Transcriptomic, proteomic and metabolomic analysis of UV-B signaling in maize
© Casati et al; licensee BioMed Central Ltd. 2011
Received: 28 September 2010
Accepted: 16 June 2011
Published: 16 June 2011
Skip to main content
© Casati et al; licensee BioMed Central Ltd. 2011
Received: 28 September 2010
Accepted: 16 June 2011
Published: 16 June 2011
Under normal solar fluence, UV-B damages macromolecules, but it also elicits physiological acclimation and developmental changes in plants. Excess UV-B decreases crop yield. Using a treatment twice solar fluence, we focus on discovering signals produced in UV-B-irradiated maize leaves that translate to systemic changes in shielded leaves and immature ears.
Using transcriptome and proteomic profiling, we tracked the kinetics of transcript and protein alterations in exposed and shielded organs over 6 h. In parallel, metabolic profiling identified candidate signaling molecules based on rapid increase in irradiated leaves and increased levels in shielded organs; pathways associated with the synthesis, sequestration, or degradation of some of these potential signal molecules were UV-B-responsive. Exposure of just the top leaf substantially alters the transcriptomes of both irradiated and shielded organs, with greater changes as additional leaves are irradiated. Some phenylpropanoid pathway genes are expressed only in irradiated leaves, reflected in accumulation of pathway sunscreen molecules. Most protein changes detected occur quickly: approximately 92% of the proteins in leaves and 73% in immature ears changed after 4 h UV-B were altered by a 1 h UV-B treatment.
There were significant transcriptome, proteomic, and metabolomic changes under all conditions studied in both shielded and irradiated organs. A dramatic decrease in transcript diversity in irradiated and shielded leaves occurs between 0 h and 1 h, demonstrating the susceptibility of plants to short term UV-B spikes as during ozone depletion. Immature maize ears are highly responsive to canopy leaf exposure to UV-B.
Under normal solar fluence, UV-B damage to macromolecules is balanced by their subsequent repair or replacement. Sporadic ozone depletion results in local "ozone holes" and spikes in terrestrial UV-B exposure. These periodic, but unpredictable UV-B spikes increase intensity up to 10-fold in both the polar and temperate zones . Furthermore, the ozone shield against UV-B is not expected to stabilize at 1950 levels until ~2050 ; consequently, determining the molecular bases for acclimation to normal fluence and tolerance of higher fluence UV-B are important factors in sustaining crop yield as the world's population continues to increase.
Previously, we established that maize lines have different UV-B tolerance, primarily because higher flavonoid sunscreens are correlated with fewer stress responses [3, 4]. Additionally, high altitude (> 2000 m) landraces naturally exposed to greater UV-B exhibit higher UV-B tolerance because they have both higher flavonoids and greater chromatin remodeling capacity [4, 5]. Conversely, temperate maize with knockdowns in chromatin remodeling factors exhibit adult tissue hypersensitivity and seedling lethality after mild UV-B supplementation [5, 6]. These and studies on other plants implicate both metabolite and gene expression responses as critical for short-term acclimation to UV-B and as examples of plant adaptation to this environmental variable [7–9].
In a pilot experiment we discovered that shielded organs, such as leaves wrapped in UV-B filters and immature ears encased in the husk leaves, show transcriptome changes within an hour or two after canopy leaves receive UV-B . Because such systemic responses can impact yield by modulating ear or kernel growth, identifying signals produced in sunlit leaves that alter reproductive organs should elucidate how UV-B decreases plant yield beyond what is predicted from the modest impact of UV on photosynthesis. Now we report a 1 to 6 hour time course of transcriptome and proteome responses in irradiated leaves, shielded leaves, and immature ears to unravel the systemic physiological and developmental responses in exposed and shielded organs. In parallel, metabolic profiling was used to search for candidate signaling molecules by charting increases in irradiated and then shielded organs. Integrating the datasets, we determined whether the biosynthetic, sequestration, or degradative pathways for candidate signaling molecules are regulated, at least in part, by UV-B exposure effects on transcript or protein abundance levels.
Figure S1 in Additional file 1 shows the GO classification of these transcripts. Genes regulated at least 2-fold (Table S1 in Additional file 2) include 6 transcripts encoding enzymes in the flavonoid pathway and 3 transcripts for enzymes in DNA repair, confirming processes often uncovered during plant acclimation to UV-B. A myo-inositol-1-phosphate synthase transcript is down-regulated by UV-B as reported previously [3, 10]. Using a 1.5-fold cutoff, more gene classes are represented (Table S1 in Additional file 2): up-regulated types include ribosomal proteins and translation factors, enzymes for detoxification of reactive oxygen species, chromatin remodeling proteins, and transcription factors (TFs). Among the down-regulated types are photosynthetic enzymes, TFs, and enzymes in starch biosynthesis. The repression and induction of TFs as well as the specific metabolic processes altered by UV-B confirm previous reports [3, 6, 10].
Canopy leaf irradiation (CLI) plants were screened by polyester (PE) that absorbs UV-B with the first, top two, or top three leaves threaded through the filter to receive UV-B-irradiation from lamps covered with cellulose acetate (CA) to filter radiation < 290 nm (Figure 1). One leaf exposure caused a small net increase of 513 transcripts versus the NI control with a slightly higher number of transcripts turned on (1397) versus off (884), plus an additional 1181 transcripts were either up- or down-regulated. In contrast, two and three leaf exposures resulted in a decrease of over 2500 transcripts and 3400 transcripts, respectively, while the total up- and down-regulated genes approximately doubled with each additional exposed leaf (2097 and 4039). Two or three exposed leaves similarly impacted shielded tissues (Figure 2a and 2b). In a time course, total transcript diversity in irradiated leaves decreased by > 5000 transcript types in 1 h compared to the no-UV-B (NI) control leaves, increased from 1 to 4 h, and then declined sharply at 6 h (Figure 2a and 2c), reinforcing previous observations that distinctive responses occur after 4 h . In shielded leaves, transcript diversity dropped by 4000 transcript types in 1 h and then increased over the time course to restore 94% of transcript diversity at 6 h. Ears have higher transcriptome complexity than leaves (Figure 2b) and sustain this through 1 and 2 h exposures; at 4 h, two and three leaf CLI decreases transcript diversity similarly (~2000) and then transcript diversity is decreased further from 4 to 6 h (Figure 2d). In summary, transcript diversity decreases > 10% in irradiated and shielded tissues, and each tissue type is distinctive over a 1 to 6 h time course.
In Figure 3a comparing all 3 CLI treatments with a 2-fold cutoff, 41% of transcripts (170 of 416) are commonly UV-B-regulated. Within these, all 6 genes in the flavonoid pathway (Table S3 in Additional file 2) and a transcript for a putative DNA repair enzyme, RadA-like protein, are up-regulated. We infer that mechanisms for avoidance and repair of UV-B damage act within exposed leaves irrespective of the irradiated area. In addition, two cysteine proteases and 17 TFs are induced after WPI and all CLI treatments (Table S4 in Additional file 2). As in the WPI results, the myo-inositol-1-phosphate synthase transcript shows a decreased level for all three CLI treatments; myo-inositol-1-phosphate may play a role in programmed cell death in Arabidopsis, however, in UV-B-treated maize it may be part of a metabolite signaling pathway . Eight HSPs are down-regulated in all three CLI treatments (Table S3 in Additional file 2).
Two transcripts encoding maize homologues of UVR8, a UV-B-specific signaling component that orchestrates expression of vital UV-protective functions in Arabidopsis [9, 14], are down-regulated in all irradiated maize leaves. Although UVR8 has not been reported to be UV-B-regulated, our inspection of Arabidopsis microarray data with Genevestigator© uncovered down-regulation of UVR8 by UV-B (data not shown).
PE-covered leaves and immature ears should respond to UV-B radiation only if exposed leaves transmit a signal(s). Shielded leaf samples were collected from the two leaves immediately below the PE filter after 4 h irradiation. We consider the shielded leaves of these mature plants (leaf expansion completed) to be physiologically similar to irradiated leaves. As shown in Figure 3b, at the 2-fold cutoff, shielded leaves from single leaf CLI plants show the fewest (132) transcript changes, with 73% of them shared with the 2 and 3 leaf treatments. Sixty one transcript alterations (12% of the total) are shared by all shielded leaf samples and represent core responses triggered by UV-B perception. Among these, 25 are increased including 4 DNA binding proteins of which at least two are TFs (Table S4 in Additional file 2). Interestingly, a serine:threonine kinase and a cysteine protease are also increased in both irradiated and shielded samples (Tables S1 and S4 in Additional file 2). UVR8 and myo-inositol-1-phosphate synthase are among the 36 down-regulated transcripts (Table S4 in Additional file 2) shared with directly irradiated leaves. Although a number of HSPs have been reported to be up-regulated by UV-B , a decrease in the levels of HSPs 82, 101 and clpB have been reported in the W23 maize inbred , and here we find two HSPs decreased in shielded organs (Table S4 in Additional file 2). Using a 1.5-fold cut-off, 120 transcripts levels were altered in shielded leaves from 4 h single leaf CLI treatments, increasing slightly to 134 for 2 leaves (Table S4 in Additional file 2). Using this lower cut-off, genes for stress responses, a hypersensitive-induced response protein, and some TFs are increased. Down-regulated genes include additional HSPs, nitrate transporters, and TFs (Table S4 in Additional file 2). Of the 416 transcripts changed more than 2-fold in WPI plants, 53 were also differentially expressed in shielded leaves on 1 or 2 leaf CLI plants (Figure 3c, center element).
In summary, after 4 h UV-B exposure of canopy leaves, shielded leaves have 521 modulated transcripts. A subset of altered transcripts is shared with WPI leaves. We hypothesize that signals from currently UV-B-irradiated leaves are transmitted to currently shielded leaves where some acclimation factors are regulated similarly to irradiated leaves. These triggered responses may contribute to leaf acclimation when sunlight reaches currently shaded leaves.
Despite the parallels in affected processes, few transcripts were shared between immature ears and shielded leaves from the 2-leaf CLI treatment: 9 transcripts at a 2-fold cutoff (1.5% of total regulated transcripts in both conditions) and 52 transcripts (2.4%) at a 1.5-fold cutoff (Figure S2 in Additional file 3 and Table S6 in Additional file 2). These low values are within the false discovery range in assigning expressed transcripts, and we conclude that shielded ears and leaves express distinctive gene suites after CLI.
In shielded leaves, the total number of transcripts changed after 1 h of UV-B was lower than in irradiated leaves (348 transcripts, Figure 5b); 270 of these changes persist at 2 h (central element, Figure S3b in Additional file 4). The 78 uniquely modulated after 1 h (Figure S3b in Additional file 4) are candidate components of initial UV-B signaling in shielded tissues; this group includes transcripts for signal transduction proteins, TFs, and transporters (Table S7 in Additional file 2). A UVR8 homolog is up-regulated in shielded leaves after 1 h of UV-B but not at longer UV-B exposures nor in any irradiated samples (Table S7 in Additional file 2). This transcript may be induced at earlier times in irradiated leaves as an initial response to UV-B and then in shielded leaves with a time delay after transmission of an inducing signal. Thus, the down-regulation of UVR8 measured at longer UV-B exposures (Table S7 in Additional file 2) may be a feedback response to the treatment. Comparing 1 h irradiated versus shielded leaves, 106 transcripts are changed in a similar way at least 2-fold (p < 0.05) in both conditions: 56 are increased, including transcripts for signaling proteins and TFs, and 50 decreased (Table S7 in Additional file 2). A wound and phytochrome signaling involved receptor-like kinase is increased in both irradiated and shielded leaves after 1 h, emphasizing its potential as a general participant in UV-B signaling.
After 6 h of UV-B, more transcripts increased significantly in both irradiated (1584, 2-fold, p < 0.05) and shielded (1280, 2-fold, p < 0.05) leaves. This is a 2.5-fold increase with respect to the 1 h treatment for irradiated and a 3.7-fold increase for shielded leaves (Figure 5a and 5b). Fewer than 20% of these transcripts are shared with the 4 h treatment in irradiated (19%) and shielded (17%) leaves (central element, Figure S3a and b in Additional file 4). This observation reinforces the distinction between initial (up to 4 h) and later responses of 6 h or longer . Genes in this category are described in Table S7 in Additional file 2.
Examining the time course for immature ears (Figure 5c) we find fewer total transcripts changed 2-fold (p < 0.05) by UV-B than in leaves; ears retain transcript diversity at 1 and 2 h because there are more up- than down-regulated genes. In ears, 82% of altered transcripts at 1 h (500) are also changed after 2 h of UV-B (Figure S3c in Additional file 4). Similar to shielded leaves, the 1 h CLI treatment induced transporters, TFs and signaling proteins (Table S8 in Additional file 2). Surprisingly, transcripts for multiple HSPs are transiently induced, and then down-regulated as a class after 4 h UV-B (Table S8 in Additional file 2); thus, as suggested for UVR8 in shielded leaves, the down-regulation observed at longer exposure times may be part of a feedback response to suppress physiological changes induced by initial UV-B perception.
Some early-induced transcripts remain differentially regulated throughout the time course: 82 in irradiated leaves, 19 in shielded leaves, and 103 in immature ears (Tables S7 and S8 in Additional file 2); these probably correspond to transcripts encoding proteins necessary for stress acclimation. Up-regulated examples for irradiated leaves include three phenylpropanoid pathway genes, a peroxidase PXC2, TFs of the WRKY type, and a PR protein (Table S7 in Additional file 2). In immature ears, 86 transcripts are induced > 2-fold at all time points (p < 0.05), including two phenylalanine ammonia-lyases, a peroxidase 13, and GSTs 20 and 22. Also up-regulated are transcripts encoding enzymes in cell wall degradation (Table S8 in Additional file 2). The up-regulation of these genes is specific for immature ears; it is possible UV-B induces some specific cell wall reorganization in these organs.
The impact of UV-B on fourteen major cellular processes was assessed by GO classification of transcripts from four expression types (those that were turned on or off or that were up- or down-regulated versus the NI control) for each of the 4 time points in the 6 h time course experiment (Figure S4 in Additional file 5). UV-B perception and systemic signaling to shielded organs has a major impact in all fourteen GO categories. The proportion of genes in each of the four expression types is distinctive for irradiated leaf, shielded leaf, and immature ears. For each of the fourteen categories and for each of the 4 time points, responses in the two types of leaf samples share more similarities than either does with the immature ear.
As a first step in identifying potential signal molecules moving from irradiated leaves to shielded organs, we conducted metabolic profiling using GC-MS to find metabolites altered by a 4 h UV-B treatment. We identified 84 compounds, 30 of which had a statistically significant change (Figure S5 in Additional file 6, one way ANOVA). Levels of six metabolites were only increased under WPI treatment (valine, serine, asparagine, glycerate, hexa/heptadecanoic acid and an unidentified metabolite, Additional file 6). Levels of thirteen metabolites were changed by UV-B in both WPI and CLI treatments, whereas altered levels of the following were restricted to irradiated leaves: alanine, fructose, galactose, glucose, glucaric/galactaric acid, xylose, dopamine, dihydroascorbic acid dimer, 2-ketoglutaric acid, and three compounds in the phenylpropanoid pathway (shikimic acid, quinic acid, and trans-caffeoylquinic acid) (Additional file 6). All six UV-B-regulated genes in the flavonoid pathway show increased levels exclusively in directly exposed leaves (Tables S3 and S4 in Additional file 2), suggesting that these metabolites are not translocated to shielded tissues nor do mobile signals induce them in shielded organs.
Because a signaling metabolite(s) must increase quickly in irradiated leaves to trigger transcriptome changes in shielded organs within an hour, we predicted that such molecules would show 1) high concentrations in treated leaves relative to untreated plants and 2) increases in shielded organs. In the first of 2 protocols to test these criteria, one adult leaf per plant was covered with a PE filter; a second leaf was covered with CA (allows UV-B transmittance) as a control for differences in temperature and humidity inside the sheath. After 1 and 6 h of UV-B-irradiation, leaf metabolites were compared; the PE-covered leaf should respond to UV-B only if there is a signal transmitted from exposed leaves. We also compared metabolites from PE-covered leaves in plants exposed to UV-B to those from PE-covered leaves in unirradiated plants; only the PE-covered leaf on an irradiated plant should exhibit transcript changes. Metabolites were also analyzed after WPI for 1 and 6 h and in control plants without UV-B. Figure S6 in Additional file 7 shows 27 metabolites changed by UV-B in at least one comparison. At 1 h, 26 of these show altered levels versus control plants; 9 are only changed in irradiated leaves (e.g. phosphate and fructose), while 17 are also changed in shielded leaves (e.g. shikimic, trans-caffeic, malic) (Additional file 7). Thus, metabolite changes do occur with 1 h UV-B and two-thirds of these candidate signal molecules are also detected in shielded tissues. Of the 26 metabolites that are specifically induced by UV-B, 8 show no difference to controls at 6 h whereas 11 are still altered in shielded leaves (Additional file 7).
The two complementary protocols generated a list of metabolites changed by UV-B in irradiated leaves and either translocated to shielded leaves or elevated there by an unknown signal. One or more of these may act as signaling compounds in shielded organ responses to UV-B in maize.
Only a subset of the changes monitored in 4 h WPI leaves were also detected in the CLI shielded leaves: 68% of the protein changes were measured in shielded leaves with 3 leaves irradiated, 58% with 2 leaves, and only 3% with a single leaf (Table S9 in Additional file 2). One kinase-like protein is increased by UV-B in shielded leaves from CLI plants; this candidate protein for UV-B signaling (Table S9 in Additional file 2) is sensitive to even low UV-B doses.
To analyze the kinetics of proteome responses to UV-B, the two leaf CLI samples from the 1, 2 or 6 h time course of irradiated and shielded leaves were analyzed along with the data from the 4 h exposure. After only 1 h of UV-B, 18 protein spots are increased and 42 are decreased by 1.5-fold (p < 0.05) in irradiated leaves; this corresponds to 92% of the total proteins changed after 4 h of UV-B exposure. Ten of the increased spots and 43 of the decreased spots are also changed in shielded leaves (Table S9 in Additional file 2). With most protein changes occurring within 1 h, it is likely that many correspond to post-translational modifications, or, in the case of decreases, protein degradation. For example, most of the decreased proteins correspond to photosynthetic proteins, particularly photosystem II components; degradation of these proteins after UV-B has been previously reported [21, 22]. Many protein changes persist over the duration tested; 52 of the 65 proteins changed after 4 h of UV-B are detectably altered at 6 h (Table S9 in Additional file 2).
In ears, 11 protein spots showed increased and 59 showed decreased levels after UV-B-irradiation (Table S10 in Additional file 2); most alterations occurred after both WPI and CLI treatments (Table S10 in Additional file 2), indicating that even modest UV-B exposure triggers proteomic changes in this reproductive organ. Plant processes highlighted by differential protein accumulation are summarized in Figure 8b; classes include enzymes of primary metabolism (24%), proteins in stress responses (18.6%), HSPs (7%), transcriptional regulation and signal transduction (5.7%) and other functions (4.2%); 35% correspond to proteins with unknown function. When analyzing the time course of responses, 73% of the protein changes measured occurred within 1 h of UV-B, while 76% of the total remains changed after 6 h (Table S10 in Additional file 2). As seen with shielded leaves, most protein changes occur quickly and a subset persist throughout longer exposures. It is interesting to note that as shown in Figures 8a and 8b, proteins changed by UV-B vary by organ, similar to transcriptome profiling observations (Additional file 8). As identified by transcriptome profiling and confirmed in the proteome, many HSPs show decreased levels in immature ears (Table S10 in Additional file 2).
Our focus was to identify systemic responses after UV-B-irradiation that elicit physiological changes in shielded organs. We also sought to correlate transcriptome changes to alterations in the proteome and metabolome both as cross-validation and to extract a more meaningful analysis of maize responses to UV-B. The first question addressed was how many canopy leaves must receive UV-B to alter transcripts in shielded leaves and ears. Assessing total transcripts, we found that exposure of just the top leaf substantially alters transcriptomes in irradiated leaf and shielded organs with greater changes as additional leaves are irradiated. Because the magnitude of responses was similar for two or three leaves, we conclude that exposure of only 10% of leaf area in adult maize plants suffices to elicit most responses (61%) with further responses (39% of total) requiring WPI (Additional file 1).
Transcript types specific to irradiated leaves include several genes in the flavonoid pathway, and concomitantly some phenylpropanoid precursors, such as shikimic, quinic, and trans-caffeoylquinic acids are increased by UV-B only in irradiated leaves (Additional file 6). Changes in other phenylpropanoid precursors, including cinnamic and trans-caffeic acids, occur in both irradiated and shielded leaves as does the increase in an isoflavone reductase-like1 protein (Table S9 in Additional file 2). Consequently, synthesis of specific secondary metabolites appears to result from both shared and distinctive regulation in shielded and irradiated tissues.
A number of transcripts that encode genes that participate in cell wall metabolism are induced by UV-B (3-deoxy-D-arabino heptulosonate-7-phosphate synthase, hydroxycinnamoyl CoA quinate transferase, UDP-glucuronosyltransferase; Table S2 in Additional file 2); these also correlate with the changes in phenylpropanoid precursors. Although we did not analyze metabolites in immature ears in this study, it is interesting that a number of transcripts that encode enzymes in cell wall metabolism are also changed by UV-B in these organs (cinnamyl alcohol dehydrogenase, cinnamoyl CoA reductase, cellulose synthase, beta-fructofuranosidase, etc.; Table S5 in Additional file 2). Thus, changes in cell wall structure may be a general acclimation response to UV-B in maize.
Transcripts for a myoinositol-1-phosphate synthase show complex regulation during UV-B responses. After 1 and 2 h treatments, myo-inositol-1-phosphate synthase (GRMZM2G155242, Table S7 in Additional file 2) transcripts increase in both irradiated and shielded leaves, but then decrease at 4 h. Myoinositol is synthesized from glucose-6-P, which is converted to myoinositol-1-phosphate by myoinositol-1-phosphate synthase, and this compound is dephosphorylated to produce myoinositol. The step catalyzed by the synthase is rate-limiting for myoinositol biosynthesis in plants [23, 24]. Myoinositol is a candidate UV-B signaling compound, because it is rapidly increased in irradiated and shielded leaves. Rapid synthesis after brief exposure is expected for a signal, while longer irradiation (4 h or longer) could provoke a down-regulation to modulate metabolite levels. Myoinositol has been proposed as important in stress protection , and may also regulate programmed cell death in Arabidopsis. Mutants in one myoinositol-1-phosphate synthase (AtIPS1) exhibit accelerated cell death ; the encoded protein has a nuclear localization sequence, suggesting that nuclear pools of myoinositol may be critical .
UVR8 is a UV-B-specific signaling component in Arabidopsis that mediates low fluence photomorphogenic responses, and it is required for UV-B induced expression of the gene encoding the HY5 transcription factor in co-operation with COP1 [8, 9]. UVR8 and COP1 interact directly and rapidly in the nucleus in planta after UV-B exposure , and this very early step in UV-B signaling is proposed to initiate UV-B acclimation . The action spectrum for the induction of HY5 transcripts by UVR8 has a maximum near 280 nm with significant action at longer UV-B wavelengths; UVR8 may act as a plant photoreceptor that can mediate UV-B-specific responses . Although Arabidopsis UVR8 has not been reported to be UV-B-regulated, interrogating microarray data with Genevestigator© shows that UVR8 is down-regulated by UV-B (data not shown), in agreement with our maize results. In contrast, after 1 h canopy exposure, shielded maize leaves show up-regulation of UVR8 (Table S7 in Additional file 2). We speculate that if UVR8 is a UV-B sensor in maize, it may be induced very quickly in irradiated leaves and then down-regulated within 1 h. Systemic signaling results in up-regulation in shielded organs after a delay, a step that may be an acclimation to increase subsequent sensitivity to UV-B
Shielded immature ears are highly responsive to canopy irradiation. Interestingly, although the classes of genes affected in shielded leaves and ears are similar, different genes are regulated. An exception is that 15 nucleic acid binding proteins, including TFs, are UV-B-regulated in a similar manner in these shielded organs (Table S6 in Additional file 2). We propose that UV-B signaling from irradiated to shielded organs elicits common regulation of some TFs to modulate similar processes via organ-specific gene family members. Other transcripts that are similarly regulated in shielded organs encode four proteins in signal transduction (Table S6 in Additional file 2), and these are candidates for processing the UV-B signaling from irradiated to shielded tissues. Moreover, both shielded leaves and immature ears show down-regulation of HSPs, a result validated by the proteomic analysis (Tables S4, S5 and S10 in Additional file 2).
In a two leaf CLI time course, there were significant transcriptome, proteome, and metabolome changes under all conditions studied in both shielded and irradiated organs. Some required only 1 h of exposure, defining likely responses in early UV-B signaling. One transcript up-regulated after 1 h is similar to Arabidopsis PFT1, a subunit of the Mediator complex. PFT1 was first described as a positive regulator of shade avoidance in Arabidopsis thaliana, and it was hypothesized that it could act downstream of phytochrome B to promote flowering in response to shade ; however, PFT1 is now considered to be a gene that negatively regulates the phytochrome signaling pathway . In addition, PFT1 is a positive regulator of jasmonic acid signaling during fungal pathogen infection . We hypothesize that PFT1 acts in early UV-B signaling in irradiated maize leaves, perhaps contributing to the dramatic and rapid reduction in transcript diversity.
Shielded leaf transcriptomes are also dramatically remodeled during UV-B responses, although to a lesser extent than in directly irradiated leaves. There are 106 responses in common, even with 1 h exposure, including a wound and phytochrome signaling involved receptor-like kinase. Transcripts are increased both in irradiated and shielded leaves after 1 h, emphasizing its potential as a general participant in UV-B signaling. In immature ears, there are fewer total transcripts changed by UV-B than in leaves. The shielded leaf blades are within 10 cm of the irradiated canopy, whereas the immature ears are ~30 cm distant. The delayed decrease in transcriptome diversity in ears most likely reflects the longer time required for signals to reach this organ. Although the focus was on early responses, after 6 h of UV-B exposure we identified a number of newly induced transcripts. For example, up-regulated transcripts correspond to enzymes in the phenylpropanoid pathway for the synthesis of flavonoids and cell wall, stress responses, TFs of the WRKY type, and PR proteins, indicating that there are continuing acclimation responses for the duration of UV-B treatments.
Our criteria for identifying candidate signaling metabolite(s) were that the compound must increase quickly in irradiated leaves relative to untreated plants and must also increase in shielded organs. After 1 h of UV-B in two irradiation protocols, 20 metabolites met these criteria (Figure 7 and Additional file 8). In the future, pharmacological and genetic knockout strategies can be used to test the relevance of these candidate molecules to UV-B systemic signaling.
Based on 2D gels, 65 and 70 protein spots were changed in leaves and immature ears, respectively. Consequently, even a small canopy exposure suffices to modulate some abundant cellular proteins; 27% of these proteins were changed when only one leaf was UV-B-irradiated. The proteomics assessment validates the microarray results in demonstrating that for those responses detected in CLI protocols, two leaves are sufficient to elicit most proteome changes. The leaf proteome after WPI shares only a subset of responses with shielded leaves in the CLI protocols after 4 h irradiation. One kinase-like protein is increased by UV-B in shielded leaves even with just one leaf exposure; this may be another candidate protein in UV-B signaling. Another key observation is that 92% of the proteins changed after 4 h in leaves and 73% in immature ears were also changed after 1 h of UV-B. Thus, most protein changes detected occur quickly. For novel spots, post-translational modification is likely; previously, we identified pyruvate phosphate dikinase to be phosphorylated after UV-B of WPI , and this protein was also identified in multiple spots in irradiated and shielded leaf samples. For decreased protein abundance, rapid proteolysis must be triggered by UV-B.
Most identified proteins changed by UV-B correspond to abundant proteins, such as photosynthetic proteins in leaves (Table S9 in Additional file 2), enzymes participating in general metabolism, and heat shock and 14-3-3-like proteins (Tables S9 and S10 in Additional file 2). In this group of highly abundant proteins, we found a good correlation between microarray and proteomic results. For example, in leaves, UV-B induces both transcript and protein levels of ferredoxin, pyruvate phosphate dikinase and fructose 1,6 biphosphate aldolase; while RuBisCO is decreased both at transcript and protein levels (Tables S1 and S9 in Additional file 2). A similar correlation exists for the down regulation of HSP70 at both transcript and protein levels in immature ears (Tables S5 and S10 in Additional file 2). For other genes, however, a correlation between protein and transcript level changes is only seen when different time points are compared, probably because new proteins require longer times to be produced. Examples of this temporally delayed correlation are oxygen-evolving enhancer protein, a chlorophyll a-b binding protein in leaves (Tables S1, S7, and S9 in Additional file 2), and an inorganic phosphatase (Tables S5 and S10 in Additional file 2). Finally, a few proteins showed opposite UV-B regulation compared to transcript changes; this is the case in leaves for 50S ribosomal protein L12 (down regulation at the transcript level, up regulation at the protein level) and glycosyl transferase family 8 (up regulation at the transcript level, down regulation at the protein level; Tables S1 and S9 in Additional file 2). With respect to metabolite changes, it is important to note that changes measured at the transcript and proteome level of photosynthetic proteins and enzymes in primary metabolism, are reflected in changes in the total pools of soluble metabolites. For example, pools of molecules that are intermediates or products of glycolysis, the Krebs cycle and fermentation pathways (lactic acid, succinic acid, glycerate, 2-ketoglucaric acid, etc.) are changed by UV-B as a consequence of metabolic pathway fluxes (Additional files 6, 7 and 8). In addition, soluble sugars, products of photosynthesis are also significantly changed after the UV-B treatments (glucose, fructose, galactose, etc.; Additional files 6, 7 and 8).
W23 maize was grown for 5 weeks in the greenhouse during July and August 2008 using the same protocol as described previously . The afternoon prior to treatment, 12 plants were moved underneath UV lamps, and the topmost leaves (either 1, 2 or 3) were threaded through slits in PE plastic; after acclimating overnight, these leaves received 1, 2, 4, or 6 h UV-B exposure . Exposure times were centered on 11 am, 5h after sunrise and supplemental greenhouse lighting (33% of summer noon solar fluence). Two different controls were utilized: an untreated (no irradiation treatment, NI) and whole plant irradiation (WPI); these plants were moved into the irradiation apparatus as described above to minimize detection of any gene expression response resulting from moving the plants into position in the grid the afternoon prior. The four hour, 2-leaf irradiated treatment was also included as a linking control between the two types of experiments. A biological replicate consisted of tissue samples pooled from 4 plants. Four biological replicates for each treatment (a total of 16 plants) were used in all assays, except for the WPI and 6 h treatment, where only two biological replicates were used in the microarray experiments (Figure 1a). The three sample types (irradiated leaf, shielded leaf, and immature ear) were immediately flash-frozen in liquid nitrogen after harvesting. Irradiated leaf CLI samples were collected from each irradiated leaf while WPI leaf samples were taken from the top 2 canopy leaves. Shielded leaf samples were taken from the 2 leaves immediately below the PE plastic. Because of space limitations in the UV-B irradiation apparatus (Figure 1b), only 12 plants could be processed at a time. Therefore one biological replicate for each sample and treatment type was always done on a subsequent day, but always during the same time period. Leaves were harvested by one person while two researchers dissected and harvested immature ears. Irradiated leaves were cut before cutting the plants and taking them to the work area, where ears and both leaf sample types were harvested and flash frozen. Total time to harvest all samples was approximately 10 min; leaves were harvested within about 5 min and ears took about 10 min. Leaves were separated from the midrib before being frozen. Immature ears (1-6 cm) were cut in half and distributed randomly among the different collection tubes for RNA or proteomics analysis.
RNA extraction and microarray hybridization were done as described in Casati and Walbot . Data acquisition, image processing, and spot flagging and removal were performed as described in Skibbe et al . The median foreground values for each channel from the Agilent Feature Extraction software were first normalized using the lowess method of the limma package in R  (within each array) and then using limma's quantile method (between all arrays). Probes were classified as "on" if their expression value was more than 3.0 standard deviations above the average foreground intensity of the Agilent negative controls, providing a 0.13% FDR (~ 50 probes). For differential expression, we used the unadjusted p-value generated by limma for FDR. Differentially expressed probes were identified using either a 1.5 or 2 fold cutoff for expression ratios with a limma-assigned p-value < 0.05. Probes were included in the analysis if at least 75% of the replicate expression values (i.e. 6 of 8, 3 of 4, or 2 of 2) were classified as "on". Microarray data were deposited in GEO under ID GSE25038.
Protein extraction, labeling, 2D gel electrophoresis, gel image analysis, MS, and database search were done as described in Falcone-Ferreyra et al. .
Extraction, liquid partition, and derivation prior to GC-MS analysis were performed as described by Lisec et al. . For analysis, four biological replicates per treatment with a second group of technical replicates were utilized (8 total data points used during analysis). GC-MS analysis was performed using an autosystem XL Gas Chromatograph and a Turbo Mass Spectrometer (Perkin Elmer) in the Facultad de Ciencias Bioquímicas y Farmacéuticas - UNR facilities. One μL split injection (split ratio 1:40) was injected at 280°C. The capillary column used was aVF-5ms column (Varian, Darmstadt, Germany) with the following dimensions: 30 m × 0.25 mm inner diameter and a 0.25 μm film with helium as carrier gas with constant flow at 1 mL/min. The temperature program was 5 min at 70ºC, 5 min ramp to 310ºC and final heating for 2 min at 310ºC. The transfer line to the MS was set to 280ºC. Spectra were monitored in the mass range m/z = 70-600. Tuning and all other settings were according to manufacturer's recommendations.
Chromatograms were acquired with TurboMass 4.1 software (Perkin Elmer). The NIST98mass spectral search program (http://www.nist.gov/srd/mslist.htm, National Institute of Standards and Technology, Gaithersburg, MD, USA) was the software platform. The MS and retention time index were compared with the collection of the Golm Metabolome Database [36, 37]. MS matching was manually supervised and matches accepted with thresholds of match > 650 (with maximum match equal to 1000) and retention index deviation < 1.0%. Peak heights were normalized using the amount of the sample fresh weight and ribitol for internal standardization. Relative metabolite contents were determined and statistical analyses were performed using ANOVA tests in Sigma Stat 3.1.
whole plant irradiation
Canopy leaf irradiation
heat shock proteins
phytochrome and flowering time 1 protein
tandem mass spectrometry.
We thank Lorena Falcone-Ferreyra and Sebastián Rius for manuscript critique, and Mónica Hourcade for GC/MS technical support. This project was supported by the National Research Initiative or Agriculture and Food Research Initiative Competitive Grants Program grant no. 2008-35100-04578 to V.W. from the USDA National Institute of Food and Agriculture and by FONCyT grants PICT-2006-00957 and PICT-2007-00711 to P.C. P.C. is a member of the Researcher Career and M.C. is a doctoral fellow of CONICET.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.