Investigation of in vitro histone H3 glycosylation using H3 tail peptides

Abstract

Posttranslational modifications (PTMs) on histone tails regulate eukaryotic gene expression by impacting the chromatin structure and by modulating interactions with other cellular proteins. One such PTM has been identified as serine and threonine glycosylation, the introduction of the ß-N-acetylglucosamine (GlcNAc) moiety on histone H3 tail at position Ser10 and Thr32. The addition of the ß-O-GlcNAc moiety on serine or threonine residues is facilitated by the O-GlcNAc transferase (OGT), and can be removed by the action of O-GlcNAcase (OGA). Conflicting reports on histone tail GlcNAc modification in vivo prompted us to investigate whether synthetic histone H3 tail peptides in conjunction with other PTMs are substrates for OGT and OGA in vitro. Our enzymatic assays with recombinantly expressed human OGT revealed that the unmodified and PTM-modified histone H3 tails are not substrates for OGT at both sites, Ser10 and Thr32. In addition, full length histone H3 was not a substrate for OGT. Conversely, our work demonstrates that synthetic peptides containing the GlcNAc functionality at Ser10 are substrates for recombinantly expressed human OGA, yielding deglycosylated histone H3 peptides. We also show that the catalytic domains of human histone lysine methyltransferases G9a, GLP and SETD7 and histone lysine acetyltransferases PCAF and GCN5 do somewhat tolerate glycosylated H3Ser10 close to lysine residues that undergo methylation and acetylation reactions, respectively. Overall, this work indicates that GlcNAcylation of histone H3 tail peptide in the presence of OGT does not occur in vitro.

Figure 1

(A) OGT catalyzes the transfer of UDP-GlcNAc to serine/threonine residues. OGA hydrolyses the GlcNAc-possessing serine/threonine residues resulting in native serine (R = H)/threonine (R = CH₃), (B) Identified sites of O-GlcNAc modification of histone H3, (C) Histone peptides and controls used in this study. Me = methyl, Me₂ = dimethyl, Me_2a = asymmetric dimethyl, Me₃ = trimethyl, Ac = acetyl.

Figure 2

MALDI-TOF MS data showing that histone peptides (A) H3(1–15), (B) H3(23–37) and (C) full length histone H3 are not glycosylated by OGT, (D) while the control peptide shows glycosylation under the same conditions. Red spectra show 1 h reactions at 37 °C with 25 µM peptide/protein, 5 µM OGT, 50 µM UDP-GlcNAc and black spectra show the no-enzyme controls.

Figure 3

MALDI-TOF MS data showing glycosylation in the presence of OGT with (A) Control peptide. No glycosylation observed with peptides: (B) H3K9Me, (C) H3K9Me2, (D) H3K9Me3, (E) H3K4Me3, (F) H3R2Me2a, (G) H3R8Me2a, (H) H3K4Me3R8Me2a, (I) H3K4AcK9Ac, (J) H3K4AcK14Ac, (K) H3K9Ac, (L) H3K14Ac. OGA efficiently hydrolyses (M) H3S10GlcNAc, (N) H3K9Me3S10GlcNAc, (O) H3K9AcS10GlcNAc, (P) H3R8Me2aS10GlcNAc. Red spectra show 1 h reactions at 37 °C with 5 µM OGT, 50 µM UDP-GlcNAc (A–L) or with 1 µM OGA (M–P) and black spectra the no-enzyme controls.

Figure 4

(A) Mono-acetylation of H3S10ßGlcNac by lysine acetyltransferases GCN5 or PCAF, (B) Di- and tri-methylation of H3S10ßGlcNac by lysine methyltransferases GLP or G9a, (C) Mono-methylation of H3S10ßGlcNac by lysine methyltransferase SETD7. Red spectra show 3 h reactions at 37 °C with KATs (2 uM), histone peptide (100 µM) and Ac-CoA (300 µM) (A) or 1 h incubations with KMTs (2 µM), SAM (500 µM for GLP/G9a and 200 µM for SETD7) (B,C) of histone peptide 15 (100 µM) and black spectra the no-enzyme controls.