Hexose/H⁺ Symporters in Lower and Higher Plants

Although green plants are carbon autotrophic organisms, synthesizing their sugars as well as all other organic compounds photosynthetically, plant cells nevertheless require sugar uptake mechanisms. Higher plants are physiological mosaics of autotrophic green and heterotrophic non-green cells and tissues. The roots, stems and reproductive organs are supplied with organic compounds and the leaves deliver sucrose and other oligosaccharides to them. Green algae, in contrast, can frequently grow autotrophically but also mixotrophically, making use of sugars, amino acids and organic acids (Droop, 1974).

Chlorella kessleri, a unicellular green algae, can be grown under autotrophic as well as under heterotrophic conditions; in the latter case, glucose serves as sole carbon source. During the shift from carbon autotrophy to heterotrophy, hexose transport activity increases more than 200-fold (Tanner, 1969; Haaß and Tanner, 1973). The transport system induced by sugars or non-metabolizable sugar analogues has been characterized as an H⁺ symporter system (Komor, 1973; Komor and Tanner, 1974a, 1976), which transports sugars and protons with a stoichiometry of 1:1 and is able to accumulate analogues such as 6-deoxyglucose more than 1000-fold (Komor et al. 1973a). This accumulation has partly been explained by distinctly different K_m values for sugar influx and efflux and partly by differences in the estimated velocity constants contributing to influx compared with those contributing to efflux (Komor et al. 1973a; Komor and Tanner, 1974b).

Sugar transport and accumulation in this lower eukaryote differ in a number of aspects from those in bacteria. For example, when uncouplers are added to cells that have reached the steady state of sugar analogue accumulation, they lose the sugar only very slowly. The same phenomenon is observed under anaerobic conditions when sugar accumulation is energized by light and this light is subsequently switched off (Komor et al. 1972, 1973b). Sugar efflux is also very slow when cells with a high internal sugar analogue concentration are resuspended in sugar-free medium, but it can be stimulated by external sugar (a positive trans effect for efflux of more than 50-fold; Komor et al. 1972). Finally, the transporter is completely inactive for all fluxes (influx, efflux and exchange flux) when the intracellular pH is 6.0 or below, although it is optimally active at an extracellular pH of 6.0 (Komor et al. 1979). These observations led to the conclusion that the Chlorella hexose transport protein acts in an asymmetric way. Although this asymmetric action is not understood in any mechanistic detail, the gene for the corresponding transport protein has been cloned (Sauer and Tanner, 1989) and an in vitro transport test has been established (Opekarová et al. 1994), so the questions of how sugars are transported through membranes, how the carriers sense and make use of energy and how the transport step may be regulated may eventually be answered.

As discussed above, hexose uptake in Chlorella kessleri is highly inducible, which suggested that the gene coding for the transport protein could be cloned by differential screening of cDNA from induced versus that from non-induced cells. The HUP1 cDNA (hexose uptake protein) obtained in this way (Sauer and Tanner, 1989) was shown by heterologous expression in Schizosaccharomyces pombe to code for a transport protein responsible for hexose uptake, for sensing a pH gradient and for the accumulation of sugar analogues such as 3-O-methylglucose (Sauer et al. 1990a). The mRNA of the HUP1 gene, absent in photosynthetically grown cells, appears within 5 min after addition of sugars (Hilgarth et al. 1991; R. Stadler, K. Wolf, C. Hilgarth, W. Tanner and N. Sauer, in preparation). The hydropathy plot suggested a structure with 12 transmembrane domains, which has been postulated for a large family of sugar transporters (Marger and Saier, 1993); the model for the Chlorella transporter is shown in Fig. 1. Heterologous expression in a S. pombe mutant not able to grow on, or to take up, D-glucose (M. Höfer and B. Milbradt, unpublished results) allowed a first investigation of the structure/function relationship using site-directed mutagenesis (Caspari et al. 1994).

Among other observations, it was shown that the three histidyl residues (His-73, His-170 and His-495) could be replaced by arginine without a marked decrease in activity. Moreover, the 27 C-terminal amino acids, but not the 43 C-terminal ones, could be removed without affecting transporter activity. Interestingly, the amino acids Gln-179 and Gln-299, conserved in all hexose transporters sequenced so far, when replaced by asparagines resulted in K_m changes for D-glucose by a factor of 10 (1×10^×4 to 2×10^×4 mol l⁻¹ instead of 1.5×10^×5 mol l⁻¹).

To look for other amino acids that may be responsible for substrate binding and specificity, random mutants of HUP1 cDNA were prepared using the polymerase chain reaction (PCR) technique. S. pombe cells grown on gluconate were than transformed with a pool of mutated HUP1 cDNA and transformants were selected which showed decreased sensitivity towards the toxic sugar analogue 2-deoxyglucose (Will et al. 1994). Whereas most transformants showed a drastically decreased rate of glucose uptake, four mutants were obtained which possessed a clearly increased K_m value for D-glucose: Q179E (helix V), Q298R (helix VII) and V433L and N436Y (both in helix XI). Since the two mutants Q298R and N436Y showed an increased K_m value of 1000-fold (although reasonable V_max values), whether more conservative amino acid exchanges in these two positions would still give rise to increased K_m values was investigated. This was indeed the case, as demonstrated with the directed mutations Q298N and N436Q; the K_m values increased 10-to 20-fold (Will et al. 1994). It is proposed, therefore, that helices V, VII and XI of the Chlorella HUP1 transporter line the sugar translocation path (see inset of Fig. 1) and that the hexose specificity may, in part, be determined by a number of amino acids, four amides (Q179, Q298, Q299 and N436) and one valine (V433), which are located in the putative centre of these helices. The observation that yeast hexokinase B crystallized with a derivatized glucose forms hydrogen bonds to five amino acids, four of which are amides (Anderson et al. 1978), may give some support to this hypothesis.

When HUP1 cDNA was expressed in frog oocytes and substrate specificity was determined by following sugar-dependent membrane depolarization (Aoshima et al. 1993), it was noticed that D-galactose was not transported by the product of HUP1 cDNA, although induced Chlorella kessleri does take up D-galactose (Tanner et al. 1969). Chlorella, however, possesses two more hexose transporter genes (HUP2 and HUP3), which are co-induced by D-glucose (R. Stadler, K. Wolf, C. Hilgarth, W. Tanner and N. Sauer, in preparation). Recently, it was shown that HUP2 cDNA expressed in S. pombe is clearly a D-galactose transporter (relative rates were galactose 100 and glucose 71), whereas the HUP1 cDNA product transports galactose at less than 10% of the rate of D-glucose transport (R. Stadler, K. Wolf, C. Hilgarth, W. Tanner and N. Sauer, in preparation).

With the help of probes derived from HUP1 cDNA, it has been possible to clone and characterize the first hexose transporter STP1 (sugar transport protein) of a higher plant, Arabidopsis thaliana (Sauer et al. 1990b). In the meantime, it has become clear from PCR studies that Arabidopsis contains at least 12 genes highly related to STP1 (K. Baier and N. Sauer, unpublished results). Moreover, a large family of hexose transporter genes has been described from Ricinus communis (Weig et al. 1994).

Although it is not yet clear whether all the highly related transporter genes of Arabidopsis really transport hexoses, it has at least been shown for four (STP1–STP4) and may well be true for all of them, with possible differences in substrate specificities. The sucrose transporters cloned and sequenced so far (Riesmeier et al. 1992; Sauer and Stolz, 1994) are very different from the hexose transporter family (Sauer and Tanner, 1993) and, therefore, are not expected to be present among STP5–STP12. Why would the small Arabidopsis plant possess 12 hexose transporters when, in humans, as far as we know, seven seem to be sufficient (Bell et al. 1993)? And even more puzzling: why do plants require hexose transporters at all, if the main sugars sent from photosynthesizing leaves (source tissue) to carbon consuming or storing organs (sink tissues) are either sucrose or, less frequently, galactosides of sucrose (sugars of the raffinose family)? Answers to the first question are rapidly coming into sight. The various transporters are expressed in a highly tissue-specific manner: STP1 in the ovary, STP2 in the anthers, STP3 in the sepals and the stigma, STP4 in the anthers but also in the root tips (E. Truernit and N. Sauer, unpublished results). In addition, it seems likely that some of the transporters are not constitutively expressed, but that their genes may be switched on in response to external (e.g. stress) or internal (hormonal) signals. Thus, a set of hexose transporter genes, each with a specific promotor, may be responsible for constitutive, but tissue-specific, expression or for developmentally or environmentally regulated expression. These transporters may, therefore, constitute the link between developmental information and metabolism and may thereby determine the formation of the various plant cells, tissues and organs in time and space.

What about the second question? Why do plants possess hexose transporters at all? It is true that the main sugars for long-distance translocation within plants are either sucrose or galactosides of sucrose (Zimmermann and Ziegler, 1975). However, it has been speculated for a long time (Glasziou and Gayler, 1972; Eschrich, 1980) that sucrose, after leaving the long-distance translocation system, may be split extracellularly by cell wall invertase and that the monosaccharides arising may be the actual substrates to be taken up by the cells to be fed. This suggestion has recently obtained support from the observation that overexpression of cell wall invertase in leaves of transgenic plants leads to an inhibition of sucrose export from these leaves and an increased re-use of monosaccharides by them (van Schaewen et al. 1990). In addition, a maize mutant, defective in kernel filling, was shown to lack a specific cell wall invertase (Miller and Chourey, 1992). Finally, a Chenopodium tissue culture, which can be grown photoautotrophically as well as mixo-or heterotrophically (Hüsemann and Barz, 1977), expresses three hexose transporter genes (out of at least seven found in this plant), but the degree of expression does not differ under the three widely differing growth conditions (Roitsch and Tanner, 1994), whereas an extracellular invertase is expressed considerably more extensively when these cells depend on extracellular sugars for growth (T. Roitsch, M. Bittner and D. Godt, in preparation). It seems likely, therefore, that the cooperation of hexose transporters and extracellular cell wall invertase is important for feeding non-photosynthesizing plant cells with organic compounds (Fig. 2) and it is tempting to speculate that the coupling of these two proteins may determine fully, or in part, what has long been termed ‘sink strength’ in plant physiology (Ho, 1988).

Transport studies with membrane vesicles were first introduced by Kaback in the 1960s and have been intensively applied to investigate the molecular details of the Escherichia coli lactose/H⁺symport (Kaback, 1989). An in vitro system established in Konings’ laboratory (Driessen et al. 1993), using right-side-out vesicles energized by cytochrome c oxidase, is widely used to study H⁺ symporters. In vitro studies of secondary active transporters of plants have so far been performed with plasma membrane vesicles isolated by the aqueous two-phase partitioning method (Bush, 1989; Buckhout, 1989; Lemoine and Delrot, 1989; Williams et al. 1992). This uptake system was energized by the formation of a transient membrane potential difference. Recently, it became possible to measure transport mediated by the HUP1 gene product of Chlorella kessleri (Opekarová et al. 1994) in the vesicle system developed by Driessen et al. (1993). Plasma membranes of S. pombe transformed with the HUP1 cDNA were fused with proteoliposomes containing cytochrome c oxidase. After energization with ascorbate/TMPD/cytochrome c, a proton-motive force of more than 130 mV was built up for at least 2 h. D-Glucose was accumulated in such vesicles up to 30-fold; this was not the case in control vesicles prepared from cells transformed with the plasmid not containing the HUP1 cDNA (Fig. 3). Using this in vitro system, it was possible to demonstrate that the degree of accumulation was determined by both kinetic and by thermodynamic parameters. Since the system also allows the manipulation of both components of the proton-motive force individually, the extent to which they contributed to sugar accumulation could be evaluated (Opekarová et al. 1994).

In the meantime, this in vitro system has also been successfully applied to the study of sucrose/H⁺symporters of higher plants (Stolz et al. 1994; M. Gahrtz, J. Stolz and N. Sauer, in preparation).

The importance of studying plant transporters in a well-defined in vitro system is obvious. First, any purification of fully active transporter proteins requires reconstitution and activity measurement in vitro. Purification of hexose and sucrose transporters has been successfully started (Stolz et al. 1994). Second, to resolve and to understand some of the unusual properties of the Chlorella hexose transporter mentioned above, in vitro studies are essential. The 20-year-old observation that the Chlorella glucose transporter in vivo never acts as a facilitator except in the presence of polyene antibiotics, which interact with sterols (Komor et al. 1974), can now be reinvestigated in vitro by changing the lipid composition of the vesicles (Opekarová and Tanner, 1994).

The original work reported in this review was supported by the Deutsche Forschungsgemeinschaft (Normalverfahren and SFB 43) and by the Fonds der Chemischen Industrie.