Phosphatidylinositol-4,5-Bisphosphate Enhances Anionic Lipid Demixing by the C2 Domain of PKCα

Abstract

The C2 domain of PKCα (C2α) induces fluorescence self-quenching of NBD-PS in the presence of Ca2+, which is interpreted as the demixing of phosphatidylserine from a mixture of this phospholipid with phosphatidylcholine. Self-quenching of NBD-PS was considerably increased when phosphatidylinositol-4,5-bisphosphate (PIP2) was present in the membrane. When PIP2 was the labeled phospholipid, in the form of TopFluor-PIP2, fluorescence self-quenching induced by the C2 domain was also observed, but this was dependent on the presence of phosphatidylserine. An independent indication of the phospholipid demixing effect given by the C2α domain was obtained by using 2H-NMR, since a shift of the transition temperature of deuterated phosphatidylcholine was observed as a consequence of the addition of the C2α domain, but only in the presence of PIP2. The demixing induced by the C2α domain may have a physiological significance since it means that the binding of PKCα to membranes is accompanied by the formation of domains enriched in activating lipids, like phosphatidylserine and PIP2. The formation of these domains may enhance the activation of the enzyme when it binds to membranes containing phosphatidylserine and PIP2.

About the Authors:

Antonio L. Egea-Jiménez

Affiliation: Departamento de Bioquímica y Biología Molecular-A, Facultad de Veterinaria, Regional Campus of International Excellence “Campus Mare Nostrum”, Universidad de Murcia, Murcia, Spain

Ana M. Fernández-Martínez

Affiliation: Departamento de Bioquímica y Biología Molecular-A, Facultad de Veterinaria, Regional Campus of International Excellence “Campus Mare Nostrum”, Universidad de Murcia, Murcia, Spain

Ángel Pérez-Lara

Affiliation: Departamento de Bioquímica y Biología Molecular-A, Facultad de Veterinaria, Regional Campus of International Excellence “Campus Mare Nostrum”, Universidad de Murcia, Murcia, Spain

Ana de Godos

Affiliation: Departamento de Bioquímica y Biología Molecular-A, Facultad de Veterinaria, Regional Campus of International Excellence “Campus Mare Nostrum”, Universidad de Murcia, Murcia, Spain

Senena Corbalán-García

Affiliation: Departamento de Bioquímica y Biología Molecular-A, Facultad de Veterinaria, Regional Campus of International Excellence “Campus Mare Nostrum”, Universidad de Murcia, Murcia, Spain

Juan C. Gómez-Fernández

* E-mail: [email protected]

Affiliation: Departamento de Bioquímica y Biología Molecular-A, Facultad de Veterinaria, Regional Campus of International Excellence “Campus Mare Nostrum”, Universidad de Murcia, Murcia, Spain

Introduction

Protein kinase C (PKC) is a family of related protein kinases that participate in cell control. The PKCs family includes eleven different mammalian isoenzymes, which are usually subdivided into three groups. Those in the first group are called classical or conventional (α, βI, βII and γ) and they are activated by anionic phospholipids, diacylglycerol (or phorbol esters) and by Ca2+. The second group includes the isozymes called novel or new (δ, ε, η and θ), which are also regulated by anionic phospholipids and diacylglycerols (or phorbol esters) but not by calcium. Finally there is another group which is relatively different, which includes the isoenzymes called atypical (ζ and ι/λ), which are not regulated neither by diacylglycerol nor by calcium [1]–[3]. See for comprehensive reviews about this family of enzymes the following references [4]–[7].

Classical PKCs are activated, at least, at three different sites; by diacylglycerols interacting with the C1 domain [6], by Ca2+ acting at the C2 domain, which bridges with membrane phospholipids, especially phosphatidylserine [8]–[10] and by PIP2, which also acts at the C2 domain [11], [12]. In addition, membrane effects arising from different molecules like diacylglycerols may also activate these enzymes [13]. Interestingly, it has been shown that the site to where PIP2 (phosphatidylinositol-4,5-bisphosphate) binds in the C2 domain of classical PKCs is a groove formed by strands β3 and β4, which is rich in lysine residues and is sometimes called the polylysine site and the three-dimensional structure of the domain bound to soluble PIP2 was determined in pioneering studies [11], [12], [14]. This site has been shown to exhibit high specificity for PIP2 [12], [15]–[23] although it may also bind other lipids like phosphatidic acid or phosphatidylserine [11] or even retinoic acid [24].

It has been observed that negatively charged lipids are recruited to some protein binding sites as a result of their affinity for the protein [25]. It has been shown recently that another C2 domain, like that of synaptotagmin 1 [26], is able to induce demixing of phosphatidylserine and the aggregation of different membranes, such aggregation being attributed to the combined effect of C2A and C2B domains.

We used fluorescent probes in this work to show that the C2 domain of PKCα may induce the demixing of POPS (1-palmitoyl-2-oleoyl-sn-phosphatidylserine) in a membrane in which POPC (1-palmitoyl-2-oleoyl-sn-phosphatidylcholine) was also present, an effect that is Ca2+-dependent. Surprisingly, the demixing of phosphatidylserine by the C2 domain was very much increased if PIP2 was also present. Similarly, the C2 domain may also induce the demixing of PIP2 from (POPC) when Ca2+ is present but this effect requires the presence of POPS. In addition 2H-NMR showed that POPC is demixed from a POPC/POPS/PIP2 membrane in the presence of Ca2+ and no demixing was detected in the absence of PIP2.

Experimental

Materials

1-Palmitoyl-2-oleoyl-sn-glycero-3-phosphoserine (POPS); 1-palmitoyl (d31)-2-oleoyl-sn-glycero-3-phosphoserine (POPS-d31); 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC); 1-palmitoyl (d31)-2-oleoyl-sn- glycero-3-phosphocholine (POPC-d31), phosphatidylinositol 4,5-bisphosphate (PIP2); 1-oleoyl-2-[6-[(7-nitro-2-1,3-benzoxadiazol-4-yl) amino] hexanoyl]-sn-glycero-3-phosphoserine (NBD-PS); 1-oleoyl-2-[6-[4-(dipyrrometheneborondifluoride)butanoyl]amino]hexanoyl-sn-glycero-3-phosphoinositol-4,5-bisphosphate (TopFluor- PIP2, a derivative of BODIPY) were purchased from Avanti Polar Lipids (Birmingham, LA, U.S.A.). All other chemicals were highly pure and were obtained from Sigma Chemical Co. (Madrid, Spain).

Construction of Expression Plasmids

PKCα cDNA was a kind gift from Drs. Nishizuka and Ono (Kobe University, Kobe, Japan).

The cDNA encoding the C2 domain of PKCα (residues 158–285) was amplified by PCR using the following primers:

C2α:5′ CAAGAATTCAAGAGGGGGCGGATTTAC

C2α:3′ CAAAAGCTTGTATTCACCCTCCTCTTG

The resulting 381 bp PCR fragment was subcloned into the HindIII and EcoRI sites of the bacterial expression vector pET28c(+), in which the insert is fused to a six-histidine tag. The construct was confirmed by DNA sequencing.

Expression and purification of the His-PKCα-C2 domain

Expression and purification of C2 domains were performed essentially as described in [27]. Briefly, the pET28c(+) plasmid containing PKC-C2 domains were transformed into BL21 (DE3) Escherichia coli cells. Once the bacterial cultures had reached an OD 600 of approximately 0.6, they were induced by adding 0.5 mM isopropyl 1-thio-β-D-galactopyranoside (IPTG) for 5 h at 30°C. The cells were resuspended in lysis buffer containing 25 mM Hepes pH 7.4 and 300 mM NaCl and protease inhibitors (10 mM benzamidine, 1 mM PMSF, and 10 µg/mL-1 trypsin inhibitor) and lysed by sonication on ice for 10 seconds 10 times. The clarified supernatants were incubated with Ni-NTA agarose for 2 h at 4°C. The Ni beads were then washed with 20 ml of lysis buffer containing 20 mM imidazole, and the bound proteins were eluted by applying the same buffer containing 50, 250, and 500 mM imidazole. The sixhistidine tag was removed by thrombin digestion, finally, the C2α domain was washed with the same buffer and concentrated using an Ultrafree-5 centrifugal filter unit (Millipore Inc, Bedford, MA).

Protein concentrations were determined at 280 nm [28]. The purity of the samples was checked by a SDS-15% (w/v) polyacrylamide gel electrophoresis [29] and Coomassie blue (Sigma, St. Louis, MO, USA) staining, which revealed that it was higher than 95%. The native folding of the recombinant C2 domain produced in our laboratory was widely confirmed by our previous X ray crystallographic studies with the PKCα-C2 domain [11], [14], [24] and also by FTIR spectroscopy [30], [31].

Preparation of lipid vesicles

Lipid vesicles for FRET experiments were generated by mixing chloroform lipid solutions in the desired proportions and then dried from the organic solvent under a stream of oxygen-free nitrogen. The last traces of organic solvent were removed under vacuum for at least 3 h. Dried phospholipids mixtures were resuspended in the corresponding buffers by vigorous vortexing. Large unilamellar phosholipid vesicles (LUV) (100 nm in diameter) were generating by extruding the lipid emulsion through two stacked polycarbonate membranes (Millipore Inc., Bedford, USA) with a pore size of 100 nm (11 passes).

Samples for NMR were prepared using a total quantity of 15 mg of phospholipid, dissolving the desired amounts of phospholipids in chloroform before being mixed. Samples were then dried under a nitrogen stream and finally under vacuum. When the sample consisted of lipid only, this mixture was dispersed in 100 mM NaCl, 25 mM Hepes pH 7.4 buffer, 1 mM CaCl2 using deuterium-depleted water. If the sample also included protein, a sufficient amount of lyophilized C2 domain protein to give the final lipid/protein molar ratio of 40∶1 was dissolved in the buffer prepared with water, or with deuterium-depleted water in the case of 2H-NMR, and added to the dried lipid to prepare multilamellar vesicles. The sample was then centrifuged to eliminate unbound protein and the pellet dispersed in 300 µl of buffer and transferred to NMR glass tubes.

Fluorescence measurements

Steady-state fluorescence measurements were carried out on a Fluoromax-3 fluorimeter (Jobin Yvon, Edison, NJ) at 25°C in an assay buffer composed of 25 mM Hepes pH 7.4, 100 mM NaCl and 250 µM CaCl2. All measurements were performed at 25°C. Fluorescence intensities were determined according to Lakowicz [32]. For each measurement a total volume of 1.5 ml was used and desired amounts of protein were titrated into the LUV suspension and incubated for 10 min before each measurement. NBD-PS emission spectra was obtained using a 50 µM lipid suspension, by scanning from 490 nm to 600 nm with the excitation wavelength set to 460 nm (4 nm excitation and 4 nm emission slit widths). TopFluor-PIP2 emission spectra was obtained using a 50 µM lipid suspension by scanning from 498 nm to 600 nm with the excitation wavelength set to 490 nm (4 nm excitation and 4 nm emission slit widths).The fluorescence decreases due to the change in volume produced by the addition of protein solution were corrected by subtracting the value obtained by adding the same volume of buffer without protein as a blank. The percent quenching was calculated according with the equation: where Fp and F0 are the fluorescence intensity in the presence or in the absence of protein respectively.

2H –NMR experiments

2H-NMR experiments were carried out on a Bruker Avance 600 instrument (Bruker, Etlingen, Germany) at 92.123 MHz using the standard quadrupole echo sequence [33]. The spectral width was 150 KHz, with a 10 µs 90° pulse, 40 µs pulse spacing, 3.35 µs dwell time, 1 s recycling time and 50 Hz line broadening, with an accumulation of 15000 transients. Spectra were acquired at temperatures ranging from −14°C to 22°C, raising the temperature in 2°C steps. The first moment, M1, was calculated for each spectrum of the different samples and at each temperature using the following equation [34]. where ω is the frequency shift from the central (Larmor) frequency, f(ω) is the spectral intensity, x is the frequency shift range (between −60 and 60 kHz) and A is defined as:

Results

PKCα C2 domain induces demixing of POPS as seen by fluorescence quenching

Figure 1 shows that C2α induces a moderate quenching of the fluorescence of POPS labeled with NBD in one of their fatty-acyl chains. NBD-PS is used as a fluorescent analogue of POPS and it has been reported to be a self- quenching fluorophore [26]. The extent of self-quenching depends on the distance between fluorophores and it will increase if POPS molecules are demixed due to the action of the C2 domain. In a first assay LUV membranes were composed of POPC/POPS/NBD-PS in molar ratio 80∶19.5∶0.5 (Fig. 1A). Increasing concentrations of the C2 domain gave place to increasing fluorescence quenching (Fig. 1A), although as seen in Fig. 2, maximum quenching is about 13% at the maximum lipid to protein ratio tested. On the other hand, addition of EGTA produced a recovery of the fluorescence indicating that the effect depends on the presence of Ca2+ so that in its absence the level of quenching is about 3% only, approaching the experimental error (Fig. 2). The observed quenching is attributed to self-quenching due to aggregation of POPS as a consequence of the interaction with the C2 domain of PKCα.

[Figure omitted. See PDF.]

Figure 1. Fluorescence spectra of NBD-PS in LUVs in the presence of different molar ratios of total lipid to protein (C2α domain) at different lipid compositions.

The molar ratio of total lipid to protein are indicated. (A) POPC/POPS/NBD-PS (molar ratio 80/19.5/0.5), (B) POPC/POPS/PIP2/NBD-PS in molar ratio 75∶19.5∶5∶0.5. (C) POPC/NBD-PC (molar ratio 99.5/0.5). A Ca2+ concentration of 100 µM was present in the assays, but control samples to which 3 mM EGTA was added instead of Ca2+ are shown.

https://doi.org/10.1371/journal.pone.0095973.g001

[Figure omitted. See PDF.]

Figure 2. Fractional fluorescence quenching of NBD-PS as a function of the molar ratio of total lipid to protein (C2α domain) and the indicated mutated forms of the domain as shown in each case.

A Ca2+ concentration of 250 mM was present in the assays, but a control sample to which 3 mM EGTA was added instead of Ca2+ is shown. Error bars indicate the S.E. (n = 3).

https://doi.org/10.1371/journal.pone.0095973.g002

A clearly bigger effect of the C2 domain was observed when PIP2 was incorporated in the membrane (Fig. 1B) so that the concentration was POPC/POPS/PIP2/NBD-PS in molar ratio 75∶19.5∶5∶0.5. It can be observed that the spectra obtained for this last sample (Fig, 1B) were noisier than those in the absence of PIP2 (Fig. 1A) and this is probably due to aggregation of the vesicles. Nevertheless this aggregation did not affect substantially to the result, since a control experiment in which water soluble NBD-glycine was added to POPC/POPS/PIP2 (non-fluorescent) vesicles showed that there was a small increase in apparent fluorescence, due to scattering, which was lower of 5% of the signal even with the maximum concentration of C2 (results not shown).

As it can be seen in Fig. 2, in the presence of PIP2 a higher quenching effect was obtained, going in this case up to 32% at the maximum lipid to protein molar ratio tested (30.5∶1, total lipid to protein) and this is interpreted as if PIP2 would facilitate aggregation of NBD-PS induced by C2 domain. It can be observed that this effect also depends on Ca2+ and the addition of EGTA produced a recovery of fluorescence, although a residual quenching of about 8% remains in the absence of Ca2+ (Fig. 2) probably because the interaction of PIP2 with the C2α domain is not totally dependent on Ca2+ [18]. Fig. 1C shows a control experiment in which C2α domain was added to a POPC membrane containing 0.5 mol% of PC-NBD. It can be observed that no quenching effect was observed in this case in which anionic lipids are not present.

Fig. 3A shows that if site-directed mutations were introduced in the calcium binding site (D246/248N) [14] the addition of C2α domain to membranes containing POPC/POPS/NBD-PS (79∶20∶0.5, molar ratio) did not produce any substantial self-quenching effect with a maximum observed quenching of about 4% when the total molar lipid/protein was increased to 1∶1 (Figs. 2 and 3). The same lack of substantial effect was observed when a membrane containing POPC/POPS/PIP2/NBD-PS (74.5∶20∶5∶0.5, molar ratio) was used (Fig. 3B), with a maximum quenching of about 9% when the molar total lipid to protein ratio was 1∶1 (Fig. 2). This shows that the calcium binding site of the C2α domain is essential to induce lipid demixing.

[Figure omitted. See PDF.]

Figure 3. Fluorescence spectra of NBD-PS in LUVs in the presence of different molar ratios of total lipid to protein concentrations (C2α domain) and some mutated forms, the molar ratios of total lipid to protein are indicated.

(A) POPC/POPS/NBD-PS (molar ratio 80/19.5/0.5) in the presence of C2α D246N/D248N, (B) POPC/POPS/PIP2/NBD-PS in molar ratio 75∶19.5∶5∶0.5 in the presence of C2α D246N/D248N. (C) POPC/POPS/NBD-PS (molar ratio 80/19.5/0.5) in the presence of C2α K209A/K211A, (D) POPC/POPS/PIP2/NBD-PS in molar ratio 75∶19.5∶5∶0.5 in the presence of C2α. A Ca2+ concentration of 100 µM was present in the assays.

https://doi.org/10.1371/journal.pone.0095973.g003

Similar results were observed when the site-directed mutation (Fig. 3C) was introduced onto the PIP2–binding site, i.e. K209A/K211A, which has been shown to form part of this site [12]. When this modified C2α domain was added to POPC/POPS/NBD-PS (80∶20∶0.5 molar ratio) (Fig. 3C) the maximum observed quenching was about 6% when the molar total lipid to protein was 1∶1 (see Fig. 2) and with POPC/POPS/PIP2/NBD-PS (75∶20∶5∶0.5 molar ratio) (Fig. 3D) membranes, the self-quenching effect was very much reduced to about 7% (Fig. 2). These results underline the lack of capacity of the mutated protein to induce this effect when either of the two lipid binding sites of the protein was disabled.

In order to check if C2α domain may also induce the aggregation of PIP2, a fluorescent analogue as TopFluor-PIP2 was used, labelled in one of the fatty acyl chains. BODIPY fluorescent probes have been also found to suffer self-quenching in lipid vesicles [35]. If POPS was not present, with membranes composed of POPC/PIP2/TopFluor-PIP2 (95∶4.5∶0.5 molar ratio), Fig. 4A shows that the quenching was not occurring within experimental error when C2α was added (see also Fig. 5). When the lipid vesicles were composed of POPC/POPS/PIP2/TopFluor-PIP2 (75∶20∶5∶0.5 molar ratio) it was observed that some quenching occurred, being about 19.5% when the molar ratio of total lipid to protein was 30.5 (Fig. 4A and Fig. 5) and therefore it seems that the C2α domain is also able of demixing this phosphoinositide. This effect was also dependent on Ca2+ and the quenching decreased to about 1% in the presence of EGTA (see Fig. 5).

[Figure omitted. See PDF.]

Figure 4. Fluorescence spectra of TopFluor-PIP2 in LUVs containing POPC/PIP2/TopFluor- PIP2 in molar ratio 95∶4.5∶0.5 (A) and POPC/POPS/PIP2/TopFluor- PIP2 in molar ratio 75∶20∶4.5∶0.5 (B) with different molar ratios of total lipid to protein (C2α domain) as indicated.

A Ca2+ concentration of 250 µM was present in the assays, but control samples to which 3 mM EGTA was added instead of Ca2+ are shown.

https://doi.org/10.1371/journal.pone.0095973.g004

[Figure omitted. See PDF.]

Figure 5. Fractional fluorescence quenching of PIP2-TopFluor as a function of the molar ratios of total lipid to protein.

Molar ratios are indicated in each case. A Ca2+ concentration of 250 µM was present in the assays, but a control sample to which 3 mM EGTA was added instead of Ca2+ is shown. Error bars indicate the S.E. (n = 3).

https://doi.org/10.1371/journal.pone.0095973.g005

2H-NMR spectroscopy studies

To further confirm the phospholipid demixing indicated by the FRET experiments, 2H-NMR experiments were designed in the presence and in the absence of C2α domain. In some of these experiments deuterated POPC (POPC-d31) was used and in others the deuterated phospholipid was POPS (POPS-d31). Spectra were measured at different temperatures. When pure POPC, in a mixture POPC/POPCd31 (molar ratio 2.5∶1) was studied (Fig. 6A) at temperatures below the transition temperature (Tc) of the pure phospholipid, the spectra were powder patterns characteristic of a gel phase. At higher temperatures (0°C and higher), the spectra corresponded to a fluid-phase bilayer and were axially symmetric with some resolved quadrupole splittings arising from methylene segments in the acyl chains. In order to better discern the phase transition and other effects of the C2α domain on these membranes, the first spectral moment M1 was calculated for each spectrum. M1 measures the average spectral width, and since each phase has a distinct spectral width it will also have a characteristic M1, which is proportional to the average order parameter. The variations in M1 with temperature can be used to characterize membrane phase transitions and the molecular order of a membrane. The phase transition has its onset at −8°C and it ends at −2°C (Fig. 6A and Fig. 7). In this case, the addition of C2α domain did not have any effect on the phase transition (Fig. 6B and Fig. 7A).

[Figure omitted. See PDF.]

Figure 6. 2H-NMR representative spectra obtained at the temperatures shown.

(A) POPC/POPC-d31 (2.5∶1, molar ratio); (B) POPC/POPC-d31 (2.5∶1, molar ratio) in the presence of C2α domain (40∶1 phospholipid/protein molar ratio); (C) POPS/POPS d31 (3∶1 molar ratio) and (D) POPS/POPS d31 (3∶1 molar ratio) in the presence of C2α domain (40∶1 phospholipid/protein molar ratio).

https://doi.org/10.1371/journal.pone.0095973.g006

[Figure omitted. See PDF.]

Figure 7. Temperature-dependent changes of the M1 first moment versus temperature, for different samples as shown in each case.

(A) POPC/POPC-d31 (2∶1, molar ratio); POPC/POPC-d31 (2∶1, molar ratio) in the presence of C2α domain (40∶1 phospholipid/protein molar ratio); POPS/POPS d31 (2∶1 molar ratio) and POPS/POPS d31 (2∶1 molar ratio) in the presence of C2α domain (40∶1 phospholipid/protein molar ratio). (B) POPC/POPC d31/POPS (40∶40∶20, molar ratio); POPC/POPC d31/POPS (40∶40∶20, molar ratio) in the presence of C2α (molar ratio phospholipid/protein 40∶1); POPC/POPS d31 (80∶20, molar ratio); POPC/POPS d31 (80∶20, molar ratio) in the presence of C2α (molar ratio phospholipid/protein 40∶1). (C) POPC/POPC-d31/POPS/PIP2 (35∶40∶20∶5, molar ratio); POPC/POPC-d31/POPS/PIP2 (35∶40∶20∶5, molar ratio); POPC/POPS-d31/PIP2 (75∶20∶5, molar ratio and POPC/POPS-d31/PIP2 (75∶20∶5, molar ratio) in the presence of C2α domain (40∶1 phospholipid/protein molar ratio). Error bars indicate the S.E. (n = 3).

https://doi.org/10.1371/journal.pone.0095973.g007

When POPS was the labeled phospholipid (POPS/POPSd31, 3∶1 molar ratio), spectra that indicate gel state of the membrane were observed up to 8°C (Fig. 6C) and as shown by the M1 dependence on temperature (Fig. 7A), the transition temperature of gel to fluid begins at 8°C and it is very cooperative being completed at 10°C. The addition of C2α to POPS broadens the transition (Fig. 6D) and this can be clearly observed by the variation in the spectral moment M1 (Fig. 7A) and although the onset remains at 8°C it tails out until 20°C.

In the next step the mixture POPC/POPCd31/POPS (40∶40∶20, molar ratio) was studied and a gel pattern of the spectra was observed until −4°C and a fluid pattern at higher temperatures (Fig. 8A). Fig. 7B shows a plot of M1 versus temperature that indicates an onset of the phase transition at −4°C ending at −2°C. As seen in Fig. 8B the addition of C2α did not change the pattern of the spectra or the phase transition (Fig. 7B). If the labeled phospholipid was POPS (POPC/POPSd31, 80∶20 molar ratio) (Fig. 8C) the pattern of the spectra was similar to those seen above, when POPC was the labeled one in the absence of protein, demonstrating that before the addition of C2α there is a good mixing of the phospholipids and the onset of the phase transition is at −4°C, i.e. the same than it can be observed for the mixture in which the deuterated lipid is POPC, in the absence of protein (Fig. 7B). It should be noted, however, that when C2α was added to this mixture (POPC/POPSd31, 80∶20 molar ratio), the spectra showed a decrease in intensity as indicated by the appearance of the isotropic peak coming from the contaminating deuterated water (Fig. 8D). This is probably due to the smearing out of the signal coming from the POPS molecules interacting with the protein. However enough POPS molecules remain free of interaction with protein to leave the onset of the transition at −4°C, i.e. the same than in the absence of protein.

[Figure omitted. See PDF.]

Figure 8. 2H-NMR representative spectra obtained at the temperatures shown.

(A) POPC/POPC d31/POPS (40∶40∶20, molar ratio); (B) POPC/POPC d31/POPS (40∶40∶20, molar ratio) in the presence of C2α (molar ratio phospholipid/protein 40∶1); (C) POPC/POPS d31 (80∶20, molar ratio) and (D) POPC/POPS d31 (80∶20, molar ratio) in the presence of C2α (molar ratio phospholipid/protein 40∶1).

https://doi.org/10.1371/journal.pone.0095973.g008

When the mixture POPC/POPS-d31/PIP2 was studied (with a 75∶20∶5, molar ratio), the effect of adding PIP2 was insignificant since the pattern of the spectra (Fig. 9A) remain similar to those observed above for POPC/POPS mixtures (Fig. 8) in the absence of protein and also the transition temperature detected by plotting the M1 values has again its onset at −4°C (Fig. 7C). The presence of C2α did not appreciably alter the transition behaviour of the system with respect to the same phospholipid mixture in the absence of protein and this can be said for both the spectra pattern (Fig. 8B) and the onset of the transition at −4°C as seen by the plot of the M1 values (Fig. 7C). The addition of C2α domain to this last sample produced a decrease in the intensity of the spectra as it was also seen with POPC/POPSd31 as indicated by the appearance of the isotropic peak of the contaminant deuterated water. Nevertheless, the onset of the transition of the observed POPS molecules, remained at the same temperature, indicating that these spectra correspond to phospholipid which was not interacting with the protein.

[Figure omitted. See PDF.]

Figure 9. 2H-NMR representative spectra obtained at the temperatures shown.

(A) POPC/POPC-d31/POPS/PIP2 (35∶40∶20∶5, molar ratio); (B) POPC/POPC-d31/POPS/PIP2 (35∶40∶20∶5, molar ratio) in the presence of C2α (molar ratio phospholipid/protein 40∶1); (C) POPC/POPS-d31/PIP2 (75∶20∶5, molar ratio) and (D) POPC/POPS-d31/PIP2 (75∶20∶5, molar ratio) in the presence of C2 domain (40∶1 phospholipid/protein molar ratio).

https://doi.org/10.1371/journal.pone.0095973.g009

A different result of adding C2α was observed when studying samples that included PIP2. The sample POPC/POPC-d31/POPS/PIP2, (35∶40∶20∶5, molar ratio) (Fig. 9A) shows a similar pattern of spectra than the sample without PIP2 (Fig. 7A). Also the onset of the transition was located at −4°C (Fig. 8E). However, although the spectra obtained after the addition of C2α domain to this membrane (Fig. 9B) were very similar to those obtained for pure phospholipid, the phase transition occurred at a lower temperature, for example at −4°C the pattern of the spectrum had already changed, showing that the transition had taken place at a lower temperature. As can be seen in Fig. 7C the phase transition studied by plotting M1 values, was located at −8°C exactly as it happens for pure POPC (Fig. 7A) and in coincidence with the value observed before by using this same NMR technique [36]. This indicates that the association of the anionic phospholipid POPS with C2á demixed it from POPC when PIP2 was present. As the process leads to pure POPC, and as it was observed that PIP2 is also demixed by the protein domain (Fig. 5) it may be concluded that the PIP2 is also segregated.

Discussion

We have studied the effect of adding C2α to a membrane system composed of POPC plus anionic phospholipids like POPS and PIP2, which are known to activate PKCα. We show using fluorescence quenching, that the C2α domain induces NBD-POPS demixing and that Ca2+ is indispensable for this effect. It is known that Ca2+ bridges the binding of the C2α domain to POPS [11], [14]. Additionally Ca2+ neutralizes negatively charged residues present in the Ca2+-binding site facilitating the approximation of the protein to the membrane where negatively charged are present [14], [37]. Thus, due to these reasons, it is explained why Ca2+ is so important, although the C2α domain also interacts through some amino acid residues directly with POPS [38]. Some proteins or peptides such as polylysine [39] are able to induce the capping of negatively charged phospholipids such as POPS, demixing them from amphipathic lipids, as POPC, and producing domains rich in polylysine and POPS and others poor in both. The aggregation of PIP2 by annexin A2 has been described recently [40].

The C2α domain is known to interact not only with phosphatidylserine through its calcium-binding site but it also binds one molecule of PIP2 with a high degree of affinity in another binding site located in a site rich in lysine residues [12], [18], [19]. Other phospholipids such as POPS and POPA [11] and also other molecules like retinoic acid [24] may also bind to this site but this site shows the highest affinity for PIP2 [12], [18], [19].

The very interesting observation, made in this work is that, in the absence of PIP2, little self-quenching was detected. However, in the presence of PIP2 the self-quenching of POPS-NBD was considerably higher. When TopFluor-PIP2 was used the C2α domain produced self-quenching but only when POPS is also present, indicating that the interaction of the domain with both POPS and PIP2 are necessary to produce anionic phospholipid demixing, although Ca2+ is also needed.

The site-directed mutagenesis of the C2α domain showed that both the Ca2+-binding site and the site rich in lysine residues are necessary for the self-quenching of POPS-NBD and the mutation of any of them almost abolish the lipid demixing. Therefore the fluorescence experiments indicate that demixing of anionic phospholipids is very much enhanced when both lipid binding sites of C2α are occupied, the Ca2+ binding site with this cation and phosphatidylserine and the site rich in lysine residues with PIP2. When POPC-NBD was used no self-quenching was detected, thus demonstrating that there is not a specific interaction of the NBD fluorophore moiety with the C2α domain and that the C2α domain establishes an electrostatic interaction with anionic lipids, which leads to their demixing and that an amphoteric lipid like POPC is not affected by this protein. When pure POPC was studied by 2H-NMR, C2α did not affect the phase transition, confirming that the domain does not interact with POPC. However it interacts with POPS as seen in the experiment in which it is added to pure phosphatidylserine, widening the phase transition.

When the samples containing POPC/POPSd31 and POPC/POPSd31/PIP2 were studied it was observed that the addition of the C2α domain reduced the intensity of the spectra as indicated by the isotropic peak produced by the contaminating deuterated water, due to the widening of the signal corresponding to POPS molecules interacting with the protein. The 2H-NMR experiments indicated that POPC is demixed with respect to the anionic lipids, and shows a phase transition at the same temperature than pure POPC-d31. This can be explained if most PIP2 molecules are also excluded from POPC and form part of a lipid domain rich in POPS and PIP2, where the C2α domain molecules are bound.

In the case of the sample with POPC/POPSd31/PIP2, it was observed that the presence of the C2α domain does not apparently affect the transition temperature of this phospholipid. We interpret this as competing influences on the transition temperature, since although the segregation of POPC should increase this temperature, PIP2 segregated with POPS should decrease it. The result is that no significant change was observed in the mixture deuterated POPS/POPC and PIP2 in the presence of C2α domain.

When POPCd31 and POPS are mixed, but in the absence of PIP2, the NMR experiments did not allow to observe any demixing, although the fluorescence experiments showed that this may happens although in a limited extent. This is probably due to the lower sensitivity of the NMR technique for this type of event.

We hypothesize that the reason why POPS demixing by the addition of C2α was greatly increased by the presence of PIP2 could be due to the great increase in the protein binding to the membrane that is observed when PIP2 is added to a POPC/POPS membrane [15], [18], [27] and to the different docking of the C2α protein in the presence of PIP2, which has been shown to change with respect to a membrane without this phosphoinositide, by both ATR-infrared spectroscopy [41] and EPR site-directed spin-labeling and relaxation [42].

The physiological significance of the observed anionic lipid demixing can be suggested, since the concentrations of phospholipids used in these experiments are close to the concentrations found in animal plasma membranes. It is known at this respect that the concentration of POPS in the inner monolayer of eukaryotic plasma membranes, such as in erythrocyte or platelet cells, is roughly 20 mol% [43]–[45]. The physiological concentration of PIP2 has been described to be around 1 mol% of the total lipid of plasma membranes [46], [47] and it is likely to be concentrated in the inner monolayer at 2 mol%, which increase locally if it forms clusters or patches [48]. Therefore, it is likely that demixing of POPS and PIP2 may take place in the plasma membrane induced by the presence of C2α domain.

It has recently been described that the C2 domains of synaptotagmin 1 are capable of segregating POPS within complex membranes [26], and changes have been detected in the physical properties of POPS molecules interacting with the C2 domains which probably lead to their demixing. Although there is a high homology between these domains, it seems that the C2 domains of synaptotagmin 1 do not need of the simultaneous presence of PIP2 to induce the demixing of anionic phospholipids; this can be explained, at least partially, by a recent demostration that shows that the C2A domain of sypnatotagmin 1 does not conserve a lysine residue critical for the interaction with PIP2 [4], [49], [50]. In the case of synaptotagmin C2AB or C2B, it was reported using ATR-IR that these domains disordered the lipid acyl chains of phosphatidylserine [26]. Our 2H-NMR results show (see Fig. 7B and 7C) that C2α disorders the acyl chains of POPS since M1 values of POPC/POPS-d31 decreased in the presence of C2α at temperatures above the phase transition, when they are compared with the sample of just lipid. Nevertheless the decrease in the case of C2α is low and in some temperatures not significant.

In the case of synaptotagmin 1, the observed demixing may serve to facilitate membrane fusion. In the case of the C2α, the physiological implications are different and are probably linked to the activation of the enzyme, an effect supported by the establishment of domains rich in POPS and PIP2.

Author Contributions

Conceived and designed the experiments: JCG-F SC-G. Performed the experiments: ALE-J AMF-M AP-L AdG. Analyzed the data: JCG-F ALE-J AMF-M AP-L AdG. Contributed reagents/materials/analysis tools: JCG-F ALE-J AMF-M AP-L AdG SC-G. Wrote the paper: JCG-F SC-G.

Citation: Egea-Jiménez AL, Fernández-Martínez AM, Pérez-Lara Á, de Godos A, Corbalán-García S, Gómez-Fernández JC (2014) Phosphatidylinositol-4,5-Bisphosphate Enhances Anionic Lipid Demixing by the C2 Domain of PKCα. PLoS ONE 9(4): e95973. https://doi.org/10.1371/journal.pone.0095973

References

1. Mellor H, Parker PJ (1998) The extended protein kinase C superfamily. Biochem J 332 (Pt 2): 281–292.

2. Newton AC (2001) Protein kinase C: structural and spatial regulation by phosphorylation, cofactors, and macromolecular interactions. Chem Rev 101: 2353–2364.

3. Nishizuka Y (1995) Protein kinase C and lipid signaling for sustained cellular responses. FASEB J 9: 484–496.

4. Corbalan-Garcia S, Gomez-Fernandez JC (2006) Protein kinase C regulatory domains: the art of decoding many different signals in membranes. Biochim Biophys Acta 1761: 633–654.

5. Coussens L, Parker PJ, Rhee L, Yang-Feng TL, Chen E, et al. (1986) Multiple, distinct forms of bovine and human protein kinase C suggest diversity in cellular signaling pathways. Science 233: 859–866.

6. Nishizuka Y (1992) Intracellular signaling by hydrolysis of phospholipids and activation of protein kinase C. Science 258: 607–614.

7. Corbalan-Garcia S, Gomez-Fernandez JC (2014) Classical protein kinases C are regulated by concerted interaction with lipids: the importance of phosphatidylinositol-4,5-bisphosphate. Biophys Rev 6: 3–14.

8. Bolsover SR, Gomez-Fernandez JC, Corbalan-Garcia S (2003) Role of the Ca2+/phosphatidylserine binding region of the C2 domain in the translocation of protein kinase Calpha to the plasma membrane. J Biol Chem 278: 10282–10290.

9. Conesa-Zamora P, Lopez-Andreo MJ, Gomez-Fernandez JC, Corbalan-Garcia S (2001) Identification of the phosphatidylserine binding site in the C2 domain that is important for PKC alpha activation and in vivo cell localization. Biochemistry 40: 13898–13905.

10. Corbalan-Garcia S, Rodriguez-Alfaro JA, Gomez-Fernandez JC (1999) Determination of the calcium-binding sites of the C2 domain of protein kinase Calpha that are critical for its translocation to the plasma membrane. Biochem J 337 (Pt 3): 513–521.

11. Ochoa WF, Corbalan-Garcia S, Eritja R, Rodriguez-Alfaro JA, Gomez-Fernandez JC, et al. (2002) Additional binding sites for anionic phospholipids and calcium ions in the crystal structures of complexes of the C2 domain of protein kinase calpha. J Mol Biol 320: 277–291.

12. Guerrero-Valero M, Ferrer-Orta C, Querol-Audi J, Marin-Vicente C, Fita I, et al. (2009) Structural and mechanistic insights into the association of PKCalpha-C2 domain to PtdIns(4,5)P2. Proc Natl Acad Sci U S A 106: 6603–6607.

13. Gomez-Fernandez JC, Corbalan-Garcia S (2007) Diacylglycerols, multivalent membrane modulators. Chem Phys Lipids 148: 1–25.

14. Verdaguer N, Corbalan-Garcia S, Ochoa WF, Fita I, Gomez-Fernandez JC (1999) Ca(2+) bridges the C2 membrane-binding domain of protein kinase Calpha directly to phosphatidylserine. EMBO J 18: 6329–6338.

15. Corbalan-Garcia S, Garcia-Garcia J, Rodriguez-Alfaro JA, Gomez-Fernandez JC (2003) A new phosphatidylinositol 4,5-bisphosphate-binding site located in the C2 domain of protein kinase Calpha. J Biol Chem 278: 4972–4980.

16. Marin-Vicente C, Gomez-Fernandez JC, Corbalan-Garcia S (2005) The ATP-dependent membrane localization of protein kinase Calpha is regulated by Ca2+ influx and phosphatidylinositol 4,5-bisphosphate in differentiated PC12 cells. Mol Biol Cell 16: 2848–2861.

17. Rodriguez-Alfaro JA, Gomez-Fernandez JC, Corbalan-Garcia S (2004) Role of the lysine-rich cluster of the C2 domain in the phosphatidylserine-dependent activation of PKCalpha. J Mol Biol 335: 1117–1129.

18. Sanchez-Bautista S, Marin-Vicente C, Gomez-Fernandez JC, Corbalan-Garcia S (2006) The C2 domain of PKCalpha is a Ca2+ -dependent PtdIns(4,5)P2 sensing domain: a new insight into an old pathway. J Mol Biol 362: 901–914.

19. Guerrero-Valero M, Marin-Vicente C, Gomez-Fernandez JC, Corbalan-Garcia S (2007) The C2 domains of classical PKCs are specific PtdIns(4,5)P2-sensing domains with different affinities for membrane binding. J Mol Biol 371: 608–621.

20. Marin-Vicente C, Nicolas FE, Gomez-Fernandez JC, Corbalan-Garcia S (2008) The PtdIns(4,5)P2 ligand itself influences the localization of PKCalpha in the plasma membrane of intact living cells. J Mol Biol 377: 1038–1052.

21. Corbin JA, Evans JH, Landgraf KE, Falke JJ (2007) Mechanism of specific membrane targeting by C2 domains: localized pools of target lipids enhance Ca2+ affinity. Biochemistry 46: 4322–4336.

22. Falke JJ (2012) Lipid targeting domain with dual-membrane specificity that expands the diversity of intracellular targeting reactions. Proc Natl Acad Sci U S A 109: 1816–1817.

23. Evans JH, Murray D, Leslie CC, Falke JJ (2006) Specific translocation of protein kinase Calpha to the plasma membrane requires both Ca2+ and PIP2 recognition by its C2 domain. Mol Biol Cell 17: 56–66.

24. Ochoa WF, Torrecillas A, Fita I, Verdaguer N, Corbalan-Garcia S, et al. (2003) Retinoic acid binds to the C2-domain of protein kinase C(alpha). Biochemistry 42: 8774–8779.

25. Snel MM, de Kruijff B, Marsh D (1994) Interaction of spin-labeled apocytochrome c and spin-labeled cytochrome c with negatively charged lipids studied by electron spin resonance. Biochemistry 33: 7146–7156.

26. Lai AL, Tamm LK, Ellena JF, Cafiso DS (2011) Synaptotagmin 1 modulates lipid acyl chain order in lipid bilayers by demixing phosphatidylserine. J Biol Chem 286: 25291–25300.

27. Torrecillas A, Laynez J, Menendez M, Corbalan-Garcia S, Gomez-Fernandez JC (2004) Calorimetric study of the interaction of the C2 domains of classical protein kinase C isoenzymes with Ca2+ and phospholipids. Biochemistry 43: 11727–11739.

28. Pace CN, Vajdos F, Fee L, Grimsley G, Gray T (1995) How to measure and predict the molar absorption coefficient of a protein. Protein Sci 4: 2411–2423.

29. Laemmli UK (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227: 680–685.

30. Garcia-Garcia J, Corbalan-Garcia S, Gomez-Fernandez JC (1999) Effect of calcium and phosphatidic acid binding on the C2 domain of PKC alpha as studied by Fourier transform infrared spectroscopy. Biochemistry 38: 9667–9675.

31. Torrecillas A, Corbalan-Garcia S, Gomez-Fernandez JC (2003) Structural study of the C2 domains of the classical PKC isoenzymes using infrared spectroscopy and two-dimensional infrared correlation spectroscopy. Biochemistry 42: 11669–11681.

32. Lakowicz JR (1999) Principles of Fluorescence Spectroscopy. Kluwer Academic/Plenum Publishers, New York.

33. Davis JH (1983) The description of membrane lipid conformation, order and dynamics by 2H-NMR. Biochim Biophys Acta 737: 117–171.

34. Hsueh YW, Gilbert K, Trandum C, Zuckermann M, Thewalt J (2005) The effect of ergosterol on dipalmitoylphosphatidylcholine bilayers: a deuterium NMR and calorimetric study. Biophys J 88: 1799–1808.

35. Elvington SM, Nichols JW (2007) Spontaneous, intervesicular transfer rates of fluorescent, acyl chain-labeled phosphatidylcholine analogs. Biochim Biophys Acta 1768: 502–508.

36. Hsueh YW, Chen MT, Patty PJ, Code C, Cheng J, et al. (2007) Ergosterol in POPC membranes: physical properties and comparison with structurally similar sterols. Biophys J 92: 1606–1615.

37. Murray D, Honig B (2002) Electrostatic control of the membrane targeting of C2 domains. Mol Cell 9: 145–154.

38. Conesa-Zamora P, Gomez-Fernandez JC, Corbalan-Garcia S (2000) The C2 domain of protein kinase calpha is directly involved in the diacylglycerol-dependent binding of the C1 domain to the membrane. Biochim Biophys Acta 1487: 246–254.

39. Franzin CM, Macdonald PM (2001) Polylysine-induced 2H NMR-observable domains in phosphatidylserine/phosphatidylcholine lipid bilayers. Biophys J 81: 3346–3362.

40. Drucker P, Pejic M, Galla HJ, Gerke V (2013) Lipid segregation and membrane budding induced by the peripheral membrane binding protein annexin A2. J Biol Chem.

41. Ausili A, Corbalan-Garcia S, Gomez-Fernandez JC, Marsh D (2011) Membrane docking of the C2 domain from protein kinase Calpha as seen by polarized ATR-IR. The role of PIP. Biochim Biophys Acta 1808: 684–695.

42. Landgraf KE, Malmberg NJ, Falke JJ (2008) Effect of PIP2 binding on the membrane docking geometry of PKC alpha C2 domain: an EPR site-directed spin-labeling and relaxation study. Biochemistry 47: 8301–8316.

43. Verkleij AJ, Zwaal RF, Roelofsen B, Comfurius P, Kastelijn D, et al. (1973) The asymmetric distribution of phospholipids in the human red cell membrane. A combined study using phospholipases and freeze-etch electron microscopy. Biochim Biophys Acta 323: 178–193.

44. Chap HJ, Zwaal RF, van Deenen LL (1977) Action of highly purified phospholipases on blood platelets. Evidence for an asymmetric distribution of phospholipids in the surface membrane. Biochim Biophys Acta 467: 146–164.

45. Leventis PA, Grinstein S (2010) The distribution and function of phosphatidylserine in cellular membranes. Annu Rev Biophys 39: 407–427.

46. McLaughlin S, Wang J, Gambhir A, Murray D (2002) PIP(2) and proteins: interactions, organization, and information flow. Annu Rev Biophys Biomol Struct 31: 151–175.

47. Di Paolo G, De Camilli P (2006) Phosphoinositides in cell regulation and membrane dynamics. Nature 443: 651–657.

48. Huang S, Lifshitz L, Patki-Kamath V, Tuft R, Fogarty K, et al. (2004) Phosphatidylinositol-4,5-bisphosphate-rich plasma membrane patches organize active zones of endocytosis and ruffling in cultured adipocytes. Mol Cell Biol 24: 9102–9123.

49. Guillen J, Ferrer-Orta C, Buxaderas M, Perez-Sanchez D, Guerrero-Valero M, et al. (2013) Structural insights into the Ca2+ and PI(4,5)P2 binding modes of the C2 domains of rabphilin 3A and synaptotagmin 1. Proc Natl Acad Sci U S A 110: 20503–20508.

50. Corbalan-Garcia S, Gomez-Fernandez JC (2014) Signaling through C2 domains: More than one lipid target. Biochim Biophys Acta. Jan 16 . pii: S0005-2736(14)00010-8. doi: 10.1016/j.bbamem.2014.01.008. [Epub ahead of print].