Figure 1.
Low formaldehyde concentration and short fixation induce PAR formation.
(A) HeLaS3 cells were fixed with increasing formaldehyde concentrations and for indicated times. Formaldehyde concentrations are indicated on the left side, fixation time on top. DAPI = nucleus; PAR = poly(ADP-ribose). Scale bars represent 10 µM. (B) Evaluation of three independent experiments (N = 3) with at least 50 cells per N as in (A) including 10% formaldehyde (FA) values. Increasing fixation time using 2% FA and 3.7% FA leads to significant decrease in PAR formation. Means were compared to the respective 5 min value. Error bar represent mean±s.e.m., **P<0.01; data were analyzed with one-way ANOVA and Dunnett's Multiple Comparison Test.
Figure 2.
Fixation by ChIP protocols induces PARylation.
(A) PAR staining after fixation by three different ChIP protocols or methanol. ChIP fixations induce PAR staining (I–III). Methanol fixation shows no PAR formation (IV). H2O2 and methanol fixation (V) induces granular PAR staining. Focal PAR formation in JLI fixation (I) is reverted to normal distribution if H2O2 is applied in advance (VI). Procedures are indicated below microscopic pictures. Scale bars represent 10 µM. (B) Statistical evaluation of data obtained in (A). Three independent experiments (N = 3) were analyzed with at least 50 cells each data point of one experiment. % PAR positive cells were calculated and analyzed by one-way ANOVA and Bonferroni's Multiple Comparison Test; ***P<0.001. Only methanol fixation is significantly different from the others. Error bars represent mean±s.e.m. (C) Mouse 3T3 cells were fixed by JLI protocol. PAR formation was detected in all three lines tested, i.e. wild type (wt), PARP1 knockout (P1ko) and PARP2 knockout (P2ko). PAR-fluorescence intensities of cells from four randomly chosen microscopic fields per cell line were analyzed by ImageJ and normalized to intensity in wt cells (RFU: relative fluorescence units). At least six independent experiments were used for statistical analysis by One-Way-ANOVA with Bonferroni's Multiple Comparisons Test; error bars represent mean±s.e.m. (N≥6), ***P<0.001 wt vs. P1ko, ****P<0.0001 wt vs. P2ko, P1ko vs. P2ko not significant.
Figure 3.
PARP inhibition suppresses PAR formation only if present in every step.
(A) Detection of PAR by immunofluorescence after different fixation strategies. H2O2 in combination with methanol fixation induces an even distribution of PAR-staining within the nucleus (I). Pretreatment of cells with 2 µM PJ34 6 h in advance of damage induction and methanol fixation completely suppresses PAR formation (II). Fixation of cells with JLI protocol induces polymer synthesis without H2O2 (III), but in contrast to (II), PJ34 is not able to block PARP activity completely (IV). Methanol fixation directly after the formaldehyde step reduces PAR staining (V), but not if the cells were fixed after PBS washing (VI). PJ34 is able to suppress PAR formation only if present in all steps until lysis (VII). Scale bars represent 10 µM. (B) Flow chart of the different fixation strategies. Standard JLI fixation (III) encompasses all steps until lysis/permeabilization for immunofluorescence detection. Preincubation with 2 µM PARP inhibitor PJ34 (IV) is otherwise identical to (III). Methanol is used to fix cells either directly after formaldehyde treatment (V), or after PBS washes (VI). (VII) 2 µM PJ34 is used for preincubation and continuous treatment of cells during all steps until lysis/permeabilization for immunofluorescence detection.
Figure 4.
Low-dose formaldehyde induces γH2AX formation.
Cells were fixed with 1% or 4% paraformaldehyde and analyzed by confocal microscopy. (A) γH2AX foci were counted using one confocal slice after reducing background staining by ImageJ software. Reduction parameters were identical for respective pictures. Cells were split into three groups with (i) less than 5, (ii) between 5 and 20, (iii) more than 20 foci, and percent of total cells was calculated. 10 min 1% paraformaldehyde (1% FA) induces more than sevenfold increase in cells with more than 20 foci, and a decrease in cells with less than 5 foci compared to 4% paraformaldehyde (4% FA). Error bars represent mean±s.e.m. (N = 3), ***P<0.001; data were analyzed with one-way ANOVA and Dunnett's Multiple Comparison Test. (B) All pictures from a z-stack were analyzed for γH2AX foci intensity and normalized to cells fixed for 20 min with 4% paraformaldehyde. Intensity increases eightfold with 1% FA compared to 4% FA. Error bars represent mean±s.e.m. (N = 3), *P = 0.016; data were analyzed by unpaired t-test.
Figure 5.
Suppression of PARylation impacts on ChIP efficiency.
Evaluation of ChIP efficiency of PARP1 bound to PARP1 promoter. Both modifications improve significantly ChIP efficiency. Error bars represent mean±s.e.m. (N = 3), *P<0.05, **P<0.01; data were analyzed with one-way ANOVA and Bonferroni's Multiple Comparison Test.
Figure 6.
PARylation affects binding of transcription factors and histone H1 differently dependent on the binding site.
Preparation of chromatin and ChIP was done by JLI and BMB protocol, respectively. Three independent chromatin preparations were analyzed by three independent PCRs each (panels A, C), or by only one PCR each (panel B) due to lack of material. (A) Suppression of PARylation improves ChIP efficiency in general, but with some specificity. Column color code: blue: ChIP with anti-CTCF antibody; white: ChIP with anti-E2F1 antibody; green: ChIP with anti-p65/RELA (NFκB) antibody; black: ChIP with anti-NFYB antibody. Respective binding sites and promoters are indicated on Y-axis of panel C. J = JLI protocol; B = BMB protocol. Error bars represent mean±s.e.m. (N = 3), exact P-values are indicated; data were analyzed by unpaired t-test. Note that CTCF binding is affected by PARylation at the H19_ICR locus, but not at the BRCA1 promoter. (B) PARylation impacts on histone H1 binding independent of transcription factors. ChIP was performed with anti-H1 antibody and analyzed for binding at the same positions as in (A). Respective binding sites and promoters are indicated on Y-axis of panel C. Coding was maintained to simplify comparison. Error bars represent mean±s.e.m. (N = 3), exact P-values are indicated; data were analyzed by unpaired t-test. Note that H1 binding at the CTCF sites is oppositely affected by PARylation. (C) Increased formaldehyde concentration during fixation does not impact on PCR efficiency. Product signal intensities from input PCRs were compared. Respective binding sites and promoters are indicated on Y-axis. J = JLI protocol, B = BMB protocol. Coding was maintained to simplify comparison between panels. Only NFκB/RELA binding to HIF1A promoter displayed border-line significance in PCR efficiency. Error bars represent mean±s.e.m. (N = 3), *P = 0.045; data were analyzed by unpaired t-test.
Figure 7.
Model of PAR-dependent chromatin remodeling during ChIP fixation.
On the left side, standard ChIP protocol leads to PARP (brown) activation and subsequent poly(ADP-ribose) formation (orange lines) by damaging DNA (red asterisk). This either dislodges proteins (blue, X) from DNA or attracts proteins (green, Y) to DNA with subsequent crosslinking (red arc). Therefore, after lysis and sonication, immunoprecipitated proteins can be present either in wrong amounts (reduced efficiency), or proteins are crosslinked that are not present on DNA in physiological conditions (false positive). On the right side, using 3.7% formaldehyde for 10 min as fixation protocol or treatment with the PARP inhibitor PJ34 throughout the experiment until lysis abolishes PAR formation completely. Thus, chromatin composition is unaltered and reflects in vivo situation.