Crystal Structure of Homo Sapiens Thialysine Nϵ‐acetyltransferase (hsssat2) in Complex with Acetyl Coenzyme A

HsSSAT1 is an important enzyme that converts polyamines such as spermidine or spermine into acetylpolyamines. Because polyamines are major regulators of cell growth and differentiation,1-3 HsSSAT1 has been considered as a target for cancer treatment. Gene locus BC011751 from Homo sapiens has been annotated as HsSSAT24 because of its high sequence identity (46%) and similarity (61%) to HsSSAT1. Recent biochemical studies showed that HsSSAT2 does not transfer acetyl group of AcCoA to polyamines, such as spermidine or spermine, but rather thialysine [2-amino-3-(2-aminoethylsulfanyl)propanoic acid] which is a much better substrate of this enzyme.5 However, the physiological role of HsSSAT2 is still unclear. Thialysine is a structural analog of L-lysine, and has its role as an antimetabolite by competing with L-lysine for the incorporation into polypeptides, which makes proteins inactive.6, 7 Therefore, high thialysine to lysine ratios block the growth of prokaryotic and eukaryotic cells.8-11 In addition to the antimetabolite function, metabolites formed from thialysine have been characterized and identified in mammalian tissues including brain.12, 13 Thialysine can be further converted to the cyclic ketimine, AECK,12 TMA,14 and AECK-DD.15, 16 The functions of AECK, TMA, and AECK-DD in the cell are not clear. However, it has been proposed that AECK and TMA serve neurochemical roles, and AECK-DD has a role in the modulation of oxidative processes in vivo.17 Herein, we report the three-dimensional structure of HsSSAT2 in complex with AcCoA at a resolution of 1.8 Å. SSAT, spermidine/spermine N1-acetyltransferase; AcCoA, acetyl coenzyme A; AECK, S-(aminoethyl)-L-cysteine ketimine; TMA, 4-thiomorpholine-3,5-dicarboxylic acid; AECK-DD, AECK-decarboxylated dimer; SeMet, selenomethionine; TCEP, tris(2-carboxylethyl) phosphine; Bis-Tris, bis(2-hydroxyethyl)imino-tris(hydroxymethyl)methane; MEPEG, methyl ether polyethylene glycol; MOPS, 3-(N-morpholino) propanesulfonic acid; PDB, Protein Data Bank; MAD, multi-wavelength anomalous diffraction; GNAT, GCN5-related N-acetyltransferases; ScHpa2, histone acetyltransferase from Saccharomyces cerevisiae; SeAac(6′)-Iy, aminoglycoside N-acetyltransferase from Salmonella enteritidis; ScGNA1, glucosamine-6-phosphate N-acetyltransferase 1 from Saccharomyces cerevisiae; APS, Advanced Photon Source; ANL, Argonne National Laboratory; DALI, Distance mAtrix aLIgnment; VAST, Vector Alignment Search Tool. The gene encoding the HsSSAT2 was cloned and the SeMet-labeled proteins were expressed and purified following the standard CESG pipeline protocol for cloning,18 protein expression,19 protein purification,20 and overall information management.21 Crystals of HsSSAT2 were grown by the hanging-drop vapor diffusion method from 10 mg/mL protein solution in buffer (50 mM NaCl, 3 mM NaN3, 0.3 mM TCEP, 5 mM Bis-Tris pH 6.0) mixed with an equal amount of reservoir solution containing 16.4% MEPEG 5,000, 100 mM MOPS pH 7.0, and 150 mM potassium glutamate at 277 K. Crystals grew approximately as tetragonal bipyramids with dimensions of approximately 100 × 100 × 50 μm. The selenomethionyl crystals of HsSSAT2 belong to space group P212121, with unit-cell parameters a = 54.6, b = 83.6, c = 87.5 Å. One crystal was transferred to the same reservoir solution where the crystal was grown, followed by stepwise additions of ethylene glycol up to 25%. Subsequently, the crystal was transferred to the final cryoprotectant solution (16.4% MEPEG 5 K, 100 mM MOPS pH 7.0, 150 mM potassium glutamate, and 25% ethylene glycol) for 2 s, and it was frozen by direct immersion in liquid nitrogen at 100 K. X-ray diffraction data were collected at synchrotron beam line 22-ID at the APS of the ANL. The diffraction images were integrated and scaled using HKL2000.22 The selenium substructure of SeMet-labeled HsSSAT2 crystal was determined using Hyss23 and SHELXD.24 The protein structure was phased and the initial phase information was further improved by electron-density modification using two-wavelength MAD data in autoSHARP.25 The automatic tracing procedure of ARP/wARP26 in autoSHARP25 produced an initial model with approximately 86% of residues placed, of which 79% had side-chains assigned. The structure was completed using alternate cycles of manual model building in Coot27 and Xfit,28 and refinement in REFMAC5.29 All steps were monitored using an Rfree value based on 5.0% of the independent reflections. The stereochemical quality of the final model was assessed using PROCHECK30 and MolProbity.31 The structure of HsSSAT2 in complex with AcCoA has been determined to a resolution of 1.8 Å. Data collection, phasing, refinement, and model statistics are summarized in Table I. The final model includes two HsSSAT2 monomers (residues 3–30, 35–60, and 68–169 for monomer A; residues 2–59 and 70–170 for monomer B), one AcCoA molecule, and 305 water molecules. The coordinates and structure factor files have been deposited in the PDB with accession number 2BEI. The three-dimensional structure of HsSSAT2 is a mixed α/β fold containing eight α-helices and seven β-strands [Fig. 1(a)]. The catalytic domain of HsSSAT2 has the common fold of GNAT family which is known as a central, mixed β-sheet that is mainly built up of antiparallel strands34 with one exception between parallel β4 and β5. a: Domain-swapped dimeric structure. The secondary structure elements are numbered in the order of appearance in the primary structure. One monomer with AcCoA is represented in cyan and the other monomer is in green. AcCoA is colored by atom types (carbon: gray; oxygen: red; nitrogen: blue; phosphorous: orange; sulfur: yellow). b: ScHpa2 in complex with AcCoA, the best structural homolog in the PDB as determined by DALI and VAST. ScHpa2 is colored in magenta and HsSSAT2 shows one monomer in cyan as in part (a). c: The electron densities of AcCoA in the 2Fo–Fc map at 1.5 σ level. AcCoA is colored the same as in part (a). Figures were prepared using the PyMol program (http://pymol.sourceforge.net/). We found significant extra electron density from the map calculated from the initial protein model using autoSHARP. The extra electron density looked like an AcCoA in a large hydrophobic cleft located at the site where the two parallel strands, β4 and β5, diverge because of a β-bulge in strand β4 [Fig. 1(c)]. Leu91 and Glu92 in this β-bulge are well conserved among GNAT families including Homo sapiens SSAT1 and other SSAT2 sequences [Fig. 2(b)]. HsSSAT2 acquired the AcCoA from Escherichia coli cells used for overexpression of HsSSAT2. Only one of two molecules in asymmetric unit contained the AcCoA, and this feature is unique among all acetyltransferase structures that contain coenzyme A or AcCoA in the PDB. This naturally acquired AcCoA in only one active site of the dimer suggests that each monomer of HsSSAT2 dimer may participate in "half-the-sites" reactivity, but we have no experimental evidence to support this. The AcCoA interactions with HsSSAT2 and the sequence alignment of SSATs. a: The interactions between HsSSAT2 and AcCoA. AcCoA is colored the same as in part (a) and hydrogen bonds are marked by green arrows. The red dot represents a water. b: The overall sequence alignment results of HsSSAT2, HsSSAT1, and other SSAT2s. Conserved, identical, similar, and different residues are colored in red, blue, cyan, and black, respectively. AcCoA interacting sites are identified by filled circles, and the proposed key thialysine contact residue (Ala128) is identified by a filled triangle. Residues forming the conserved β-bulge are marked with a green triangle, and four GNAT sequence motifs are boxed and labeled. The abbreviations for SSAT species are used: Hs, Homo sapiens (human); Bt, Bos taurus (domestic cow); Cf, Canis familiaris (domestic dog); Ss, Sus scrofa (wild pig); Mm, Mus musculus (house mouse); Rn, Rattus norvegicus (Norway rat). HsSSAT2 shares the {R/Q}-X-X-G-X-G sequence motif (R101-X-X-G104-X-G106 in HsSSAT2) for AcCoA recognition and binding35 with other GNAT superfamily, and AcCoA bends at the pyrophosphate moiety and at the pantetheine moiety. The pyrophosphate moiety of AcCoA mainly interacts with the GNAT sequence motif, and the pantetheine moiety of AcCoA interacts with β4 (I94 and V96) and α6 (N133 and Y140) through hydrogen bonds. The interactions between HsSSAT2 and AcCoA are described in Figure 2(a). Structural homology searches using the DALI server36 and VAST (http://www.ncbi.nlm.nih.gov/Structure/VAST/vast.shtml) frequently produced the structure of the ScHpa2 in complex with AcCoA (PDB ID: 1QSM)37 as the closest structural homolog of HsSSAT2, with Z = 19.5, RMSD = 2.1 Å over 143 aligned residues and 22% sequence identity, and Figure 1(b) shows the structure alignment between HsSSAT2 and ScHpa2. The SeAac(6′)-Iy in complex with coenzyme A and ribostamycin (PDB ID: 1S3Z)38 is the second best structural homolog of HsSSAT2 from DALI, with Z = 18.0, RMSD = 2.3 Å over 141 aligned residues, and 17% sequence identity. After our HsSSAT2 structure was released, HsSSAT1 structures were deposited to the PDB, which have high structural homology to HsSSAT2 (Z = 21.2, RMSD = 1.9 Å over 154 aligned residues, and 47% sequence identity for entry 2B5G, for example). From the structure alignment with the SeAac(6′)-Iy which contains the substrate ribostamycin, we suggest that the backbone carbonyl of Ala128 has an important role in recognizing the nucleophilic amine group of thialysine by a hydrogen bond that enhances the nucleophilic character of the amine. Interestingly, the intermolecular interaction between two monomers of HsSSAT2 shows a pattern of domain swapping39; β7 (residues 154–160) of one molecule inserts between β5 and β6 of the other molecule, and vice versa [Fig. 1(a)]. This domain swapping β-strand exchange between subunits in the dimer is not unusual among GNATs, and has been observed in PDB structures of ScHpa2, SeAac(6′)-Iy, and ScGNA1 (PDB ID: 1I12). The interface of HsSSAT2 dimer is extensive, burying 3,409 Å2 or 30.6% of the total monomeric surface. The dimeric interface is quite nonpolar, with 66.5% of the buried interface area from carbon atoms. The number of residues that are buried by the other molecule by domain swapping comprises 15 amino acids from the N-terminal end of β7 to the C-terminal end of the molecule (residues 154–169). Based on the domain swapping and the surface area calculation results, we conclude that HsSSAT2 is probably a dimer in vivo. The authors acknowledge the Southeast Regional Collaborative Access Team (SER-CAT) for use of beamline 22-ID and 22-BM at the Advanced Photon Source (APS), Argonne National Laboratory (ANL). Supporting institutions may be found at http://www.ser-cat.org/members.html. Use of the APS was supported by the US Department of Energy, Office of Science, Office of Basic Energy Sciences. The authors also acknowledge all CESG members, especially Eduard Bitto, Euiyoung Bae, Dave Aceti, Craig S. Newman, Zhaohui Sun, Russell L. Wrobel, Eric Steffan, Zachary Eggers, Megan Riters, Ronnie O. Frederick, John Kunert, Hassan Sreenath, Brendan T. Burns, Kory D. Seder, Holalkere V. Geetha, Frank C. Vojtik, Won Bae Jeon, Jason M. Ellefson, Andrew C. Olson, Janet E. McCombs, Janelle T. Warick, Bryan Ramirez, Zsolt Zolnai, Peter T. Lee, Mike Runnels, John Cao, Jianhua Zhang, John G. Primm, Donna M. Troestler, Michael R. Sussman, Brian G. Fox, and John L. Markley. A publication on the spermidine/spermine acetyltransferases has recently appeared.40