|Home | About | Journals | Submit | Contact Us | Français|
Protein palmitoylation is emerging as an important post-translational modification in development as well as in the establishment and progression of diseases such as cancer. This chapter describes the use of fluorescent lipidated peptides to characterize palmitoyl acyltransferase (PAT) activities in vitro and in intact cells. The peptides mimic two motifs that are enzymatically palmitoylated, i.e. C-terminal farnesyl and N-terminal myristoyl sequences. These substrate peptides can be separated from the palmitoylated product peptides by reversed-phase HPLC, detected and quantified by the fluorescence of their NBD label. Through these methods, the activities of PATs toward these alternate substrates in isolated membranes or intact cells can be quantified. The in vitro assay has been used to characterize human PATs and to identify inhibitors of these enzymes. The cellular assay has been useful in elucidating the kinetics of protein palmitoylation by PATs in situ, and the sub-cellular distribution of the palmitoylated products.
In a classical biochemical approach, proteins are purified from cell or tissue lysates by activity-based fractionation, and subsequently sequenced to reveal their molecular identity. Knowing what an enzyme does and where it is located in the cell, frequently aids in determining its endogenous substrates, and hence its biological function. In the case of protein palmitoylation catalyzed by palmitoyl acyltransferases (PATs), this process has been reversed. Initially, specific proteins isolated from eukaryotic cells were shown to contain thioester-linked palmitate. Next, it was demonstrated that this modification plays functional roles for the proteins (for a review see ). For example, palmitoylation of certain proteins has been shown to target them to peripheral membranes . More specifically, palmitoylation in conjunction with a second lipid modification, such as prenylation or myristoylation, targets proteins to sub-domains of the plasma membrane enriched in signaling molecules [3,4].
Unlike other lipid modifications of proteins, palmitoylation can be enzymatically reversed by thioesterases that cleave the palmitate from proteins [5,6]. Therefore, palmitoylation can be considered to be a molecular switch capable of regulating protein function similar to other reversible modifications, such as phosphorylation. The dynamic nature of palmitoylation and the proteins that are modified by this lipid make the PATs attractive new targets for drug development. For example, oncoproteins, including H-Ras, N-Ras, and K-Ras2A and most Src-related tyrosine kinases, require palmitoylation to correctly function in cells [7–9]. These proteins are nonfunctional if the cysteine residues at their sites of palmitoylation are mutated, even if the proteins are constitutively activated by mutation . Therefore, pharmacologic inhibition of PATs may provide a new approach to blocking aberrant signaling in cancer and other diseases.
Surprisingly, this sophisticated understanding of the function of the PATs and their substrates was achieved prior to the identification of the enzymes responsible for the modification. It is only recently that PATs have been cloned and characterized. A critical breakthrough came from the yeast community with the description of AKR1 and ERF2/ERF4 as PATs [11,12]. These papers allowed us and others to use structural and sequence information to identify the human homologues of these enzymes [13–17]. However, to better characterize these enzymes biochemically, it was necessary to develop assays that utilized standardized substrates and that are amenable to medium-throughput testing. To this end, we described the use of PAT peptide substrates that can be efficiently synthesized and analyzed. These peptides mimic two motifs commonly palmitoylated in vivo (Fig. 1). The (NBD)-CLC(OMe)-Farn peptide mimics the modified C-terminus of proteins like H- and N-Ras. It contains a C-terminal carboxymethylated and farnesylated cysteine residue linked to a stretch of amino acids that contains a palmitoylatable cysteine. The Myr-GC(NBD) peptide mimics the modified N-terminus of proteins such as the Src-related tyrosine kinases. This peptide contains an N-myristoylated glycine followed by a palmitoylatable cysteine residue. These substrate peptides can be separated from the palmitoylated product peptides by reversed-phase HPLC, and detected and quantified by the fluorescence of their NBD label as described below.
The fluorescent peptides (NBD)-CLC(OMe)-Farn and Myr-GC(NBD) (Fig. 1) are synthesized as t-butyl-disulfide-protected precursors by solution-phase chemistry using mild conditions to maintain chemically labile functional groups, e.g., the farnesyl-cysteine thioether linkage [18,19]. The peptides are stored under argon at −80 °C, and deprotected as described below immediately prior to their use.
The assay extracts are dissolved in 30 µl DMSO and peptides are resolved on a reversed-phase Chromolith RP-8e column using a methanol gradient with a flow rate of 1 ml/min. Initially, the mobile phase is maintained as 30% MeOH: 70% water for 1 min, followed by a 5 min linear gradient from 30 to 100% MeOH. The mobile phase is then maintained at 100% MeOH for 10 min and followed by a linear gradient from 100 to 30% MeOH over 3 min. The NBD-labeled peptides are detected by their fluorescence at their optimal excitation and emission wavelengths of 465 and 531 nm, respectively. Typical elution patterns are shown in Fig. 2. For either peptide, the non-palmitoylated substrate peptide elutes prior to the more hydrophobic palmitoylated peptide product, with baseline resolution being attained between peaks to allow their quantification. The limit of detection is approximately 5 pmol of NBD-labeled peptide, as determined using serial dilutions of 2-(12-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)amino)dodecanoyl-1-hexadecaboyl-sn-glycero-3-phosphocholine (NBD C12-HPC) from Molecular Probes, Inc. (Eugene, OR).
The methods described above can be adapted to measure in situ PAT activity in a variety of cell types. The lipidated peptide substrates are deprotected immediately before use as described above. Cells (typically SKOV3 human ovarian adenocarcinoma cells) are grown to confluence and incubated for 5–60 min at 37 °C with one of the peptides at a final concentration varying from 0.1 to 80 µM (usually 1 µM) for times varying from 5 to 120 min. Following this incubation, the cells are washed twice with ice-cold PBS, killed with 50% methanol, scraped and transferred to a glass test tube. The peptide substrate and the palmitoylated peptide product are then extracted using potassium carbonate-buffered CH2Cl2/water/methanol (2:1:1, by volume), and analyzed by fluorescence HPLC as described above. At high concentrations of peptide, as-yet unidentified fluorescent products are often seen between the substrate and palmitoylated product peaks. These may represent non-enzymatic acylation by activated fatty acids other than palmitoyl-CoA, or may be disulfides with cellular thiols. In our experience, the substrate peptides are rapidly and efficiently palmitoylated in intact cells, such that at concentrations below approximately 10 µM, more than 85% of the peptide is palmitoylated within 15 min .
The NBD label of the peptides allows efficient monitoring of their intracellular location, e.g., for studies of lipidated protein trafficking. For these experiments, SKOV3 (or other) cells are incubated with a deprotected substrate peptide as described above, except that phenol red-free DMEM medium is used. The incubations should be conducted under conditions shown by HPLC analyses to result in nearly complete palmitoylation of the substrate peptides, so that the molecular identity of the fluorescent label is unambiguous. The cells are rapidly washed with cold PBS, and imaged by confocal microscopy using fluoresceine optics while still living. Comparison of the distribution of the palmitoylated peptides with markers of various subcellular compartments allows the sequential analysis of peptide trafficking within an individual cell. For example, SKOV3 cells can be incubated with 1 µM peptide for 15–120 min, conditions under which the peptides are almost completely palmitoylated for the entire time course . Fig. 3 shows that at 15 min, the palmitoylated peptides are primarily associated with the plasma membrane; whereas, longer incubations result in greater distribution into the Golgi apparatus (not shown).
Most previous attempts to characterize protein palmitoylation have been performed by isotopic labeling of cellular proteins or peptide constructs with radiolabeled palmitate or palmitate analogs. However, these techniques have several complications: the isotopes become diluted with endogenous palmitate during cell labeling; the palmitoyl-CoA substrate may bind to or be metabolized by proteins that are not involved in the palmitoylation reaction, such as acyl binding proteins; the palmitate group can be released from proteins prior to analysis; and visualization of isotope-labeled proteins by autoradiography often requires exposure to film for weeks or months. Perhaps the greatest disadvantage of characterizing enzymatic palmitoylation with isotope labeling experiments is that these techniques do not easily distinguish between non-enzymatic and enzymatic palmitoylation, making the characterization of PATs very difficult. We can now overcome many of these problems using the assays described herein to unambiguously identify and quantify specific fluorescently labeled palmitoylated peptides.
Both the cellular and the in vitro assay rely on the subsequent separation of the modified peptide by reversed-phase HPLC. The fluorescent label on the peptide allows for the quantification of the amount of the peptide modified based on the ratio of palmitoylated to total peptide measured, i.e., peak b to peaks (a + b) in Fig. 2. The mass of peptide converted to the palmitoylated form can then be determined, allowing for the determination of the kinetics of turnover of the peptide. Because there is a low rate of chemical palmitoylation of the peptides, it is necessary to determine the rate of autoacylation under each in vitro assay condition. Enzymatic palmitoylation can be calculated by subtracting the level of autoacylation from the level of total palmitoylated peptide.
We have used these assays to demonstrate that multiple PATs exist in cells [18,21]. Additionally, we have shown that different cell lines have varying activities towards the two peptide substrates, and that Ras-transformed NIH/3T3 cells have a significant increase in activity towards the Ras–mimetic substrate, (NBD)-CLC(OMe)-Farn . In this same study, we showed that this activity is dependent on the PAT HIP14. Therefore, these assays are useful for the molecular identification of PATs within human cells.
Another significant use for these assays is for the identification of PAT inhibitors that could function as anticancer therapies. To this end, we have used the assays in medium-throughput screens of small molecule libraries . Examples of PAT inhibitors that can be classified into five chemotypes are shown in Fig. 4. The in vitro assay described above has allowed us to not only verify the inhibitory activity of these compounds, but also to determine the specificity of the inhibitors. Most of the hits from the screen did not demonstrate selective inhibition of palmitoylation of either peptide. However, as also shown in Fig. 4, some compounds demonstrate good selectivity toward one peptide substrate or the other. The effects of the five representative compounds on the palmitoylation of either peptide substrate at a single dose of 25 µg/ml (approximately 30–40 µM) are shown. These selective PAT inhibitors should be very useful in further defining the properties and roles of individual PATs, and ultimately may provide new therapeutics for diseases associated with aberrant PAT activity.