Oxygenase-catalyzed ribosome hydroxylation occurs in prokaryotes and humans

Abstract

The finding that oxygenase-catalyzed protein hydroxylation regulates animal transcription raises questions as to whether the translation machinery and prokaryotic proteins are analogously modified. Escherichia coli ycfD is a growth-regulating 2-oxoglutarate oxygenase catalyzing arginyl hydroxylation of the ribosomal protein Rpl16. Human ycfD homologs, Myc-induced nuclear antigen (MINA53) and NO66, are also linked to growth and catalyze histidyl hydroxylation of Rpl27a and Rpl8, respectively. This work reveals new therapeutic possibilities via oxygenase inhibition and by targeting modified over unmodified ribosomes.

Figure 1

E.coli ycfD is an Arg hydroxylase of the 50S ribosomal protein Rpl16 required for growth. (a) Reaction scheme for Arg hydroxylation (red). (b) Endogenous Rpl16 Arg81 hydroxylation requires ycfD. Rpl16 was purified from either wildtype (BW25113) or ycfD-null (JW1114) E.coli reconstituted with wild-type GFP-ycfD or inactive GFP-ycfD (H125A/D127A, Supplementary Fig. S8) and Arg81 hydroxylation measured by LC-MS. Data represent mean values ± s.d. (c) ycfD hydroxylates a 20mer Arg81 peptide in vitro. ycfD was incubated with Rpl16 peptide plus cosubstrate prior to LC-MS. In the absence of ycfD the triply-charged ([M+H]³⁺) Rpl16 peptide showed the predicted mass (m/z 717.8). Following ycfD incubation the triply-charged Rpl16 peptide mass is 723.2 (+16Da [M+H]³⁺), consistent with hydroxylation.

Figure 2

N066 catalyses β-carbon His hydroxylation of 60S ribosomal protein Rpl8. (a) NO66 hydroxylates a peptide derived from Rpl8 (CNPVEHPFGGGNHQHIGKPST). Recombinant NO66 was incubated with the above Rpl8 peptide prior to chymotrypsinolysis and LC-MS. Without NO66 the doubly-charged ([M+H]²⁺) chymotryptic Rpl8 peptide has the predicted mass (m/z 644.8). Following incubation the doubly-charged chymotryptic peptide displayed a mass of m/z 652.8 (+16Da [M+H]²⁺) consistent with hydroxylation, plus a mass of m/z 643.8 (-2Da [M+H]²⁺), consistent with hydroxylation followed by dehydration. (b) Reaction scheme showing His hydroxylation (red). (c) Partial 2D ¹H-¹³C HSQC spectra of Rpl8 peptide (CNPVEHPFGGGNHQHIGKPST; left panel) and hydroxylated (CNPVEHPFGGGNH(OH)QHIGKPST; right panel) Rpl8 peptide labelled with amino acid assignments for the alpha resonances. The NMR assignment is consistent with beta-hydroxylation at His. Blue labels highlight His residues. Red labels highlight alpha and beta residues of the hydroxylated His residue. The purple labelled residue (N) shows significant proton deshielding on hydroxylation as this precedes the modified His. See Supplementary Fig. 12c for the Hα-Hβ coupling correlation of the alpha and beta residues of the hydroxylated His residue in a region of the 2D ¹H-¹H COSY experiment.

Figure 3

Histidyl hydroxylation of endogenous human Rpl8 and Rpl27a is enzyme-dependent and specific. (a) Endogenous Rpl8 hydroxylation is NO66-dependent. HeLa cells expressing inducible control or NO66 ShRNA were treated with doxycycline prior to immunoblot (left). Endogenous Rpl8 was purified from control (upper right) or NO66 knockdown (lower right) cells as described in methods prior to trypsinolysis and LC-MS quantitation of His216 hydroxylation. (b) Endogenous Rpl27a hydroxylation is MINA53-dependent. A549 cells expressing control or MINA53 ShRNA were immunoblotted as shown (left). Endogenous Rpl27a was purified from control (upper right) of MINA53 knockdown (lower right) cells prior to whole protein MS quantitation of hydroxylation. (c) NO66 binds Rpl8, but not Rpl27a. Complexes were immunoprecipitated from HA-tagged NO66 or empty vector control (EV) HEK293T cells prior to immunoblot. The interaction was specific as Rpl8 was not detected in control pulldowns (anti-V5). (d) MINA53 binds Rpl27a, but not Rpl8. Complexes were immunoprecipitated from doxycycline-treated FLAG-tagged MINA53 or empty vector control (EV) U2OS inducible cells prior to immunoblot. The interaction was specific as Rpl27a was not detected in control pulldowns (anti-V5). Uncropped immunblots from panels c and d are presented in Supplementary Fig. S17).