- Email: [email protected]
Characterization of ARD1 variants in mammalian cells
BBRC Biochemical and Biophysical Research Communications 340 (2006) 422–427 www.elsevier.com/locate/ybbrc
Characterization of ARD1 variants in mammalian cells
q,qq
Se-Hee Kim a,1, Jeong Ae Park a,1, Jin Hyoung Kim a, Ji-Won Lee a, Ji Hae Seo a, Bo-Kyung Jung a, Kwang-Hoon Chun a, Joo-Won Jeong b, Moon-Kyoung Bae c, Kyu-Won Kim a,* a
Research Institute of Pharmaceutical Sciences, College of Pharmacy, Seoul National University, Seoul 151-742, Republic of Korea b Molecular Neurobiology Laboratory, McLean Hospital, Harvard Medical School, Belmont, MA 02478, USA c College of Dentistry, Pusan National University, Busan 609-735, Republic of Korea Received 28 November 2005 Available online 15 December 2005
Abstract Mouse ARD1 (mARD1) has been reported to negatively regulate the hypoxia-inducible factor 1a (HIF-1a) protein by acetylating a lysine residue and enhancing HIF-1a ubiquitination and degradation. However, it was recently reported that human ARD1 (hARD1) does not affect HIF-1a stability. To further explore the activities of the two orthologs, three mouse (mARD1198, mARD1225, mARD1235) and two human (hARD1131, hARD1235) variants were identified and characterized. Among these, mARD1225 was previously reported as a novel negative regulator of HIF-1a. Amino acid sequence analysis showed that the C-terminal region (aa 158–225) of mARD1225 completely differs from those of mouse and human ARD1235, although all three proteins share a well-conserved N-acetyltransferase domain (aa 45–130). The effects of ARD1 variants were evaluated with respect to HIF-1a stability and acetylation activity. Interestingly, mARD1225 strongly decreased the level of HIF-1a and increased the extent of acetylation, whereas mARD1235 and hARD1235 variants had a much weaker effect on HIF-1a stability and acetylation. These results suggest that ARD1 variants might have different effects on HIF-1a stability and acetylation, which may reflect diverse biological functions that remain to be determined. 2005 Elsevier Inc. All rights reserved. Keywords: ARD1; HIF-1a; Acetylation
ARD1 was first described in Saccharomyces cerevisiae, where its activity is required for full repression of the silent mating type locus HML, for sporulation, and for entry into the G0 phase of the cell cycle [1–4]. Yeast ARD1 is a regulatory or catalytic component of the NAT complex and
q This work was supported by Grant No. FG-2-1 of the 21C Frontier Functional Human Genome Project and the Creative Research Initiatives Program, the Ministry of Science and Technology, Korea (to K.-W. Kim) and by the Post-doctoral Fellowship Program of the Korea Science & Engineering Foundation (KOSEF) (to J.-W. Jeong). qq Abbreviations: HIF-1, hypoxia-inducible factor-1; NLS, nuclear localization signal; ODD, oxygen-dependent degradation; RT-PCR, reverse transcription-polymerase chain reaction; VEGF, vascular endothelial growth factor. * Corresponding author. Fax: +82 2 872 1795. E-mail address: [email protected] (K.-W. Kim). 1 These authors contributed equally to this work.
0006-291X/$ - see front matter 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.bbrc.2005.12.018
forms homodimers or heterodimers with NAT1 [4]. Mouse homologs of ARD1 and NAT1 are expressed in regions of cell division and migration throughout brain development, and their expressions are downregulated as neurons mature and form appropriate synaptic connections [5]. The human and mouse homologs are located on the respective X chromosome [6]. Furthermore, amino acid alignments have assigned ARD1 to the GNAT (GCN5-related N-terminal acetyltransferase) family, of which there are over 50 members, and the protein possesses the acetyl-CoA binding motif [7,8] that is conserved from yeast to humans. Previously, we reported that mouse ARD1 (GenBank Accession No. BC027219) interacted with HIF-1a in a yeast two-hybrid assay in which the oxygen-dependent degradation (ODD) domain of HIF-1a was used as a bait [9]. The ODD domain is implicated in regulating the half-life of HIF-1a. MALDI-TOF MS analysis shows that lysine
S.-H. Kim et al. / Biochemical and Biophysical Research Communications 340 (2006) 422–427
532 in the ODD domain is acetylated in vitro in the presence of ARD1. Acetylated HIF-1a is detected in cell lysates, and mutation of lysine 532 to arginine stabilizes overexpressed HIF-1a [10,11]. Unlike the prolyl hydroxylase domain enzymes (PHDs), the activity of acetyltransferase is not dependent on oxygen concentration. Therefore, it was assumed that ARD1 acetylates HIF-1a in both normoxic and hypoxic conditions [12]. Recently, Fisher et al. [13] suggested that human ARD1 (GenBank Accession No. NM_003491) is downregulated in several cell lines in response to hypoxia and hypoxiamimicking compounds and that it might be involved in cell proliferation and in regulating metabolic pathways that respond to hypoxic conditions. However, they reported that inhibition of ARD1 does not affect the HIF-1a protein level under normoxia. Moreover, Bilton et al. [14] showed that overexpressing or silencing of human ARD1 has no impact on HIF-1a protein stability. They further reported that neither mRNA nor protein levels of ARD1 are regulated by hypoxia in several cell types. To clarify these discrepancies, we hypothesized that the mouse and human homologs of ARD1 differ with respect to their influence on HIF-1a stability and their biological functions. Therefore, we identified mammalian ARD1 variants and determined their effects on HIF-1a stability and acetylation. The results presented here demonstrate that there are several ARD1 variants in mammalian cells and that they differentially regulate HIF-1a stability and acetylation. Materials and methods Reagents and antibodies. MG132 was purchased from Calbiochem. Mouse anti-HIF-1a antibody was purchased from BD Pharmingen. Anti-GFP and acetyl-lysine antibodies were purchased from Santa Cruz Biotechnology and Cell Signaling, respectively. A polyclonal antibody to ARD1 was produced by Dinona. The immunogens correspond to amino acids 1–17 of mARD1225 (GenBank Accession No. BC027219) and hARD1235 (GenBank Accession No. NM_003491). Antibody specificity was confirmed using siRNA targeting ARD1 mRNA (data not shown). Plasmids. The pEGFP-HIF-1a expression vector was prepared as previously described [15]. ARD1 variant expression cassettes were constructed by PCR and subcloned into the GFP-tagged pCS2+ vector. The mARD1225, mARD1235, and hARD1235 expression constructs were subcloned into the EcoRI and XhoI sites of the GFP-tagged pCS2+ vector. Cell culture and hypoxic conditions. HeLa, HT1080, H1299, and NIH3T3 cells were maintained in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum (Gibco-BRL) and 1% antibiotics (Gibco-BRL). For hypoxic conditions, cells were incubated in 5% CO2/1% O2 balanced with N2 in a hypoxic chamber (Forma). Immunoprecipitation and Western blot analysis. HeLa cells were transfected using Lipofectamine Plus (Invitrogen). Transfected cells were lysed in whole cell extract buffer (10 mM Hepes at pH 7.9, 400 mM NaCl, 0.1 mM EDTA, 5% glycerol, 1 mM DTT, and protease inhibitors). Antiacetyl-Lys antibody (1 lg) was added to the lysate, followed by the addition of protein-A–agarose (Upstate Biotech) in TEG reaction buffer (20 mM Tris–HCl at pH 7.4, 1 mM EDTA, 10% glycerol, 1 mM DTT, and 150 mM NaCl), and the mixture was stirred overnight at 4 C. The immunoprecipitates were washed in TEG washing buffer (TEG reaction buffer containing 0.1% Triton X-100), subjected to SDS–PAGE, and transferred onto a nitrocellulose membrane (Amersham Bioscience). The
423
membrane was probed with primary antibodies followed by a secondary antibody conjugated with horseradish peroxidase and detected by the ECL plus system (Amersham Bioscience). Transient transfection. For overexpression of ARD1 variants, subconfluent cells in 60 or 100 mm dishes were transiently transfected with 3 or 6 lg empty vector or ARD1 variant plasmids using Lipofectamine Plus. Briefly, the day before transfection, exponentially growing cells were trypsinized, replated at a density of 5 · 105 per 60 mm dish or 106 per 100 mm dish, and incubated for 24 h. Liposome–DNA precipitates were added dropwise onto a plate. The plate was swirled to uniformly distribute the mixture and incubated for 3 h at 37 C. The medium was replaced with normal growth medium, and the cells were further incubated for 24 h. In vivo acetylation assay. HeLa cells were transfected with GFPmARD1225, GFP-mARD1235, and GFP-hARD1235, treated with MG132 (10 lM) and subjected to hypoxia for 16 h. Total cell extracts were isolated and immunoprecipitated with 1 lg of acetyl-Lys antibody in TEG reaction buffer overnight at 4 C. The immunoprecipitates were washed in TEG washing buffer and subjected to SDS–PAGE and Western blot analysis with the HIF-1a-specific antibody. RT-PCR analysis. Total RNA was extracted using an RNA extraction kit (Invitrogen). Complementary DNA was synthesized from 4 lg total RNA using an oligo(dT) primer. Primers used for PCR were: VEGFforward—5 0 GAGAATTCGGCCTCCGAAACCATGAACTTTCTGCT 3 0 and reverse—5 0 GAGCATGCCCTCCTGCCCGGCTCACCGC 3 0 ; bactin-forward—5 0 GACTACCTCATGAAGATC 3 0 and reverse—5 0 GATCCACATCTGCTGGAA 3 0 . Thirty cycles of PCR was carried out to amplify VEGF and b-actin, and signals were detected by autoradiography.
Results Identification of mammalian ARD1 variants A search of the NCBI database and DNA sequence alignment allowed identification of mouse ARD1 variants (Fig. 1A). The mARD1198 (GenBank Accession No. AK078700), mARD1225 (GenBank Accession No. BC027219), and mARD1235 (GenBank Accession No. NM_019870) sequences encode proteins of 198, 225, and 235 amino acids, respectively. The mARD1225 and mARD1235 proteins have the well-conserved N-acetyltransferase domain (aa 45–130), although mARD1198 has only a partial domain. Human ARD1 variants were isolated as cDNAs from HeLa and HEK293T cells with RT-PCR primers flanking the hARD1 coding region. Two hARD1 variants were identified (Fig. 1B) that hARD1131 and hARD1235 were previously reported as GenBank Accession Nos. BC063377 and NM_003491, respectively. Previously, we reported that mouse ARD1 interacted with the ODD domain of HIF-1a using yeast two-hybrid assay [9]. We sequenced the mouse cDNA clones and compared the sequence data of NCBI. Thereafter, we reported our clones as a homolog of mARD1235 in the previous report (GenBank Accession No. NM_019870) [9]. After then, we realized that the full sequences of our clones were exactly the same as mARD1225 (GenBank Accession No. BC027219). Several reports have indicated that mARD1225 differs from hARD1235 in regulating HIF-1a stability [9,13,14]. To address this issue, we first analyzed the mARD1225, hARD1235, and mARD1235 sequences. The nucleotide
424
S.-H. Kim et al. / Biochemical and Biophysical Research Communications 340 (2006) 422–427
Fig. 1. Mouse and human ARD1 variants. (A) Sequence alignment of mouse ARD1 variants. Identical residues are shaded in black or gray. (B) Sequence alignment of human ARD1 variants. Identical residues are shaded in black. (C) Alignment of mARD1225, mARD1235, and hARD1235. Identical residues are shaded in black or gray. (D) Schematic representation of mARD1225, mARD1235, and hARD1235 nucleotide sequences. The acetyltransferase domain at nucleotides 133–390, the putative NLS at nucleotides 232–249, and the acetyl-CoA binding domain at nucleotides 244–261 are indicated. Nucleotides 472–562 might be deleted by alternative splicing in mARD1235 and hARD1235. The stop codons of mARD1225 and ARD1235 at nucleotides 678 and 799 are indicated.
sequences of mARD1225 are identical to those of mARD1235 except nucleotides 472–562 of exon 8, and amino acids of the mARD1225 protein share 69.8% identity
with those of mARD1235 (Figs. 1C and D). Therefore, it is intriguing that the C-terminal region (aa 158–225) of mARD1225 differs from those of both mouse and human
S.-H. Kim et al. / Biochemical and Biophysical Research Communications 340 (2006) 422–427
ARD1235 (Fig. 1C), although all three proteins have the well-conserved N-acetyltransferase domain (aa 45–130) and all three genes are located in the region q28 of the respective X chromosome [6,16]. GenBank analysis of the hARD1235 introns and exons suggests that the C-terminal amino acid differences between mARD1225 and mARD1235/hARD1235 are probably due to alternative splicing of exon 8 (Fig. 1D), which alters the reading frame and introduces a stop codon at nucleotides 678 and 799, respectively. Arnesen et al. [17] predicted that the acetyltransferase domain is located between nucleotides 133 and 390, and a putative NLS (nuclear localization signal) (KRSHRR) and acetyl-CoA binding domain [(Q/ R)XXGX(G/A)] [18] are located at nucleotides 232–249 and 244–261, respectively (Fig. 1D). Effect of ARD1 variants on HIF-1a stability We previously reported that mARD1225 decreases the level of the HIF-1a protein [9]. To further examine the effect of other ARD1 variants on HIF-1a stability, we transfected them into HeLa cells and incubated the cells under normoxic and hypoxic conditions. As shown in Fig. 2A, mARD1225 decreased the level of HIF-1a under hypoxic conditions by approximately 60%, whereas hARD1235 and mARD1235 inhibited HIF-1a stability by 5–10%. We previously reported that mARD1225 downregulates the expression of VEGF mRNA and protein under hypoxic conditions [9]. To further test the influence of ARD1 variants on VEGF expression, we transfected HeLa cells with these variants, incubated the cells under normoxic or hypoxic conditions, and performed RT-PCR analysis. Similar to the results shown in Fig. 2A, mARD1225 strong-
425
ly decreased VEGF mRNA levels under hypoxia, as compared with the effect of hARD1235 and mARD1235 (Fig. 2B). These data indicated that mARD1225 strongly decreases HIF-1a stability and downregulates the level of VEGF mRNA, whereas hARD1235 and mARD1235 had only weak effects. Evaluation of ARD1 variants on HIF-1a acetylation mARD1225 mediates the e-acetylation of a HIF-1a lysine residue and thereby enhances the ubiquitination and degradation of HIF-1a [9]. To further evaluate the acetyltransferase activity of ARD1 variants with respect to HIF-1a, we transfected HeLa cells with these variants and performed an in vivo acetylation assay using an antiacetyl-Lys antibody after treatment with MG132, a proteasomal inhibitor, to be equal HIF-1a protein level. Consistent with the results shown in Fig. 2, mARD1225 strongly acetylated HIF-1a but hARD1235 and mARD1235 had much weaker effects (Fig. 3A). From repeated experiments, we concluded that mARD1225 increased the level of HIF1a acetylation by more than 100% as compared with the level observed in control cells but that hARD1235 and mARD1235 showed a little effect (Fig. 3B). Expression of ARD1 proteins in mammalian cell lines To determine the expression pattern of ARD1 variants in mammalian cells, we performed Western blot analysis using the ARD1 antibody. As shown in Fig. 4, immunoblots of proteins from all human cell lines (HeLa, HT1080, and H1299) tested showed major intense band of 32 kDa, which corresponds to hARD1235, whereas
Fig. 2. Effect of ARD1 variants on the expression of HIF-1a protein. (A,B) HeLa cells were transfected with plasmids encoding GFP alone (mock), GFPmARD1225 (m225), GFP-mARD1235 (m235) or GFP-hARD1235 (h235) and incubated under normoxia (N, 21% O2) and hypoxia (H, 1% O2) for 16 h. (A, left) Total cell extracts were subjected to Western blot analysis with anti-HIF-1a, anti-GFP, or anti-a-tubulin antibodies. (Right) The relative HIF-1a protein levels were determined in five independent assays and quantified by densitometry. The HIF-1a protein level under hypoxia was set to 100%. (B) RT-PCR analysis to detect the expression of VEGF and b-actin mRNA using specific primers and total RNA as a template.
426
S.-H. Kim et al. / Biochemical and Biophysical Research Communications 340 (2006) 422–427
weak bands of smaller sizes were present in mammalian cells, suggesting that these might be novel ARD1 variants (Fig. 4). Discussion
Fig. 3. Effects of ARD1 variants on HIF-1a acetylation. (A) HeLa cells were transfected with plasmids encoding GFP alone (mock), GFPmARD1225 (m225), GFP-mARD1235 (m235) or GFP-hARD1235 (h235) and treated with 10 lM MG132 for 16 h at hypoxia (1% O2) (H + MG132). Extracts were immunoprecipitated with the anti-acetylLys antibody and the immunoprecipitates were subjected to Western blot analysis to detect HIF-1a. Total cell lysates were similarly analyzed with anti-HIF-1a antibody and anti-GFP antibody as a loading control and to determine transfection efficiency. (B) The relative level of acetylated HIF1a was expressed as ratios of acetylated HIF-1a against total expressed HIF-1a, as determined in three independent assays by immunoprecipitation and Western blot analysis followed by densitometry.
Fig. 4. Expression of hARD1 proteins in mammalian cell lines. Immunoblot analysis of lysates in different cell lines, including human cervical adenocarcinoma HeLa cells, human fibrosarcoma HT1080 cells, human lung adenocarcinoma H1299 cells, and murine fibroblast NIH3T3 cells, using the anti-ARD1 antibody.
mouse NIH3T3 cells strongly expressed a 30 kDa protein corresponding to mARD1225. These results were confirmed by RT-PCR experiments with primer specificity on the variants cDNA (data not shown). Consistent with the ARD1 variants identified by RT-PCR (Fig. 1), two or three
We previously reported that mouse ARD1 interacts with the ODD domain of HIF-1a and mediates acetylation of Lys532, thereby enhancing HIF-1a ubiquitination and degradation [9]. The ARD1 protein that negatively regulates HIF-1a is mARD1225 (GenBank Accession No. BC027219), which shares the N-terminal region (aa 1– 157) with mARD1235, previously reported as GenBank Accession No. NM_019870. In contrast to our findings, recent studies have shown that hARD1235 does not affect HIF-1a stability or acetylation [13,14]. From these results, we hypothesized that ARD1 variants might differentially modulate HIF-1a stability. Consistent with this prediction, we identified several ARD1 variants by sequence alignment and RT-PCR analysis in mammalian cell lines; three mouse variants (mARD1198, mARD1225, and mARD1235) and two human variants (hARD1131 and hARD1235). Furthermore, we found that mARD1225 was highly expressed compared with mARD1235 in mouse NIH3T3 cells, whereas hARD1235 was strongly expressed in human cell lines. The mARD1235 mRNA lacks a part of exon 8 of the ARD1225 mRNA, which results in an altered reading frame that specifies 235-amino acid polypeptides. Therefore, the C-terminal region (aa 158–225) of the mARD1225 protein is completely different from that of the mARD1235 protein. Thereafter, we validated the effects of mARD1225, mARD1235, and hARD1235 in the regulation of HIF-1a stability. In HeLa cells, mARD1225 strongly decreased the expression of HIF-1a protein under hypoxic conditions, as compared with hARD1235 and mARD1235, although all three proteins have a conserved N-acetyltransferase domain (aa 45–130). This result suggests that the Cterminal region of ARD1 may be important in the regulation of HIF-1a stability. In addition, only mARD1225 strongly mediated e-acetylation of HIF-1a. These results support the idea that the acetyltransferase activity of mARD1225 downregulates HIF-1a protein stability, whereas hARD1235 and mARD1235 have a weak activity. It was also reported that ARD1 binds NAT1 forming a heterodimer in yeast and mammalian cells. Since the human NAT1 homolog Tubedown-1 has been reported to be an important regulator of retinal neovascularization [18], ARD1 may also be involved in retinal angiogenesis. Furthermore, both mARD1235 and mNAT1 are expressed in regions of cell division with a downregulation of expression as neurons mature and form appropriate synaptic connections [5]. In addition, hARD1235 and human NAT1 are regulated by caspase-mediated cleavage in stressed cells such as those undergoing apoptosis [17]. These results indicate that ARD1 variants function in a range of processes, including angiogenesis, cell proliferation, differentiation, and apoptosis, possibly through posttranslational modifi-
S.-H. Kim et al. / Biochemical and Biophysical Research Communications 340 (2006) 422–427
cation of target proteins. However, the relationships among ARD1 variants are not yet clear. Therefore, future experiments will focus on the specific roles played by ARD1 variants in different cell lines and the consequences of ARD1-dependent regulation. References [1] F.J.S. Lee, L.W. Lin, J.A.J. Smith, Na acetylation is required for normal growth and mating of Saccharomyces cerevisiae, Bacteriology 171 (1989) 5795–5802. [2] J.R. Mullen, P.S. Kayne, R.P. Moerschell, S. Tsunasawa, M. Gribskov, M. Colarito-Shepanski, M. Grunstein, F. Shetman, R. Sternglanz, Identification and characterization of genes and mutants for an N-terminal acetyltransferase from yeast, EMBO J. 8 (1989) 2067–2075. [3] E.C. Park, J.W. Szostak, Point mutations in the yeast histone H4 gene prevent silencing of the silent mating type locus HML, Mol. Cell. Biol. 10 (1990) 4932–4934. [4] E.C. Park, J.W. Szostak, ARD1 and NAT1 proteins form a complex that has N-terminal acetyltransferase activity, EMBO J. 11 (1992) 2087–2093. [5] N. Sugiura, S.M. Adams, R.A. Corriveau, An evolutionarily conserved N-terminal acetyltransferase complex associated with neuronal development, J. Biol. Chem. 278 (2003) 40113–40120. [6] C. Tribioli, M. Mancini, E. Plassart, S. Bione, S. Rivella, C. Sala, G. Torri, D. Toniolo, Isolation of new genes in distal Xq28: transcriptional map and identification of a human homologue of the ARD1 Nacetyltransferase of Saccharomyces cerevisiae, Hum. Mol. Genet. 3 (1994) 1061–1067. [7] A.F. Neuwald, D. Landsman, GCN5-related histone N-acetyltransferases belong to a diverse superfamily that includes the yeast SPT10 protein, Trends Biochem. Sci. 22 (1997) 154–155. [8] R. Strernglanz, H. Schindelin, Structure and mechanism of action of the histone acetyltransferase Gcn5 and similarity to other N-acetyltransferases, Proc. Natl. Acad. Sci. USA 96 (1999) 8807–8808.
427
[9] J.W. Jeong, M.K. Bae, M.Y. Ahn, S.H. Kim, T.K. Sohn, M.H. Bae, M.A. Yoo, E.J. Song, K.J. Lee, K.W. Kim, Regulation and destabilization of HIF-1a by ARD1-mediated acetylation, Cell 111 (2002) 709–720. [10] K. Tanimoto, Y. Makino, T. Pereira, L. Poellinger, Mechanism of regulation of the hypoxia-inducible factor-1a by the von Hippel– Lindau tumor suppressor protein, EMBO J. 19 (2000) 4298–4309. [11] Y.S. Chun, K.H. Lee, E. Choi, S.Y. Bae, E.J. Yeo, L.E. Huang, M.S. Kim, J.W. Park, Phorbol ester stimulates the nonhypoxic induction of a novel hypoxia-inducible factor 1a isoform: implications for tumor promotion, Cancer Res. 15 (2003) 8700–8707. [12] Z.N. Demidenko, A.M. Rapisarda, M. Garayoa, P. Giannakakou, G. Melillo, M.V. Blagosklonny, Accumulation of hypoxia-inducible factor-1a is limited by transcription-dependent depletion, Oncogene 24 (2005) 4829–4838. [13] T.S. Fisher, S.D. Etages, L. Hayes, K. Crimin, B. Li, Analysis of ARD1 function in hypoxia response using retroviral RNA interference, J. Biol. Chem. 280 (2005) 17749–17757. [14] R. Bilton, N. Mazure, E. Trottier, M. Hattab, M.A. Dery, D.E. Richard, J. Pouyssegur, M.C. Brahimi-Horn, Arrest-defective-1 protein, an acetyltransferase, does not alter stability of hypoxiainducible factor (HIF)-1a and is not induced by hypoxia or HIF, J. Biol. Chem. 280 (2005) 31132–31140. [15] M.K. Bae, M.Y. Ahn, J.W. Jeong, M.H. Bae, Y.M. Lee, S.K. Bae, J.W. Park, K.R. Kim, K.W. Kim, Jab1 interacts directly with HIF-1a and regulates its stability, J. Biol. Chem. 277 (2002) 9–12. [16] V. Brenner, G. Nyakatura, A. Rosenthal, M. Platzer, Genomic organization of two novel genes on human Xq28: compact head to head arrangement of IDHc and TRAPd is conserved in rat and mouse, Genetics 44 (1997) 8–14. [17] T. Arnesen, D. Anderson, C. Baldersheim, M. Lanotte, J.E. Varhaug, J.R. Lillehaug, Identification and characterization of the human ARD1–NATH protein acetyltransferase complex, Biochem. J. 386 (2005) 433–443. [18] D.S. Wall, R.L. Gendron, W.V. Good, E. Miskiewicz, M. Woodland, K. Leblanc, H. Paradis, Suppressed expression of Tubedown-1 in retinal neovascularization of proliferative diabetic retinopathy, Invest. Ophthalmol. Vis. Sci. 45 (2004) 3704–3712.
Characterization of ARD1 variants in mammalian cells
q,qq
Se-Hee Kim a,1, Jeong Ae Park a,1, Jin Hyoung Kim a, Ji-Won Lee a, Ji Hae Seo a, Bo-Kyung Jung a, Kwang-Hoon Chun a, Joo-Won Jeong b, Moon-Kyoung Bae c, Kyu-Won Kim a,* a
Research Institute of Pharmaceutical Sciences, College of Pharmacy, Seoul National University, Seoul 151-742, Republic of Korea b Molecular Neurobiology Laboratory, McLean Hospital, Harvard Medical School, Belmont, MA 02478, USA c College of Dentistry, Pusan National University, Busan 609-735, Republic of Korea Received 28 November 2005 Available online 15 December 2005
Abstract Mouse ARD1 (mARD1) has been reported to negatively regulate the hypoxia-inducible factor 1a (HIF-1a) protein by acetylating a lysine residue and enhancing HIF-1a ubiquitination and degradation. However, it was recently reported that human ARD1 (hARD1) does not affect HIF-1a stability. To further explore the activities of the two orthologs, three mouse (mARD1198, mARD1225, mARD1235) and two human (hARD1131, hARD1235) variants were identified and characterized. Among these, mARD1225 was previously reported as a novel negative regulator of HIF-1a. Amino acid sequence analysis showed that the C-terminal region (aa 158–225) of mARD1225 completely differs from those of mouse and human ARD1235, although all three proteins share a well-conserved N-acetyltransferase domain (aa 45–130). The effects of ARD1 variants were evaluated with respect to HIF-1a stability and acetylation activity. Interestingly, mARD1225 strongly decreased the level of HIF-1a and increased the extent of acetylation, whereas mARD1235 and hARD1235 variants had a much weaker effect on HIF-1a stability and acetylation. These results suggest that ARD1 variants might have different effects on HIF-1a stability and acetylation, which may reflect diverse biological functions that remain to be determined. 2005 Elsevier Inc. All rights reserved. Keywords: ARD1; HIF-1a; Acetylation
ARD1 was first described in Saccharomyces cerevisiae, where its activity is required for full repression of the silent mating type locus HML, for sporulation, and for entry into the G0 phase of the cell cycle [1–4]. Yeast ARD1 is a regulatory or catalytic component of the NAT complex and
q This work was supported by Grant No. FG-2-1 of the 21C Frontier Functional Human Genome Project and the Creative Research Initiatives Program, the Ministry of Science and Technology, Korea (to K.-W. Kim) and by the Post-doctoral Fellowship Program of the Korea Science & Engineering Foundation (KOSEF) (to J.-W. Jeong). qq Abbreviations: HIF-1, hypoxia-inducible factor-1; NLS, nuclear localization signal; ODD, oxygen-dependent degradation; RT-PCR, reverse transcription-polymerase chain reaction; VEGF, vascular endothelial growth factor. * Corresponding author. Fax: +82 2 872 1795. E-mail address: [email protected] (K.-W. Kim). 1 These authors contributed equally to this work.
0006-291X/$ - see front matter 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.bbrc.2005.12.018
forms homodimers or heterodimers with NAT1 [4]. Mouse homologs of ARD1 and NAT1 are expressed in regions of cell division and migration throughout brain development, and their expressions are downregulated as neurons mature and form appropriate synaptic connections [5]. The human and mouse homologs are located on the respective X chromosome [6]. Furthermore, amino acid alignments have assigned ARD1 to the GNAT (GCN5-related N-terminal acetyltransferase) family, of which there are over 50 members, and the protein possesses the acetyl-CoA binding motif [7,8] that is conserved from yeast to humans. Previously, we reported that mouse ARD1 (GenBank Accession No. BC027219) interacted with HIF-1a in a yeast two-hybrid assay in which the oxygen-dependent degradation (ODD) domain of HIF-1a was used as a bait [9]. The ODD domain is implicated in regulating the half-life of HIF-1a. MALDI-TOF MS analysis shows that lysine
S.-H. Kim et al. / Biochemical and Biophysical Research Communications 340 (2006) 422–427
532 in the ODD domain is acetylated in vitro in the presence of ARD1. Acetylated HIF-1a is detected in cell lysates, and mutation of lysine 532 to arginine stabilizes overexpressed HIF-1a [10,11]. Unlike the prolyl hydroxylase domain enzymes (PHDs), the activity of acetyltransferase is not dependent on oxygen concentration. Therefore, it was assumed that ARD1 acetylates HIF-1a in both normoxic and hypoxic conditions [12]. Recently, Fisher et al. [13] suggested that human ARD1 (GenBank Accession No. NM_003491) is downregulated in several cell lines in response to hypoxia and hypoxiamimicking compounds and that it might be involved in cell proliferation and in regulating metabolic pathways that respond to hypoxic conditions. However, they reported that inhibition of ARD1 does not affect the HIF-1a protein level under normoxia. Moreover, Bilton et al. [14] showed that overexpressing or silencing of human ARD1 has no impact on HIF-1a protein stability. They further reported that neither mRNA nor protein levels of ARD1 are regulated by hypoxia in several cell types. To clarify these discrepancies, we hypothesized that the mouse and human homologs of ARD1 differ with respect to their influence on HIF-1a stability and their biological functions. Therefore, we identified mammalian ARD1 variants and determined their effects on HIF-1a stability and acetylation. The results presented here demonstrate that there are several ARD1 variants in mammalian cells and that they differentially regulate HIF-1a stability and acetylation. Materials and methods Reagents and antibodies. MG132 was purchased from Calbiochem. Mouse anti-HIF-1a antibody was purchased from BD Pharmingen. Anti-GFP and acetyl-lysine antibodies were purchased from Santa Cruz Biotechnology and Cell Signaling, respectively. A polyclonal antibody to ARD1 was produced by Dinona. The immunogens correspond to amino acids 1–17 of mARD1225 (GenBank Accession No. BC027219) and hARD1235 (GenBank Accession No. NM_003491). Antibody specificity was confirmed using siRNA targeting ARD1 mRNA (data not shown). Plasmids. The pEGFP-HIF-1a expression vector was prepared as previously described [15]. ARD1 variant expression cassettes were constructed by PCR and subcloned into the GFP-tagged pCS2+ vector. The mARD1225, mARD1235, and hARD1235 expression constructs were subcloned into the EcoRI and XhoI sites of the GFP-tagged pCS2+ vector. Cell culture and hypoxic conditions. HeLa, HT1080, H1299, and NIH3T3 cells were maintained in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum (Gibco-BRL) and 1% antibiotics (Gibco-BRL). For hypoxic conditions, cells were incubated in 5% CO2/1% O2 balanced with N2 in a hypoxic chamber (Forma). Immunoprecipitation and Western blot analysis. HeLa cells were transfected using Lipofectamine Plus (Invitrogen). Transfected cells were lysed in whole cell extract buffer (10 mM Hepes at pH 7.9, 400 mM NaCl, 0.1 mM EDTA, 5% glycerol, 1 mM DTT, and protease inhibitors). Antiacetyl-Lys antibody (1 lg) was added to the lysate, followed by the addition of protein-A–agarose (Upstate Biotech) in TEG reaction buffer (20 mM Tris–HCl at pH 7.4, 1 mM EDTA, 10% glycerol, 1 mM DTT, and 150 mM NaCl), and the mixture was stirred overnight at 4 C. The immunoprecipitates were washed in TEG washing buffer (TEG reaction buffer containing 0.1% Triton X-100), subjected to SDS–PAGE, and transferred onto a nitrocellulose membrane (Amersham Bioscience). The
423
membrane was probed with primary antibodies followed by a secondary antibody conjugated with horseradish peroxidase and detected by the ECL plus system (Amersham Bioscience). Transient transfection. For overexpression of ARD1 variants, subconfluent cells in 60 or 100 mm dishes were transiently transfected with 3 or 6 lg empty vector or ARD1 variant plasmids using Lipofectamine Plus. Briefly, the day before transfection, exponentially growing cells were trypsinized, replated at a density of 5 · 105 per 60 mm dish or 106 per 100 mm dish, and incubated for 24 h. Liposome–DNA precipitates were added dropwise onto a plate. The plate was swirled to uniformly distribute the mixture and incubated for 3 h at 37 C. The medium was replaced with normal growth medium, and the cells were further incubated for 24 h. In vivo acetylation assay. HeLa cells were transfected with GFPmARD1225, GFP-mARD1235, and GFP-hARD1235, treated with MG132 (10 lM) and subjected to hypoxia for 16 h. Total cell extracts were isolated and immunoprecipitated with 1 lg of acetyl-Lys antibody in TEG reaction buffer overnight at 4 C. The immunoprecipitates were washed in TEG washing buffer and subjected to SDS–PAGE and Western blot analysis with the HIF-1a-specific antibody. RT-PCR analysis. Total RNA was extracted using an RNA extraction kit (Invitrogen). Complementary DNA was synthesized from 4 lg total RNA using an oligo(dT) primer. Primers used for PCR were: VEGFforward—5 0 GAGAATTCGGCCTCCGAAACCATGAACTTTCTGCT 3 0 and reverse—5 0 GAGCATGCCCTCCTGCCCGGCTCACCGC 3 0 ; bactin-forward—5 0 GACTACCTCATGAAGATC 3 0 and reverse—5 0 GATCCACATCTGCTGGAA 3 0 . Thirty cycles of PCR was carried out to amplify VEGF and b-actin, and signals were detected by autoradiography.
Results Identification of mammalian ARD1 variants A search of the NCBI database and DNA sequence alignment allowed identification of mouse ARD1 variants (Fig. 1A). The mARD1198 (GenBank Accession No. AK078700), mARD1225 (GenBank Accession No. BC027219), and mARD1235 (GenBank Accession No. NM_019870) sequences encode proteins of 198, 225, and 235 amino acids, respectively. The mARD1225 and mARD1235 proteins have the well-conserved N-acetyltransferase domain (aa 45–130), although mARD1198 has only a partial domain. Human ARD1 variants were isolated as cDNAs from HeLa and HEK293T cells with RT-PCR primers flanking the hARD1 coding region. Two hARD1 variants were identified (Fig. 1B) that hARD1131 and hARD1235 were previously reported as GenBank Accession Nos. BC063377 and NM_003491, respectively. Previously, we reported that mouse ARD1 interacted with the ODD domain of HIF-1a using yeast two-hybrid assay [9]. We sequenced the mouse cDNA clones and compared the sequence data of NCBI. Thereafter, we reported our clones as a homolog of mARD1235 in the previous report (GenBank Accession No. NM_019870) [9]. After then, we realized that the full sequences of our clones were exactly the same as mARD1225 (GenBank Accession No. BC027219). Several reports have indicated that mARD1225 differs from hARD1235 in regulating HIF-1a stability [9,13,14]. To address this issue, we first analyzed the mARD1225, hARD1235, and mARD1235 sequences. The nucleotide
424
S.-H. Kim et al. / Biochemical and Biophysical Research Communications 340 (2006) 422–427
Fig. 1. Mouse and human ARD1 variants. (A) Sequence alignment of mouse ARD1 variants. Identical residues are shaded in black or gray. (B) Sequence alignment of human ARD1 variants. Identical residues are shaded in black. (C) Alignment of mARD1225, mARD1235, and hARD1235. Identical residues are shaded in black or gray. (D) Schematic representation of mARD1225, mARD1235, and hARD1235 nucleotide sequences. The acetyltransferase domain at nucleotides 133–390, the putative NLS at nucleotides 232–249, and the acetyl-CoA binding domain at nucleotides 244–261 are indicated. Nucleotides 472–562 might be deleted by alternative splicing in mARD1235 and hARD1235. The stop codons of mARD1225 and ARD1235 at nucleotides 678 and 799 are indicated.
sequences of mARD1225 are identical to those of mARD1235 except nucleotides 472–562 of exon 8, and amino acids of the mARD1225 protein share 69.8% identity
with those of mARD1235 (Figs. 1C and D). Therefore, it is intriguing that the C-terminal region (aa 158–225) of mARD1225 differs from those of both mouse and human
S.-H. Kim et al. / Biochemical and Biophysical Research Communications 340 (2006) 422–427
ARD1235 (Fig. 1C), although all three proteins have the well-conserved N-acetyltransferase domain (aa 45–130) and all three genes are located in the region q28 of the respective X chromosome [6,16]. GenBank analysis of the hARD1235 introns and exons suggests that the C-terminal amino acid differences between mARD1225 and mARD1235/hARD1235 are probably due to alternative splicing of exon 8 (Fig. 1D), which alters the reading frame and introduces a stop codon at nucleotides 678 and 799, respectively. Arnesen et al. [17] predicted that the acetyltransferase domain is located between nucleotides 133 and 390, and a putative NLS (nuclear localization signal) (KRSHRR) and acetyl-CoA binding domain [(Q/ R)XXGX(G/A)] [18] are located at nucleotides 232–249 and 244–261, respectively (Fig. 1D). Effect of ARD1 variants on HIF-1a stability We previously reported that mARD1225 decreases the level of the HIF-1a protein [9]. To further examine the effect of other ARD1 variants on HIF-1a stability, we transfected them into HeLa cells and incubated the cells under normoxic and hypoxic conditions. As shown in Fig. 2A, mARD1225 decreased the level of HIF-1a under hypoxic conditions by approximately 60%, whereas hARD1235 and mARD1235 inhibited HIF-1a stability by 5–10%. We previously reported that mARD1225 downregulates the expression of VEGF mRNA and protein under hypoxic conditions [9]. To further test the influence of ARD1 variants on VEGF expression, we transfected HeLa cells with these variants, incubated the cells under normoxic or hypoxic conditions, and performed RT-PCR analysis. Similar to the results shown in Fig. 2A, mARD1225 strong-
425
ly decreased VEGF mRNA levels under hypoxia, as compared with the effect of hARD1235 and mARD1235 (Fig. 2B). These data indicated that mARD1225 strongly decreases HIF-1a stability and downregulates the level of VEGF mRNA, whereas hARD1235 and mARD1235 had only weak effects. Evaluation of ARD1 variants on HIF-1a acetylation mARD1225 mediates the e-acetylation of a HIF-1a lysine residue and thereby enhances the ubiquitination and degradation of HIF-1a [9]. To further evaluate the acetyltransferase activity of ARD1 variants with respect to HIF-1a, we transfected HeLa cells with these variants and performed an in vivo acetylation assay using an antiacetyl-Lys antibody after treatment with MG132, a proteasomal inhibitor, to be equal HIF-1a protein level. Consistent with the results shown in Fig. 2, mARD1225 strongly acetylated HIF-1a but hARD1235 and mARD1235 had much weaker effects (Fig. 3A). From repeated experiments, we concluded that mARD1225 increased the level of HIF1a acetylation by more than 100% as compared with the level observed in control cells but that hARD1235 and mARD1235 showed a little effect (Fig. 3B). Expression of ARD1 proteins in mammalian cell lines To determine the expression pattern of ARD1 variants in mammalian cells, we performed Western blot analysis using the ARD1 antibody. As shown in Fig. 4, immunoblots of proteins from all human cell lines (HeLa, HT1080, and H1299) tested showed major intense band of 32 kDa, which corresponds to hARD1235, whereas
Fig. 2. Effect of ARD1 variants on the expression of HIF-1a protein. (A,B) HeLa cells were transfected with plasmids encoding GFP alone (mock), GFPmARD1225 (m225), GFP-mARD1235 (m235) or GFP-hARD1235 (h235) and incubated under normoxia (N, 21% O2) and hypoxia (H, 1% O2) for 16 h. (A, left) Total cell extracts were subjected to Western blot analysis with anti-HIF-1a, anti-GFP, or anti-a-tubulin antibodies. (Right) The relative HIF-1a protein levels were determined in five independent assays and quantified by densitometry. The HIF-1a protein level under hypoxia was set to 100%. (B) RT-PCR analysis to detect the expression of VEGF and b-actin mRNA using specific primers and total RNA as a template.
426
S.-H. Kim et al. / Biochemical and Biophysical Research Communications 340 (2006) 422–427
weak bands of smaller sizes were present in mammalian cells, suggesting that these might be novel ARD1 variants (Fig. 4). Discussion
Fig. 3. Effects of ARD1 variants on HIF-1a acetylation. (A) HeLa cells were transfected with plasmids encoding GFP alone (mock), GFPmARD1225 (m225), GFP-mARD1235 (m235) or GFP-hARD1235 (h235) and treated with 10 lM MG132 for 16 h at hypoxia (1% O2) (H + MG132). Extracts were immunoprecipitated with the anti-acetylLys antibody and the immunoprecipitates were subjected to Western blot analysis to detect HIF-1a. Total cell lysates were similarly analyzed with anti-HIF-1a antibody and anti-GFP antibody as a loading control and to determine transfection efficiency. (B) The relative level of acetylated HIF1a was expressed as ratios of acetylated HIF-1a against total expressed HIF-1a, as determined in three independent assays by immunoprecipitation and Western blot analysis followed by densitometry.
Fig. 4. Expression of hARD1 proteins in mammalian cell lines. Immunoblot analysis of lysates in different cell lines, including human cervical adenocarcinoma HeLa cells, human fibrosarcoma HT1080 cells, human lung adenocarcinoma H1299 cells, and murine fibroblast NIH3T3 cells, using the anti-ARD1 antibody.
mouse NIH3T3 cells strongly expressed a 30 kDa protein corresponding to mARD1225. These results were confirmed by RT-PCR experiments with primer specificity on the variants cDNA (data not shown). Consistent with the ARD1 variants identified by RT-PCR (Fig. 1), two or three
We previously reported that mouse ARD1 interacts with the ODD domain of HIF-1a and mediates acetylation of Lys532, thereby enhancing HIF-1a ubiquitination and degradation [9]. The ARD1 protein that negatively regulates HIF-1a is mARD1225 (GenBank Accession No. BC027219), which shares the N-terminal region (aa 1– 157) with mARD1235, previously reported as GenBank Accession No. NM_019870. In contrast to our findings, recent studies have shown that hARD1235 does not affect HIF-1a stability or acetylation [13,14]. From these results, we hypothesized that ARD1 variants might differentially modulate HIF-1a stability. Consistent with this prediction, we identified several ARD1 variants by sequence alignment and RT-PCR analysis in mammalian cell lines; three mouse variants (mARD1198, mARD1225, and mARD1235) and two human variants (hARD1131 and hARD1235). Furthermore, we found that mARD1225 was highly expressed compared with mARD1235 in mouse NIH3T3 cells, whereas hARD1235 was strongly expressed in human cell lines. The mARD1235 mRNA lacks a part of exon 8 of the ARD1225 mRNA, which results in an altered reading frame that specifies 235-amino acid polypeptides. Therefore, the C-terminal region (aa 158–225) of the mARD1225 protein is completely different from that of the mARD1235 protein. Thereafter, we validated the effects of mARD1225, mARD1235, and hARD1235 in the regulation of HIF-1a stability. In HeLa cells, mARD1225 strongly decreased the expression of HIF-1a protein under hypoxic conditions, as compared with hARD1235 and mARD1235, although all three proteins have a conserved N-acetyltransferase domain (aa 45–130). This result suggests that the Cterminal region of ARD1 may be important in the regulation of HIF-1a stability. In addition, only mARD1225 strongly mediated e-acetylation of HIF-1a. These results support the idea that the acetyltransferase activity of mARD1225 downregulates HIF-1a protein stability, whereas hARD1235 and mARD1235 have a weak activity. It was also reported that ARD1 binds NAT1 forming a heterodimer in yeast and mammalian cells. Since the human NAT1 homolog Tubedown-1 has been reported to be an important regulator of retinal neovascularization [18], ARD1 may also be involved in retinal angiogenesis. Furthermore, both mARD1235 and mNAT1 are expressed in regions of cell division with a downregulation of expression as neurons mature and form appropriate synaptic connections [5]. In addition, hARD1235 and human NAT1 are regulated by caspase-mediated cleavage in stressed cells such as those undergoing apoptosis [17]. These results indicate that ARD1 variants function in a range of processes, including angiogenesis, cell proliferation, differentiation, and apoptosis, possibly through posttranslational modifi-
S.-H. Kim et al. / Biochemical and Biophysical Research Communications 340 (2006) 422–427
cation of target proteins. However, the relationships among ARD1 variants are not yet clear. Therefore, future experiments will focus on the specific roles played by ARD1 variants in different cell lines and the consequences of ARD1-dependent regulation. References [1] F.J.S. Lee, L.W. Lin, J.A.J. Smith, Na acetylation is required for normal growth and mating of Saccharomyces cerevisiae, Bacteriology 171 (1989) 5795–5802. [2] J.R. Mullen, P.S. Kayne, R.P. Moerschell, S. Tsunasawa, M. Gribskov, M. Colarito-Shepanski, M. Grunstein, F. Shetman, R. Sternglanz, Identification and characterization of genes and mutants for an N-terminal acetyltransferase from yeast, EMBO J. 8 (1989) 2067–2075. [3] E.C. Park, J.W. Szostak, Point mutations in the yeast histone H4 gene prevent silencing of the silent mating type locus HML, Mol. Cell. Biol. 10 (1990) 4932–4934. [4] E.C. Park, J.W. Szostak, ARD1 and NAT1 proteins form a complex that has N-terminal acetyltransferase activity, EMBO J. 11 (1992) 2087–2093. [5] N. Sugiura, S.M. Adams, R.A. Corriveau, An evolutionarily conserved N-terminal acetyltransferase complex associated with neuronal development, J. Biol. Chem. 278 (2003) 40113–40120. [6] C. Tribioli, M. Mancini, E. Plassart, S. Bione, S. Rivella, C. Sala, G. Torri, D. Toniolo, Isolation of new genes in distal Xq28: transcriptional map and identification of a human homologue of the ARD1 Nacetyltransferase of Saccharomyces cerevisiae, Hum. Mol. Genet. 3 (1994) 1061–1067. [7] A.F. Neuwald, D. Landsman, GCN5-related histone N-acetyltransferases belong to a diverse superfamily that includes the yeast SPT10 protein, Trends Biochem. Sci. 22 (1997) 154–155. [8] R. Strernglanz, H. Schindelin, Structure and mechanism of action of the histone acetyltransferase Gcn5 and similarity to other N-acetyltransferases, Proc. Natl. Acad. Sci. USA 96 (1999) 8807–8808.
427
[9] J.W. Jeong, M.K. Bae, M.Y. Ahn, S.H. Kim, T.K. Sohn, M.H. Bae, M.A. Yoo, E.J. Song, K.J. Lee, K.W. Kim, Regulation and destabilization of HIF-1a by ARD1-mediated acetylation, Cell 111 (2002) 709–720. [10] K. Tanimoto, Y. Makino, T. Pereira, L. Poellinger, Mechanism of regulation of the hypoxia-inducible factor-1a by the von Hippel– Lindau tumor suppressor protein, EMBO J. 19 (2000) 4298–4309. [11] Y.S. Chun, K.H. Lee, E. Choi, S.Y. Bae, E.J. Yeo, L.E. Huang, M.S. Kim, J.W. Park, Phorbol ester stimulates the nonhypoxic induction of a novel hypoxia-inducible factor 1a isoform: implications for tumor promotion, Cancer Res. 15 (2003) 8700–8707. [12] Z.N. Demidenko, A.M. Rapisarda, M. Garayoa, P. Giannakakou, G. Melillo, M.V. Blagosklonny, Accumulation of hypoxia-inducible factor-1a is limited by transcription-dependent depletion, Oncogene 24 (2005) 4829–4838. [13] T.S. Fisher, S.D. Etages, L. Hayes, K. Crimin, B. Li, Analysis of ARD1 function in hypoxia response using retroviral RNA interference, J. Biol. Chem. 280 (2005) 17749–17757. [14] R. Bilton, N. Mazure, E. Trottier, M. Hattab, M.A. Dery, D.E. Richard, J. Pouyssegur, M.C. Brahimi-Horn, Arrest-defective-1 protein, an acetyltransferase, does not alter stability of hypoxiainducible factor (HIF)-1a and is not induced by hypoxia or HIF, J. Biol. Chem. 280 (2005) 31132–31140. [15] M.K. Bae, M.Y. Ahn, J.W. Jeong, M.H. Bae, Y.M. Lee, S.K. Bae, J.W. Park, K.R. Kim, K.W. Kim, Jab1 interacts directly with HIF-1a and regulates its stability, J. Biol. Chem. 277 (2002) 9–12. [16] V. Brenner, G. Nyakatura, A. Rosenthal, M. Platzer, Genomic organization of two novel genes on human Xq28: compact head to head arrangement of IDHc and TRAPd is conserved in rat and mouse, Genetics 44 (1997) 8–14. [17] T. Arnesen, D. Anderson, C. Baldersheim, M. Lanotte, J.E. Varhaug, J.R. Lillehaug, Identification and characterization of the human ARD1–NATH protein acetyltransferase complex, Biochem. J. 386 (2005) 433–443. [18] D.S. Wall, R.L. Gendron, W.V. Good, E. Miskiewicz, M. Woodland, K. Leblanc, H. Paradis, Suppressed expression of Tubedown-1 in retinal neovascularization of proliferative diabetic retinopathy, Invest. Ophthalmol. Vis. Sci. 45 (2004) 3704–3712.