Serine is the major residue for ADP-ribosylation upon DNA damage

Abstract

Poly(ADP-ribose) polymerases (PARPs) are a family of enzymes that synthesise ADP-ribosylation (ADPr), a reversible modification of proteins that regulates many different cellular processes. Several mammalian PARPs are known to regulate the DNA damage response, but it is not clear which amino acids in proteins are the primary ADPr targets. Previously, we reported that ARH3 reverses the newly discovered type of ADPr (ADPr on serine residues; Ser-ADPr) and developed tools to analyse this modification (Fontana et al., 2017). Here, we show that Ser-ADPr represents the major fraction of ADPr synthesised after DNA damage in mammalian cells and that globally Ser-ADPr is dependent on HPF1, PARP1 and ARH3. In the absence of HPF1, glutamate/aspartate becomes the main target residues for ADPr. Furthermore, we describe a method for site-specific validation of serine ADP-ribosylated substrates in cells. Our study establishes serine as the primary form of ADPr in DNA damage signalling.

Keywords: ADP-ribosylation; ARH3; DNA damage; DNA repair; PARP; biochemistry; chemical biology; enzyme; human.

Figure 1.. HPF1-dependent Ser-ADPr is the major form of ADPr upon genotoxic stress.

(A) Control, ARH3 KO (ARH3^−/−), HPF1 KO (HPF1^−/−), and PARP1 KO (PARP1^−/−) U2OS cells were treated with 2 mM H₂O₂. After treatment/recovery, cells were lysed and proteins were separated by SDS-PAGE, analysed by western blot and probed for pan-ADPr, mono-ADPr, PAR, PARP1, ARH3, H3, and HPF1 antibodies. Additionally, Ponceau-S staining was used as loading control. (B) Control, ARH3 KO (ARH3^−/−) and two independent clones of HPF1 KO (HPF1^−/−−1 and HPF1^−/−−2) U2OS cells were treated with 2 mM H₂O₂. After treatment/recovery, cells were lysed and proteins were separated by SDS-PAGE, analysed by western blot and probed for pan-ADPr, PAR, ARH3, and HPF1 antibodies. Ponceau-S staining was used as loading control. (C) Control, ARH3 KO (ARH3^−/−), and HPF1 KO (HPF1^−/−) U2OS cells were treated with 2 mM MMS. After the induction of DNA damage, the cells were left to recover from genotoxic stress for the indicated time points. After treatment/recovery, cells were lysed and proteins were separated by SDS-PAGE, analysed by western blot and probed for pan-ADPr, ARH3, and HPF1 antibodies. Ponceau-S staining was used as loading control. (D) Control and HPF1 KO (HPF1^−/−) HEK293 cells were treated with 2 mM H₂O₂. After treatment/recovery, cells were lysed and proteins were separated by SDS-PAGE, analysed by western blot and probed for pan-ADPr, ARH3, and HPF1 antibodies. Ponceau-S staining was used as loading control.

Figure 2.. HPF1-dependent ADPr is resistant to hydroxylamine.

(A) Autoradiogram shows serine ADPr of two synthetic peptides (wild type (WT) or Ser10Ala (S10A) mutant) corresponding to amino acids 1–21 of human H3 by wild type PARP1 or PARP1 E988Q in the presence of HPF1, with or without treatment with 1M NH₂OH (hydroxylamine). Imperial Blu staining was used to show equal loading of samples. (B) Autoradiogram shows auto-ADPr of PARP1 E988Q (at glutamate residues) and the effect of the treatment with 1M NH₂OH. Imperial Blue staining was used to show equal loading of samples. (C) Whole cell extracts were prepared from pre-damaged U2OS wild type or HPF1 KO (HPF1^−/−) cells. Extracts were either left untreated or treated with 1M hydroxylamine (NH₂OH) for 3 hr prior to separation on SDS-PAGE gel and immunoblotting with pan-ADPr, PAR or HPF1 antibodies. Ponceau-S staining was used as loading control. (D) Whole cell extracts were prepared from U2OS wild type or HPF1 KO (HPF1^−/−) cells following treatment with 2 mM H₂O₂ for 10’. Extracts were either left untreated or treated with 1M hydroxylamine (NH₂OH) for 3 hr prior to separation on SDS-PAGE gel and immunoblotting with pan-ADPr, PAR, H3 or HPF1 antibodies. Ponceau-S staining was used as loading control.

Figure 3.. Serine six is the main ADPr site of histone H2B induced by DNA damage.

(A) Control and HPF1 KO (HPF1^−/−) HEK293 cells were transfected or not with Flag-H2B wild type (wt) and Flag-H2B Ser6Ala mutant construct (S6A). 24 hr post-transfection, cells were treated with 2 mM H₂O₂. After treatment/recovery, cells were lysed and proteins were separated by SDS-PAGE, analysed by western blot and probed for pan-ADPr, Flag, and HPF1 antibodies. Ponceau-S staining was used as loading control. The black star marks the ADP-ribosylated Flag-tagged H2B protein in the whole cell wild type extract, which is absent in other extracts. (B) Flag-tagged H2B wild type (wt) and Ser6Ala mutant (S6A) were immunoprecipitated (IP) by using anti-Flag antibody from the lysates generated in Figure 3A. IPs were separated by SDS-PAGE, analysed by western blot and probed for pan-ADPr and Flag antibodies. Ponceau-S staining was used to stain light chains of Immunoglobulins (IgG) as loading control of the IP.

Figure 4.. Serine 10 and serine 28 are the main ADPr sites of histone H3 induced by DNA damage.

(A) HEK293 cells were transfected or not with Flag-H3.1 (Flag-H3) wild type (wt), Flag-H3.1 Ser10Ala (S10A), Flag-H3.1 Ser28Ala (S28A), and Flag-H3.1 Ser10Ala Ser28Ala double mutant (S10A S28A) constructs. 24 hr post-transfection, cells were treated with 2 mM H₂O₂. After treatment/recovery, cells were lysed and proteins were separated by SDS-PAGE, analysed by western blot and probed for pan-ADPr and Flag antibodies. Ponceau-S staining was used as loading control. (B) Flag-tagged H3.1 (Flag-H3) wild type (wt), Ser10Ala (S10A), Ser28Ala (S28A), and Ser10Ala Ser28Ala double mutants (S10A S28A) were immunoprecipitated (IP) by using anti-Flag antibody from the lysates generated in Figure 4A. IPs were separated by SDS-PAGE, analysed by western blot and probed for pan-ADPr and Flag antibodies. Ponceau-S staining was used to stain light chains of immunoglobulins (IgG) as loading control of the IP.

Author response image 1.