Transfer cells are specialised transport cells containing invaginated wall ingrowths that generate an amplified plasma membrane surface area with high densities of transporter proteins. They trans-differentiate from differentiated cells at sites at which enhanced rates of nutrient transport occur across apo/symplasmic boundaries. Despite their physiological importance, little is known of the molecular mechanisms regulating construction of their intricate wall ingrowths. We investigated the genetic control of wall ingrowth formation in phloem parenchyma transfer cells of leaf minor veins in Arabidopsis thaliana. Wall ingrowth development in these cells is substantially enhanced upon exposing plants to high-light or cold treatments. A hierarchical bioinformatic analysis of public microarray datasets derived from the leaves of plants subjected to these treatments identified GIGANTEA (GI) as one of 46 genes that are commonly up-regulated twofold or more under both high-light and cold conditions. Histological analysis of the GI mutants gi-2 and gi-3 showed that the amount of phloem parenchyma containing wall ingrowths was reduced 15-fold compared with wild-type. Discrete papillate wall ingrowths were formed in gi-2 plants but failed to develop into branched networks. Wall ingrowth development in gi-2 was not rescued by exposing these plants to high-light or cold conditions. In contrast, over-expression of GI in the gi-2 background restored wall ingrowth deposition to wild-type levels. These results indicate that GI regulates the ongoing development of wall ingrowth networks at a point downstream of inputs from environmental signals.
Transfer cells (TCs) are functionally specialised for enhanced nutrient transport in plants. They develop highly invaginated wall ingrowths, a modification that provides a scaffold to greatly amplify the surface area of their plasma membranes, which are enriched in nutrient transporters (Offler et al., 2003). TCs are formed by trans-differentiation of differentiated cells at sites where enhanced rates of nutrient flow occur across apo/symplasmic boundaries (Offler et al., 2003). For example, TCs can form at loading and unloading regions of the vascular system (Haritatos et al., 2000), in vascular parenchyma cells of stem nodal complexes to redirect nutrient flows between xylem and phloem (Pate et al., 1970), and at the maternal/filial interface in seeds to facilitate nutrient exchange between generations (Thompson et al., 2001). The physiological importance of TC development is clearly demonstrated by the pea seed mutant E2748, in which disruption of normal embryo growth during maturation of the seed correlates with the failure of cotyledon epidermal cells to differentiate into TCs (Borisjuk et al., 2002).
Despite the physiological importance of TCs in nutrient exchange processes contributing to plant development (Offler et al., 2003), little is known of the molecular mechanisms regulating induction and subsequent construction of their intricate wall ingrowths. Developing maize kernels have been used as an experimental system to identify genes expressed specifically in basal endosperm TCs. Four defensin-like genes, BETL1–4 (Hueros et al., 1995, 1999; Serna et al., 2001; Thompson et al., 2001), a novel cell wall-related gene, MEG1 (Gutiérrez-Marcos et al., 2004), and a MYB-related transcriptional activator, ZmMRP-1 (Gómez et al., 2002), have been identified as TC-specific genes in basal endosperm TCs. Interestingly, ZmMRP-1 trans-activates BETL and MEG1 promoters (Gutiérrez-Marcos et al., 2004; Barrero et al., 2006), and the ZmMRP-1 promoter is active in other species in which active transport between source and sink tissues is presumed to occur (Barrero et al., 2009). Indeed, Gómez et al. (2009) used ectopic expression of ZmMRP-1 in maize endosperm to demonstrate that this gene is a key regulator controlling TC differentiation in this tissue.
In addition to these findings, two recent transcript profiling studies have demonstrated that TC development is accompanied by large-scale transcriptional regulation. Dibley et al. (2009) used cDNA-amplified fragment length polymorphism (AFLP) to estimate that trans-differentiation of epidermal cells in Vicia faba cotyledons to form TCs may involve differential expression of approximately 650 genes. Similarly, Thiel et al. (2008) identified 815 genes on a 12K macroarray as being differentially expressed during formation of nucellar projection and endosperm TCs in growing barley (Hordeum vulgare) grains.
We investigated the genetic control of wall ingrowth formation in the model species Arabidopsis thaliana. Phloem parenchyma (PP) in minor veins of leaves develop a TC morphology where they function in phloem loading (Haritatos et al., 2000; Amiard et al., 2007). These cells deposit ‘reticulate’ (Talbot et al., 2001; McCurdy et al., 2008) wall ingrowths that are locally deposited to wall domains adjacent to neighbouring sieve elements, thereby facilitating apoplasmic loading of sugars delivered symplasmically to the PP (Maeda et al., 2006; Amiard et al., 2007). Importantly for experimental analysis using Arabidopsis, the deposition of wall ingrowths in PP TCs is enhanced substantially when plants are transferred to high-light (Amiard et al., 2007) or cold (Maeda et al., 2006, 2008) conditions. This enhancement of wall ingrowth development by abiotic stress possibly operates through a stress signalling response involving jasmonic acid (JA) (Amiard et al., 2007).
To identify Arabidopsis genes regulating wall ingrowth development in PP TCs, we performed a bioinformatic analysis of publicly available microarray datasets, and identified GIGANTEA (GI) as one of 46 genes whose expression is commonly up-regulated twofold or more in leaf tissue within 24 h after transfer to either high-light or cold conditions. Histological analysis of the GI mutants gi-2 and gi-3 revealed that the amount of PP containing wall ingrowths was significantly reduced compared to wild-type (WT). Transmission electron microscopy (TEM) showed that individual wall ingrowth papillae in gi-2 were deposited in PP TCs but they did not progress to form the typical branched networks of fenestrated wall material seen in the WT. Furthermore, normal reticulate wall ingrowth formation in PP TCs was not rescued when gi plants were exposed to high light or other abiotic stresses. These results indicate that GI regulates wall ingrowth development downstream of the stress signalling pathways induced by these abiotic signals, and therefore functions in a regulatory pathway determining wall ingrowth development in TCs.
Calcofluor White staining of PP TCs in leaf minor veins
Because of their location deep within vascular bundles, PP TCs are not readily accessible for experimental observation, and consequently these cells have only been studied by TEM in Arabidopsis (Haritatos et al., 2000; Maeda et al., 2006, 2008; Amiard et al., 2007). To circumvent this limitation and provide a robust and rapid means of assessing wall ingrowth development in PP TCs, we developed a simple procedure involving Calcofluor White (CW) staining of cleared leaf material (see Experimental Procedures). Using this procedure, we observed linear regions of bright fluorescence associated with the reticulate venation network, defined here as both the major vein network (up to three orders of branching from the mid-rib; see Haritatos et al., 2000) and the minor vein network (four or more orders of branching), minus the mid-rib. Examples of this fluorescence in major and minor veins are shown in Figure 1(a,b). We subsequently refer to major and minor veins comprising the reticulate venation network simply as ‘veins’. The staining consisted of thin bands of bright fluorescence running either continuously for some distance along veins (Figure 1a,c,d) or restricted to discrete regions or patches, particularly in terminating veins (Figure 1b). Often two or more bands of fluorescence were seen within veins (data not shown), and CW-stained material was often more extensive where veins branched (Figure 1d). Higher-magnification views showed that the strands of CW-stained material did not correspond to xylem elements (Figure 2) but were positioned deeper within the vascular bundle, at a location consistent with the bright staining corresponding to phloem tissue (see Haritatos et al., 2000). Interestingly, Figure 2 shows that CW does not stain cellulose-rich xylem elements in cleared leaf material. At this higher magnification, the CW staining in phloem cells was often seen as a central band of fluorescence with an irregular or patchy appearance (Figure 3a). The width of this central band of CW-stained material was 2.7 ± 0.7 μm (mean ± SE, n =10). These dimensions are consistent with scanning electron microscope (SEM) images of PP TCs exposed after fresh leaf material was torn paradermally and then processed for SEM. In these instances, wall ingrowths in the PP TCs were seen as a central band of entangled reticulate wall projections of irregular density (Figure 3b). The width of this central band of wall ingrowth material viewed by SEM was 2.1 μm ± 0.3 (mean ± SE, n =6), which is comparable to the bands of CW-stained material seen in cleared leaves. Based on these observations, we conclude that the CW-stained material seen by fluorescence microscopy corresponds to wall ingrowths deposited in PP TCs of veins.
To test this conclusion experimentally, we performed CW staining of mature rosette leaves from plants exposed to high light, methyl jasmonate (MeJA) or cold (see Experimental Procedures). Scoring the length of veins displaying PP TCs containing CW-stained wall ingrowths revealed that, in low light-grown plants, 33% of the vein length contained PP TCs with stained wall ingrowths (Table 1). This figure increased nearly twofold for plants analysed 7 days after transfer to either high-light or cold (4°C) conditions, or sprayed daily with 10 μm MeJA (Table 1). The magnitude of this increase is consistent with the results of a study by Amiard et al. (2007), who used quantitative TEM to measure increased wall ingrowth development in PP TCs of minor veins in response to high light and MeJA, and also with the results of a study by Maeda et al. (2006), who reported substantial increases in wall ingrowth deposition in PP TCs in response to cold treatment. Based on these data (Figures 1–3 and Table 1), we conclude that CW staining of cleared leaf tissue provides a simple and semi-quantitative analysis of PP TC development in veins of Arabidopsis.
Table 1. Comparison of the extent of wall ingrowth deposition in PP TCs in response to abiotic stresses
Percentage vein length exhibiting PP with CW-stained wall ingrowths
Wild-type (Col-0) plants were subjected to various stresses (high light, MeJA and cold; see Experimental Procedures), and mature leaves were then stained with CW. ImageJ was used to calculate the percentage of vein length displaying CW-stained PP (see Experimental Procedures). Data are means ± SE, n ≥5. *P < 0.001 (Student’s t test) compared to control.
Low light (control)
33.2 ± 6.1
53.4 ± 5.9*
55.7 ± 6.3*
52.3 ± 4.9*
Bioinformatic analysis and CW staining identifies GI as a gene regulating wall ingrowth deposition in PP TCs
Wall ingrowth deposition in PP TCs is enhanced in plants exposed to high-light (Amiard et al., 2007) or cold (Maeda et al., 2006, 2008) conditions. To identify candidate genes putatively involved in regulating wall ingrowth deposition in Arabidopsis, we analysed publicly available microarray datasets documenting changes in gene expression in the leaves of plants exposed to high-light or cold growth conditions (see Experimental Procedures and Table S1). This analysis identified a common cohort of 46 genes whose expression in leaf tissue increased twofold or more within the first 24 h of exposure to either high-light or cold conditions (Table S2). A similar number of genes (42) were identified as being down-regulated twofold or more over this period in response to these treatments (Table S3). As a proof-of-concept test of the biological relevance of the lists generated, we selected 11 mutants from the cohort of up-regulated genes (Table S2) for analysis of wall ingrowth deposition in PP TCs of veins. Ten of these mutants were chosen based on the availability from the Arabidopsis Biological Resource Center of homozygous lines carrying an exon-positioned T-DNA insertion in the nominated gene, while one mutant, gi-2, is a well-characterised ethylmethane sulfonate-induced point mutation in GIGANTEA (GI) (Fowler et al., 1999). Staining mature leaves from these plants with CW showed that, of the 11 lines examined, 10 were indistinguishable from WT (Col-0) when comparing levels of CW staining in leaf PP TCs (Figure S1). In contrast, however, the GI mutant, gi-2, showed little if any CW staining of PP TCs in veins (Figure S1). The gi-2 line (in the Col-0 background) carries a point mutation in GI causing a premature stop codon and consequently production of a truncated polypeptide representing less than 20% of the GI protein (Fowler et al., 1999). We analysed this mutant further by CW staining of transverse thin sections of resin-embedded leaf material. Under these conditions, sectioned WT leaves showed strong staining of xylem elements in minor veins in addition to single localised patches of strong fluorescence in PP TCs (Figure 4a). Each patch of CW staining in a PP TC was located adjacent to a smaller, oblong or rhomboid-shaped sieve element (Figure 4a, and see Figure 3 in Haritatos et al., 2000 for TEM image of an equivalent vein), consistent with this bright fluorescence corresponding to staining of the highly localised wall ingrowth material deposited adjacent to sieve elements in PP TCs (see Amiard et al., 2007). In contrast, similar patches of CW-stained material were absent from an analysis of PP cells from multiple sections of gi-2 leaves (Figure 4b). The venation patterns of gi-2 leaves from 4-week-old plants grown under long days (16 h light/8 h dark) were not substantially different from those of the WT (Figure S2), indicating that the PP TC phenotype seen in gi-2 plants is not caused by secondary effects on leaf vascular development caused by disruption of functional GI expression.
Over-expression of GI fusion proteins complements the gi-2 phenotype
To assess complementation, we analysed gi-2 plants expressing either a haemagglutinin (HA)- or tandem affinity purification (TAP)-tagged GI fusion protein under the control of the CaMV 35S promoter (seeds provided by Jo Putterill, School of Biological Sciences, University of Auckland, New Zealand). Both GI- usion proteins were functional, as evidenced by rescue of the late-flowering phenotype of gi-2 plants grown under long days (David et al., 2006). Quantification of the percentage of vein length exhibiting CW-stained PP TCs in either 35S::HA-GI or 35S::GI-TAP lines revealed that expression of either recombinant GI polypeptide in the gi-2 background restored the levels of PP TCs displaying wall ingrowth deposition to that seen in WT (Table 2). Furthermore, similar to WT, the number of PP TCs displaying CW-stained wall ingrowths nearly doubled when these lines were exposed to high light (Table 2). These data provide strong support for the conclusion that functional GI expression is required for normal wall ingrowth deposition in PP TCs.
Table 2. Extent of wall ingrowth deposition in PP TCs of leaf veins in response to abiotic stresses
Percentage vein length exhibiting PP with CW-stained wall ingrowths
gi-2 plants expressing GI fusions
Wild-type (Col-0), Ler, gi-2, gi-3 and gi-2 plants expressing either HA- or TAP-tagged GI fusion proteins were grown under low light or transferred to high light. Imagej was used to calculate the percentage of vein length displaying CW-stained PP. Data are means ± SE, n ≥5. *P <0.001 (Student’s t test) compared to control.
26.2 ± 3.1
4.2 ± 1.4*
36.1 ± 5.3
1.8 ± 1.2*
24.2 ± 3.2
26.4 ± 6.4
46.1 ± 5.1
3.3 ± 1.0*
55.2 ± 6.1
3.1 ± 2.6*
43.4 ± 6.2
44.1 ± 5.0
Wall ingrowth deposition in gi mutants is not rescued in response to abiotic stress
GI is well known for its role in promoting flowering and regulating circadian rhythms in Arabidopsis (Mizoguchi et al., 2005). In addition, the GI gene is also up-regulated in response to cold (Fowler and Thomashow, 2002), and gi mutants show increased tolerance of oxidative stress (Kurepa et al., 1998). We therefore tested whether the reduced wall ingrowth phenotype of gi-2 could be rescued by exposure to treatments known to enhance wall ingrowth deposition in PP TCs (Amiard et al., 2007). As previously described (Table 1), exposure of WT plants to high light, MeJA or cold resulted in an almost twofold increase in vein length containing PP TCs. However, WT levels of wall ingrowth deposition were not detected in gi-2 plants subjected to high light (Table 2), cold or MeJA (data not shown).
Arabidopsis gi mutants display considerable allelic variability in flowering responses and other traits (Martin-Tryon et al., 2007). Therefore, we also examined wall ingrowth deposition in PP cells in gi-3 (Ler background). These plants carry a point mutation causing production of a truncated polypeptide containing approximately 80% of the GI polypeptide (Fowler et al., 1999). Compared to Ler as WT, staining of mature gi-3 leaves showed that the percentage of vein length containing PP TCs was significantly reduced (Table 2). Furthermore, similar to gi-2, the extent of wall ingrowth deposition in gi-3 plants did not increase substantially in response to high light (Table 2).
The primary effect in gi-2 is arrested branching of wall ingrowths
TEM was used to analyse the morphology of the small percentage (<5%) of PP TCs that did form wall ingrowths in gi-2 (see Table 2) as determined by CW staining. Initial TEM observations of these plants indicated that some PP TCs displayed early-stage papillate wall ingrowth deposition, but lacked more developed reticulate ingrowth networks. We therefore scored the percentage of PP cells displaying either discrete, finger-like, papillate wall ingrowth projections, or branched and/or more elaborate ingrowths, in both WT and gi-2. Figure 5(c) shows that, in WT, similar numbers of PP cells (30–35%) were smooth-walled (no wall ingrowths), contained discrete, finger-like papillae (Figure 5a), or contained branched (Figure 5b) or more developed wall ingrowths. In contrast, gi-2 plants had the same percentage of PP TCs with discrete papillae as WT, but only 5% (P < 0.001) of the cells contained branched or more elaborate wall ingrowths (Figure 5c). Correspondingly, gi-2 plants contained increased numbers of PP cells with no wall ingrowths (smooth cell walls) (Figure 5c). We interpret these results as indicating that the initial deposition of finger-like papillate wall ingrowth projections is not disrupted in gi-2, but further progression to create a more elaborate and branched reticulate wall ingrowth labyrinth is blocked.
GI is expressed in PP of minor veins
If GI regulates wall ingrowth development in PP TCs, the GI gene should be expressed in phloem tissue. Although GI is known to be expressed in all major organs (Fowler et al., 1999), a detailed analysis of GI expression at the tissue and cellular levels in Arabidopsis has not been reported. We therefore performed a histological analysis of plants expressing GUS, driven by the GI promoter (pGI::GUS construct expressed in Ler background; seeds provided by George Coupland, Max Planck Institute for Plant Breeding, Cologne, Germany). Stained leaf material from several independently transformed lines showed GUS expression throughout the vein network, including terminating minor veins. Representative images of this GUS expression in major and terminating minor veins are shown in Figure 6(a,b). The inferred GI expression was located in the phloem tissue of vascular bundles, as the GUS staining seen in some images was restricted to phloem running parallel to but distinct from xylem elements (Figure 6a), and even extending past xylem strands in terminating minor veins (Figure 6b). These images showed that xylem elements, commonly known to generate false-positive staining for GUS activity, did not accumulate GUS reaction product in these plants (Figure 6a,b). WT leaves (Ler background) did not stain for GUS (Figure 6c,d). We conclude from these results that GI is expressed in leaf phloem tissue.
Our study used bioinformatics and histological analysis to identify GI as a gene that participates in the regulatory pathway controlling wall ingrowth deposition in leaf PP TCs in Arabidopsis. As documented by both CW staining and TEM, GI mutants showed dramatic reductions compared to WT with respect to the levels of wall ingrowth deposition in PP TCs. This study therefore links mutation in a specific gene (GI) to disrupted wall ingrowth deposition in TCs.
We used bioinformatics to identify a cohort of genes that are commonly up-regulated in leaf tissue in response to exposure to high-light or cold conditions. These abiotic stresses cause increased development of wall ingrowths in PP TCs in minor veins of Arabidopsis leaves (Maeda et al., 2006, 2008; Amiard et al., 2007). On this basis, we reasoned that genes involved in regulating wall ingrowth deposition may be up-regulated under these conditions. The bioinformatics analysis was necessarily rudimentary because the microarray experiments that we analysed only very approximately replicated the experimental growth conditions used by Amiard et al. (2007) and Maeda et al. (2006, 2008) (see Experimental Procedures and Table S1) to induce enhanced wall ingrowth deposition. Furthermore, the microarray experiments sampled whole shoots or leaves, but the response of increased wall ingrowth deposition is restricted to a single cell type, namely PP TCs, in leaves. Consequently, the majority of genes identified as commonly up-regulated twofold or more in response to high-light or cold treatments are presumably unlikely to play a role in regulating wall ingrowth deposition in PP TCs. Despite these limitations, however, the bioinformatics approach, used in combination with CW staining of selected mutants, was successful in identifying GI as a gene putatively involved in regulating wall ingrowth development. Furthermore, since the analysis of only 11 of the 46 up-regulated genes listed in Table S2 was sufficient to identify GI as a confirmed candidate gene in this process (see Figure S1), we are continuing to screen the remaining 35 genes on this list for disrupted wall ingrowth deposition using confirmed homozygous T-DNA insertional mutants as they become available.
CW staining of cleared leaf material provided a rapid means to score wall ingrowth development in PP TCs. Prior to this study, PP TCs in Arabidopsis minor veins had only been studied by TEM (Haritatos et al., 2000; Maeda et al., 2006, 2008; Amiard et al., 2007). Our semi-quantitative analysis using CW staining revealed that only approximately 30% of PP TCs in plants grown under low-light conditions contained wall ingrowths (Tables 1 and 2), a result that agrees with those of Amiard et al. (2007) and Haritatos et al. (2000), who reported that PP TCs with wall ingrowths occurred only intermittently along the length of a minor vein. Despite the apparent inability of CW staining to detect early-stage papillate wall ingrowths (see Figure 5), the development of a high-throughput fluorescence-based procedure to screen for the presence of wall ingrowth complexes now offers the opportunity to use Arabidopsis as a model to investigate genetic regulation of wall-ingrowth deposition in TCs.
The CW staining procedure demonstrated that PP TCs also occur in third-order major veins (Figure 1a) and occasionally in second-order major veins (data not shown), as well as in minor veins of Arabidopsis leaves. This observation is consistent with the proposal by Haritatos et al. (2000), who concluded, based on the intimate association of both major and minor veins with mesophyll cells, that all veins in the Arabidopsis leaf probably play a role in phloem loading and can therefore be considered as minor veins.
Analysis of the lines over-expressing HA- or TAP-tagged GI supports a role for GI in regulating PP TC development. As the levels of HA- or TAP-tagged GI in these lines change in response to light in a manner similar to endogenous levels of GI in WT plants (David et al., 2006), and the photoperiod-insensitive flowering of gi-2 is suppressed, it appears that these lines display restored functionality of GI (David et al., 2006). Therefore, the presence of WT levels of PP TCs in a gi-2 line that exhibits restored GI functionality provides further support for a role for GI in regulating wall ingrowth development in these cells. The promoter–reporter gene studies (Figure 6) demonstrating that GI is expressed in the phloem of veins further support this conclusion.
GI is a well-studied circadian clock-regulated gene that is involved in numerous developmental processes including regulation of flowering time in response to photoperiod (Koornneef et al., 1991; Fowler et al., 1999), phytochrome signalling (Huq et al., 2000), carbohydrate metabolism (Eimert et al., 1995) and breaking seed dormancy (Penfield and Hall, 2009). The function of GI in regulating these processes is unknown, but mutations in GI cause a pleiotrophic phenotype affecting each of these developmental processes. Although GI has no recognised functional domains (Fowler et al., 1999), it interacts directly with FLAVIN-BINDING, KELCH REPEAT, F-BOX 1 (FKF1, Sawa et al., 2007), ZEITLUPE (ZTL, Kim et al., 2007) and SPINDLY (SPY, Tseng et al., 2004) to regulate flowering in Arabidopsis. In rice (Oryza sativa), at least seven proteins have been identified as interacting directly with OsGI (Abe et al., 2008). Collectively, these observations suggest that GI may interact with many partners to form diverse regulatory complexes involved in controlling various aspects of plant development, including wall ingrowth deposition in PP TCs, as the present study demonstrates.
The mechanism by which GI regulates wall ingrowth formation is unknown. Wall ingrowth development in TCs appears to be linked in many instances to stress responses (see Offler et al., 2003). These include exposure to high light and cold (this study and Amiard et al., 2007 and Maeda et al., 2006), but also nutrient deficiency (Schikora and Schmidt, 2002) and pathogen attack, during which giant feeding cells with wall ingrowth morphologies develop in response to nematode infection (Hammes et al., 2005). The production of early-stage discrete papillate wall ingrowths was unaffected in gi-2, but development of branched ingrowth networks was significantly reduced compared to WT (Figure 5c). This observation implies that GI is not involved in induction of wall ingrowth deposition, but is involved in the on-going formation of fenestrated layers of wall ingrowth, a process that is up-regulated in response to stress signals generated by high-light and cold conditions. A role for GI as part of a regulatory pathway required for on-going deposition of branched wall ingrowths appears to contrast with that of ZmMRP-1, a MYB-related transcription factor from Zea mays, which, when ectopically expressed in endosperm, drives the deposition of flange-type (Talbot et al., 2002) wall ingrowths in aleurone layer cells that normally do not trans-differentiate into TCs (Gómez et al., 2009).
GI mutants show increased tolerance to oxidative stress (Kurepa et al., 1998), presumably via constitutive activation of SOD and APX genes encoding superoxide dismutase and ascorbate peroxidase, respectively (Cao et al., 2006). Deposition of wall ingrowths in epidermal TCs in V. faba cotyledons involves reactive oxygen species (ROS) as part of an inductive signalling cascade (F.A., unpublished results). Thus, the enhanced ability of GI mutants to detoxify oxygen free radicals and thereby potentially reduce ROS levels in response to stress may be responsible for the apparent stalled branching of wall ingrowths observed in gi-2 plants. This possibility is supported by changes in wall ingrowth morphology observed in plants exhibiting altered ROS levels. For example, Maeda et al. (2006) reported that the Arabidopsis vte2-1 mutant, which is deficient in the production of tocopherol, a key ROS scavenger, exhibited greatly enhanced deposition of cell-wall material in PP TCs. Additionally, co-ordinated production of high levels of ROS is required for formation of nematode synctia in tomatoes (Melillo et al., 2006). Further experimentation involving exposure of gi plants to high exogenous levels of ROS may be useful in determining a possible regulatory role for ROS production in the development of wall ingrowths in Arabidopsis PP TCs.
In summary, we have identified GI, a gene that is normally studied with respect to its role in regulating flowering time in response to photoperiod, as a putative regulator of wall ingrowth deposition in PP TCs in Arabidopsis leaves. The mechanism by which GI regulates this process is unknown, but may involve co-ordination of inputs from stress signalling pathways acting through ROS signalling. Our development of a simple high-throughput histological procedure to identify wall ingrowth deposition in PP TCs will enable Arabidopsis to be used as a genetic model to investigate mechanisms regulating wall ingrowth development in plant TCs.
Plant growth and treatments
Unless otherwise stated, all seed lines were obtained from the Arabidopsis Biological Resource Center (Ohio State University, Columbus, OH). Arabidopsis seeds were sown directly onto pasteurised soil mix and stratified for 2–3 days in darkness at 4°C. Plants were then grown under low-light conditions in a growth cabinet (100–120 μmol m−2 sec−1, 22°C, 8 h photoperiod) for 3 weeks, and fully expanded rosette leaves were collected for analysis. For treatment under high-light conditions (approximately 600 μmol m−2 sec−1, 12–14 h photoperiod, 22–25°C), 2-week-old plants grown under low-light conditions were transferred to a glasshouse, and fully expanded rosette leaves were collected after 7 days. For cold treatment, 2-week-old plants grown under low-light conditions were transferred to 4°C (100 μmol m−2 sec−1, 8 h photoperiod), and fully expanded leaves were collected after 7 days. For MeJA treatment, rosette leaves of 2-week-old plants grown under low-light conditions were sprayed daily with 10 μm MeJA (Sigma, http://www.sigmaaldrich.com/) in 0.05% v/v Tween-20, and fully expanded rosette leaves were collected after 7 days. Leaves of control plants were sprayed with 0.05% v/v Tween-20.
Microarray datasets were selected on the basis of experimental growth conditions that most closely matched those used by Amiard et al. (2007) and Maeda et al. (2006, 2008). In Amiard et al. (2007), plants grown under low-light conditions (100 μmol photons m−2 sec−1, 8 h photoperiod, 25°C) were transferred to high-light conditions (1000 μmol photons m−2 sec−1, 8 h photoperiod, 25°C) and analysed after 7 days. In Maeda et al. (2006, 2008), 3–4-week-old plants grown at 120 μmol m−2 s−1 (12 h photoperiod, 22°C) were transferred to 12 h at 75 μmol m−2 sec−1 and 12 h darkness, both at 7.5°C for 3–14 days before analysis.
Raw data for each microarray set were retrieved from the Arabidopsis Information Resource (TAIR) (http://www.arabidopsis.org), and MAS5.0-processed data for each set were retrieved from NASCArrays (http://affymetrix.arabidopsis.info/narrays/). Details of the six datasets selected (three for cold treatment and three for transition from low light to high light) are presented in Table S1. MAS5.0-processed data were loaded into Microsoft Excel™ (http://www.microsoft.com/), and the present/absent calls generated by the MAS5.0 algorithm were used to exclude from further analysis any genes that were not called as present across all replicate microarrays at each time point. Mean expression values for each gene at each time point were calculated across the treatment and control arrays, and those displaying a twofold or greater increase in expression in treatments compared to control were considered ‘up-regulated’, while those displaying a twofold or greater decrease in expression were considered ‘down-regulated’.
Raw data files obtained from TAIR were loaded into BRBArrayTools (http://linus.nci.nih.gov/BRB-ArrayTools.html) (Simon and Lam, 2006) and analysed using the Robust Multiarray Analysis algorithm to calculate a mean array of gene expression levels across each array, and any gene that did not display at least a twofold increase or decrease from the mean array for at least one time point were excluded from further analysis. The expression values were converted to log2 values, and the difference between treatments and controls was calculated by subtraction. A difference of 1 or greater was considered ‘up-regulated’, while a difference of -1 or lower was considered ‘down-regulated’.
Lists of up- and down-regulated genes from both sources were listed as AGI numbers, and then uploaded to VirtualPlant (http://www.virtualplant.org). The SunGear interface (Poultney et al., 2007) was used to determine common genes present in each list. As the lists generated from the raw data processed with BRBArrayTools were contained almost wholly within the lists generated from the MAS5.0-processed data, both lists (processed and raw data) were combined in VirtualPlant into a single list for each designated time period. As the sampling times for each microarray experiment varied substantially (see Table S1), we focused our analysis on expression data obtained from leaf tissue harvested within the first 24 h after treatment. This procedure identified 46 genes that were commonly up-regulated (Table S2) and 42 genes that were commonly down-regulated (Table S3).
Calcofluor White staining
The lower epidermis of rosette leaves was scraped away using fine tweezers or abrasive paper. Leaf tissue was then cleared in 1% w/v sodium hypochlorite (WhiteKing™ commercial bleach, diluted 1:4; http://www.whiteking.com.au/) for at least 1 h and rinsed thoroughly in distilled water. The cleared leaf material was then stained for 1 h in 0.1% w/v CW, rinsed well in distilled water and mounted in 50% v/v glycerol for viewing.
For CW staining of leaf sections, mature rosette leaves were fixed in 4% w/v formaldehyde plus 0.5% w/v glutaraldehyde in 50 mM sodium phosphate buffer, pH 7.0, for 2–3 h at room temperature. The fixed tissue pieces were then dehydrated through an ethanol series and embedded in LR White resin. Thin transverse sections were cut and stained with 0.5% w/v CW for 1 h, and washed briefly in distilled water before mounting in 50% v/v glycerol.
Scanning electron microscopy
Mature rosette leaves were torn paradermally with forceps to expose PP cells of minor veins. The leaves were then processed for SEM as described by Talbot et al. (2007), which involves extracting the tissue in 2% w/v sodium hypochlorite, dehydration through an ethanol series, and critical point drying prior to mounting on double-sided carbon tabs. Specimens were sputter-coated with gold to a thickness of 20 nm, and viewed at 15 kV using a Philips XL30 scanning electron microscope.
Cleared and stained tissue, as well as sectioned material, was viewed by bright-field and fluorescence microscopy using a Zeiss Axiophot microscope (http://www.zeiss.com/) equipped with a 50 W short-arc mercury lamp. CW labelling was visualised using an FT580 filter set, and images were recorded with a Zeiss AxioCam HCr camera using Axiovision software, and then imported into Adobe Photoshop (http://www.adobe.com/) for display.
Quantification of PP TC development in CW-stained leaves
Five digital images for each CW-stained leaf were taken at 100 x magnification and imported into imagej (http://rsbweb.nih.gov/ij/). The total length of CW-stained PP was measured as a percentage of the total length of veins (major and minor) in each image, using the freehand ‘measure’ tool. In cases where more than one row of PP TCs per region of vein was clearly visible by CW staining, only the length of the combined CW-stained material was recorded. Three fully expanded leaves from a minimum of five individual plants were measured in this way to obtain the percentage of vein length exhibiting PP with CW-stained wall ingrowths (Tables 1 and 2).
GI expression analysis using histochemical localisation of GUS activity
Mature rosette leaf samples were fixed in 90% v/v acetone for 1 h on ice, rinsed in 5-bromo-4-chloro-3-indolyl-β-d-glucuronide (X-gluc) buffer (50 mm NaPO4, pH 7.2, 0.5 mm K3Fe(CN)6, 0.5 mm K4Fe(CN)6, 0.5% v/v Triton X-100), then placed in X-gluc buffer containing 0.5 mm X-gluc. The tissue was vacuum-infiltrated for 4 h, and then incubated at 37°C for at least 48 h. The stained leaf tissue was then rinsed in distilled water and post-fixed in ethanol:acetic acid (3:1 v/v) for 1 h, rinsed in 80% v/v ethanol, and cleared overnight in saturated chloral hydrate before mounting in the same solution. Microscopy was performed as described above.
Mature rosette leaves from five individual WT or gi-2 plants grown under low-light conditions were fixed and embedded in LR White resin as described above, and ultrathin transverse sections were cut and stained with uranyl acetate following standard procedures. TEM micrographs of minor vein sections were scored for the presence of either single papillae or branched, more elaborate wall ingrowth networks in PP TCs. A minimum of 25 cells were scored for each leaf section. Means and standard deviations were calculated, and a two-tailed t test was used to determine significance levels between WT and gi-2 plants within each category of cell wall ingrowth morphology.
We thank the Arabidopsis Biological Resource Center for supply of homozygous seeds, Professor Jo Putterill (University of Auckland, New Zealand) for seeds of gi-2, 35S::HA-GI/gi-2 and 35S::TAP-GI/gi-2, Professor George Coupland (Max Planck Institute for Plant Breeding, Cologne, Germany) for seeds of pGI-GUS, and Dr Xing-Ding Wang (School of Environmental and Life Sciences, University of Newcastle) for light and electron microscopy. This work was supported by an Australian Research Council Discovery Project grant (DP0664626) to J.W.P., D.W.McC. and C.E.O., and a University of Newcastle Faculty Enhancement Project Grant to D.W.McC.