tie-dyed1 Functions Non-Cell-Autonomously to Control Carbohydrate Accumulation in Maize Leaves 1

The tie-dyed1 (tdy1) mutant of maize (Zea mays) produces chlorotic, anthocyanin accumulating regions in leaves due to the hyperaccumulation of carbohydrates. Based on the nonclonal pattern, we propose that the accumulation of sucrose or another sugar induces the tdy1 phenotype. The boundaries of regions expressing the tdy1 phenotype frequently occur at lateral veins. This suggests that lateral veins act to limit the expansion of tdy1 phenotypic regions by transporting sucrose out of the tissue. Double mutant studies between tdy1 and chloroplast impaired mutants demonstrate that functional chloroplasts are needed to generate the sucrose that induces the tdy1 phenotype. However, we also found that albino cells can express the tdy1 phenotype and overaccumulate sucrose imported from neighboring green tissues. To characterize the site and mode of action of Tdy1 , we performed a clonal mosaic analysis. In the transverse dimension, we localized the function of Tdy1 to the innermost leaf layer. Additionally, we determined that if this layer lacks Tdy1 , sucrose can accumulate, move into adjacent genetically wild type layers and induce tdy1 phenotypic expression. In the lateral dimension, we observed that a tdy1 phenotypic region did not reach the mosaic sector boundary, suggesting that wild type Tdy1 acts non-cell-autonomously and exerts a short-range compensatory effect on neighboring mutant tissue. A model proposing that Tdy1 functions in the vasculature to sense high concentrations of sugar, up-regulate sucrose transport into veins and promote tissue differentiation and function is discussed. 1998), phenotypes resembling a tdy1 region. Future studies will investigate potential associations between TDY1 and maize SUTs in controlling carbohydrate accumulation in leaves. Understanding how Tdy1 regulates carbon partitioning may have applicability towards engineering biomass feedstocks for biofuels production.


INTRODUCTION
As well as being the primary products of carbon assimilation, sugars have been shown to act as signals that regulate plant growth and development, influencing embryogenesis, seed germination, seedling growth, organ initiation, flowering and senescence (Koch, 2004;Gibson, 2005;Raines and Paul, 2006;Rolland et al., 2006). In addition to their role in regulating plant growth, sugars play important roles in controlling gene expression. Numerous studies in a wide variety of plants have found evidence for sugars both inducing and repressing gene expression (Koch, 1996;Chiou and Bush, 1998;Aoki et al., 1999;Fujiki et al., 2001;Lloyd and Zakhleniuk, 2004;Blasing et al., 2005;Crevillen et al., 2005). Perhaps the best known examples of this are high sugar levels repressing photosynthetic gene expression (Sheen, 1990;Jang and Sheen, 1994;Sheen, 1994;Moore et al., 2003). This can be seen in cases where transport of sugars out of photosynthetic tissues is impaired, and the excess carbohydrates down-regulate photosynthesis and photosynthetic gene expression, thereby resulting in chlorosis (Goldschmidt and Huber, 1992;Riesmeier et al., 1994;Krapp and Stitt, 1995;Burkle et al., 1998;Gottwald et al., 2000;Jeannette et al., 2000).
Maize (Zea mays) leaf blades use C 4 carbon assimilation to synthesize carbohydrates (Langdale and Nelson, 1991). The photosynthetic cells display Kranz anatomy, with mesophyll cells surrounding a ring of bundle sheath cells, which in turn surround the vein (Esau, 1977). This arrangement reflects functional differences in carbohydrate synthesis, storage and transport. For instance, sucrose (Suc) is synthesized in the cytoplasm of mesophyll cells (Lunn and Furbank, 1999) while starch transiently accumulates only in bundle sheath cells (Rhoades and Carvalho, 1944). For transport out of the photosynthetic tissues, Suc is loaded into phloem cells in the minor veins by Suc transporters (Lalonde et al., 2004). Long distance transport of assimilates out of leaves to other plant tissues is principally accomplished by lateral veins (Fritz et al., 1989;van Bel, 2003).
Except near the leaf margins, clonal cell lineages within maize leaves are arranged in parallel 6 photosynthetic pigments (Fig. 1A). We utilized the stereotypical division pattern in maize leaves to identify mutants that develop chlorotic regions that violate clonal lineages. The formation of these regions can not be explained by clonal inheritance, and suggests that a mobile signal is responsible for the phenotype. We recently characterized the recessive tie-dyed1 (tdy1) mutant which develops nonclonal, chlorotic regions in leaves (Fig. 1B) (Braun et al., 2006). In tdy1 mutant leaves, green tissues are essentially like wild type, while chlorotic regions accumulate excess soluble sugars and starch. Expression of the tdy1 phenotype requires high light as the leaf blade emerges from the whorl and, once formed, the variegation pattern is permanent. In a genetic background capable of producing anthocyanin in vegetative tissues, tdy1 leaves accumulate anthocyanin exclusively in chlorotic regions, presumably as an osmotic stress response to the excess carbohydrate levels ( Fig. 1C) (Chalker-Scott, 1999). We took advantage of this anthocyanin accumulation pattern as a marker to detect the presence of tdy1 phenotypic regions within albino or pale-green mutant tissues in several experiments described in this paper.
Several features of the tdy1 phenotype suggest that the accumulation and spread of a sugar determines the tissue phenotype. First, within a clonal lineage, cells display different tissue phenotypes (either chlorotic or green), excluding a lineage-dependent mechanism to explain their formation (Fig. 1A, B). Second, within a tdy1 phenotypic region all cells display uniform pigmentation, in contrast to what might be predicted if the phenotype of each cell was determined independently. Finally, we observe chlorotic regions surrounded by green tissues, which develop at the same time in a similar environment. This suggests that the external environment alone does not determine the tissue phenotype. Based on these and other observations, we postulated that Tdy1 integrates developmental and physiological information to induce transport of Suc into the veins, thereby preventing the formation of carbohydrate hyperaccumulating regions in leaves (Braun et al., 2006).
We previously found that the first detectable sign of a tdy1 phenotypic region is the hyperaccumulation of starch. This indicates that chlorosis is a secondary effect of excess carbohydrate accumulation, and suggests carbohydrates play a role in generating the tdy1 phenotype (Braun et al., 2006). In this paper, we provide evidence supporting the hypothesis that a sugar is the substance inducing the tdy1 phenotype. We show that the boundaries of a tdy1 7 region frequently occur at a lateral vein, suggesting that the transport of Suc out of the tissue through veins restricts expansion of the region. Because a tdy1 region is caused by the accumulation of excess carbohydrates, we hypothesized that chloroplasts are required for expression of the mutant phenotype. To test this, we performed double mutant analyses of tdy1 with iojap1 (ij1) and Yellow green-striate*-Mutator (Yg-str*), two mutations that give rise to lineages of photosynthetically impaired tissue. The results demonstrate that functional chloroplasts are needed to produce the sugars that induce the tdy1 phenotype. In addition, we found that albino tissues can express the tdy1 phenotype if they are located adjacent to green tissues expressing the mutant phenotype.
To further investigate where and how Tdy1 functions, we performed a clonal mosaic analysis.
We constructed genetic stocks linking an albino mutation to tdy1 as a means to distinguish mutant from wild type tissue. Radiation was used to induce somatic chromosome breaks and the mosaic sectors analyzed for tdy1 phenotypic expression. If a gene product is able to influence the phenotype of cells outside those in which it is expressed, the gene is considered to act non-cellautonomously. Conversely, if a gene functions only in the cell in which it is produced, it acts cell-autonomously (Hake and Sinha, 1994 ). In the transverse (adaxial-abaxial) dimension we found that Tdy1 acts within the innermost layer of leaves to limit the accumulation of Suc.
Within the lateral (midrib-margin) dimension, we observed that wild type tissue adjacent to mutant tissue was able to compensate for the lack of Tdy1 function over a short distance, indicating that Tdy1 functions in a limited non-cell-autonomous manner. Our results support a model whereby Tdy1 acts in the veins to promote Suc export.

tdy1 Region Boundaries Are Often Delineated by Lateral Veins
tdy1 leaves exhibit a nonclonal pattern of chlorotic regions that accumulate carbohydrates ( Fig.   1B) (Braun et al., 2006). This suggests that a mobile substance such as Suc accumulates and spreads within a local area of a developing leaf to induce the phenotype (Fig. 1B, C). In examining tdy1 leaves, we frequently observed very sharp demarcations at tdy1 region boundaries (Fig. 1C, arrows). To understand what was responsible for limiting the expansion of a tdy1 region and the nature of the sharp boundaries, we investigated the boundaries for morphological features.
Examination of transverse leaf sections from tdy1 region boundaries revealed that lateral veins, the largest vein class in the maize leaf, frequently separated chlorotic and green tissues ( Fig. 2 and Table I). As previously shown, chlorophyll levels in the tdy1 regions were greatly diminished relative to those in green tissue ( Fig. 2A). Ultraviolet (u.v.) illumination verified that the tdy1 regions contain reduced chlorophyll autofluorescence and confirmed that the lateral vein was the border (Fig. 2B). As tdy1 regions hyperaccumulate starch (Braun et al., 2006), we examined whether the starch accumulation phenotype coincided with the lateral vein boundary.
We found that tdy1 regions contain far greater levels of starch, present in both mesophyll and bundle sheath cells, than in green tissue where starch was located only in bundle sheath cells (Fig. 2C,D). Further, at the lateral vein, bundle sheath cells on the chlorotic side express the starch hyperaccumulation phenotype, while those on the green side contain normal starch levels.
These data indicate that the discrete boundaries frequently correspond to lateral veins. The data also suggest that lateral veins may limit the lateral expansion of a tdy1 region by transporting Suc out of the tissue. However, we note that tdy1 regions can encompass multiple lateral veins, indicating that lateral veins do not always limit the expansion of a region (Fig. 1B, C). In addition, boundaries were rarely located at intermediate veins (Table I). In the proximal-distal axis, there were no obvious morphological features located at tdy1 region boundaries.

9
The formation of a tdy1 region requires high light, potentially for the production of high levels of Suc (Braun et al., 2006). Further, these regions accumulate excess carbohydrates, implying that chloroplast function may be needed for this process. To determine if chloroplasts are necessary for the development of a tdy1 region, double mutants of tdy1 and ij1 were constructed. ij1 is a recessive mutation that causes defects in chloroplasts in some cells early in development, resulting in longitudinal stripes of albino tissue in an otherwise largely green leaf (  3G). In ij1 albino tissue, starch was never observed within any cells (Fig. 3H, I), consistent with albino leaf tissue not accumulating starch (Cox and Dickinson, 1971). In ij1; tdy1 albino tissue expressing anthocyanin, starch accumulation was seen in mesophyll and bundle sheath cells, albeit reduced compared to a tdy1 region (Fig. 3G, J, K). Because albino tissue can accumulate starch and anthocyanin it demonstrates that functional chloroplasts are not required for a cell to express the tdy1 phenotype.

Functional Chloroplasts Generate Sugars that Overaccumulate in tdy1 Regions
In ij1; tdy1 double mutant plants, we did not observe tdy1 regions solely in white tissue. Albino tissues expressing the tdy1 phenotype were always found next to tdy1 regions in green tissues, suggesting functional chloroplasts are necessary to generate the Suc that overaccumulates and induces tdy1 phenotypic expression. To test this hypothesis, double mutants were constructed between tdy1 and Yg-str*. Yg-str* is a dominant, transposon-induced pale-green mutant with revertant sectors of wild type (yg) dark green tissue due to somatic excision of the Mutator transposable element (Fig. 4A). In the absence of Mutator activity, Yg-str* mutants are palegreen seedling lethals, indicating that the mutation abrogates chloroplast function (D.M.B., unpublished). In Yg-str*; tdy1 double mutant leaves, tdy1 phenotypic regions are indicated by the accumulation of anthocyanin. We found that these regions localize to the revertant green yg tissue (Fig. 4B). The only exceptions to this observation occurred if a tdy1 region was sufficiently large (Fig. 4B, arrows).
To verify the cellular expression of the tdy1 phenotype we examined tissues from Yg-str*; tdy1 double mutant leaves for chlorophyll and starch accumulation ( Yg-str* pale-green mutant tissue displays reduced chlorophyll abundance ( Fig. 4I) yet increased chlorophyll autofluorescence (Fig. 4J) similar to some high chlorophyll fluorescence mutants (Miles and Daniel, 1974). Yg-str* pale-green mutant tissues do not accumulate starch (Fig. 4K).
Occasionally, a tdy1 region spread into adjacent Yg-str* pale-green mutant tissue (Fig. 4B arrows, 4L), and the chlorophyll autofluorescence seen in pale-green mutant tissue was reduced (compare Fig. 4J, M). These tissues expressed anthocyanin and exhibit starch accumulation in mesophyll cells similar to tdy1 regions in yg revertant tissue (Fig. 4H, N). As tdy1 regions only initiate within yg revertant tissue, these data suggest functional chloroplasts are needed to generate the Suc that overaccumulates and induces tdy1 phenotypic expression. Additionally, if sufficient levels accumulate, Suc can move into neighboring pale-green mutant cells, causing them to express the tdy1 phenotype (Fig. 4B, L-N), as seen in ij1; tdy1 plants.

Mosaic Analysis of Tdy1
As chloroplasts are needed to generate the sugars that induce tdy1 phenotypic expression, one possibility is that Tdy1 functions in photosynthetic cells. To test this hypothesis and determine the cell autonomy of Tdy1, we performed a clonal mosaic analysis. Based on the ij1; tdy1 double mutant studies, we knew that tdy1 regions can be detected in albino tissues by anthocyanin and starch accumulation (Fig. 3). Genetic stocks were constructed in which tdy1 was linked in coupling to the proximally located albino mutant white14 (w14), and the homologous chromosome carried wild type functional alleles of both genes (Figs. 5A, S1). Gamma irradiation of germinating seeds induced chromosome breakage and uncovered albino, aneuploid w14 tdy1/sectors present in otherwise wild type green plants. We analyzed white tissues marked by anthocyanin pigmentation to determine which cell layers lacked wild type Tdy1 function, and thereby resulted in expression of the tdy1 phenotype (Fig. 5B). In this experiment, we observed tdy1 regions within albino tissue (Figs. 5B; 8A, S2). This suggests that wild type cells containing functional chloroplasts produce the Suc that moves into adjoining mutant tissue, in which the absence of TDY1 results in Suc overaccumulation and a mutant phenotype.
For describing tissue layers (TL) in the transverse dimension, we utilized the numbering system of Nelson et al., (2002) (Fig. 5D). In brief, adaxial and abaxial epidermal layers are respectively termed TL1 and TL5, subepidermal mesophyll layers TL2 and TL4, and the innermost layer, containing the veins, bundle sheath cells and interveinal mesophyll cells, TL3. Chimeric sectors analyzed in the mosaic analysis are summarized in Figure 5E and Supplemental Table S1. Ten genotypic classes (i-x) of chimeric sectors expressed a tdy1 phenotype. In examining the distribution of tdy1 regions within mosaic leaves, there was no effect whether the albino sectors were located on lower (juvenile) or upper (adult) leaves, nor were differences found between a single albino sector on an isolated leaf or on a meristematic sector encompassing multiple leaves.
Likewise, sector position within the lateral dimension of the leaf had no effect on tdy1 phenotypic expression (Table S1). Examination of w14 hemizygous tissue from control plants not carrying tdy1 showed aneuploidy for chromosome six did not result in anthocyanin deposition ( Fig. 5C) or produce any adverse affects on leaf development, as previously reported (Walker and Smith, 2002).

Overaccumulation of Sucrose Induces Wild Type Cells to Express the tdy1 Phenotype
To genetically dissect the site and mode of Tdy1 function, we investigated whether a wild type epidermis prevented expression of the mutant phenotype. For each albino sector, the epidermal genotype was determined by examining guard cells for chlorophyll autofluorescence, as these are the only epidermal cells containing chloroplasts (Fig. 6A  However, we identified multiple sectors in which a genetically wild type epidermal layer(s) overlaid entirely mutant internal layers and exhibited a tdy1 phenotype, as marked by anthocyanin accumulation (Figs. 5E (classes iv-vi); 6E, F). Consistent with this, we observed an epidermal mericlinal sector where wild type and mutant tissues are juxtaposed and both strongly display the tdy1 anthocyanin accumulating phenotype regardless of their genotype (Fig. 6G, H).
Other periclinal chimeric sectors with a wild type epidermal layer and different internal layers composed of genetically wild type or mutant layers also expressed a tdy1 mutant phenotype ( In every case examined, wild type Tdy1 function in the epidermis was not sufficient to prevent a tdy1 region from forming. Thus, independent of its genotype, the epidermis phenotypically reflected the phenotype of the internal layers. These data also demonstrate that genotypically wild type Tdy1 epidermal cells can be induced to express the tdy1 mutant phenotype, presumably due to accumulation and movement of Suc from underlying mutant cells.

Tdy1 Acts in the Innermost Layer of Leaves
To investigate in which internal tissue layers Tdy1 functions and determine its cell autonomy, aneuploid w14 tdy1/-chimeric sectored leaves expressing tdy1 regions were examined. As shown previously, wild type tissue displays abundant chlorophyll levels and shows starch accumulation only in bundle sheath cells (Fig. 7A-D). White w14 tdy1/-sectors not expressing anthocyanin lacked both chlorophyll and starch ( Fig. 7E-H). In contrast, w14 tdy1/-albino sectors containing tdy1 phenotypic regions marked by anthocyanin lacked chlorophyll (Fig. 7I, J) and accumulated starch in both mesophyll and bundle sheath cells (Fig. 7K, L). These results show aneuploid leaf sectors express the tdy1 phenotype. In periclinal chimeras, tdy1 regions were observed only in instances where the TL3 was mutant ( Fig. 5E (classes ii-x); Table S1). This can be seen in an example where part of the TL4 was genetically wild type, as evidenced by the presence of chlorophyll, but exhibited the diminished chlorophyll abundance of a tdy1 region (Fig. 7M, N; compare with Figs. 2B, 7B). Moreover, the TL4 mesophyll cells accumulated starch, which is a hallmark of tdy1 phenotypic expression ( Fig.   7O, P). Hence, internal wild type cells can also be induced to express the tdy1 mutant phenotype.
In every tdy1 phenotypic region analyzed we observed that all five tissue layers expressed the tdy1 anthocyanin and starch accumulating phenotypes, regardless of whether the tissue layers were genetically mutant or wild type. We never observed instances of mixed phenotypic layers in which one or more tissue layers expressed a tdy1 mutant phenotype while the others did not.
Significantly, in every chimeric sector analyzed that expressed a tdy1 phenotype, the TL3 was mutant. This suggests that the site of Tdy1 function is within the innermost leaf layer (Fig. 5E). This is supported by the observation that the genotype of the epidermal and subepidermal mesophyll cell layers had no influence on determining the phenotype of the tissue (Figs. 5E; 6G, H; 7D, H, L, P; Table S1).

Wild Type Tissue Has a Short-Range Compensatory Effect in the Lateral Dimension
If the Tdy1 gene acts cell-autonomously, tdy1 phenotypic regions should form in the albino mutant sectors and extend to the border of green, wild type tissue. Alternatively, if Tdy1 functions completely non-cell-autonomously, no tdy1 regions should occur in the white tissue as Tdy1 function in neighboring wild type cells would generate a mobile product that can complement the mutation in albino tissues. As evidenced by the pronounced anthocyanin accumulation, the tdy1 phenotype was expressed in albino tissue (Figs. 5B; 8A, B; S2). This indicates Tdy1 does not generate a signal that is able to move laterally over large distances.
However, the tdy1 phenotypic regions never reached the mosaic sector border. We always

DISCUSSION
tdy1 mutant leaves display chlorotic regions resulting from the hyperaccumulation of carbohydrates. These regions form during a limited period in leaf development as the leaf emerges from the whorl (Braun et al., 2006). Therefore, TDY1 must function at or prior to this stage to prevent excess carbohydrate accumulation. As elaborated below, one possibility for TDY1 function is to promote the activity of veins to transport Suc out of the tissue and lower its concentration. Consistent with this, our results demonstrate that a chloroplast-derived product, likely Suc, induces the tdy1 phenotype, and that Tdy1 acts in the tissue layer containing the veins.

Lateral Veins as tdy1 Region Boundaries
At a tdy1 phenotypic region boundary, cells on one side of the lateral vein exhibit the features of a tdy1 chlorotic region, and those on the other side the characteristics of normal appearing green tissue. This organization may reflect the "half-vein" model for maize leaf vascular ontogeny (Langdale et al., 1989), and indicates that different sides of a lateral vein can develop and function independently from each other. In fact, lateral veins have often been found to mark the boundaries between leaf sectors (Cerioli et al., 1994, and references therein). We hypothesize that the transport function of veins may be responsible for their association with tdy1 phenotypic region boundaries. Two explanations for this association are that lateral veins have the greatest transport capacity due to their larger amount of phloem tissues (Russell and Evert, 1985), and they have hypodermal sclerenchyma fibers between the bundle sheath and epidermis which form a barrier to apoplastic solute movement (Esau, 1977). Intermediate veins have limited hypodermal sclerenchyma, which might also explain the occasional boundary occurring at these veins.
If lateral veins can act as region boundaries, how do tdy1 regions encompass multiple lateral veins? We previously proposed that a threshold level of Suc determines whether the tdy1 phenotype is expressed (Braun et al., 2006). Within a developing tdy1 region, we hypothesized that the concentration of Suc exceeds this threshold. However, at a region boundary the Suc concentration drops below the threshold and results in neighboring tissue not expressing the phenotype. At the region borders, either Suc never accumulated to a high enough level to induce the phenotype, or the veins loaded sufficient amounts to drop its concentration below the threshold needed to induce the phenotype. Hence, within a tdy1 region, we suggest that the concentration of Suc internal to the borders exceeded the capacity of veins to load and remove the sugar, resulting in the tissue expressing the tdy1 phenotype.

Functional Chloroplasts Are Needed to Generate Sugars That Induce the tdy1 Phenotype
If Suc or another sugar induces the mutant phenotype, we expect functional chloroplasts would be required for its production. To test this hypothesis, we constructed double mutants between tdy1 and Yg-str*. In Yg-str*; tdy1 double mutants, tdy1 phenotypic regions initiated in revertant yg tissue. tdy1 regions did not initiate in pale-green mutant tissue, despite it occupying a greater percentage of leaf area than green, revertant tissue. This restriction of tdy1 phenotypic expression demonstrates that functional chloroplasts are needed to generate the Suc required for inducing the tdy1 phenotype. Occasionally a tdy1 region was large enough that the Suc could spread into adjacent pale-green mutant tissue. In these cases, the fact that pale-green mutant tissue was capable of accumulating starch and anthocyanin indicates functional chloroplasts are not required to express the tdy1 mutant phenotype. This is consistent with our findings from ij1; tdy1 double mutants and albino mosaic sectors. These data support the hypothesis that the inducing substance is a chloroplast byproduct, likely Suc.

Tdy1 Functions in the Innermost Leaf Layer
To understand how Tdy1 functions and to determine its focus of action, we performed a mosaic analysis. In the transverse dimension, we found that Tdy1 function within the epidermal or subepidermal mesophyll layers was not sufficient to prevent tdy1 phenotypic expression. In the absence of Tdy1 function in the TL3, Suc accumulates and moves into all tissue layers, establishing a tdy1 phenotype throughout the tissue. Therefore, TDY1 functions to limit the accumulation of Suc. Further, the data indicate that Tdy1 acts within the innermost leaf layer.
Though it was not possible in this experiment to distinguish where in the TL3 Tdy1 acts, we speculate that Tdy1 functions in the vein. In support of this, we recently cloned Tdy1 and found it encodes a novel protein (Y.Ma and D.M.B., unpublished results). Preliminary expression studies detect the Tdy1 transcript only in veins, confirming the mosaic analysis results. We conclude that Tdy1 functions in the TL3 to limit the accumulation of Suc.

A Model for TDY1 Function
In the lateral dimension, tdy1 phenotypic regions never reached the mosaic sector border, indicating that wild type tissue produces a non-cell-autonomous compensatory effect over the distance of several veins. That wild type tissue is apparently both the source of the Suc that induces tdy1 phenotypic expression, and produces a short-range compensatory effect, suggests that these are separate processes. We found that functional chloroplasts are needed to produce sugars that overaccumulate and induce the mutant phenotype. We suggest that the function of vascular tissues, specifically phloem loading of Suc, is responsible for the compensatory effect.
Two alternatives could account for the ability of wild type tissue to prevent neighboring mutant tissue from expressing a tdy1 phenotype (Fig. 9). The first possibility is that wild type tissue generates a short-range mobile signal (either the TDY1 gene product or a downstream transmissible signal) that moves into adjacent mutant tissue and insulates it from the overaccumulation of Suc (Fig. 9A). This would prevent expression of the tdy1 mutant phenotype in the neighboring mutant veins, possibly by promoting their ability to load and transport Suc.
Suc transport out of the tissue through the veins would reduce its concentration in the albino tissue adjacent to the mosaic sector border and prevent these tissues from expressing the tdy1 phenotype. The second possibility is that wild type tissue indirectly affects nearby tdy1 phenotypic expression. For example, veins located in the bordering wild type tissue might siphon the Suc from neighboring mutant tissue (Fig. 9B). This would lower its concentration below the level needed to induce a tdy1 phenotypic region in the albino tissue located near the mosaic sector border. In either case, the strength with which wild type tissue could insulate or reduce the Suc concentration in mutant tissue would extend only for a limited distance. We previously proposed that TDY1 functions as a sugar or osmotic stress sensor that upregulates an inducible export pathway, resulting in removal of excess sugar from a region of tissue and leading to its normal development (Braun et al., 2006). Data presented in this paper that lateral veins are frequently associated with tdy1 region boundaries, that chloroplasts are needed to generate the sugar that induces tdy1 phenotypic expression, and that Tdy1 functions in the tissue layer containing the veins support this model. To incorporate the limited non-cellautonomous action of Tdy1 into the model, we favor the second possibility that TDY1 rescues  et al., 1998), phenotypes resembling a tdy1 region. Future studies will investigate potential associations between TDY1 and maize SUTs in controlling carbohydrate accumulation in leaves.
Understanding how Tdy1 regulates carbon partitioning may have applicability towards engineering biomass feedstocks for biofuels production.

Growth Conditions and Genetic Stocks
Maize ( For all other experiments, tdy1 plants in a W23-derived stock capable of anthocyanin expression in leaves were used. In this line, anthocyanin specifically marks tdy1 regions (Braun et al., 2006) and was used to identify tdy1 regions in albino and pale-green mutant tissues. ij1 and Yg-str* stocks were obtained from the Maize Genetics Cooperation Stock Center. The F1 progeny of ij1 and tdy1 mutant parents was self-fertilized and the F2 individuals cross-pollinated to generate lines segregating both mutants. Plants heterozygous for the dominant Yg-str* mutant were crossed to tdy1 plants, and F1 Yg-str* plants backcrossed to tdy1 mutants to generate families segregating Yg-str*; tdy1 double mutants.
The albino w14 mutation was obtained from the Maize Genetics Cooperation Stock Center and maps to the long arm of chromosome six. We fine mapped w14 3.6 cM proximal from bnlg1702 (2 recombinants/56 chromosomes) and 10.7 cM distal to bnlg2249 (6 recombinants/56 chromosomes) placing w14 approximately 45 cM proximal to tdy1. The crossing scheme used to generate stocks for the Tdy1 mosaic analysis is diagramed in Supplemental Fig. S1.

Microscopy
Transverse free-hand sections were cut from leaf material using a razor, mounted in water and observed on a Nikon Eclipse 80i fluorescent microscope. Observations were made under brightfield, then subsequently under u.v. using a 365 nm excitation filter and a 420 nm long pass emission filter, or FITC using a 470 nm excitation filter and a 515 nm long pass emission filter, with epifluorescent illumination provided by a 100W mercury lamp. To characterize starch accumulation in photosynthetic cells, sections were stained on the slide for approximately ten seconds using a 1.25% IKI solution (Ruzin, 1999) 20 captured using a DXM1200F Nikon Digital Camera. All images within a figure were taken using identical exposure conditions.

tdy1 Region Boundary Analysis
Sharp boundaries of tdy1 regions were examined to determine the frequency at which boundaries coincided with a particular class of veins. A region boundary was frequently discontinuous along its length such that different parts might define more than one class. Frequency was defined as the percentage a particular class occurred among all boundaries scored. One hundred and four distinct boundaries from twenty-five tdy1 regions on ten leaves were evaluated.

Starch Staining of Leaves
Leaves were decolorized by boiling in 95% ethanol, rinsed in water, stained in 1.25% IKI for approximately five minutes, rinsed and photographed.

Tdy1 Mosaic Analysis
Maize seeds were imbibed in darkness for forty-five hours at room temperature on moist paper towels. For the Tdy1 mosaic analysis, 5000 germinating seeds were irradiated. Tdy1 stocks were constructed using the anthocyanin accumulating genetic background to specifically mark tdy1 regions, as described previously (Braun et al., 2006). As controls, 1000 seed carrying w14 in the absence of tdy1 were also irradiated. Seeds were irradiated with 1500 rads of gamma radiation from a 60 Co source at the Penn State Radiation Science and Engineering Center (University Park, PA), and then planted into a freshly prepared field. Plants were observed for the presence of albino sectors throughout their development, and sectored leaves were harvested after anthocyanin accumulated marking a tdy1 region. Leaves were photographed, and leaf number, the position and widths of the albino sector as well as the tdy1 regions were recorded on a standard diagram. Though the amount of anthocyanin accumulation varied leaf to leaf, this marker was diagnostic for tdy1 regions, as w14 only aneuploid control sectors never accumulated anthocyanins. Forty-three leaves with white aneuploid sectors expressing tdy1 regions were analyzed in detail. Some leaves contained mosaic sectors composed of different genotypes at different locations, and by including these, we analyzed a total of sixty-four informative cases.
To determine which inner leaf layers expressed the w14 phenotype, chlorophyll abundance was observed in free-hand sections under bright-field, and chlorophyll autofluorescence scored under u.v. To inspect chlorophyll presence in epidermal layers, pieces of leaf were mounted in water and guard cells examined for chlorophyll autofluorescence under blue light. In periclinal chimeric sectors, two important markers for tdy1 phenotypic expression in internal layers were 1) reduced chlorophyll autofluorescence, and 2) starch accumulation in mesophyll cells. To determine tdy1 phenotypic expression in these periclinal chimeric sectors, tissue sections immediately proximal and distal to the anthocyanin accumulating tissue of interest, as well as adjacent wild type tissue, were inspected to compare the degree to which chlorophyll autofluorescence was reduced, and for determining relative starch abundance and location following IKI staining.

Supplemental Material
Supplemental Table S1. Chimeric sectors analyzed in Tdy1 mosaic analysis. Figure S1. Crossing scheme used to generate stocks for mosaic analysis. Figure S2. Compensatory effect on mutant tissue is short-range and variable.       genotypically wild type layers, while white indicates tissue layers that are genotypically mutant for w14 and tdy1. Numbers below each class represent the frequency this genotypic configuration was observed.    Figure S1. Crossing scheme for Tdy1 mosaic analysis. w14/+ heterozygotes were self-pollinated and outcrossed as males to homozygous tdy1 mutant and B73 females (P1 generation). The progeny from self-pollinations were scored to determine which male parents were heterozygous. Outcrosses of w14 heterozygotes to B73 were used to generate the chromosome six aneuploid control seed. F1 progeny of the outcross to tdy1 were self-fertilized to generate F2 families segregating tdy1 and w14. tdy1 individuals were self-pollinated to identify

Supplemental
A, Marginally located albino, aneuploid tissue expressing a tdy1 phenotype. Note the lateral distance between the anthocyanin accumulating and wild type tissue is variable along the sector length. B, Mosaic sector expressing the tdy1 phenotype in the middle of a leaf. C, Close up of sector within box in B. Both sides of the tdy1 phenotypic region are uneven and show variable distance to wild type tissue along length of sector. D, Close up of an albino sector expressing the tdy1 phenotype at the leaf margin showing the compensatory effect along its length.