Metabolic Engineering of Plants for Osmotic Stress Resistance Michael L Nuccio*, David Rhodes†, Scott D Mcneil* and Andrew D Hanson*‡

128

Metabolic engineering of plants for osmotic stress resistance Michael L Nuccio*, David Rhodes†, Scott D McNeil* and Andrew D Hanson*‡

Genes encoding critical steps in the synthesis of raise osmotic pressure and thereby maintain both turgor osmoprotectant compounds are now being expressed in and the driving gradient for water uptake [3]. Many transgenic plants. These plants generally accumulate low levels microorganisms also produce osmoprotectants [4]. of osmoprotectants and have increased stress tolerance. The Figure 1 shows the structures of some common osmo- next priority is therefore to engineer greater osmoprotectant protectant compounds; they fall into three synthesis without detriment to the rest of metabolism. This will groups — amino acids (e.g. proline), onium compounds require manipulation of multiple genes, guided by thorough (e.g. glycine betaine, dimethylsulfoniopropionate), and analysis of metabolite fluxes and pool sizes. polyols/sugars (e.g. mannitol, D-ononitol, trehalose). Being non-toxic, osmoprotectants can accumulate to Addresses osmotically significant levels without disrupting metabo- *Department of Horticultural Sciences, University of Florida, lism; some of them can also protect enzymes and Gainesville, FL 32611-0690, USA membranes against damage from high salt concentra- † Department of Horticulture, Purdue University, West Lafayette, tions [4] and others (especially polyols) protect against IN 47907-1165, USA ‡e-mail: [email protected] reactive oxygen species [5].

Current Opinion in Plant Biology 1999, 2:128–134 The accumulation of osmoprotectants has been a target for http://biomednet.com/elecref/1369526600200128 plant genetic engineering for more than 15 years [6], and work is now in progress on all the compounds in Figure 1. © Elsevier Science Ltd ISSN 1369-5266 In several cases, introduction of a single foreign gene into Abbreviations a transgenic plant has led to modest accumulation of an GlyBet glycine betaine osmoprotectant and, apparently in consequence, a small MCA metabolic control analysis increase in stress tolerance. The physiological and agricultural implications of these experiments have been Introduction thoroughly reviewed [2•,5,7–9,10•]. We will focus here on Improving crop resistance to osmotic stresses is a long- the metabolic implications of the genetic engineering standing goal of agricultural biotechnology [1,2•]. work, and on what needs to be done to drive more flux Drought, salinity and freeze-induced dehydration consti- towards osmoprotectant synthesis, with emphasis on tute direct osmotic stresses; chilling and hypoxia can glycine betaine, polyols and trehalose. indirectly cause osmotic stress via effects on water uptake and loss. Soil salinity alone affects some 340 mil- Overview of progress in engineering lion hectares of cultivated land [2•]. To withstand osmoprotectant synthesis osmotic stresses, certain plants have evolved a high Table 1 catalogs the osmoprotectant work published to capacity to synthesize and accumulate non-toxic solutes date, carried out with tobacco, Arabidopsis, rice and potato. (osmoprotectants or compatible solutes), predominantly The table illustrates several important points. First, many in the cytoplasm, as part of an overall mechanism to of the transgenes are of a non-plant origin, which reflects

Figure 1

CH2OH OH + O CH2OH CH OH N COO- 2

OH OH H2COH

HO O O OH OH

OH OH OH Glycine betaine Trehalose OH OH OH OH OH - OH COO OH OH + S - CH3O + COO CH2OH N OH CH2OH

3-dimethylsulfoniopropionate Proline D-ononitol Mannitol Sorbitol

Current Opinion in Plant Biology

Structures of various osmoprotectants found in plants. Metabolic engineering of plants for osmotic stress resistance Nuccio et al. 129

Table 1

Experiments designed to metabolically engineer osmoprotectant biosynthesis

Osmo- Gene constructs Observed phenotype* Product level† Enzyme activity‡ Metabolic measurements§ protectant Source Promoter Stress Side- (% of physiol.) Assayed Localized Precursor Intermediate Pathway resistance effects pool pools flux Proline Mothbean P5CS 35S √ 180 √ [35] Mannitol E. coli mtldh 35S √ 16 [36,37] NOS Mannitol E. coli mtldh 35S √ 16 [38] Mannitol E. coli mtldh 35S √ 8–16 [39] Sorbitol Apple s6pdh 35S 0.7–1.3 √ [40] Sorbitol Apple s6pdh 35S √ 0.7–300 [16•]

D-Ononitol Ice plant imtl 35S √ 10–70 √√ [41,21••] Trehalose Yeast tpsl atslA √√0.8–3.4 √√ [42] Trehalose E. coli tpsl 35S √ 0.8–28 √√ [17••] tpp Patatin Trehalose Yeast tpsl 35S √√ 1.2 [43] GlyBet E. coli cdh 35S √√ [18] GlyBet A. globiformis 35S √ 5 √√ [19•,20] cod A GlyBet Spinach cmo 35S 1 √√√√ √ [12••]

*A tick under ‘stress resistance’ indicates that the transgenic plant proline 36mM [35]; polyols 50mM [44]; trehalose 50mM [17••]; displayed more resistance than controls. A tick under side-effects GlyBet 24mM [45]. ‡Indicates that the enzyme activity of the indicates that transgenic plants had a growth defect. †This column transgene product was assayed and localized to a subcellular reports the level of osmoprotectant synthesis in osmotically- compartment. §Indicates measurements were made of the stressed transgenic plants as a percentage of the level found in a endogenous levels of the osmoprotectant’s precursor or of any representative plant that naturally accumulates the osmoprotectant pertinent intermediate, or whether flux through the pathway was in similar conditions. These levels (given in mM in tissue water) are: measured with isotopic tracer methods. the lack of investment in plant biochemistry. The plant and further progress requires that the metabolic constraints gene pool should not be overlooked, as plants have be identified. evolved unique genes to synthesize osmoprotectants [11], and plant genes can possess particularly useful characteris- What are these constraints likely to be? First, certain meta- tics. For example, in transgenic tobacco, choline bolic networks can be rigid in that they have evolved to monooxygenase (CMO) is stabilized by salt-stress via a maintain metabolite flux distributions that are optimal for post-transcriptional mechanism, which leads to higher growth, and oppose any flux redistribution following CMO activity when it is most needed [12••]. Second, most expression of a transgene [15]. Second, the engineered of the transgenes have been expressed from a constitutive pathway may divert flux away from primary metabolism promoter. Tissue-specific and inducible promoters eventu- and so create undesirable side effects [16•]. Finally, a for- ally will be necessary, and could be developed from genes eign metabolite may be degraded, limiting its known to be induced when and where high osmoprotec- accumulation [17••]. We will now use published results to tant synthesis is required [13,14]. Third, most of the plants illustrate these points. described express a single transgene and so represent only the first step in the engineering process, which is by nature Glycine betaine synthesis iterative [1]. Finally, very few of the transgenic plants pro- Glycine betaine (GlyBet) is synthesized in the chloroplast duced to date have been thoroughly analyzed at the in two steps from choline (Figure 2a) and can accumulate metabolic level. This is critical because most of them accu- to high levels (>20 mM on a tissue water basis, Table 1) in mulate only small amounts of the desired osmoprotectant, plants such as spinach and sugar beet under osmotic stress. 130 Plant biotechnology

Figure 2

(a) Chloroplast Cho Bet ald GlyBet cmo badh codA codA ptd-EA ptd-MME ptd-DME ptd-Cho

CDP-EA CDP-MME CDP-DME CDP-Cho Cho Bet ald GlyBet cdh cdh

P-EA P-MME P-DME P-Cho peamt Storage Cho

(b) Sucrose

UDP Glycolysis/ Sucrose-6P Glucose Fructose Krebs cycle UDP Fructose PPi UTP F6P Triose UDPG G1P G6P F6P F16P 2 phosphates

tps1 s6pdh mtldh

Trehalose-6P myo-inositol-1P Sorbitol-6P Mannitol-1P

tpp * **

Trehalose myo-inositol Sorbitol Mannitol imt1

D-ononitol Current Opinion in Plant Biology

Metabolic pathways associated with GlyBet, polyol and trehalose dehydrogenase activity; cdh, Cho dehydrogenase; codA, Cho oxidase; synthesis. (a) GlyBet synthesis (modified from [12••]) and (b) polyol peamt, P-EA-N-methyltransferase; mtldh, mannitol dehydrogenase; and trehalose synthesis. The osmoprotectants are given in bold. s6pdh, sorbitol dehydrogenase; imt1, inositol methyltransferase; Abbreviations are CDP-, cytidyldiphospho-; EA, ethanolamine; MME, tps1, trehalose-6-phosphate synthase; tpp, trehalose-6-phosphate monomethylEA; DME, dimethylEA; Bet ald, betaine aldehyde; cmo, phosphatase. Reactions with an asterisk (*) are non-specific choline monooxygenase; nsdh, endogenous non-specific aldehyde phosphatase activity.

Several groups have reported engineering GlyBet synthe- tobacco de novo choline synthesis is probably constrained at sis by introducing choline-oxidizing genes from E. coli [18], the step catalyzed by phosphoethanolamine-N-methyl- Arthrobacter spp. [19•,20], and spinach [12••]. GlyBet levels transferase [12••]. In this context, a comparison of the in the transgenic plants represented just a few percent of demand for choline moieties between a plant that lacks those found in plants that naturally accumulate it. Two GlyBet and a plant that synthesizes it is instructive. groups have shown that supplying choline in axenic cul- Figure 3 makes clear that GlyBet synthesis represents a tured plants enhanced GlyBet synthesis, indicating a huge demand on choline synthesis, requiring more than constraint in choline synthesis ([12••], Huang et al. abstract 90% of the choline moieties. This suggests that the choline in Plant Physiol 1997, 114S:120). Detailed analysis of synthesis pathway is far more active in GlyBet accumula- choline metabolism in tobacco demonstrates that it is tors. Increasing the choline supply in tobacco will require embedded in a rigid network in which choline is directed additional engineering steps. One possibility is to increase almost exclusively to phosphatidyl-choline synthesis, mak- choline synthesis Via up-regulation of phospho- ing it difficult to divert choline to GlyBet [12••]. Also, in ethanolamine-N-methyltransferase activity. Metabolic engineering of plants for osmotic stress resistance Nuccio et al. 131

Polyol synthesis Figure 3 In some plants polyol synthesis is upregulated in response to osmotic stress, leading to significant polyol accumulation (up to 50 mM on a tissue water basis, Table 1). Polyols 20 Tobacco are derived from sugar phosphates by reduction and Spinach dephosphorylation (Figure 2b). Mannitol, sorbitol and D- ononitol synthesis have been introduced into transgenic fw

plants. Two reports are of particular interest from a meta- -1 bolic standpoint; one illustrates a competitive or 10

antagonistic effect between transgene activity and host- mol g plant metabolism, and the other a synergistic effect. First, µ analysis of sorbitol levels in transgenic tobacco showed that high sorbitol accumulation was associated with growth defects and necrosis. This was attributed to myo-inositol 0 depletion, although sorbitol toxicity and cytosolic P deple- Cho i P-Cho ptd-Cho GlyBet tion due to sorbitol-6-P build-up were not ruled out [16•]. G-P-Cho Second, in tobacco, the myo-inositol pool expands as the Molecule plants are salt- or drought-stressed, and in plants engi- Current Opinion in Plant Biology neered to synthesize D-ononitol from myo-inositol, the Glycine betaine synthesis represents a large demand on choline increased precursor supply makes D-ononitol production synthesis. Total demand for choline moieties in a GlyBet accumulator •• stress inducible [21 ]. (spinach) and a non-accumulator (tobacco). Abbreviations are Cho, choline; P-Cho, phosphocholine; G-P-Cho, glycerophosphorylcholine; Trehalose synthesis ptd-Cho, phosphatidylcholine; GlyBet, glycine betaine. The data are •• The non-reducing disaccharide trehalose is believed to derived from [12 ,45–48] and also from the authors’ unpublished results. enable desiccation resistant organisms to survive dehydration stress [17••]; it is synthesized from glucose-6-phosphate and uridine-diphosphoglucose and pool sizes (see [23]). Once developed, the model can (Figure 2b). In plants engineered to synthesize trehalose, be tested experimentally to assess its accuracy. The most only very small amounts accumulate (Table 1). In two important feature of models is their potential to predict the reports trehalose-6-P-synthase (tps) alone was introduced, impact of a modification (e.g. inserting a transgene to leaving dephosphorylation of trehalose-6-P to an endoge- upregulate a step in the pathway) on pathway flux. This can nous non-specific phosphatase activity. In one report both guide the engineering process by simulating the outcome tps and trehalose-6-P phosphatase (tpp) were inserted, but of various experimental strategies. For instance, a model for even with both genes, a dramatic increase in trehalose did choline metabolism in tobacco has been developed to sim- not occur due to its degradation by trehalase [17••]. This ulate strategies for engineering GlyBet synthesis. It can be was demonstrated by the use of a trehalase inhibitor, which accessed at the world wide web site (http://www.hort.pur- increased trehalose accumulation and identified a con- due.edu/cfpesp/models/models.htm) for constructing straint on trehalose synthesis in plants that do not naturally interactive metabolic models. accumulate it [17••]. A related, but distinct, approach to understanding and pre- Metabolic engineering requires a guided, dicting metabolic flux is metabolic control analysis (MCA). iterative approach MCA is basically concerned with the steady state of a As noted above, Table 1 highlights the relative lack of metabolic pathway, in which the enzyme activities and attention to metabolic analysis of transgenic plants, in par- metabolite pool sizes remain constant. MCA shows that ticular in vivo estimates of pathway fluxes and intermediate control of pathway flux is typically shared among all the pool sizes made using isotopic tracer methods. Such data enzyme activities in the pathway, in sharp contrast with the can be incorporated into metabolic models, which are very traditional concept of a single ‘rate-limiting’ step [24]. helpful tools to describe and predict the behavior of the tar- Each enzyme’s contribution to control of flux can vary and get pathway (Figure 4). Models have a long track record of MCA defines this behavior as it relates to an individual application to the metabolic engineering of microorganisms step, or the entire pathway. The analysis can be used in [22••], and to metabolic pathway characterization in plants simulation experiments similar to those outlined above. [23]. Model development begins with a metabolic map as MCA has contributed both to microbial metabolic engi- shown for GlyBet synthesis (Figure 2a) or polyol synthesis neering [22••], and to basic understanding of the control of (Figure 2b). The input data consist of intermediate and metabolism [25••]. For example, a recent report, using end-product pool size measurements, and flux-rates calcu- MCA to evaluate transgenic plants, clearly demonstrates lated from timed in vivo isotope labeling data. The that the control of pathway flux is shared among the com- modeling program is used to fit the data to the described ponent reactions within the pathway and is not limited to pathway, which is done interactively by varying flux rates an individual step [26••]. 132 Plant biotechnology

Figure 4

A generic biochemical model as displayed at the modeling website. This web page is part of a modeling site which has been established to allow scientists to understand and work with biochemical models (http://www.hort. purdue.edu/cfpesp/models/models.htm). By adjusting fluxes and pool sizes with literature values as a guide, model curves can be generated and compared with experimental radiolabeling data. When the model and experimental results are in agreement, the fluxes and pool sizes determined can be very informative. Manipulating the flux or pool of interest in the model provides a guide to the effects of altering that particular flux or pool size. Each letter in the generic model represents a metabolic pool (E, A, B, C and D). The numbers beside the letters designate: specific activity of that pool at time zero (nCi/nmol), 1; the initial pool size (nmol/gfw), 2; the flux into the pool (nmol/min/gfw), 3; the combined rate of utilization from the pool (nmol/min/gfw), 4. The variable (k) determines the rate of uptake of the radiolabeled precursor, E. As the pool of E is drawn down, its rate of uptake declines proportionately. TS is the total simulation time in minutes. MR is the maximum radioactivity (nCi/gfw) for the left-hand graph and MP is the maximum pool size (nmol/gfw) for the right-hand graph. DS is the scaling factor for pool D to allow the curve to fit on the right graph. This modeling procedure was used in the analysis of 3- dimethylsulfoniopropionate synthesis in S. alterniflora [23].

Advancing technology to aid the one at a time. Such an approach was used to manipulate engineering process secondary metabolite production in maize cells [32]. The Two useful tools to help diagnose problems in transgenic basic principle is to manipulate the expression of a regu- plants are DNA micro-array technology and in vivo NMR latory gene (e.g. a specific transcription factor), which in spectroscopy. Micro-array technology was developed to turn alters the expression of the entire regulon. Over- evaluate the expression of many genes at once. The ana- expression of CBF1, a transcription factor that controls lytical power of this technology has been demonstrated in the expression of cold-responsive genes in Arabidopsis, yeast [27••], and is being applied to plants [28]. It provides demonstrated the feasibility of this approach in a plant high-resolution gene expression data which can be used to stress context [31••]. Other distinct transcription factors identify (unanticipated) changes in expression that follow that may function in a similar way to CBF1 have also been the insertion of a transgene. In vivo NMR spectroscopy can described [33]. An Arabidopsis regulatory locus (esk1) that be used to detect and quantitate several metabolites at governs genes involved in both the synthesis and break- once. In some cases the compartmentation of a metabolite down of proline has been identified [34•], and is, can be distinguished [29••]. Although rarely as sensitive as therefore, an attractive candidate for regulon engineer- standard in vivo biochemical or mass spectral methods, in ing. It is important to note, however, that ectopic vivo NMR techniques can be used to make otherwise expression of regulatory genes can only modify a plant’s impossible measurements — for instance, high resolution natural capabilities; it cannot confer new ones (e.g. the two-dimensional 31P-NMR spectroscopy was recently ability to synthesize a novel osmoprotectant). used to directly measure metabolic flux rates [30•]. Conclusions Engineering metabolic change is not a trivial endeavor. It Regulon engineering — manipulating entire will typically require iterative rounds of transformation pathways using regulatory genes guided by thorough analysis of each transgenic generation. Expression of regulatory genes that control several steps Interactive metabolic modeling promises to streamline this in a pathway or in related pathways (i.e. a regulon [31••]), process by helping predict which step(s) will require alter- may provide an alternative to introducing pathway genes ation through additional rounds of engineering. Emerging Metabolic engineering of plants for osmotic stress resistance Nuccio et al. 133

technology including DNA micro-arrays and in vivo NMR 13. Nelson DE, Rammesmayer G, Bohnert HJ: Regulation of cell- specific inositol metabolism and transport on plant salinity could also speed and simplify the analysis of genetically tolerance. Plant Cell 1998, 10:753-764. engineered plants. Only when a biochemical trait, such as 14. Russell BL, Rathinasabapathi B, Hanson AD: Osmotic stress osmoprotectant biosynthesis, is successfully engineered induces expression of choline monooxygenase in sugar beet and can the physiological consequences of that trait be amaranth. Plant Physiol 1998, 116:859-865. assessed. Metabolic engineering research will teach us not 15. Stephanopoulos G, Vallino JJ: Network rigidity and metabolic engineering in metabolite overproduction. Science 1991, only how to engineer biochemical change but also much 252:1675-1681. about the metabolic pathways themselves. 16. Sheveleva EV, Marquez S, Chmara W, Zegeer A, Jensen RG, • Bohnert HJ: Sorbitol-6-phosphate dehydrogenase expression in transgenic tobacco: high amounts of sorbitol lead to necrotic Acknowledgements lesions. Plant Physiol 1998, 117:831-839. Work in the authors’ laboratories is supported by the National Science This report highlights a downside in metabolic engineering. It shows that Foundation (IBN 9816075 and IBN 9813999), the United States extensive redirection of metabolic flux to sorbitol depletes the myo-inositol Department of Agriculture National Research Initiative Competitive Grants pool, and that this correlates with necrosis and growth defects. Clearly, Program (95-37100-1596 and 98-35100-6149), the Office of Naval Research transgene expression created an imbalance in primary metabolism; this (N00014-96-1-0364 and N000149-96-1-0366) and the CV Griffin Sr. could perhaps be corrected through additional rounds of engineering. Foundation and the Florida Agricultural Experiment Station. Journal series 17. Goddijn OJM, Verwoerd TC, Voogd E, Krutwagen RWHH, number R-06645. •• de Graaf PTHM, Poels J, van Dun K, Ponstein AS, Damm B, Pen J: Inhibition of trehalase activity enhances trehalose accumulation References and recommended reading in transgenic plants. Plant Physiol 1997, 113:181-190. The analytical experiments in this paper show that an endogenous trehalase Papers of particular interest, published within the annual period of review, activity is one factor limiting trehalose synthesis in transgenic plants. have been highlighted as: Specific inhibition of trehalase increased trehalose accumulation. It also • of special interest reports the expression of trehalose-6-P synthase both alone and with tre- •• of outstanding interest halose-6-P phosphatase (tpp), and that the presence of tpp has a small pos- itive effect on trehalose accumulation. 1. McCue KF, Hanson AD: Drought and salt tolerance: towards understanding and application. Trends Biotech 1990, 8:358-362. 18. Lilius G, Holmberg N, Bülow L: Enhanced NaCl stress tolerance in transgenic tobacco expressing bacterial choline dehydrogenase. 2. Jain RK, Selvaraj G: Molecular genetic improvement of salt Biotechnology 1996, 14:177-180. • tolerance in plants. Biotechnol Annu Rev 1997, 3:245-267. An excellent review of recent progress towards engineering salt-resistance 19. Hayashi H, Alia, Mustardy L, Deshnium P, Ida M, Murata N: in plants. It covers the physiological aspects of salt-tolerance and includes a • Transformation of Arabidopsis thaliana with the codA gene for discussion of the adaptive strategies found in nature. choline oxidase; accumulation of glycinebetaine and enhanced tolerance to salt and cold stress. Plant J 1997, 12:133-142. 3. Rhodes D, Samaras Y: Genetic control of osmoregulation in plants. This was the first report on the expression of a foreign choline oxidizing gene In Cellular and Molecular Physiology of Cell Volume Regulation. in chloroplasts. The transformed plants synthesized GlyBet and were slight- Edited by Strange K. Boca Raton: CRC Press; 1994:347-361. ly more tolerant to salt and cold stress. 4. Yancey PH: Compatible and counteracting solutes. In Cellular and 20. Alia, Hayashi H, Sakamoto A, Murata N: Enhancement of the Molecular Physiology of Cell Volume Regulation. Edited by tolerance of Arabidopsis to high temperatures by genetic Strange K. Boca Raton: CRC Press; 1994:81-109. engineering of the synthesis of glycine betaine. Plant J 1998, 16:155-162. 5. Smirnoff N: Plant resistance to environmental stress. Curr Opin Biotechnol 1998, 9:214-219. 21. Sheveleva EV, Chmara W, Bohnert HJ, Jensen RG: Increased salt •• and drought tolerance by D-ononitol production in transgenic 6. LeRudulier D, Strom AR, Dandekar AM, Smith LT, Valentine RC: Nicotiana tabacum L. Plant Physiol 1997, 115:1211-1219. Molecular biology of osmoregulation. Science 1984, This paper reports the production of D-ononitol in transgenic plants. The 224:1064-1068. myo-inositol pool size increased with stress, which led to greater D-ononitol 7. Nelson DE, Shen B, Bohnert HJ: Salinity tolerance-mechanisms, production and osmotic stress resistance. It illustrates how metabolic engi- models and the metabolic engineering of complex traits. Genet neering can lead to insight into endogenous metabolic characteristics. Engineering 1998, 20:153-176. 22. Bailey JE: Mathematical modeling and analysis in biochemical 8. Bohnert HJ, Sheveleva E: Plant stress adaptations — making •• engineering: past accomplishments and future opportunities. metabolism move. Curr Opin Plant Biol 1998, 1:267-274. Biotechnol Prog 1998, 14:8-20. A historical discussion of mathematical modeling and its application to meta- 9. Holmberg N, Bülow L: Improving stress tolerance in plants by gene bolic problems. It highlights the importance of modeling complex metabolic transfer. Trends Plant Sci 1998, 3:61-66. networks which, in turn, facilitate the engineering process. 10. Yeo A: Molecular biology of salt tolerance in the context of whole 23. Kocsis MG, Nolte KD, Rhodes D, Shen T-L, Gage DA, Hanson AD: • plant physiology. J Exp Botany 1998, 49:915-929. Dimethylsulfonioproprionate biosynthesis in Spartina alterniflora: A whole-plant perspective on molecular approaches to engineering stress evidence that S-methylmethionine and dimethylsulfonio tolerance which points out that osmotic stress resistance encompasses a propylamine are intermediates. Plant Physiol 1998, 117:273-281. number of mechanisms, many of which are unknown. It cautions against focusing on a single trait such as osmoprotectant synthesis as a solution to 24. ap Rees T: Prospects of manipulating plant metabolism. Trends introducing resistance, and favors the identification and exploitation of quan- Biotech 1995, 13:375-378. titative trait loci associated with stress resistance. 25. Fell DA: Understanding the Control of Metabolism. London: 11. Hanson AD, Rathinasabapathi B, Rivoal J, Burnet M, Dillon DO, •• Portland Press; 1997. Gage DA: Osmoprotective compounds of the Plumbaginaceae: a This book provides a thorough and highly readable introduction to metabol- natural experiment in metabolic engineering of stress tolerance. ic control analysis (MCA). It bridges the gap between traditional enzymolo- Proc Natl Acad Sci USA 1994, 91:306-310. gy and MCA, and demonstrates how MCA overcomes some of the descriptive limitations of traditional enzymology. 12. Nuccio ML, Russell BL, Nolte KD, Rathinasabapathi B, •• Gage DA, Hanson AD: The endogenous choline supply limits 26. Thomas S, Mooney PJF, Burrell MM, Fell DA: Finite change analysis glycine betaine synthesis in transgenic tobacco expressing •• of glycolytic intermediates in tuber tissue of lines of transgenic choline monooxygenase. Plant J 1998, 16:487-496. potato (Solanum tuberosum) overexpressing This paper illustrates the utility of metabolic analysis in the engineering phosphofructokinase. Biochem J 1997, 322:111-117. process. It demonstrates that the spinach choline monooxygenase enzyme An instructive demonstration of the use of MCA. Phosphofructokinase has functions in tobacco, and describes experiments that show how the choline been traditionally thought of as a ‘rate-limiting’ enzyme in glycolysis. supply limits GlyBet synthesis in these plants. It also suggests engineering However, overexpressing it in transgenic plants does not result in an steps that may enhance GlyBet synthesis in the next generation of trans- increase in glycolytic flux. The authors use MCA to show that flux control is genic plants. shared by the many steps between fructose-6-P and the Krebs cycle. 134 Plant biotechnology

27. DeRisi JL, Iyer VR, Brown PO: Exploring the metabolic and genetic genes and downregulate degradative genes, implying that it is near the top •• control of gene expression on a genomic scale. Science 1997, of a signaling cascade. 278:680-686. DNA micro-array technology enables the expression of many genes to be 35. Kavi Kishor PB, Hong Z, Miao G-H, Hu C-AA, Verma DPS: measured at once. This paper is the first to report bringing this analytical pro- Overexpression of D1-pyrroline-5-carboxylate synthetase cedure to the genomic scale, in yeast. The experiments clearly demonstrate increases proline production and confers osmotolerance in how alterations in growth, mutations and genetic manipulations can affect transgenic plants. Plant Physiol 1995, 108:1387-1394. the expression of many genes. 36. Tarczynski MC, Jensen RG, Bohnert HJ: Expression of a bacterial 28. Ruan Y, Gilmore J, Conner T: Towards Arabidopsis genome mtlD gene in transgenic tobacco leads to production and analysis: monitering expression profiles of 1400 genes using accumulation of mannitol. Proc Natl Acad Sci USA 1992, cDNA microarrays. Plant J 1998, 15:821-833. 89:2600-2604. 29. Aubert S, Curien G, Bligny R, Gout E, Douce R: Transport, 37. Tarczynski MC, Jensen RG, Bohnert HJ: Stress protection of •• compartmentation, and metabolism of homoserine in higher plant transgenic tobacco by production of the osmolyte mannitol. cells: carbon-13- and phosphorus-31-nuclear magnetic Science 1993, 259:508-510. resonance studies. Plant Physiol 1998, 116:547-557. A good example of how in vivo NMR spectroscopy can complement classi- 38. Thomas JC, Sepahi M, Arendall B, Bohnert HJ: Enhancement of cal biochemical analysis. 13C- and 31P-NMR were used to monitor the seed germination in high salinity by engineering mannitol uptake and metabolism of homoserine. The experiments demonstrate how expression in Arabidopsis thaliana. Plant Cell Environ 1995, metabolite pool size data and enzymatic rate data can be derived from NMR 18:801-806. spectra. An experiment to determine the distribution of phosphohomoserine 39. Karakas B, Ozias-Akins P, Stushnoff C, Suefferheld M, Reiger M: between the cytosol and the vacuole is also described. Salinity and drought tolerance of mannitol-accumulating 30. Roscher A, Emsley L, Raymond P, Roby C: Unidirectional steady transgenic tobacco. Plant Cell Environ 1997, 20:609-616. • state rates of central metabolism enzymes measured Sorbitol synthesis in transgenic simultaneously in a living plant tissue. J Biol Chem 1998, 40. Tao R, Uratsu SL, Dandekar AM: 273:25053-25061. tobacco with apple cDNA encoding NADP-dependent sorbitol-6- This paper describes how in vivo NMR spectroscopy can be used to mea- phosphate dehydrogenase. Plant Cell Physiol 1995, 36:525-532. sure the flux of phosphate from substrates to products. The methods to mon- 41. Vernon DM, Tarczynski MC, Jensen RG, Bohnert HJ: Cyclitol itor four enzymatic reactions in central metabolism are described and used production in transgenic tobacco. Plant J 1993, 4:199-205. to show how each reaction responds to changes in external stimuli. 42. Holmström K-O, Mäntylä E, Wellin B, Mandal A, Palva ET, Tunnela OE, 31. Jaglo-Ottosen KR, Gilmour SJ, Zarka DG, Schabenberger O, Londesborough J: Drought tolerance in tobacco. Nature 1996, •• Arabidopsis CBF1 overexpression induces COR Thomashow MF: 379:683-684. genes and enhances freezing tolerance. Science 1998, 280:104-106. 43. Romero C, Bellés JM, Vayá JL, Serrano R, Culiáñez-Macià FA: This is the first report to demonstrate ‘regulon engineering’ in a plant stress Expression of the yeast trehalose-6-phosphate synthase gene in context. CBF1 had been shown to interact with a number of cold-reponsive transgenic tobacco plants: pleiotropic phenotypes include genes during the ‘cold-acclimation’ process, which leads to an increase in drought tolerance. Planta 1997, 201:293-297. freezing tolerance. The experiments show that overexpressing CBF1 induces the acclimation effect and enhances freezing tolerance. 44. Popp M, Smirnoff N: Polyol accumulation and metabolism during water deficit. In Environment and Plant Metabolism: Flexibility and 32. Grotewold E, Chamberlin M, Snook M, Siame B, Butler L, Swenson J, Acclimation. Edited by Smirnoff N. Oxford: Bios Scientific Publishers Maddock S, St Clair G, Bowen B: Engineering secondary Limited; 1995:199-215. metabolism in maize cells by ectopic expression of transcription factors. Plant Cell 1998, 10:721-740. 45. Storey R, Wyn Jones RG: Salt stress and comparative physiology in the Gramineae. II: glycinebetaine and proline accumulation in 33. Liu Q, Kasuga M, Sakuma Y, Abe H, Miura S, Yamaguchi-Shinozaki K, two salt- and water-stressed barley cultivars. Aust J Plant Physiol Shinozaki K: Two transcription factors, DREB1 and DREB2, with an 1978, 5:817-829. EREBP/AP2 DNA binding domain separate two cellular signal transduction pathways in drought- and low-temperature- 46. Hanson AD, May AM, Grumet R, Bode J, Jamieson GC, Rhodes D: responsive gene expression, respectively, in Arabidopsis. Plant Betaine synthesis in chenopods: localization in chloroplasts. Cell 1998, 10:1391-1406. Proc Natl Acad Sci USA 1985, 82:3678-3682. 34. Xin Z, Browse J: eskimo1 mutants of Arabidopsis are constitutively 47. Bligny R, Gardestrom P, Roby C, Douce R: 31P NMR studies of • freezing tolerant. Proc Natl Acad Sci USA 1998, 95:7799-7804. spinach leaves and their chloroplasts. J Biol Chem 1990, This paper describes the identification and characterization of an 265:1319-1326. Arabidopsis mutant that is defective in a key regulatory gene. It is a new locus, distinct from other cold-acclimation pathways. The data show that, in 48. Roughan PG, Batt RD: The glycerolipid composition of leaves. proline metabolism, this gene can simultaneously upregulate synthesis Phytochemistry 1969, 8:363-369.