SummaryThe Arabidopsis thaliana chlorophyll a/b-binding protein underexpressed 1 (cue1) mutant shows a reticulate leaf phenotype and is defective in a plastidic phosphoenolpyruvate (PEP)/phosphate translocator (AtPPT1). A functional AtPPT1 providing plastids with PEP for the shikimate pathway is therefore essential for correct leaf development. The Arabidopsis genome contains a second PPT gene, AtPPT2. Both transporters share similar substrate speci®cities and are therefore able to transport PEP into plastids. The cue1 phenotype could partially be complemented by ectopic expression of AtPPT2 but obviously not by the endogeneous AtPPT2. Both genes are differentially expressed in most tissues: AtPPT1 is mainly expressed in the vasculature of leaves and roots, especially in xylem parenchyma cells, but not in leaf mesophyll cells, whereas AtPPT2 is expressed ubiquitously in leaves, but not in roots. The expression pro®les are corroborated by tissue-speci®c transport data. As AtPPT1 expression is absent in mesophyll cells that are severely affected in the cue1 mutant, we propose that the vasculature-located AtPPT1 is involved in the generation of phenylpropanoid metabolism-derived signal molecules that trigger development in interveinal leaf regions. This signal probably originates from the root vasculature where only AtPPT1, but not AtPPT2, is present.

a b s t r a c tThe Arabidopsis thaliana gene At1g74030 codes for a putative plastid phosphoenolpyruvate (PEP) enolase (ENO1). The recombinant ENO1 protein exhibited enolase activity and its kinetic properties were determined. ENO1 is localized to plastids and expressed in most heterotrophic tissues including trichomes and non-root-hair cells, but not in the mesophyll of leaves. Two T-DNA insertion eno1 mutants exhibited distorted trichomes and reduced numbers of root hairs as the only visible phenotype. The essential role of ENO1 in PEP provision for anabolic processes within plastids, such as the shikimate pathway, is discussed with respect to plastid transporters, such as the PEP/phosphate translocator.

Restriction of phosphoenolpyruvate (PEP) supply to plastids causes lethality of female and male gametophytes in Arabidopsis thaliana defective in both a phosphoenolpyruvate/phosphate translocator (PPT) of the inner envelope membrane and the plastid-localized enolase (ENO1) involved in glycolytic PEP provision. Homozygous double mutants of cue1 (defective in PPT1) and eno1 could not be obtained, and homozygous cue1 heterozygous eno1 mutants [cue1/eno1 (+/2)] exhibited retarded vegetative growth, disturbed flower development, and up to 80% seed abortion. The phenotypes of diminished oil in seeds, reduced flavonoids and aromatic amino acids in flowers, compromised lignin biosynthesis in stems, and aberrant exine formation in pollen indicate that cue1/eno1(+/2) disrupts multiple pathways. While diminished fatty acid biosynthesis from PEP via plastidial pyruvate kinase appears to affect seed abortion, a restriction in the shikimate pathway affects formation of sporopollonin in the tapetum and lignin in the stem. Vegetative parts of cue1/eno1(+/2) contained increased free amino acids and jasmonic acid but had normal wax biosynthesis. ENO1 overexpression in cue1 rescued the leaf and root phenotypes, restored photosynthetic capacity, and improved seed yield and oil contents. In chloroplasts, ENO1 might be the only enzyme missing for a complete plastidic glycolysis. INTRODUCTIONPhosphoenolpyruvate (PEP) plays a central role in plant metabolism. As an intermediate of glycolysis, PEP is indispensable for energy metabolism in the cytosol and delivers ATP and pyruvate by the action of cytosolic pyruvate kinase (PK) (Plaxton, 1996;Givan, 1999). Pyruvate can be fed into the citric acid cycle, yielding NADH for respiratory ATP generation (Fernie et al., 2004). Inside the plastids, PEP may act as a precursor for at least four metabolic pathways ( Figure 1A).Together with erythrose 4-phosphate, PEP is fed into the shikimate pathway, which delivers essential aromatic amino acids and a large number of secondary plant products. The initial steps of the shikimate pathway are exclusively localized within the plastid stroma (Herrmann, 1995;Schmid and Amrhein, 1995;Herrmann and Weaver, 1999). Inside the stroma, PEP can also be sequentially metabolized to pyruvate and acetyl-CoA by plastid PK and the pyruvate dehydrogenase complex (Reid et al., 1977;Elias and Givan, 1979; Lernmark and Gardeströ m, 1994) and thus enter the biosynthesis of fatty acids (Dennis, 1989;Ohlrogge and Jaworski, 1997), which are quantitatively important for triacylglycerol production in oil seeds (e.g., Voelker and Kinney, 2001;Rawsthorne, 2002;Ruuska et al., 2002). Like the shikimate pathway, the de novo biosynthesis of fatty acids for membranes and storage lipids is localized to the plastids (Ohlrogge et al., 1979;Ohlrogge and Jaworski, 1997). Moreover, stromal pyruvate can act as a precursor for the synthesis of branched-chain amino acids (Schulze-Siebert et al., 1984) and together with glyceraldehyde 3-phosphate for the mevalonate-independent way (2-C-methyl-...

Phosphoenolpyruvate (PEP) serves not only as a high energy carbon compound in glycolysis, but it acts also as precursor for plastidial anabolic sequences like the shikimate pathway, which produces aromatic amino acids (AAA) and subsequently secondary plant products. After conversion to pyruvate, PEP can also enter de novo fatty acid biosynthesis, the synthesis of branched-chain amino acids, and the non-mevalonate way of isoprenoid production. As PEP cannot be generated by glycolysis in chloroplasts and a variety of non-green plastids, it has to be imported from the cytosol by a phosphate translocator (PT) specific for PEP (PPT). A loss of function of PPT1 in Arabidopsis thaliana results in the chlorophyll a/b binding protein underexpressed1 (cue1) mutant, which is characterized by reticulate leaves and stunted roots. Here we dissect the shoot- and root phenotypes, and also address the question whether or not long distance signaling by metabolites is involved in the perturbed mesophyll development of cue1. Reverse grafting experiments showed that the shoot- and root phenotypes develop independently from each other, ruling out long distance metabolite signaling. The leaf phenotype could be transiently modified even in mature leaves, e.g. by an inducible PPT1RNAi approach or by feeding AAA, the cytokinin trans-zeatin (tZ), or the putative signaling molecule dehydrodiconiferyl alcohol glucoside (DCG). Hormones, such as auxins, abscisic acid, gibberellic acid, ethylene, methyl jasmonate, and salicylic acid did not rescue the cue1 leaf phenotype. The low cell density1 (lcd1) mutant shares the reticulate leaf-, but not the stunted root phenotype with cue1. It could neither be rescued by AAA nor by tZ. In contrast, tZ and AAA further inhibited root growth both in cue1 and wild-type plants. Based on our results, we propose a model that PPT1 acts as a net importer of PEP into chloroplast, but as an overflow valve and hence exporter in root plastids.