HemK, a class of protein methyl transferase with similarity to DNA methyl transferases, methylates polypeptide chain release factors, and hemK knockout induces defects in translational termination

Abstract

HemK, a universally conserved protein of unknown function, has high amino acid similarity with DNA-(adenine-N6) methyl transferases (MTases). A certain mutation in hemK gene rescues the photosensitive phenotype of a ferrochelatase-deficient (hemH) mutant in Escherichia coli. A hemK knockout strain of E. coli not only suffered severe growth defects, but also showed a global shift in gene expression to anaerobic respiration, as determined by microarray analysis, and this shift may lead to the abrogation of photosensitivity by reducing the oxidative stress. Suppressor mutations that abrogated the growth defects of the hemK knockout strain were isolated and shown to be caused by a threonine to alanine change at codon 246 of polypeptide chain release factor (RF) 2, indicating that hemK plays a role in translational termination. Consistent with such a role, the hemK knockout strain showed an enhanced rate of read-through of nonsense codons and induction of transfer-mRNA-mediated tagging of proteins within the cell. By analysis of the methylation of RF1 and RF2 in vivo and in vitro, we showed that HemK methylates RF1 and RF2 in vitro within the tryptic fragment containing the conserved GGQ motif, and that hemK is required for the methylation within the same fragment of, at least, RF1 in vivo. This is an example of a protein MTase containing the DNA MTase motif and also a protein-(glutamine-N5) MTase.

Figure 1

(a) Structures of the key constructs used in this study. Name, schematic structure of the construct, selection marker, and type of replicon (copy number per cell) is shown for each plasmid. Sequences derived from chromosomes are shown by thick lines and those from vectors by dashed lines. (b) The gene map around hemK in E. coli. The coding region of each gene (open boxes) and the locations of promoters (ref. , bent arrows) are shown. All genes are transcribed from left to right.

Figure 2

Effect of hemK knockout and its suppressor mutation on bacterial growth. A fresh overnight culture of each strain was diluted to about 10 klett units (with filter no. 66) in 5 ml of medium and shaken at 37°C. The cell density was determined at intervals, and representative results of the experiments are shown. (a) The growth defect of the hemK knockout strain and complementation by hemK in an expression plasmid. CA293 (hemK⁺) and CK783 (ΔhemK) were transformed with pAR-hemK or its vector control, pAR3K. The growth of each transformant was examined in casamino acid broth. Essentially the same results were obtained in LB broth. Strains: ■, CA293 pAR3K; □, CK783 pAR3K; ○, CK783 pAR-hemK; x, CK783 pAR-hemK (for x, hemK was induced by adding 50 mg/liter arabinose to the medium at time 0). (b) The growth of the suppressor mutant was compared with that of the hemK⁺ and ΔhemK strains in LB broth. Strains: ■, LE392 (hemK⁺); □, LK783 (ΔhemK); ○, LLR201 (one of the suppressor mutants of ΔhemK). (c) The growth of ΔhemK strain transformed with wild-type or mutant prfB in a plasmid was compared with that of the hemK⁺ or ΔhemK strain in LB broth. Strains: ■, CA293 pMW218 (vector control); □, CK783 pMW218; ▵, CK783 pMW-prfB; ◊, CK783 pMW-prfB(A737G).

Figure 3

tmRNA-tagged proteins in ΔhemK and other cells were detected by immunoblot analysis. Chromosomal ssrA in each strain (lane 1, WK103; lane 2, W3110; lane 3, CA274; lane 4, LE392; lane 5, C600) was inactivated by transducing ΔssrA∷Km, and a plasmid encoding the FLAG-tagging ssrA variant was introduced. Then, cells grown in LB broth were harvested at OD₆₀₀ of about 0.7, suspended in SDS sample buffer (200 μl/OD₆₀₀-unit, and boiled for 5 min, and 10 μl of each cell extract was then analyzed by SDS/PAGE (12% acrylamide) and immunoblotting with anti-FLAG antibody.

Figure 4

Methylation of RFs by HemK. (a) Matrix-assisted laser desorption ionization–time of flight MS analysis of a tryptic digest of RF1 from CA293(hemK⁺) or CK783(ΔhemK). A section of the MS spectrum of the tryptic digest of RF1-FLAG is shown. The predicted m/z value for each fragment after correction for the isotopes is indicated. A peak matching the predicted mass of the 229–245 peptide (1657.8 m/z) was present in the sample from ΔhemK cells, but this peak was apparently shifted by 14 points (to 1671.9 m/z) in the sample from hemK⁺ cells. (b) SDS/PAGE analysis of in vitro-methylated RFs. The methylation reaction was performed on a 10-fold larger scale than described in Table 3 and the samples were subjected to SDS/PAGE (12% polyacrylamide). ³H-labeled protein was then visualized by autoradiography. In each reaction, 1 μg of 6×His-HemK was used as enzyme, and the substrate was: lane 1, 5 μg of BSA; lane 2, 10 μg of BSA; lane 3, 5 μg of RF1-FLAG; lane 4, 10 μg of RF1-FLAG; lane 5, 5 μg of 6×His-RF2; and lane 6, 10 μg of 6×His-RF2. The positions of the size markers are shown on the left. (c) One microgram of RF1-FLAG (lane 1) or 6×His-RF2 (lane 2) was analyzed by SDS/PAGE with the same size markers as in b and stained by using Coomassie brilliant blue.