Spermidine/Spermine-N1-Acetyltransferase 2 Is an Essential Component of the Ubiquitin Ligase Complex That Regulates Hypoxia-inducible Factor 1α*

Hypoxia-inducible factor 1 (HIF-1) is a heterodimeric transcription factor that functions as a master regulator of oxygen homeostasis. The HIF-1α subunit is subjected to O2-dependent prolyl hydroxylation leading to ubiquitination by the von Hippel-Lindau protein (VHL)-Elongin C ubiquitin-ligase complex and degradation by the 26 S proteasome. In this study, we demonstrate that spermidine/spermine-N1-acetyltransferase (SSAT) 2 plays an essential role in this process. SSAT2 binds to HIF-1α, VHL, and Elongin C and promotes ubiquitination of hydroxylated HIF-1α by stabilizing the interaction of VHL and Elongin C. Multivalent interactions by SSAT2 provide a mechanism to ensure efficient complex formation, which is necessary for the extremely rapid ubiquitination and degradation of HIF-1α that is observed in oxygenated cells.

Oxygen homeostasis represents a critical organizing principle of metazoan evolution and biology (1,2). Hypoxia-inducible factor 1 (HIF-1) 2 functions as a master regulator of oxygen homeostasis in metazoan species as diverse as Caenorhabditis elegans, an organism of Ͻ10 3 cells with no specialized systems for O 2 delivery, to Homo sapiens, an organism with complex respiratory and circulatory systems to capture O 2 and deliver it to each of Ͼ10 13 cells (1,(3)(4)(5)(6)(7). Precise moment-to-moment matching of O 2 supply and demand is required for maintenance of cellular energetics and redox equilibrium that in turn is necessary for the survival of individual cells and of the organism (2, 8). To achieve this essential task, HIF-1 regulates the transcription of hundreds of human genes (9, 10) encoding proteins that are required for proper development of the respiratory and circulatory systems (3,11,12) and play essential roles in adaptive physiological responses to hypoxia and ischemia in postnatal life (13)(14)(15).
Delineation of the complex network of interacting proteins that regulates HIF-1 activity has progressed rapidly since the discovery (16), biochemical purification (17), and determination of the nucleic acid sequences encoding HIF-1 (18). HIF-1 is a heterodimer composed of a constitutively expressed HIF-1␤ subunit and a HIF-1␣ subunit, the expression of which increases dramatically as cellular O 2 concentration declines (17)(18)(19). HIF-1␣ levels are low in oxygenated cells because of the rapid proteasomal degradation of the protein in response to its modification by an E3 ubiquitin-protein ligase complex containing the von Hippel-Lindau (VHL) protein (20,21). VHL binds to HIF-1␣ and to Elongin C, which recruits Elongin B, Cullin 2, Rbx 1, and other components of the ubiquitin ligase complex (22).
VHL only binds after the hydroxylation of HIF-1␣ on Pro 402 and/or Pro 564 (4,(23)(24)(25)(26) by HIF-1␣ prolyl hydroxylase domain proteins (PHDs) 1-3, which are dioxygenases that utilize O 2 to hydroxylate HIF-1␣ in a reaction that also consumes ␣-ketoglutarate and generates succinate and CO 2 as by-products (27). Hydroxylase activity is inhibited as cellular O 2 concentration declines, either as a result of substrate limitation (4,28,29) or because of the effect of increased mitochondrial production of reactive oxygen species in hypoxic cells (30 -32). The activity of PHD2 is primarily responsible for the maintenance of low HIF-1␣ levels in oxygenated cells (33). Although PHD2 and HIF-1␣ bind to each other, efficient PHD2 activity is dependent on OS-9, a protein that binds to both PHD2 and HIF-1␣, thereby insuring stable complex formation (34). In addition to PHD inhibition, the stability of HIF-1␣ under hypoxic conditions is also dependent upon the binding of HSP90. 17-Allyl-17-aminomethylgeldanamycin, an anti-cancer agent that inhibits HSP90, induces ubiquitination and proteasomal degradation of HIF-1␣ by an alternative pathway involving RACK1 (receptor for activated C kinase 1), which in parallel to VHL binds to both HIF-1␣ and Elongin C but unlike VHL does so by an O 2 -independent mechanism (35).
Finally, the ability of HIF-1␣ to interact with the co-activator proteins p300 and CBP is blocked by O 2 -dependent * This work was supported by Public Health Service Grant R01-HL55338 from the National Heart, Lung, and Blood Institute and by the Johns Hopkins Institute for Cell Engineering. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1  hydroxylation of Asn 803 by factor inhibiting HIF-1 (36,37), which binds to HIF-1␣ and to VHL (38). Factor inhibiting HIF-1 and OS-9 were identified in a yeast two-hybrid screen for human proteins that interact with amino acid residues 576 -826 of HIF-1␣ (34,38). In the present study, we performed a yeast two-hybrid screen using HIF-1␣ residues 17-299 as bait to identify a novel and essential component of the O 2 -dependent degradation pathway that is mediated by OS-9, PHD2, and VHL.

EXPERIMENTAL PROCEDURES
Yeast Two-hybrid Screening-Bait vector pGAL4-HIF-1␣ (17-299) was constructed as described (34,38). Prey vectors were derived from a human thymus cDNA library cloned into pACT II (Clontech). Interaction of bait and prey proteins reconstitutes active GAL4, resulting in transcription of genes that mediate histidine auxotrophy (his ϩ ) and ␣-galactosidase activity. Purified his ϩ /␣-galactosidase ϩ colonies were grown in liquid medium lacking leucine to select for the presence of the prey vector and inoculated onto medium lacking leucine and supplemented with 10 g/ml cycloheximide to cure clones of the bait vector and identify prey vectors encoding a protein capable of autonomous activation of the ␣-galactosidase reporter gene (i.e. false positives). Prey vector was isolated by the glass bead method (52) for transformation of Escherichia coli DH5␣ cells and plasmid DNA isolation.
In Vitro Binding Assay Using GST Fusion Protein-E. coli BL21-Gold(DE3)pLysS (Stratagene) was transformed with pGEX expression vectors and treated for 4 h with 0.5 mM isopropyl-D-thiogalactoside. Cell lysates were applied to glutathione-Sepharose 4B beads (GE Healthcare). Glutathione S-transferase (GST) fusion proteins were eluted with 10 mM reduced glutathione in 50 mM Tris-HCl (pH 8.0). [ 35 S]Methionine-labeled proteins were generated in reticulocyte lysates using the TNT T7 coupled transcription/translation system (Promega). Ten l of in vitro translated 35 S-labeled protein was mixed with 4 g of GST or GST-HIF-1␣ fusion protein (34,38) in PBS-T binding buffer (Dulbecco's PBS, pH 7.4, 0.1% Tween 20) at 4°C for 2 h, followed by the addition of 20 l of glutathione-Sepharose 4B beads. After 30 min of mixing, the beads were washed with PBS-T. The proteins were eluted in Laemmli sample buffer and analyzed by SDS-PAGE followed by autoradiography using Molecular Imager FX (Bio-Rad).
Tissue Culture-Human 293 and 293T cells were cultured as described (53). The cells were maintained at 37°C in a 5% CO 2 , 95% air incubator. For hypoxic exposures, the cells were placed in a modulator incubator chamber (Billups-Rothenberg) that was flushed with 1% O 2 , 5% CO 2 , balance N 2 , sealed, and incubated at 37°C.

Construction of Expression Vectors-
The open reading frame of SSAT2 was amplified from cDNA prepared from human umbilical vein endothelial cell mRNA using specific primers based on the nucleotide sequence of GenBank TM accession number NM_133491. The PCR product was inserted into pcDNA3.1(D)/V5-His-TOPO vector (Invitrogen). For generation of SSAT2 deletion mutants, PCRs were performed using pcDNA3.1(D)/V5-His-SSAT2 as template and cloned into pcDNA3.1(D)/V5-His-TOPO vector. The coding sequences of SSAT2 were amplified with primers containing SmaI and XhoI restriction enzyme sites using pcDNA3.1(D)/V5-His-SSAT2 as template and ligated into SmaI-and XhoI-digested pGEX-5X-1 (GE Healthcare). The expression vectors for wild type and deletion mutants of HIF-1␣ were described previously (39,54). The open reading frame of VHL was amplified by PCR using pcDNA3.1-FLAG-VHL as template (46) (kindly provided by J. Frydman, Stanford University) and ligated into HindIII-and XbaI-digested p3XFLAG-CMV-7 (Sigma).
shRNA Assays-The mammalian expression vector, pSR. retro.GFP.Neo.circular.stuffer (Oligo Engine) was used for expression of shRNA in 293 cells. The SSAT2 shRNA insert consisted of the 19-nucleotide sequence tgtgatgccggaatatcgg, corresponding to nucleotides 284 -302 of SSAT2 mRNA (GenBank TM accession number NM_133491), which was separated by a spacer (ttcaagaga) from the reverse complement of the same 19-nucleotide sequence. A scrambled negative control vector, constructed using a 19-nucleotide sequence (acgcatgcatgcttgcttt) with no significant similarity to any mammalian gene sequence, served as a control. Oligonucleotides were annealed and ligated into BglII-and Hin-dIII-digested vector. 293 cells were analyzed by fluorescence microscopy and lysed for RNA and protein isolation 24 h after transfection.
Reverse Transcriptase-PCR Assays-Total RNA was extracted from cells using RNeasy Mini Kit (Qiagen) and treated with DNase. Five g of total RNA were used for firststrand synthesis with iScript cDNA Synthesis system (Bio-Rad). cDNA was used for PCR analysis of SSAT2, HIF-1␣, and VEGF mRNA and 18 S rRNA (primer sequences available upon request). PCR products were analyzed by 2% agarose/ethidium bromide gel electrophoresis.
Site-directed Mutagenesis-PCR-based mutagenesis was performed using pcDNA3.1(D)/V5-His-SSAT2 as template. The cycling parameters were 95°C for 60 s, 58°C for 60 s, and 68°C for 6 min for 18 cycles using Pfu polymerase (Stratagene). Subsequently, plasmid DNA was digested with 10 units DpnI (Stratagene) for 1 h at 37°C, before an a 10-l aliquot was transformed into E. coli DH-5␣ cells. The mutations were verified by nucleotide sequencing.
Statistical Analysis-The data are presented as the means Ϯ S.D. Differences between experiments were analyzed for statis-tical significance (p Ͻ 0.05) by analysis of variance or two-sample t test.

RESULTS
Interaction of SSAT2 with HIF-1␣-The bait for two-hybrid screening was a fusion protein containing the DNA-binding domain of the yeast transcription factor GAL4 fused to the basic helix-loop-helix and PER-ARNT-SIM homology domains of HIF-1␣ (Fig. 1A), which are required for dimerization and DNA binding of the HIF-1␣:HIF-1␤ heterodimer (18,39). From a screen of 1 ϫ 10 6 yeast transformants, 23 prey fusion proteins, consisting of the GAL4 transactivation domain and amino acid sequences encoded by cDNAs prepared from human thymus mRNA, were identified that interacted with the bait fusion protein. The human sequence in one prey protein was that of fulllength (170 amino acid residues) spermidine/spermine-N 1 -acetyltransferase 2 (SSAT2; Unigene accession Hs.10846). SSAT2 is a protein that was identified based on 46% amino acid identity with SSAT1 (Unigene accession Hs.28491), which is a bona fide spermidine/spermine-N 1acetyltransferase (40). Unlike SSAT1, SSAT2 does not function in polyamine catabolism, but instead acetylates thialysine (S-(2-aminoethyl)-L-cysteine), a naturally occurring modified amino acid (41)(42)(43).
Effect of SSAT2 on HIF-1␣ Expression-Compared with cells expressing FLAG-HIF-1␣ alone, HIF-1␣ protein levels were markedly reduced in cells in which V5-SSAT2 was co-expressed ( Fig. 2A). The inhibitory effect of SSAT2 on HIF-1␣ expression was blocked by treatment with MG132 (Fig. 2B), indicating that SSAT2 induced the proteasomal degradation of HIF-1␣. Overexpression of SSAT2 also inhibited transcription of a HIF-1-regulated reporter gene in a dose-dependent man-FIGURE 1. SSAT2 interacts with HIF-1␣ in vitro and in human cells. A, for yeast two-hybrid screening, the bait vector encoded a chimeric protein consisting of the DNA-binding domain from the yeast GAL4 transcription factor (GAL4 DBD) fused to amino acid residues 17-299 of HIF-1␣. The prey vectors encoded the GAL4 transactivation domain (TAD) fused to amino acid residues that were encoded by cDNAs prepared from human thymus mRNA. B and C, identification of HIF-1␣ sequences required for binding to SSAT2 in vitro. GST alone and GST fusion proteins containing the indicated residues of HIF-1␣ were purified from E. coli, incubated with 35 S-labeled in vitro translated SSAT2, captured on glutathione-Sepharose beads, and analyzed by SDS-PAGE and autoradiography (top panels). T/I, 10% of total input of 35 S-labeled in vitro translated SSAT2 was analyzed directly by SDS-PAGE. GST and fusion proteins were analyzed by SDS-PAGE and immunoblot assay using anti-GST antibody as input control (bottom panels). D, human 293 cells were transfected with expression vectors encoding FLAG-HIF-1␣, V5-SSAT2, or empty vector and treated 10 M of MG132 for 6 h. WCL were prepared, and immunoprecipitation (IP) was performed using anti-V5 antibody-conjugated agarose beads. Immunoblot assays were performed with antibodies against FLAG (top panels) or V5 (bottom panels) using 80 g of WCL directly (panels on right) or after IP of 500 g (panels on left). E, 293 cells were transfected with expression vector encoding V5-SSAT2 or empty vector and exposed to 1% or 20% O 2 for 4 h after pretreatment with 10 M MG132 for 2 h. Immunoblot assays were performed with anti-HIF-1␣ (top panels) and anti-V5 (bottom panels) antibodies using 150 g of WCL directly (panels on right) or after IP from 750 g of WCL (panels on left). bHLH, basic helix-loop-helix.

SSAT2 Promotes VHL-mediated HIF-1␣ Ubiquitination
ner (Fig. 2C). Conversely, transfection of cells with an expression vector encoding both green fluorescent protein (Fig. 2D) and a shRNA that targeted SSAT2 mRNA for degradation (Fig.  2E) induced transcription of the HIF-1-dependent reporter gene in oxygenated cells (Fig. 2F). Thus, both gain-of-function and loss-of-function studies implicate SSAT2 as a negative regulator of HIF-1␣ protein levels under nonhypoxic conditions. SSAT2 overexpression also inhibited the induction of endogenous HIF-1␣ protein expression (Fig. 3A) and HIF-1-dependent gene transcription (Fig. 3B) under hypoxic conditions. Conversely, SSAT2 knockdown by shRNA increased HIF-1␣ protein expression (Fig. 3C) and HIF-1-dependent gene transcription (Fig. 3D) under hypoxic conditions. SSAT2 overexpression blocked the induction of endogenous VEGF mRNA expression in response to hypoxia (Fig. 3E, top panels). In contrast, neither hypoxia nor SSAT2 overexpression affected HIF-1␣ or ␤-actin mRNA levels (Fig. 3E, bottom panels). Taken together, these results indicate that SSAT2 reduces the levels of HIF-1␣ protein that are available to associate with HIF-1␤ to form the transcriptionally active heterodimer, thus reducing the expression of HIF-1-regulated genes such as VEGF.
Regulation of HIF-1␣ by SSAT2 Is Dependent upon Prolyl Hydroxylation-To determine whether the activity of SSAT2 was linked to the O 2 -dependent degradation pathway, we analyzed HIF-1␣ protein levels and HIF-1␣-dependent gene transcription in cells co-expressing SSAT2 and full-length or deletion mutants of HIF-1␣. The levels of full-length HIF-1␣ and deletion mutants 1-754 and 1-608, which retain the property of O 2 -dependent degradation (44), were reduced by SSAT2 overexpression (Fig. 4A), as was their ability to activate transcription of a HIF-1dependent reporter gene (Fig. 4B). In contrast, the expression and transcriptional activity of deletion mutant 1-391/576 -826, which is not subject to O 2 -dependent degradation (44), were not significantly reduced by SSAT2 overexpression. Both PHD2 and SSAT2 reduced transcription mediated by wild type HIF-1␣, but they did not inhibit transcription mediated by the P402A/P564A double mutant or P402A/P564A/N803A triple mutant form of HIF-1␣ (Fig. 4C). Remarkably, the reduction in HIF-1␣-dependent gene transcription mediated by PHD2 overexpression was completely blocked by expression of shRNA directed against SSAT2 (Fig. 4D). These data demonstrate that negative regulation of HIF-1␣ by PHD2 is SSAT2-dependent and that negative regulation of HIF-1␣ by SSAT2 requires the presence of proline residues that are hydroxylated by PHD2.
Next, we investigated whether SSAT2 was required for VHL activity. As expected, the effect of SSAT2 overexpression on HIF-1␣ expression and HIF-1␣-dependent gene transcription was blocked by SSAT2 shRNA (Fig. 5). However, SSAT2 shRNA also blocked the effect of FLAG-VHL on HIF-1␣ expression and HIF-1␣-dependent gene transcription. Thus, negative regulation of HIF-1␣ by both PHD2 and VHL is SSAT2-dependent.
Mutation of Conserved Residues in SSAT2 Impairs Regulation of HIF-1␣-SSAT2 has been reclassified as a thialysine N ⑀ -acetyltransferase and homologous proteins in Schizosaccharomyces pombe, C. elegans, and Leishmania major have been identified (45). Ser 82 and Thr 83 of SSAT2 are conserved in all three orthologues, and mutation of these residues in the  (1 and 2 g). D, after 37 h, transfected cells were analyzed by phase contrast microscopy to visualize cells and fluorescence microscopy to detect GFP expression. E, the transfected cells were exposed to 20% or 1% O 2 for 24 h. VEGF and SSAT2 mRNA and 18 S rRNA expression were analyzed by reverse transcription-PCR. F, cells were co-transfected with pSV-Renilla, p2.1, and expression vector encoding both GFP and either SNC shRNA (2 g) or SSAT2 shRNA (1 and 2 g). After 37 h, luciferase activity was determined as described above. *, p Ͻ 0.05. AUGUST 10, 2007 • VOLUME 282 • NUMBER 32

JOURNAL OF BIOLOGICAL CHEMISTRY 23575
Leishmania protein resulted in a loss of thialysine N ⑀ -acetyltransferase activity (45). Compared with wild type SSAT2, the S82D/T83A double mutant was significantly less effective at reducing HIF-1␣ protein levels and inhibiting HIF-1-dependent reporter gene transcription (Fig. 6). However, cells transfected with SSAT2(S82D/ T83A) expression vector had significantly reduced HIF-1␣ protein levels and HIF-1␣-dependent reporter gene transcription compared with cells transfected with empty vector. In contrast, the mutation did not affect the interaction of SSAT2 with HIF-1␣ protein (Fig.  6C). These results suggest that SSAT2 acetyltransferase activity may contribute to its ability to reduce HIF-1␣ protein levels, but it does not affect SSAT2-HIF-1␣ protein-protein interaction.
SSAT2 Promotes VHL-Elongin C Interaction and HIF-1␣ Ubiquitination-To investigate other mechanisms by which SSAT2 may promote O 2 -dependent degradation of HIF-1␣, the effect of SSAT2 on the interaction of HIF-1␣ and VHL was analyzed. Incubation of GST-VHL with lysates from cells overexpressing SSAT2 did not increase the specific binding of in vitro translated HIF-1␣ to GST-VHL as compared with control lysates (Fig. 7A). Similarly, incubation of GST-HIF-1␣ with lysates from cells overexpressing SSAT2 did not promote the binding of in vitro translated VHL to GST-HIF-1␣ as compared with control lysates (Fig. 7B). Finally, the results of co-immunoprecipitation assays indicated that SSAT2 did not promote the interaction of HIF-1␣ and VHL in 293 cells (Fig. 7C). Taken together, these results demonstrate that SSAT2 does not promote hydroxylation of HIF-1␣ or the interaction of hydroxylated HIF-1␣ with VHL. In contrast, the incubation of in vitro translated Elongin C with in vitro translated SSAT2 increased the interaction of Elongin C with GST-VHL as compared with Elongin C that was incubated with FIGURE 3. Effect of SSAT2 on induction of HIF-1␣ in response to hypoxia. A, 293 cells were transfected with EV or plasmid encoding V5-SSAT2, and exposed to 20% or 1% O 2 for 24 h. WCL were subjected to immunoblot assay using anti-HIF-1␣, V5, or actin antibodies. B, cells were co-transfected with pSV-Renilla, p2.1, and EV or vector encoding V5-SSAT2, and exposed to 20% or 1% O 2 for 24 h. The ratio of firefly:Renilla luciferase activity was determined. The results were normalized to those from cells transfected with EV and exposed to 20% O 2 . The means and S.D. based on three independent transfections are shown. *, p Ͻ 0.05 compared with cells transfected with EV at 1% O 2 . C, cells were transfected with plasmid encoding SNC shRNA or SSAT2 shRNA and exposed to 20% or 1% O 2 for 24 h. WCL were subjected to immunoblot assay using anti-HIF-1␣, HIF-1␤, V5, or ␤-actin antibodies. A short and a long chemiluminescence exposure (exp) are shown to compare differences in HIF-1␣ expression at 1% and at 20% O 2 , respectively. D, cells were co-transfected with pSV-Renilla, p2.1, and vector encoding SNC shRNA or SSAT2 shRNA, and exposed to 20% or 1% O 2 for 24 h. Luciferase activity was determined as described above. *, p Ͻ 0.05 compared with cells transfected with SNC shRNA at 1% O 2 . E, cells were transfected with EV or vector encoding V5-SSAT2 and exposed to 20% or 1% O 2 for 24 h. VEGF, SSAT2, HIF-1␣, and ␤-actin mRNA expression were analyzed by reverse transcription-PCR.  1-826), or HIF-1␣ deletion mutant containing the indicated residues. Whole cell lysates were subjected to immunoblot assays to detect HIF-1␣, V5, and ␤-actin. B, 293 cells were co-transfected with: p2.1; pSV-Renilla; EV or vector encoding V5-SSAT2; and EV, full-length HIF-1␣, or the HIF-1␣ deletion mutant containing the indicated residues. After 24 h, the ratio of luciferase activities was determined. The results were normalized to those from cells transfected with EV. The means and S.D. based on three independent transfections are shown. C, cells were co-transfected with pSV-Renilla, p2.1, and EV or vector encoding V5-SSAT2, wild type HIF-1␣ (WT), HIF-1␣ double mutant (DM) P402A/ P564A, or triple mutant (TM) P402A/P564A/N803A and analyzed as described above. *, p Ͻ 0.05. D, cells were co-transfected with: pSV-Renilla; p2.1; EV or vector encoding PHD2; and SNC shRNA (shSNC) or SSAT2 shRNA (shSSAT2) and analyzed as described above. *p Ͻ 0.05. in vitro translation reaction programmed by empty vector (Fig.  7D). Furthermore, when V5-SSAT2 was co-expressed with FLAG-VHL, the recovery of endogenous Elongin C in anti-FLAG immunoprecipitates was increased compared with cells in which FLAG-VHL was expressed alone (Fig. 7E). Similarly, when SSAT2 was co-expressed with FLAG-HIF-1␣, the recov-ery of Elongin C in anti-HIF-1␣ immunoprecipitates was increased compared with cells in which FLAG-HIF-1␣ was expressed alone (Fig. 7F). In contrast, Elongin C was not recovered in anti-HIF-1␣ immunoprecipitates from cells expressing the FLAG-tagged double mutant (P402A/P564A) form of HIF-1␣, which is not competent for binding to VHL (Fig. 7F). These data demonstrate that SSAT2 increases the stability of VHL-Elongin C complexes in oxygenated cells by direct interaction with both proteins.
We hypothesized that by stabilizing the VHL-Elongin C interaction, SSAT2 promotes the ubiquitination of HIF-1␣. To test this hypothesis, we analyzed the extent of HIF-1␣ ubiquitination in cells transfected with expression vectors encoding HA-tagged ubiquitin and FLAG-HIF-1␣ plus empty vector or vector encoding PHD2 or V5-SSAT2. The cells were treated with the proteasome inhibitor MG132 to block degradation of ubiquitinated FLAG-HIF-1␣, which was identified in anti-FLAG immunoprecipitates by immunoblot assay using anti-HA antibodies. SSAT2 overexpression was comparable with PHD2 overexpression in promoting the ubiquitination of HIF-1␣ (Fig. 9). Thus, as in the case of PHD2, SSAT2 overexpression is sufficient to increase the ubiquitination of HIF-1␣.

DISCUSSION
In this paper we demonstrate that SSAT2 is a novel and essential component of the multi-protein ubiquitin ligase complex that regulates HIF-1␣ stability as a function of the cellular O 2 concentration. Just as OS-9 promotes the interaction of HIF-1␣ with PHD2 by interaction with both proteins to facilitate hydroxylation, SSAT2 promotes the interaction of Elongin C with VHL to facilitate ubiquitination of HIF-1␣ (Fig. 10). The multivalent interactions of SSAT2 with HIF-1␣, VHL, and Elongin C suggest the formation of higher order complexes in which VHL-Elongin C binding is specifically stabilized. It is unclear whether SSAT2 functions in the general assembly of the VHL-Elongin C ubiquitin ligase complex, as does  Whole cell lysates were analyzed by immunoblot assay with antibodies that recognize FLAG, V5 or ␤-actin. C, cells were transfected with EV or vector encoding wild type V5-SSAT2 (WT), or mutant V5-SSAT2 (S82D/T83A) and treated with 10 M MG132 for 6 h. WCL were prepared, and immunoprecipitation (IP) was performed using control IgG (con) or anti-FLAG Ab. IP (panels on top) and WCL (panels on bottom) were subjected to immunoblot assays using Ab against FLAG, V5, or ␤-actin. the chaperonin TriC (45), or whether it acts only after VHL has bound to hydroxylated HIF-1␣. The data in Fig. 7D demonstrate that SSAT2 overexpression can stabilize the interaction of VHL and Elongin C in the absence of HIF-1␣. The remarkable complexity of protein interactions required for the efficient hydroxylation, ubiquitination, and subsequent proteasomal degradation of HIF-1␣ in oxygenated cells points to the critical importance of this process for cellular homeostasis and provides a molecular basis for the extremely rapid hypoxic induction and post-hypoxic decay of HIF-1␣ (47,48).
Interestingly, SSAT2 was recently shown to interact with the p65 subunit of NF-B and function as a co-activator to stimulate NF-Bdependent transcriptional activation (49), although the precise molecular mechanisms were not determined. Thus, SSAT2 inhibits HIF-1␣ and activates NF-B, two of the most important stress-induced transcription factors in metazoan species, by regulating protein stability and transactivation function, respectively.
A major unanswered question is whether, in addition to the physical interactions of SSAT2 with HIF-1␣, VHL, and Elongin C, the acetyltransferase activity of SSAT2 contributes to its mechanism of action as a negative regulator of HIF-1␣ protein stability. SSAT2 shares 46% amino acid identity with SSAT1 but is only distantly related to other known members of the GCN5-related N-acetyltransferase family (42). Unlike SSAT1, SSAT2 levels are not induced by polyamines, and SSAT2 cannot efficiently acetylate polyamines (41,42). A candidate physiological substrate for SSAT2 is thialysine (S-aminoethyl-L-cysteine), which is formed by the reaction of L-serine and cysteamine that is catalyzed by cystathionine-␤-synthase (50). Thialysine, which is structurally similar to L-lysine, with a sulfur atom instead of a 4-methylene group in its side chain, acts as an FIGURE 7. SSAT2 potentiates the interaction between VHL and Elongin C. A and B, GST and GST-VHL or GST-HIF-1␣ were incubated for 5 min at 30°C with lysate (WCL) from cells transfected with empty EV or SSAT2 vector, then incubated with 35 S-labeled in vitro transcribed and translated (IVTT) HIF-1␣ or VHL, captured on glutathione-Sepharose beads, and analyzed by SDS-PAGE and autoradiography. Input, 10% of total input of HIF-1␣ or VHL was analyzed directly by SDS-PAGE. GST proteins were analyzed by SDS-PAGE and immunoblot assay using anti-GST antibody (bottom panels). C, cells were transfected with EV or vector encoding HIF-1␣, FLAG-VHL, or V5-SSAT2 and treated with 10 M MG132 for 6 h. WCL were prepared, and immunoprecipitation was performed using control IgG (con) or anti-FLAG Ab. Immunoprecipitation (panels on right) and WCL (panels on left) were subjected to immunoblot assays using Ab against HIF-1␣, FLAG, V5, or ␤-actin. D, IVTT-Elongin C was incubated for 30 min at 30°C with IVTT reaction products programmed by EV or SSAT2 vector, then incubated with GST or GST-VHL, captured on glutathione-Sepharose beads, and analyzed by SDS-PAGE and autoradiography. GST proteins and IVTT products from EV or SSAT2 vector were analyzed by SDS-PAGE and immunoblot assay using anti-GST, V5, or ␤-actin antibody. E, cells were transfected with EV or vector encoding FLAG-VHL or V5-SSAT2. WCLs were prepared, and immunoprecipitation was performed using control IgG (con) or anti-FLAG Ab. Immunoprecipitation (panels on right) and WCL (panels on left) were subjected to immunoblot assays using Ab against Elongin C, FLAG, V5, or ␤-actin. F, cells were transfected with EV or vector encoding V5-SSAT2, wild type (WT) FLAG-HIF-1␣, or mutant (P402/564A) FLAG-HIF-1␣ and treated with 10 M MG132 for 6 h. WCLs were prepared, and immunoprecipitation was performed using control IgG (con) or anti-HIF-1␣ Ab. Immunoprecipitation (panels on right) and WCL (panels on left) were subjected to immunoblot assays using Ab against FLAG, Elongin C, V5, or ␤-actin.  1-329), GST-VHL, or GST-Elongin C was incubated with lysates from the cells transfected with empty vector or expression vector encoding full-length (1-170) SSAT2 or SSAT2 deletion mutant containing the indicated residues. GST fusion proteins were captured on glutathione-Sepharose beads and analyzed by SDS-PAGE. A, whole cell lysates were subjected to immunoblot assays using antibodies that recognize V5 and ␤-actin. B, proteins pulled down from whole cell lysates by the indicated GST fusion protein were subjected to immunoblot assays using antibodies that recognize V5 and GST.
anti-metabolite by competing with L-lysine for incorporation into polypeptides and altering their tertiary structure and function (45). Thialysine is metabolized to cyclic ketimine derivatives, which are present in human plasma, urine, and cultured cells (51).
Ser 82 and Thr 83 of SSAT2 are conserved in all known thialysine N ⑀ -acetyltransferases (but not in SSAT1), and mutation of these residues in the Leishmania orthologue of SSAT2 resulted in the loss of thialysine acetyltransferease activity (45). Mutation of these residues in SSAT2 reduced but did not eliminate the ability of SSAT2 to promote HIF-1␣ degradation, suggesting that the inhibitory activity of SSAT2 may be due to the combined effect of its acetyltransferase activity and physical interactions with HIF-1␣, VHL, and Elongin C. The orthologues of SSAT2 in Leishmania and C. elegans can also utilize as a substrate 5-hydroxy-L-lysine (45), a modified amino acid that is generated in by the action of lysine hydroxylases, which are mechanistically similar to the PHDs in utilizing O 2 and ␣-keto-glutarate as co-substrates. Further studies are required to investigate the intriguing possibility that, just as VHL recognizes a modified prolyl residue in HIF-1␣, SSAT2 may recognize a modified amino acid in HIF-1␣, VHL, or Elongin C to promote the ubiquitination of HIF-1␣ by the VHL-Elongin C ubiquitinligase complex. FIGURE 9. SSAT2 promotes the ubiquitination of HIF-1␣. 293 cells were co-transfected with vectors encoding HA-ubiquitin, FLAG-HIF-1␣, and empty vector or expression vector encoding V5-SSAT2 or PHD2. After treatment with 10 M MG132 for 6 h, WCL were prepared, and immunoprecipitation (IP) was performed using anti-FLAG or anti-HA Ab. WCL (A) and IP (B) were subjected to immunoblot (IB) assays using anti-HA, FLAG, PHD2, V5, or ␤-actin Ab. FIGURE 10. Multivalent protein complexes regulate hydroxylation and ubiquitination of HIF-1␣. A, OS-9 promotes HIF-1␣ hydroxylation through trivalent complex formation with PHD2 and HIF-1␣. B, SSAT2 promotes HIF-1␣ ubiquitination through tetravalent complex formation with HIF-1␣, VHL, and Elongin C, which enhances the interaction of VHL and Elongin C, leading to efficient ubiquitination of hydroxylated (OH) HIF-1␣.