Analysis of the Contribution of Saturated and Polyunsaturated

Jan 6, 2011 - E-mail: [email protected]. ...... Liberty Francois-Moutal , Myriam M. Ouberai , Ofelia Maniti , Mark E. Welland , Agniesz...
0 downloads 0 Views 870KB Size
Download PDF
pubs.acs.org/Langmuir © 2011 American Chemical Society

Analysis of the Contribution of Saturated and Polyunsaturated Phospholipid Monolayers to the Binding of Proteins†  Demers, Elodie  Philippe Calvez, Eric Boisselier, and Christian Salesse* LOEX/CUO - Recherche, Centre Hospitalier Affili e Universitaire de Qu ebec and D epartement d’Ophtalmologie, Facult e de M edecine, Universit e Laval, Qu ebec, Qu ebec, Canada G1S 4L8 Received October 11, 2010. Revised Manuscript Received November 30, 2010 The binding of peripheral proteins to membranes results in different biological effects. The large diversity of membrane lipids is thought to modulate the activity of these proteins. However, information on the selective binding of peripheral proteins to membrane lipids is still largely lacking. Lipid monolayers at the air/water interface are useful model membrane systems for studying the parameters responsible for peripheral protein membrane binding. We have thus measured the maximum insertion pressure (MIP) of two proteins from the photoreceptors, Retinitis pigmentosa 2 (RP2) and recoverin, to estimate their binding to lipid monolayers. Photoreceptor membranes have the unique characteristic that more than 60% of their fatty acids are polyunsaturated, making them the most unsaturated natural membranes known to date. These membranes are also thought to contain significant amounts of saturated phospholipids. MIPs of RP2 and recoverin have thus been measured in the presence of saturated and polyunsaturated phospholipids. MIPs higher than the estimated lateral pressure of biomembranes have been obtained only with a saturated phospholipid for RP2 and with a polyunsaturated phospholipid for recoverin. A new approach was then devised to analyze these data properly. In particular, a parameter called the synergy factor allowed us to highlight the specificity of RP2 for saturated phospholipids and recoverin for polyunsaturated phospholipids as well as to demonstrate clearly the preference of RP2 for saturated phospholipids that are known to be located in microdomains.

Introduction Peripheral proteins undergo either permanent or temporary interactions with membranes. They play critical roles and are involved in many biological functions such as ion conductivity, cell adhesion, membrane trafficking, and cell signaling.1,2 There is a large diversity of membrane lipids in terms of polar headgroups and fatty acyl chains that are thought to modulate the activity of membrane and peripheral proteins.1,3-5 Although a large amount of data is available on the structure of membranes, information on the selectivity of the interactions between proteins and membranes is still largely lacking. Local variations of the membrane lateral pressure will lead to differences in the membrane lateral density, organization, and structure.6,7 The use of lipid monolayers at the air/water interface is an interesting model membrane system for studying proteinmembrane interactions because it allows one to control many parameters, including the surface pressure and thus the lipid † Part of the Supramolecular Chemistry at Interfaces special issue. *Corresponding author. Phone (418) 682-7569. Fax (418) 682-8000. E-mail: [email protected].

(1) Cho, W.; Stahelin, R. V. Annu. Rev. Biophys. Biomol. Struct. 2005, 34, 119– 151. (2) Go~ni, F. M. Mol. Membr. Biol. 2002, 19, 237–245. (3) Dumas, F.; Lebrun, M. C.; Tocanne, J.-F. FEBS Lett. 1999, 458, 271–277. (4) DiNitto, J. P.; Cronin, T. C.; Lambright, D. G. Sci. STKE 2003, re16. (5) Kinnunen, P. K. J.; K~oiv, A.; Lehtonen, J. Y. A.; Ryt€omaa, M.; Mustonen, P. Chem. Phys. Lipids 1994, 73, 181–207. (6) Lemmon, M. A. Nat. Rev. Mol. Cell. Biol. 2008, 9, 99–111. (7) Van Den Brink-van der Laan, E.; Killian, A. J.; de Kruijff, B. Biochim. Biophys. Acta 2004, 1666, 275–288. (8) Brezesinski, G.; M^ohwald, H. Adv. Colloid Interface Sci. 2003, 100-102, 563–584.  Salesse, C. Biochimie 2009, 91, 718–733. (9) Calvez, P.; Bussieres, S.; Demers, E.; (10) Cho, W.; Bittova, L.; Stahelin, R. V. Anal. Biochem. 2001, 296, 153–161.11. (11) Dynarowicz-Latka, P.; Dhanabalan, A.; Oliveira, O. N. J. Adv. Colloid Interface Sci. 2001, 91, 221–93.

Langmuir 2011, 27(4), 1373–1379

lateral packing density.8-12 There is a direct thermodynamic relationship between monolayers and bilayers,13-16 and it has thus been extensively used to study lipid-protein interactions. It is also a useful approach for comparing the relative surface activity of different types of proteins or a particular protein with different lipid monolayers or subphase compositions as well as the activity of lipolytic enzymes (for reviews, see refs 9, 11-13, and 17 -22). The maximum insertion pressure (MIP) of proteins in lipid monolayers has been shown to be useful in characterizing protein adsorption and lipid specificity without the need for radiolabels or other tags.9 However, although these measurements are rather straightforward, their detailed analysis raised different unsolved questions such as how to interpret the fact that different slopes can provide the same value of MIP,9 which kinds of interactions are responsible for this observation, and the influence of the diversity of acyl chains in the insertion of proteins into membranes. The proteins Retinitis pigmentosa 2 (RP2) and recoverin from photoreceptors have been expressed and purified to clarify these ~ /β issues. The crystal structure of RP2 contains a C-terminal R (12) Maget-Dana, R. Biochim. Biophys. Acta 1999, 1462, 109–140. (13) Brockman, H. Curr. Opin. Struct. 1999, 9, 438–443. (14) Feng, S. S. Langmuir 1999, 15, 998–1010. (15) MacDonald, R. C.; Simon, S. A. Proc. Natl. Acad. Sci. U.S.A. 1987, 84, 4089–4093. (16) Marsh, D. Langmuir 2006, 22, 2916–2919. (17) Baszkin, A. Adv. Colloid Interface Sci. 2006, 128-130, 111–120. (18) Boucher, J.; Trudel, E.; Methot, M.; Desmeules, P.; Salesse, C. Colloids Surf., B 2007, 58, 73–90. (19) Winget, J. M.; Pan, Y. H.; Bahnson, B. J. Biochim. Biophys. Acta 2006, 1761, 1260–1269. (20) Brockman, H. L. Biochimie 2000, 82, 987–995. (21) Douchet, I.; De Haas, G.; Verger, R. Chirality 2003, 15, 220–226. (22) Ransac, S.; Ivanova, M.; Panaiotov, I.; Verger, R. Methods Mol. Biol. 1999, 109, 279–302.

Published on Web 01/06/2011

DOI: 10.1021/la104097n

1373

Article

Calvez et al.

domain whereas its N-terminal domain forms a β helix.23 RP2 was shown to localize to the membrane and to be predominantly associated with membrane microdomains.23-25 Recoverin is an N-myristoylated protein that comprises two functional EF-Hand motifs. Each functional EF-Hand is able to bind Ca2þ reversibly, which induces a structural change in recoverin.26-28 In the presence of Ca2þ, the N-myristoyl moiety of recoverin is extruded from the core of the protein that favors membrane binding.29,30 Photoreceptor membranes comprise more than 50% polyunsaturated fatty acids, mainly docosahexaenoic acid (22:6), as well as a significant amount of saturated fatty acids, mainly stearic acid (18:0).31 These membranes contain phosphatidylethanolamine (45%), phosphatidylcholine (38%), phosphatidylserine (14%), and a small percentage of phosphatidylinositol and sphingomyelin.32 Photoreceptor membranes were also shown to contain microdomains.33 In this article, we have thus determined the MIPs of RP2 and recoverin, which behave diametrically differently, in the presence of monolayers of saturated and polyunsaturated phospholipids. A new parameter, the synergy factor, was then described to analyze the MIP data further and allowed us to highlight the binding specificiy of RP2 and recoverin toward phospholipids and to draw conclusions about the interactions responsible for their membrane binding.

Materials and Methods Materials. E. coli strain BL21 (DE3) pLysS cells were from Novagen (Darmstadt, Germany). The deionized water used for the buffer solutions was prepared from a Barnstead Nanopure system (Barnstead, Dubuque, IA). Its resistivity and surface tension were 18.2 MΩ 3 cm and 72 mN/m at 20 °C, respectively. Ethylene glycol tetraacetic acid (EGTA), sodium myristate, Hepes, butylated hydroxytoluene (BHT), and β-mercaptoethanol were purchased from Sigma (St Louis, MO). CaCl2 and ultrapure NaCl (99.9%) were from J. T. Baker (Phillipsburg, NJ). Highgrade sodium phosphate was from Merck (Darmstadt, Germany). The resins used for protein purification were from GE Healthcare (Uppsala, Sweden). 1,2-Dimyristoyl-sn-glycero3-phosphocholine (DMPC), 1,2-distearoyl-sn-glycero-3-phosphocholine (DSPC), and 1,2-didocosahexaenoyl-sn-glycero3-phosphocholine (DDPC) were purchased from Avanti Polar Lipids (Alabaster, AL). All chemicals were used as received. The phospholipid solutions were prepared in chloroform at a concentration of 0.1 mg/mL. BHT (5 μg/mL) was added to the DDPC chloroform solution to prevent the oxidation of its fatty acyl chains. Methods. Expression and Purification of Proteins. The RP2 construct cloned in the pGEX-4T3 plasmid to express a GST fusion protein was a kind gift from Dr. Alfred Wittinghofer (23) K€uhnel, K.; Veltel, S.; Schlichting, I.; Wittinghofer, A. Structure 2006, 14, 367–378. (24) Chapple, J. P.; Grayson, C.; Hardcastle, A. J.; Bailey, T. A.; Matter, K.; Adamson, P.; Graham, C. H.; Willison, K. R.; Cheetham, M. E. Biochem. J. 2003, 372, 427–433. (25) Chapple, J. P.; Hardcastle, A. J.; Grayson, C.; Spackman, L. A.; Willison, K. R.; Cheetham, M. E. Hum. Mol. Genet. 2000, 9, 1919–1926. (26) Ames, J. B.; Ishima, R.; Tanaka, T.; Gordon, J. I.; Stryer, L.; Ikura, M. Nature 1997, 389, 198–202. (27) Ames, J. B.; Tanaka, T.; Stryer, L.; Ikura, M. Curr. Opin. Struct. 1996, 6, 432–438. (28) Burgoyne, R. D.; Weiss, J. L. Biochem. J. 2001, 353, 1–12. (29) Desmeules, P.; Grandbois, M.; Bondarenko, V. A.; Yamazaki, A.; Salesse, C. Biophys. J. 2002, 82, 3343–3350. (30) Desmeules, P.; Penney, S. E.; Desbat, B.; Salesse, C. Biophys. J. 2007, 93, 2069–82. (31) Salesse, C.; Boucher, F.; Leblanc, R. M. Anal. Biochem. 1984, 142, 258–266. (32) DeGrip, W. J.; Damens, F. J.; Bonting, S. L. Methods in Enzymology; Academic Press: New York, 1980. (33) Elliott, M. H.; Nash, Z. A.; Takemori, N.; Fliesler, S. J.; McClellan, M. E.; Naash, M. I. J. Neurochem. 2008, 104, 336–352.

1374 DOI: 10.1021/la104097n

(Max-Planck-Institut f€ ur Molekulare Physiologie, Germany). RP2 was expressed in E. coli BL21 (DE3) RIPL during 5 h at 37 °C. Bacteria were then formed into pellets by centrifugation. RP2 was purified by a modified method described by K€ uhnel et al.23 Briefly, pellets were resuspended in load buffer (50 mM Tris at pH 7.4, 100 mM NaCl, 5 mM MgCl2, 3 mM β-mercaptoethanol) at 4 °C. Bacteria lysis was done by sonication followed by centrifugation. Cell lysate supernatant was loaded on a GST Trap column that was pre-equilibrated with the load buffer. Contaminants were removed by washing the column with 10 volumes of load buffer. Thrombin cleavage overnight at 4 °C was then performed directly on the column. This protease was eliminated by the use of a HiTrap benzamidine FF (high sub) column located directly after the GST Trap column. The elution of pure RP2 was performed with a modified load buffer containing 500 mM NaCl. Amicon Ultra15 (Millipore) was then used to concentrate RP2 and to change the buffer (5 mM phosphate at pH 7.4 and 3 mM β-mercaptoethanol). RP2 was kept at 4 °C until used. Recoverin was expressed and purified essentially as previously reported.34 Briefly, recoverin was expressed in E. coli strain BL21 (DE3) pLysS containing plasmids encoding for recoverin (pET11a-REC) and N-myristoyl transferase (pBB131). To allow the N-myristoylation of recoverin, sodium myristate (0.08 mM) was added 20 min before the induction. Cells were then harvested by centrifugation and resuspended in a buffer containing 50 mM Hepes at pH 7.5, 100 mM NaCl, 5 mM β-mercaptoethanol, and 1 mM CaCl2. After sonication and centrifugation, the cleared lysate was loaded onto a column containing 5 mL of phenyl sepharose 6 fast flow (low sub). Protein was eluted with 5 mM Hepes at pH 7.5, 100 mM NaCl, 5 mM β-mercaptoethanol, and 5 mM EGTA. The concentration of recoverin and RP2 was determined using the Bradford method, and a myristoylation level of close to 100% recoverin was quantified by RP-HPLC as described.34 The identities of pure recoverin and pure RP2 were confirmed by mass spectrometry. Surface Pressure Measurements. Surface pressure isotherms have been measured with a Teflon-covered glass trough made by Kibron Inc. (Helsinki, Finland) with dimensions of 20.8 cm length, 5.7 cm width, and 20 mL total volume. The surface pressure (Π) was measured by the Wilhelmy method using a tensiometer from Nima Technology (Coventry, U.K.) for RP2 and a DeltaPi4 microtensiometer from Kibron Inc. for recoverin. The experimental setup was placed in a Plexiglas box with humidity control at room temperature. A 1200 μL home-built round Teflon trough and a 500 μL glass trough from Kibron Inc. were used for the RP2 and recoverin binding measurements, respectively. The subphase buffer was 5 mM phosphate buffer at pH 7.4 and 100 mM NaCl in the case of RP2 and 10 mM Hepes at pH 7.5, 100 mM NaCl, 1 mM CaCl2, and 5 mM β-mercaptoethanol for recoverin. The phospholipid monolayer was prepared by spreading a few microliters of a solution of phospholipids until the desired initial surface pressure (Πi) was reached. Depending on the initial surface pressure, a waiting period that varies with the type of lipid, the speading volume, and the lipid concentration is then provided for the film to reach equilibrium. This waiting period could vary between 20 and at most 60 min. Figures S1-S4 showing that the same optical properties and film morphology are obtained when preparing a DSPC monolayer either by compressing the DSPC monolayer from 0 mN/m or by successive DSPC spreading until the desired surface pressure of 30 mN/m is reached. Then, RP2 or recoverin was injected underneath the lipid monolayer until an optimal, saturating final concentration of 20 or 2 μg/mL was achieved, respectively. The kinetics of protein binding onto the phospholipid monolayer was monitored until the equilibrium surface pressure (Πe) was reached. No difference was observed when measurements were performed in the presence or the absence of N2 or Ar when using polyunsaturated DDPC. In addition, the (34) Desmeules, P.; Penney, S.; Salesse, C. Anal. Biochem. 2006, 349, 25–32.

Langmuir 2011, 27(4), 1373–1379

Calvez et al.

Article

Figure 1. Π-A isotherms of the phospholipids used in this study. Isotherms were obtained at compression rates of 10, 10, and 20 A˚2 molecule-1 min-1 for DMPC, DSPC, and DDPC, respectively. The measuring temperature was 21 °C for DDPC (2) and DSPC (() whereas it was 24 (b) or 4 °C (9) for DMPC. The subphase was pure water.

same protein adsorption isotherms were obtained in the presence or the absence of BHT in the DDPC solution. Moreover, no change in the surface pressure isotherm of DDPC shown in Figure 1 was observed in the absence of BHT. This is consistent with the observations of Chaiyasit et al. who have shown that increasing the concentrations of BHT in hexadecane did not change the interfacial tension of water, indicating that BHT has very low surface activity.35 Surface Pressure Kinetics. The surface pressure increase (ΔΠ) after the injection of the proteins corresponds to Πe - Πi, where Πi is the initial surface pressure. The curves of surface pressure as a function of time recorded during the adsorption of proteins onto phospholipid monolayers were fitted using the stretched exponential equation adapted to surfaceβ pressure measurements by Pitcher et al., Πt = Πe - Πie-(kt) , where Πt is the surface pressure of the monolayer at time t, k is the rate coefficient, t is the time, and β is the exponential scaling factor that is fixed with a starting value of 0.001 to optimize the fitting of the adsorption isotherms.36

Determination of the Maximum Insertion Pressure (MIP).

The determination of MIP is realized as described previously.9 Briefly, protein injection is performed at different Πi values of the lipid monolayer. Then, the plot of the surface pressure increase (ΔΠ) as a function of Πi allows the determination of the MIP by extrapolating the regression of the plot to the x axis. The uncertainty in the MIP is calculated from the covariance of the experimental data on the linear regression as described previously.9 The uncertainty in the slopes a of the curves of ΔΠ as a function of Πi (Figure 1) is calculated from the equation (σ(Πe)  (1 - r2)1/2)/(σ(Πi)  (n - 2)1/2), where σ is the standard deviation, r is the correlation coefficient, and n is the number of points.

Results Maximum Insertion Pressure of RP2 and Recoverin. The MIP of RP2 and recoverin was measured with two saturated phospholipids (DMPC and DSPC) and one polyunsaturated phospholipid (DDPC). The measurements with DMPC were performed at a temperature above (24 °C) and much lower (35) Chaiyasit, W.; McClements, D. J.; Decker, E. A. J. Agric. Food Chem. 2005, 53, 4982–8. (36) Pitcher, W. H.; Keller, S. L.; Huestis, W. H. Biochim. Biophys. Acta 2002, 1564, 107–113.

Langmuir 2011, 27(4), 1373–1379

Figure 2. Determination of the maximum insertion pressure (MIP) of proteins retinis pigmentosa 2 (RP2) and recoverin in the presence of different phospholipids: (b) DMPC (LE), (9) DMPC (LC), (() DSPC, and (2) DDPC. (A) Plot of the surface pressure increase (ΔΠ) as a function of the initial surface pressure (Πi) to determine the MIP of RP2 with DMPC in the liquid-expanded (LE) and liquid-condensed (LC) states and DSPC and DDPC monolayers by extrapolating the curve to the x axis. (Inset) Histograms of the MIP of RP2 with the same phospholipids. The subphase was 5 mM phosphate at pH 7.4 and 100 mM NaCl. The final RP2 concentration was 20 μg/mL. (B) Plot of ΔΠ as a function of Πi to determine the MIP of recoverin with DMPC in the LE and SC states and DSPC and DDPC monolayers. (Inset) Histograms of the MIP of recoverin with the same phospholipids. The subphase was 1 mM Hepes at pH 7.5, 100 mM NaCl, 1 mM CaCl2, and 5 mM β-mercaptoethanol. The final concentration of recoverin was 2 μg/mL.

(4 °C) than its phase-transition temperature (23 °C).37 As shown in Figure 1, the isotherm of DMPC is in the liquid-expanded (LE) state at 24 °C whereas it shows a phase transition from between the LE and the liquid-condensed (LC) states from ∼5.5 and 9.4 mN/m at 4 °C, which is consistent with previously reported data.37 In contrast, the isotherms of DSPC and DDPC at 21 °C are showing a single solid-condensed (SC) and LE state, respectively (Figure 1). These isotherms are very similar to those previously reported for DSPC and DDPC.38,39 The MIP obtained (37) Albrecht, O.; Gruler, H.; Sackmann, E. J. Phys. 1978, 39, 301. (38) Mingotaud, A. F.; Mingotaud, C.; Patterson, L. K. Handbook of Monolayers; Academic Press: London, 1993. (39) Brockman, H. L.; Applegate, K. R.; Momsen, M. M.; King, W. C.; Glomset, J. A. Biophys. J. 2003, 85, 2384–2396.

DOI: 10.1021/la104097n

1375

Article

Calvez et al.

by extrapolating the regression of the plot of ΔΠ as a function of Πi to the x axis is shown in Figure 2. As can be seen in Figure 2A, the MIP of RP2 is very different depending on the type of phospholipid used. MIP values of 16.6 ( 0.6, 20.9 ( 1.3, 25.5 ( 1.7, and 36.4 ( 2.4 mN/m have been respectively obtained for RP2 in the presence of DDPC, DMPC (LC), DMPC (LE), and DSPC, which are all significantly different. The lowest MIP has been obtained with polyunsaturated DDPC. These data contrast with those obtained with recoverin (Figure 2B). Indeed, the largest value of MIP of 48.6 ( 5.6 mN/m has been observed with DDPC. Moreover, similar MIP values of ∼20 mN/m have been obtained for DMPC (LC), DMPC (LE), and DSPC. It is noteworthy that only the MIP values obtained for RP2 with saturated DSPC (36 mN/m) and for recoverin with polyunsaturated DDPC (48 mN/m) are larger than the membrane lateral pressure that has been estimated to be in the range of 3035 mN/m.40-46 Moreover, these data also show that RP2 and recoverin have a preference for phospholipids in different physical states. Indeed, RP2 and recoverin preferentially bind phospholipids in the SC (DSPC) and LE (DDPC) states, respectively. However, similar results have been obtained for these two proteins with DMPC in the LE and LC states. Nevertheless, in the case of RP2, there is a large difference in the slope of the curve used to determine the values of MIP for DMPC in the LE and SC states (Figure 2A). Further analysis of the data allowed us to clarify this issue (see below). New Approach to Determine What Governs Protein Binding onto Lipid Monolayers. As can be seen in Figure 3, a relationship can be found between the equilibrium adsorption pressure Πe and the initial monolayer surface pressure Πi, which follows the linear regression Πe ¼ aΠi þ b

ð1Þ

By deriving eq 1, it can be shown that dΠe ¼ a

ð2Þ

Πe ¼ ΔΠ þ Πi

ð3Þ

and given that

a relationship can be found to quantify the evolution of ΔΠ and Πi: dΔΠ ¼ a - dΠi

ð4Þ

Therefore, the observed decrease in ΔΠ with the increase in Πi (Figure 2) is related by the parameter a, which modulates protein adsorption. It is indeed interesting to plot Πe as a function of Πi to determine the value of a that corresponds to the slope of this linear regression (Figure 3). Three different behaviors have been observed. A slope of a < 0 is observed for RP2 in the presence of DDPC (Figure 3A) whereas a slope of a > 0 has been obtained for recoverin and RP2 in the presence of DDPC (Figure 3B) and DSPC (Figure 3A), respectively. Finally, a slope of a = 0 can be (40) Demel, R. A.; Geurts van Kessel, W. S. M.; Zwaal, R. F. A.; Roelofsen, B.; van Deenen, L. L. M. Biochim. Biophys. Acta 1975, 406, 97–107. (41) Blume, A.; Eibl, H. Biochim. Biophys. Acta 1979, 558, 13–21. (42) Moreau, H.; Pieroni, G.; Jolivet-Reynaud, C.; Alouf, J. E.; Verger, R. Biochemistry 1988, 27, 2319–2323. (43) Seelig, A. Biochim. Biophys. Acta 1987, 899, 196–204. (44) Silvius, J. R. Biochim. Biophys. Acta 2003, 1610, 174–183. (45) Boguslavsky, V.; Rebecchi, M.; Morris, A. J.; Jhon, D. Y.; Rhee, S. G.; McLaughlin, S. Biochemistry 1994, 33, 3032–3037. (46) Marsh, D. Biochim. Biophys. Acta 1996, 1286, 183–223.

1376 DOI: 10.1021/la104097n

Figure 3. (A) Plot of the equilibrium adsorption pressure (Πe) of RP2 as a function of the initial surface pressure (Πi) of different phospholipids: (9) DSPC and (b) DDPC. (B) Plot of the equilibrium adsorption pressure (Πe) of recoverin as a function of the initial surface pressure (Πi) of different phospholipids: (9) DDPC and (b) DMPC (LE). Other conditions are the same as in the legend of Figure 2.

seen in Figure 3B for the binding of recoverin onto a DMPC monolayer in the LE state. The value of a < 0 is observed solely for RP2 binding onto DDPC (Figure 3A) for which the smallest value of MIP of ∼17 mN/m was observed (Figure 2A). It can also be seen that the Πe of RP2 decreases proportionally to the Πi of DDPC (Figure 3A). Given eq 3, it can be concluded that, in this case, protein insertion is governed by Πi and thus by a negative synergy between the protein and the lipid that is related to a. The value of a = 0 obtained for recoverin with a DMPC monolayer in the LE state means that Πe remains almost unchanged whatever the value of Πi (Figure 3B). Therefore, the value of Πe ∼ 20 mN/m (Figure 3B) corresponds approximately to that of ∼20 mN/m for the MIP (Figure 2B). In the case of a = 0, the insertion of the protein into the monolayer is thus influenced only by the Πe of the monolayer. The value of a > 0 obtained for the binding of RP2 and recoverin onto DSPC (Figure 3A) and DDPC (Figure 3B), respectively, corresponds to the most favorable conditions for protein monolayer binding, which are further demonstrated by the fact that their MIP values are the only ones that are larger than the estimated membrane lateral pressure. It can also be seen that the Πe of RP2 and recoverin increases proportionally to Πi. Therefore, given eq 3, it can be concluded that protein insertion into these phospholipid monolayers is governed by Πi as well as by a positive synergy between the protein and the lipid monolayer that is related to a. Because a is related to the synergy between the protein and the lipid monolayer, we propose to use the name “synergy factor” to describe a. Analysis of the Synergy of the Binding of Proteins onto Phospholipid Monolayers. Figure 3 shows that synergy factor a allowed us to improve our understanding of protein binding onto lipid monolayers and to draw conclusions on the parameters governing these interactions. A more quantitative understanding Langmuir 2011, 27(4), 1373–1379

Calvez et al.

Article

Figure 4. (A) Synergy factor (a) of RP2 with DMPC in the LE and SC states and DSPC and DDPC monolayers. (B) Synergy factor (a) of recoverin with DMPC in the LE and SC states and DSPC and DDPC monolayers.

of protein monolayer binding can be achieved from Figure 4, which compares the value of synergy factor a for RP2 (Figure 4A) and recoverin (Figure 4B) for each phospholipid monolayer studied. It can be seen that RP2 binding onto the DSPC and DDPC monolayers is characterized by slopes of 0.45 (a > 0) and -0.24 (a < 0), respectively (Figure 4A). It can be concluded that DSPC largely favors whereas DDPC largely prevents RP2 binding, which is consistent with the large MIP of 36 mN/m of RP2 obtained with DSPC and the small value of ∼17 mN/m measured with DDPC. Moreover, a much larger value of a = 0.51 has been obtained for DMPC in the LC state compared to 0.21 for DMPC in the LE state. RP2 binding is thus highly favored by phospholipids in the liquid-condensed (DMPC (LC)) or solidcondensed (DSPC) states whereas phospholipids in a fluid state (DMPC (LE), DDPC) prevent its binding (Figure 4A). Synergy factor a thus provides additional information that was unavailable with the MIP. Indeed, a small, although significant, difference can be observed between the values of MIP for DMPC in the LE (25.5 ( 1.7) and LC (20.9 ( 1.3) states (Figure 2A) whereas large differences in the synergy factor are observed (Figure 4A). The MIP of RP2 in the presence of the DMPC (LC) monolayer (21.6 mN/m ( 1.3) is significatively different from that of the DSPC monolayer (36.4 mN/m ( 2.4) (Figure 2A). In contrast, similar synergy factors of a = 0.51 ( 0.03 and 0.45 ( 0.04 have been obtained with DMPC (LC) and DSPC (Figure 4A), respectively. Because of the difference of four carbons in the fatty acyl chain length between DMPC and DSPC, it can be postulated that the affinity of RP2 for the lipid monolayer is modulated by the physical state of these phospholipids whereas the MIP is governed by the phospholipid fatty acyl chain length. No phospholipid shows a negative contribution to the monolayer binding of recoverin (Figure 4B). Nevertheless, small values of a are obtained for saturared DMPC (LE or LC) and DSPC (Figure 4B), which are consistent with the small values of MIP obtained with these phospholipids. In addition, compared to DMPC and DSPC, a much larger value of a = 0.57 has been obtained for DDPC (Figure 4B), which thus highly favors recoverin monolayer binding and is also consistent with the largest value of MIP obtained for recoverin with this phosphoLangmuir 2011, 27(4), 1373–1379

Figure 5. (A) Kinetics of adsorption of recoverin in the presence of DDPC at different initial surface pressures: (b) 4, (9) 16, and (() 33 mN/m. (B) Kinetics of adsorption of recoverin in the presence of DMPC (LE) at different initial surface pressures: (b) 7, (9) 11, and (() 15 mN/m.

Figure 6. Rate of adsorption of recoverin in the presence of DDPC at different initial surface pressures. (Inset) Values of the rates of adsorption have been calculated by fitting the kinetics of adsorption to the equation of Pitcher.31

lipid (Figure 2B). Moreover, the kinetics of the monolayer binding of recoverin is also very particular in the presence of DDPC. Indeed, as can be seen in Figure 5A, the higher the initial surface pressure, the faster the kinetics of recoverin binding onto the DDPC monolayer. As shown in Figure 6, the rate of recoverin adsorption in the presence of DDPC increases exponentially from ∼2  10-4 to ∼37  10-4 s-1 when the initial surface pressure is increased from 4 to 33 mN/m (table in Figure 6). This contrasts with the data obtained with DMPC in the same physical state (LE) where the rate of adsorption of recoverin remains almost DOI: 10.1021/la104097n

1377

Article

Calvez et al.

Figure 7. Schematic representation of the different proposed limiting surface pressure of protein adsorption in monolayers on the basis of the value of a. A maximum insertion pressure is observed for a > 0, an exclusion pressure is observed for a < 0, and a stationary surface pressure is observed for a = 0.

unchanged with the increase in surface pressure (Figure 5B). This behavior is, however, particular to recoverin because no such effect has been observed for the other phospholipids and for RP2 with any of the lipids studied. Indeed, kinetics similar to those shown in Figure 5B have been obtained with RP2. This very particular behavior of recoverin might thus be promoted by its acylation.

Discussion Influence of Phospholipid Physical States and Unsaturation on the Binding of RP2 and Recoverin onto Phospholipid Monolayers. A large number of different lipids with varying fatty acyl chains are found in natural membranes. It is thus important to determine the effect of different lipids on the association of proteins with membranes. A useful approach to determining the extent of protein membrane binding and its preference for given lipids consists of measuring their maximum insertion pressure. The physical parameters modulating the maximum insertion pressure of two proteins have thus been dissected in detail, and a new approach was devised to analyze these data. Synergy factor a was thus found to provide new, additional information on the interaction taking place between saturated and polyunsaturated phospholipids with peripheral proteins RP2 and recoverin, respectively, which are known to bind membranes. Measurements were performed with phospholipids bearing the same polar headgroup (phosphatidylcholine) but with different saturated or polyunsaturated fatty acyl chains to determine selectively the effect of these fatty acyl chains on protein binding. Phospholipids with short (14:0, myristoyl) and long saturated (stearoyl, 18:0) and polyunsaturated (22:6, docosahexaenoyl) fatty acyl chains have thus been used to study RP2 and recoverin monolayer binding. The differences observed in the value of a should thus solely be due to hydrophobic interactions related to the nature of the phospholipid fatty acyl chains. Nevertheless, the physical state of the lipids must also be taken into account in rationalizing protein membrane binding. Indeed, the phase transitions of DMPC, DSPC, and DDPC are located at 23, 54.5, and -68 °C, respectively.47,48 The surface pressure isotherm of DMPC shows a phase transition between the LE and LC states at surface pressures that depend on temperature (Figure 1). For example, at 24 °C, a single LE state is observed whereas a phase transition is observed from ∼5.5 mN/m at 4 °C for DMPC (Figure 1). In contrast, the surface pressure isotherm of DSPC (47) Marsh, D. In CRC Handbook of Lipid Bilayers; CRC Press: Boca Raton, FL, 1990. (48) Kariel, N.; Davidson, E.; Keough, K. M. W. Biochim. Biophys. Acta 1991, 1062, 70–76.

1378 DOI: 10.1021/la104097n

and DDPC at room temperature shows a single SC and LE state, respectively (Figure 1). Contrasting observations have been obtained with the polyunsaturated DDPC monolayer where preferential binding is measured for recoverin and unfavorable binding is measured for RP2 (Figure 4). Conversely, preferential binding is observed for RP2 with saturated DSPC whereas this phospholipid (as well as DMPC) does not favor recoverin monolayer binding (Figure 4). This is also well illustrated when comparing the values of MIP shown in Figure 2. Indeed, MIP values much smaller than the estimated membrane lateral pressure have been obtained for recoverin and RP2 with DSPC and DDPC, respectively (Figure 2); conversely, much larger values than the estimated membrane lateral pressure have been obtained for recoverin and RP2 with DDPC and DSPC, respectively (Figure 2). The physical state of the lipids does not seem to influence recoverin binding because almost identical MIP values have been obtained with DMPC in the LE and LC states (as well as with DSPC) (Figure 2A). In contrast, a strong difference can be seen for the synergy factor of RP2 when binding to DMPC and DDPC in the LE state is compared to those when binding to DMPC in the LC state and DSPC in the SC state (Figure 4A). This is consistent with the previous observation that RP2 is associated with microdomains made of saturated phospholipids.33 Significance of Synergy Factor a on the Binding of Protein onto Phospholipid Monolayers. We have previously proposed to use the term maximum insertion pressure to describe the surface pressure up to which a protein can insert into the monolayer and beyond which no insertion takes place.9 However, on the basis of the present data, the terms to be used must be extended to take the present observations into account as illustrated in Figure 7. The MIP obtained when a > 0 would indeed correspond to the maximum insertion pressure because there is a positive interaction between the protein and the lipids. However, the MIP value obtained when a = 0 could be called the stationary surface pressure because it does not favor or disfavor protein monolayer binding. Finally, the MIP with a slope of a < 0 could be called the exclusion pressure because it corresponds to a repulsion between the protein and the monolayer. Parameters Influencing the Surface Pressure Increase (ΔΠ). ΔΠ can be related to both the extent of protein insertion in the monolayer and/or to the quantity of protein bound to the monolayer. The decrease in ΔΠ with the increase in Πi observed in Figure 2 could thus be due to either a progressive decrease in the number of proteins bound to the monolayer or a decrease in the extent of protein monolayer insertion, because less free area is available,49 or both. It has been previously shown for apolipoprotein A-IV that the relationship between ΔΠ and Πi depends directly on the quantity of protein at the air/water interface,50 which might also be the case for RP2 and recoverin. It is deemed necessary to work at a saturating protein concentration that allows us to compare the value of a of one protein with that of another protein and thus to generalize the use of a. Also, eq 4 showed that ΔΠ depends on Πi and on synergy factor a. This synergy factor provides information on the affinity of proteins for lipid monolayers as shown in Figures 3 and 4. However, the integration of eq 4 resulted in ΔΠ ¼ aΠi - Πi þ b

ð5Þ

(49) Sugar, I. P.; Mizuno, N. K.; Brockman, H. L. Biophys. J. 2005, 89, 3997– 4005. (50) Weinberg, R. B.; Ibdah, J. A.; Phillips, M. C. J. Biol. Chem. 1992, 267, 8977–8983.

Langmuir 2011, 27(4), 1373–1379

Calvez et al.

which shows that ΔΠ also depends on a constant b that can be determined from the plot of ΔΠ as a function of Πi. Indeed, when Πi = 0 mN/m, ΔΠ = b. Because b can be obtained at an initial surface pressure of 0 mN/m, it could be called ΔΠ0, which is described by the intercept of the curves with the y axis shown in Figures 2 and 3. The insertion of proteins into monolayers can be described by three different parameters: the maximum insertion pressure (the x intercept of the curves in Figure 2), synergy factor a (the slope of the curves in Figure 3), and ΔΠ0 (the y intercept of the curves in Figure 2), which is more difficult to describe. One could postulate that ΔΠ0 corresponds to the surface tension of the pure protein at the air/water interface in the absence of a phospholipid monolayer. However, Figure 2A shows that ΔΠ0 also depends on the nature of the phospholipid monolayer. In fact, values of ΔΠ0 ∼ 20 and 10 mN/m are respectively observed for RP2 with the DMPC (LE) and DMPC (LC) monolayers (Figure 2A) whereas the surface tension of pure RP2 is ∼10 mN/m (data not shown). Similarly, the surface tension of pure recoverin is ∼15 mN/m (data not shown) whereas ΔΠ0 with all phospholipid monolayers is ∼20 mN/m (Figure 2B). Therefore, ΔΠ0 is different from the surface tension of these pure proteins except for RP2 in the presence of DMPC (LC). One can thus conclude that ΔΠ0 is not solely governed by the surface tension of the protein. ΔΠ0 might thus be described as the trend of the monolayer of modifying the surface activity of the protein. How could the monolayer influence the surface activity of a protein, and why, in some cases, would the monolayer have no influence on this surface activity? To evaluate the significance of ΔΠ0 properly, one has to keep in mind that our data mostly provide information on the macroscopic level. It can be postulated that when ΔΠ0 is equal to the protein surface tension, protein would not be inserted more deeply than the polar headgroup of the phospholipid monolayer.

Langmuir 2011, 27(4), 1373–1379

Article

In contrast, when ΔΠ0 is larger than the protein surface tension, protein would insert within the phospholipid fatty acyl chains. Ellipsometric measurements could allow us to correlate the extent of protein insertion with ΔΠ0. The values of the synergy factor and the differences observed between ΔΠ0 and the surface tension of the pure proteins could allow us to highlight the role of the surface activity of proteins in their binding onto lipid monolayers. In fact, such binding is solely governed by the surface activity only when ΔΠ0 is equal to the surface tension of the protein and when a is equal to 0. However, this has not been observed in the present study. Conversely, when a is different from 0 and ΔΠ0 is different from the surface tension of the protein, protein binding should be governed by hydrophobic and electrostatic interactions in addition to protein surface activity. In conclusion, this study allowed us to clarify that, in addition to the maximum insertion pressure, the synergy factor and ΔΠ0 are also important parameters that provide very useful information and that selective interactions can be demonstrated between proteins and lipids using the monolayer model membrane system. Acknowledgment. This work was supported by the Natural Sciences and Engineering Research Council of Canada. E.D. is the recipient of a studentship from the Canadian Institutes of Health Research (CIHR) and the E. A. Baker Foundation and a travel fellowship from the Reseau FRSQ de Recherche en Sante du Quebec. Supporting Information Available: Evolution of the grey level of BAM for the Π-A isotherm of DSPC and after spreading of the DSPC monolayer. BAM images obtained during the compression and spreading of the DSPC monolayer. This material is available free of charge via the Internet at http://pubs.acs.org.

DOI: 10.1021/la104097n

1379