Lipid Phase Control and Secondary Structure of Viral Fusion Peptides

Aug 22, 2017 - Lipid Phase Control and Secondary Structure of Viral Fusion Peptides Anchored in Monoolein Membranes. Artem Levin, Christoph Jeworrek, ...
0 downloads 9 Views 3MB Size
Article pubs.acs.org/JPCB

Lipid Phase Control and Secondary Structure of Viral Fusion Peptides Anchored in Monoolein Membranes Artem Levin, Christoph Jeworrek, Roland Winter, Katrin Weise, and Claus Czeslik* Department of Chemistry and Chemical Biology, TU Dortmund University, D-44221 Dortmund, Germany S Supporting Information *

ABSTRACT: The fusion of lipid membranes involves major changes of the membrane curvatures and is mediated by fusion proteins that bind to the lipid membranes. For a better understanding of the way fusion proteins steer this process, we have studied the interaction of two different viral fusion peptides, HA2-FP and TBEV-FP, with monoolein mesophases as a function of temperature and pressure at limited hydration. The fusion peptides are derived from the influenza virus hemagglutinin fusion protein (HA2-FP) and from the tick-borne encephalitis virus envelope glycoprotein E (TBEV-FP). By using synchrotron X-ray diffraction, the changes of the monoolein phase behavior upon binding the peptides have been determined and the concomitant secondary structures of the peptides have been analyzed by FTIR spectroscopy. As main results we have found that the fusion peptides interact differently with monoolein and change the pressure and temperature dependent lipid phase behavior to different extents. However, they both destabilize the fluid lamellar phase and favor phases with negative curvature, i.e. inverse bicontinuous cubic and inverse hexagonal phases. These peptide-induced phase changes can partially be reversed by the application of high pressure, demonstrating that the promotion of negative curvature is achieved by a less dense packing of the monoolein membranes by the fusion peptides. Pressure jumps across the cubic-lamellar phase transition reveal that HA2-FP has a negligible effect on the rates of the cubic and the lamellar phase formation. Interestingly, the secondary structures of the fusion peptides appear unaffected by monoolein fluid−fluid phase transitions, suggesting that the fusion peptides are the structure dominant species in the fusion process of lipid membranes.

1. INTRODUCTION When enveloped viruses attack cells, lipid membrane fusion is an essential step that is mediated by viral fusion proteins, which are found on the virion surface.1−4 Fusion proteins become fusogenic by a trigger. Then, a hydrophobic segment which is known as a “fusion peptide” or “fusion loop” is exposed and inserted into the target cell membrane, thereby forming a bridging intermediate. In the subsequent fusion process, the fusion protein refolds and pulls both membranes together. When the opposing (cis) lipid monolayers are already merged, but the distal (trans) lipid monolayers are still separated, a hemifusion stalk is formed, which finally evolves into a full fusion pore (Figure 1). Hemifusion stalks and fusion pores are characterized by a high curvature of the membrane monolayers. At the waist, a monolayer is curved like a saddle surface where the two principle curvatures, c1 and c2, have opposite signs. In this way, the mean curvature, H = (c1 + c2)/2, is small, whereas the Gaussian curvature, K = c1·c2, is large and negative. Lipid mesophases with these properties are inverse bicontinuous cubic phases, QII.5 The structural analogy between QII phase formation and the fusion pathway has been pointed out in the literature and used to estimate the energetics underlying fusion pore formation.6−9 Indeed, QII phases can be viewed as ordered arrays of hemifusion intermediates. Moreover, there is a series © 2017 American Chemical Society

of examples, where inverse cubic phases and membrane curvature are induced by the addition of fusion peptides to lipid membranes.6,10−13 In these studies, only the fusion peptide, i.e. the short hydrophobic segment of the whole fusion protein that is inserted into the target membrane, has been used. It is suggested that a decreased activation free energy of the inverse cubic phases results in an increased rate of hemifusion stalk and fusion pore formation. Thus, the perturbation of lipid membranes by fusion peptides is probably needed to assist viral membrane fusion but is not sufficient for this process.14−16 Whereas many studies deal with the temperature dependent phase behavior of lipids with added fusion peptides, little is known about the concomitant structures of the peptides. In particular, it is rather unclear if the structures change on passing through lipid phase transitions, such as the lamellar-cubic transition. Furthermore, the molecular origin of lipid membrane curvature promotion by fusion peptides is not fully understood. To address these issues, we have studied the temperature and pressure dependent phase behavior of monoolein (MO) at limited hydration in the absence and the Received: June 29, 2017 Revised: August 18, 2017 Published: August 22, 2017 8492

DOI: 10.1021/acs.jpcb.7b06400 J. Phys. Chem. B 2017, 121, 8492−8502

Article

The Journal of Physical Chemistry B

Figure 1. Schematic drawing of the fusion pore formation mediated by viral fusion proteins. Fusion proteins are triggered (A), fusion peptides or fusion loops, FP, are inserted into the target membrane (B), and a hemifusion stalk (C) and a fusion pore (D) are formed.

2. MATERIALS AND METHODS The fusion peptides, GLFGAIAGFIENGWEGMIDGGCGKKKK (HA2-FP) and DRGWGNHCGLFGKGGCGKKKK (TBEV-FP), were obtained from the German Cancer Research Center and Centic Biotec, Heidelberg, Germany. They are synthesized with a free C-terminus and an unprotected N-terminus. Peptide purities were checked by mass spectrometry and were ≥90% for all samples. According to the manufacturer, they are free of TFA. Monoolein (MO, 1oleoyl-rac-glycerol, cat. No. M7765), and D2O were obtained from Sigma-Aldrich. Tris(hydroxymethyl)-aminomethane (Tris) and chloroform were obtained from Merck Millipore. For the SAXS experiments, 10 mM Tris solution in H2O at pH = 5.0 was used. Samples were prepared by adding the peptide to MO, followed by the addition of the buffer solution (17 wt %). For the FTIR experiments, the peptide-MO mixture was dissolved in D2O and lyophilized for H/D exchange two times before adding the buffer solution. The buffer solution was prepared with D2O, and the pD-value was adjusted to 5.4 (pD = pH-meter reading +0.4).34 Finally, all samples were subjected to extensive freeze−thaw cycles and stored at −25 °C until use. The SAXS experiments were performed at the high brilliance beamline ID02 at the European Synchrotron Radiation Facility (ESRF) in Grenoble, France. The wavelength was about 1.0 Å. Sample exposure times were chosen in the range 0.05−2 s depending on the intensity of the observed Bragg reflections. The scattering data are scaled as a function of wavevector transfer, q = (4π/λ) sin θ, where λ is the wavelength and θ is half of the scattering angle. The Bragg reflections of silver behenate, which are found at q = 2πn/58.38 Å with n = 1, 2, 3, ..., were used for the calibration of the detector.35 The scattering of water was used as background that was subtracted from all data. Samples were contained in a home-built highpressure cell, which is already described in the literature.36 Briefly, it is designed for a maximum operating pressure of 4000 bar and is temperature-controlled by circulating water flow. It is made of Nimonic 90 alloy and has flat diamond windows of 0.7 mm thickness. The sample is separated from the pressurizing medium by using a sample holder consisting of a Teflon ring which is sealed with Mylar films on both sides. Pressure− temperature phase diagrams were constructed by collecting numerous SAXS images with increasing pressure in steps of 100 bar at constant temperature. Pressure jumps with variable amplitudes were achieved within 5 ms using a home-built highpressure jump apparatus.37 Three SAXS images were collected before triggering the pressure jump, followed by 30 images every 150 ms and further images every 1 s. The FTIR experiments were carried out using the Nicolet 6700 Fourier transform infrared spectrometer from Thermo Fisher Scientific operating with a liquid nitrogen-cooled MCT detector. The whole spectrometer was continuously purged with dry air. Pressure dependent measurements at constant

presence of two different fusion peptides. MO is highly suitable for these experiments, because it forms virtually all different kinds of lipid phases in water, such as the lamellar Lα-phase, inverse cubic QII-phases with space groups Ia3d and Pn3m, and the inverse hexagonal HII-phase. The MO−water phase diagram has already been determined as a function of water fraction, temperature ,and pressure.17−23 With increasing temperature, the general phase sequence Lα, QII, and HII can be observed, whereas the reverse sequence applies with increasing pressure, because high pressure favors structures with high packing densities.18,24 Although MO is not found in biological membranes, it has the advantage of a rich phase behavior. Considering that all lipids, including MO, share the general phase sequence as a function of temperature and pressure, the results of this study are relevant for biological membranes. Thus, the high pressure experiments carried out in this study can reveal the contribution of packing effects to the lipid phase control of fusion peptides. The fusion peptides used in this study are chosen from class I and class II viral fusion proteins, which are defined according to their structure.2 The used fusion peptide from the influenza virus hemagglutinin fusion protein (class I), denoted as HA2FP, has the sequence GLFGAIAGFIENGWEGMIDG.25 The used fusion loop of the tick-borne encephalitis virus envelope glycoprotein E (class II), denoted as TBEV-FP, has the sequence DRGWGNHCGLFGKG.26,27 In both cases, we have added the sequence GCGKKKK to increase the solubility, as suggested in the literature.25,28 The secondary structure of HA2-FP has been investigated in great detail, and a significant fraction of helices has been reported in the presence of lipids.25,29−31 The fusion peptides were analyzed at pH = 5, since they stem from viruses that enter cells by endocytosis; that is, they induce fusion in the low pH environment of endosomes. In this study, the phase diagrams of MO−water in the absence and presence of HA2-FP or TBEV-FP are recorded applying small-angle X-ray scattering (SAXS) as a function of temperature and pressure at limited hydration. SAXS reveals the lipid phase identity by characteristic patterns of Bragg reflections.5,32,33 The secondary structures of the fusion peptides incorporated in MO−water membranes have been determined by Fourier-transform infrared (FTIR) spectroscopy utilizing the diamond anvil technique. As we will show in this study, the fusion peptides induce inverse nonlamellar phases in MO−water mixtures by a less dense packing of the membranes. Furthermore, we observe that the secondary structures of HA2FP and TBEV-FP are not controlled by the MO phases, because they do not change upon lipid fluid−fluid phase transitions. 8493

DOI: 10.1021/acs.jpcb.7b06400 J. Phys. Chem. B 2017, 121, 8492−8502

Article

The Journal of Physical Chemistry B

Figure 2. Selected SAXS intensities, I, as a function of wavevector transfer, q, and pressure, p. Left, MO with 17 wt % water at 60 °C shows a pressure-induced transition from the cubic QII(Ia3d) to the lamellar Lα phase. Right, MO with 2 wt % TBEV-FP and 17 wt % water at 6 °C shows pressure-induced transitions from the hexagonal HII to the cubic QII(Ia3d) and to a lamellar Lx phase.

mm-thick sample is pressurized within about 1 μs, whereas temperature jumps of 10−30 K can be induced by iodine pulsed lasers within nanoseconds in small aqueous samples.) Therefore, we have studied the QII(Ia3d) phase boundaries of MO at limited hydration as a function of temperature and pressure, and we analyze how these are affected by the addition of the two different fusion peptides. It must be noted, however, that lipid samples at limited hydration are not easy to work with. An error of about ±5 wt % water content is estimated in this study. 3.1. Effects of fusion peptides on the p,T-phase behavior of MO at limited hydration. SAXS is a highly useful tool to identify the structure of lipid mesophases. Each phase is clearly characterized by a pattern of a few Bragg reflections.5,33 In Figure 2, selected SAXS images are shown, which were obtained with the MO samples of this study. For example, without fusion peptide, we can identify the QII(Ia3d) phase of MO at 60 °C and low pressures, which exhibits Bragg reflections at q = (2π/acub)·(h2 + k2 + l2)1/2, where acub is the cubic lattice constant and h, k, l are the Miller indices (Figure 2). Some Bragg reflections of Ia3d structures are extinguished, leaving q-ratios of √6:√8:√14:√16:√20:√22:√24:√26.... From these data, we can calculate a pressure-independent lattice constant of acub = 104 Å for MO. At 1.15−1.35 kbar, we observe a pressure-induced phase transition to the Lα phase, which has Bragg peaks at q = (2π/alam)·n, where alam is the lamellar spacing and n is the order of the reflection, equal to 1 and 2 in this case. With increasing pressure, the lattice constant is found to slightly increase from alam = 41 Å at 1.2 kbar to 42 Å at 2.0 kbar. These observations are consistent with the effect of pressure on the lipid chain conformation and packing. Pressure is reducing the number of gauche conformations, which reduces the chain volume and the negative curvature of a monolayer. Thus, the lamellar phase is favored over the inverse cubic phase at higher pressures. Within the Lα phase, pressure enhances the dense packing of the lipid chains even further, resulting in a higher number of trans conformations and thicker monolayers. In the presence of 2 wt % TBEV-FP, the phase behavior of MO at limited hydration is changed drastically (Figure 2). We observe an HII phase at low pressures with Bragg peaks at q = (4π/31/2ahex)·(h2 + k2 + hk)1/2, where ahex is the hexagonal lattice constant and h, k are the Miller indices. Thus, the first Bragg peaks exhibit q-ratios of 1:√3:√4. For example, a lattice

temperature were performed using the P-series diamond anvil cell from High Pressure Diamond Optics (Tucson, AZ, USA). It has two type IIa diamonds with a surface diameter of 0.6 mm. They are separated by a 20 μm thick steel gasket with a central hole of 0.5 mm diameter containing the sample solution (about 4 nL). The temperature of the diamond anvil cell was set by a circulating water flow. α-Quartz was added to the sample as an internal pressure sensor. The shift of the 695 cm−1 band, Δν̃, to higher wavenumbers is proportional to the applied pressure, p, according to p/kbar = 1.2062 Δν̃/cm−1 + 0.015054 Δν̃2/ cm−2.38 Spectral analysis was carried out using the Grams/AI 8.1 software from Thermo Fisher Scientific. The diamond absorbance was fitted in the range 2000−1500 cm−1 with a polynomial, as shown in Figure S1, and subtracted as background from all FTIR spectra.

3. RESULTS AND DISCUSSION As outlined in the Introduction, MO at limited hydration forms many of the lamellar and nonlamellar phases known from various lipid systems. In particular, there is a broad phase region of the inverse cubic QII phase with Ia3d space group as a function of temperature and hydration.17,19 It is noted that MO in excess water only forms the QII(Pn3m) phase from 0−95 °C.18 Even under pressures up to 1500 bar, no phase transition occurs with the exception of a lamellar crystalline phase below 10 °C.18 Thus, no fluid lamellar-cubic phase transition of MO can be studied in excess water. In this study, a water content of 17 wt % has been chosen which locates the Lα−QII(Ia3d) phase transition at about 30 °C and the QII(Ia3d)−HII phase transition at about 90 °C.19 Of course, according to Gibbs’s phase rule, these phase transitions are ranging over short temperature intervals, where phase coexistence is observed, because we are dealing with a two-component system. Pressure is used as a further parameter in this study, because temperature-induced lipid phase transitions are generally reversed upon pressurization. Moreover, application of pressure has the advantage to change only the density of the system without simultaneously changing the thermal energy, in contrast to temperature dependent measurements. Also, pressure propagates rapidly, preventing any sample inhomogeneity in kinetic experiments, and pressure jumps can be carried out bidirectionally with positive or negative amplitude. (The speed of sound in water is about 1500 m s−1, meaning that a 18494

DOI: 10.1021/acs.jpcb.7b06400 J. Phys. Chem. B 2017, 121, 8492−8502

Article

The Journal of Physical Chemistry B

Figure 3. Tentative phase diagrams of MO, MO with 2 wt % HA2-FP, and MO with 2 wt % TBEV-FP at limited hydration (17 wt %) as a function of temperature, T, and pressure, p. The line-shaded areas cover p,T-data points, where phase coexistence has been observed.

constant of ahex = 53 Å at 1 bar and 54 Å at 0.7 kbar can be determined at 6 °C. In the range 0.3−0.7 kbar, there is a gradual transition from the HII to the QII(Ia3d) phase with acub = 113 Å at 0.7 kbar. Upon further pressurization at 6 °C, the cubic phase eventually gives way to a lamellar phase with alam = 53 Å at 1.4 kbar (Figure 2). Unfortunately, the precise nature of this lamellar phase is not clear, because we have no access to the wide-angle scattering of the chain packing in our highpressure experiments. Thus, we will denote this phase as Lx in this study. However, it has been reported that MO in excess water shows a lamellar crystalline (Lc) phase at 8 °C and 1.4 kbar.18 Furthermore, a transient lamellar gel (Lβ) phase has been observed under the same conditions.18 Moreover, the FTIR experiments discussed below suggest the formation of the Lβ phase at 40 °C under high pressures. Therefore, Lx may represent Lβ or Lc in this study. However, whereas the lattice constant of the Lα phase is increasing with pressure, the one of the Lx phase is staying constant, because here the hydrocarbon chains of MO adopt the all-trans conformation with maximum extension. This property allows for the differentiation between the Lα and Lx phases. From numerous pressure scans at given temperatures, the pressure−temperature phase diagrams of MO at limited hydration in the absence and presence of the two fusion peptides have been constructed. In Figure S2 (Supporting Information), the pressure ranges of phase coexistence are given. From this information, the phase diagrams shown in Figure 3 are drawn, where the shaded areas indicate phase coexistence. For MO without fusion peptides, the inverse cubic phase is observed at high temperatures and low pressures, where the lipid chains require a large volume, and a negative curvature of the MO monolayer is induced. At higher pressures

and lower temperatures, the Lα phase is induced, where the lipid chains occupy a smaller volume appropriate for zero curvature. A further lamellar phase, denoted as Lx, is formed at even higher pressures and lower temperatures. Upon addition of 2 wt % of the fusion peptide HA2-FP, the phase behavior of MO at limited hydration is changing only marginally. The phases QII(Ia3d), Lα, and Lx can be found at almost the same pressures and temperatures (Figure 3). However, we observe a broadening of the QII(Ia3d)−Lα phase coexistence region to higher pressures and lower temperatures, suggesting a stabilization of the QII(Ia3d) phase. In contrast, upon adding 2 wt % TBEV-FP, the phase diagram of MO at limited hydration is very different (Figure 3). Here, we observe the HII phase with a relatively large negative curvature at high temperatures and low pressures, which is replaced by the QII(Ia3d) phase with an intermediate negative curvature at higher pressures and lower temperatures. Again, further pressurization at low temperatures induces a lamellar Lx phase, where the chains are in all-trans conformation and densely packed. Overall, the phase diagrams shown in Figure 3 clearly reveal the general phase sequence of lipids with increasing pressure, HII → QII → Lα → Lx, which is the reverse sequence observed with increasing temperature.5,24,33 The effect of adding fusion peptides is simply a shift of the phase transitions to higher pressures and lower temperatures. Whereas only a broadening of the QII(Ia3d)−Lα coexistence region to higher pressures is induced with HA2-FP, TBEV-FP is moving the HII−QII(Ia3d) phase boundary of MO from about 90 °C (see ref 19) to about 0 °C at 1 bar (Figure 3). As a consequence, the QII(Ia3d) phase is located at temperatures and pressures where the Lα phase is found in the absence of the fusion peptide (Figure 3). Therefore, the interaction of the 8495

DOI: 10.1021/acs.jpcb.7b06400 J. Phys. Chem. B 2017, 121, 8492−8502

Article

The Journal of Physical Chemistry B

Figure 4. Changes of the lattice constants and the Bragg peak intensities of MO (17 wt % water) upon pressure jumps from 1000 to 2000 bar (top) and from 2000 to 1000 bar (bottom) at 64 °C. The data show the formation of the lamellar phase (▲), when the pressure is increased (top), and the formation of the cubic phase (■) upon pressure release (bottom).

Figure 5. Changes of the lattice constants and the Bragg peak intensities of MO + HA2-FP (17 wt % water) upon pressure jumps from 700 to 1700 bar (top) and from 1600 to 700 bar (bottom) at 58 °C. The data show the formation of the lamellar phase (▲), when the pressure is increased (top), and the formation of the cubic phase (■) upon pressure release (bottom).

8496

DOI: 10.1021/acs.jpcb.7b06400 J. Phys. Chem. B 2017, 121, 8492−8502

Article

The Journal of Physical Chemistry B

at different pressures. The corresponding Gibbs energies of activation are themselves functions of pressure:

TBEV-FP with MO is strongly favoring the nonlamellar phases HII and QII(Ia3d) with negative curvature over the lamellar Lα phase. 3.2. Kinetics of the QII(Ia3d)−Lα phase transition of MO with HA2-FP upon pressure jumps. The QII(Ia3d)−Lα phase boundary is observed for pure MO and MO with HA2FP at limited hydration (Figure 3). Although the phase diagram of MO does not change in a major way upon interaction with HA2-FP, the kinetics and the mechanism of the QII(Ia3d)−Lα phase transformation might be different in the absence and the presence of the fusion peptide. In particular, the activation energy for fusion pore formation might be lowered by HA2-FP, which would be reflected in a faster QII(Ia3d) phase formation of MO with HA2-FP. Therefore, we carried out a series of pressure jumps across the QII(Ia3d)−Lα phase boundary and recorded the SAXS patterns with a time resolution of 150 ms. In Figures 4 and 5, the obtained lattice constants and the integrated intensities of the Bragg reflections are plotted as a function of time. The pressure jump is always triggered at t = 0 s. When the QII(Ia3d) → Lα transition of pure MO at limited hydration is induced by a positive pressure jump, the Bragg reflection of the Lα phase can be detected immediately, i.e. within the deadtime of the experiment (Figure 4 top). Over about 4 s, its intensity is growing with little change of the lattice constant. Simultaneously, the cubic phase vanishes, and no intermediate structure can be detected. It is interesting to see that the degradation of the cubic phase is associated with a sigmoidal increase of its lattice constant from about 110 Å to 115 Å. This might be explained by the diffusion-controlled adjustment of the hydration level of the QII(Ia3d) phase to the new pressure. The backward Lα → QII(Ia3d) transition of MO has been induced by a negative pressure jump (Figure 4 bottom). Remarkably, the transition proceeds over about 450 ms (three data points). The first recorded SAXS pattern of the cubic phase reveals a lattice constant of about 114 Å, which is relaxing to 110 Å over time. Thus, the rate constants for the forward and backward transition differ by a factor of about 10. For comparison, the QII(Ia3d) → Lα and Lα → QII(Ia3d) transition kinetics of MO with HA2-FP at limited hydration are shown in Figure 5. Overall, the data appear very similar to those of pure MO (Figure 4). Again, the formation of the Lα phase (pressure jump up) takes a few seconds, whereas the formation of the QII(Ia3d) phase (pressure jump down) takes a few 100 ms for MO with HA2-FP (Figure 5). The differences between the transition times observed without or with HA2-FP (Figures 4 and 5) will be due to the experimental error, which is apparent from additional pressure jump experiments shown in the Supporting Information (Figures S3 and S4). However, it is important to note that no intermediate phase can be detected in the presence of the fusion peptide, suggesting the same transition mechanism as that applying to pure MO. Moreover, HA2-FP does not exert major effects on the activation Gibbs energies of the QII(Ia3d) → Lα and Lα → QII(Ia3d) transitions. The much slower rate of Lα formation as compared to the QII(Ia3d) formation (Figure 4 and 5) reveals a higher activation Gibbs energy for the Lα formation. As we have performed pressure jumps across a phase boundary, we deal with relaxation kinetics. For the QII(Ia3d) → Lα transition, we jump to the higher pressure, p2, and then, the forward rate constant k1(p2) determines the transition rate. The reverse transition is characterized by the backward rate constant k2(p1) at lower pressure, p1. Notably, both rate constants are measured

ΔG1≠(p2 ) = ΔG1≠(p°) + ΔV1≠·(p2 − p°)

(1)

ΔG2≠(p1 ) = ΔG2≠(p°) + ΔV 2≠·(p1 − p°)

(2)

where p° is a reference pressure, e.g., 1 bar. The activation volumes, ΔV≠1 and ΔV≠2 , for the forward and backward phase transition are illustrated in Figure 6. Since the phase transition

Figure 6. Schematic volume diagram for the QII(Ia3d) → Lα phase transition with increasing pressure. Pressure jumps are carried out between p1 and p2. The rate constants, k1 and k2, refer to the forward and the backward phase transition. ΔV is the phase transition volume; ΔV≠1 and ΔV≠2 are the corresponding activation volumes.

volume, ΔV, can be assumed to be very small for Lα → QII phase transitions,18 we can write ΔV ≪ ΔV≠1 ≈ ΔV≠2 = ΔV≠1,2 (Figure 6). This relation implies that the slow Lα formation and the fast QII(Ia3d) formation (Figure 4 and 5) can simply be explained by the different pressure conditions and not by different mechanisms or activation energies. Using the transition state theory, the ratio of the rate constants for the forward and backward transition is given by ⎛ ΔG ≠(p ) − ΔG ≠(p ) ⎞ 1 2 1 2 ⎟ = exp⎜⎜ − ⎟ k 2(p1 ) RT ⎝ ⎠ k1(p2 )

⎛ ΔV ≠ ·(p − p ) ⎞ 1,2 2 1 ⎟ ≈ exp⎜⎜ − ⎟ RT ⎝ ⎠

(3)

Here, the pre-exponential factors of k1 and k2 are canceling each other, because k1 and k2 refer to the same temperature and both phases are composed of liquid-crystalline bilayers with similar dynamics. Furthermore, we approximate ΔG≠1 (p°) ≈ ΔG≠2 (p°). Then, according to our model (eq 3), the ratio of the rate constants is only determined by the pressure amplitude, p2 − p1, and the activation volume, ΔV≠1,2. We can roughly estimate the activation volume from the data of Figure 4. With k1/k2 = 0.1, p2 − p1 = 1000 bar, T = 64 °C, we obtain ΔV≠1,2 = 65 mL mol−1. Apparently, there is a substantial volume barrier between the QII(Ia3d) and the Lα phase. The structure of the QII(Ia3d) phase can be considered as bilayers that are connected by pores, such as depicted in Figure 1. Its formation proceeds via hemifusion stalks, where the cis monolayers are already fused, but the trans monolayers are still separated (Figure 1C). It is compelling that there is some transient void volume created when the cis and the trans monolayer of a bilayer detach to form a stalk. It is very likely that the activation volume estimated here can be ascribed to the stalk formation. 3.3. Secondary structure of HA2-FP and TBEV-FP in MO membranes. After having characterized the structure and transition kinetics of MO phases with fusion peptides, 8497

DOI: 10.1021/acs.jpcb.7b06400 J. Phys. Chem. B 2017, 121, 8492−8502

Article

The Journal of Physical Chemistry B

Figure 7. FTIR spectra of MO without fusion peptide and with HA2-FP or TBEV-FP (17 wt % water) at 40 °C as a function of pressure. Phase transitions can be detected by shifts of the CO and CC stretching bands of MO.

complementary FTIR spectra were recorded at 40 °C as a function of pressure for pure MO and MO with HA2-FP or TBEV-FP at limited hydration (Figure 7). The main aim of these measurements was to reveal the effects of MO and its phase transitions on the secondary structures of the fusion peptides. The amide I′ band (the prime indicates D2O as the solvent) of proteins or peptides is found at 1700−1600 cm−1 and is sensitive to the various secondary structure elements, such as α-helix, β-strand, and random coil.39 Because these structures have different peak wavenumbers, a decomposition of the amide I′ band into sub-bands allows for the estimation of the secondary structure element distribution in the protein or peptide. However, the strong CO stretching band of MO at 1750−1700 cm−1 is partially overlapping with the amide I′ band of the fusion peptides (Figure 7). Furthermore, as shown in Figure S1, the diamond background was subtracted by a polynomial, leaving some uncertainty in the range 1700−1600 cm−1. As a consequence, an increased fraction of 8 wt % of fusion peptide was found to be necessary for detecting a significant amide I′ band. Although this increased fraction may shift the phase boundaries given in Figure 3, meaningful insights can be drawn nonetheless from the FTIR spectra in Figure 7 regarding the secondary structures of the fusion peptides. As reference, the FTIR spectrum of pure MO at limited hydration and at 40 °C is shown in Figure 7 as a function of pressure. The CO stretching band of MO has two components. Whereas unhydrated CO groups are characterized by the high-wavenumber component, the low-wavenumber component represents CO groups that form

hydrogen bonds with water, which weakens the CO bond.40 Upon increasing the pressure, we detect a strong drop of the CO band intensity and a small shift to lower wavenumbers at about 3000 ± 400 bar (Figure 7). This change corresponds to the Lα → Lx phase transition in Figure 3, where the unordered hydrocarbon chains of MO adopt an ordered conformation and the packing of the lipid headgroups is increased. It has been reported that the Lc phase of MO is characterized by a correlation field splitting of the CO band with narrow components observed at 1741 and 1730 cm−1.40 Here, this splitting is not observed up to 10 kbar. Thus, we can exclude the formation of the Lc phase for MO at 40 °C. In addition, we detect the CC stretching vibration of MO at 1654 cm−1 below 3000 bar (Figure 7).41 This FTIR band is also indicating the Lα → Lx phase transition at 3000 ± 400 bar. At higher pressures, the band is located at the smaller wavenumber of about 1645 cm−1. Notably, the CO and the CC stretching bands of MO do not change upon passing through the QII(Ia3d) → Lα phase transition observed around 500 bar and 40 °C (Figure 3). Such fluid−fluid phase transitions involve weak conformational changes of the lipids only and are hardly detectable by FTIR spectroscopy.40 Moreover, both phases are based on a bilayer structure. When HA2-FP is added to MO at limited hydration, the pressure and temperature dependent phase behavior is only changing marginally (Figure 3). In the same way, the FTIR spectra of MO with HA2-FP are very similar to those of pure MO (Figure 7). We can detect the Lα → Lx phase transition at 40 °C and about 3700 ± 400 bar by a strong drop of the CO band intensity and a shift of the CC band wavenumber, in 8498

DOI: 10.1021/acs.jpcb.7b06400 J. Phys. Chem. B 2017, 121, 8492−8502

Article

The Journal of Physical Chemistry B

Figure 8. Normalized amide I′ bands of HA2-FP and TBEV-FP in MO−water after CO band subtraction of MO. The peak at 1673 cm−1 may reflect some TFA impurity.

Figure 9. Decomposition of the amide I′ bands of HA2-FP and TBEV-FP in aqueous buffer solution (pD = 5.4). The dashed line represents the sum of all sub-bands.

intensity is observed at 2100 bar, suggesting a phase transition at about 1800 ± 300 bar and 40 °C (Figure 7). This spectral change is in agreement with the HII → QII(Ia3d) transition observed by SAXS (Figure 3). Furthermore, the intensity of the lower-wavenumber component of the CO stretching band is increasing above 2100 bar (Figure 7), indicating a higher hydration of the MO headgroups, which is consistent with a larger headgroup area due to the less negative monolayer curvature in the QII(Ia3d) phase relative to the HII phase. Over the whole pressure range studied at 40 °C, a strong amide I′ band and a peak at 1673 cm−1 can be observed (Figure 7). It is also dropping in intensity up to 1500 bar, but this drop is likely to be caused by the neighboring CO band changing in the same way. From 2100 bar, a rather constant intensity is

good agreement with the phase transition located in the phase diagram of MO + HA2-FP as derived by SAXS (Figure 3). An increased absorbance over the range 1700 cm−1 − 1600 cm−1, as compared to pure MO, indicates the amide I′ band (Figure 7). Two additional peaks appear in the amide I′ band region at 1675 and 1625 cm−1, which are assigned below. The FTIR bands of MO + HA2-FP do not change upon the QII(Ia3d) → Lα phase transition, where the monolayer curvature is reduced, but the fluid state of the lipid molecules is retained. When TBEV-FP is added to MO, the FTIR spectra look very different from those of pure MO (Figure 7), in agreement with the huge difference of the corresponding phase diagrams (Figure 3). From 1 to 1500 bar, the intensity of the CO stretching band of MO is dropping, whereas a strong gain in 8499

DOI: 10.1021/acs.jpcb.7b06400 J. Phys. Chem. B 2017, 121, 8492−8502

Article

The Journal of Physical Chemistry B

hemifusion stalks and fusion pores, but they are not generally sufficient to drive this process. Moreover, we see that the induction of MO phases with negative monolayer curvature by the fusion peptides is associated with a lower packing density in the MO membranes, because this effect can at least partially be reversed by the application of high pressure. Of note, this might have implications regarding membrane fusion in organisms living at extreme pressure conditions, such as in the deep sea, where pressures up to the 1 kbar level are encountered. The existence of life under extreme pressures suggests that nature has found membrane fusion mechanisms that overcome the diminishing effect of pressure on negative monolayer curvature at least up to 1 kbar. However, membrane fusion under even higher pressures might be suppressed, even though not biologically relevant. The pressure dependent experiments of this study have also revealed a substantial volume barrier between the lamellar and cubic phases of MO, which is likely caused by the detachment of the two monolayers of a membrane upon stalk formation. Within the experimental error, the activation volume does not change in the presence of HA2-FP, suggesting a similar geometry of the stalk. Moreover, the secondary structures of the fusion peptides seem to be unaffected by the interaction with the MO membranes, and they even do not change upon lipid phase transitions. Thus, the fusion peptides control the structure of the lipid membrane and not vice versa when interacting with lipid membranes.

observed, but otherwise no clear change of the amide I′ band is apparent. To further evaluate the secondary structure of the fusion peptides interacting with MO, the CO band has been subtracted. The obtained amide I′ bands of HA2-FP and TBEV-FP are shown in Figure 8. In all spectra, a peak at 1673 cm−1 is observed that might indicate some TFA impurity (although the peptides were labeled as TFA free by the manufacturer), but it can also be assigned to turn structures of the fusion peptides.39,42 Furthermore, the CC stretching band of MO is visible at 1655 or 1645 cm−1 in the HA2-FP spectra. However, as can be clearly seen in Figure 8, there is little change of the amide I′ band shape with increasing pressure in the case of both fusion peptides. Thus, HA2-FP and TBEV-FP which are added to MO at limited hydration do not change their secondary structures in a major way, when passing through the various lipid phases. For comparison, the amide I′ bands of the fusion peptides have also been recorded in aqueous buffer solution in the absence of MO (Figure 9). Remarkably, they look very similar to the corresponding spectra of MO + HA2-FP and MO + TBEV-FP (Figure 8), suggesting no major conformational change of the fusion peptides upon the interaction with MO. A decomposition of the amide I′ band of HA2-FP yields subbands at 1673 cm−1, 1654 cm−1, 1638 cm−1, and 1623 cm−1 (Figure 9), representing probably turns/TFA (1673 cm−1), αhelices (1654 cm−1), and β-structures (1638 and 1623 cm−1).39 In the case of TBEV-FP, the amide I′ band can be fitted by subbands representing turns/TFA (1673 cm−1) and unordered structures (1647 cm−1). The secondary structures of the fusion peptides observed here are in agreement with other studies. For example, by using circular dichroism (CD) spectroscopy, the secondary structure of HA2-FP has been reported to consist of mainly coil and turn structures (84%) with some β-sheets (16%) in aqueous solution.43 In a further study, a dominating β-structure with some α-helices is proposed.31 In contact with phospholipid vesicles, the analysis of the amide I′ band of the HA2 fusion peptide yields mainly β-sheet and random coil structures in addition to about 20% α-helical conformations.30 The secondary structure of TBEV-FP has also been studied by CD spectroscopy and found to consist mainly of coil and turn structures.43 Notably, the intrinsic fluorescence of fusion peptides senses their insertion into membranes by a blue shift of the Trp band,14,44 and NMR spectroscopy reveals the structure of fusion peptides in membranes with atomic resolution. For example, the structure of HA2-FP has been found to consist of two short helices in a V or hairpin arrangement.25,29



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpcb.7b06400. FTIR spectrum of MO at limited hydration with background of the diamond anvil cell. Phase diagrams of MO without and with fusion peptides including data points. SAXS data of MO + HA2-FP (17 wt % water) during pressure jumps at 49 °C. (PDF)



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

Roland Winter: 0000-0002-3512-6928 Claus Czeslik: 0000-0003-0612-3602 Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS Financial support from the Deutsche Forschungsgemeinschaft (DFG, Forschergruppe 1979) is gratefully acknowledged. We thank the ESRF (Grenoble, France) for providing synchrotron beamtime. We also thank Dr. Michael Sztucki (ESRF), Dr. Mirko Erlkamp (TU Dortmund University), and Dr. Sebastian Grobelny (TU Dortmund University) for assistance with the SAXS experiments.

4. CONCLUSIONS The effects of two different fusion peptides, HA2-FP and TBEV-FP, on the phase behavior and phase transition kinetics of MO at limited hydration have been studied. The results obtained clearly demonstrate that both peptides are very different in their potential to affect MO membranes. When HA2-FP is added to MO membranes, the phase behavior is only marginally changed and the Gibbs activation energies of the inverse bicontinuous cubic to lamellar phase transition in the forward and backward directions remain unchanged. In contrast, the interaction of TBEV-FP with MO induces the inverse hexagonal phase at ambient pressure and largely destabilizes the lamellar phase. These findings support the view that fusion peptides can assist in the formation of



REFERENCES

(1) Harrison, S. C. Viral membrane fusion. Virology 2015, 479−480, 498−507. (2) White, J. M.; Delos, S. E.; Brecher, M.; Schornberg, K. Structures and mechanisms of viral membrane fusion proteins: Multiple variations on a common theme. Crit. Rev. Biochem. Mol. Biol. 2008, 43, 189−219. 8500

DOI: 10.1021/acs.jpcb.7b06400 J. Phys. Chem. B 2017, 121, 8492−8502

Article

The Journal of Physical Chemistry B (3) Jardetzky, T. S.; Lamb, R. A. A class act. Nature 2004, 427, 307− 308. (4) Schroth-Diez, B.; Ludwig, K.; Baljinnyam, B.; Kozerski, C.; Huang, Q.; Herrmann, A. The role of the transmembrane and of the intraviral domain of glycoproteins in membrane fusion of enveloped viruses. Biosci. Rep. 2000, 20, 571−595. (5) Seddon, J. M. Structure of the inverted hexagonal (HII) phase, and non-lamellar phase transitions of lipids. Biochim. Biophys. Acta, Rev. Biomembr. 1990, 1031, 1−69. (6) Darkes, M. J. M.; Davies, S. M. A.; Bradshaw, J. P. X-ray diffraction study of feline leukemia virus fusion peptide and lipid polymorphism. FEBS Lett. 1999, 461, 178−182. (7) Siegel, D. P. Fourth-order curvature energy model for the stability of bicontinuous inverted cubic phases in amphiphile-water systems. Langmuir 2010, 26, 8673−8683. (8) Siegel, D. P. The Gaussian curvature elastic energy of intermediates in membrane fusion. Biophys. J. 2008, 95, 5200−5215. (9) Siegel, D. P.; Kozlov, M. M. The Gaussian curvature elastic modulus of N-monomethylated dioleoylphosphatidylethanolamine: Relevance to membrane fusion and lipid phase behavior. Biophys. J. 2004, 87, 366−374. (10) Siegel, D. P.; Epand, R. M. Effect of influenza hemagglutinin fusion peptide on lamellar/inverted phase transitions in dipalmitoleoylphosphatidylethanolamine: implications for membrane fusion mechanisms. Biochim. Biophys. Acta, Biomembr. 2000, 1468, 87−98. (11) Colotto, A.; Epand, R. M. Structural study of the relationship between the rate of membrane fusion and the ability of the fusion peptide of influenza virus to perturb bilayers. Biochemistry 1997, 36, 7644−7651. (12) Colotto, A.; Martin, I.; Ruysschaert, J.-M.; Sen, A.; Hui, S. W.; Epand, R. M. Structural study of the interaction between the SIV fusion peptide and model membranes. Biochemistry 1996, 35, 980− 989. (13) Tristam-Nagle, S.; Nagle, J. F. HIV-1 fusion peptide decreases bending energy and promotes curved fusion intermediates. Biophys. J. 2007, 93, 2048−2055. (14) Nieva, J. L.; Agirre, A. Are fusion peptides a good model to study viral cell fusion? Biochim. Biophys. Acta, Biomembr. 2003, 1614, 104−115. (15) Epand, R. M. Fusion peptides and the mechanism of viral fusion. Biochim. Biophys. Acta, Biomembr. 2003, 1614, 116−121. (16) Martin, I.; Ruysschaert, J.-M. Common properties of fusion peptides from diverse systems. Biosci. Rep. 2000, 20, 483−500. (17) Hyde, S. T.; Andersson, S.; Ericsson, B.; Larsson, K. A cubic structure consisting of a lipid bilayer forming an infinite periodic minimum surface of the gyroid type in the glycerolmonooleat-water system. Z. Kristallogr. 1984, 168, 213−219. (18) Czeslik, C.; Winter, R.; Rapp, G.; Bartels, K. Temperature- and pressure-dependent phase behavior of monoacylglycerides monoolein and monoelaidin. Biophys. J. 1995, 68, 1423−1429. (19) Briggs, J.; Chung, H.; Caffrey, M. The temperature-composition phase diagram and mesophase structure characterization of the monoolein/water system. J. Phys. II 1996, 6, 723−751. (20) Qiu, H.; Caffrey, M. The phase diagram of the monoolein/water system: metastability and equilibrium aspects. Biomaterials 2000, 21, 223−234. (21) Pisani, M.; Bernstorff, S.; Ferrero, C.; Mariani, P. Pressure induced cubic-to-cubic phase transition in monoolein hydrated systems. J. Phys. Chem. B 2001, 105, 3109−3119. (22) Conn, C. E.; Ces, O.; Squires, A. M.; Mulet, X.; Winter, R.; Finet, S. M.; Templer, R. H.; Seddon, J. M. A pressure-jump timeresolved X-ray diffraction study of cubic-cubic transition kinetics in monoolein. Langmuir 2008, 24, 2331−2340. (23) Kulkarni, C. V.; Wachter, W.; Iglesias-Salto, G.; Engelskirchen, S.; Ahualli, S. Monoolein: a magic lipid? Phys. Chem. Chem. Phys. 2011, 13, 3004−3021. (24) Tang, T.-Y. D.; Brooks, N. J.; Jeworrek, C.; Ces, O.; Terrill, N. J.; Winter, R.; Templer, R. H.; Seddon, J. M. Hydrostatic pressure

effects on the lamellar to gyroid cubic phase transition of monolinolein at limited hydration. Langmuir 2012, 28, 13018−13024. (25) Tamm, L. K.; Han, X.; Li, Y.; Lai, A. L. Structure and function of membrane fusion peptides. Biopolymers 2002, 66, 249−260. (26) Bressanelli, S.; Stiasny, K.; Allison, S. L.; Stura, E. A.; Duquerroy, S.; Lescar, J.; Heinz, F. X.; Rey, F. A. Structure of a flavivirus envelope glycoprotein in its low-pH-induced membrane fusion conformation. EMBO J. 2004, 23, 728−738. (27) Allison, S. L.; Schalich, J.; Stiasny, K.; Mandl, C. W.; Heinz, F. X. Mutational evidence for an internal fusion peptide in flavivirus envelope protein E. J. Virol. 2001, 75, 4268−4275. (28) Han, X.; Tamm, L. K. A host-guest system to study structurefunction relationships of membrane fusion peptides. Proc. Natl. Acad. Sci. U. S. A. 2000, 97, 13097−13102. (29) Lorieau, J. L.; Louis, J. M.; Bax, A. The complete influenza hemagglutinin fusion domain adopts a tight helical hairpin arrangement at the lipid:water interface. Proc. Natl. Acad. Sci. U. S. A. 2010, 107, 11341−11346. (30) Lüneberg, J.; Martin, I.; Nüßler, F.; Ruysschaert, J.-M.; Herrmann, A. Structure and topology of the influenza virus fusion peptide in lipid bilayers. J. Biol. Chem. 1995, 270, 27606−27614. (31) Takahashi, S. Conformation of membrane fusion-active 20residue peptides with or without lipid bilayers. Implication of α-helix formation for membrane fusion. Biochemistry 1990, 29, 6257−6264. (32) Mariani, P.; Luzzati, V.; Delacroix, H. Cubic phases of lipidcontaining systems. Structure analysis and biological implications. J. Mol. Biol. 1988, 204, 165−189. (33) Winter, R.; Czeslik, C. Pressure effects on the structure of lyotropic lipid mesophases and model biomembrane systems. Z. Kristallogr. - Cryst. Mater. 2000, 215, 454−474. (34) Covington, A. K.; Paabo, M.; Robinson, R. A.; Bates, R. G. Use of the glass electrode in deuterium oxide and the relation between the standardized pD (paD) scale and the operational pH in heavy water. Anal. Chem. 1968, 40, 700−706. (35) Huang, T. C.; Toraya, H.; Blanton, T. N.; Wu, Y. X-ray powder diffraction analysis of silver behenate, a possible low-angle diffraction standard. J. Appl. Crystallogr. 1993, 26, 180−184. (36) Kraineva, J.; Narayanan, R. A.; Kondrashkina, E.; Thiyagarajan, P.; Winter, R. Kinetics of lamellar-to-cubic and intercubic phase transitions of pure and cytochrome c containing monoolein dispersions monitored by time-resolved small-angle X-ray diffraction. Langmuir 2005, 21, 3559−3571. (37) Woenckhaus, J.; Köhling, R.; Winter, R.; Thiyagarajan, P.; Finet, S. High pressure-jump apparatus for kinetic studies of protein folding reactions using the small-angle synchrotron x-ray scattering technique. Rev. Sci. Instrum. 2000, 71, 3895−3899. (38) Siminovitch, D. J.; Wong, P. T. T.; Mantsch, H. H. Effects of cis and trans unsaturation on the structure of phospholipid bilayers: A high-pressure infrared spectroscopic study. Biochemistry 1987, 26, 3277−3287. (39) Byler, D. M.; Susi, H. Examination of the secondary structure of proteins by deconvolved FTIR spectra. Biopolymers 1986, 25, 469− 487. (40) Lendermann, J.; Winter, R. Interaction of cytochrome c with cubic monoolein mesophases at limited hydration conditions: The effects of concentration, temperature and pressure. Phys. Chem. Chem. Phys. 2003, 5, 1440−1450. (41) Sinclair, R. G.; McKay, A. F.; Myers, G. S.; Jones, R. N. The infrared absorption spectra of unsaturated fatty acids and esters. J. Am. Chem. Soc. 1952, 74, 2578−2585. (42) Fabian, H.; Szendrei, G. I.; Mantsch, H. H.; Otvos, L. Comparative analysis of human and Dutch-type Alzheimer β-amyloid peptides by infrared spectroscopy and circular dichroism. Biochem. Biophys. Res. Commun. 1993, 191, 232−239. (43) Weise, K.; Reed, J. Fusion peptides and transmembrane domains of fusion proteins are characterized by different but specific structural properties. ChemBioChem 2008, 9, 934−943. (44) Gray, C.; Tatulian, S. A.; Wharton, S. A.; Tamm, L. K. Effect of the N-terminal glycine on the secondary structure, orientation, and 8501

DOI: 10.1021/acs.jpcb.7b06400 J. Phys. Chem. B 2017, 121, 8492−8502

Article

The Journal of Physical Chemistry B interaction of the influenza hemagglutinin fusion peptide with lipid bilayers. Biophys. J. 1996, 70, 2275−2286.

8502

DOI: 10.1021/acs.jpcb.7b06400 J. Phys. Chem. B 2017, 121, 8492−8502