Transcription factor-dependent chromatin remodeling at heat shock and copper-responsive promoters in Chlamydomonas reinhardtii

Abstract

How transcription factors affect chromatin structure to regulate gene expression in response to changes in environmental conditions is poorly understood in the green lineage. To shed light on this issue, we used chromatin immunoprecipitation and formaldehyde-assisted isolation of regulatory elements to investigate the chromatin structure at target genes of HSF1 and CRR1, key transcriptional regulators of the heat shock and copper starvation responses, respectively, in the unicellular green alga Chlamydomonas reinhardtii. Generally, we detected lower nucleosome occupancy, higher levels of histone H3/4 acetylation, and lower levels of histone H3 Lys 4 (H3K4) monomethylation at promoter regions of active genes compared with inactive promoters and transcribed and intergenic regions. Specifically, we find that activated HSF1 and CRR1 transcription factors mediate the acetylation of histones H3/4, nucleosome eviction, remodeling of the H3K4 mono- and dimethylation marks, and transcription initiation/elongation. By this, HSF1 and CRR1 quite individually remodel and activate target promoters that may be inactive and embedded into closed chromatin (HSP22F/CYC6) or weakly active and embedded into partially opened (CPX1) or completely opened chromatin (HSP70A/CRD1). We also observed HSF1-independent histone H3/4 deacetylation at the RBCS2 promoter after heat shock, suggesting interplay of specific and presumably more generally acting factors to adapt gene expression to the new requirements of a changing environment.

Figure 1.

Analysis of Protein and Transcript Levels in HSF1-Underexpressing Strains Prior to ChIP Analysis. (A) HSF1 abundance is reduced in HSF1-amiRNA and -RNAi strains. Control, HSF1-amiRNA, and HSF1-RNAi lines were kept under nonstress conditions (CL) or subjected to heat shock (HS) for 30 min. Whole-cell proteins were extracted, and proteins corresponding to 2 μg chlorophyll were separated by SDS-PAGE and analyzed by immunoblotting using antisera against HSF1 and CF1β (as loading control). (B) Accumulation of selected transcripts in control and HSF1-RNAi/amiRNA cells. RNA was extracted from nonstressed cells and cells subjected to a 30-min heat shock for analysis by qRT-PCR using the comparative CT method with CBLP2 as control gene. Primer efficiencies and qRT-PCR end products (amplicons) for all target transcripts are presented in Supplemental Figure 3 online. Shown are fold changes in transcript accumulation between stressed versus nonstressed conditions in control (black) and HSF1-RNAi/amiRNA lines (white), respectively. Three technical replicates each from HSF1-RNAi line #10 (triangles) and HSF1-amiRNA line #5 (diamonds) were performed.

Figure 2.

Regions Amplified from Chromatin Immunoprecipitates by qPCR. Shown are the six genes investigated in this study. Promoter regions are indicated by gray boxes, transcriptional start sites (TS) by arrows, translated regions by black boxes, untranslated regions by white boxes, and introns by thin lines. The HSP70A promoter has two transcriptional start sites designated TSA1 and TSA2 (von Gromoff et al., 2006). Vertical black lines designate putative HSEs in the HSP70A and HSP22F promoter regions and putative CuREs in the CYC6, CPX1, and CRD1 promoters. Gray bars designate the regions amplified by qPCR (if not indicated otherwise, region I was used by default for HSP22F and CYC6).

Figure 3.

HSF1 Binds to Promoters HSP70A and HSP22F. ChIP was done on control (black bars) and HSF1-underexpressing strains (gray bars) grown under nonstress conditions or subjected to a 30-min heat shock. From DNA fragments precipitated with αHSF1 antibodies the promoter regions shown in Figure 2 were amplified by qPCR. The enrichment relative to 10% input DNA was calculated and normalized to the values obtained for the CYC6 promoter. Error bars indicate standard errors of the mean of two biological replicates, with each analyzed in triplicate. HSF1-underexpressing strains analyzed were HSF1-RNAi line #10 and HSF1-amiRNA line #5. Asterisks indicate the significance of change compared with the CYC6 promoter in control cells under nonstress conditions (t test, P value ≤ 0.01).

Figure 4.

Analysis of Nucleosome Occupancy and Histone Modifications at Heat Shock–Responsive and Control Promoters. (A) Nucleosome occupancy declines at heat shock gene promoters after heat shock. ChIP was done as described in Figure 3 but using antibodies against the unmodified C terminus of histone H3 to determine nucleosome occupancy at the indicated promoters in control (black bars) and HSF1-underexpressing strains (gray bars). (B) HSF1 promotes acetylation of histone H3 at the HSP70A and HSP22F promoters. ChIP was done using antibodies against acetylated Lys-9 and -14 of histone H3. (C) HSF1 promotes acetylation of histone H4 at the HSP22F promoter. ChIP was done using antibodies against acetylated Lys-5, -8, -12, and -16 of histone H4. (D) HSF1 has no effect on histone H3 dimethylation at the heat shock gene promoters. ChIP was done using antibodies against dimethylated Lys-4 at histone 3 (H3K4). (E) HSF1 might reduce histone H3 monomethylation at the HSP22F promoter. ChIP was done using antibodies against monomethylated Lys-4 at histone 3 (H3K4). qPCR data from the experiments in (B) to (E) are given relative to the nucleosome occupancy at the respective promoter region (data from [A]). Error bars indicate standard errors of the mean of two biological replicates, with each analyzed in triplicate. Asterisks indicate the significance of change at the respective promoter compared with control cells under nonstress conditions (t test, P value ≤ 0.05).

Figure 5.

Analysis of the Sequence of Events at the HSP22F Promoter within the First 10 min after Onset of Heat Stress. (A) HSF1 binding precedes chromatin remodeling and transcription. Control cells were subjected to heat stress, and samples for RNA extraction and ChIP were taken immediately prior to the temperature shift and at the indicated time points after shift from 25 to 40°C. HSP22F mRNA levels were quantified by qRT-PCR as described in Figure 1B. Shown are fold changes in transcript accumulation relative to the nonstressed state. Values derive from two biological replicates, with each analyzed in triplicate. ChIP was done as described in Figure 3, again amplifying region I of the HSP22F promoter (Figure 2). The enrichment relative to 10% input DNA was calculated and normalized to the values obtained for the CYC6 promoter. Error bars indicate standard errors of two biological replicates, each analyzed in triplicate. (B) Graphical overview of the sequence of events at the HSP22F promoter after onset of heat stress. The data from (A) are given as percentage of the respective maximal values.

Figure 6.

Accumulation of Selected Transcripts in Control and crr1 Mutant Cells. Transcript accumulation in copper-replete versus copper-deprived control (black) and crr1 mutant cells (white) was assessed by qRT-PCR as described in Figure 1B. Values shown are from two biological replicates (triangles and diamonds), each analyzed in triplicate.

Figure 7.

Analysis of Nucleosome Occupancy and Histone Modifications at Copper-Responsive and Control Promoters. (A) Nucleosome occupancy declines at copper-responsive gene promoters under copper depletion. ChIP was done on control (black bars) and crr1 knockout cells (gray bars) grown under copper-replete or copper deprivation conditions. From DNA fragments precipitated with antibodies against the unmodified C terminus of histone H3, the promoter regions shown in Figure 2 were amplified by qPCR. The enrichment relative to 10% input DNA was calculated and normalized to the values obtained for the RBCS2 promoter. Error bars indicate standard errors of two biological replicates, each analyzed in triplicate. (B) CRR1 promotes histone H3 acetylation at the CYC6 and CPX1 promoters after copper depletion. ChIP was done using antibodies against acetylated Lys-9 and -14 of histone H3. (C) CRR1 promotes histone H4 acetylation at the CYC6 and CPX1 promoters after copper depletion. ChIP was done using antibodies against acetylated Lys-5, -8, -12, and -16 of histone H4. (D) CRR1 promotes reduction of H3K4 dimethylation at promoters CYC6 and CPX1 after copper depletion. ChIP was done using antibodies against dimethylation of Lys-4 at histone H3 (H3K4). (E) CRR1 promotes reduction of H3K4 monomethylation at promoters CYC6 and CPX1 after copper depletion. ChIP was done using antibodies against monomethylation of Lys-4 at histone H3 (H3K4). qPCR data from the experiments in (B) to (E) are given relative to the nucleosome occupancy at the respective promoter region (data from [A]). Asterisks indicate the significance of change at the respective promoter compared with control cells under copper replete conditions (t test, P value ≤ 0.01).

Figure 8.

FAIRE Indicates Chromatin Remodeling after Transcriptional Activation. (A) Nucleosome occupancy declines at heat shock gene promoters after heat shock. FAIRE was performed with control (black bars) and HSF1-underexpressing cells (gray bars) grown under nonstress conditions or subjected to a 30-min heat shock. DNA fragments present in the supernatant after phenol/chloroform extraction of formaldehyde-cross-linked chromatin were precipitated, and the promoter regions of HSP70A and HSP22F (Figure 2) were amplified by qPCR. The enrichment relative to input DNA, after reversion of the cross-link, was calculated. Error bars indicate standard errors from two biological replicates, each analyzed in duplicate. Asterisks indicate the significance of change at the respective promoter compared with control cells under nonstress conditions (t test, P value ≤ 0.05). (B) Nucleosome occupancy declines at the CYC6 promoter after copper depletion. FAIRE was performed with control and crr1 knockout cells grown in the presence or absence of copper. qPCR analyses were done as in (A). Asterisks indicate the significance of change at the respective promoter compared with control cells under copper replete conditions (t test, P value ≤ 0.01).

Figure 9.

Gene-Wide Overview of the Relative Abundance of Histone Occupancy and Modifications. ChIP was done on control cells grown under the following conditions: nonstress in copper-replete medium (CL or Cu+), 30 min heat shock (HS), and medium depleted from copper (Cu−). ChIP was done using antibodies against the unmodified C terminus of histone H3 (A), acetylated Lys-9 and -14 of histone H3 (B), acetylated Lys-5, -8, -12, and -16 of histone H4 (C), dimethylation of Lys-4 at histone H3 (H3K4me₂) (D), and monomethylation of Lys-4 at histone H3 (H3K4me₁) (E). Fragments corresponding to transcribed regions shown in Figure 2 and intergenic regions separating genes au5.g14265_t1/P23 (IGR-1) and RBCS2/au5.g9204_t1 (IGR-2) were amplified by qPCR. The enrichment relative to 10% input DNA was calculated and normalized to the values obtained for the CYC6 promoter (heat stress experiments) or the RBCS2 promoter (copper depletion experiments). The data on the promoter regions correspond to that shown in Figures 4 and 7. In case of ChIP analysis with antibodies against modified histones, an additional normalization was necessary due to different levels of modification within control promoters (RBCS2 and CYC6) that lead to different absolute values between the copper starvation response data set and the heat shock data set. Therefore, the maximum value of each data set was set to 100%. Error bars indicate standard errors of the mean of two biological replicates, each analyzed in triplicate.

Figure 10.

Schematic Description of How Transcription Factors Affect Chromatin State and Activity of the Promoters Studied. Promoters are schematically depicted by a transcription factor binding site, one nucleosome, and the transcriptional start site (TS). Histone modifications are given on top of the nucleosome, where ac_H3 stands for acetylation at H3K9 and H3K14; ac_H4 for acetylation at H4K5, H4K8, H4K12, and H4K16; me₁ for H3K4 monomethylation; and me₂ for H3K4 dimethylation. HSF1 is shown as constitutive trimer, CRR1 as monomer, and polymerase II as multiprotein complex. The darker the symbols for proteins and modifications are drawn, the higher their levels under the respective condition. As we have no data on the occupancy of the CuREs by CRR1, the latter is drawn in white.