Letter pubs.acs.org/ac
Cholesterol Rich Domains Identified in Unilamellar Supported Biomimetic Membranes via Nano-Viscosity Measurements Imad Younus Hasan and Adam Mechler* Department of Chemistry and Physics, La Trobe Institute for Molecular Science, La Trobe University, Melbourne, Victoria 3086, Australia ABSTRACT: Understanding the distribution of cholesterol in phospholipid membranes is of key importance in membrane biophysics, primarily since cholesterol enriched regions, rafts, are known to play a special role in protein function. In this work, quartz crystal microbalance with dissipation (QCM)based viscosity measurements were used to study cholesterol-induced domain formation in partially suspended single bilayer membranes. 1,2-Dimyristoyl-snglycero-3-phosphocholine (DMPC) and its mixtures with different amounts of cholesterol were studied. QCM temperature ramping experiments identified domains of different phase transition temperatures in the mixed membranes. The phase transition of DMPC shifted from 23.4 °C toward lower temperatures with increasing cholesterol content. A second, continuous but much broader, transition peak has been observed for the DMPC: cholesterol mixtures suggest that a separate cholesterol rich domain coexists with the DMPC rich domain. Importantly, the sharp DMC phase transition peak gradually diminished and eventually disappeared over 15% cholesterol content, suggesting that the cholesterol rich domain has a definite stoichiometry and once this cholesterol concentration is reached the DMPC-rich domain disappears. DSC control experiments do not show the second domain, suggesting that the phase separation only happens in nontensioned (flat) membranes.
C
chain melting of 1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC) and their mixtures with 1,2-dimyristoyl-sn-glycero-3phospho-rac-1-glycerol (DMPG). We noted that, unlike neat DMPC and DMPC/DMPG mixtures, the DMPC/cholesterol mixtures appeared to give unclear phase transitions. In this work, we revisit this phenomenon. We demonstrate that the viscosity profiles reveal the coexistence of domains of different cholesterol content. From the phase transition temperatures, we estimate the cholesterol content of these domains. Liposome suspensions were created by a method described before16,17 that produces mostly unilamellar vesicles of a broad size distribution. 50 μM lipid suspensions were prepared in 20 mM phosphate buffer containing 100 mM sodium chloride. Differential scanning calorimetry (DSC) has been used as a reference method to measure the phase transition temperatures of DMPC/cholesterol mixtures in liposome suspensions. A SETARAM μDSC evo3 instrument was used with 0.7 mL hastelloy batch cells. In each DSC experiment, three zones were recorded with the same parameters: the system was equilibrated at 15 °C for 30 min; then, the temperature was increased at a rate of 0.33 °C min−1 to 30 °C. Next, the system was equilibrated for 30 min at 35 °C, and then, temperature was returned to 15 °C at a rate of 0.33 °C min−1 followed by a 30 min equilibration. Hence, the phase transition temperature was
holesterol-enriched regions of plasma membranes, also known as lipid rafts, attract substantial attention for their different affinities toward integral and peripheral membrane proteins1 and hence play an important role in membrane trafficking and signaling.2 It is believed that they exert exceptional physiological function due to their distinct physicochemical properties.3 In these domains, the presence of cholesterol imposes increased order on the saturated acyl chains of phospholipids (and sphingolipids) forming less fluid domains than the surrounding plasma membrane. The cholesterol-rich region, or raft, is also known as a liquid ordered (Lo) domain.4 Rafts are frequently discussed in terms of different thermodynamic phases. Coexistence of domains of different thermodynamic phases has been observed in both biological5 and biomimetic membranes,6,7 primarily with microscopic methods,8−10 although attempts were also made to characterize the composition and physicochemical properties by using highresolution secondary ion mass spectrometry,11 ellipsometry,12 and fluorescence.13,14 However, due to the dimensions of the systema single bilayer membrane is ∼5 nm thick and domain sizes vary from a few nanometers to tens of micrometers10,14,15the composition and the thermodynamic phase of the domains could not be identified unequivocally. In a previous work,16 we demonstrated the ability of the quartz crystal microbalance “with dissipation” (QCM) method to measure phase transition temperatures of partially suspended single bilayer membranes from the viscoelastic changes upon © XXXX American Chemical Society
Received: March 16, 2016 Accepted: May 3, 2016
A
DOI: 10.1021/acs.analchem.6b01045 Anal. Chem. XXXX, XXX, XXX−XXX
Letter
Analytical Chemistry measured six times in each experiment. As shown in Figure 1, neat DMPC gives a sharp phase transition peak with an onset at
Figure 1. Differential scanning calorimetry measurements for DMPC and its mixtures with cholesterol where (a) neat DMPC, (b) DMPC/ 3% cholesterol, (c) DMPC/5% cholesterol, (d) DMPC/10% cholesterol, (e) DMPC/15% cholesterol, and (f) DMPC/20% cholesterol.
Figure 2. Effect of temperature ramping on the f and D signals in the case of a neat DMPC bilayer. The gray line indicates the temperature profile; the blue line is Δf; the red line refers to ΔD. (A) sensograms of Δf and ΔD of liposome deposition against time for neat DMPC on MPA-modified gold at 25 °C. (B) The applied temperature sweep profile. (C, D) Δf and ΔD for neat DMPC and the reference channel, respectively, upon temperature sweep. (E) Corrected Δf and ΔD for neat DMPC after subtraction of the reference. (F) First-order time derivatives of Δf and ΔD from (E).
23.65 ± 0.03 °C and the position of this peak shifts to lower temperatures as a function of the cholesterol content to reach ∼21.50 ± 0.19 °C at 15 mol % of cholesterol (Figure 1e). Furthermore, the integral area of the peak, that is, the heat of phase transition, also decreases with increasing cholesterol content until the peak diminishes at 15 mol % of cholesterol. For the QCM experiments, liposome deposition method18−21 was used to form supported unilamellar bilayer membranes on a gold surface modified with 3-mercaptopropionic acid (MPA). MPA is known to promote liposome fusion by increasing the strength of the vesicle−surface interaction.22 QCM was used to monitor lipid deposition on the surface, where the changes in oscillation frequency (Δf) and energy dissipation (ΔD) were recorded to characterize the resulting membranes.18 Experiments were repeated at least three times. Figure 2A shows a typical deposition sensogram (neat DMPC). A few seconds after injecting the liposomes into the measurement chamber, a negative frequency change (blue solid line) was observed and simultaneously the dissipation (red solid line) increased. Both signals stabilized at −16 Hz and ∼2.7 arb. u. respectively, after a phosphate buffer wash at ∼20 min. Almost the same sensograms have been observed for depositing DMPC/cholesterol mixtures. Table 1 shows a summary of the final values of frequency and dissipation changes for each mixture. All depositions were performed at temperatures above the main phase transition temperature for either of the lipid mixtures in order to promote liposome rupture and fusion on the surface and reduce the amount of intact liposomes.20,21 The methodology employed to analyze the results rests on a previous work.16 Temperature sweeping was performed in QCM by a protocol mirroring the parameters of the DSC experiments (Figure 2B). Δf and ΔD sensograms that were measured upon temperature sweeping for a bilayer membrane and the blank reference (MPA modified gold surface in buffer solution) are shown in Figure 2C,D, respectively. Both
Table 1. Summary of the Final Values of Frequency and Dissipation Change for the Deposition of Lipid Mixtures on MPA Modified Golda MPA-modified gold
a
lipids mixtures
temp °C
Δf [Hz]
sd
ΔD [arb. u.]
sd
DMPC DMPC/3% Cho DMPC/5% Cho DMPC/10% Cho DMPC/15% Cho
25 25 25 25 25
−16 −22 −18.6 −24.5 −21
2.18 4.56 4.13 3.85 2.62
2.7 2.5 3.1 4.8 4.3
0.4 0.5 0.6 0.6 0.2
All values reported for the 7th eigenmode (35 MHz).
frequency and dissipation signals were constant at 20 °C, and these values were chosen as arbitrary zero. In both cases, a monotonous frequency increase and dissipation decrease was observed with increasing temperature. However, in the presence of the membrane, the changes in both Δf and ΔD were higher than the blank reference and also there was marked deviation from linearity in both f and D signals at ∼23 °C. After subtracting the respective blank curves from frequency and dissipation, the trend became clear (Figure 2. E). The temperature ramping started after 30 min of equilibration; initially, there were no significant changes until ∼23 °C where a steep change started in both QCM parameters, which gradually slowed down after ∼25 °C to reach constant values when the temperature reached the maximal value at 35 °C. The negative temperature sweep mirrored the trend of the positive sweep for both parameters, and the second repeat showed the same trend. Following the previously established method, the temperature at which the frequency and dissipation changes are the steepest corresponds to the phase B
DOI: 10.1021/acs.analchem.6b01045 Anal. Chem. XXXX, XXX, XXX−XXX
Letter
Analytical Chemistry transition temperature. This was identified by plotting the first derivative of frequency (dF/dT) and dissipation (dD/dT) against time (that corresponds to temperature)16,19,23 as shown in Figure 2F. Importantly, while the conceptually correct way to identify the viscoelastic changes is by using the dissipation signal, the frequency and dissipation are interdependent in a viscoelastic system;24 hence, here the frequency change is also shown for comparison. For neat DMPC, well-defined peaks are seen at ∼23.4 °C in a good agreement with the DSC experiments. It should be mentioned that repeating the same sweeping profile has shown a shift in peak position to 22.6 °C likely due to an increase in the membrane−surface contact area in the fluid phase that was preserved in the gel phase, affecting the consecutive sweeps. Temperature sweeping experiments were performed on a range of DMPC/cholesterol mixtures (Figure 3). For 3%
the trend of the positive sweep with a slight hysteresis. Nearly identical sensograms have been observed for 5% and 10% cholesterol content (Figure 3C,E) with the exception that the steep changes start at lower temperatures. On the other hand, the steep change diminished for 15% cholesterol content and a nearly linear trend is observed in both frequency and dissipation signals (Figure 3G). The first derivatives of the signals characteristically differ from neat DMPC (Figure 3, right panels). For 3%, 5%, and 10% cholesterol (Figure 3B,D,F), well-defined asymmetric peaks exist that have a sharp edge at 20−23 °C and a long tailing edge toward higher temperatures. The position of the sharp peak shifted from 22.55 °C for 3% cholesterol to 22.1 and 20.1 °C for 5% and 10% mixtures, respectively. For 15% cholesterol content, however, only dD/dT shows a broad peak extending to the entire temperature sweep: a continuous, linear viscosity change without a well-defined phase transition (Figure 2H). The asymmetry of the peaks suggests that there are two processes affecting viscosity: a sharp phase transition as in neat DMPC and a slow change resulting in a broad peak in the derivative. Hence, Gaussian fitting was applied to analyze dF/ dT and dD/dT results. Figure 4 shows the analysis of 5%
Figure 4. Gaussian fitting of dF/dT and dD/dT results for DMPC/5% cholesterol identifies two domains of different viscosity.
cholesterol membrane. When the temperature starts increasing (Figure 4A,B), dD/dT (and dF/dT) gives a sharp peak at 22.17 °C. This peak is consistent with the phase transition of DMPC from the gel (Lβ) to the liquid-crystalline (Lα) phase, commonly called chain melting or main transition; albeit, it appears at somewhat lower temperature.16 This sharp peak coexists with a very broad peak (over 4.8 °C fwhm for dD/dT). Hence, two domains can be identified with characteristically different phase behavior. In contrast to QCM results, DSC measurements showed that the main effect of the incorporation of increasing amounts of cholesterol into lipid bilayers was to broaden the phase transition peak and reduce the enthalpy and the temperature of the Lβ → Lα phase transition, in a good agreement with literature reports.25,26 Hence, the QCM results highlight a different aspect of the membrane structure as reflected in the viscoelastic properties. The transition of neat DMPC from the gel (Lβ) to the liquid-crystalline (Lα) state happens at a welldefined temperature (sharp peak); however, 3 mol % cholesterol is enough to introduce another continuous phase
Figure 3. Left panels show the frequency (blue line) and dissipation change (red line) upon temperature sweeping for a single bilayer membrane of DMPC with 3, 5, 10, 15, and 20 mol % cholesterol content. Right panels show the first-order time derivatives.
cholesterol content, initially, ΔD shows only a small decrease (Figure 3A red line), parallel with a similarly small increase in Δf, until the temperature reaches ∼22 °C where a steep change starts in both signals that is the fastest at ∼23 °C; after that, the change becomes gradually less steep until the sweep reaches its maximum at 30 °C. The negative temperature sweep mirrors C
DOI: 10.1021/acs.analchem.6b01045 Anal. Chem. XXXX, XXX, XXX−XXX
Letter
Analytical Chemistry change that overlaps with the sharp Lα → Lβ transition that appears as a result of the restrained mobility of the aliphatic chains of the lipid molecules in cholesterol rich domains of the membrane. This restrained mobility prohibits chain melting; hence, instead of a primary phase transition, the slow change in viscosity over a temperature range is consistent with a “glassy” transition of the cholesterol-rich domains. The difference between DSC and QCM results may be explained with the curvature tension of the liposome suspensions that promote even cholesterol distribution, or else the role of surface contact in domain separation in the supported membranes. Given that domain separation was suggested before for bulk lamellar phase of phospholipid−cholesterol mixtures,26,27 the former is more likely. The presence and the size of these domains depend strongly on the concentration of cholesterol (and the temperature). When the mol % of cholesterol is increased up to 10%, Lα → Lβ transition is still observed but is less sharp than at lower cholesterol percentages, suggesting an increase in the cholesterol rich domains at the expense of phospholipid rich domains. However, no clear phase transition is detected at cholesterol content over 15 mol %; thus, only the Lo phase exists at all temperatures. These results suggest that in a single bilayer membrane cholesterol forms an optimal mixture with DMPC at ∼15 mol % (Figure 5) as opposed to measurements
DMPC domains that largely exclude cholesterol, analogous to the exclusion of salts from frozen water; the exclusion is limited by the miscibility of cholesterol and DMPC that is found to be ∼15%. However, since the gel phase lipid is still a dynamic system, an equilibrium cholesterol distribution exists between the two domains. In the presence of excess cholesterol (in our case, over 15%), both the main and the glass transitions are lost as the system becomes nonideally mixed, essentially an emulsion. In summary, in a flat membrane below the main transition temperature of the main constituent phospholipid, cholesterol is found mostly in domains of a specific, ∼15 mol % composition, revealed by nanoviscosity measurements. Domain separation was not detected in DSC measurements; hence, it is feasible to assume that in liposomes the curvature tension enforces even distribution of cholesterol.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. Author Contributions
The manuscript was written through contributions of both authors. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS I.Y.H. expresses his special thanks to the Higher Committee for Education Development in Iraq (HCED) for supporting this work.
■
REFERENCES
(1) Pike, L. J. J. Lipid Res. 2006, 47, 1597−1598. (2) Simons, K.; Ikonen, E. Nature 1997, 387, 569−572. (3) Gandhavadi, M.; Allende, D.; Vidal, A.; Simon, S.; McIntosh, T. Biophys. J. 2002, 82, 1469−1482. (4) Samsonov, A. V.; Mihalyov, I.; Cohen, F. S. Biophys. J. 2001, 81, 1486−1500. (5) Pralle, A.; Keller, P.; Florin, E.-L.; Simons, K.; Hörber, J. J. Cell Biol. 2000, 148, 997−1008. (6) Veatch, S. L.; Keller, S. L. Biophys. J. 2003, 85, 3074−3083. (7) Marques, J. T.; de Almeida, R. F. M.; Viana, A. S. Electrochim. Acta 2014, 126, 139−150. (8) de Almeida, R. F. M.; Borst, J.; Fedorov, A.; Prieto, M.; Visser, A. Biophys. J. 2007, 93, 539−553. (9) Johnston, L. J. Langmuir 2007, 23, 5886−5895. (10) Yuan, C. B.; Furlong, J.; Burgos, P.; Johnston, L. J. Biophys. J. 2002, 82, 2526−2535. (11) Kraft, M. L.; Weber, P. K.; Longo, M. L.; Hutcheon, I. D.; Boxer, S. G. Science 2006, 313, 1948−1951. (12) Howland, M. C.; Szmodis, A. W.; Sanii, B.; Parikh, A. N. Biophys. J. 2007, 92, 1306−1317. (13) Stottrup, B. L.; Heussler, A. M.; Bibelnieks, T. A. J. Phys. Chem. B 2007, 111, 11091−11094. (14) de Almeida, R. F. M.; Loura, L. M. S.; Fedorov, A.; Prieto, M. J. Mol. Biol. 2005, 346, 1109−1120. (15) Lewis, B. A.; Engelman, D. M. J. Mol. Biol. 1983, 166, 211−217. (16) Hasan, I. Y.; Mechler, A. Soft Matter 2015, 11, 5571−5579. (17) Mechler, A.; Praporski, S.; Piantavigna, S.; Heaton, S. M.; Hall, K. N.; Aguilar, M. I.; Martin, L. L. Biomaterials 2009, 30, 682−689. (18) Richter, R.; Mukhopadhyay, A.; Brisson, A. Biophys. J. 2003, 85, 3035−3047. (19) Jing, Y.; Trefna, H.; Persson, M.; Kasemo, B.; Svedhem, S. Soft Matter 2014, 10, 187−195.
Figure 5. Schematics of the phase behavior of the DMPC/cholesterol single bilayer membrane. At temperatures lower than the main transition temperature (Tm), two coexistent phases are observed: a gel (Lβ) for DMPC rich domains and a liquid ordered (Lo) state for cholesterol rich domains. At Tm, the melting of the hydrocarbon chains yields a sharp peak in the time derivatives of the dissipation signal (yellow region). A broad peak (blue region) reveals the presence of cholesterol rich domains.
on bulk lamellar lipids where it was indicated at 25%.25−28 The difference could be related to the principle of measurement. For detection in QCM, it is necessary that domains exhibit continuum behavior as viscosity is a continuum property. Continuum behavior might be lost for very small domains that could still be identified on the basis of calorimetric measurements. However, given that our DSC results on liposome suspensions did not show any domain separation in these mixtures, it is more likely that bulk lamellar lipids exhibit different phase behavior from single bilayer or vesicular lipids. Consistently, AFM imaging and force spectroscopy results on DPPC/cholesterol membrane29 have provided evidence that, contrary to the phase diagram of the bulk lamellar phase,27 in single bilayer membranes domain separation exists only below the phase transition temperature. Hence, it is feasible to assume that the domain separation is driven by the lower free energy of the directly interacting aliphatic chains in the gel phase of neat D
DOI: 10.1021/acs.analchem.6b01045 Anal. Chem. XXXX, XXX, XXX−XXX
Letter
Analytical Chemistry (20) Reimhult, E.; Hook, F.; Kasemo, B. Phys. Rev. E: Stat. Phys., Plasmas, Fluids, Relat. Interdiscip. Top. 2002, 66, 051905. (21) Reimhult, E.; Hook, F.; Kasemo, B. Langmuir 2003, 19, 1681− 1691. (22) Steinem, C.; Janshoff, A.; Ulrich, W. P.; Sieber, M.; Galla, H. J. Biochim. Biophys. Acta, Biomembr. 1996, 1279, 169−180. (23) Wargenau, A.; Tufenkji, N. Anal. Chem. 2014, 86, 8017−8020. (24) Voinova, M. V.; Rodahl, M.; Jonson, M.; Kasemo, B. Phys. Scr. 1999, 59, 391−396. (25) Benesch, M. G.; Lewis, R. N.; Mannock, D. A.; McElhaney, R. N. Chem. Phys. Lipids 2015, 188, 10−26. (26) Mcmullen, T. P. W.; Mcelhaney, R. N. Biochim. Biophys. Acta, Biomembr. 1995, 1234, 90−98. (27) Vist, M. R.; Davis, J. H. Biochemistry 1990, 29, 451−464. (28) Mcmullen, T. P. W.; Lewis, R. N. A. H.; Mcelhaney, R. N. Biophys. J. 1994, 66, 741−752. (29) Redondo-Morata, L.; Giannotti, M. I.; Sanz, F. Langmuir 2012, 28, 12851−12860.
E
DOI: 10.1021/acs.analchem.6b01045 Anal. Chem. XXXX, XXX, XXX−XXX
