Holocarboxylase synthetase is a chromatin protein and interacts directly with histone H3 to mediate biotinylation of K9 and K18

Abstract

Holocarboxylase synthetase (HCS) mediates the binding of biotin to lysine (K) residues in histones H2A, H3 and H4; HCS knockdown disturbs gene regulation and decreases stress resistance and lifespan in eukaryotes. We tested the hypothesis that HCS interacts physically with histone H3 for subsequent biotinylation. Co-immunoprecipitation experiments were conducted and provided evidence that HCS co-localizes with histone H3 in human cells; physical interactions between HCS and H3 were confirmed using limited proteolysis assays. Yeast two-hybrid (Y2H) studies revealed that the N-terminal and C-terminal domains in HCS participate in H3 binding. Recombinant human HCS was produced and exhibited biological activity, as evidenced by biotinylation of its known substrate, recombinant p67. Recombinant histone H3.2 and synthetic H3-based peptides were also good targets for biotinylation by recombinant HCS (rHCS) in vitro, based on tracing histone-bound biotin with [(3)H]biotin, streptavidin and anti-biotin antibody. Biotinylation site-specific antibodies were generated and revealed that both K9 and K18 in H3 were biotinylated by HCS. Collectively, these studies provide conclusive evidence that HCS interacts directly with histone H3, causing biotinylation of K9 and K18. We speculate that the targeting of HCS to distinct regions in human chromatin is mediated by DNA sequence, biotin, RNA, epigenetic marks or chromatin proteins.

Fig. 1. Schematic presentation of HCS expression plasmids and HCS domains

(A) Plasmid HCS-GFP; (B) plasmid HCS-pET41a; and (C) diagram of HCS domains. Abbreviations: CD = central domain; L = linker domain; and CT = C-terminal domain.

Fig. 2. HCS interacts physically with histone H3.2 in HEK293 cells

Panel A: Nuclear extracts from HEK293 HCS-GFP cells were precipitated with anti-GFP and probed with anti-histone H3. Nuclear extracts precipitated with non-specific IgG was used as negative control. IgG precipitates (without anti-GFP) were probed with anti-histone H3 (“IgG”); Nuclear extracts without antibody treatment (input), and 0.1 μg of recombinant human histone H3.2 (rH3.2) were used as positive controls. Panel B: Nuclear extracts from HEK293 cells were precipitated with anti-HCS and probed with anti-histone H3 (“HCS pulldown”). As positive control, nuclear extracts collected before treatment with anti-HCS were probed with anti-histone H3 (“Input”); as negative control, protein A precipitates (without anti-HCS) were probed with anti-histone H3 (“Protein A”). Panel C: Purified rHCS was probed with anti-human HCS (lane 1), anti-poly·his tag (lane 2), and coomassie blue (lane 3). Panel D: rHCS was incubated with p67 and cofactors for enzymatic biotinylation; negative controls were generated by omission of individual compounds from reaction mixtures. p67-bound biotin was probed using streptavidin. Panel E: Preincubation of H3 with HCS protects H3 against proteolysis by trypsin in limited proteolysis assays. Left = recombinant histone H3.2 alone; middle = H3.2 pre-incubated with HCS-GST; right = H3.2 pre-incubated with GST alone. H3 was probed with coomassie blue (top panel) and anti-histone H3.2 (bottom panel).

Fig. 3. The N-, C-, and the linker domains in human HCS interact with histone H3.2 in yeast-two-hybrid assays

(A) plate layout of HCS interactions; the interaction between p53 and T antigen was used as positive control; (B) activation of reporter genes and secretion of a-galactosidase, mediated by HCS-H3 interactions (arrows); and (C) successful co-transformation of test plasmids was verified by growing AH109 host strain on SD/−Leu, −Trp, +Kan plates;.

Fig. 4. Recombinant human HCS catalyzes biotinylation of histone H3.2

Panel A: rHCS was incubated with recombinant human histone H3.2 and cofactors for enzymatic biotinylation for up to 12 h; negative controls were generated by omission of rHCS and H3.2. Samples were collected at timed intervals and histone-bound biotin was probed using anti-biotin. Equal loading of histone H3.2 was confirmed by staining with Coomassie blue. Panel B: rHCS was incubated with recombinant human histone H3.2 and cofactors for enzymatic biotinylation for 12 h; negative controls were generated by omission of rHCS and H3.2. Histone-bound biotin was probed using streptavidin. Staining with Coomassie blue was used as loading control. Panel C: Same as panel B, but a synthetic peptide representing amino acids 1 to 25 (N_1–25) in human histone H3.2 was used as substrate. Negative controls were generated by omission of rHCS and peptide.

Fig. 5. Recombinant human HCS catalyzes biotinylation of K9 and K18 in human histone H3

Recombinant human histone H3 was incubated with rHCS and cofactors; negative controls were prepared by omission of rHCS. Transblots of histone were probed with anti-H3K9bio (panel A) and anti-H3K18bio (panel B). Histone H3 based peptide N_1–25 was incubated with rHCS and cofactors; negative control was prepared by omission of rHCS. Transblots of peptide were probed with anti-H3K9bio (panel C) and anti-H3K18bio (panel D