DHHC-7 and -21 are palmitoylacyltransferases for sex steroid receptors

DHHC-7 and -21 are ER PATs

To identify potential PATs for ERα, we expressed single plasmids coding for each of the known 23 mammalian DHHC proteins in MCF-7 breast cancer cells. Only expression of DHHC-7 or -21 significantly enhanced endogenous ERα palmitoylation (Figure 1A). All DHHC proteins were detected using hemagglutinin (HA) antibody, indicating that the unbiased screen probably detected all candidate PATs.

We then used small interfering RNAs (siRNAs) to candidate DHHC PATs endogenously expressed in MCF-7 cells to further support ERα as substrate. Protein knockdown of DHHC-7 and -21 was specific (Figure 1B, left), and DHHC-22 transcripts were determined, since no commercial antibody is available. Endogenous ERα's at the PM in MCF-7 cells were only decreased from DHHC-7 or -21 knockdowns (Figure 1B, right). We counted 200 cells in each condition and found PM ERα in 92% of the control siRNA-transfected cells (white arrows). Membrane localization of ER was reduced to 24 and 33%, respectively, from DHHC-7 or -21 knockdown, but was insignificantly affected by DHHC-22 knockdown (85%; Table 1). The selective effects for PM localization were confirmed by Western blotting (Figure 1B, far right), also indicating that localization of nuclear ERα was unaffected. The latter result reflects the fact that nuclear ERα is not palmitoylated (Pedram et al., 2007). Knockdown of either DHHC-7 or -21 (but not DHHC-22) significantly decreased endogenous ERα palmitoylation (Figure 1C), and the residual acylation probably reflects incomplete knockdown and redundancy of the function of the two DHHC proteins.

Rapid signaling by 17-β-estradiol (E2) occurs through PM and not nuclear ER (Pedram et al., 2006, 2009b). We therefore determined the impact of the PATs on these functions. Only the knockdown of DHHC-7 or -21 significantly impaired the ability of E2 to stimulate extracellular signal-regulated kinase (ERK) and phosphatidylinositol 3-kinase (PI3K)/AKT kinases (Figure 2A), and to generate cAMP (Figure 2B). In contrast, knockdown of DHHC-22 did not affect E2 activation of ERK and PI3K/AKT (Figure 2C). Importantly, reduction of either putative ER PAT significantly, but incompletely, impaired E2 signaling that enhanced cell proliferation/viability, as shown by the MTT assay (Figure 2D). This finding is consistent with previous work that showed that both nuclear and PM ER pools contribute to breast cancer cell proliferation (Razandi et al., 2003a). Interestingly, knockdown of both DHHC-7 and -21 was not significantly more potent than the individual PAT protein knockdowns in inhibiting rapid signaling by E2 (Figure 2A). This finding may be due to cooperation between DHHC-7 and -21 proteins to palmitoylate ER, and the knockdown of one protein significantly disrupted the steroid receptor localization and function.

We also determined whether DHHC-7 or -21 influenced epidermal growth factor (EGF) signaling, since transactivation of the EGF receptor is required for membrane ER signaling in hormone-responsive cancers (Razandi et al., 2003b). EGF stimulated ERK and PI3K/AKT activity and increased the viability of MCF-7 cells. These actions were unaffected by DHHC knockdown, indicating specificity of the PATs for ER (Figure 2E). The results also suggest that EGF receptor signaling does not require cross-talk to membrane-localized ERα.

Structure–function interactions and localization of DHHC proteins and ERα

We previously determined that a nine–amino acid palmitoylation motif in the ligand-binding (E) domains of all sex steroid receptors is important for the acylation that drives the receptors to the PM (Pedram et al., 2007). The cysteine at amino acid 451 is the site of palmitic acid attachment to ERα. We therefore expressed wild-type (WT) or a C451A mutant ERα and DHHC-7 or DHHC-21 in ER-null CHO cells and determined protein interactions. Expressed DHHC-7 and -21 proteins each strongly associated with WT ERα but failed to significantly interact with C451A ERα (Figure 3A). The results indicate that an intact palmitoylation site is required for PAT binding to ERα. Importantly, only WT, but not C451A ERα, was robustly palmitoylated upon DHHC-7 or -21 expression (Figure 3B). In contrast, DHHC-22 expression did not stimulate ER palmitoylation beyond the basal palmitoylation seen from endogenous PAT(s) (Figure 3B, lanes 1 and 7). We conclude that the PATs promote palmitoylation of ERα specifically at cysteine 451, consistent with the inability of C451A ERα to traffic to the PM (Pedram et al., 2007).

Palmitoylation of ERα appears to occur by a reversible thioester linkage of palmitic acid at the internal C451 site (Pedram et al., 2007). We examined this by expressing WT ERα in CHO cells with DHHC control vector or DHHC-7 or -21 protein-encoding plasmids. Palmitoylation was determined in the absence or presence of hydroxylamine, which significantly cleaves thioester bonding of fatty acyl groups (Drisdel et al., 2006; Pedram et al., 2007). Hydroxylamine inhibited both basal ERα palmitoylation that resulted from the endogenous PATs and the additional palmitoylation promoted by expressing DHHC-7 or -21 (Figure 3C).

Membrane-bound, but not nuclear, ERα is palmitoylated. Palmitoylation likely occurs at the Golgi (Pedram et al., 2007), where many proteins are palmitoylated by DHHC proteins (Fukata and Fukata, 2010). We determined where endogenous DHHC-7 and -21 proteins colocalize with ER. In MCF-7 cells, ERα is densely localized to the nucleus and is seen at extranuclear sites (Figure 3D). DHHC-7 and -21 proteins were found to be exclusively extranuclear and colocalized with some ERα in cytoplasm. We also determined that endogenous DHHC-7 and -21 colocalize with a Golgi marker (Figure 2E, top panels) and that some ERα also localizes to the Golgi (Figure 3E, bottom panels). These results support the idea that the ERα–PAT protein interaction (Figure 3A) occurs in cytoplasm (Figure 3D), at least in part at the Golgi organelle (Figure 3E).

Signaling by membrane ER affects epigenetic transcription

Palmitoylation and the resulting trafficking of ERα are relevant only for the membrane-localized pool (Pedram et al., 2006, 2009b). Thus the loss of palmitoylation should not significantly affect the transcription of genes regulated by nuclear ERα binding to promoter/enhancer estrogen-response elements (EREs). To test this, we knocked down DHHC-7 or -21 and determined endogenous pS2 gene abundance in MCF-7 cells. pS2 transcription is ERE-mediated (Harrington et al., 2006), and E2 comparably stimulated endogenous gene expression under control or DHHC-specific siRNA expression conditions (Figure 4A). Thus E2 modulation of the pS2 gene does not require membrane ERα signaling.

In contrast, ERK activation by membrane PR or ER collaborates with nuclear-localized sex steroid receptors to enhance transcription of some genes (Madak-Erdogan et al., 2008; Subtil-Rodriguez et al., 2008) Interestingly, several estrogenic compounds stimulate PR gene expression, in part by signaling through AKT to phosphorylate Ser-21 of EZH2. Phosphorylation of EZH2 at this site inhibits the histone trimethyltransferase activity of this protein (Bredfeldt et al., 2010). EZH2 is the primary histone 3 trimethyltransferase for Lys-27 (repressive mark; Margueron et al., 2008), and the PR gene is epigenetically derepressed as a result of estrogen signaling (Bredfeldt et al., 2010). AKT signaling by ER in the study by Bredfeldt et al. presumably occurred through PM-localized receptors, as we have shown in MCF-7 cells (Pedram et al., 2009a).

To investigate the impact of DHHC proteins on this important function, we determined that E2-stimluated AKT activity was diminished from DHHC knockdown or by a soluble inhibitor of PI3K, LY 290042 (Pedram et al., 2009a; Figure 4B, top panel). Inhibition of E2-induced AKT activity by either approach prevented sex steroid–induced phosphorylation of EZH2 at the inhibitory site (Ser-21) (middle panel). E2 also inhibited the H3K27Me3 (trimethylation) repressive mark (bottom panel), which was reversed by DHHC protein knockdown or LY290042. This signaling pathway importantly contributed to E2-stimulated endogenous PR gene expression, since either inhibiting the DHHC proteins or blocking AKT activation significantly reversed this steroid effect (Figure 4C). In this way, DHHC-induced trafficking of ER to the PM results in rapid signaling that induces epigenetic modulation of gene expression.

DHHC-7 and -21 are PATs for PRs and ARs

All sex steroid receptors must be palmitoylated for localization to the PM (Pedram et al., 2007). We therefore sought to identify the PATs that promote PR and AR trafficking to and functioning at the PM in hormone-responsive cancer cells. Expression in MCF-7 cells of the individual DHHC-encoding plasmids from a cDNA library is seen in Figure 1A, an experiment that allowed us to also probe the endogenous PR palmitoylation shown here. Only expression of DHHC-7 and -21 clearly promoted the enhanced palmitoylation of endogenous PRB (Figure 5A). We previously showed that both PRA and PRB are found at the PM of MCF-7 cells (Pedram et al., 2007), but PRB is dominant. Therefore we used an antibody that only identifies PRB. Knockdown of DHHC-7 and -21 in the MCF-7 cells significantly prevented endogenous PR palmitoylation (Figure 5B) and PM localization (Figure 5C). As a result, progesterone signaling through ERK and PI3K was inhibited, and the enhanced cell viability/proliferation seen in response to the steroid ligand decreased significantly (Figure 6, A and B).

Similarly, expression of the individual plasmids in C4-2 prostate cancer cells resulted in enhanced palmitoylation of endogenous AR only by the same PATs for palmitoylating ER and PR (Figure 7A). Knockdown of DHHC-7 or -21 with siRNA diminished AR localization only at the PM (Figure 7B) and diminished AR palmitoylation (Figure 7C). PAT knockdown inhibited dihydrotestosterone (DHT) signaling to ERK and PI3K/AKT (Figure 8A) and decreased cell proliferation and viability (Figure 8B). These latter results indicate that membrane AR signaling in prostate cancer cells contributes to the hormone-induced biology. Thus the DHHC-7 and -21 proteins are conserved as PATs for all classes of sex steroid receptors.

Cell culture and reagents

MCF-7 and CHO-K cells were from ATCC and C4-2 prostate cancer cells (AR positive) were a gift from Ganesh Raj (University of Texas, Southwestern, Dallas, TX). Cells were cultured in phenol red–free DMEM-F12, 10% charcoal-stripped fetal bovine serum (FBS), 1% antibiotic–antimycotic (15240–062; Life Technologies, Carlsbad, CA) at 37°C to 75% confluency in 100-mm dishes or six-well plates and exposed to 10 nM E2, 100 nM progesterone acetate (P) or 10 nM DHT (Steraloids, Newport, RI) for the times indicated in the experiments. ERα (MC-20), PRB (C-19), and AR (C-19) antibodies (all polyclonal, raised in rabbits); green fluorescent protein antibody (GFP; polyclonal, rabbit); and HA antibodies (monoclonal, mouse) were from Santa Cruz Biotechnology (Santa Cruz, CA). Golgi marker antibody (AE-6, mouse monoclonal) from Santa Cruz is directed against the isolated Golgi cisternae functional regions composed of various enzymes and is derived from a human lymphoma cell line, per the manufacturer. Antibodies to DHHC-7 and -21 (monoclonal, mouse), phospho-EZH2 Ser-21, histone H3, and H3K27Me3 (polyclonal, rabbit) antibodies were from Abcam (Cambridge, MA), and were used for immunoblots in subcellular fractions and for immunofluorescence microscopy. We also used a DHHC-7 polyclonal antibody raised in rabbits (Novus Biologicals, Littleton, CO) for the Golgi colocalization study. Phospho-AKT Ser-473 antibodies (monoclonal, rabbit) were from Cell Signaling Biotechnology (Danvers, MA). Double-stranded control and RNAs for the DHHC-7, -21 and -22 proteins (two siRNAs each) were obtained from Santa Cruz. Each siRNA was used for all key experiments, and the data shown are from experiments using either of the two siRNAs. Construction and transfection/expression of all ER, AR, PR, and GFP (pEGFP vector; Clontech, Mountain View, CA) plasmids were previously described (Pedram et al., 2007). The DHHC cDNA plasmid library used the pEF-BOS-HA backbone vector and was provided by Y. Fukata and M. Fukata (Fukata et al., 2004). PD 98059, a MEK inhibitor, was a gift from Alan Saltiel (Parke-Davis, Ann Arbor, MI), and the LY290042 (PI3K inhibitor) was from Calbiochem (San Diego, CA).

Cell fraction isolation

Cells were homogenized using 15–20 strokes of a tight-fitting Dounce homogenizer until ∼ 90% of the cells were broken. Homogenates were centrifuged at 4000 rpm for 10 min to pellet the nuclear fractions. Supernatants were layered onto a discontinuous sucrose gradient (1.0–2.5 M) made up in buffer containing 10 mM Tris (pH 7.6), 2 mM EDTA, 2 mM dithiothreitol, protease inhibitor cocktail, and 1 mM phenylmethylsulfonyl fluoride (PMSF). The mixtures were centrifuged at 2000 rpm for 30 min at 4°C to produce a top layer (cytosol). Cytosolic fractions were centrifuged at 100,000 × g for 1 h. The pellet was centrifuged again to isolate the membrane fraction, which was separated by SDS–PAGE on a 10% gel, transferred to nitrocellulose, and then immunoblotted with the indicated antibodies (Santa Cruz), which were detected using the ECL kit (Amersham/GE Healthcare, Waukesha, WI). Western blots were also used to normalize total receptor or other proteins.

Steroid receptor palmitoylation

Palmitoylation of sex steroid receptors was carried out as previously described (Pedram et al., 2007). MCF-7, CHO, or C4-2 cells were labeled with [³H]palmitic acid (0.5 mCi/ml) for 4 h in the absence of sex steroids. Cytoplasmic extracts were immunoprecipitated with antibodies to ERα, PR, or AR, each of which was conjugated to protein A Sepharose. Separation of samples by SDS–PAGE was followed by autoradiography with emulsifying enhancer.

ERK and PI3K/AKT activity

ERK was immunoprecipitated from lysates of MCF-7 cells exposed to various steroids or EGF (10 ng/ml). Equal ERK protein amounts from each experimental condition were added to separate tubes containing [³²P]ATP and substrate peptide (myelin basic protein; Sigma-Aldrich, St. Louis, MO) in buffer to determine kinase activity. Activity was determined by autoradiography of phosphorylation of substrate protein. (Pedram et al., 2006). PI3K activity was determined as phosphorylation of AKT at Ser-473 (p-AKT) by Western blotting (Pedram et al., 2006) after 15 min of exposure of the synchronized cells to steroid or EGF. Total ERK or AKT protein, as identified by Western blotting, was used for sample activity normalization. In control DHHC signaling experiments, MCF-7 cells were incubated with DHHC-22 siRNA, recovered overnight in DMEM-F12 medium without serum or phenol red, and then incubated with 10 nM E2 for 15 min, with ERK or PI3K/AKT activation determined at that time.

cAMP generation

Adenylate cyclase activity in the membrane suspension was determined by measuring cAMP generation in the presence of IBMX (phosphodiesterase inhibitor) after cells were incubated for 5 min with 10 nm E2 in DMEM-F12 medium. After incubation, the cells were washed and 0.1N HCl was added to the cell monolayers for 20 min at room temperature to extract cAMP. The supernatants were collected and recentrifuged for use in the cAMP radioimmunoassay according to the manufacturer's protocol (Perkin Elmer-Cetus, Waltham, MA; Pedram et al., 2006).

MTT assay

Cell proliferation/viability was quantified by a colorimetric assay according to the manufacturer's instructions (Sigma Chemical Corporation). The kit measures mitochondrial dehydrogenase activity in viable cells based on the conversion of the MTT to MTT–formazan crystal by mitochondrial enzyme. In brief, synchronized MCF-7 and C4-2 cells were exposed to various treatments in DMEM-F12 medium for 24 h. At the end of incubation, the cells were washed and then incubated with medium containing 1 mg/ml MTT for 4 h at 37°C. MTT–formazan was extracted by overnight incubation at 37°C with 100 μl extraction buffer (20% SDS, 50% formamide adjusted to pH 4.7 with 0.02% acetic acid, and 0.025 N HCl]. Optical densities at 570 nm were measured using extraction buffer as a blank.

Cell localization of steroid receptor proteins

Steroid receptor localization was determined by immunofluorescent cell microscopy (Drisdel et al., 2006). After MCF-7 or C4-2 cells were cultured on glass bottom dishes (MatTek, Ashland, MA), the cells underwent siRNA transfection or other treatments, and then they were recovered over 24 h, fixed with 3.5% paraformaldehyde, permeabilized with 0.2% Triton X-100, and then incubated with sex steroid receptor antibody overnight at 4°C. Nonspecific immunoglobulin G (IgG) antibody served as control. Incubation with the first antibody was followed by incubation with fluorescein isothiocyanate (FITC)-conjugated second antibody (green color) or IgG antibody (no visualization) for 2 h at room temperature. For colocalization studies, human-specific antibodies to Golgi markers (Santa Cruz Biotechnology) were used for separate red and green fluorescence. A similar method was used for fluorescence colocalization of endogenous ERα with DHHC proteins or Golgi in MCF-7 cells using monoclonal antibodies against each protein. Identically treated MCF-7 cells were also used for plasma membrane and nuclear fraction isolation. The fractions were separated by SDS–PAGE on a 10% gel, transferred to nitrocellulose, immunoblotted with the indicated antibodies (Santa Cruz Biotechnology), and detected using the ECL kit.

Epigenetic modulation studies

MCF-7 cells were cultured in DMEM-F12 medium and transfected with DHHC siRNAs (3 μg/well of cells in six-well plates), which was followed by 24 h of recovery and overnight synchronization. The cells were incubated with 10 nM E2 ± 10 μM LY294002 for 30 min to determine AKT and EZH2 phosphorylation, and for 24 h to determine PR gene expression using real-time PCR (RT-PCR). MCF-7 cell monolayers were washed in ice-cold phosphate-buffered saline supplemented with 5 mM sodium butyrate to retain histone acetylation. After centrifugation at 2000 rpm for 5 min in the cold, cell pellets were resuspended in lysis buffer (10 mM Tris-HCl, pH 7–7.4, 10 mM NaCl, 3 mM MgCl₂, 1 mM PMSF, 1 mM Na₃VO₄, 1 mM NaF, and complete protease inhibitor cocktail) with 0.5% NP-40 on ice. Cell lysates were centrifuged for 5 min at 6000 rpm at 4°C. Phospho-AKT (Ser-473) was determined from supernatants by immunoblotting as readout for PI3K activity (Pedram et al., 2006). The activity-inhibiting phosphorylation at Ser-21 of the histone methyltransferase EZH2 was also investigated. Total AKT and EZH2 served as loading controls. Pellets were resuspended in media with 5 mM MgCl₂ and 0.8M HCl, sonicated at low speed for 20 s, and incubated on ice for 1 h. The solution of histone proteins was centrifuged at 14,000 rpm for 10 min at 4°C. Supernatants were transferred to a new tube, and histones were precipitated with a 50% solution of trichloroacetic acid dissolved in ddH₂O. Precipitated histones were collected by centrifugation for 20 min at 14,000 rpm at 4°C. The resultant pellet was washed with ice-cold acetone, dried, and subsequently resuspended with ddH₂O (150 μl) and 1.5 M Tris-HCl (pH 8.8; 2 μl). Histone proteins (10 μl) were resolved by SDS–PAGE on a 10–20% Tris-tricine gel (Bio-Rad, Hercules, CA), and changes in trimethylation of histone 3 at Lys-27 relative to total histone 3 were investigated by immunoblotting with antibodies from Abcam.

Quantitative RT-PCR

MCF7 cells were grown in phenol red–free DMEM with 10% charcoal dextran–treated FBS for greater than 24 h before transfection. Cells were transiently transfected with control or DHHC siRNAs plasmids for 24 h, then washed and incubated for 24 h in either 10 nm E2 ± LY290042 or an ethanol vehicle. Total RNA was extracted using the RNeasy kit from Qiagen (Valencia, CA). cDNA was synthesized using the Improm-II reverse transcription system (Promega, Madison, WI). Quantitative RT-PCR (qRT-PCR) was used to determine pS2 and PR gene expression, and RT-PCR was used to determine DHHC-22 expression, normalized to the housekeeping gene, GAPDH. Primers were designed with Primer3 (http://frodo.wi.mit.edu) and were blasted for specificity.

pS2:F5′-ATACCATCGACGTCCCTCCA-3′,R5′-AAGCGTGT­CTGAGGTGTCCG-3′; PR:F5′-CGCGCTCTACCCTGCACTC-3′,R5′-TGAATCCGG­CCTCAGGTAGTT-3′; ZDHHC22:F5′-GCCTACAT­CTCC­GCTGTCCTTT-3′, R5′ATGGCG­­AACCAGAGGTAGAGCA-3′; GAPDH:F 5′-CCACAGTCCATGCCATCA-3′, R5′-GGATGACCTTGCCCACAG-3′ primers were designed for annealing temperature of 60°C and to amplify regions of ∼95–120 base pairs. PCR amplicon sizes were confirmed by agarose gel electrophoresis prior to qRT-PCR analysis. For qRT-PCR, 500 ng of cDNA was used in a 50-μl reaction consisting of 25 μl of SYBR Green, qPCR Supermix (Invitrogen, Carlsbad, CA), 1 μl of 10 μM forward/reverse primer stocks, and nuclease-free water. Thermocycling was carried out using the iCycler (Bio-Rad) with a melting-curve temperature of 60°C. Relative mRNA levels were calculated using the C_t method.

Image acquisition

The microscopic images were obtained using a Nikon Eclipse TE-200 microscope with magnification from 200–400×, at room temperature. Imaging medium (Vectashield) with antifade properties was from Vector Laboratories (Burlingame, CA). A Nikon Diagnostic Instruments camera (Model 3.2.0) was used in conjunction with Spot Advance software (Sterling Heights, MI) to capture and transfer images to the computer. Rhodamine-conjugated (red) and FITC-conjugated (green) secondary antibodies (Vector Laboratories) were used for fluorescence visualization.

Data analysis

Data from multiple experiments were used to calculate a mean ± SEM and were analyzed by analysis of variance (ANOVA) plus Schefe's test at a p < 0.05 level of significance.