Deconvoluting the Effect of the Hydrophobic and Hydrophilic Domains

Oct 21, 2015 - Leonie van 't Hag , Liliana de Campo , Christopher J. Garvey , George C. Feast , Anna E. Leung , Nageshwar Rao Yepuri , Robert Knott , ...
0 downloads 0 Views 3MB Size
Article pubs.acs.org/Langmuir

Deconvoluting the Effect of the Hydrophobic and Hydrophilic Domains of an Amphiphilic Integral Membrane Protein in Lipid Bicontinuous Cubic Mesophases Leonie van ’t Hag,†,‡,⊥ Hsin-Hui Shen,∥,# Jingxiong Lu,∥ Adrian M. Hawley,$ Sally L. Gras,†,‡,§ Calum J. Drummond,*,⊥,^ and Charlotte E. Conn*,^ †

Department of Chemical and Biomolecular Engineering, ‡Bio21 Molecular Science and Biotechnology Institute, and §The ARC Dairy Innovation Hub, The University of Melbourne, Parkville, Victoria 3010, Australia ⊥ CSIRO Manufacturing Flagship, Clayton, Victoria 3169, Australia ∥ Department of Microbiology, and #Faculty of Engineering, Monash University, Melbourne, Victoria 3800, Australia $ Australian Synchrotron, Clayton, Victoria 3168, Australia ^ School of Applied Sciences, College of Science, Engineering and Health, RMIT University, Melbourne, Victoria 3001, Australia S Supporting Information *

ABSTRACT: Lipidic bicontinuous cubic mesophases with encapsulated amphiphilic proteins are widely used in a range of biological and biomedical applications, including in meso crystallization, as drug delivery vehicles for therapeutic proteins, and as biosensors and biofuel cells. However, the effect of amphiphilic protein encapsulation on the cubic phase nanostructure is not well-understood. In this study, we illustrate the effect of incorporating the bacterial amphiphilic membrane protein Ag43, and its individual hydrophobic β43 and hydrophilic α43 domains, in bicontinuous cubic mesophases. For the monoolein, monoalmitolein, and phytantriol cubic phases with and without 8% w/w cholesterol, the effect of the full length amphiphilic protein Ag43 on the cubic phase nanostructure was more significant than the sum of the individual hydrophobic β43 and hydrophilic α43 domains. Several factors were found to potentially influence the impact of the hydrophobic β43 domain on the cubic phase internal nanostructure. These include the size of the hydrophobic β43 domain relative to the thickness of the lipid bilayer, as well as its charge and diameter. The size of the hydrophilic α43 domain relative to the water channel radius of the cubic mesophase was also found to be important. The secondary structure of the Ag43 proteins was affected by the hydrophobic thickness and physicochemical properties of the lipid bilayer and the water channel diameter of the cubic phase. Such structural changes may be small but could potentially affect membrane protein function.



INTRODUCTION Transmembrane proteins (TMPs) are predicted to represent 20−30% of the proteins produced in most organisms.1,2 TMPs are involved in many essential cellular functions, including transport of molecules across the cell membrane, signal transduction, and energy conversion.3 Lipid bilayer systems can be exploited to encapsulate amphiphilic proteins in vitro to provide information on protein-lipid interactions, on fundamental biological processes, and for a range of biological and biomedical applications like drug delivery and in meso crystallization. Self-assembled lipid bilayer systems that mimic the lipid bilayer environment in the cell membrane are commonly used for this purpose.4 A good model system to study membrane-protein interactions is the lipidic bicontinuous cubic phase. Cubic phases have been observed in vivo in the endoplasmatic reticulum and in the nuclear and mitochondrial membranes of cells.5 Bicontinuous cubic phases (QII) consist of a continuous three-dimensional periodic structure of surfactant © 2015 American Chemical Society

bilayers and water channels with a high surface area to volume ratio.6 The three experimentally observed cubic phases are the primitive cubic phase (QIIP), diamond cubic phase (QIID) and gyroid cubic phase (QIIG) (Figure 1). The amphiphilic nanostructure of cubic mesophases is particularly useful for the encapsulation of amphiphilic proteins and peptides: the hydrophobic domain is incorporated within the lipid bilayer with the hydrophilic domain(s) encapsulated within the water channels.7−9 Cubic mesophases are widely used in a range of biological and biomedical applications including in meso crystallization of TMPs for structure determination.3 Nanoparticles with an internal cubic phase nanostructure (cubosomes) are also used as drug delivery vehicles for amphiphilic and hydrophobic Received: August 30, 2015 Revised: October 17, 2015 Published: October 21, 2015 12025

DOI: 10.1021/acs.langmuir.5b03256 Langmuir 2015, 31, 12025−12034

Article

Langmuir

variants: the full length Ag43 protein and the separate hydrophobic β43, and hydrophilic α43 domains.21 Modeling of the hydrophobic integral outer membrane β43 domain (50 kDa),26 Figure 2(A), showed that it forms an

Figure 1. Schematic representations of (A) primitive cubic phase, QIIP; (B) diamond cubic phase, QIID; (C) gyroid cubic phase, QIIG. Sheets represent the lipid bilayers.

therapeutic proteins and peptides and imaging agents for medical diagnostics.10 These systems are of particular use for poorly water-soluble proteins and imaging agents, which are unsuited for solution-based delivery methods. Other prospective applications of the cubic phase include use in biosensors11,12 and biofuel cells13,14 upon encapsulation of enzymes. Cubic phase retention is important for all these applications. However, TMPs can significantly impact the cubic phase internal structure upon incorporation, including effecting a phase transition to a noncubic nanostructure in some cases.15 To advance the use of the hybrid lipid−protein systems in a range of applications, we must understand the effect of incorporated proteins and peptides on the internal nanostructure of the lipidic cubic phase. Despite recent studies investigating the phase transitions caused by incorporation of TMPs within a bicontinuous cubic phase,15−18 we are not yet able to predict the phase transitions caused by incorporation of an amphiphilic TMP within the cubic phase.15 In addition, the physical characteristics of the lipid matrix can play an important role in the functionality of peptides, proteins, and receptors that are embedded in the lipid membrane and may assist in the retention of protein functionality.19,20 This study aims to improve our fundamental understanding of the complex interplay between the protein and the lipid matrix using the amphiphilic integral membrane protein Antigen 43 (Ag43). Ag43 is an outer membrane protein composed of two proteinaceous subunits. This 110 kDa protein can be cleaved in situ,21 allowing the direct effect of the separate protein domains on the structure of the bicontinuous cubic phase to be deconvoluted for the first time. In addition, the effect of lipid physicochemical characteristics on the secondary structure of Ag43 has been investigated. Ag43, belonging to the autotransporter family, is an abundant outer membrane protein in Escherichia coli (E. coli). Autotransporter proteins are involved in the pathogenesis of Gram-negative bacteria; the functions of Ag43 (a selfassociating surface-located protein) include biofilm formation and cell aggregation.22,23 These functions are associated with chronic and persistent bacterial infections, which are difficult to treat. Examples of such infections include infections of the ear, gums, heart, urinary tract, and on medical devices.23,24 The structural information available on autotransporter proteins, on a molecular level, is very limited despite their importance in bacterial virulence. From N to C terminus, the Ag43 protein is 1039 amino acids (AA) long and contains a signal peptide (AA 1−52), a passenger domain (α43) that is secreted (AA 53− 552), an autochaperone domain, and a β-barrel domain (β43) that forms a transmembrane domain (AA 553−1039).22,23 The functional α43 domain and the translocation β43 domain remain in close-contact after processing in E. coli via noncovalent interactions.25 Herein, we have investigated three protein

Figure 2. Illustration of secondary structures and dimensions of: (A) the hydrophobic β43 domain. A modeled structure for the translocation unit and autochaperone domain of Ag43; amino acids 553−1039, adapted from the literature.22 The lipid-exposed surface of β-barrels in the outer membrane of Gram-negative bacteria is approximately 27 Å thick.27,28 (B) Crystal structure and the dimensions of the passenger domain of Ag43 (amino acids 55−554, PDB-ID 4KH3).23 Ribbon representations are shown from N (top) to C terminus (bottom) and in both structures the β-strands are colored yellow/orange; the loops are depicted in gray in (A) and in blue in (B). In (A), the central αhelix is blue.

antiparallel β-helical structure that is surface-exposed with an autochaperone domain that is required for folding of the passenger domain.22 Typically, the lipid-exposed surface of βbarrels in the outer membrane of Gram-negative bacteria is approximately 27 Å thick, which commonly contains a central α-helix.27,28 The hydrophilic passenger domain α43 (60 kDa)26 forms a parallel β-helical rod, which is also common for members of the autotransporter family.25,29 The recently revealed crystal structure of the β-helix shows that it is Lshaped with a length of ∼100 Å as shown in Figure 2(B).23 The cubic phase forming lipids monoolein, monopalmitolein, and phytantriol were used to incorporate Ag43, β43, and α43. To date, monoolein (MO, Figure 3(A)) has been used fairly ubiquitously for in meso crystallization and drug delivery

Figure 3. Molecular models of the three lipids used in this study produced using Chem 3D pro v10 using its MM2 energy-minimization routine for (A) monoolein, (B) monopalmitolein, and (C) phytantriol. 12026

DOI: 10.1021/acs.langmuir.5b03256 Langmuir 2015, 31, 12025−12034

Article

Langmuir

Figure 4. One-dimensional diffraction plots of intensity against q on a double logarithmic scale for QIID. In all diffraction plots, the √2, √3, √4, √6, √8, and √9 reflections are visible. (A) MO with α43 at 0.1 mg mL−1 (7.9 × 10−5 mole %); lp = 11.0 nm. (B) MP with Ag43 at 0.7 mg mL−1 (2.8 × 10−4 mole %); lp = 12.1 nm. (C) PT with β43 at 0.2 mg mL−1 (1.8 × 10−4 mole %); lp = 7.0 nm.

applications.30 It is readily available, and it forms robust cubic phases at room temperature and in excess water (lattice parameter ∼102 Å). However, MO does not form a thermodynamically stable cubic phase at low temperatures and is hydrolyzable at high and low pH.31,32 TMPs with varying hydrophilic and hydrophobic domain sizes might benefit from encapsulation within cubic phases with different internal mesophase nanostructures.16 Depending on the size of the membrane proteins, varying the water-channel diameter and the bilayer thickness have been used to improve success rates for in meso crystallization.33,34 We therefore tested two other lipids that form the cubic phase at room temperature and in excess water. Monopalmitolein (MP, Figure 3(B)) forms a cubic phase with slightly larger lattice parameters (∼125 Å) than those of MO.35 Its molecular structure is similar to that of MO, but it has a slightly shorter alkyl chain. Phytantriol (PT, Figure 3(C)) forms a cubic phase with significantly smaller lattice parameters (∼66 Å), and it contains a branched alkyl chain.36 Using MO, MP, and PT allows us to investigate the effect of the internal cubic mesophase nanostructure and alkyl chain identity on protein incorporation in the cubic mesophases and on the secondary structure of the three Ag43 proteins. Lipidic cubic phases are commonly doped with 5−10% w/w cholesterol for in meso crystallization.37 In this study MO, MP, and PT cubic phases were studied both in the presence and absence of 8% w/w cholesterol. Small angle X-ray scattering (SAXS) is used to study the effect of Ag43, β43, and α43 and incorporation on the structure of the bicontinuous cubic phases formed by MO, MP, and PT.38 Circular dichroism spectroscopy (CD) is also used to study the effect of the cubic mesophases on the secondary structure of the three Ag43 proteins.39,40 Comparing the hydrophobic β43 and hydrophilic α43 domains with the full length Ag43 protein will lead to an improved understanding of the effect of amphiphilic integral membrane protein incorporation in bicontinuous cubic phases and a complementary understanding of the effect of the lipid bilayers on protein secondary structures.



nonionic detergent for purification of membrane-bound proteins, and it is composed of a mixture of alkyl glucosides (∼300 g mol−1). SD-2 96-well crystallization plates were obtained from IDEX Corporation (Rohnert Park, CA, USA). Hamilton syringes were purchased from Hamilton Company (Reno, NV, USA) and syringe couplers from TTP Labtech (Cambridge, MA, USA). Ag43, β43, and α43 Expression and Purification. Ag43 with 10 His-tags on the C terminus was expressed by Top10 cells. The cells were cultured at 37 °C before being induced by 0.1% L-arabinose at OD = 0.6 followed by expression at 30 °C for 4 h. Cells were harvested by centrifugation at 8,000 rpm for 20 min at 4 °C and disrupted by ultrasonication (VWR International-102C) in cell lysis buffer containing 20 mM Tris, 150 mM NaCl, 0.1% v/v of 2mercaptoethanol, 10 μg/mL of DNase I, 0.2 mg/mL of lysozyme, 2 mM PMSF, and protease inhibitor (1 tablet for 50 mL). The cell lysate was centrifuged at 15,000 rpm for 20 min at 4 °C, and the membrane fraction in the supernatant was pelleted down by further centrifugation at 35,000 rpm for 1 h at 4 °C. After the pellet was dissolved in 20 mM Tris, 150 mM NaCl, and 0.5% v/v of Elugent, Ni-NTA-agarose beads were added to bind Ag43 at 4 °C overnight. Ag43 conjugated beads were washed twice with 20 mM Tris, 150 mM NaCl, 0.5% v/v of elugent, and 20 mM imidazole and Ag43 molecules were eluted with elution buffer 20 mM Tris, 150 mM NaCl, 0.5% v/v of Elugent, and 500 mM imidazole. α43 and β43 were prepared by cleaving Ag43 with 20 mM Tris, 150 mM NaCl, 0.5% v/v of Elugent, and 8 M urea. α43 was separated by running through Ni-NTA-agarose beads, and then β43 was eluted with elution buffer. Plate Setup. Samples for small angle X-ray scattering (SAXS) measurements were prepared in standard 96-well SD-2 crystallization plates using a high-throughput method, as described in detail in a recent article.38 MO, MP, and PT solutions in ethanol were initially prepared at 200 mg mL−1. For samples with cholesterol, MO, MP, and PT were initially mixed with 8% w/w of cholesterol and then dissolved at 200 mg mL−1 in ethanol with a few drops of chloroform. A sample of 1.05 μL of lipid solution (correlating to 0.21 mg of dry lipid) was dispensed onto the top and bottom subwells of the 96-well SD-2 plates using a Mosquito robot (TTP Labtech) based at the C3 Collaborative Crystallization Centre, CSIRO, Parkville, Australia. The plates were dried under vacuum for 24 h and subsequently stored in a fume hood for 5 days to remove solvent. Next, 20 μL of 20 mM Tris, 150 mM NaCl, and 0.5% v/v of Elugent was added to the reservoir for each well. Using the Mosquito robot, 0.28 μL of protein solution at 0−3.5 mg mL−1 in 20 mM Tris, 150 mM NaCl, and 0.5% v/v of Elugent was dispensed on top of the dried lipid film, which means the samples were in 57% w/w (excess) water. The concentrations of Ag43 protein studied were up to 3.5 mg mL−1 for the full-length Ag43 protein (110 kDa), 2 mg mL−1 for β43 (50 kDa), and 1.5 mg mL−1 for α43 (60 kDa). Small-Angle X-ray Scattering (SAXS) Measurements. The structure of the lipid mesophase was determined in situ within the 96well plates 1 day after the addition of the protein solution at 25 °C. A temperature-controlled custom-designed plate holder was used to

EXPERIMENTAL SECTION

Materials. MO, cholesterol, Tris, and NaCl were purchased from Sigma-Aldrich (St. Louis, MO, USA). MP was obtained from NuCheck Prep, Incorporated (Waterville, MN, USA), and 3,7,11,15tetramethyl-1,2,3-hexadecanetriol (PT) (96%) was purchased from DSM Nutritional Products (Germany). Ethanol and chloroform were bought from Merck & Co., Incorporated (Whitehouse Station, NJ, USA). Elugent detergent was purchased from VWR International, Pty Ltd. (Murarrie, QLD, Australia). Elugent is a commercially available 12027

DOI: 10.1021/acs.langmuir.5b03256 Langmuir 2015, 31, 12025−12034

Article

Langmuir mount the plates directly onto the SAXS/WAXS beamline at Australian Synchrotron.41 Automated scans across the plates were performed by using position variables based on the well positions. The experiments used a beam of wavelength λ = 1.1271 Å (11.0 keV) with dimensions 300 × 200 μm, a typical flux of 1 × 1013 photons s−1, and an exposure time of 0.5 s. Two-dimensional (2D) diffraction images were recorded on a Pilatus 1 M detector, which offers very low noise and rapid data collection over a large active area. One-dimensional (1D) scattering profiles were obtained by radial averaging of the 2D images produced. SAXS Data Analysis. Ordered nanostructured lyotropic liquid crystalline diamond cubic QIID mesophases with distinct diffraction patterns were studied. Diffraction patterns were obtained that may be used as unambiguous identification for the phase, provided an adequate number of reflections are observed. Representative 1D diffraction patterns obtained for diamond cubic QIID mesophases of MO, MP, and PT are shown in Figure 4. For the majority of samples, many orders of diffraction were observed despite the small sample size. The high-throughput analysis program RapidPhaseIdent, developed at the SAXS/WAXS beamline of the Australian Synchrotron and described in a recent publication,42 was used for identification of lipid mesophases and lattice parameters along with the IDL-based AXCESS software package developed at Imperial College London.43 Circular Dichroism Spectroscopy (CD) Analysis. Far ultraviolet circular dichroism spectra were collected on an AVIV 410SF Protein Solutions CD spectrometer (Lakewood Township, NJ, USA) using a 0.1 mm path length demountable quartz cuvette (Starna Pty Ltd., Baulkham Hills, NSW, Australia). The CD spectra of Ag43, β43, and α43 proteins were collected at 0.57−0.88 mg mL−1 in a solution of 20 mM Tris and 0.5% v/v of Elugent. A buffer without 150 mM NaCl was used for CD analysis to improve signal quality. Cubic phase samples without excess buffer, at 37% w/w and 25% w/w water for MO and PT, respectively, were prepared using Hamilton syringes.40 The final concentrations of Ag43, β43, and α43 studied within the lipid phase were 0.65−0.88 mg mL−1. These proteins were prepared initially as stocks in 20 mM Tris and 0.5% v/v of Elugent. A circulating water bath was used to keep the temperature of all samples constant at 25 °C during CD measurements. CD spectra were collected between 190 and 260 nm with a 0.5 nm interval and an averaging time of 4 s. For proteins in solution and the PT cubic phase, a bandwidth of 1 nm was used. For proteins in the MO cubic phase, a bandwidth of 2 nm was used. Data analysis was performed using the online DICHROWEB server44 using the CDSSTR algorithm.45 The reference database used is SP175, which is thought to be most accurate for membrane proteins.46 The data presented are the averages from three separate samples, where four to six replicate scans were initially averaged for each sample. A background of 20 mM Tris and 0.5% v/v of Elugent (collected with 1 or 2 nm bandwidth) was also subtracted from the data. The intensity of the data was normalized relative to the intensity of the detergent solution spectra, aiding comparison between spectra. For protein in the lipid phases, three spectra were again collected. Only data collected with a dynode voltage lower than 750 V are shown.



lattice parameter (lp), bilayer thickness, and water channel radius for the cubic phases with buffered detergent solution determined in the absence of protein is presented in Table 1. Table 1. Literature Values of QIID Lattice Parameters (nm) of MO, MP, and PT at 20−25 °C in Excess Water and Experimentally Observed Values of All Three Lipids with and without 8% w/w of Cholesterol with Buffered Detergent Solution Consisting of 20 mM Tris, 150 mM NaCl, and 0.5% v/v of Elugent at 57% w/w (Excess) Water and 25 °Ca lipid MO MO, 8% chol. MP MP, 8% chol. PT PT, 8% chol.

lp literature (nm)

lp with detergent solution (nm)

bilayer thickness (Å)

water channel diameter (Å)

10.231,32

10.9 10.9

33.4

51.8

12.535

12.0 11.7

32.5

61.2

6.636

7.0 6.8

29.8

24.8

a

The experimental values are the average of the protein or protein fragment 0 mg mL−1 values of all six plates (see Supporting Information S1). On the basis of the QIID lattice parameters of MO, MP, and PT without cholesterol, the bilayer thicknesses and water channel diameters (Dw) were calculated.32,47 The water contents used in the calculations were the excess water points: 47% for MO, 51% for MP, and 28% for PT.31,32,35,36 Lipid densities used in the calculations were 0.942 g cm−3 for MO, 0.982 g cm−3 for MP, and 0.940 g cm−3 for PT.

The bilayer thicknesses and water channel diameters were calculated32,47 based on the QIID lattice parameters of MO, MP, and PT without cholesterol. MP has the largest lattice parameter and water channel diameter but a slightly thinner bilayer than MO. PT has the smallest bilayer thickness and water channel diameter of the three cubic phases. Unlike MO and MP, which contain unsaturated hydrocarbon chains, PT has a highly branched isoprenoid type hydrocarbon chain, allowing us to investigate the effect of the chain architecture on protein incorporation. Such branched chains are known to induce a much higher lateral pressure in the bilayer.48 The presence of 8% w/w of cholesterol did not significantly influence the QIID lattice parameters. In previous studies, cholesterol was found to induce a small increase in lattice parameter upon addition to the monoolein cubic phase.49,50 In this study, this was not observed, potentially due to the presence of the Elugent detergent in the buffered detergent solution. The 18 combinations of protein and lipid samples were run in duplicate (top and bottom subwells within each well of the 96-well plate), and reproducibility of the phase behavior was confirmed, as described in previous work.42,51 A full list of the mesophases identified and associated lattice parameters is provided in Supporting Information S1. Plots of the lattice parameter variation between the top and bottom subwells are also provided in Supporting Information S2. Samples were found to be reproducible in more than 85% of wells. In some cases, no diffraction was observed in one or both subwells; in these cases, the data are considered irreproducible and are not considered further. We note that data were more likely to be irreproducible at the highest protein concentrations studied, consistent with the disrupting effect of the protein on the cubic

RESULTS AND DISCUSSION

Effect of Ag43, β43, and α43 Incorporation on the Bicontinuous Cubic Mesophases of MO, MP, and PT. Ag43 and its hydrophobic and hydrophilic domains, β43 and α43, were incorporated within the bicontinuous cubic phase formed by three different lipids: MO, MP, and PT. Although Ag43 is implicated in cell aggregation, we anticipate that the Ag43, β43, and α43 proteins were present in the monomeric form at the low concentrations used in this study, as was shown previously by Heras et al.23 Representative 1D diffraction patterns obtained for the diamond cubic QIID mesophases of MO, MP, and PT, each with one of the three proteins in buffer, are shown in Figure 4. The 12028

DOI: 10.1021/acs.langmuir.5b03256 Langmuir 2015, 31, 12025−12034

Article

Langmuir

Figure 5. Incorporation of Ag43 (black filled squares), β43 (blue open circles), and α43 (red half open triangles) into QIID of (A) MO, (B) MP, (C) PT, (D) MO + 8% w/w of cholesterol, (E) MP + 8% w/w of cholesterol, and (F) PT + 8% w/w of cholesterol. All data points were the average of the top and bottom subwells (see Supporting Information S1 and S2) and are shown as the percent (%) increase in lattice parameter relative to the lattice parameter of lipid (+ 8% w/w of cholesterol) with detergent solution in each plate.

the incorporation of the hydrophobic (β43) and hydrophilic (α43) domains separately within the same three lipid systems. The results obtained for the individual domains are discussed below. β43 Domain. We look first at the effect of the β43 domain. The model structure of β43 is shown in Figure 2(A). The length of the transmembrane β-barrel of Ag43 is expected to be ∼27 Å (consistent with the standard for the lipid-exposed surface of βbarrels in the outer membrane of gram-negative bacteria).27,28 Incorporation of the hydrophobic β43 domain resulted in a steady increase in lattice parameter for MO and MP (Figure 5). However, this increase is smaller than was seen upon incorporation of the full length Ag43 protein. A concentration of 1.8 × 10−3 mole % of β43 resulted in approximately a 5.5% increase in QIID lattice parameter with MO (Figure 5(A)) and around a 5% increase with MP (Figure 5(B)). The lattice parameter of PT did not increase significantly upon incorporation of β43 (0.5% increase in QIID lattice parameter at 1.8 × 10−3 mole %, Figure 5(C)). The differing effect of the β43 domain on the three cubic phases may be explained by a consideration of hydrophobic mismatch between the lipid bilayer and the lipid-exposed surface of the protein. The increase in the lattice parameter of MO and MP upon incorporation of the β43 domain is consistent with a negative hydrophobic mismatch of the β43 domain (27 Å) with the lipid bilayer (MO, 33.4 Å; MP, 32.5 Å; Table 1). The negative hydrophobic mismatch is less for MP than for MO, consistent with the smaller increase in lattice parameters observed upon incorporation of β43 into QIID of MP compared to MO. The bilayer thickness for PT is approximately 29.8 Å (Table 1); thus, in this case, there is almost no hydrophobic mismatch of the lipid-exposed β43 surface, which is approximately 27 Å thick, with the lipid bilayer. This observation is consistent with the negligible effect of β43 incorporation on QIID lattice parameters of PT. The negative hydrophobic mismatch between the hydrophobic domain length of β43 (Lβ43) and the lipid bilayer

phase. For limiting any dehydration issues with the samples, the top (A) and bottom (H) rows of the six SD-2 plates were not used. Protein concentrations were randomized within the plates to eliminate any deposition effects based on their location. All three proteins were dissolved within a buffered detergent solution containing 20 mM Tris, 150 mM NaCl, and 0.5% v/v (∼17 mM) of Elugent. The effect of the detergent solution on the three cubic phase nanostructures was minor, as shown in Table 1. A slight increase in lattice parameter relative to the lattice parameter observed in excess water was observed for MO and PT with a small decrease observed for MP. Although both salts and detergents can impact the internal cubic phase nanostructure, this typically occurs at much higher concentrations than those present within the buffer.42,51,52 This observation is consistent with the mesophase behavior of alkyl glucosides.53 Ag43. The increase in lattice parameter following incorporation of all three proteins (compared to the control samples containing lipid and detergent solution) is shown in Figure 5 for MO, MP, and PT with and without 8% w/w of cholesterol, where the protein concentrations are presented in mole %, representing the number of protein molecules present relative to the total amount of protein and lipid molecules. The incorporation of the full length Ag43 protein into the cubic phase of MO, MP, and PT resulted in an increase in lattice parameters for all lipid systems, Figure 5(A−C). The percentage increase in lattice parameter rose steadily with increasing protein concentration with no plateau observed up to the highest protein concentration studied (1.4 × 10−3 mole %). The QIID phase was retained at all protein concentrations with no coexisting phases observed. The maximum increase observed was in the range 15−21% (20.9% for MO, 14.6% for MP, and 16.5% for PT). The effect of cholesterol on the observed structural trends, shown in Figure 5(D−F), was found to be negligible and will therefore not be discussed separately. To allow us to deconvolute the effect of the hydrophobic and hydrophilic domains within the Ag43 moiety, we investigated 12029

DOI: 10.1021/acs.langmuir.5b03256 Langmuir 2015, 31, 12025−12034

Article

Langmuir

Figure 6. Incorporation of β43 (model adapted from ref 22) in QIID: (A) the hydrophobic domain length of β43 (Lβ43) is smaller than the lipid bilayer thickness (Lb), resulting in negative hydrophobic mismatch and a local thinning of Lb. (B) The hydrophobic domain length is similar to the lipid bilayer thickness, and there is a hydrophobic match.

integral membrane proteins, it was proposed that there is an effect of geometry, charge, and chemical structure and that it is currently not possible to predict individual phase transitions based on the encapsulation of the proteins in lipid mesophases.15 The results presented here support the idea that hydrophobic mismatch is an important factor in determining the effect of the hydrophobic transmembrane domains of membrane proteins on lipidic cubic mesophases once incorporated in the bilayers but that charge and diameter might also have an effect. α43 Domain. The hydrophilic β-helical rod of α43 has a section of ∼12 Å by 23 Å in size and a length of ∼100 Å (Figure 2(B)).23 In the case of the hydrophilic α43 domain, the effect of incorporation on QIID lattice parameters was also larger for MO (∼5.0% increase at 1.1 × 10−3 mole %) than for MP (∼2.5% increase at 1.1 × 10−3 mole %) and PT (∼2.5% increase at 1.1 × 10−3 mole %). We suggest that this is due to the difference in water channel diameter; MP (Dw = 61.2 Å) has larger water channels than MO (Dw = 51.8 Å), as shown in Table 1. It is expected that α43 exerts less pressure on the bilayer in the QIID phase of MP than of MO upon incorporation in the water channels, resulting in a reduced increase in the lattice parameter for this lipid. An increase in lattice parameter, which results in an increase in water channel radius, reduces geometrical mismatch between α43 and the water channels of the MO cubic phase. In the case of PT (Dw = 24.8 Å), α43 is not expected to fit inside the QIID water channels because of the large size discrepancy. The excess water point of PT is at 28% w/w of water,36 and the samples were at 57% w/w of water, which suggests that some of the α43 molecules could probably be accommodated in the excess water region. For MO and MP, there is less excess water with excess water points of 47%31,32 and 51%35 w/w, respectively, and the water channels are significantly larger, so α43 is more likely to be accommodated within the water channels. The 2.5% increase in QIID lattice parameters for α43 at 1.1 × 10−3 mole % with PT suggests that some molecules were encapsulated in the water channels of the cubic phase despite the large size discrepancy. It has been shown previously that even if the water channel diameter is smaller than the molecular dimension of a protein, they can sometimes be encapsulated within the cubic phase. In this case, there might be regions of slight disorder, referred to as nanopockets.7,11 In summary, the effect of the hydrophilic α43 domain on the lattice parameters of QIID of MO, MP, and PT could be

thickness (Lb), illustrated in Figure 6(A), is likely to cause a local thinning of the bilayer.20,27,54 This results from the lipid molecules closest to the protein being compressed to cover the hydrophobic core of the β-barrel.20 This probably results in a local flattening of the lipid membrane, corresponding to a decrease in interfacial curvature. An increase in lattice parameter, corresponding to a decrease in surface curvature, was observed for MO and MP consistent with this hypothesis. Hydrophobic matching of β43 with the lipid bilayer is illustrated in Figure 6(B). Presumably, there was almost no change in lattice parameter upon incorporation of the β43 protein in QIID of PT because of the hydrophobic match. No local differences in bilayer thickness are expected when the hydrophobic core of the protein matches the bilayer thickness. However, there might also be an effect from the alkyl chain architecture.48 The highly branched hydrocarbon chains of PT are stiffer than the unsaturated hydrocarbon chains of MO and PT and for that reason are less susceptible to changes in internal nanostructure and mesophase upon incorporation of the transmembrane protein. In previous studies, hydrophobic mismatch has also been found to be important in determining the morphology of resultant membrane structures, for example, upon incorporation of model α-helical WALP peptides within different flat lipid bilayer systems.17,55 In these systems, relatively short peptides were found to promote the formation of phases with high negative curvature: cubic (QII) or hexagonal (HII) phases with the particular phase adopted being dependent on the extent of hydrophobic mismatch. However, in this study, a decrease in negative curvature, or an increase in QIID lattice parameter, was found upon negative hydrophobic mismatch. This suggests that other factors, for example charge, are also playing a role. Even though the overall charge of the β43 protein is expected to be approximately zero in the buffer used (pI = 6.9, Tris pH = 7.5), local charges are present. For example, charge was found to be important in lidocaine incorporation in bilayers of the MO cubic phase.56 Another factor that could play a role is peptide/protein diameter; the section of the βbarrel of β43 is significantly larger than for the single α-helical WALP peptides previously studied. Hydrophobic mismatch has also been found to be important for the incorporation of integral membrane proteins and peptides in lipid mesophases.15 Nevertheless, it has been suggested that all protein−lipid systems must be screened individually, as incorporation of proteins into the cubic phase of several lipids has caused a different phase change.16,18 For 12030

DOI: 10.1021/acs.langmuir.5b03256 Langmuir 2015, 31, 12025−12034

Article

Langmuir

In summary, it was found that the size of the hydrophobic domain relative to the lipid bilayer thickness of the cubic mesophase was important for the changes in QIID lattice parameters observed. The size of the hydrophilic domain relative to the water channel radius of the cubic mesophase was also found to be an important factor influencing the trends observed. Secondary Structure of Ag43, β43, and α43 in Solution and in the Cubic Phases of MO and PT. Circular dichroism spectroscopy was used to study the effect of cubic mesophases on the secondary structure of the three Ag43 proteins. Ag43, β43, and α43 in buffered detergent solution and in the cubic phases of MO and PT were studied to investigate the effect of incorporation into bilayers with varying thickness and chain architecture and incorporation into water channels with varying diameter. CD spectra of the three proteins in 20 mM Tris and 0.5% v/v of Elugent solution are shown in Figure 7. Results of Ag43, β43,

explained by the difference in water channel size of the cubic phase nanostructure. When the water channels were too small, α43 was probably mainly accommodated in the excess water region. When the water channels were large enough to accommodate α43, the increase in QIID lattice parameters depended on the water channel size. In the literature, most studies focus on protein loading in MO cubic phases. In some cases, structural transitions were observed upon incorporation of soluble proteins in cubic phases of MO.15 The loading of soluble proteins in MO cubic mesophases has been shown to depend on the size of the protein.57 As mentioned above, it has been shown that even if the water channel diameter is smaller than the molecular dimension of a protein, they can sometimes be encapsulated within a cubic phase with regions of slight disorder.11 MO has previously been shown to be able to retain its initial cubic phase structures upon incorporation of soluble proteins of similar molecular weight (48, 66, 67, and 75 kDa) to α43 and at low concentrations similar to this study.15 The results presented here support the hypothesis that the relative sizes of the soluble protein and water channel radii of cubic mesophases are important for protein encapsulation. Deconvolution of the Effect of the Hydrophilic and Hydrophobic Domains within an Amphiphilic Protein. We have rationalized the effect of the full length Ag43 protein on the diamond cubic QIID phase of MO, MP, and PT based on the structural impact of its hydrophobic and hydrophilic domains. For all lipid systems, the effect of the full length Ag43 protein was more significant than the sum of the individual β43 and α43 domains. For MO, the β43 and α43 domains both tended to increase the lattice parameter of the QIID phase (by approximately 5% at 1.1 × 10−3 mole % protein for both domains). The full length Ag43 protein resulted in a 15% increase in lattice parameter at 1.1 × 10−3 mole %. This result suggests that for MO both the hydrophobic mismatch of the transmembrane β-barrel with the lipid bilayer and the hydrophilic mismatch of the β-helical rod with the water channel size contribute to the increase in lattice parameter of QIID upon incorporation of Ag43. For MP, a similar effect was observed with a 3.5% increase in lattice parameter for β43 and a 2.5% increase in lattice parameter for α43 at 1.1 × 10−3 mole %. Ag43 incorporation resulted in a 14% increase in lattice parameter at the same concentration. For MP with Ag43, the hydrophobic mismatch of the transmembrane β-barrel with the lipid bilayer and the hydrophilic mismatch of the hydrophilic β-helical rod with the water channel size were slightly reduced compared to MO, because of the thinner bilayers and larger water channels of MP. For PT, the β43 domain individually had a negligible effect on the QIID lattice parameters because of the hydrophobic match of the transmembrane β-barrel with the lipid bilayer. The α43 domain resulted in a 2.5% increase in lattice parameter at 1.1 × 10−3 mole %. The separate hydrophilic domain is expected to be accommodated in the excess water region as well as in slightly disordered regions of the cubic phase with larger water channels. Changes in lattice parameters of QIID of PT upon incorporation of the full length Ag43 protein (approximately 13% at 1.1 × 10−3 mole %) are significantly larger and are probably caused by the size of the hydrophilic domain, which is expected to be forced into the water channels when covalently bound to the transmembrane β43 domain. In this case, slight regions of disorder might also be formed to accommodate the hydrophilic domain in the PT QIID water channels.

Figure 7. Circular dichroism (CD) spectra of Ag43, β43, and α43 in 20 mM Tris and 0.5% v/v of Elugent. Spectra are the average of three samples, where four to six scans were averaged for each sample, and error bars are the standard deviation between the three samples. Ag43 spectra were recorded at 0.70 mg mL−1 for all three samples, β43 spectra at 0.88 mg mL−1 (one sample) and 0.82 mg mL−1 (two samples), and α43 spectra at 0.88, 0.74, and 0.57 mg mL−1.

and α43 in buffered detergent solution were similar; the spectra had a negative peak at approximately 218 nm and a positive peak near 195 nm. This is typical for spectra arising from βsheets,58 as was expected for Ag43, β43, and α43. The spectrum for the full length Ag43 protein was almost identical to the β43 spectrum from 200 to 260 nm, but from 190 to 200 nm, it was intermediate between the β43 and α43 spectra. The estimated secondary structures obtained by deconvolution of the spectra in Table 2 are also highly similar for the three proteins, indicating a mixture of secondary structures. The well-ordered secondary structure of the proteins as observed by CD in the buffered detergent solution is a positive indicator, showing that the proteins are folded appropriately in β-sheet structures with a small contribution of turns, α-helices, and unordered Table 2. Secondary Structure of Ag43, β43, and α43 in 20 mM Tris and 0.5% of Elugent as estimated from the Deconvolution of CD Spectra in Figure 7 using the DICHROWEB Online Server44 and CDSSTR Algorithm45

Ag43 β43 α43 12031

α-helix

β-strand

turn

unordered structure

0.4 0.2 0.3

0.3 0.4 0.3

0.1 0.1 0.1

0.2 0.3 0.3 DOI: 10.1021/acs.langmuir.5b03256 Langmuir 2015, 31, 12025−12034

Article

Langmuir

Figure 8. Circular dichroism spectra of Ag43, β43, and α43 in 20 mM Tris and 0.5% v/v of Elugent, QIID of PT, and QIID of MO. Spectra are the average of three samples, where four to six scans were averaged for each sample, and error bars are the standard deviation between the three samples. Solution spectra were recorded at protein concentrations as specified in Figure 7. Ag43 spectra in PT and MO were recorded at 0.71 mg mL−1; β43 spectra in PT and MO were recorded at 0.88 mg mL−1, and α43 spectra in PT and MO were recorded at 0.65 mg mL−1.

Ag43 and β43 in the PT cubic phase compared to the structures in buffered detergent solution, and the effect of these changes on function, requires techniques such as FTIR and NMR. However, much higher protein concentrations would be required than those used here. In the case of Ag43 and α43, the hydrophilic β-helix is expected to be present in the water channels of the cubic phases. Unlike SAXS measurements, which were in excess water, samples for CD measurements were at limited hydration. There is a large size discrepancy between α43 (Figure 2; 23 Å × 100 Å) and the water channels of PT (Table 1; Dw = 25 Å), and this could conceivably result in pressure on the proteins, potentially giving rise to the observed conformational change. Again, the data down to 210 nm for MO did not show any significant differences. Further investigations would be needed to determine the detailed structural changes of Ag43 and α43 upon incorporation in the PT cubic phase, as opposed to the structure in buffered detergent solution. Nevertheless, for all three proteins in this study, Ag43, β43, and α43, the secondary structures are significantly altered when present in the PT cubic phase compared to the structure in buffered detergent solution.

structures as is expected from observation of the known crystal and modeled structures (Figure 2). The CD spectra indicated that protein aggregation in solution was minimal. The CD spectra of Ag43, β43, and α43 in buffered detergent solution in Figure 7 are compared to the spectra for these same proteins in PT and MO diamond cubic QIID phases in Figure 8. For PT, spectra are shown only for 200−260 nm, where there was low absorption from the lipid. For MO, all wavelengths below approximately 210 nm were discarded due to the strong background absorption of the lipid, as described in the literature.40 The significantly higher absorption in the far UV for MO compared to PT is most likely caused by its unsaturated carbon−carbon bond (Figure 3). MP demonstrated strong absorbance in the region of interest and was excluded from analysis. For Ag43, β43, and α43 (Figure 8(A−C, respectively)), there was a slight change in conformation when present in the PT cubic phase compared to the spectra in buffered detergent solution, as shown by the differences in CD signal from 200 to 215 nm. There was little difference in the structure of the three proteins in MO compared to in PT. This could reflect the limited wavelength range that resulted in reliable data for MO or could indicate that the largest structural perturbation occurred within PT. In the case of Ag43 and β43, the hydrophobic β-barrel is expected to be embedded within the lipid bilayer of the cubic phase. In the PT diamond cubic QIID phase, the β-barrel conformation could be slightly changed compared to the solution state because of the high lateral pressure in the bilayer of this branched chain lipid.48 For MO, the data down to 210 nm did not show any significant differences in Ag43 and β43 conformation compared to PT and solution results, most likely due to the limited wavelength range. In the literature, there is also no significant change in CD signal reported for the β-barrel BtuB in detergent solution and in the MO cubic phase from 210 to 250 nm.40 It was shown for the denatured β-barrel OmpA, that it could refold in a lipid bilayer with matching hydrophobic thickness using CD.59 Transmembrane β-barrel CD spectra of these proteins were similar to the spectra reported here. In addition to changes in secondary structure, the tertiary structure of Ag43 and β43 may change upon varying hydrophobic mismatch with a lipid bilayer due to tilting of the hydrophobic domains.60 There is a significant amount of experimental evidence in the literature suggesting that the hydrophobic thickness of the bilayer and lipid bilayer curvature stress are important for membrane protein function and that they function optimally at certain bilayer thicknesses and curvatures.20 Further exploring the conformational changes of



CONCLUSIONS In this study, we illustrate the effect of incorporating the bacterial amphiphilic membrane protein Ag43, and its individual hydrophobic β43 and hydrophilic α43 domains, in bicontinuous cubic mesophases formed by monoolein, monopalmitolein, and phytantriol with and without 8% w/w of cholesterol. It was found that the size of the hydrophobic domain relative to the thickness of the lipid bilayer of the cubic mesophase may be important in relation to the observed structural changes to the lipidic cubic phase upon protein incorporation. The surface charge of the protein and its diameter may also be significant. Additionally, the size of the hydrophilic domain relative to the water channel radius of the cubic mesophase was found to be important for determining structural changes to the cubic phase. The effect of the full length Ag43 protein was more significant than the sum of the individual hydrophobic β43 and hydrophilic α43 domains. The results suggest that retention of the lipid bicontinuous cubic mesophase structure may be more likely when the geometric parameters of the lipidic cubic phase match those of the incorporated protein or peptide (hydrophobic and hydrophilic). These results will help facilitate predictions on whether amphiphilic integral membrane proteins with particular hydrophobic and hydrophilic domain sizes will be compatible with the internal nanostructure of lipidic cubic phases formed by a 12032

DOI: 10.1021/acs.langmuir.5b03256 Langmuir 2015, 31, 12025−12034

Article

Langmuir

(8) Angelov, B.; Angelova, A.; Ollivon, M.; Bourgaux, C.; Campitelli, A. Diamond-type lipid cubic phase with large water channels. J. Am. Chem. Soc. 2003, 125, 7188−7189. (9) Angelov, B.; Angelova, A.; Filippov, S. K.; Drechsler, M.; Stepanek, P.; Lesieur, S. Multicompartment Lipid Cubic Nanoparticles with High Protein Upload: Millisecond Dynamics of Formation. ACS Nano 2014, 8, 5216−5226. (10) Mulet, X.; Boyd, B. J.; Drummond, C. J. Advances in drug delivery and medical imaging using colloidal lyotropic liquid crystalline dispersions. J. Colloid Interface Sci. 2013, 393, 1−20. (11) Angelova, A.; Ollivon, M.; Campitelli, A.; Bourgaux, C. Lipid cubic phases as stable nanochannel network structures for protein biochip development: X-ray diffraction study. Langmuir 2003, 19, 6928−6935. (12) Razumas, V.; Kanapieniene, J.; Nylander, T.; Engstrom, S.; Larsson, K. Electrochemical Biosensors for Glucose, Lactate, Urea, and Creatinine Based on Enzymes Entrapped in a Cubic Liquid-Crystalline Phase. Anal. Chim. Acta 1994, 289, 155−162. (13) Nazaruk, E.; Smolinski, S.; Swatko-Ossor, M.; Ginalska, G.; Fiedurek, J.; Rogalski, J.; Bilewicz, R. Enzymatic biofuel cell based on electrodes modified with lipid liquid-crystalline cubic phases. J. Power Sources 2008, 183, 533−538. (14) Rowinski, P.; Kang, C.; Shin, H.; Heller, A. Mechanical and chemical protection of a wired enzyme oxygen cathode by a cubic phase lyotropic liquid crystal. Anal. Chem. 2007, 79, 1173−1180. (15) Conn, C. E.; Drummond, C. J. Nanostructured bicontinuous cubic lipid self-assembly materials as matrices for protein encapsulation. Soft Matter 2013, 9, 3449−3464. (16) Conn, C. E.; Darmanin, C.; Sagnella, S. M.; Mulet, X.; Greaves, T. L.; Varghese, J. N.; Drummond, C. J. Incorporation of the dopamine D2L receptor and bacteriorhodopsin within bicontinuous cubic lipid phases. 2. Relevance to in meso crystallization of integral membrane proteins in novel lipid systems. Soft Matter 2010, 6, 4838− 4846. (17) Siegel, D. P.; Cherezov, V.; Greathouse, D. V.; Koeppe, R. E.; Killian, J. A.; Caffrey, M. Transmembrane peptides stabilize inverted cubic phases in a biphasic length-dependent manner: Implications for protein-induced membrane fusion. Biophys. J. 2006, 90, 200−211. (18) Liang, Y. L.; Conn, C. E.; Drummond, C. J.; Darmanin, C. Uptake of the butyrate receptors, GPR41 and GPR43, in lipidic bicontinuous cubic phases suitable for in meso crystallization. J. Colloid Interface Sci. 2015, 441, 78−84. (19) Cantor, R. S. Lateral pressures in cell membranes: A mechanism for modulation of protein function. J. Phys. Chem. B 1997, 101, 1723− 1725. (20) Mouritsen, O. G. Membrane Protein-Lipid Match and Mismatch. Comprehensive Biophysics, Vol 5: Membranes 2011, 245− 260. (21) Shen, H. H.; Leyton, D. L.; Shiota, T.; Belousoff, M. J.; Noinaj, N.; Lu, J. X.; Holt, S. A.; Tan, K.; Selkrig, J.; Webb, C. T.; Buchanan, S. K.; Martin, L. L.; Lithgow, T. Reconstitution of a nanomachine driving the assembly of proteins into bacterial outer membranes. Nat. Commun. 2014, 5, 5078. (22) van der Woude, M. W.; Henderson, I. R. Regulation and Function of Ag43 (Flu). Annu. Rev. Microbiol. 2008, 62, 153−169. (23) Heras, B.; Totsika, M.; Peters, K. M.; Paxman, J. J.; Gee, C. L.; Jarrott, R. J.; Perugini, M. A.; Whitten, A. E.; Schembri, M. A. The antigen 43 structure reveals a molecular Velcro-like mechanism of autotransporter-mediated bacterial clumping. Proc. Natl. Acad. Sci. U. S. A. 2014, 111, 457−462. (24) Ulett, G. C.; Valle, J.; Beloin, C.; Sherlock, O.; Ghigo, J. M.; Schembri, M. A. Functional analysis of antigen 43 in uropathogenic Escherichia coli reveals a role in long-term persistence in the urinary tract. Infect. Immun. 2007, 75, 3233−3244. (25) Henderson, I. R.; Navarro-Garcia, F.; Desvaux, M.; Fernandez, R. C.; Ala’Aldeen, D. Type V protein secretion pathway: the autotransporter story. Microbiol Mol. Biol. R 2004, 68, 692. (26) Caffrey, P.; Owen, P. Purification and N-Terminal Sequence of the Alpha-Subunit of Antigen-43, a Unique Protein Complex

range of lipids. The hydrophobic thickness and physicochemical properties of the lipid bilayer and the water channel diameter of the cubic phase were also shown to affect the secondary structure of the hydrophobic β43 and hydrophilic α43 domains of the Ag43 protein. Although such structural changes may be small, they could potentially affect the function of the membrane protein and impact the use of the hybrid protein− lipid material for end applications.



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.langmuir.5b03256. Tables showing (1) all phases and lattice parameters (lp) for MO, MP, and PT with and without 8% w/w of cholesterol with Ag43, α43, and β43 at all protein concentrations studied and (2) reproducibility of SAXS results (PDF)



AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS We thank Dr. Stephen T. Mudie and Dr. Nigel M. Kirby for their assistance with SAXS experiments, and we acknowledge use of the SAXS/WAXS beamline at the Australian Synchrotron. We thank Dr. Janet Newman and Dr. Shane A. Seabrook for help with the robotic setup of plates at the C3 Collaborative Crystallization Centre, CSIRO, Parkville, Australia. We acknowledge Dr. Yee-Foong Mok for help with CD measurements at the Bio21 Molecular Science and Biotechnology Institute, The University of Melbourne, Parkville, Australia. Sally L. Gras is supported by the ARC Dairy Innovation Hub IH120100005.



REFERENCES

(1) Doerr, A. Membrane protein structures. Nat. Methods 2009, 6, 35−35. (2) Wallin, E.; von Heijne, G. Genome-wide analysis of integral membrane proteins from eubacterial, archaean, and eukaryotic organisms. Protein Sci. 1998, 7, 1029−1038. (3) Landau, E. M.; Rosenbusch, J. P. Lipidic cubic phases: A novel concept for the crystallization of membrane proteins. Proc. Natl. Acad. Sci. U. S. A. 1996, 93, 14532−14535. (4) Seddon, A. M.; Curnow, P.; Booth, P. J. Membrane proteins, lipids and detergents: not just a soap opera. Biochim. Biophys. Acta, Biomembr. 2004, 1666, 105−117. (5) Almsherqi, Z. A.; Landh, T.; Kohlwein, S. D.; Deng, Y. R. Cubic Membranes: The Missing Dimension of Cell Membrane Organization. Int. Rev. Cell Mol. Biol. 2009, 274, 275−342. (6) Luzzati, V.; Tardieu, A.; Gulikkrz, T.; Rivas, E.; Reisshus, F. Structure of Cubic Phases of Lipid-Water Systems. Nature 1968, 220, 485−488. (7) Angelova, A.; Angelov, B.; Mutafchieva, R.; Lesieur, S.; Couvreur, P. Self-Assembled Multicompartment Liquid Crystalline Lipid Carriers for Protein, Peptide, and Nucleic Acid Drug Delivery. Acc. Chem. Res. 2011, 44, 147−156. 12033

DOI: 10.1021/acs.langmuir.5b03256 Langmuir 2015, 31, 12025−12034

Article

Langmuir Associated with the Outer-Membrane of Escherichia-Coli. J. Bacteriol. 1989, 171, 3634−3640. (27) Tamm, L. K.; Hong, H.; Liang, B. Y. Folding and assembly of beta-barrel membrane proteins. Biochim. Biophys. Acta, Biomembr. 2004, 1666, 250−263. (28) Wimley, W. C. The versatile beta-barrel membrane protein. Curr. Opin. Struct. Biol. 2003, 13, 404−411. (29) Meng, G. Y.; Spahich, N.; Kenjale, R.; Waksman, G., St; Geme, J. W. Crystal structure of the Haemophilus influenzae Hap adhesin reveals an intercellular oligomerization mechanism for bacterial aggregation. EMBO J. 2011, 30, 3864−3874. (30) Kulkarni, C. V.; Wachter, W.; Iglesias-Salto, G.; Engelskirchen, S.; Ahualli, S. Monoolein: a magic lipid? Phys. Chem. Chem. Phys. 2011, 13, 3004−3021. (31) Qiu, H.; Caffrey, M. The phase diagram of the monoolein/water system: metastability and equilibrium aspects. Biomaterials 2000, 21, 223−234. (32) 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. (33) Li, D. F.; Lee, J.; Caffrey, M. Crystallizing Membrane Proteins in Lipidic Mesophases. A Host Lipid Screen. Cryst. Growth Des. 2011, 11, 530−537. (34) Zabara, A.; Mezzenga, R. Plenty of room to crystallize: swollen lipidic mesophases for improved and controlled in-meso protein crystallization. Soft Matter 2012, 8, 6535−6541. (35) Briggs J. Phase behaviour of hydrated monoacylglycerols and the design of an X-ray compatible scanning calorimeter. Ph.D. Thesis, The Ohio State University, Columbus, OH, 1994; pp 174−195. (36) Barauskas, J.; Landh, T. Phase behavior of the phytantriol/water system. Langmuir 2003, 19, 9562−9565. (37) Caffrey, M.; Li, D. F.; Dukkipati, A. Membrane Protein Structure Determination Using Crystallography and Lipidic Mesophases: Recent Advances and Successes. Biochemistry 2012, 51, 6266− 6288. (38) Darmanin, C.; Conn, C. E.; Newman, J.; Mulet, X.; Seabrook, S. A.; Liang, Y. L.; Hawley, A.; Kirby, N.; Varghese, J. N.; Drummond, C. J. High-Throughput Production and Structural Characterization of Libraries of Self-Assembly Lipidic Cubic Phase Materials. ACS Comb. Sci. 2012, 14, 247−252. (39) Angelov, B.; Angelova, A.; Papahadjopoulos-Sternberg, B.; Hoffmann, S. V.; Nicolas, V.; Lesieur, S. Protein-containing PEGylated cubosomic particles: freeze-fracture electron microscopy and synchrotron radiation circular dichroism study. J. Phys. Chem. B 2012, 116, 7676−7686. (40) Caffrey, M.; Cherezov, V. Crystallizing membrane proteins using lipidic mesophases. Nat. Protoc. 2009, 4, 706−731. (41) Kirby, N. M.; Mudie, S. T.; Hawley, A. M.; Cookson, D. J.; Mertens, H. D. T.; Cowieson, N.; Samardzic-Boban, V. A lowbackground-intensity focusing small-angle X-ray scattering undulator beamline. J. Appl. Crystallogr. 2013, 46, 1670−1680. (42) van ’t Hag, L.; Darmanin, C.; Le, T.; Mudie, S.; Conn, C. E.; Drummond, C. J. In Meso Crystallization: Compatibility of Different Lipid Bicontinuous Cubic Mesophases with the Cubic Crystallization Screen in Aqueous Solution. Cryst. Growth Des. 2014, 14, 1771−1781. (43) Seddon, J. M.; Squires, A. M.; Conn, C. E.; Ces, O.; Heron, A. J.; Mulet, X.; Shearman, G. C.; Templer, R. H. Pressure-jump X-ray studies of liquid crystal transitions in lipids. Philos. Trans. R. Soc., A 2006, 364, 2635−2655. (44) Whitmore, L.; Wallace, B. A. DICHROWEB, an online server for protein secondary structure analyses from circular dichroism spectroscopic data. Nucleic Acids Res. 2004, 32, W668−W673. (45) Compton, L. A.; Johnson, W. C. Analysis of Protein CircularDichroism Spectra for Secondary Structure Using a Simple Matrix Multiplication. Anal. Biochem. 1986, 155, 155−167. (46) Lees, J. G.; Miles, A. J.; Wien, F.; Wallace, B. A. A reference database for circular dichroism spectroscopy covering fold and secondary structure space. Bioinformatics 2006, 22, 1955−1962.

(47) Gulik, A.; Luzzati, V.; Derosa, M.; Gambacorta, A. Structure and Polymorphism of Bipolar Isopranyl Ether Lipids from Archaebacteria. J. Mol. Biol. 1985, 182, 131−149. (48) Polak, A.; Tarek, M.; Tomsic, M.; Valant, J.; Ulrih, N. P.; Jamnik, A.; Kramar, P.; Miklavcic, D. Structural Properties of Archaeal Lipid Bilayers: Small-Angle X-ray Scattering and Molecular Dynamics Simulation Study. Langmuir 2014, 30, 8308−8315. (49) Cherezov, V.; Clogston, J.; Misquitta, Y.; Abdel-Gawad, W.; Caffrey, M. Membrane protein crystallization in meso: Lipid typetailoring of the cubic phase. Biophys. J. 2002, 83, 3393−3407. (50) Tyler, A. I. I.; Barriga, H. M. G.; Parsons, E. S.; McCarthy, N. L. C.; Ces, O.; Law, R. V.; Seddon, J. M.; Brooks, N. J. Electrostatic swelling of bicontinuous cubic lipid phases. Soft Matter 2015, 11, 3279−3286. (51) Conn, C. E.; Darmanin, C.; Mulet, X.; Hawley, A.; Drummond, C. J. Effect of lipid architecture on cubic phase susceptibility to crystallisation screens. Soft Matter 2012, 8, 6884−6896. (52) Misquitta, Y.; Caffrey, M. Detergents destabilize the cubic phase of monoolein: Implications for membrane protein crystallization. Biophys. J. 2003, 85, 3084−3096. (53) Boyd, B. J.; Drummond, C. J.; Krodkiewska, I.; Grieser, F. How chain length, headgroup polymerization, and anomeric configuration govern the thermotropic and lyotropic liquid crystalline phase behavior and the air-water interfacial adsorption of glucose-based surfactants. Langmuir 2000, 16, 7359−7367. (54) Killian, J. A.; Nyholm, T. K. M. Peptides in lipid bilayers: the power of simple models. Curr. Opin. Struct. Biol. 2006, 16, 473−479. (55) Killian, J. A.; Salemink, I.; dePlanque, M. R. R.; Lindblom, G.; Koeppe, R. E.; Greathouse, D. V. Induction of nonbilayer structures in diacylphosphatidylcholine model membranes by transmembrane alpha-helical peptides: Importance of hydrophobic mismatch and proposed role of tryptophans. Biochemistry 1996, 35, 1037−1045. (56) Engstrom, S.; Engstrom, L. Phase-Behavior of the LidocaineMonoolein-Water System. Int. J. Pharm. 1992, 79, 113−122. (57) Clogston, J.; Caffrey, M. Controlling release from the lipidic cubic phase. Amino acids, peptides, proteins and nucleic acids. J. Controlled Release 2005, 107, 97−111. (58) Greenfield, N. J. Using circular dichroism spectra to estimate protein secondary structure. Nat. Protoc 2006, 1, 2876−2890. (59) Kleinschmidt, J. H.; Tamm, L. K. Secondary and tertiary structure formation of the beta-barrel membrane protein OmpA is synchronized and depends on membrane thickness. J. Mol. Biol. 2002, 324, 319−330. (60) Tamarit, B.; Bugault, F.; Pillet, A. H.; Lavergne, V.; Bochet, P.; Garin, N.; Schwarz, U.; Theze, J.; Rose, T. Membrane microdomains and cytoskeleton organization shape and regulate the IL-7 receptor signalosome in human CD4 T-cells. J. Biol. Chem. 2013, 288, 8691− 8701.

12034

DOI: 10.1021/acs.langmuir.5b03256 Langmuir 2015, 31, 12025−12034