Sucrose Efflux Mediated by SWEET Proteins as a Key Step for Phloem Transport

See allHide authors and affiliations

Science  13 Jan 2012:
Vol. 335, Issue 6065, pp. 207-211
DOI: 10.1126/science.1213351


Plants transport fixed carbon predominantly as sucrose, which is produced in mesophyll cells and imported into phloem cells for translocation throughout the plant. It is not known how sucrose migrates from sites of synthesis in the mesophyll to the phloem, or which cells mediate efflux into the apoplasm as a prerequisite for phloem loading by the SUT sucrose–H+ (proton) cotransporters. Using optical sucrose sensors, we identified a subfamily of SWEET sucrose efflux transporters. AtSWEET11 and 12 localize to the plasma membrane of the phloem. Mutant plants carrying insertions in AtSWEET11 and 12 are defective in phloem loading, thus revealing a two-step mechanism of SWEET-mediated export from parenchyma cells feeding H+-coupled import into the sieve element–companion cell complex. We discuss how restriction of intercellular transport to the interface of adjacent phloem cells may be an effective mechanism to limit the availability of photosynthetic carbon in the leaf apoplasm in order to prevent pathogen infections.

Breeding has led to marked increases in crop yield. Increased yield potential has mainly been attributed to improvements in allocation efficiency, defined as the amount of total biomass allocated into harvestable organs (1, 2). Despite the critical importance of sucrose translocation in this process, we do not understand mechanistically how changes in translocation efficiency may have contributed to an increase in harvestable products. Allocation of photoassimilates in plants is conducted by transport of sucrose from the photosynthetic “sources” (predominantly leaves) to the heterotrophic “sinks” (meristems, roots, flowers, and seeds) (35). Sucrose, the predominant transport form of sugars in many plant species [see (6) for an overview of the different sugars and translocation mechanisms], is produced in leaf mesophyll cells, particularly in the palisade parenchyma of dicots and the bundle sheath of monocots.

In apoplasmic loaders, sucrose is loaded into the sieve element–companion cell complex (SE/CC) in the phloem by the sucrose-H+ cotransporter SUT1 (named SUC2 in Arabidopsis) from the apoplasm (cell wall space) (711). However, sucrose must effuse from inside the cell into the cell wall, either directly from mesophyll cells (after which it travels to the phloem in the apoplasm) or from cells closer to the site of loading (having traveled cell-to-cell through plasmodesmata). Both the site and the mechanism of sucrose efflux remain to be elucidated, although it has been argued that a site in the vicinity of the site of phloem loading is most probable (4, 5).

We identified proteins that can transport sucrose across the plasma membrane: AtSWEET10 to 15 in Arabidopsis and OsSWEET11 and 14 in rice (Oryza sativa). We found that AtSWEET11 and 12 are expressed in phloem cells, and that inhibition by mutation reduces leaf assimilate exudation and leads to increased sugar accumulation in leaves. Thus, apoplasmic phloem loading occurs in a two-step model: Sucrose exported by SWEETs from phloem parenchyma cells feeds the secondary active proton-coupled sucrose transporter SUT1 in the SE/CC.

The sucrose efflux transporters were identified by means of a Förster resonance energy transfer (FRET) sensor-based screen. Because humans do not seem to possess sucrose transporters, we reasoned that human cell lines should lack endogenous sucrose transport activity and should thus represent a suitable functional expression system for heterologous sucrose transporters. A preliminary set of ~50 candidate genes encoding membrane proteins with unknown function as well as members of the recently identified SWEET glucose effluxer family (12) were coexpressed with the FRET sucrose sensor FLIPsuc90μ∆1V (13) in human embryonic kidney (HEK) 293T cells. AtSWEET10 to 15, which all belong to clade III of the AtSWEET family (12), enabled HEK293T cells to accumulate sucrose, as detected by a negative ratio change in sensor output (Fig. 1A). To corroborate these findings, we tested the clade III orthologs OsSWEET11 and 14 from rice (Fig. 1B) and found that they also transport sucrose. By contrast, proteins from the other SWEET clades did not show detectable sucrose uptake into HEK293T cells (Fig. 1A). Clade III SWEETs showed preferential transport activity for sucrose over glucose and did not appear to transport maltose (Fig. 1C and fig. S2). The ability of clade III SWEETs to export sucrose was shown by time-dependent efflux of [14C]sucrose injected into Xenopus oocytes (Fig. 1D and fig. S2D) and was further supported by the reversibility of sucrose accumulation as measured by optical sensors in mammalian cells (Fig. 1E and fig. S3). SWEETs function as low-affinity sucrose transporters (affinity constant Km for sucrose uptake by AtSWEET12 was ~70 mM, Km for efflux was >10 mM; Fig. 1F and fig. S4, A to C). The largely pH-independent transport activity supports a uniport mechanism (fig. S4D). The observed transport characteristics are compatible with those of the low-affinity components for sucrose transport detected in vivo (14, 15).

Fig. 1

Identification of sucrose transporters. (A and B) HEK293T cell–FRET sensor uptake assay. (A) Of ~50 membrane protein genes tested, AtSWEET10 to 15 showed sucrose influx as measured with the sucrose sensor FLIPsuc90μ∆1V; HEK293T cells transfected with sensor only (“control”) or the sensor and the H+/sucrose cotransporter StSUT1 served as controls (fig. S1) (7) (±SEM, n ≥ 11). (B) The rice transporters OsSWEET11 and 14 mediate sucrose transport in HEK293T cells (±SEM, n ≥ 11). (C) Oocyte uptake assay: OsSWEET11 and 14 and AtSWEET11 and 12 mediate [14C]sucrose uptake (1 mM sucrose; ±SEM, n ≥ 7). (D) Oocyte efflux assay: [14C]sucrose efflux by OsSWEET11 in Xenopus oocytes injected with 50 nl of a solution containing 50 mM [14C]sucrose; the truncated version OsSWEET11_F205* served as control (±SEM, n ≥ 7). (E) HEK293T cell–FRET sensor transport assay: Reversible accumulation of sucrose in HEK293T cells by AtSWEET11 (±SEM, n ≥ 10; Fc/D, FRET index). (F) Oocyte uptake assay: Kinetics of AtSWEET12 for sucrose uptake in Xenopus oocytes (±SEM, n ≥ 14).

AtSWEET11 and 12 are highly expressed in leaves [microarray data and translatome data (16, 17) (figs. S5A and S6)] and were found to be coexpressed with genes involved in sucrose biosynthesis and phloem loading (e.g., sucrose phosphate synthase, SUC2, and AHA3; fig. S5, B and C). The tissue-specific expression and cellular localization of AtSWEET11 and 12 and the phenotypes of sweet mutants were analyzed to determine the physiological role of the sucrose transporters.

AtSWEET11 and 12 are close paralogs, with 88% similarity at the amino acid level. Lines carrying single T-DNA (transfer DNA; Agrobacterium tumefaciens) insertions (fig. S7) in the AtSWEET11 and 12 loci did not show any obvious altered morphological phenotype when compared to the wild-type Col-0 or wild-type siblings segregated from the same mutant populations. However, at higher light levels, the double mutant line was smaller relative to wild-type controls (20 to 35% reduction in rosette diameter depending on light conditions; Fig. 2A and fig. S8) and contained elevated starch levels at the end of the diurnal dark period (Fig. 2, B and C). Moreover, mature leaves of the double mutant contained higher sucrose levels both at the end of the light period and the end of the dark period (Fig. 2D). Leaves also accumulated higher levels of hexoses, as also observed in plants exposed to sucrose (18) or plants in which phloem loading has been blocked (9, 19). Accumulation of free sugars is expected to lead to down-regulation of photosynthesis through sugar signaling networks (20). The starch accumulation phenotype was partially complemented by expression of either AtSWEET11 or 12 under their respective promoters in the double mutant (fig. S9). Together, these data indicate an impaired ability of the mutants to export sucrose from the leaves. Direct [14CO2]-labeling experiments indicated that under low light conditions (in which the double mutant did not accumulate high starch levels; fig. S10), the double mutant exported ~50% of fixed 14C relative to controls (Fig. 2E). It is noteworthy that the mutant was affected with respect to leaf size, photosynthetic capacity, and steady-state sugar levels; thus, the apparent efflux rates may be compounded by these parameters.

Fig. 2

Phenotypic characterization of AtSWEET11 and 12 mutants. (A) Reduced growth of the atsweeet11;12 double mutant relative to Col-0 wild type and isogenic wild type (control). (B and C) Elevated starch accumulation in atsweeet11;12 single and double mutants at the end of the light and dark periods (high light conditions). (D) Sugar levels in mature leaves at the end of the light and dark periods (±SEM, n ≥ 6; identical letters indicate significance between pairs (daytime) according to t test (P ≤ 0.001; high light conditions). Glc, glucose; Frc, fructose; Suc, sucrose; Tre, trehalose; Mal, maltose; Raf, raffinose; Xyl, xylose; c, control; 11;12, atsweet11;12. (E) Cumulative exudation of [14C]-derived assimilates from cut petioles of leaves fed with 14CO2 (14C in exudate shown as percentage in exudate plus exudate from the previous exudation period for each time point; ±SEM, n = 5; *t significant at P < 0.05, **t significant at P < 0.01) (low light conditions). (F and G) Impaired root growth of atsweet11;12 seedlings grown on sugar-free media and media supplemented with sucrose [±SEM, n ≥ 60; two-way analysis of variance indicates a significant interaction (P < 0.0001) between genotype and sucrose treatment].

Reduced efflux of sugars from leaves is expected to lead to reduced translocation of photoassimilates to the roots, thus negatively affecting root growth and the ability to acquire mineral nutrients (9, 10). When germinated in the light on sugar-free media, atsweet11;12 mutants exhibited reduced root length (Fig. 2, F and G). Addition of sucrose to the media rescued the root growth deficiency of atsweet11;12 mutants (Fig. 2, F and G). A similar sucrose-dependent root growth deficiency has also been observed for the Arabidopsis sucrose/H+ cotransporter suc2 mutant (11). Both the suc2 and the atsweet11;12 mutants are apparently able to acquire sucrose or sucrose-derived hexoses from the medium to restore root growth that had been restricted by a carbohydrate deficiency.

The growth phenotype for atsweet11;12 is not as severely affected as described previously for the suc2 mutant (911). The Arabidopsis genome encodes several SWEET paralogs, including the closely related transporters AtSWEET10, 13, 14, and 15, which we show here to function as sucrose transporters. Quantitative polymerase chain reaction analyses showed that AtSWEET13, which is typically expressed at low levels in leaves, is induced by a factor of ~16 in the atsweet11;12 double mutant (fig. S11). Thus, in contrast to the secondary active SE/CC loaders SUT1/SUC2, SWEETs function as redundant elements of phloem loading. It is noteworthy that ossweet14 rice mutants display stunted growth, possibly as a result of reduced sugar efflux from leaves as well (21).

Taken together, the data indicate that clade III SWEETs are involved in export of sucrose and are responsible for the previously undescribed first step in phloem loading. The efflux of sucrose to the apoplasm could theoretically occur directly at the site of production in mesophyll cells, from bundle sheath cells, or from phloem parenchyma cells. Localization of AtSWEET11 and 12 driven by their native promoters, as translational enhanced green fluorescent protein (eGFP) or GUS fusions, revealed that both proteins are present in the vascular tissue including minor and major veins, which in Arabidopsis are considered “to participate in phloem loading” (22) (Fig. 3, A to D, and fig. S12). The subcellular localization of eGFP-tagged AtSWEET11 and 12 was consistent with localization to the plasma membrane [Fig. 3, E and F; further supported by data from cauliflower mosaic virus (CaMV) 35S-SWEET-YFP plants, fig. S13]. AtSWEET11 and 12 were both expressed in select cells in the phloem, which form cell files along the veins (Fig. 3, C, D, and F, and fig. S12). Most likely, these cells correspond to phloem parenchyma. However, there are no known markers that would allow us to unambiguously identify these cells. Data from cell-specific translatome studies show that AtSWEET11/12-expressing cells have a clearly distinct translatome when compared to SUC2-expressing companion cells (fig. S6) (17). These data exclude the possibility that SWEET11 and 12 are expressed to high levels in companion cells, supporting a localization in phloem parenchyma cells as the only remaining cell type in the phloem besides the enucleate sieve elements.

Fig. 3

GUS and eGFP localization of AtSWEET11 and 12 promoter-reporter fusions. (A to D) GUS histochemistry analysis in rosette leaves of transgenic Arabidopsis plants expressing translational GUS fusions of AtSWEET11 [(A), (C), and (D)] or 12 (B) from their native promoters. (A) and (B), GUS staining detected in leaf vein network; (C), high-resolution images of expression in one cell file of an individual vein; (D), cross section of Arabidopsis leaf showing cell-specific localization of AtSWEET11. (E and F) Confocal images of eGFP fluorescence in sepal vein cell files of transgenic Arabidopsis plants expressing translational AtSWEET11-eGFP fusions under control of its native promoter. Insets in (F) show eGFP channel in black and white; red dashed line in upper inset indicates position of z-scan shown in lower inset. eGFP accumulation is observed in static puncta, which may be caused by accumulation of AtSWEET11 in membranes in cell wall ingrowths, a characteristic feature of phloem parenchyma cells (33). The presence of cell wall ingrowth was confirmed by electron microscopy (fig. S14).

Further, OsSWEET11/Xa13, encoding a sucrose uniporter and functioning as a rice susceptibility (S) gene (Xa13) for specific pathovars of Xanthomonas oryzae pv. oryzae, was found to be expressed in the phloem of uninfected rice leaves (23), indicating that OsSWEET11 may play a similar role in phloem loading. Coimmunolocalization of SUT1/SUC2 and SWEET11/12 to an extent detectable by transmission electron microscopy will be required to unambiguously define the cell type in which the SWEETs are functioning.

Our findings are compatible with a model suggested by Geiger (24), in which sucrose moves symplasmically via plasmodesmata toward the phloem and then effluxes close to the site of apoplasmic loading (fig. S15). We predict that communication is needed to coordinate the efflux from phloem parenchyma with the uptake into the SE/CC to prevent spillover and limit the availability of nutrient resources for pathogens in the apoplasm of the leaf. Invertases and glucose/H+ cotransporters that are induced during pathogen infection may serve in retrieval of sugars spilled at the infection site (25). It is tempting to speculate that sugar- and turgor-controlled regulatory mechanisms involved in post-phloem unloading may also apply to sucrose efflux in the phloem loading process (26, 27). The availability of SWEET sucrose transporters, together with FRET sensors (28), provides valuable tools for studying the regulatory networks coordinating local and long-distance transport and metabolism.

Clade III SWEETs had previously been implicated as key targets of biotrophic pathogens. OsSWEET11 and 14 are co-opted during infection of rice by Xanthomonas oryzae pv. oryzae (Xoo) (12, 21, 29, 30). Pathovar-specific effectors secreted by Xoo activate the transcription of clade III SWEET genes, and mutations in the effector binding sites in SWEET promoters lead to resistance to Xoo in a wide spectrum of rice lines (21, 2931). Our finding that these SWEETs are key elements of the phloem translocation machinery indicates that the pathogen retools a critical physiological function (i.e., a cellular sucrose efflux mechanism in the phloem) to gain access to the plant’s energy resources at the site of infection.

However, this function is redundant in the plant. Such redundancy in both pathogen and host functions has been attributed to increased system robustness and may have evolved to allow the plant to survive mutations in essential functions that create pathogen resistance (32). It is possible that the highly localized transfer of sucrose between phloem parenchyma and SE/CC has evolved to limit sucrose release into the apoplasm to a limited interface of adjacent cells inside the phloem, and thus to reduce the availability of sucrose in the apoplasm to pathogens. Pathogens can overcome this first line of defense by targeting exactly this efflux mechanism in order to gain access to sugars in cells surrounding the infection site—for example, in the epidermis or mesophyll. Invertase and monosaccharide transporters, which are also typically induced during infection, may then serve as a secondary line of defense to reduce apoplasmic sugar levels at the infection site (25). The work presented here adds a crucial item to the list of machinery essential for carbon allocation: the transport proteins responsible for the efflux of sucrose to apoplasm in preparation for phloem loading.

Supporting Online Material

Materials and Methods

Figs. S1 to S15

References (3460)

References and Notes

  1. Acknowledgments: We thank G. Grossmann and D. Ehrhardt for advice and help with confocal imaging; J. Bailey-Serres for help with the translatome analyses; K. Barton and T. Liu for plastic embedding and sectioning help and advice; and V. Lanquar and A. Jones for critical reading of the manuscript. Supported by U.S. Department of Energy grant DE-FG02-04ER15542 and National Institute of Diabetes and Digestive and Kidney Diseases grant 1RO1DK079109 (W.B.F.); the Carnegie Institution and the Scholarship Program of the Chinese Scholarship Council (grant 2009635108) (X.-Q.Q.); and the Max-Planck-Gesellschaft (S.O. and A.R.F.). Author contributions: W.B.F. and L.-Q.C. conceived and designed the experiments. L.-Q.C., X.-Q.Q., D.S., B.-H.H., and S.O. performed the experiments. W.B.F., L.-Q.C., X.-Q.Q., D.S., B.-H.H., S.O., and A.R.F. analyzed the data. L.-Q.C. and W.B.F. wrote the manuscript.
View Abstract

Navigate This Article