Vernalization Alters Sink and Source Identities and Reverses Phloem Translocation from Taproots to Shoots in Sugar Beet

Cold treatment prior to bolting transforms sugar beet taproots from sucrose-storing sink organs to sucrose-mobilizing source organs.

Abstract

During their first year of growth, overwintering biennial plants transport Suc through the phloem from photosynthetic source tissues to storage tissues. In their second year, they mobilize carbon from these storage tissues to fuel new growth and reproduction. However, both the mechanisms driving this shift and the link to reproductive growth remain unclear. During vegetative growth, biennial sugar beet (Beta vulgaris) maintains a steep Suc concentration gradient between the shoot (source) and the taproot (sink). To shift from vegetative to generative growth, they require a chilling phase known as vernalization. We studied sugar beet sink-source dynamics upon vernalization and showed that before flowering, the taproot underwent a reversal from a sink to a source of carbohydrates. This transition was induced by transcriptomic and functional reprogramming of sugar beet tissue, resulting in a reversal of flux direction in the phloem. In this transition, the vacuolar Suc importers and exporters TONOPLAST SUGAR TRANSPORTER2;1 and SUCROSE TRANSPORTER4 were oppositely regulated, leading to the mobilization of sugars from taproot storage vacuoles. Concomitant changes in the expression of floral regulator genes suggest that these processes are a prerequisite for bolting. Our data will help both to dissect the metabolic and developmental triggers for bolting and to identify potential targets for genome editing and breeding.

INTRODUCTION

Plants have a remarkable ability to store energy, starting with the energy from photons, which they use to fix carbon in photosynthetic tissues (source organs). Fixed carbon, in turn, is transported to different plant tissues (sink organs) to drive growth and development. Moreover, to drive future growth, plants store fixed carbon in storage tissues such as the seed endosperm, starch granules, and taproots of perennial and biennial plants. Because the carbohydrates stored in these organs constitute a major source of human nutrition, understanding the mechanisms by which source and sink tissues operate remains a major concern for crop improvement.

Suc is the primary sugar transported by the phloem from source to sink organs. After it is unloaded into the sink organ, Suc can be used either as an energy source or as a building block for growth and for the biosynthesis of storage compounds. Nonphotosynthetic storage organs like tubers and taproots must maintain a steep Suc concentration gradient between source (high) and sink (low) organs to maintain the source-to-sink direction of transport. To accomplish this, Suc imported into sink tissues is quickly converted into relatively inert storage compounds like starch or is compartmentalized into large intracellular vacuoles.

Sink organs, rich in stored carbon, can later become source organs as they mobilize their storage products to feed newly emerging sink organs. The identity of a plant organ as either sink or source is a dynamic process that is specified both by endogenous developmental signals (Turgeon, 1989) and in response to specific environmental stimuli (Roitsch, 1999). In addition, the switch from a sink to a source organ implies the dynamic regulation of gene expression and associated enzymatic activities involved in carbohydrate metabolism and sugar import to mobilize stored molecules and release them into the phloem (Viola et al., 2007; O’Neill et al., 2013; Liu et al., 2015; Boussiengui-Boussiengui et al., 2016). The balance between sink and source can be adjusted by the activity of two groups of proteins: (1) sucrose-synthesizing and -degrading enzymes (Herbers and Sonnewald, 1998) and (2) sucrose-loading proteins located in the plasma membrane of phloem cells (Imlau et al., 1999; Gottwald et al., 2000; Srivastava et al., 2008; Chen et al., 2012). Sucrose-metabolizing enzymes and transporters therefore represent relevant targets for breeding and biotechnological strategies aimed at increasing crop yield by releasing more stored sugars for growth (Ludewig and Sonnewald, 2016; Fernie et al., 2020; Lu et al., 2020).

Sugar beet (Beta vulgaris) is a major crop grown commercially for Suc production in the temperate zones of Europe and North America. Sugar beets have a biennial life cycle. During their first growth year, they form a large taproot that stores large amounts of Suc (up to 20% of their fresh weight, which can be 1 kg or more). The sugar beet vacuolar Suc loader TONOPLAST SUGAR TRANSPORTER2;1 (BvTST2;1) plays an essential role for sugar accumulation in the taproot (Jung et al., 2015). During the second growth year, the taproot switches to a source organ and thus provides previously stored Suc to sustain the growth of a large inflorescence and seed production.

The establishment of flowering (or bolting) competence in sugar beet requires prolonged exposure to cold temperatures below 15°C (Boudry et al., 2002), known as vernalization, between the two growth cycles. The development and growth of flowering structures leads to a significant loss of taproot sugar and biomass and, therefore, a net yield drop. To avoid this yield loss, sugar beets are cultivated as an annual crop. They are sown in the spring and harvested in the late fall of the same year just before photosynthesis stops supplying taproots with sugars. A prolonged cultivation period (particularly from fall to fall) would extend the growing period and thus sugar loading of the taproot, but it would also necessitate the identification of bolting-resistant varieties. This has become a primary goal in sugar beet breeding (Hoffmann and Kluge-Severin, 2011; Hoffmann and Kenter, 2018).

Two major early-bolting loci, designated B and B2, were identified in segregating populations derived from crosses between the biennial domesticated sugar beet and an annual wild relative. Map-based cloning revealed the identity of both genes: B is the pseudoresponse regulator gene BOLTING TIME CONTROL1 (BTC1; Pin et al., 2012) and B2 is DOUBLE B-BOX TYPE ZINC FINGER (BBX19; Dally et al., 2014). In annual beets, the expression of both genes leads to the repression of the floral repressor FLOWERING LOCUS T1 (FT1) and the subsequent induction of the floral inducer FT2 and vernalization-independent flowering in long days (Pin et al., 2010; Dally et al., 2014). Biennial beets are homozygous for nonfunctional btc1 and bbx19 alleles that are unable to prevent the repression of flowering imposed by FT1 (Pfeiffer et al., 2014). Accordingly, biennial sugar beets require vernalization for BTC1- and BBX19-independent FT1 repression and flowering (Pin et al., 2010). Because the sugar beet taproot feeds the development of the inflorescence, the regulation of floral induction and the sink-source transition must be tightly interconnected. The coordinated network of floral inducers and repressors that initiates the transition to bolting following vernalization has been well described (Ream et al., 2012). Although adjustment of metabolism appears equally important for the morphological and physiological changes taproots undergo prior to the formation of inflorescences, little is known about the molecular and physiological processes early in the vernalization process in sugar beet.

In this work, we therefore sought to understand how chilling temperatures, which are indispensable for bolting of biennial sugar beet plants, influence sugar metabolism, photosynthesis, phloem transport, and ultimately source and sink identities of shoots and taproots. We combined comprehensive transcriptome and proteome analyses with careful recording of organ growth characteristics, photosynthetic parameters, and metabolite profiles. We found that, despite the cessation of photosynthesis caused by cold temperatures, shoot biomass increased at the expense of taproot Suc content. We measured (1) substantial export of taproot sugar in the cold that correlated with altered activities of sugar exporters and importers and (2) markedly altered transcript levels for genes involved in either Suc biosynthesis or degradation. We speculate that this metabolic reprogramming is a prerequisite for the initiation of flowering, alongside a corresponding redirection of sugar flux from the taproot to the shoot. However, this process might also contribute to the pronounced frost sensitivity of sugar beet. Thus, our findings provide a molecular and physiological explanation for the well-known problem of sugar beet cultivation—the significant loss of yield due to the biennial life cycle—and offer new targets to achieve bolting resistance and winter hardiness in this crop species.

RESULTS

Cold Exposure Causes a Rapid Loss of Shoot and Root Water but Not a Loss of Shoot Biomass Production

We monitored the cold-dependent growth dynamics of sugar beet source and sink organs in three biennial hybrid genotypes (GT1, GT2, and GT3), which performed differently in field trials for various traits unrelated to bolting. We first grew plants under short-day conditions (10 h of light/14 h of darkness) for 6 weeks in control conditions (20°C), acclimated them for 1 week at 12°C, and then transferred them to cold temperature conditions (4°C) for 19 d. At this age, which corresponded to about stage 19 to 31 of the Biologische Bundesanstalt, Bundessortenamt, und CHemische Industrie scale (Meier et al., 1993), plants had already developed a taproot of the size of a human thumb.

During the course of the exposure to 4°C, we recorded shoot and taproot weights after 5, 10, and 20 d for a subset of plants (Figures 1A to 1C). Shoot dry weight but not fresh weight continued to increase during exposure to cold (Figures 1A and 1B). In agreement, shoot water content gradually decreased by almost 50% by the end of the phenotyping period (Figure 1C). Simultaneously, taproot fresh weight and dry weight decreased together with taproot water content during the cold exposure period (Figures 1A to 1C). These results demonstrated that the growth of taproots was more affected by the cold than the growth of shoots, and they further suggested distinct physiological and metabolic responses to cold exposure in shoot versus root tissues.

Figure 1.

Biomass and Sugar Contents in Cold-Treated Sugar Beet Shoots and Taproots.

(A) to (C) Response of shoots and taproots to cold temperatures using three different sugar beet genotypes. Recording of biomass and sugar accumulation started concurrently with the 4°C treatment. Fresh weight (FW; [A]), dry weight (DW; [B]), and water content (C) of shoots and roots are shown.

(D) to (G) Contents of glucose (D), Fru (E), Suc (F), and starch (G) during the course of the chilling (4°C) period in shoots and taproots.

Data points show means from n = 6 to 10 plants ± sd. Significant changes relative to the control condition (first data point) were calculated using one-way ANOVA with posthoc Tukey’s HSD (P < 0.05; Supplemental File 1).

Sugar Levels Respond to Cold Differently in Shoots and Taproots

The accumulation of soluble sugars in shoots is a common response to low temperatures and a feature of the cold acclimation process in many plant species (Steponkus, 1971; Strand et al., 1997; Wolfe and Bryant, 1999). Consistent with these observations, we detected a clear increase in Glc and Fru levels after transfer to 4°C in the same leaves we used for biomass and water content calculation (Figures 1D and 1E) and to a lesser extent for the disaccharide Suc (Figure 1F). In contrast to these soluble sugars, leaf starch content decreased rapidly after transfer to 4°C in all three genotypes, reaching only 20 to 33% of the starting value prior to cold transfer (Figure 1G).

The pattern of sugar accumulation behaved very differently between shoots and taproots. Glc and Fru levels increased in taproots upon exposure to the cold, but they reached only 10 to 20% of the total monosaccharide concentration seen in leaves at the same time point (Figures 1D and 1E). Prior to transfer to cold, taproot Suc levels exceeded those of monosaccharides by 30- to 100-fold (Figure 1F). Taproot starch levels were extremely low in all genotypes and hardly changed during cold treatment (Figure 1G). The three genotypes analyzed did, however, exhibit different sugar and starch accumulation dynamics in the cold: GT2 and GT3 taproot Suc levels decreased during cold exposure, but those of GT1 remained fairly constant. Interestingly, the steep drop in Suc concentration in taproots of GT3 (by ∼400 µmol/g dry weight) and to a lesser extend of GT2 (by ∼200 µmol/g dry weight) was not accompanied by a proportionate rise in monosaccharides, as might be expected for an exclusive hydrolysis of Suc. This massive net loss in taproot Suc suggested that this sugar was either (1) increasingly respired, (2) converted into molecules other than the monosaccharides Glc and Fru, or (3) exported into other organs. In the following sections, we aimed to elucidate the fate of Suc with respect to these possibilities.

Cold Exposure Affects Photosynthetic Rates and Carbon Dioxide Assimilation

Cold-tolerant plants such as Arabidopsis (Arabidopsis thaliana) accumulate sugars in leaves in response to cold, maintain photosynthetic activity, reduce Suc phloem loading, and increase sugar import into the vacuoles of leaf mesophyll cells (Strand et al., 1997; Wingenter et al., 2010; Pommerrenig et al., 2018). We analyzed the effect of cold temperatures on sugar beet photosynthesis using three complementary techniques: pulse amplitude-modulated (PAM) fluorometry, CO₂ assimilation with gas-exchange measurements, and expression profiling of photosynthesis-related genes (Figure 2). Our measurements in sugar beets revealed that plants exposed to cold responded with a decline in photosynthetic efficiency (Figures 2A and 2B) and that PSII quantum yield [Y(II)], leaf CO₂ concentration, CO₂ assimilation rate, and leaf transpiration rate were dependent on ambient temperature. In detail, our three genotypes showed a slight but significant reduction in Y(II) during the acclimation phase (1 week at 12°C). In parallel, nonphotochemical quenching [Y(NPQ)] increased and nonregulated energy dissipation [Y(NO)] decreased after acclimation to 12°C, again in all three genotypes. That Y(NPQ) quantum yield was higher at 12°C compared with 20°C indicated an increase in electron flow toward the Mehler ascorbate peroxidase pathway (Asada et al., 1998), which is responsible for thermal energy dissipation at PSII reaction centers. After transfer to 4°C, Y(II) decreased further and did not recover over the time period tested. However, the decrease in Y(NPQ) quantum yield and the concomitant rise in Y(NO) quantum yield suggested that electrons underwent unregulated energy dissipation, which can produce free radicals and result in membrane damage at this low temperature (Figure 2A).

Figure 2.

Photosynthetic Parameters, CO₂ Assimilation, and Expression Data of Sugar Beet Leaves after Cold Exposure.

(A) PAM measurements using leaves of the three different genotypes.

(B) Gas-exchange measurements. Intercellular leaf CO₂ concentration (C_i), CO₂ assimilation rate (A), and transpiration rate (E) are shown.

Data points in (A) and (B) show means from four independent plants used throughout the entire time period. Temperature intervals are highlighted in light orange (12°C) or light blue (4°C). Asterisks indicate significant changes to the 20°C condition (0 days after transfer) according to Students t-test (P < 0.05)

(C) Percentage of RNA-seq reads annotated as genes coding for photosynthesis (PS)-related proteins. Pie charts represent averaged means from three different genotypes at 20°C (control) and after 14 d at 4°C.

(D) Expression of sugar beet RCA (Bv2_025300_tzou.t1), RBCS (Bv2_026840_jycs.t1), CAB (Bv_002570_dmif.t1), and PC (Bv_004160_hgjn.t1). Data represent mean normalized cpm values of three independent RNA-seq analyses per genotype and temperature condition ± sd. Significant changes relative to the control condition (20°C) were calculated using Student’s t test (*, P < 0.05; Supplemental File 1).

Measurements of CO₂ gas exchange showed that reduced PSII activity, as determined by PAM fluorometry, was accompanied by a drastic decline of the CO₂ assimilation rate at 4°C but not at 12°C. Transpiration rates also increased in all three genotypes, starting when plants were acclimated to 12°C and continuing to rise transiently at 4°C. This elevated transpiration rate coincided with a cold-dependent increase in leaf CO₂ concentration, indicating that the activities of Calvin cycle enzymes were greatly reduced despite increased stomata opening (Figure 2B).

To determine the global patterns of gene expression in response to cold in sugar beet source and sink tissues, we performed deep sequencing of the transcriptome (RNA-seq) using leaf and taproot tissues for all three genotypes exposed to cold (4°C) or control (20°C) conditions. We collected samples 14 d after transfer from 12 to 4°C, at a time with maximum contrast between metabolic accumulation of sugars (Figure 1) and photosynthetic rate (Figure 2A). We mapped raw RNA-seq reads to the sugar beet reference genome (Dohm et al., 2014) and found that exposure to cold induced a global rearrangement of gene expression in both shoot and taproot tissues (Supplemental Figures 1 and 2).

We extracted transcript levels for genes involved in photosynthesis and observed that all genotypes exhibited the same behavior, as principal component analysis failed to clearly separate genotypes (Supplemental Figure 1A). In samples collected at 20°C, ∼9% of all transcript reads could be assigned to Gene Ontology terms involved in photosynthesis, specifically the subgroups “photosynthesis, light reaction,” “photosynthesis, Calvin cycle,” and “photosynthesis, photorespiration.” After exposure to 4°C, this group was represented by only 3% of all reads, indicating a drastic downregulation of photosynthesis-related genes in the cold (Figure 2C). For example, we observed a strong downregulation for transcript levels of genes encoding Rubisco activase (RCA), Rubisco small subunit (RBCS), a chlorophyll a/b binding protein (CABA), and plastocyanin (PC; Figure 2D). In summary, the data demonstrated that sugar beet photosynthesis was extremely sensitive to chilling temperatures below 12°C. Our results further suggested that the increase in biomass and sugar seen in leaves of cold-treated sugar beet cannot be explained by CO₂ assimilation rates, which are negligible at 4°C (Figures 1 and 2).

Cold Temperatures Alter Major Carbohydrate Metabolism in Shoots and Taproots

We next investigated whether the lower taproot Suc concentrations in the cold might be due to increased respiration. We also asked whether cold temperatures would alter transcript levels of genes encoding enzymes involved in major carbohydrate metabolism (Figure 3). Respiration in taproots reflected the depth at which the taproot was buried in the soil, as it gradually decreased between the upper and lower parts. We observed the same pattern at 4°C, although respiration rates were generally lower than those measured at 20°C (Figure 3B). Taproot carbohydrates therefore deliver raw energy for glycolytic and oxidative catabolism at both 4 and 20°C, but cold exposure resulted in a 50 to 60% drop in respiration relative to that seen at 20°C. By contrast, net respiration increased in source leaves of all genotypes in the cold (Figure 3A), indicating that mature leaves required a high carbohydrate supply from another source, due to a loss in CO₂ assimilation capacity from photosynthesis (Figure 2). One such source is likely starch, which decreased in leaves in the cold (Figure 1G).

Figure 3.

Changes in Major Carbohydrate Metabolism and Energy State in Response to Cold.

(A) Respiration (CO₂ production) from leaf tissue of three sugar beet genotypes under control conditions (20°C; yellow) or 1 week after transfer to 4°C (blue).

(B) Respiration (CO₂ production) of different taproot regions from GT2 under control conditions (20°C; yellow) or 1 week after transfer to 4°C (blue). FW, fresh weight.

(C) ATP, ATP/ADP ratio, and energy charge ([ATP] + 0.5 [ADP]/[ATP] + [ADP] + [AMP]).

(D) Heat-map analysis of grouped expression values extracted from RNA-seq data. Unit variance scaling was applied to rows. Rows use Manhattan distance and average linkage.

(E) Expression values for two sugar beet SPS genes (SPSA1 and SPSA2) and two SUS genes (SUS1 and SUS2) extracted from RNA-seq data.

For (A), (B), (C), and (E), data are means of at least four individual plants ± sd for (A) and (B) or of three pools each consisting of four plants for (C) to (E). Asterisks represent P < 0.05 (Supplemental File 1) using a double-sided t test in comparison with the values at the control condition (20°C).

To determine the cellular energy state of shoot and taproots, we measured adenylate levels (Figure 3C). We found that the ATP level, ATP/ADP ratio, and energy charge ([ATP] + 0.5 [ADP]/[ATP] + [ADP] + [AMP]) rose in shoots of all genotypes. This elevated energization of shoot tissue in the cold may stem from the drastic drop in ATP-consuming CO₂ assimilation (Figure 2B) and the parallel increase of respiration in shoots (Figure 3B). By contrast, energization in taproots did not change in the cold. Although ATP levels also increased, ATP/ADP ratios of GT1 and GT2 taproots indeed remained constant, while they decreased in GT3. In addition, taproot energy charge values diminished slightly in GT2 and GT3 taproots, but not in GT1 taproots, upon transfer to 4°C (Figure 3C).

A principal component analysis using transcript levels of genes assigned to the Gene Ontology term “major CHO metabolism” as input revealed organ- and temperature-dependent differences (Supplemental Figure 1B). As visualized in the heat map shown in Figure 3D, genes encoding enzymes that contribute to starch degradation in shoots were upregulated, while those involved in starch biosynthesis were downregulated, by cold exposure (Figure 3D). Despite the taproots having extremely low starch levels (Figure 1G), starch-related genes were also regulated in taproots in the same manner as in shoots (Figure 3D). This observation is consistent with a report from Turesson et al. (2014) that detected the activity of starch metabolic enzymes in taproots even though the taproots lack starch. Genes that encode Suc biosynthetic enzymes were upregulated in roots in the cold but remained unchanged in leaves. Transcript levels for genes encoding sucrose-degrading enzymes were however clearly downregulated in roots and slightly upregulated in shoots (Figure 3D).

Suc phosphate synthase (SPS) and Suc synthase (SUS) are two key enzymes involved in Suc biosynthesis and degradation; their activity dictates the identities of source and sink tissues. While source tissue identity is associated with high SPS activity, sink tissues usually have high SUS activity (Martin et al., 1993; Kovtun and Daie, 1995; Sturm, 1996; Ruan, 2014). We therefore conducted a genome-wide search in the sugar beet genome (RefBeet 1.2; Dohm et al., 2014) for SPS and SUS genes and identified two SPS and four SUS candidates. Both SPS candidates belong to the SPS A subgroup (Volkert et al., 2014) based on Bayesian analysis and were therefore designated SPSA1 and SPSA2 (Supplemental Figure 3A). In control conditions, the expression level of SPSA1 was ∼10-fold higher in the shoots of all genotypes compared with roots. Following cold exposure, SPSA1 transcript levels rose in roots up to sevenfold, but they remained constant in leaves. Under control conditions, SPSA2 displayed an opposite expression pattern: low in shoots but high in roots of all tested genotypes, in agreement with a previous report (Hesse et al., 1995) that described this gene as being taproot-specific, induced by Glc, and repressed by Suc. SPSA2 expression was either unchanged in GT1 and GT3 shoots or downregulated in GT2 leaves upon cold treatment. Just like SPSA1, however, SPSA2 was induced in taproots of all genotypes.

We also used mass spectrometry to explore the soluble protein landscape of the same taproot tissues used for the transcriptome analysis: SPSA1, but not SPSA2, was slightly more abundant following cold treatment (Supplemental Figure 3B). In agreement, we detected higher SPS activity at 4°C compared with 20°C in taproot protein extracts (Supplemental Figure 3C). Taking these observations together, cold temperatures resulted in higher UDP levels in taproots, higher sucrose-6-phosphate levels, alongside elevated levels of the allosteric SPS activator glucose-6-phosphate (Huber and Huber, 1992). These observations support a scenario whereby elevated SPS activity in taproot shifts this tissue to Suc synthesis (Supplemental Figure 4).

The expression pattern of the four SUS candidates showed tissue- and temperature-specific responses. While the SUS1 and SUS2 genes were strongly expressed in roots and their encoded proteins were highly abundant, SUS3 and SUS4 were scarcely expressed and their corresponding proteins were not detected in our soluble proteome (Figure 3E; Supplemental Figure 5). SUS1 and SUS2 transcript levels were 10-fold (SUS1) to 100-fold (SUS2) higher in taproots relative to shoots, substantiating their role as sink-determining factors. After cold exposure, mRNA levels for SUS1 and SUS2 decreased by ∼50% in taproots. By contrast, SUS2 transcript levels increased 10- to 20-fold in shoots, although never reaching the high levels observed in taproots (Figure 3E). SUS2 levels largely followed SUS2 mRNA levels in taproots, whereas the SUS1 level remained constant, indicating differential protein turnover of these two isoforms in response to cold treatment (Supplemental Figures 5B and 5C).

Taken together, these data indicate that developing taproots respond to cold by shifting from a sucrose-consuming/storing organ to a sucrose-synthesizing tissue, whereas shoots adopt some of the characteristics of a sink tissue.

Cold Temperatures Reverse Phloem Translocation of Suc and Esculin

The data presented above suggested that changes in CO₂ assimilation or starch degradation were unlikely to fully explain cold-induced sugar accumulation in shoots and supported the hypothesis that carbon used as a building block for shoot metabolites might be remobilized from taproot storage cells. To track the fate of taproot-based carbon after exposure to cold temperatures, we injected radiolabeled [¹⁴C]Suc into the fleshy parenchymatic taproot tissue of plants grown either in 20°C control conditions or after cold exposure (5 d at 12°C and then 7 d at 4°C). We kept treated plants for another 1 week in control or cold temperature conditions and then dissected them into individual leaves and taproots. We pressed and dried leaves or longitudinal thin sections of taproots and visualized the incorporated radioactivity using phosphorimaging plates and software (Figure 4; Supplemental Figures 6 and 7).

Figure 4.

Distribution of [¹⁴C]Suc and Esculin in Leaves.

(A) to (D) Autoradiography of [¹⁴C]Suc in leaves.

(A) Schematic depiction of the experiment: a source leaf from a representative plant grown for 1 week at 4°C.

(B) Blackening of veins indicates radioactivity incorporated and distributed into leaf tissue after the injection of radiolabeled Suc into taproots. mv, middle vein; p, petiole; 1°, first order lateral vein; 2°, second order lateral vein.

(C) Source leaf from a representative control plant grown at 20°C.

(D) Radioactivity in cpm was measured in isolated petioles from plants grown at either 4 or 20°C. Center lines show the medians; box limits indicate the 25th and 75th percentiles; whiskers extend 1.5 times the interquartile range from the 25th and 75th percentiles; outliers are represented by dots; crosses represent sample means; n = 16 sample points. The asterisk indicates a significant difference between the 20 and 4°C treatments using a t test (*, P < 0.05; Supplemental File 1). DW, dry weight.

(E) to (K) Esculin loadings. Yellow fluorescence indicates lignified xylem vessels; blue fluorescence indicates esculin trafficking.

(E) Schematic depiction of the experiment.

(F) and (G) Sections through a petiole of a source leaf not loaded with esculin. Cross sections of petioles from 20°C are shown as a bright-field image (F) and a UV light fluorescence image (G).

(H) and (I) Cross sections of petioles from 4°C are shown as a bright-field image (H) and a UV light fluorescence image (I). ph, phloem; xy, xylem.

(J) and (K) Longitudinal sections of a petiole from 4°C are shown as a bright-field image (J) and a UV light fluorescence image (K).

Bars = 1 cm in (B) and (C), 100 µm in (F) to (I), and 100 µm in (J) and (K).

To our surprise, plants grown subjected to cold treatment accumulated radioactivity in leaves, more specifically in leaf veins and with decreasing intensity toward the leaf tip, consistent with phloem transport (Figure 4B). Source leaves of plants grown in control conditions displayed very low radioactivity (Figure 4C), although we did detect radioactivity in the petioles of young sink leaves of the same plants (Figure 4D). This radioactive signal may represent the transport of Suc or derivatives through the xylem, due to the injury of punctured vessels caused by the invasive inoculation procedure. The drastic water loss in shoots at 4°C (Figure 1C), however, argued against Suc transport via the xylem to leaves and instead supported transport via the phloem to leaves that had switched from source to sink organs.

To test this hypothesis, we used a less invasive strategy that mirrors the actual transport of assimilates more realistically (including the prior downward transport). We used esculin, a fluorescent phloem-mobile coumarin glycoside (Knoblauch et al., 2015) that is recognized by several Suc transporters, including the sugar beet phloem loader SUCROSE TRANSPORTER1 (SUT1; Nieberl et al., 2017). We infiltrated esculin into one source leaf and assessed esculin transport routes by fluorescence imaging in thin sections of source leaf petioles from leaves that had not been loaded with esculin either in plants after transfer to cold or in plants maintained in control conditions (Figure 4E). We detected esculin fluorescence only in plants transferred to cold, where we detected it in phloem vascular bundles of source leaves. However, this fluorescence was not confined to the phloem, as we also detected fluorescence in a bundle region interspersed with the yellow fluorescence characteristic of lignified xylem vessels (Figures 4H to 4K). In control plants maintained at 20°C, we failed to detect esculin fluorescence in the phloem (Figures 4F and 4G). Esculin is likely first translocated to the base of the petiole of the loaded leaf and through (at least parts of) the taproot before migrating to other leaves.

To follow Suc flow from the site of inoculation in the taproots, we cut longitudinal thin sections of taproots injected with radiolabeled Suc and exposed the tissue to phosphorimaging plates (Figure 4A; Supplemental Figures 6 and 7). We detected concentrated radioactivity in taproots of plants exposed to 4°C in veiny or spotty structures found between the site of inoculation and the taproot top (crown tissue). Higher magnification allowed the identification of these structures as vascular bundles (Supplemental Figure 6). In taproots from plants grown in control conditions, we observed no such radioactive signal originating from vascular structures, although crown tissue did accumulate some radioactivity, illustrating the transport of radiolabeled Suc upward in the direction of the shoot (Supplemental Figure 7). In most cases, however, radioactivity detected in taproots of plants maintained at 20°C was merely confined to parenchymatic regions near the site of inoculation or concentrated in thick strands that reached from the site of inoculation toward the emergence of lateral roots.

These results indicated that radiolabeled Suc and esculin were preferentially transported from taproots into shoots in the cold but not under control conditions, the same transport route taken from Suc released from parenchymatic storage tissue.

Vacuolar Suc Importer and Exporter Genes and Proteins Show Opposite Cold-Dependent Expression

Next, we analyzed whether the transport of Suc between taproots and shoots in response to cold was mediated by differential activity of vacuolar Suc importers and exporters. In Arabidopsis, TONOPLAST SUGAR TRANSPORTERs (TSTs) and SUCROSE TRANSPORTER4 (SUT4) mediate vacuolar Suc import and export, respectively (Wormit et al., 2006; Schulz et al., 2011; Schneider et al., 2012). Sugar beet TST2;1, the putative homolog of Arabidopsis TST1, is responsible for vacuolar Suc accumulation (Jung et al., 2015). TST2;1 expression in taproots of all tested genotypes greatly exceeded its expression in leaf tissue, supporting its role as the Suc loader of taproot parenchyma vacuoles (Figures 5A and 5B). Interestingly, both mRNA and protein abundance decreased significantly in all genotypes in taproots after cold treatment (Figure 5B; Supplemental Figure 8).

Figure 5.

Cold-Dependent Accumulation of TST2;1 and SUT4 in Three Different Sugar Beet Genotypes.

(A) Illustration of cold-induced processes. The top image shows cold-dependent sugar relocations from taproots to shoots. The middle image shows a schematic of taproot vacuolar transport processes and factors. A vacuolar ATPase (V-H⁺-ATPase) establishes a proton motif force across the vacuolar membrane; TST2;1 acts as a proton/Suc antiporter using the proton-motive force to drive Suc import into vacuoles. SUT4 acts as a proton/Suc symporter using the proton-motive force for vacuolar Suc export. The bottom image shows reciprocal cold-induced regulation of TST2;1 and SUT4 mRNA levels in taproots.

(B) Transcript abundance of TST2;1 based on RNA-seq reads.

(C) Subcellular localization of SUT4-GFP in leaf mesophyll protoplasts of Arabidopsis or sugar beet. Single optical sections are shown in all images. Green color, GFP signal; red color, chlorophyll autofluorescence. Arrowheads point toward the vacuolar membrane (tonoplast). Bars = 5 µm.

(D) Transcript abundance of SUT4 based on RNA-seq reads.

Values in (B) and (D) represent means from n = 3 biological replicates ± se. Asterisks indicate significant differences between the 20 and 4°C treatments using a t test (*, P < 0.05; Supplemental File 1).

Export of Suc from the vacuole is presumably mediated by a SUT4-type transporter. We identified Bv5_124860_zpft.t1 as a putative sugar beet homolog of Arabidopsis SUT4 (AtSUT4) and accordingly termed the corresponding transporter BvSUT4 (Supplemental Figure 9). Transient expression of an N-terminal fusion of the BvSUT4 coding sequence with GFP in sugar beet or Arabidopsis mesophyll protoplasts clearly indicated a tonoplast localization for BvSUT4 (Figure 5C). BvSUT4 transcript levels increased significantly in taproots in the cold, based on our RNA-seq data (Figure 5D). These results demonstrated that vacuolar taproot Suc import decreased, while the release of vacuolar taproot Suc increased, under cold conditions. These observations further support our hypothesis that the opposite regulation of BvTST2;1 and BvSUT4 in taproots is the underlying driving force for the accumulation and delivery of sugars to shoots.

Expression of Floral Regulator Genes Is Adjusted in the Cold

The observed translocation of Suc from taproots back to the shoots that first produced it may represent a metabolic preparation before flowering initiation. We therefore queried the expression of flowering regulator genes in our RNA-seq data set and observed downregulation of the sugar beet floral repressor FT1 and upregulation of the floral activator FT2 in leaves of plants exposed to cold temperatures (Figure 6), in agreement with previous reports (Pin et al., 2012). All genotypes analyzed here have a biennial growth habit and as such carry nonfunctional alleles of BTC1 and BBX19 that do not influence FT1 expression. Nevertheless, BTC1 and BBX19 expression responded to cold treatment in an opposite manner. BTC1 transcript levels decreased in the cold in both roots and shoots, whereas BBX19 transcript levels increased. This observation was in sharp contrast to results from Pin et al. (2012), which showed that vernalized biennial sugar beets had higher BTC1 mRNA levels compared with nonvernalized plants. We hypothesize that the apparent discrepancy with our results stems from expression profiling after vernalization (Pin et al., 2012) versus at early stages of vernalization (this study). In our conditions, we detected BTC1 and BBX19 expression in both shoots and taproots, with BBX19 transcript levels in cold-treated taproots exceeding those in shoots kept at 20°C by almost threefold. However, FT1 and FT2, representing potential targets of functional BCT1 and BBX19 in annual beets, were specifically and exclusively expressed in shoots (Figure 6). In summary, these data showed that the vernalization signal was already transmitted to modulate the expression level of floral regulator genes and that transcriptional changes of related genes did occur in both shoots and taproots.

Figure 6.

Expression of Floral Regulator Genes.

Transcript abundances of sugar beet BBX19 (Bv9_216430_rwmw.t1), BTC1 (Bv2_045920_gycn.t1), FT1 (Bv9_214250_miuf.t1), and FT2 (Bv4_074700_eewx.t1) based on RNA-seq reads in shoots and taproots of three different genotypes. Values represent means from n = 3 biological replicates ± se. Asterisks indicate P < 0.05 using a double-sided t test (Supplemental File 1).

DISCUSSION

In this work, we discovered a previously unknown switch of sink and source identities of taproots and shoots upon cold exposure in sugar beet plants. In contrast to sinks like seeds, culms, or tubers, which adopt source identities after complete differentiation and subsequent separation from their nourishing source, the sink-source switch in sugar beet occurred in response to an environmental cue when both shoot and taproot tissues were still physiologically connected. The findings discussed below are schematically explained in the model we present in Figure 7.

Figure 7.

Schematic Illustration of Cold-Induced Sink-to-Source Transition.

Leaf and taproot tissues of sugar beet are reprogrammed and source and sink identities are shifted upon cold. Shoots adopt sink identity during cold treatment. Biomass and sugar concentration in the shoot increase (A) despite reduced photosynthetic activity and inactivation of carbon assimilation (B). Concomitantly, shoot respiration increases (C) and cellular starch pools decrease (D). By contrast, taproots show a decrease of Suc levels (E) but lower respiration rate (F) as well as increased Suc biosynthesis (G). Taproot sugar is remobilized in the cold due to opposite regulation of the activities of taproot-specific vacuolar Suc importer (TST2;1) and exporter (SUT4; [H]). Together, this results in a reversal of the phloem translocation stream (I) triggered by a reprogramming of source and sink identities, which might correlate with inflorescence initiation.CC, companion cell; PS, photosynthesis; RuBP, ribulose-1,5-bisphosphate SE: sieve element.

At 4°C, shoot CO₂ assimilation decreased drastically, while PSII activity stayed relatively high (Figure 2). This lack of correlation indicates that enzymes of the Calvin-Benson cycle slowed down in the cold and could not utilize electrons liberated from the photosynthetic electron transport (PET) chain. With lower CO₂ fixation rates during cold conditions, high PET rates may be detrimental because they produce harmful reactive oxygen species such as superoxide, hydrogen peroxide, or hydroxyl anions (Suzuki and Mittler, 2006; Choudhury et al., 2017; Pommerrenig et al., 2018). During a cold exposure time course (Figures 1 and 2), we recorded decreased CO₂ assimilation starting at 12°C, during the acclimation phase. As indicated by the increased Y(NPQ) percentage (Figure 2A), either the Mehler ascorbate pathway (Asada, 1999) or the xanthophyll cycle (Farber et al., 1997) may act as quenchers for PET-released electrons at this temperature. At 4°C, these scavenging pathways appear to slow down, as evidenced by the further decrease in Y(II) and Y(NPQ) as well as the concomitant increase in Y(NO) indicative of nonregulated energy dissipation, which can severely damage chloroplast membranes and plant cells. Under those sustained challenging conditions, the cold response reprograms gene expression, leading to effective downregulation of photosynthesis-related gene expression (Figures 2C and 2D).

These data are in agreement with results from Arabidopsis, where photosynthesis and expression of the AtRBCS and AtCAB genes were reduced after shifting plants first grown at 23°C to 5°C, although photosynthesis rates recovered after prolonged exposure to cold (Strand et al., 1997). In summary, these metabolic and transcriptomic changes eventually result in a drastic decrease of CO₂ incorporation into sugars, which are required for growth and protection of cell vitality in the cold.

Despite inactivation of photosynthesis, however, sugars continued to accumulate in leaves and decreased in taproots in the cold (Figures 1D to 1F). Lower Suc levels and impaired respiration in taproots revealed that Suc was not being used for energy metabolism during cold at the same rate as under control conditions (Figure 3B). Cold-tolerant plants like Arabidopsis accumulate sugars in leaves in response to cold by maintaining photosynthetic activity, reducing Suc phloem loading, and increasing sugar import into leaf vacuoles (Wingenter et al., 2010; Nägele and Heyer, 2013). Although Arabidopsis can overcome sugar repression of photosynthesis after prolonged exposure to cold (Huner et al., 1993; Strand et al., 1997), such a mechanism does not occur in sugar beet. Instead, the drastic drop in carbon assimilation and eventually photosynthesis in the shoot turns leaves into sink organs that are now supplied with sugar from taproots (Figure 4).

Under nonchilling temperatures, the reversibility of taproot sinks and the remobilization of sugars from storage vacuoles may become essential, either when leaves cope with wounding from insect feeding or after a new powerful sink like the inflorescence forms after winter. However, a short exposure to cold temperature is sufficient to trigger the remobilization of carbohydrates from taproot storage sites, as indicated by the movement of radiolabeled Suc and fluorescent esculin toward the shoot (Figure 4). While Suc biosynthesis and hydrolysis were reciprocally regulated under warm and cold conditions (Figures 3D and 3E), levels of taproot Suc decreased upon cold treatment (Figure 1F). In agreement with this process, we observed opposite regulation patterns for the major vacuolar Suc importer (BvTST2;1) and putative major exporter (BvSUT4) in the same tissue (Figure 5). BvTST2;1 expression and BvTST2;1 abundance were downregulated in response to cold exposure, while BvSUT4 expression followed an opposite trend. The role of BvSUT4 as a Suc exporter is supported by its homology to Suc transporters of the SUT family and by its homology to AtSUT4 (Supplemental Figure 9), for which both Suc export activity and vacuolar localization have been demonstrated (Schulz et al., 2011; Schneider et al., 2012). It is unlikely that sugars are released from vacuoles in the cold as monosaccharides via other transporters, such as the INTEGRAL MEMBRANE PROTEIN transporter (Klemens et al., 2014). This is because vacuolar invertase activity––a prerequisite for vacuolar monosaccharide generation and thus export––is extremely low at this developmental stage (Giaquinta, 1979; Godt and Roitsch, 2006).

It is surprising that the flux transition took place prior to bolting, that is, before the formation of an inflorescence that would then act as a new strong sink organ utilizing remobilized taproot sugars as building blocks. During the early stage of natural vernalization, warmer temperatures and relatively longer daylength do not yet signal the onset of spring (Mutasa-Göttgens et al., 2010; Ritz et al., 2010). However, at the same time as the switch in sink/source identities, cold exposure was accompanied by an adjustment of expression levels in floral regulator genes. The floral repressor BvFT1 and the activator BvFT2 (Pin et al., 2010) showed opposite patterns in response to cold in shoots (Figure 6).

The expression of the flowering-related genes BvBTC1 and BvBBX19 in taproots suggested that underground plant tissue may also perceive the vernalization signal. We cannot exclude the possibility that the combined loss of carbon assimilation in the shoot and the concomitant demand for sugars required for cold acclimation might create a sink that then sends a signal triggering taproot Suc efflux independently of bolting. It is, nevertheless, tempting to speculate that sugar beets may employ factor(s) that integrate bolting and the transition between sink and source, similar to the recently described phloem-mobile FT homolog StSPS6A (tuberigen) in potato (Solanum tuberosum; Navarro et al., 2011; Abelenda et al., 2019). This factor or factors may have already been identified (Pfeiffer et al., 2014; Broccanello et al., 2015; Tränkner et al., 2017) or may remain undiscovered.

Our study represents a comprehensive analysis of sugar beet taproot tissue during cold treatment and shows that cold temperatures induce a sink-to-source transition that establishes the accumulation of taproot-derived carbohydrates in the shoot. For this to occur, sugars have to be loaded into taproot phloem, transported from taproots to shoots, and unloaded in leaf tissue. It is currently unknown whether taproot phloem loading in the cold involves an apoplastic step, whether the same phloem vessels are being used for root- and shoot-bound sugar trafficking, and how sugar unloading is initiated in former source leaves. This latter issue possibly involves a reprogramming of transporter activity that mediates sugar efflux from the vasculature to the mesophyll and may involve both passive and active transport processes. Our transcriptomic and proteomic data sets may reveal candidate factors and transporters involved in this unloading in the cold in the future.

Our findings also have implications for agriculture and breeding, where attempts have been made to grow sugar beet during any season (Hoffmann and Kluge-Severin, 2011; Hoffmann and Kenter, 2018), a scenario that will be facilitated by the generation and deployment of bolting-resistant hybrid genotypes (Pin et al., 2010; Pfeiffer et al., 2014; Tränkner et al., 2016). Biennial growth of sugar beet might also indirectly benefit from the warmer winters precipitated by climate change in central and northern Europe (Lavalle et al., 2009), and this would expand the cultivation of sugar beet under nonfreezing, nonlethal low temperatures. However, even under nonfreezing but prolonged above-zero chilling conditions, the advantages of a longer vegetative period would be negated, at least to some extent, by the tradeoff of cold-induced taproot sugar loss described in our study. This phenomenon might also partially account for the observed reduced yield and higher marc-to-sugar ratio of fall- or early spring-sown sugar beet plants (Hoffmann and Kluge-Severin, 2011).

In the future, it will be highly valuable to analyze the sink-source transition of sugar beet taproots in bolting-resistant mutants and varieties without the activating function of FT2 (Pin et al., 2010) to reveal whether FT activity is required to trigger this transition. Equally relevant will be the generation of sut4 mutant plants of sugar beet to study the effects of the loss of vacuolar Suc efflux for floral induction and cold tolerance. Such modified plants might exhibit a diminished taproot Suc release and therefore a reduced supply of building blocks for inflorescence formation. This potential impact on bolting makes sugar beet SUT4 a highly attractive target for breeding approaches (Chiurugwi et al., 2013; Pfeiffer et al., 2014) aimed at bolting resistance and at decreasing cold-induced Suc loss from taproots.

METHODS

Plant Material and Growth Conditions

Three hybrid sugar beet (Beta vulgaris) genotypes (GT1, GT2, and GT3; Kleinwanzlebener Saatzucht) were used for this study. Plants were germinated and grown on standard soil substrate ED73 (Einheitserdwerke Patzer)/10% (v/v) sand mixture under short-day conditions (10 h of light/14 h of dark) at 60% relative humidity and 110 µmol m⁻² s⁻¹ light intensity (fluorescent tube light). For growth and sugar accumulation kinetics, plants were grown at 20°C for 6 weeks, transferred to 12°C for 1 week, and then placed at 4°C for 3 weeks. Six to 10 plants per genotype (biological replicates) were harvested 4 h after dawn at 20°C and at 5, 9, and 19 d after transfer to 4°C. Shoots were severed, and taproots were dug out of the soil and washed with tap water. For dry weight determinations, individual shoots and chopped taproots were wrapped in aluminum foil, freeze-dried for 1 week, and weighed. Water content was calculated as g water/g dry weight using the difference between fresh weight and dry weight. Dry matter was used for the quantification of sugars and starch.

For RNA-seq, proteome, and metabolite analyses, plants were grown at 20°C for 10 weeks, transferred to 12°C for 1 week, and then placed at 4°C for 2 weeks. Control plants were kept at 20°C. Plants were grown under short-day conditions (10 h of light/14 h of dark) and 110 µmol m⁻² s⁻¹ light intensity (sodium vapor lamps). For harvest, plants were dissected into shoot and taproot tissues. At least three pools of four different plants were made for each tissue and genotype. Each pool was considered as a biological replicate. Tissues were chopped with a kitchen knife, transferred to liquid nitrogen, and kept at –80°C until further processing. An independent analysis of transcriptome and metabolites using 15-week old plants transferred to 12°C for 1 week and placed at 4°C for 2 weeks was performed in the same manner as with 10-week-old plants and reproduced the results of the first experiment.

Chlorophyll Fluorescence Measurements

The photosynthetic activity of four individual plants per genotype was measured in a temporal kinetics experiment using an Imaging-PAM M-Series System (Heinz Walz). Plants were placed in the dark for 12 min to deplete the energy of PSII. The capacity of PSII was measured by saturation with 14 cycles of photosynthetically active radiation (PAR; 76 µmol photons m⁻² s⁻¹) light pulses at 0, 50, and 70 s. Recorded fluorescence was used for calculation of the effective quantum yield of PSII [Y(II) = (Fm' − F)/Fm'], of nonphotochemical quenching [Y(NPQ) = 1 − Y(II) − 1/(NPQ + 1 + qL(Fm/Fo − 1))], and of nonregulated energy dissipation [Y(NO) = 1/(NPQ + 1 + qL(Fm/Fo − 1))]. Required factors were calculated by the formulas NPQ = (Fm − Fm')/Fm', qN = (Fm − Fm')/(Fm − Fo'), Fo' = Fo/(Fv/Fm + Fo/Fm'), qP = (Fm' − F)/(Fm' − Fo'), and qL = (Fm' − F)/(Fm' − Fo') × Fo'/F = qP × Fo'/F.

Gas-Exchange Measurements

A GFS-3000 system (Heinz Walz) was used to analyze gas exchange-related parameters of four individual plants per genotype in a temporal kinetics experiment. A 2.5-cm³ gas-exchange cuvette was used to measure the CO₂ assimilation rate, respiration, leaf CO₂ concentration, and transpiration of sugar beet source leaves. Leaf regions including large central midribs were omitted. The conditions inside the cuvette were set to the same temperature, humidity, and CO₂ concentration in which the plants had been grown. The intervals were determined by a trial experiment, in which the time necessary for stabilization of the flow of CO₂ after transfer of the leaf section into the cuvette and adaption to the changed light intensities was measured. The measurements started after stabilization of the CO₂ flow, which required ∼5 min. The light intensity was set to PAR 125 for 460 s, and three consecutive CO₂/water-exchange measurements each lasting 30 s were performed. Afterwards, light intensity was again set to PAR 0 for 320 s, and three consecutive CO₂/water-exchange measurements each lasting 30 s were performed. The 30-s interval between the measurements was necessary for the leaf to return to the stabilized value.

Respiration of Sugar Beet Taproot Tissue

Respiration of taproots was measured by cutting out 0.5-cm³ tissue cubes from central taproot regions and measuring CO₂ production in a whole-plant cuvette with a volume of 60 cm³. Values were normalized to tissue weight.

RNA Extraction and Sequencing

RNA was isolated from three biological replicates per genotype, tissue (leaf and root), and treatment. About 100 mg of frozen plant material was pulverized in a tissue lyser (Qiagen) at 30 Hz for 90 s After grinding, samples were again transferred to liquid N₂, supplemented with 1.5 mL of QIAzol Lysis reagent (Qiagen), vortexed three times for 30 s, and centrifuged at 4°C for 10 min at 12,000g. Supernatants were transferred to fresh tubes, incubated at room temperature (RT) for 5 min, extracted with 300 μL of chloroform, vortexed for 15 s, and centrifuged at 4°C for 15 min at 12,000g. Aqueous supernatants were transferred to fresh tubes, and RNA was precipitated with 750 μL of isopropanol for 10 min at RT and spun down at 4°C for 10 min at 12,000g. Precipitates were washed with 75% (v/v) ethanol, and the RNA pellets were dried at 37°C for 5 to 10 min prior to resuspension in 100 μL of diethyl pyrocarbonate-water by gentle shaking at 37°C for 5 to 10 min. To remove residual contaminants, RNA was further purified using the RNeasy kit (Qiagen). The following steps were performed at RT. Per 100 μL of RNA suspension, 350 μL of RLT buffer (provided with the kit) was added and vortexed briefly. Then, 250 μL of ethanol was added, and the mixture was vortexed again. The RNA was purified using a spin column (provided with the RNeasy kit) by centrifugation for 30 s at 15,000g, followed by washing three times with 500 μL of RPE buffer (provided with the kit) and centrifugation for 30 s at 15,000g, and finally eluted from the column for a final volume of 50 μL (in diethyl pyrocarbonate-water) per sample. The RNA was quantified (NanoDrop 2000/2000c, Thermo Fisher Scientific) for each sample prior to further processing or storage at –80°C. RNA quality was confirmed using an Agilent Technologies 2100 Bioanalyzer. RNAs (2 µg per sample) were transcribed to cDNAs and sequenced using the Illumina HiSeq 2000 system. Sequencing and assembly were provided as a custom service (GATC). The statistical analysis included data normalization, graphical exploration of raw and normalized data, test for differential expression for each feature between the conditions, and raw P value adjustment. The analysis was performed using the R software (R Core Team, 2017), Bioconductor (Gentleman et al., 2004) packages including DESeq2 (Anders and Huber, 2010; Love et al., 2014), and the SARTools package developed at PF2-Institute Pasteur.

Phylogenetic Analysis

Multiple sequence alignments of amino acid sequences for SPS, SUS, and SUT4 proteins were performed using Clustal Omega (Sievers et al., 2011). Sequence alignment files are available as Supplemental Files 2 to 4. Bayesian phylogenetic analysis was performed with MrBayes version 3.2 (Ronquist et al., 2012). MrBayes always selected the best-fit models Jones (Jones et al., 1992) and WAG (Whelan and Goldman, 2001) for amino acid substitution analysis of SPS proteins and SUS proteins, respectively. MrBayes conducted two parallel Metropolis coupled Monte Carlo Markov chain analyses with four chains for 300,000 (SUS and SPS) or 20,000 (SUT4) generations. Trees were sampled every 1000 generations. The analyses were run until the sd of split frequencies was below 0.005. Consensus trees were computed after burn-in of the first 25% of trees and visualized using FigTree version 1.4.3.

Principal Component Analysis and Heat Map Analysis

For RNA-seq data, the mean cpm values were used for the analysis. Data were visualized using ClustVis (Metsalu and Vilo, 2015).

Analysis of Soluble Sugars and Starch

Leaves and taproots were harvested separately, frozen in liquid nitrogen, freeze-dried, and stored at –80°C until use. Pulverized material was extracted twice with 1 mL of 80% (v/v) ethanol at 80°C for 1 h. Combined extracts were evaporated in a vacufuge concentrator (Eppendorf), and pellets were dissolved in distilled, deionized water. For starch isolation, pellets were washed with 80% (v/v) ethanol and 1 mL of distilled, deionized water. A total of 200 μL of water was added to the pellet, and the sample was autoclaved for 40 min at 121°C. Then, 200 μL of enzyme mix (5 units of α-amylase and 5 units of amyloglucosidase in 200 mM sodium acetate, pH 4.8) was added to the pellet, and starch was hydrolytically cleaved into Glc units at 37°C for 4 h. The enzymatic digestion was stopped by heating the samples to 95°C for 10 min. After centrifugation (20,000g, 10 min, 21°C), the supernatant could be used for starch quantification. Extracted sugars and hydrolytically cleaved starch were quantified using a NAD⁺-coupled enzymatic assay (Stitt et al., 1989).

Analysis of Phosphorylated Metabolites

The contents of phosphorylated intermediates (glucose-6-phosphate, fructose-6-phosphate, sucrose-6-phosphate, UDP-glucose, and UDP) were determined according to Horst et al. (2010).

Radiolabeled Suc Translocation Assay

Ten- to 12-week-old sugar beet plants grown at 20°C under short-day conditions (10 h of light/14 h of darkness) were used for the analyses. Plants for cold treatment were grown for 1 more week at 12°C and then kept for 6 to 7 d at 4°C. Taproots from 4 and 20°C plants were partially uncovered from surrounding soil substrate, and a 1-mm hole was punched with a biopsy puncher into the upper half of the taproot (∼1 cm below the soil surface). The created pit was filled with 10 μL of 50% (v/v) radiolabeled Suc (536 mCi/mmol; Hartmann Analytic) and coated with a drop of petroleum jelly. Plants were then kept for another 10 d at 4 or 20°C (control). At the end of the treatment, all source leaves of injected plants were detached and individually pressed between blotting paper. For the detection of radioactivity in taproots, taproots were dug out, washed, cut in thin slices (∼0.5 mm thick) with a truffle slicer, and pressed between blotting paper. Radioactivity was recorded with phosphorimage plates (exposed for 4 to 5 h either to the adaxial surface of pressed and dried leaves or to dried taproot slices), and plates were analyzed with a Cyclone Storage Phosphor Screen (Packard Bioscience). For the quantification of radioactivity in petioles, source leaf petioles from the same leaves used for phosphorimaging were cut off, ground, and pulverized. Next, 5 to 10 mg of powder was mixed with 2 mL of scintillation cocktail, and cpm was recorded with a TRI-Carb 2810TR liquid scintillation analyzer (Perkin Elmer).

In Planta Esculin Transport

Ten-week-old sugar beet plants grown at 20°C under short-day conditions (10 h of light/14 h of darkness) were used for the analysis. One source leaf per plant (usually from leaf stage 10 to 12) was abraded at the adaxial side with fine sandpaper (grade 800). About 500 μL of a 100 mM esculin sesquihydrate (Carl Roth) solution was distributed over the injured leaf surface with a plastic pipette. Treated leaves were coated with plastic foil, kept for another 2 d at 20°C, and then transferred to 4°C or kept at 20°C (control). After 5 to 7 d in the cold, source leaves, which were not loaded with esculin, were detached and sections of petioles were analyzed for esculin fluorescence with a TCS SP5II confocal microscope (Leica) using an HCX PL APO lambda blue 20.0x0.70 IMM UV light objective. Slices of taproots from the same plants were analyzed for esculin fluorescence to ensure that esculin was successfully translocated into taproots in both cold-treated and control plants. The emission bandwidths were 440 to 465 nm for the detection of esculin fluorescence and 594 to 631 nm for lignin fluorescence.

Soluble Protein Extraction

Plants were harvested, washed, and separated in the cold into taproots and source leaves. Frozen leaf tissue was pulverized with liquid nitrogen using a Retsch mill (Retsch). Then, 800 μL of buffer E1 (50 mM HEPES-KOH, pH 7.5, 10 mM MgCl₂, 1 mM EDTA, pH 7.5, 2 mM DTT, 1 mM PMSF, 1 mM Pefabloc [Carl Roth], 5 mM aminohexanoic acid, 0.1% [v/v] Triton X-100, and 10% [v/v] glycerol) and 100 mg of pulverized tissue were transferred into 1.5-mL microfuge tubes. Samples were vortexed and centrifuged for 3 min at 12,000g at 4°C. A total of 500 μL of the supernatant was loaded onto a Sephadex NAP5 (G25) column (GE Healthcare) preequilibrated with buffer E1 without Triton X-100. Eluents were collected in precooled microfuge tubes and stored at –20°C. Taproot tissues were treated as above with the following alterations. Taproots were blended with buffer E1 at 4°C until a homogenous pulp was obtained. The pulp was filtered through a kitchen sieve and centrifuged. Five milliliters of the supernatant was dialyzed through a membrane with a 12-kD pore size for 48 h against 2 liters of distilled, deionized water. Water was exchanged seven to eight times. Samples were collected in precooled microfuge tubes and used for enzymatic tests or stored at –20°C. Liquid chromatography and tandem mass spectrometry were performed as described (Jung et al., 2015).

Isolation of Taproot Vacuoles and Vacuolar Proteins

Vacuoles were isolated as described (Jung et al., 2015) with the following modifications. Taproots were cut into thin slices (∼0.5 mm thickness) with a truffle slicer. Slices were cut into small cubes with a razor blade. Taproot cubes were then transferred to 130 mL of collection buffer (750 mM mannitol, 5 mM EDTA, pH 8, 50 mM Tris-HCl, pH 7.6, and 1 mM DTT) and incubated on ice for 15 min with slight agitation. The solution was filtered through a kitchen sieve and a stainless steel sieve (125 µm mesh size). Vacuoles were centrifuged (2000g, 10 min, 4°C), resuspended in 40 mL of centrifugation buffer (collection buffer + 30% [w/v] Nycodenz [AxisShield]), and centrifuged in a Sorvall SS-34 fixed-angle rotor (1000g, 15 min, 4°C). Intact vacuoles floating on the upper phase of the self-forming Nycodenz gradient were aliquoted in 1-mL fractions containing 1 μL of Pefabloc proteinase inhibitor. Vacuolar proteins were precipitated with 20% trichloroacetic acid in a 1:1 (v/v) ratio and were incubated at −20°C for 1 h. Samples were centrifuged (20,000g, 10 min, 4°C) and washed twice with 100% ethanol and 100% acetone. The protein pellet was resuspended in 8 M urea und used for mass spectrometry analysis. Liquid chromatography and tandem mass spectrometry was performed as described (Jung et al., 2015).

SPS Assay

For this assay, 80 µg of soluble protein was added to 200 μL of freshly prepared E_max (50 mM HEPES-KOH, pH 7.5, 20 mM KCl, 4 mM MgCl₂, 12 mM UDP-Glc, and 10 mM fructose-6-phosphate to glucose-6-phosphate [1 to 4]), E_lim (50 mM HEPES-KOH, pH 7.5, 20 mM KCl, 4 mM MgCl₂, 4 mM UDP-Glc, 2 mM fructose-6-phosphate to glucose-6-phosphate [1 to 4], and 5 mM KH₂PO₄), and E_blank (E_max without UDP-glucose and sugar phosphates). Samples were incubated for 20 min at 25°C, followed by 5 min at 95°C to stop the reaction, and centrifuged at 12,000g at 4°C for 5 min. Next, 100 μL of the supernatant was pipetted into 100 μL of 5 M KOH and incubated for 10 min at 95°C. The solution was mixed with 800 μL of anthrone (14.6 M H₂SO₄ and 0.14% [w/v] anthrone), and absorbance was immediately measured at 620 nm. A calibration standard was made with 0 to 5 mmol of Suc.

Subcellular Localization of BvSUT4 in Arabidopsis and Sugar Beet Mesophyll Protoplasts

The sugar beet SUT4 coding sequence (Bv5_124860_zpft.t1 = BVRB_5g124860) was amplified from sugar beet leaf RNA with the gene-specific primers BvSUT4-CACC-f (5′-CAC​CAT​GAC​AGG​CCA​GGA​CCA​AAA​TA-3′) and BvSUT4-rev (5′-TAC​ATG​CAT​CAC​ATG​AAC​TCT​GG-3′). The resulting open reading frame was cloned into pENTR/D-TOPO (Life Technologies), sequenced, and recombined into the Gateway-compatible destination vector pK7FWG,0 (Karimi et al., 2002) to obtain a 35S_pro:SUT4-GFP fusion. Transient transformation of Arabidopsis (Arabidopsis thaliana) mesophyll protoplasts was performed as described (Abel and Theologis, 1994). Isolation and transient transformation of sugar beet mesophyll protoplasts were performed as described (Nieberl et al., 2017). The transient transformation assays were repeated at least three times with Arabidopsis and two times with sugar beet protoplasts. The results from the different transformation rounds were comparable, with ∼70% of transformed Arabidopsis and 80% of transformed sugar beet protoplasts displaying tonoplast localization of BvSUT4-GFP.

Statistical Analysis

Mean values of biomass and sugars for each genotype and time point in Figure 1 and radioactivity in petioles (Figure 4F) were tested for normal distribution using the Shapiro-Wilk normality test (Shapiro and Wilk, 1965). Normal data distribution was assumed when the null hypothesis was not rejected with a threshold of P = 0.05. Normally distributed data were subjected to one-way ANOVA with posthoc Tukey’s HSD test using the One-way ANOVA with post-hoc Tukey HSD Test Calculator from Navendu Vasavada (https://astatsa.com/OneWay_Anova_with_TukeyHSD). Differences were considered as significant when Tukey’s HSD P value was below 0.05. Because of sample size (n < 5), no normality test was performed for PAM and gas-exchange data in Figure 2, gene expression in Figures 2, 3, 5, and 6, and metabolite data in Figure 3 and Supplemental Figure 4. Significance for these data was calculated using a double-sided Student's t test as provided in Microsoft Excel. Data were considered as significantly different when P values were below 0.05. Results of statistical analyses are provided in Supplemental File 1.

Accession Numbers

Transcriptome sequencing data have been deposited in the GenBank Sequence Read Archive (BioProject PRJNA602804). The following accession numbers are found in TAIR, RefBeet 1.2, and GenBank for the Arabidopsis (At), sugar beet (Bv), barley (Hordeum vulgare; Hv), Saccharum officinarum (Sof), and potato (Solanum tuberosum; St) genes analyzed in this work: AtSPSA1 (AT5G20280), AtSPSA2 (AT5G11110), AtSPSB (AT1G04920), AtSPSC (AT4G10120), AtSUC1 (AT1G71880), AtSUC2 (AT1G22710), AtSUC3 (AT2G02860), AtSUC4 (AT1G09960), AtSUC5 (AT1G71890), AtSUC9 (AT5G06170), AtSUS1 (AT5G20830), AtSUS2 (AT5G49190), AtSUS3 (AT4G02280), AtSUS4 (AT3G43190), AtSUS5 (AT5G37180), AtSUS6 (AT4G02280), AtSUT4 (AT1G09960), AtTST1 (AT1G20840), BvBBX19 (Bv9_216430_rwmw.t1), BvBTC1 (Bv2_045920_gycn.t1), BvCAB (Bv_002570_dmif.t1), BvFT1 (Bv9_214250_miuf.t1), BvFT2 (Bv4_074700_eewx.t1), BvPC (Bv_004160_hgjn.t1), BvRBCS (Bv2_026840_jycs.t1), BvRCA (Bv2_025300_tzou.t1), BvSPSA1 (Bv2_030670_mgoq.t1), BvSPSA2 (Bv8_193450_doak.t1), BvSUS1 (Bv8_190960_nnjy.t1), BvSUS2 (Bv7_163460_jmqz.t1), BvSUS3 (Bv7_173620_ffuo.t1), BvSUS4 (Bv4_084720_myet.t1), BvSUT1 (Bv1_000710_gzum.t1), BvSUT3 (Bv6_154300_yemu.t1), BvSUT4 (Bv5_124860_zpft.t1), Bv-SUT4 (Bv5_124860_zpft.t1), BvTST2;1 (Bv5_115690_zuju.t1), HvSPS1 (AF261107.1), SofSPS1F (AB001338.1), StSUS1 (NP_001275237), and StSUS2 (NP_001274911).

Supplemental Data

• Supplemental Figure 1. Principal component analysis (PCA) of PC1 versus PC2 for three genotypes.

• Supplemental Figure 2. Venn diagrams of differentially expressed genes (DEGs) in leaves and taproots.

• Supplemental Figure 3. Phylogeny of Beta vulgaris SPS isoforms and protein abundance of Bv-SPS isoforms in taproots.

• Supplemental Figure 4. Phosphorylated metabolites in shoots and taproots of sugar beet plants.

• Supplemental Figure 5. Phylogeny of Beta vulgaris SUS isoforms and protein abundance of Bv-SUS isoforms in taproots.

• Supplemental Figure 6. Radioactivity incorporated and distributed in taproot tissue in the cold.

• Supplemental Figure 7. Radioactivity incorporated and distributed in taproot tissue.

• Supplemental Figure 8. TST2;1 protein levels.

• Supplemental Figure 9. Phylogeny, sequence and predicted 2D-protein structure of Bv-SUT4.

• Supplemental File 1. Statistical analysis tables.

• Supplemental File 2. Sequence alignment used to produce the phylogenetic tree in Supplemental Figure 3.

• Supplemental File 3. Sequence alignment used to produce the phylogenetic tree in Supplemental Figure 5.

• Supplemental File 4. Sequence alignment used to produce the phylogenetic tree in Supplemental Figure 9.

DIVE Curated Terms

The following phenotypic, genotypic, and functional terms are of significance to the work described in this paper:

• FT1 Gramene: AT2G03220

• FT1 Araport: AT2G03220

• SUS4 Gramene: AT3G43190

• SUS4 Araport: AT3G43190

• SUS1 Gramene: AT5G20830

• SUS1 Araport: AT5G20830

• SUS3 Gramene: Csa1G59780

• SUS3 Araport: Csa1G59780

• SUT4 Gramene: XP_010695440.1

• SUT4 Araport: XP_010695440.1

• SUS2 Gramene: GRMZM2G318780

• SUS2 Araport: GRMZM2G318780

• BBX19 Gramene: At4g38960

• BBX19 Araport: At4g38960

• TST2;1 Gramene: XP_010678632.1

• TST2;1 Araport: XP_010678632.1