The maize Pepc gene is highly active in the leaves of illuminated plants. The activity is only slightly reduced during a normal dark period 13 . The diurnal pattern of gene expression was unknown. We therefore studied the promoter activity (amount of unspliced heterogenous nuclear transcripts [hnRNA]) and mRNA accumulation during the day (Figure 1). Promoter activity was induced two- to threefold by illumination and reached its highest levels around 4 hours after illumination (hai). Afterwards, a constant decline down to a basal level was observed that remained almost constant during the night. mRNA accumulation followed a similar profile with an offset of approximately 4 h (Figure 1B). It should be noted that the lowest levels of Pepc promoter activity obtained in the natural diurnal rhythm are still at least as high as the activity of the constitutively expressed Actin-1 gene 15 and much higher than lowest Pepc promoter activities observed in re-etiolated leaves or roots (orange line, and such-like 14 ). We verified that heterogeneous nuclear RNA (hnRNA) measurements are representative of the transcription rate by comparing with the number of RNA polymerase II molecules on the coding region of the gene as determined by chromatin immunoprecipitation (Figure 1C). Results obtained by both methods were very similar, indicating that the hnRNA method is suitable for the detection of promoter activity. The data indicate that Pepc promoter activity varies around a high level during the day. We expected this variation to be accompanied by changes in histone acetylation and so, accordingly, tested the acetylation levels on the upstream and core promoter regions. Data for all sites of H3 N-terminal acetylation (H3K9, H3K14, H3K18, H3K23 and H3K27) and for the acetylation of lysine 5 and 16 on histone H4 at two typical positions of the upstream and core promoter, respectively, are shown in Figure 2. Additionally, detailed promoter acetylation profiles for each modification are available in the Additional File 1. An antibody to an invariant domain of histone H3 (H3C) was used to estimate the nucleosome densities on the gene 1 . Additional File 2 shows that profiles obtained with this antibody are very near to those obtained with a second independent antibody directed to an invariant domain of H4. Data were recorded at 4 hai (the highest diurnal promoter activity) and 16 hai (the lowest diurnal promoter activity). For comparison, data from re-etiolated plants are shown. We refer to differences between 4 hai and the re-etiolated state as a response to prolonged darkening and to differences between 4 hai and 16 hai as a diurnal response. The black line always indicates the levels obtained on the constitutively active Actin-1 reference gene.

As summarized in Figure 2A, the response of transcription to prolonged darkening was much stronger than the diurnal response. Nucleosome densities were almost constant on the upstream promoter and close to those observed on Actin-1 (Figure 2B). On the core promoter, the differences in nucleosome densities were obvious. In the re-etiolated state, where the gene shows minimal activity, nucleosome density was similar to Actin-1. A reduction of around 40% relative to this value was observed at the basal or induced activation, state. As a control, we also measured nucleosome densities on the Zein gene, which is exclusively expressed in developing seeds but never in leaves. Here, nucleosome densities are around 30% higher than on Actin-1. These data suggest that the Pepc promoter is always in an open chromatin configuration in leaves. Nucleosome density on the core promoter is even further reduced when the gene is fully activated.

Highly individual patterns, depending on the promoter position, were observed for the tested acetylation sites (Figure 2C). On the upstream promoter region, a good correlation of acetylation and transcription was evident for only three of the modifications (H3K9ac, H3K27ac, H4K5ac). For other modifications (H3K14ac, H3K18ac, H3K23ac, H4K16ac), the diurnal response was as strong, or even stronger, than the response to prolonged darkening. On the core promoter, more modification sites showed only weak or no responses to the stimuli applied. Where a light response was evident (H3K9ac, H3K23ac, H4K5ac, H4K16ac), the diurnal response was as strong as the response to prolonged darkening. Thus, the lowest acetylation levels for these sites were already observed at 16 hai, where significant gene transcription occurred. Remarkably, the lowest acetylation levels on Pepc are still in the range of those observed on the constitutively active Actin-1 gene. The data reveal a complex diurnal control of acetylation on both promoter regions. Only some of the acetylation sites on the upstream promoter showed a positive correlation of acetylation and transcription. Acetylation of all variable sites on the core promoter is reduced to minimum levels in the diurnal rhythm, although transcription remains high. Nucleosome density appears to respond mainly to stronger promoter inactivation down to the lowest levels observed in leaves. This is reminiscent of the light-induced enhancement in chromatin accessibility to the restriction enzymes that we observed in an earlier study 16 .

In order to test whether the decrease in histone acetylation during the illumination period controls the concomitant modest reduction in promoter activity, detached leaves were treated with the deacetylase inhibitor Trichostatin A (TSA). Figure 3B exemplarily shows the effect of this treatment on H4K5 acetylation. As expected, high levels of histone acetylation were induced in both promoter regions. The reduction in acetylation in the afternoon was almost completely eliminated. Identical results were obtained for all other acetylation sites (data not shown). However, the reduction in promoter activity between 4 hai and 16 hai was completely unaffected by this treatment (Figure 3A). This indicates that high histone acetylation is not sufficient to clamp the promoter on the highest activity levels.

Diurnal gene regulation is often affected by circadian oscillators that control the variation in gene transcription during the course of a day, independent of a day/night shift 17 . In plants, circadian control is typically tested by following the promoter activity through a prolonged period of constant illumination. We recorded Pepc promoter activity and mRNA accumulation through two daily cycles where dark periods were replaced by constant light (Figure 4). Throughout the experiment, oscillation of promoter activity could be clearly observed. The period of one cycle declined from 24 h to approximately 21 h during the constant light period. This is typical for circadian control of gene transcription 18 . Therefore, diurnal cycling of Pepc transcription showed all characteristics of circadian control.

Histone acetylation was recorded at six indicative time points during constant illumination. Data are shown for a typical position on the core and the upstream promoter, respectively (Figure 5). Results for seven additional promoter positions are displayed in Additional file 3. On the upstream promoter, all tested sites showed circadian oscillation of acetylation, albeit to different extents. The best correlation to transcription, with respect to amplitude and period of oscillation, was observed for H3K9ac. H3K18ac constituted an exception as oscillation was completely suppressed in the second cycle. Very weak oscillation was observed for the invariant control H3C position. On the core promoter, less circadian oscillation of acetylation was clearly observed. Most modifications already showed a clear decline in histone acetylation in the first cycle and no oscillation of acetylation was detectable during the second cycle. As observed in the diurnal rhythm, the lowest acetylation levels on Pepc were in the range of, or slightly lower than, those observed on the Actin-1 promoter used as a standard. Importantly, the highest transcription levels at the beginning of the second cycle were induced although core promoter acetylation levels remained unchanged. This indicates that the induction of the core promoter acetylation is not mechanistically required for high promoter activity.

We have previously shown that several histone acetylation sites on the core promoter of the Pepc gene in maize are constitutively acetylated in leaves but not in roots or other organs where the gene is always inactive 14 . Light specifically induces acetylation of H3K9, H4K5 14 , H3K23 and H4K16 (this study). Light-induced acetylation of H4K16 is in disagreement with our previous data. However, in earlier studies, leaves from plants 8 hai were investigated. Pepc transcription has already decreased at this time point (Figure 1) and, therefore, we might have missed the time point of maximum acetylation in previous experiments.

Our analysis of core promoter acetylation in the diurnal and circadian rhythm revealed that light-inducible acetylation is not required for high-level transcription. In the diurnal rhythm, all light-inducible acetylation sites were deacetylated during the second half of the day. Only the leaf-specific acetylation of H3K14, H3K18 and H3K27 remained stable under these conditions (Figure 2). Gene transcription also declined in the afternoon but remained at an intermediate level that was much higher than the lowest transcription levels found in re-etiolated leaves (Figure 1). Thus, light-inducible acetylation was actively removed from the core promoter during the light period with little impact on transcription. The uncoupling of transcription and acetylation was even more evident under free-running conditions (constant illumination). Particularly at the beginning of the second daily cycle, acetylation of all tested sites on the core promoter remained low but transcription was induced (Figure 5). This indicates that the Pepc gene can be activated to maximum transcription levels at low core promoter acetylation levels. The results are particularly significant because Pepc is not a weakly expressed gene, but belongs to the highest expressed genes in plants 9 . However, when compared to the constitutively active Actin-1 gene, the promoter acetylation at the lowest activation state of Pepc was still in the same range as Actin-1 (Figures 2C and 5). A similar pattern was observed for nucleosome densities (Figure 2B). This indicates that Pepc retains an open promoter chromatin structure even at minimal transcription levels. Thus, modulation of Pepc transcription in the circadian rhythm takes place in a euchromatic chromatin context, even if acetylation levels are extremely low with respect to Pepc.

Histone acetylation has been correlated, as well as mechanistically linked to transcription (see Background). However, in the circadian rhythm we observed the highest Pepc transcription levels at the lowest Pepc acetylation. At least two mechanisms for gene activation independent of histone acetylation have been described. Kato et al. have shown that the activity of histone chaperones can stimulate promoters which are, to a certain extent, independent of histone acetylation 19 . For enhancer elements, acetylation was dispensable for the activation of associated genes if the underlying DNA sequence contained a sufficiently strong enhancer element 20 . In such cases, remodelling complexes and transcription factors were directly recruited that would normally interact with acetylated histones via their bromodomain 21 . Both described mechanisms might help in the activation of the Pepc promoter in maize, particularly as transcription is not activated from the minimum level, but from an intermediate state (Figures 1 and 4). The data underscore our previous observations that specific stimuli induce individual histone modifications on the Pepc core promoter whether or not transcription is finally induced. Such a pattern is consistent with a bona fide histone code 8 . However, the regulation of acetylation in this promoter region seems to be more complex than anticipated, because diurnal and circadian factors impact on light-induced acetylation.

Acetylation of the tested lysine residues on the upstream promoter differed considerably from core promoter acetylation as it correlated well with transcription levels. This was almost always independent of the stimulus regulating the promoter. The only clear exception to this rule was light regulation of the upstream promoter acetylation. If the promoter was deactivated by extended dark periods, H3K14 and H3K18 acetylation remained high (Figure 2). The general correlation of the upstream promoter acetylation and transcription was also reproducible in diurnal and, mostly, in circadian regulation (Figures 2 and 5). Thus, a charge neutralization model (see Background) can describe most of the interdependence of upstream promoter acetylation and transcription. In such a simple scenario, artificial hyperacetylation should be sufficient to prevent the afternoon decrease in Pepc transcription during the diurnal cycle. However, TSA treatments did not affect the diurnal transcription rhythm, albeit inducing strong acetylation of all lysine residues investigated in this study (Figure 3 and data not shown). Both in mammals and plants, positive and negative examples exist for the regulation of gene transcription by TSA. MHC class II genes 22 or the PDK4 gene 23 in mammals can be induced by TSA treatments. In Arabidopsis, only few genes were affected by TSA when transcriptome-wide studies of either leaves or germinating seedlings were performed 24 25 . This might indicate that the pleiotropic effects of this treatment were masking gene-specific induction or that the acetylation level itself did not contain the relevant information for gene regulation. Consistent with the latter model, Clayton et al. suggested that high turn-over rates, rather than steady-state acetylation levels, are important for high transcription levels 26 . We conclude that, for the Pepc promoter, diurnal fluctuations in transcription are not controlled by the level of upstream promoter acetylation.

There is significant evidence from other systems that the diurnal and circadian regulation of transcription is controlled by changes in histone acetylation. In plants, the promoter of the clock component TOC1 is regulated by acetylation/deacetylation cycles that are controlled by the endogenous clock itself 18 27 . In mammals, the central clock component CLOCK is a histone acetyltranferase that binds in a multiprotein complex to promoters containing E-box DNA elements 28 . Interestingly, a G-box (the plant functional homologue to the mammalian E-box) is present in the upstream promoter at position -1186, but not in the core promoter of Pepc. It is therefore plausible to assume that upstream promoter elements are more important for the quantitative modulation of Pepc transcription levels than the core promoter. Rather, the latter might integrate information on basal factors such as illumination status and tissue specificity. This is in agreement with the observation that Pepc core promoter sequences were sufficient to control leaf-specific and light-dependent transcription when linked to reporter genes 29 30 .

The Pepc sequence was derived from GenBank accession gi 22396. BAC clone gi 116268332 was used to add 5' sequence elements. The Actin-1 promoter was described in 15 . The Zein gene sequence was derived from Genbank accession gi 535019 16 .

Maize (Zea mays cv. Montello) was cultivated in growth chambers with a 16-h photoperiod and a day/night temperature regime of 25°/20°C. The plants were illuminated with Osram Superstar HQI-T 400W/DH lamps at a photon flux density of 120-180 μmol m^-2 s^-1. Seedlings were grown in soil (VM, Einheitserde, Sinntal-Jossa, Germany) for 10-12 days. Re-etiolated plants were derived from illuminated plants that were subjected to darkness for 72 h. For constant light experiments, the dark period was replaced by continuous illumination with constant temperature.

After 1 h of illumination, 10-12-day-old leaves were detached under water 1 cm above the laminar joint and incubated for 3 h (4 hai) or 15 h (16 hai) in tap water containing 30 μM trichostatin A (TSA; International Clinical Services, Munich, Germany). To avoid nitrogen depletion, 0.275 μM trans-Zeatin (Sigma-Aldrich, Schnellendorf, Germany) and 16 mM KNO₃ were added 14 31 . Mock control leaves were incubated in tap water with Zeatin, KNO₃ and methanol.

As described previously, 6 g of foliar leaf material were harvested and crosslinked 13 14 . ChIP was performed as described by Bowler et al. 32 with modifications as described in Haring et al. 15 . Chromatin was sheared with a Bioruptor (Diagenode) for 10 min (setting: high, interval 30/30 s) under constant cooling. Modified histones were detected with 5 μl anti-acetyl H4K5 (Upstate 07-327 Millipore, Schwalbach, Germany), 5 μl anti-acetyl H4K16 (Upstate 07-329), 5 μl anti-acetyl H3K9 (Upstate 07-352), 1 μl anti-acetyl H3K14 (Upstate 07-353), 1 μl anti-acetyl H3K18 (Upstate 07-354), 5 μl anti-acetyl H3K23 (Upstate 07-355), 5 μl anti-acetyl H3K27 (Upstate, 07-360),1 μl anti H3 C-term (Abcam ab1791) and 7.5 μl H4 C-term (Abcam 10158). RNA polymerease II was detected with 7.5 μl of an antibody directed to the C-terminal domain repeat (Abcam 817). The control serum for determination of background precipitation was derived from rabbits immunized with an unrelated protein from potato. In general, background signals never exceeded 10% of positive signals.

RNA isolation and reverse transcription were performed as described 13 . As a direct estimate for promoter activity, hnRNAs were amplified from cDNA with a primer system specific for an intron 33 34 . A dilution series of illuminated leaf cDNA was used as a standard. Oligonucleotide sequences and conditions are given in Additional File 4.

Quantitative PCR was performed on an ABI PRISM 7300 (Applied Biosystems) using SYBR Green fluorescence (Platinum SYBR Green qPCR Mix, Invitrogen) for detection. Oligonucleotide sequences and conditions are given in Additional File 4.