Topological analysis of HHAT

Hedgehog proteins are secreted morphogens that play critical roles in development and disease. During maturation of the proteins through the secretory pathway they are modified by the addition of N-terminal palmitic acid and C-terminal cholesterol moieties, both of which are critical for their correct function and localisation. Hedgehog acyltransferase (HHAT) is the enzyme in the endoplasmatic reticulum (ER) that palmitoylates Hedgehog proteins, is a member of a small subfamily of MBOAT proteins that acylate secreted proteins, and is an important drug target in cancer. However little is known about HHATs structure and mode of function. We show that HHAT is comprised of 10 transmembrane domains and 2 reentrant loops with the critical His and Asp residues on opposite sides of the ER membrane. We further show that HHAT is palmitoylated on multiple cytosolic cysteines, which maintain protein structure within the membrane. Finally, we provide evidence that mutation of the conserved His residue in the hypothesised catalytic domain results in a complete loss of HHAT palmitoylation, providing novel insight into how the protein may function in vivo. INTRODUCTION The Hedgehog (Hh) family of proteins are secreted morphogens which play a significant role during embryonic development in determining organogenesis and anteriorposterior patterning of various tissues, including the central nervous system and limb and digit formation (1,2). Aberrant activation of Hh signaling in various cancers result in promotion of cancer growth and metastasis (3,4) and small Topological analysis of HHAT 2 molecule inihbitors for this pathway are in use in various clinical trials. A unique feature of the Hh proteins are that during maturation through the secretory pathway, they are post-translationally modified by the addition of a cholesterol moiety to its Cterminus, via an ester linkage, while a palmitate (C16:0) fatty acid is also added on the conserved N-terminal cysteine residue via an amide linkage (5). Palmitoylation of Hh proteins is catalysed by the protein acyltransferase (PAT) Hedgehog acyltransferase (HHAT) (6). These modifications produce the mature and functional Hh signaling molecule and are crucial for the correct function of the protein. They not only direct the formation of large Hh multimers upon secretion from the producing cells, but they also determine the proper release and targeting of the proteins, as well as having a significant effect on the potency of the protein to activate the signaling pathway on the receiving cells (79). For these reasons, disrupting the posttranslational modification of Hh proteins, and hence inhibiting the formation of functional signaling molecules, is an attractive new method of inhibiting the Hh pathway in cancer. HHAT is a multipass transmembrane (TM) domain protein located in the endoplasmic reticulum (ER) of cells (6,10), and a member of a small sub-group of the membrane-bound Oacyltransferase (MBOAT) superfamily of proteins, that specifically acylates secreted proteins.. Other members of this important subgroup include Porcupine (PORCN) which palmitoyleoylates (C16:1) the Wnt family of proteins and ghrelin-O-acyltransferase (GOAT) which octanoylates (C8:0) the appetite-sensing peptide ghrelin (11). All MBOAT proteins are characterized by their MBOAT homology domain, a region of highly conserved residues including an invariant His residue (His379 in the case of HHAT) and a highly conserved Asn or in the case of HHAT, Asp (Asp339) residue 30-45 amino acids upstream (11,12). These amino acids are proposed to be catalytic, although this is still not clear, especially for HHAT where mutation of the His to an Ala still retains significant PAT activity (13). Recent studies by our group and others provided proof of principle that inhibiting HHAT function is a valid method of inhibiting Hh signaling in cancer (14-16). Despite the importance of HHAT in Hh signaling and its therapeutic potential in cancer, little is understood about the structure of the protein and the identity of the catalytically important amino acids, information that may guide future studies in development of small molecule inhibitors of HHAT as well as into other functions HHAT may have in the cell. In this study, we determine HHAT topology using a variety of experimental methods. Our data suggest that HHAT contains 10 TM domains and 2 re-entrant loops (RLs) and that the invariant His379 is luminal, while Asp339 is on the cytosolic side of the ER. Furthermore, we show that HHAT is itself palmitoylated at multiple cytoplasmic Cys residues, and His379 is critical for the palmitoylation of HHAT, while a conserved Cys324 appears to modulate protein topology significantly. EXPERIMENTAL PROCEDURES Cell culture and transfection HEK293a and HeLa cells were cultured in high glucose (4.5g/L) DMEM supplemented with Glutamax (Life Technologies, Paisley, UK) containing 10% foetal bovine serum (FBS, Sigma, Gillingham, UK). All cells were grown at 37 °C in a humidified incubator under 5% CO2. Cells were transfected at 70% confluence using Turbofect (Thermo Fisher Scientific, Cramlington, UK) according to the manufacturer’s specifications. Plasmid construction and mutagenesis Fulllength human HHAT cDNA (Accession No: BC117130) expression vector with C-terminal V5 and 6xHis epitopes has previously been described (15). The cloned cDNA sequence carries a missense mutation compared to the human HHAT consensus sequence, a serine to aspartate change in position 182 of the protein. Prior to proceeding with making mutants for HHAT for topology analysis, the missense mutation was corrected to serine by QuickChange II site-directed mutagenesis (forward primer: CTACTACACCAGCTTCAGCCTGGAGCTCT GCTGGCAGCAGC; reverse primer: CAGGCTGAAGCTGGTGTAGTAGAGGCAG CGAACGGTCAGCG). All subsequent HHAT mutants and truncates for topological analysis were made using the corrected vector expressing HHAT-V5-Hisx6 as template. For cysteine mapping topology analysis, selected cysteines were mutated to alanine by Q5 sitedirected mutagenesis (NEB, Hitchin, UK). Q5 mutagenesis was also used for the introduction of the TEV protease site, ENLYFQG, in the HHAT-TEV mutants. Topological analysis of HHAT 3 For the production of the V5 topology clones, the V5-6xHis epitope from the HHAT-V5-6xHis construct was removed and replaced with a FLAG epitope (DYKDDDDK) by Q5 mutagenesis, followed by insertion of the V5 epitope (GKPIPNPLLGLDST) at the required sites. For N-glycosylation analysis, an N-glycosylation site (Asp-Leu-Thr) was introduced by QuickChange II site-directed mutagenesis into the spacer region between the gene of interest cloning site and the V5-Hisx6 epitope of the empty destination vector pcDNA-DEST40 used for HHAT cloning Next, full length HHAT or HHAT truncations HHAT-157-493 and HHAT-192-493 were amplified from the HHAT expression vector described above (15) and cloned into the modified destination vector using Gateway cloning resulting in expression vectors carrying N-glycosylation, V5 and Hisx6 epitopes at the C-terminus. All clones were verified by sequencing. SDS-PAGE and immunoblotting Separation of proteins was performed by sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE), using 10 or 15% Tris gels and Tris-glycine-SDS running buffer. Samples were prepared in NuPAGE® LDS 4 x sample loading buffer (Life Technologies) and 10% (v/v) 2mercaptoethanol (unless stated otherwise). The protein ladder used for comparison of molecular weight was Precision Plus Protein® Standards All Blue (Bio-Rad, Herts, UK). Gels were run using the Mini-PROTEAN® Tetra Cell System and power supply unit (Bio-Rad). Fluorescently-tagged proteins were imaged (540 nm excitation and 595 nm emission, channel Cy3) using an Ettan DIGE Imager (GE Healthcare, Bucks, UK) and images were analysed with ImageQuantTM TL software (GE Healthcare). Proteins were transferred from SDS-PAGE gels to a PVDF membrane (Millipore, UK) using a semi-dry transfer unit (Hoefer, Holliston, MA, USA). Membranes were blocked with blocking solution (5% w/v milk powder (Marvel), dissolved in PBS) for 1 h at room temperature and washed with PBS-T (phosphate buffered saline, PBS, containing 0.05% Tween-20 (Sigma)) and probed with appropriate antibodies. Antibodies and their sources were as follows: mouse anti-V5 monoclonal antibody (1:10,000 dilution, Life Technologies); mouse anti-6xHis antibody (1:1000, AD1.1.10, R&D Systems, Abingdon, UK), rabbit anti-6xHis antibody (1:1000, ab137839, Abcam, Cambridge, UK); goat anti-Grp94 antibody (1:200 dilution, C-19, Santa Cruz Biotechnology, UK); anti-calnexin-N-terminus mouse monoclonal IgG1 (1:1000, AF18, Sigma). Secondary antibodies used were: goat antimouse IgG2a-HRP (1:20,000 dilution, Southern Biotech, UK); goat anti-mouse IgG1-HRP (1:20.000, Southern Biotech); goat anti-rabbitHRP IgG (1:20,000, Southern Biotech); or after immunoprecipitations, mouse anti-rabbit IgGHRP VeriBlot (1:1000, Abcam, UK). Visualisation was carried out by enhanced chemiluminescence kit (Pierce ECL2 Western Blotting Substrate, Thermo Scientific, UK) according to manufacturer’s instructions and on an Ettan DIGE Imager (GE Healthcare, UK), excitation at 480 nm, emission at 530 nm (channel Cy2). Images were analysed with ImageQuantTM TL software. TEV cleavage of ER microsomes HEK293a cells were transfected with wild-type or TEV mutant HHAT-V5-6xHis constructs. After 48 h, cells were washed twice with cold HCN buffer (50 mM Hepes, pH 7.5, 150 mM NaCl, 2 mM CaCl2) and pelleted at 800 × g for 2 min. Cells were passed through a 23g needle 10 times and then intact cells and nuclei were pelleted at 800 x g for 5 min at 4 o C. The supernatant was then ultracentrifuged at 100,000 x g for 1 h to obtain microsomes, which were resuspended in ProTEV protease buffer (50mM HEPES pH 7.0, 0.5mM EDTA, 1mM DTT) and protein concentration determined using the DC protein assay (BioRad). Due to lower expressi