Membrane anchoring of diacylglycerol lactones substituted with rigid hydrophobic acyl domains correlates with biological activities

Abstract

Synthetic diacylglycerol lactones (DAG lactones) are effective modulators of critical cellular signaling pathways downstream of the lipophilic second messenger diacylglycerol that activate a host of protein kinase C (PKC) isozymes as well as other non-kinase proteins that share with PKC similar C1 membrane-targeting domains. A fundamental determinant of the biological activity of these amphiphilic molecules is the nature of their interactions with cellular membranes. This study characterizes the membrane interactions and bilayer anchoring of a series of DAG lactones in which the hydrophobic moiety is a ‘molecular rod’, namely a rigid 4-[2-(R-phenyl)ethynyl]benzoate moiety in the acyl position. Use of assays employing chromatic biomimetic vesicles and biophysical techniques revealed that the mode of membrane anchoring of the DAG lactone derivatives was markedly affected by the presence of the hydrophobic diphenyl rod and by the size of the functional unit at the terminus of the rod. Two primary mechanisms of interaction were observed: surface binding of the DAG lactones at the lipid/water interface and deep insertion of the ligands into the alkyl core of the lipid bilayer. These membrane-insertion properties could explain the different patterns of the PKC translocation from the cytosol to membranes that is induced by the molecular-rod DAG lactones. This investigation emphasizes that the side residues of DAG lactones, rather than simply conferring hydrophobicity, profoundly influence membrane interactions, and thus may further contribute to the diversity of biological actions of these synthetic biomimetic ligands.

Abbreviations

cryo-TEM

cryogenic transmission electron microscopy

DAG

diacylglycerol

DMPC

dimyristoylphosphatidylcholine

DSC

differential scanning calorimetry

%FCR

percentage fluorescence chromatic response

NBD-PE

N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)1,2-dihexadecanoyl-sn-glycero-3-phosphoethanolamine triethylammonium salt

PDA

polydiacetylene

PKC

protein kinase C

TMA-DPH

1-(4-trimethylammoniumphenyl)-6-phenyl-1,3,5-hexatriene

Introduction

The lipophilic second messenger sn-1,2-diacylglycerol (DAG) is released in situ from membrane phosphatidylinositol 4,5-bisphosphate through the action of phospholipase C in response to induction of a wide range of G-protein-coupled receptors and receptor tyrosine kinases [1]. As a second messenger, DAG mediates the action of numerous growth factors, hormones and cytokines by activating members of the protein kinase C (PKC) family of enzymes, as well as several other families of signaling proteins, e.g. Ras guanyl nucleotide-releasing protein (RasGRPs) and chimaerins, that share with PKC the C1 domain as a DAG recognition motif. Many of these signaling pathways feature prominently in the development and properties of cancer cells [2,3], and, in consequence, PKC isozymes are being actively studied as possible therapeutic targets for cancer [4]. The majority of C1 binding ligands that are utilized are structurally rigid and complex natural products, such as the prototypical phorbol esters and bryostatins [5] These compounds bind their C1 receptors with nanomolar binding affinities, and are more than three orders of magnitude more effective than the very flexible natural DAG agonists. In order to overcome this affinity gap and generate structures that are simpler and easy to synthesize, we have proposed overcoming the entropic penalty associated with the flexible glycerol backbone by constructing cyclic esters of DAG with the embedded glycerol backbone in various rigid conformations. In a comprehensive review, we discussed the reasons for selecting the five-member ring lactones, which are generically described as DAG lactones [6]. Many of these DAG lactones possess affinities for PKC that approach those of the phorbol esters, and show marked diversity in the patterns of biological response that they induced as a function of the chemical nature of the side chains [6–9].

The concept that has emerged from these studies is that different patterns of substitution on the conformationally restricted DAG lactone template allow preferentially interaction with PKC isozymes within particular membrane microenvironments, promoting phosphorylation of substrates co-localized with the activated PKC.

Previous results obtained with DAG lactones containing acyl chains with an ensemble of repetitive oligo(p-phenylene-ethynylene) units that form a rigid rod showed that two units is the ideal length of the rod [10]. The synthesis of several DAG lactones fulfilling this structural constraint has already yielded important insights into the mechanisms of self-assembly and lipid interactions at the water/air interface, and the diverse effects of various lipids upon the organization and thermodynamic properties of the molecules [11]. Because the end residue of the rod (R) was shown to interact with the inner layer of the membrane and to modulate its surrounding environment, we focused in the present study on a group of compounds in which the R terminus of the rigid rod was varied from the smallest possible, i.e. a hydrogen atom, to the bulkier isopropyl (i-Pr) and tertbutyl (t-Bu) groups (Table 1). As before, we compared these compounds to a DAG lactone with a flexible decanoic acid chain (compound 1, Table 1). The experiments were designed to explore both the roles of the rod side residue, as well as the properties of the larger alkyl R units in modulating bilayer interactions.

Table 1.   Structure and properties of the DAG lactones.

Here, we investigate the interactions and association of DAG lactones 1–4 with lipid vesicles, and correlate the data with their binding affinities for PKCα and their ability to induce translocation of PKC to cellular membranes. These ligands are shown to induce different patterns and kinetics for translocation of PKC isoforms to membranes, and this study aims to examine whether membrane association might account for the biological differences. Application of several biophysical techniques, including use of biomimetic chromatic vesicles [12–14], fluorescence quenching [15], fluorescence anisotropy [16], differential scanning calorimetry (DSC) [17] and cryogenic transmission electron microscopy (cryo-TEM) [18], revealed significantly different modes of bilayer binding of the rod compounds depending on their structure, and highlighted the role of the R terminus in affecting membrane insertion and biological activity of the DAG lactones. The results shed light upon the molecular parameters affecting PKC translocation to membranes by DAG lactones.

Results

Membrane translocation

To analyze membrane interactions of the DAG lactones, we first evaluated their cellular effects. Compounds 2–4 showed similar high affinity for PKCα as did DAG lactone 1 in in vitro assays in the presence of 100 μg·mL⁻¹ phosphatidylserine (Table 1). To study the behavior of these DAG lactones in living cells, we first determined the pattern and kinetics of the translocation of overexpressed, GFP-tagged PKCα and PKCδ to the membranes of Chinese hamster ovary (CHO) cells following addition of the compounds (Fig. 1). As reported previously [10], DAG lactone 1, included in this study as a DAG lactone derivative which exhibits a highly flexible side residue, translocated both PKCα and δ almost instantaneously to the cellular membranes, within less than 2 min (Fig. 1A). Furthermore, lactone 1 induced PKCδ translocation simultaneously to the plasma membrane and to the internal membranes [10]. The translocation to the cellular membranes of both PKCα and δ was transient, unlike that caused by phorbol 12-myristate 13-acetate (the standard derivative used to characterize responses of PKC to phorbol esters or other ligands targeted to the C1 domain), or by the DAG-lactones containing rigid rod side chains described previously [10].

Figure 1.  PKC translocation. Confocal microscopy images of CHO cells overexpressing GFP–PKCδ (top) and GFP–PKCα (bottom), following treatment with (A) DAG lactone 1, (B) DAG lactone 2, (C) DAG lactone 3, and (D) DAG lactone 4. The final concentrations of all compounds were 10 μm.

Figure 1 shows that DAG-lactone 2 is more similar to DAG-lactone 1 and DAG-lactone 3 is more similar to DAG-lactone 4 for inducing PKC translocation to the membranes. Specifically, DAG-lactones 1 and 2, unlike PMA, gave rise to almost simultaneous translocation of PKC-δ to the plasma membrane, to the nuclear membrane, and to other internal membranes overall exhibiting a patchy distribution (Fig. 1A–B, top row). In contrast, 3 and 4, similarly to PMA, translocated PKC-δ in a sequential manner, initially to the plasma membrane and only later to the nuclear membrane and other internal membranes (Fig. 1C and D, top rows). Additionally, the PKC-α translocation induced by 3 and 4 appears to be somewhat slower than the corresponding process induced by 1 and 2 (Fig. 1A–D, bottom rows).

Chromatic vesicle analysis

To elucidate the mechanistic basis for the differences in the translocation patterns of PKC induced by DAG lactones 2–4 apparent in Fig. 1, we applied several biophysical techniques to characterize the membrane interactions of the molecules. Figure 2 shows the concentration dependence of the fluorescence chromatic response (%FCR, see Experimental procedures) following addition of DAG lactones 1–4 to biomimetic dimyristoylphosphatidylcholine (DMPC)/cholesterol/polydiacetylene (PDA) vesicles (mole ratio 1 : 1 : 3). Lipid/PDA vesicle assays have been used previously for analysis of lipid bilayer interactions of membrane-associated biological and pharmaceutical molecules [13,20]. Lipid/PDA vesicles consist of lipid bilayer domains (serving as biomimetic membrane docking areas) interspersed with PDA patches, which act as the chromatic reporting units [12–14]. The specific lipid composition used here, comprising DMPC and cholesterol, was designed to approximate cell-membrane environments [21].

Figure 2.  Fluorescence dose–response curves of DMPC/cholesterol/PDA vesicles based on fluorescence emissions induced in the lipid/PDA vesicles following addition of dag lactones 1–4.

The divergent chromatic dose–response curves in Fig. 2 indicate differences in membrane association of the DAG lactones 1-4. Specifically, DAG lactone 2 produced the most moderate increase in fluorescence chromatic response when incubated with the vesicles, while DAG lactone 4, in contrast, gave rise to the steepest %FCR dose–response curve (Fig. 2). The chromatic response curves of DAG lactones 1 and 3 were between those of DAG lactones 2 and 4, and were closer to that of DAG lactone 2.

Previous studies have correlated the steepness (i.e. slope) of the chromatic dose–response curves of lipid/PDA vesicles with the degree of bilayer insertion of the tested compounds [13,20]. Generally, molecules that penetrate deep into the hydrophobic core of the lipid bilayer induce lower chromatic transformations of the lipid/PDA vesicles (i.e. a more moderate increase for the chromatic dose–response curves). On the other hand, substances that show significant interactions with the lipid surface (lipid/water interface) were found to induce relatively more pronounced chromatic response (steeper dose–response curves) [22]. Thus, DAG lactone 2 most likely inserts deeper into the vesicle bilayers compared to the other DAG lactones examined, while DAG lactone 4 shows more pronounced surface interactions, inducing the steeper dose–response curve in Fig. 2.

Biophysical analysis

To further probe the interactions of DAG lactones 1–4 with lipid bilayers, particularly the extent of their localization at the vesicle interface, we performed fluorescence quenching experiments utilizing DMPC/cholesterol vesicles into which the fluorescence probe N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)1,2-dihexadecanoyl-sn-glycero-3-phosphoethanolamine triethylammonium salt (NBD-PE) was incorporated [15] (Fig. 3). The NBD dye is embedded close to the bilayer interface, providing a useful marker for surface interactions of membrane-active compounds. The experiments summarized in Fig. 3 indicated modulation of the fluorescence quenching of NBD by water-dissolved sodium dithionite following pre-incubation of the vesicles with the DAG lactones, providing a measure of membrane interactions of the compounds [15].

Figure 3.  Fluorescence quenching of NBD-PE embedded in DMPC/cholesterol vesicles. Fluorescence decay curves recorded following incubation of DAG lactones 1–4 with NBD-PE/DMPC/cholesterol vesicles (1 : 50 : 50 mole ratio), followed by addition of sodium dithionite. Control: no DAG lactone added.

Figure 3 demonstrates that incubation of the NBD-PE/DMPC/cholesterol vesicles with the DAG lactones studied yielded significant changes in the rate of dithionite-induced fluorescence quenching of the bilayer-embedded dye. Importantly, all the DAG lactones examined yielded lower quenching rates compared with the control vesicles (which were not pre-incubated with any DAG lactone prior to addition of sodium dithionite). This result suggests that the NBD dye became more ‘shielded’ from the soluble dithionite quencher as a consequence of vesicle binding by the DAG lactones.

Figure 3 further shows that DAG lactones 3 and 4 induced more moderate quenching of vesicle-embedded NBD compared to DAG lactones 1 and 2. This result is ascribed to greater shielding of the fluorescence dye by membrane interactions of DAG lactones 3 and 4, implying that these ligands are more localized at the vesicle surface compared to DAG lactones 1 and 2. The differences in fluorescence quenching profiles between DAG lactones 2 and 4, in particular, echo the results of the chromatic experiment shown in Fig. 3, which suggested the relatively deeper insertion of DAG lactone 2 into the lipid bilayer compared to the pronounced surface interaction of DAG lactone 4.

To further probe the effects of the four DAG lactones on the cooperative properties and molecular organization of the lipid bilayer, we examined the vesicles using differential scanning calorimetry (DSC) (Fig. 4 and Table 2). Figure 4 shows the effect of pre-incubating DAG lactone 2 with DMPC/cholesterol vesicles. The thermograms in Fig. 4 demonstrate that interactions of DAG lactone 2 with the phospholipids affected the peak position (i.e. the temperature at which the thermal transition, T_m, occurred [23]), the width at half-height (T_1/2, reflecting the ordering of phospholipid molecules undergoing the phase transition [23]), and the peak area (corresponding to ΔH, the overall enthalpy change associated with the thermal transition [23]).

Figure 4.  DSC analysis. Effect on the DSC traceof incubating DMPC/cholesterol vesicles with DAG lactone 2. Control: no DAG lactone added.

Table 2.   Parameters extracted from the DSC thermograms. T_m, maximum of the DSC spectrum (weighted average) (°C); ΔH, enthalpy change (cal·mol⁻¹).

DAG lactone added	Tm	ΔH
Control (no addition)	23.1	3340
1	21.1	3000
2	20.6	2340
3	21.8	1380
4	21.8	1080

The experimental parameters derived from the DSC experiments using DMPC/cholesterol vesicles incubated with DAG lactones 1–4 are shown in Table 2. All four DAG lactones significantly altered the DSC spectral parameters as a consequence of interaction of the compounds with the lipid bilayer. Both T_m (maximum of DSC spectra) and ΔH (enthalpy change calculated from the peak areas) significantly decreased as a consequence of incubation of the DMPC/cholesterol vesicles with the DAG lactones. However, the DSC parameters in Table 2 indicate that the compounds separate into two groups. Specifically, DAG lactones 1 and 2 gave rise to lower T_m values than DAG lactones 3 and 4. Even more pronounced were the changes in the ΔH values. While DAG lactones 1 and 2 reduced ΔH by 1000 cal·mol⁻¹ or less, DAG lactones 3 and 4 yielded a decrease in ΔH of more than 2000 cal·mol⁻¹ (Table 2). This disparity between the two clusters (1 and 2 versus 3 and 4) is similar to that shown by the fluorescence quenching results (Fig. 3), and likewise is probably attributable to two distinct mechanisms of membrane interactions by the DAG lactones (see Discussion).

To further elucidate the effects of the DAG lactones upon the dynamics of lipid molecules and bilayer fluidity, we measured the fluorescence anisotropy of trimethyl-ammonium-1,6-phenyl-1,3,5-hexatriene (TMA-DPH), a widely used probe that shows sensitivity to the dynamics of its lipid environment [24]. The DPH dye embedded in the lipid vesicles is located within the headgroup region close to the lipid/water interface [25], and thus provides insight into the dynamic consequences of molecular interactions at the bilayer surface [25].

Similar to the results of the fluorescence quenching (Fig. 3) and DSC analysis (Table 2), the fluorescence anisotropy data in Fig. 5 highlight two groupings among the DAG lactones examined. Incubation of the TMA-DPH/DMPC/cholesterol vesicles with DAG lactones 1 and 2 resulted in relatively small changes in the fluorescence anisotropy of the lipid-embedded dye (Fig. 5), indicating that lipid interactions of these two DAG lactones had little effect upon the bilayer fluidity. In contrast, DAG lactones 3 and 4 gave rise to a significant reduction of fluorescence anisotropy, suggesting greater lipid mobility around the DPH probe [16].

Figure 5.  Fluorescence anisotropy. Fluorescence anisotropy of DPH-TMA embedded in DMPC/cholesterol vesicles after addition of DAG lactones 1–4. Control: no compound added to vesicles.

To visualize the effect of the DAG lactones on lipid vesicles, we performed cryo-TEM experiments (Fig. 6). The cryo-TEM images in Fig. 6 reveal the pronounced morphological consequences of DAG lactone interactions with the vesicles, and highlight the different effects induced by the ligands. Prior to addition of the DAG lactones, the DMPC/cholesterol vesicles have a circular shape with relatively uniform sizes (diameters < 100 nm) (Fig. 6A) [18]. However, after incubation with DAG lactones 2 (Fig. 6B) or with 4 (Fig. 6C), the cryo-TEM images indicate dramatic structural effects. DAG lactone 2 appears to have induced internalization of vesicles within each other, giving rise to ‘onion-shape’ structures (Fig. 6B). DAG lactone 4, on the other hand, induced the formation of giant vesicular structures, each comprising a single vesicle embedded within another, that do not have precise circular structures (Fig. 6C). While we cannot speculate on the exact mechanisms leading to the distinct structural transformations induced by the two DAG lactones examined, the cryo-TEM data in Fig. 6 clearly distinguish between the bilayer interactions of DAG lactones 2 and 4, consistent with the chromatic and biophysical analyses discussed above.

Figure 6.  Morphological effects of DAG lactones probed by cryo-TEM. (A) Control DMPC/cholesterol vesicles (no DAG lactones added), (B) vesicles incubated with DAG lactone 2, and (C) vesicles incubated with DAG lactone 4. Scale bar = 100 nm.

Discussion

The DAG lactones substituted with rigid rods studied here are effective modulators of PKC both in vitro and in intact cells; however, differences in the patterns of PKC translocation to membranes were apparent between DAG lactones 1 and 2 on one hand, and DAG lactones 3 and 4 on the other hand (Fig. 1). The experimental data presented here offer a mechanistic explanation for these differences, and suggest that membrane interaction and insertion of the DAG lactones are important factors modulating the biological properties of these synthetic ligands.

Previous analysis of DAG lactone libraries clearly indicated that the nature of the hydrophobic substituents on the DAG lactones had a major influence on the specificity of their biological activities [7]. The present study provides direct evidence that DAG lactones 1–4 show two distinct modes of bilayer interactions, closely affected by the biphenyl rods and the properties of the R terminus in particular. Moreover, the two modes of membrane binding might account for the apparent differences in PKC translocation patterns and kinetics.

The experiments presented here suggest deep bilayer insertion of DAG lactones 1 and 2, compared to localization of DAG lactones 3 and 4 at the lipid bilayer surface (Fig. 7). This interpretation is supported by the chromatic vesicle data (Fig. 2, although the experiment did not unequivocally distinguished surface binding of DAG lactone 3) and the fluorescence quenching analysis (Fig. 3). The DSC experiments (Fig. 4 and Table 2) highlight the pronounced effects of the DAG lactones upon the structure of the lipid bilayers, and corroborate the two distinct membrane interaction mechanisms proposed in Fig. 7. Specifically, deep insertion of DAG lactones 1 and 2 into the bilayer is expected to modulate the lipid organization, and consequently results in significant changes in the position and width of the thermal transition (Table 2). In comparison, association of DAG lactones 3 and 4 at the lipid/water interface essentially leads to ‘pinning down’ of the lipid molecules in direct contact with the surface-attached DAG lactones. These lipids, as a result, do not participate in the phase transition, thereby significantly reducing the ΔH values obtained from the DSC thermograms (Table 2).

Figure 7.  Structural models. Schematic drawings depicting the proposed modes of association of the DAG lactones with lipid bilayers. (A) Surface binding of DAG lactone 4; (B) deep bilayer insertion of DAG lactone 2.

The fluorescence anisotropy analysis (Fig. 5) yields additional insight into the effects of the DAG lactones upon the dynamic characteristics of the lipid bilayer. Fig. 5 shows that DAG lactones 1 and 2 hardly modulate the fluorescence anisotropy of the DPH dye, consistent with deeper penetration of these compounds into the lipid bilayer, and, as a consequence, lesser disruption of the bilayer interface at which the fluorescence probe is localized. In contrast, the enhanced fluidity (i.e. lower fluorescence anisotropy) apparent following incubation of the vesicles with DAG lactones 3 and 4 most likely corresponds to the pronounced interactions of these DAG lactones with the lipid headgroup region.

Together, the experimental data demonstrate that the size of the side residue R (Table 1) is most likely the primary factor determining the extent of binding and insertion of the DAG lactones into the lipid bilayer. Paradoxically, although they increase the hydrophobicity of the molecule, bulky residues such as i-Pr (DAG lactone 3) and t-Bu (DAG lactone 4) minimize penetration of the rods into the phospholipid bilayer, most likely resulting in their accumulation at the interface between the acyl chains and headgroup region of the lipids. On the other hand, when hydrogen is present at the rod terminus (DAG lactone 2), the ligand is capable of deep insertion into the more hydrophobic alkyl core of the bilayer. Our analysis indicates that the flexibility of the DAG lactone side residue also contributes to bilayer insertion; DAG lactone 1, which does not possess a rigid diphenyl side chain but instead a saturated alkyl side chain of similar length, appears to penetrate deep into the lipid bilayer.

The two modes of DAG lactone/membrane interactions might explain the differences in the patterns and kinetics of PKC translocation from cytosol to membranes observed for these ligands. The apparent insertion of DAG lactones 1 and 2 into the lipid bilayer most likely leads to anchoring of the molecules within the membrane and more effective binding between the DAG moiety and PKC, accounting for the rapid translocation of PKCα and simultaneous translocation of PKCδ onto all cellular membranes induced by these two compounds. In contrast, the interfacial bilayer adsorption of DAG lactones 3 and 4 probably disrupts presentation of the DAG units, thereby slowing the translocation of both isoforms of PKC (apparent in Fig. 1). This interpretation might also explain the lower translocation of PKCδ to the internal membranes, which might be more sensitive to DAG lactone orientation within the lipid bilayer.

The different patterns of interaction of synthetic DAG lactone ligands with membranes point to both opportunities and challenges in drug design. We have previously described how the pattern of substitution can contribute to the orientation, either sn-1 or sn-2, of the insertion of the DAG lactone into the binding cleft of the C1 domain [9]. The nature of membrane insertion of the ligand would likewise have great influence. Indeed, previous combinatorial analysis has illustrated the pronounced sensitivity and selectivity of biological responses to the properties of the hydrophobic domain of the DAG lactone ligands [10]. Overall, this work indicates that the side residues of DAG lactones, rather than simply conferring hydrophobicity, affect membrane interactions of these synthetic ligands, thus directly modulating their biological functions.

Experimental procedures

Materials

Dimyristoylphosphatidylcholine (DMPC) was purchased from Avanti (Alabaster, AL, USA). Sodium dithionite (Na₂O₄S₂) and cholesterol were purchased from Sigma (St Louis, MO, USA). The diacetylenic monomer 10,12-tricosadiynoic acid was purchased from Alfa Aesar (Karlsruhe, Germany), and purified by dissolving the powder in chloroform, filtering the resulting solution through a 0.45 μm nylon filter (Whatman Inc., Clifton, NJ, USA), and evaporation of the solvent. The fluorescent probes N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)1,2-dihexadecanoyl-sn-glycero-3-phosphoethanolamine triethylammonium salt (NBD-PE) and 1-(4-trimethylammoniumphenyl)-6-phenyl-1,3,5-hexatriene (TMA-DPH) were purchased from Molecular Probes Inc. (Eugene, OR, USA). Buffer solutions were passed through a 0.2 μm nylon filter (Whatman Inc.) to remove impurities.

Synthesis

The synthesis of compounds 1 [10] and 2 [6] has been reported previously, and similar synthesis procedures were employed for compounds 3 and 4. The complete synthetic methodology and full characterization of these compounds are given in Docs S1–S3.

Determination of binding of compounds to PKC

The binding affinity of ligands for murine PKCα was determined by competition with [20-³H]phorbol 12,13-dibutyrate as described previously [26]. Briefly, binding of [20-³H]phorbol 12,13-dibutyrate to PKCα was determined in the presence of the competing DAG lactone. Assays were performed for 5 min at 37 °C in the presence of 100 μg·mL⁻¹ phosphatidylserine, 0.1 mm Ca²⁺, 50 mm Tris/Cl pH 7.4 and 2 mg·mL⁻¹ IGG. The reaction mixture was then chilled to 4 °C, and the protein–[20-³H]phorbol 12,13-dibutyrate complex was precipitated by addition of 35% polyethylene glycerol. Samples were subjected to centrifugation (at 8000 g in a Beckman Allegra 21R centrifuge at 4 °C for 15 min), the supernatant was removed, and radioactivity in pellet and supernatant was determined. Inhibition curves were determined using seven concentrations of DAG lactone, with triplicate determinations at each concentration in each experiment. The 50% inhibitory concentration of DAG lactone was derived from the least-squares fit of the data to a non-cooperative competition curve, and the K_i was calculated from the 50% inhibitory value. Triplicate independent experiments were performed for each DAG lactone, and the K_i values presented represent the means ± SEM of these triplicate experiments.

Measurement of translocation of GFP-tagged PKC isoforms α and δ to the plasma membrane and to internal membranes in response to ligand addition

Chinese hamster ovary (CHO) cells were purchased from the American Type Culture Collection (Manassas, VA, USA), and cultured in F12-K medium supplemented with 10% fetal bovine serum and antibiotics (penicillin at 50 units·mL⁻¹ and streptomycin at 0.05 mg·mL⁻¹). CHO cells plated onto T delta dishes (Bioptechs Inc., Butler, PA, USA) were transfected with GFP-tagged PKCα or PKCδ using Lipofectamine reagent (Invitrogen, Carlsbad, CA, USA). After 24 h, cells were visualized using a Zeiss LSM 510 confocal microscope (Carl Zeiss Inc., Thornwood, NY, USA) with a Zeiss Axiovert 100M inverted microscope operating with a 25 mW argon laser tuned to 488 nm. Cells were imaged using a 63× 1.4 NA Zeiss Plan-Apochromat oil immersion objective and with varying zooms (1.4–2). Time-lapse images were collected every 30 s before and after treatment with the indicated compounds (diluted in dimethylsulfoxide, final concentration 0.1%) using the Zeiss aim software, with green emission collected in a photomultiplier tube with an LP 505 filter. All experiments are representative of at least three independent experiments.

Vesicle preparation

Preparation of vesicles containing DMPC, cholesterol and the diacetylene monomer 10,12-tricosadiynoic acid (1 : 1 : 3 mole ratio) was performed by dissolving the constituents in chloroform/ethanol (1 : 1) and drying together in vacuo to constant weight. This was followed by addition of deionized water to a final concentration of 1 mm, and subsequently probe sonication at 70 °C for 3 min. The vesicle solution was then cooled to room temperature and kept at 4 °C overnight prior to polymerization by irradiation at 254 nm for 0.5 min, resulting in an intense blue solution. Regular unilamellar vesicles composed of DMPC and cholesterol (1/1 mole ratio) were prepared by sonication of the aqueous lipid mixtures at room temperature for 10 min.

Multi-well fluorescence spectroscopy

PDA fluorescence was measured in 96-well microplates (Greiner Bio-One GmbH, Frickenhausen, Germany) on a Fluoroscan Ascent fluorescence plate reader (Thermo Vantaa, Finland). All measurements were performed at room temperature at 485 nm excitation and 555 nm emission using LP filters with normal slits. Acquisition of data was automatically performed every 40 s for 20 min, and the last measurement is presented. Samples comprised 30 μL vesicle solution and DAG lactone (1–20 μL), followed by addition of 30 μL 50 mm Tris-base buffer (pH 8.2).

A quantitative value for the extent of the blue-to-red color transitions within the PDA-labeled vesicles is given by the fluorescence colorimetric response (%FCR), which is defined as follows [20]:

• 

where F_I is the fluorescence measurement of the vesicles after addition of the compounds, F₀ is the fluorescence of the control sample (without addition of the compounds), and F₁₀₀ is the fluorescence of a positive control sample (heated to produce the highest fluorescence emission of the red PDA phase).

Fluorescence quenching

NBD-PE was added to the DMPC/cholesterol vesicles at a molar ratio of 1 : 100 (probe:total phospholipids), and the mixture was dried in vacuo prior to sonication. Samples were prepared by mixing a selected amount of DAG lactones with 30 μL of the vesicles containing the fluorescent probe, and 30 μL of 50 mm Tris-base buffer (pH 8.2), followed by addition of distilled water to a total volume of 1.5 mL. The quenching reaction was initiated by adding sodium dithionite from a stock solution of 0.6 m in 50 mm Tris-base buffer (pH 11), to give a final concentration of 1 mm. The decrease in fluorescence emission was recorded for 5 min at room temperature using 469 nm excitation and 530 nm emission on an FL920 spectrofluorimeter (Edinburgh Instruments Ltd, Edinburgh, UK). The fluorescence decay curves were calculated as a percentage of the initial fluorescence measured before addition of dithionite [15].

Fluorescence anisotropy

The fluorescence probe TMA-DPH was incorporated into the DMPC/cholesterol vesicles by adding the dye dissolved in tetrahydrofuran (1 mm) to the vesicle solution and incubating for 30 min at room temperature. DAG lactones were added to 30 μL of the TMA-DPH/DMPC/cholesterol vesicles and 30 μL buffer (pH 8.2), followed by addition of distilled water to a total volume of 1.5 mL. TMA-DPH fluorescence anisotropy was measured at 428 nm (excitation 360 nm) on an FL920 spectrofluorimeter. Anisotropy values (r) were automatically calculated by the spectrofluorimeter software using conventional methodology [16].

Differential scanning calorimetry (DSC)

DSC experiments were performed on a VP-DSC calorimeter (MicroCal, Piscataway, NJ, USA). Vesicle concentrations used in the experiments were 2 mm. Distilled water served as a blank. Heating scans were run at a rate of 1 °C·min⁻¹. Data analysis was performed using microcalorigin 6.0 software [17].

Cryogenic transmission electron microscopy (cryo-TEM)

Specimens studied by cryo-TEM were prepared in a similar manner to the samples for the lipid/PDA vesicle analysis, described above. Sample solutions (4 μL) were deposited on perforated polymer films supported on a 300 mesh carbon-coated electron microscopy grid [copper, Ted Pella Inc. (Redding, CA, USA) Lacey substrate]. Ultrathin films (10–250 nm) were formed by removing most of the solution by blotting. The process was performed in a vitrification system in which the temperature and relative humidity were controlled, using a Vitrobot automatic system (FEI, Hillsboro, OR, USA). Cryo-TEM images were recorded on a FEI Tecnai 12 G2 twin transmission electron microscope equipped with a Gatan 626 cold stage [18].

Acknowledgements

We are grateful to Dr L. Meshi and Dr E. Nativ-Roth for help with the cryo-TEM experiments. This research was supported in part by the Intramural Research Program of the US National Institutes of Health, Center for Cancer Research, National Cancer Institute.