In previous studies we have identified eIF5A as the only cellular protein that contains an unusual amino acid, hypusine N-epsilon-(4-amino-2-hydroxybutyl)lysine, and established that hypusine biosynthesis occurs by two sequential enzymatic reactions: i) deoxyhypusine synthesis and ii) deoxyhypusine hydroxylation. We have cloned and characterized the structural and catalytic properties of the two enzymes of the hypusine pathway, deoxyhypusine synthase (DHS) and deoxyhypusine hydroxylase (DOHH). We and others have demonstrated that hypusine modification is essential for the activity of eIF-5A and for mammalian cell proliferation. eIF5A is an abundant protein with a long half life and its hypusine modification is irreversible. Therefore, we have sought a possible mechanism of regulation of eIF5A by a reversible posttranslational modification, such as phosphorylation or acetylation. Previously, we and others reported biochemical evidence for acetylation of eIF5A at Lys47 and/or Lys68. In this reporting period, we have obtained strong evidence that eIF5A is negatively regulated by Lys47 acetylation in cells. We have also obtained evidence for acetylation of eIF5A at the hypusine residue by SSAT1 (spermidine/spermine N1-acetyltransferase 1). This hypusine acetylation is predicted to be another mechanism of eIF5A inactivation. We have also investigated the effects of hypusine modification and acetylation on subcellular localization of eIF5A, the reaction mechanism of deoxyhypusine hydroxylase and the role of eIF5A-1 and deoxyhypusine synthase in moouse embryonic development. The cellular activities of eIF5A may be regulated not only by the status of its post-translational modification, such as hypusination or acetylation, but also by its subcellular localization. We have investigated the effect of hypusine modification on the subcellular localization of eIF5A. Immunocytochemical analyses showed differences in the distribution of the non-hypusinated eIF5A precursor, eIF5A(Lys) and the hypusine-containing mature eIF5A. While the precursor is found in both cytoplasm and nucleus, the hypusinated eIF5A is primarily localized in cytoplasm. eIF5A mutant proteins, defective in hypusine modification (K50A, K50R) were localized in a similar manner to the eIF5A precursor, whereas hypusine-modified mutant proteins (K47A, K47R, K68A) were localized mainly in the cytoplasm. These findings provide strong evidence that the hypusine modification of eIF5A dictates its localization in the cytoplasmic compartment where it is required for protein synthesis. Recombinant human deoxyhypusine hydroxylase (hDOHH) has been reported to have oxygen- and iron-dependent activity, an estimated iron/holoprotein stoichiometry of 2. It displays a visible band at 630 nm responsible for the blue color of the as-isolated protein. EPR, Mssbauer, and XAS spectroscopic results provide direct spectroscopic evidence that hDOHH has an antiferromagnetically coupled diiron center with histidines and carboxylates as likely ligands, as suggested by mutagenesis experiments. Resonance Raman experiments show that its blue chromophore arises from a (micro-1,2-peroxo)diiron(III) center that forms in the reaction of the reduced enzyme with O2, so the peroxo form of hDOHH is unusually stable. Nevertheless it can carry out the hydroxylation of deoxyhypusine residue of eIF5A substrate. Despite a lack of sequence similarity, hDOHH has a non-heme diiron active site that resembles both in structure and function those found in methane and toluene monooxygenases, bacterial and mammalian ribonucleotide reductases, and stearoyl acyl carrier protein Delta-9-desaturase from plants, suggesting that the oxygen activating diiron motif is a solution arrived at by convergent evolution. Notably, hDOHH is the only example thus far of a human hydroxylase with such a diiron active site. To investigate the physiological function of eIF5A isoform 11 and deoxyhypusine synthase, we performed their gene targeting in mice. We purchased the ES cell lines, RRE174 (Eif5a1 +/-) and RRM039 (Dhps +/-) (both derived from 129P2 mouse strain with yellow/white color) which have one allele of the Eif5a1 or the Dhps gene disrupted by the gene trap method. We have confirmed disruption of each gene in the intron between the first and second exons by PCR of genomic DNA isolated from these ES cells. The level of eIF5A-1 expression in RRE174 was reduced to half of that in parental ES cells. The RRE174 and RRM039 cells were injected into blastocysts of C57 and the injected blastocysts were implanted into pseudo-pregnant C57 female mice. Chimera pups were born for each clone. Male chimeras of each clone were mated with C57 females to produce agouti pups. The agouti pups were genotyped for identification of those with germ-line transmission of the targeted knockout gene. The gene-targeted heterozygous agouti mice (Eif5a1+/-, or Dhps+/-) appeared to be normal and did not show any growth defects or phenotypes. The heterozygous agouti male and female mice were crossed and the pups born from the heterozygous intercrosses were genotyped to determine if homozygous disruption of Eif5a1 or Dhps is lethal or not. Of 65 pups born from over 10 heterozygous intercrosses (Eif5a1 +/-), we could not find any pups with Eif5a1 homozygous KO genotype. Thus, Eif5a1 gene disruption must cause embryonic lethality in mice. To clarify the time point of embryonic lethality, we cultured the embryos at developmental stages (day 3.5, 6.5 and 8.5) and genotyped them by PCR. The Eif5a1 -/- homozygous embryo was identified on the blastocyst stage (E3.5), but not at later stage, indicating that Eif5a1 -/- embryo is viable up to 3.5 days, but not after 6 days. The phenotype of Dhps homozygous gene disruption in mice was also embryonic lethal. Of 52 pups born from 11 heterozygous agouti intercrosses, no pups with homozygous Dhps disruption were found. When we cultured blastocysts in vitro on a gelatin-coated plate for 7 days, the cell number from homozygous embryo was decreased significantly in comparison to the one from either wild type or heterozygous embryo. These results indicate that both eIF5A-1 and DHS play an essential role for early embryonic development in mice.
Showing the most recent 10 out of 30 publications
