Report

A Previously Unknown Maltose Transporter Essential for Starch Degradation in Leaves

See allHide authors and affiliations

Science  02 Jan 2004:
Vol. 303, Issue 5654, pp. 87-89
DOI: 10.1126/science.1091811

Abstract

A previously unknown maltose transporter is essential for the conversion of starch to sucrose in Arabidopsis leaves at night. The transporter was identified by isolating two allelic mutants with high starch levels and very high maltose, an intermediate of starch breakdown. The mutations affect a gene of previously unknown function, MEX1. We show that MEX1is a maltose transporter that is unrelated to other sugar transporters. The severe mex1 phenotype demonstrates that MEX1is the predominant route of carbohydrate export from chloroplasts at night. Homologous genes in plants including rice and potato indicate that maltose export is of widespread significance.

The conversion of starch to sucrose in leaves in the dark is one of the largest metabolic fluxes in living organisms, amounting to tens of millions of tons of carbon every night. Sucrose and starch are the primary products of photosynthetic carbon assimilation in the light. Sucrose is synthesized in the cytosol of leaf cells and exported to the nonphotosynthetic parts of the plant, and starch is accumulated within the chloroplast. In many species, up to half of the newly assimilated carbon may be used for starch synthesis. In the dark, starch degradation supplies carbon for continued sucrose synthesis and export and for the immediate metabolic requirements of the leaf cell. Despite the importance of this process, the degradative pathway itself remains largely unknown (13). Evidence increasingly favors the involvement early in the pathway of a chloroplastic beta amylase, the main product of which is maltose. For example, when the activity of this enzyme was reduced, transgenic potato plants had reduced rates of leaf starch degradation at night (4). Although it is not yet clear where in the cell or by means of which enzymes the further metabolism of maltose occurs, there is indirect evidence that it may be exported from the chloroplast for metabolism in the cytosol. First, independent in vivo approaches demonstrate that either maltose or glucose is likely to be the major exported metabolite (57). Second, maltose can cross the chloroplast envelope in vitro and may be a major product of starch degradation in these conditions (810). However, no maltose transporter has been identified and the importance of maltose export in vivo is poorly understood.

To discover the fate of the maltose produced during starch degradation, we looked for mutants with elevated maltose levels among a population of Arabidopsis thaliana mutants that were preselected for reduced rates of starch degradation at night [starch excess (sex) mutants (11, 12)]. Our previous experience has shown that lesions in the pathway of starch degradation that lead to malto-oligosaccharide accumulation also result in a reduced overall rate of degradation and hence a sex phenotype (13). Two independent mutant lines with greatly elevated levels of maltose (14) (fig. S1) were named maltose excess1-1 and -2 (mex1-1 and mex1-2). The mutations were shown by crossing to be monogenic, recessive, and allelic.

The mex1 plants were considerably smaller than wild-type plants and had lower chlorophyll contents (Fig. 1) (15). Enzyme assays and native gels revealed little difference in enzymes of starch metabolism between mutant and wild-type plants (table S1) (14, 15). The maltose levels of the mutants were at least 40 times as high as those of wild-type leaves. The amount of maltose fell during the day and increased at night. Starch levels were elevated, and the rates of both synthesis and degradation were reduced (Fig. 1) (14). Levels of sucrose and hexoses were greatly elevated during the day but fell to levels comparable to those of wild-type leaves at night (fig. S2). These data indicated that mex1 plants lack a component of the pathway of starch degradation necessary for maltose metabolism. In its absence, the initial release of glucans from the starch granule is inhibited and there are perturbations in the diurnal pattern of primary carbohydrate metabolism.

Fig. 1.

The role and location of the MEX1 protein. (A) Wild-type (Col-0, left) and mex1-1 (right) plants grown in 12-hour-day, 12-hour-night conditions. Chlorophyll content of mature leaves of mex1-1 was 60% of that of wild-type leaves under these conditions. (B and C) Levels of maltose (B) and starch (C) in leaves of wild-type (closed circles) and mex1-1 (open circles) plants in 12-hour-day, 12-hour-night conditions. Values are means ± SEM of measurements on five plants. In (B), note the discontinuous y axis.

The MEX1 gene was identified by map-based cloning (14). In mex1-1 plants, a point mutation in the locus At5g17520 on chromosome V gives a premature stop codon in the fourth exon of this gene (TGG to TGA). In mex1-2 plants, the replacement of 61 base pairs (bp) of At5g17520 with 10 bp of unrecognized DNA is predicted to lead to miss-splicing and a stop codon in the aberrant mRNA (Fig. 2A). To check whether the mutations in At5g17520 were indeed responsible for the mex1 phenotype, we expressed a full-length MEX1 cDNA on a 35S promoter in mex1-1 plants. This experiment confirmed that At5g17520 is responsible for all elements of the phenotype. Transformants that contained the transferred DNA construct were like wild-type rather than mex1 plants with respect to growth rate, leaf color, starch content (15), and maltose content (fig. S3).

Fig. 2.

Analysis of the MEX1 gene and protein. (A) The MEX1 locus in Arabidopsis. Exons are depicted as solid boxes and the mutation sites in mex1-1 and mex1-2 are shown. (B) Transmembrane domains in MEX1 are shown as solid and hatched orange boxes. The consensus prediction for AtMEX1 is for nine transmembrane domains (TM2-10) (16). The predictions correspond well to the hydrophobic domains in OsMEX1 with the exception of TM10. An additional region (TM1), not reliably predicted as a transmembrane domain in AtMEX1, is shared by OsMEX1. The green box depicts the predicted 47–amino acid transit peptide of AtMEX1. OsMEX1 has a predicted transit peptide of 67 amino acids. (C and D) Subcellular location of MEX1. (C) Confocal microscopic image, showing chlorophyll fluorescence in a leaf of a mex1-1 mutant plant transformed with a 35S::MEX1-YFP construct. Scale bar, 20 μm. (D) Confocal microscopic image of the same area of leaf as that in (C), showing YFP fluorescence. The green YFP fluorescence coincides with the regions of chlorophyll fluoresence shown in (C). Scale bar, 20 μm. (E) Complementation of the E. coli strain KU98 (lacking the malF component of the maltose transport system) with LacZ::MEX1 (dashed curve) or the empty pSU18 vector (solid curve). Bacteria were grown with 0.5% (w/v) maltose as the only carbon source, and growth was monitored at 600 nm. Each point represents the mean ± SEM of measurements on five bacterial cultures. (F) The phenotype of the mex1-1/dpe1-1 double mutant. A wild-type plant (left) and four double mutant segregants from the F2 population (right). Plants were grown in 16-hour-light, 8-hour-dark conditions and photographed at the same scale when they were 39 days old.

MEX1 is a single-copy gene encoding a predicted membrane protein with a chloroplast transit peptide (Fig. 2B and fig. S4) (16). The gene is expressed in both the leaves and the roots of Arabidopsis but had no known function (17). The predicted protein is not notably similar to any other Arabidopsis protein. However, a predicted plastid-targeted protein encoded in the rice genome (OsMEX1, compared with AtMEX1 in Fig. 2B) is 70% identical in amino acid sequence, exactly the same length, and almost identical in predicted structure. There are excellent matches between large parts of the MEX1 sequence and peptides predicted from expressed sequence tags (ESTs) from plant species including Angiosperms, Gymnosperms, and mosses (fig. S5). The class of proteins defined by AtMEX1, OsMEX1, and these ESTs is unique, with no notable sequence similarity to any other known proteins.

The phenotype and predicted subcellular location of the MEX1 protein indicate that the protein is a chloroplast envelope maltose transporter. To discover the location of MEX1, a C-terminal fusion between full-length MEX1 cDNA and yellow fluorescent protein (YFP) was expressed in mex1-1 plants (14). Similar to the full-length MEX1 cDNA alone, this MEX1-YFP fusion protein complemented the mex1 phenotype. Maltose levels in seven independent transformants were within the range expected for wild-type plants, and transformants were indistinguishable in appearance from wild-type plants (15). This indicates that the fusion protein is correctly targeted. YFP fluorescence in the leaves of transformants was specifically associated with chloroplast envelopes (Fig. 2, C and D, and fig. S6).

To discover whether the protein can facilitate movement of maltose across membranes, we observed the effect of expression of the mature protein (minus the transit peptide) in a mutant of Escherichia coli that lacked MalF, a membrane protein that forms part of the endogenous maltose transporter (18). malF mutants cannot grow on maltose as a sole carbon source. Expression of the MEX1 protein allowed the growth of the malF mutant on maltose, whereas the introduction of the empty vector did not (Fig. 2E). Thus, it is highly likely that MEX1 is a novel type of maltose transporter.

The very high levels of maltose in mex1 plants indicate that MEX1 is the main route by which the products of starch degradation are exported from the chloroplast at night. To investigate further the involvement of MEX1 in starch degradation, we crossed mex1-1 to a mutant lacking the ability to synthesize starch [pgm, lacking chloroplastic phosphoglucomutase (19)]. The double mutant was identical to pgm, lacking the high maltose and pale color of mex1 (table S2 and fig. S7). This shows that the high maltose in mex1 is derived from starch and that the specific function of MEX1 is to export it from the chloroplast. Our previous work indicated that glucose, as well as maltose, is a product of starch degradation in Arabidopsis chloroplasts. Glucose is produced by the action of the chloroplastic disproportionating enzyme on maltotriose, a minor product of beta amylase (13). The existence of a glucose transporter in the chloroplast envelope (20) potentially allows the export of glucose to the cytosol. To investigate this potential route of starch degradation, we crossed mex1-1 to a mutant lacking chloroplastic disproportionating enzyme (dpe1-1). The mex1/dpe1 double mutants grew very slowly compared with the wild type and parent lines (Fig. 2F). This suggests that there may be two parallel routes for the export of starch degradation products from the chloroplast at night: a major flux of maltose by means of MEX1 and a more minor flux of glucose derived from the disproportionation of maltriose (fig. S8 and supporting online material text).

The distribution of MEX1 transcript in both photosynthetic and nonphotosynthetic tissues of Arabidopsis and of MEX1-like genes in diverse species of plants indicates that the occurrence of maltose export from plastids is widespread. The discovery of MEX1 thus resolves the long-standing debate over the chloroplastic pathway of starch degradation for Arabidopsis and opens the way to the resolution of this problem in the plant kingdom as a whole.

Supporting Online Material

www.sciencemag.org/cgi/content/full/303/5654/87/DC1

Materials and Methods

SOM Text

Figs. S1 to S8

Tables S1 and S2

References

References and Notes

View Abstract

Navigate This Article