anti-GS antibodies (Fig. 4B). The GS immunoreactive bands were quantified and the values were plotted. The analysis was done three times and only data from a representative experiment is shown here. In spite of a significant accumulation of GS₁ transgene transcripts in the leaves of non-nitrate-fed transgenic plants (Fig. 2A), no increase in GS activity or GS₁ polypeptide level was detected in the leaves of the transformants compared with the control (Fig. 4, A and B). In general, the roots and the leaves showed a trend of a slightly reduced GS activity and GS₁ polypeptide level in the transformants, regardless of the nitrogen feeding regime. The leaves of the nitrate-fed plants, control and the transformants, showed a 2-fold drop in the level of GS₁ polypeptide and a slight increase in the level of GS₂ polypeptide. However, no significant change in GS activity was observed in the leaves or the roots of nitrate-fed plants.

View larger version (47K): [in this window] [in a new window]   Figure 4.   Effect of nitrogen on the GS activity and GS polypeptides in alfalfa plants transformed with pGS20Q (CaMV 35S-Gmgln1) gene construct. A, GS enzyme activity was measured in protein extracts from the roots and leaves of control (C1) and transgenic (T12 and T35) alfalfa plants fed with KNO3 or with KCl using the transferase activity assay (Ferguson and Sims, 1971). The activity was plotted as transferase units (micromoles -glutamyl hydroxamate per minute) per milligram of protein. Values are the mean value from at least three experiments ± SE. B, Total soluble protein from the same extracts (1.25 µg per lane for roots and 2.5 µg per lane for leaves) was fractionated by SDS-PAGE, transferred to nitrocellulose, and the membrane was probed with GS antibodies. The GS1 and GS2 immunoreactive bands were quantified and the band intensities were plotted. C, Total soluble protein, after desalting (10 µg) from the leaves of a control and a transgenic plant, was subjected to two-dimensional gel electrophoresis followed by western-blot analysis with anti-GS antibodies. The spots corresponding to the GS1 and GS2 forms are indicated. Msb represents the major alfalfa GS1 polypeptide. The Gmgln1 gene product comigrates with the spot Msb.

To determine if there is any qualitative difference in the GS polypeptide profile due to the presence of the functional transgene in the transformants, leaf proteins from non-nitrate-fed control and transformed plants were subjected to two-dimensional immunoblot analysis using the GS antibodies. The more abundant GS₂ polypeptides separated as three main spots, whereas the minor fraction of GS₁ polypeptides fractionated as one major and two minor spots in control and transgenic plants (Fig. 4C). Based on two-dimensional gel analysis of alfalfa leaf proteins (by immunostaining) and translation products (by autoradiography) corresponding to Gmgln₁ hybrid selected mRNA from soybean roots, we have determined that the major GS₁ spot of alfalfa leaves comigrates with the Gmgln₁ gene product (data not shown). No quantitative or qualitative difference was observed in the GS polypeptide profile between the control and the transformed plant.

The Gmgln₁ Transcripts in the Transformants Are Translatable

Since no increase in GS₁ polypeptide was detected in the transformants, the question arises as to whether the transcript for the transgene can be translated. To check the translatability of the transcript for the transgene, the Gmgln₁ coding region and 3'-UTR were used to hybrid-select RNA from the leaves of the transformant and the translation products were subjected to one-dimensional SDS-PAGE, along with the hybrid select translation (HST) products of RNA selected from the leaves of the control plants. As a positive control, the translation products corresponding to soybean root RNA hybrid-selected with the GS₁ coding region and the Gmgln₁ 3'-UTR were also analyzed. The HST products of RNA selected by the GS₁ coding region from control and transformed alfalfa plants comigrated as a band of approximately 39 to 40 kD on SDS-PAGE in the same position as the lower band of the doublet (GS₁ and GS₂) obtained as the HST product for soybean root mRNA (Fig. 5). The translation products of RNA selected with Gmgln₁ 3'-UTR from the leaves of the transformant and the soybean roots comigrated with the lower band of the doublet (GS₁) described above (Fig. 5). The Gmgln₁ 3'-UTR did not select any mRNA from the control alfalfa leaves, as indicated by the absence of a translation product. Furthermore, tobacco plants transformed with the CaMV 35S-Gmgln₁ showed accumulation of the Gmgln₁ polypeptide and an increase in GS activity in the leaves when compared with leaves of non-transformed tobacco (Data not shown). The results suggest that the transcript for the transgene is translatable.

View larger version (41K): [in this window] [in a new window]   Figure 5.   Electrophoretic analysis of GS hybrid-select translation products from control (non-transformed) and transgenic alfalfa leaves, and from 4-d-old roots of soybean. GS mRNA was hybrid-selected from control and transgenic alfalfa leaves and from 4-d-old roots of soybean with immobilized plasmid DNA from plasmid pGmgln1 (coding) or plasmid pGmGS16D (Roche et al., 1993), which corresponds to the 3'-UTR of the Gmgln1 gene (3'-UTR). Hybrid-selected mRNAs were translated in vitro in the rabbit reticulocyte system and the translation products were analyzed by SDS-PAGE followed by autoradiography. The positions of the soybean GS isoforms GS1 and GS2 and the corresponding Mr standards are indicated. The lane labeled as () is translation mix without exogenous RNA.

The Transcripts Corresponding to the GS₁ Transgene Are Recruited into the Polysomes

Our data suggests that the Gmgln₁ gene is transcribed in the leaves of transgenic alfalfa and the RNA can be translated in vitro. However, there is no increase in the GS₁ polypeptide level or GS activity in the leaves of the alfalfa transformants. These results could be interpreted to mean that the Gmgln₁ transcripts are not recruited for protein synthesis in alfalfa or that the transcripts are translated, but the protein is not stable. To address this question, polysomal RNA and total RNA was isolated from the leaves of control and transgenic alfalfa grown under non-nitrate-fed conditions and subjected to RNA-blot analysis, using the Gmgln₁ 3'-UTR and MsGS100 3'-UTR as probes. The hybridization signal was measured and the band intensity values were standardized for the RNA load as determined by ethidium bromide staining of the 28S rRNA. The plotted values showed that the pattern of hybridization for the polysomal and total RNA was similar with the two probes (Fig. 6), suggesting that the recruitment of transcripts for the transgene into polysomes follows the same pattern as the transcripts for the endogenous MsGS100 gene. These results rule out the possibility of Gmgln₁ transcript not being recruited into polysomes for translation as a mechanism to account for the absence of any increase in GS₁ polypeptides in the transgenic plants.

View larger version (63K): [in this window] [in a new window]   Figure 6.   Analysis of GS1 transcripts in polysomal RNA from the leaves of control and pGS20Q transformed alfalfa plants. A, Total and polysomal RNA (20 µg per lane) isolated from the leaves of non-transformed control (C1 and C2) and transgenic (T12 and T35) plants were fractionated on formaldehyde-agarose gels followed by RNA-blot hybridization using a 32P-labeled full-length cDNA clone for Gmgln1 gene, the 3'-UTR of the soybean transgene (Gmgln1), and the 3'-UTR of the constitutive alfalfa gene (MsGS100). B, Band intensities were quantified, standardized against the 28S ribosomal RNA band (28S rRNA), as seen by ethidium bromide staining, and plotted.

	DISCUSSION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Earlier reports in the literature have shown an increase in GS activity and GS₁ polypeptide level in the leaves of transgenic tobacco transformed with gene constructs consisting of the CaMV 35S promoter driving different GS₁ genes (Eckes et al., 1989; Hemon et al., 1990; Hirel et al., 1992; Temple et al., 1993). It was shown that the GS₁ polypeptide corresponding to the transgene was fairly labile and the amount of GS₁ polypeptide and activity decreased dramatically without any change in the level of the transgene mRNA in mature phototrophically soil-grown plants (Hemon et al., 1990; Temple et al., 1993; Temple and Sengupta-Gopalan, 1997). A 50% to 80% increase in GS activity was detected in the leaves of birdsfoot trefoil transformed with a soybean GS₁ gene driven by the CaMV 35S promoter (Hirel et al., 1997). In a similar manner, L. japonicus plants transformed with an alfalfa GS₁ gene driven by the CaMV 35S promoter showed a 2- to 3-fold increase in GS activity and in the level of GS₁ polypeptide (Temple et al., 1994). However, our attempts to overexpress GS₁ genes in alfalfa in a constitutive manner have, to date, resulted in no increase in the GS₁ polypeptide level or GS activity. This is interesting in view of the fact that the first report of plant cells overexpressing GS₁ was in an alfalfa cell line that was selected for resistance to L-phosphinothricin, a competitive inhibitor of GS (Donn et al., 1984). The cell line was found to contain a 3- to 7-fold elevation in the GS activity resulting from a 4- to 11-fold amplification of a specific GS₁ gene.

Our data suggests that besides transcriptional regulation, there may be two additional levels of regulation that play a major role in controlling the accumulation of the GS₁ peptide in alfalfa: one involves the turnover of the transcript and the other point of control is at the level of protein turnover. The complete absence of the GS₁ transcripts corresponding to the transgene in the nodules (Fig. 1), but not in the leaves or the roots would suggest that the first level of control in the expression of the GS₁ transgene in the transformed alfalfa lines is at the level of transcript stability. We have shown in this paper that the CaMV 35S promoter is capable of expressing the GUS gene in the nodules (Fig. 3) and yet in plants with the CaMV 35S promoter driving different GS₁ genes, MsGS100 and Gmgln₁, it failed to promote accumulation of the transcript corresponding to the transgene in this organ. Furthermore, in an earlier paper (Bagga et al., 1992) it was shown that in alfalfa plants transformed with a CaMV 35S--phaseolin gene-NOS 3' terminator construct, equivalent amounts of the -phaseolin transcript was found to accumulate in the leaves and the nodules. Because the -phaseolin gene construct and the GUS construct had the same NOS 3' terminator as the Gmgln₁ gene construct, we can rule out the possibility of differences in accumulation pattern of the transcripts on the terminator sequences.

A 3- to 4-fold drop in the level of the Gmgln₁ transcript in transgenic alfalfa plants that were fed nitrate over their non-nitrate-fed counterpart (Fig. 2) would suggest that the transcripts are unstable in the presence of nitrate or downstream metabolites formed during nitrate assimilation. This difference in the Gmgln₁ transcript level is not attributed to differential promoter activity because no significant difference in the level of GUS transcripts was observed between the nitrate-fed and non-nitrate-fed transformants containing the CaMV 35S-GUS construct (Fig. 3C). What is even more intriguing is that the transcript for the alfalfa endogenous GS₁ gene, represented by the MsGS100 subclass, also showed a 3-fold drop in levels in the leaves of plants treated with nitrate compared with non-nitrate-fed plants. The actin mRNA under similar conditions showed no change in the level and the GS₂ transcript level showed a 2-fold increase in the NO₃-fed plants compared with control. Although we cannot rule out the possibility that a drop in the level of endogenous GS₁ transcript in nitrate-fed plants is due to differences in the activity of the MsGS100 gene promoter in response to nitrate treatment, it is most likely due to differential turnover of the GS₁ transcript, as is the case with the Gmgln₁ transcript in transgenic alfalfa where the transgene is driven by the CaMV 35S promoter.

Although there is no precedence for mRNA turnover as a regulatory step for GS in bacteria, regulation through differential stability of the same mRNA under different growth conditions is well documented in eukaryotic cells (Atwater et al., 1990; Brodl and Ho, 1991; Zhang et al., 1993). Analysis of Arabidopsis mutants that overaccumulate soluble Met revealed that the gene for cystathionine -synthase, the key enzyme in Met biosynthesis, is regulated at the level of mRNA stability (Chiba et al., 1999). Furthermore, it was also shown that an amino acid sequence encoded by the first exon of the cystathionine -synthase gene acts in cis to destabilize its own mRNA in a process that is activated by Met or one of its metabolites (Chiba et al., 1999). Iron-dependent destabilization of the transferrin receptor mRNA in mammalian cells is attributed to iron-responsive elements in the 3'-UTR of the TfR mRNA (Müllner and Kühn, 1988; Casey et al., 1989). The -amylase in germinating rice embryos, where it plays an important role in the degradation of starch, has also been shown to be regulated at the level of transcript turnover (Sheu et al., 1996). More recently, Chan and Yu (1998), with the use of chimeric gene constructs, have identified regulatory sequences in the -amylase 3'-UTR that may act as potent determinants of mRNA stability in response to sugar availability. Experiments are in progress to test the role of the 3'-UTR of GS₁ genes in the turnover of the corresponding transcripts.

An additional level of regulation in the expression of the GS₁ transgene appears to be at the level of enzyme turnover. In spite of a significant increase in the level of GS₁ transcripts in the leaves of the transformants, no significant change in the level of GS₁ polypeptides could be detected in the transformants over control samples (Fig. 4B). The absence of the Gmgln₁ polypeptide cannot be attributed to translational control since the GS₁ transcripts corresponding to the transgene can be translated in vitro and is recruited for translation into polysomes in the leaves similar to the transcripts of the endogenous GS₁ gene (Fig. 6). Since no increase in GS₁ polypeptides could be observed in these transformants, it would follow that the assembled holoprotein is degraded. No qualitative or quantitative differences could be observed in the two-dimensional SDS-PAGE profile of GS protein in the leaves of control and transgenic plant (Fig. 4C), suggesting that only the GS₁ transgene subunits are targeted for turnover. However, since the transgene product comigrates with the major alfalfa GS₁ subunit (data not shown), we cannot rule out the possibility of the accumulation of the transgene product and a corresponding drop in the level of the endogenous GS₁ protein in the transformants.

The major GS isoform found in the mesophyll cells of leaves is chloroplast-localized GS₂ that assimilates NH₃ produced by nitrate reduction or by photorespiration, whereas GS₁ is found only in the vascular tissues where it functions in transport. In the alfalfa plants transformed with a CaMV 35S-GS₁ gene, however, the GS₁ gene is transcribed in the mesophyll cells, but GS₁ does not play a significant role in the cytosol of the mesophyll cells and consequently, is unstable and does not accumulate. We have shown that GS in plants, as in bacteria, is subject to a two-step turnover process, the first step involving oxidation of a specific amino acid residue in the active site followed by the proteolytic degradation of the oxidized GS (Ortega et al., 1999). If, however, the active site of the GS enzyme is occupied by the substrates, it is protected from oxidative modification and hence from proteolytic turnover (Ortega et al., 1999). Thus, we can postulate that in the absence of the GS substrates, the active sites of the GS₁ enzyme that is made in the cytosol of the mesophyll cells of the transformant is more prone to oxidative modification and thus rapid turnover. This postulate is supported by the finding that alfalfa transformants containing a GS₂ gene driven by the CaMV 35S promoter accumulates high levels of GS₂ polypeptide in the leaves (M. Zozaya-Garza and C. Sengupta-Gopalan, unpublished data).

Post-translational control of GS is further supported by earlier work from our laboratory where we showed that a reduction of about 80% in the GS₁ mRNA levels in transgenic alfalfa by antisense RNA technology was not accompanied by any change in the levels of total GS₁ polypeptides (Temple et al., 1998b). We postulated that the GS₁ mRNA was probably not limiting and, as such, big reductions in the level of the mRNA were not accompanied by any changes in the corresponding protein level. Furthermore, we had also shown that ineffective nodules of soybean had significantly reduced levels of GS activity and GS₁ polypeptides and this was ascribed to turnover of the holoprotein in the absence of the substrate resulting from the absence of N₂-fixation (Temple et al., 1996). Similarly ineffective pea nodules showed lower levels of GS polypeptide and GS activity than the effective nodules without any significant change in the level of the GS mRNA. The enzyme activity and the GS polypeptide level, however, could be enhanced by the exogenous application of ammonia, suggesting that the GS activity/stability in the nodules is regulated by its substrate, ammonia (Suganuma et al., 1999).

We have presented data in this paper that would suggest that GS is regulated at multiple steps and we present a preliminary model: the first step in regulation of GS is at the transcriptional level and mechanistically little is known about it at this time. The second step of regulation is at the level of transcript stability and this step may be controlled by the Gln/Glu ratio, ATP/ADP ratio, or the redox balance. In the presence of excess nitrogen substrate, the carbon skeletons and ATP may become limited, which in turn may have a negative feedback control on the GS transcript. There is evidence in the literature suggesting crosstalk between signals derived from carbon and nitrogen metabolism in E. coli (Merrick and Edwards, 1995); however, the means by which these signals are sensed is not known. There is also evidence in the literature that would suggest that such crosstalk exists in plants (Faure et al., 1994; Lam et al., 1995; Watanabe et al., 1997; Hsieh et al., 1998). The third level of regulation would be at the level of enzyme stability, and it would involve the inactivation of GS by oxygen radicals generated by redox reactions, particularly during conditions of an excess of carbon skeletons or nitrogen substrate limitation. One unresolved issue that remains is why alfalfa plants exhibit such stringent control in the expression of GS₁ transgene, whereas Lotus species and tobacco do not. Plants differ in their site of NO₃ assimilation and also differ in the proportions of GS₁ and GS₂, depending on the plant or the organ (Lam et al., 1996; Woodall and Forde, 1996). This could be attributed to basic physiological differences between plant species, like the availability of carbon skeletons and the source of ammonia in the different plant compartments and to the particular role that each GS isoenzyme has in the different plant parts.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Recombinant DNA Techniques

Standard procedures were used for all recombinant DNA manipulations (Sambrook et al., 1989). Plasmid pMsGS100 contains a constitutively expressed class of alfalfa (Medicago sativa) GS₁ cDNA that was isolated from a alfalfa cell culture line (Das Sarma et al., 1986) and was a gift from Dr. H.M. Goodman (Department of Molecular Biology and Genetics, Harvard Medical School, Boston). A soybean (Glycine max) full-length GS₁ cDNA clone (pGmgln₁) was isolated by screening a 14-d-old soybean seedling cDNA library in gt11 using the pMsGS100 coding region as the probe. The cDNA fragment was released as a 1.5-kb Bsi WI fragment containing part of the phage left and right arms and ligated into pGEM3Zf(-) SmaI-linearized vector to create the plasmid pGS51B. A 1.4-kb partial EcoRI fragment containing the full-length GS₁ cDNA was released from pGS51B and subcloned into the EcoRI site of pSP73 vector to produce pGmgln₁. This plasmid was sequenced using the Di-deoxy sequencing kit (United States Biochemical, Cleveland). Sequence analysis showed that the cDNA was homologous to the pGS20 gene representing the ammonia-inducible ₁ isoform of soybean GS (Marsolier et al., 1995). The 1,400-bp cDNA containing the 282-bp 3'-UTR was isolated as a ClaI-KpnI fragment and was ligated into the ClaI/KpnI sites of pMON 316 in between the 35S promoter and the NOS 3' terminator to produce the gene construct pGS20Q.

Plant Transformation

pGS20Q was mobilized from Escherichia coli strain DH5a into Agrobacterium tumefaciens receptor strain A206 containing the Ti plasmid pTiT37ASE by triparental mating essentially as described by Rogers et al. (1987). The Agrobacterium strain with the 35S-GUS-NOS3 gene construct was kindly provided to us by Dr. John Kemp (New Mexico State University, Las Cruces). Transformation of alfalfa Regen-SY was carried out according to the procedure of Austin et al. (1995). Trifoliates from greenhouse-grown plants were surface sterilized and leaf segments approximately 0.5 cm² were immersed for 5 min in an overnight culture of A. tumefaciens. Following blotting, the leaf discs were cocultivated for 2 d on sterile filter paper placed on top of TM-1 media (Bagga et al., 1992). Following cocultivation, the leaf segments were washed extensively with a solution of 1,000 mg L¹ cefotaxime in sterile distilled water and transferred to TM-1 medium supplemented with 50 mM 2,4-dichlorophenoxyacetic acid, 2.2 mM benzylaminopurine, 0.5 mM naphthylacetic acid, 25 mg L¹ kanamycin, and 500 mg L¹ cefotaxime. The kanamycin-resistant embryogenic calli that developed were transferred to hormone-free TM-1 medium for further embryo development and plant regeneration. Plantlets with roots were allowed to develop further on hormone-free TM-1 media in magenta boxes before being transferred into pots containing a mixture of soil:perlite:vermiculite (3:1:1). Transformed plants were maintained in the greenhouse.

Plant Treatments

All experiments described in this paper were performed multiple times and only the results from representative experiments are shown here. All the transformants and the control plants have identical background except for the presence and position of the transgenes in the genome since all transformation is performed on the same clonal material. To nodulate plants, the primary transformants and control plants were rooted and then inoculated with a culture of Sinorhizobium meliloti 1021 and were watered with nitrogen-free nutrient solution. The nodules and leaves were harvested 30 d after inoculation. For the NO₃ feeding experiments, plants were vegetatively propagated from the original transformants that had been grown in soil and fertilized with 10 mM KNO₃. Cuttings were planted in vermiculite and watered with distilled water for 4 weeks during which period they developed a good root system. At this stage the plants were fertilized with nutrient solution containing 10 mM nitrate and after 1 week they were transferred to new pots with vermiculite. The root systems of these newly transferred plants were then thoroughly washed with distilled water to remove all traces of nitrate. The plants were then divided up into two groups, and one group was fed with nitrogen-free nutrient solution containing 10 mM KNO₃, and the second group was fed with ntirogen-free nutrient solution containing 10 mM KCl for 10 d. The leaves and roots from these plants were harvested and used for RNA and protein analysis. The NO₃-fed and the non-NO₃-fed plants are clonal and, as such, any changes in gene expression can be attributed to the particular treatment.

Histochemical Localization of GUS Activity

Histochemical localization of GUS activity was performed using 5-bromo-3 indolyl -D-GlcUA as a chromogenic substrate. A reaction mixture consisting of 1 mM 5-bromo-3 indolyl -D-GlcUA dissolved in 50 mM sodium phosphate buffer (pH 7) was used. The freshly cut plant parts were incubated for 12 h in this mixture at 37°C, rinsed with water, and then fixed in 10% (v/v) formaldehyde, 42% (v/v) ethanol, and 5% (v/v) acetic acid, followed by rinsing in 42% (v/v) ethanol and 5% (v/v) acetic acid before photographing.

Nucleic Acid Isolation and Analysis

Total RNA was isolated using the LiCl precipitation procedure (De Vries et al., 1982). Polysomal RNA was isolated from alfalfa leaves by using the Suc cushion polysomal RNA extraction procedure (Sengupta-Gopalan et al., 1986). RNA was fractionated in 1.3% (w/v) agarose/formaldehyde gels and blotted onto nitrocellulose. DNA probes were prepared from plasmid inserts isolated from agarose using the Wizard mini-prep columns (Promega, Madison, WI) and labeled by the random primer method as described by Temple et al. (1998a). All filters were prehybridized for a minimum of 4 h and hybridized for 20 to 24 h in 50% (w/v) formamide, 5× SSC (1× SSC is 0.15 M NaCl, 0.015 M sodium citrate), 5× Denhardt's, 50 mM sodium phosphate (pH 7.0), 0.1% (w/v) SDS, 0.1 mg mL¹ denatured calf thymus DNA, and 0.04 mg mL¹ poly(A) at 42°C. Following hybridization the filters were washed three times with 2× SSC, 0.1% (w/v) SDS at 42°C for 15 min each followed by two washes with 0.5× SSC, 0.1% (w/v) SDS at 42°C for 20 min each, and exposed to x-ray film. The hybridization signals were subjected to band quantitation analysis using the BioImage Intelligent quantifier software (Genomic Solutions, Ann Arbor, MI).

HST

This was carried out essentially as described by Roche et al. (1993). Plasmid DNA containing the full-length cDNA or the 3'-UTR (plasmid pGS16, Roche et al., 1993) of the Gmgln₁ gene were immobilized on nitrocellulose discs and hybridized with 250 µg of target total RNA. The hybrid-selected RNA was translated in vitro using the rabbit reticulocyte system (Promega) with ³⁵S-Met as the tracer amino acid. Prior to sample preparation for two-dimensional SDS-PAGE, aliquots of the in vitro translations were mixed with an appropriate aliquot of soluble alfalfa proteins and the samples were denatured by boiling in 2% (w/v) SDS before the addition of 9.5 M urea. Following two-dimensional SDS-PAGE the polypeptides were transferred to nitrocellulose as described above and were subjected to western analysis using GS antibody. The filter was then air-dried and exposed to x-ray film, allowing the detection of the radiolabeled HST products.

Protein Extraction and GS Enzyme Activity Assay

All procedures were carried out at 4°C. The different tissues were ground in liquid nitrogen with 15% (w/w) insoluble polyvinylpolypyrrolidone and homogenized with two (roots) or five (leaves) volumes of extraction buffer (50 mM Tris-Cl, pH 8.0, 5 mM EDTA, 5% [v/v] ethyleneglycol, 20% [v/v] glycerol, 1 mM Mg acetate, 1 mM dithiothreitol [DTT], and a mixture of protease inhibitors [50 µg mL¹ antipain, 1 µg mL¹ cystatin, 10 µg mL¹ chymostatin, 2 µg mL¹ leupeptin, and 1 mM phenylmethylsulfonyl fluoride]). The homogenate was centrifuged for 15 min at 20,000g. For two-dimensional SDS-PAGE analysis, the tissue extracts were desalted in Sephadex G25 columns against the same buffer as described above.

Protein concentration was measured by the Bradford (1976) protein assay using bovine serum albumin as protein standard. GS activity was measured spectrophotometrically at 500 nm by the transferase assay reported by Ferguson and Sims (1971). Transferase units were calculated from a standard curve of -glutamyl hydroxamate. One unit of transferase activity is equivalent to 1 µmol of -glutamyl hydroxamate produced per min at 30°C. GS activity data presented are the average of at least three independent experiments.

PAGE

Two different PAGE systems were employed. One was SDS-PAGE using 12% (w/v) slab mini-gels, and the other was two-dimensional SDS-PAGE carried out essentially as described (Temple et al., 1996) using 1.6% (w/v) of pH 5 to 7 and 0.4% (w/v) of pH 3.5 to 10 Ampholites (Sigma), 2% (w/v) CHAPS {3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonic acid} replaced Nonidet P-40, and 10 mM DTT replaced -mercaptoethanol in all isoelectric focusing (IEF) solutions. To assist with sample solubilization and eliminate proteolytic sample degradation prior to IEF, the protein samples were denatured in 2% (w/v) SDS, placed in a boiling water bath for 5 min, and cooled to room temperature before the addition of 1 mg urea/µL sample. The IEF tube gels were run overnight (16 h) at 400 V, followed by 1 h at 800 V. The IEF gels were equilibrated for 30 min in 62.5 mM Tris-Cl (pH 6.8), 2.0% (w/v) SDS, 10 mM DTT, and 10% (w/v) glycerol before being mounted on 12% (w/v) SDS-PAGE slab gels. For immunoblot analysis, proteins were electroblotted onto nitrocellulose in 25 mM Tris, 192 mM Gly, and 20% (w/v) methanol (pH 8.8). The nitrocellulose was blocked with 2% (w/v) bovine serum albumin in Tris-buffered saline containing 0.1% (w/v) Tween 20 and was probed with antibody against Phaseolus vulgaris nodule GS₁ (1:2,000 dilution). Cross-reacting polypeptides were made visible using an alkaline phosphatase-linked second antibody employing the substrates nitroblue tetrazolium and 5-bromo-4-chloro-3-indoyl-phosphate, used according to the suppliers instructions (Promega).