Skip to main content
NCBI home page
Acta Crystallographica Section F: Structural Biology Communications logoLink to Acta Crystallographica Section F: Structural Biology Communications
. 2015 Jun 27;71(Pt 7):838–846. doi: 10.1107/S2053230X15009516

Use of dynamic light scattering and small-angle X-ray scattering to characterize new surfactants in solution conditions for membrane-protein crystallization

Mohamed Dahani a, Laurie-Anne Barret a,b, Simon Raynal a, Colette Jungas b,‡, Pétra Pernot c, Ange Polidori a, Françoise Bonneté a,*
PMCID: PMC4498704  PMID: 26144228

The influence of fluorinated segment length on surfactant micelle form factors, intermicellar interaction parameters and the surfactant phase diagram in solution conditions for membrane-protein crystallization is reported.

Keywords: membrane protein, surfactant micelle, scattering techniques, phase diagram, crystallization

Abstract

The structural and interactive properties of two novel hemifluorinated surfactants, F2H9-β-M and F4H5-β-M, the syntheses of which were based on the structure and hydrophobicity of the well known dodecyl-β-maltoside (DD-β-M), are described. The shape of their micellar assemblies was characterized by small-angle X-ray scattering and their intermicellar inter­actions in crystallizing conditions were measured by dynamic light scattering. Such information is essential for surfactant phase-diagram determination and membrane-protein crystallization.

1. Introduction  

The crystallization of membrane proteins (MPs) for structure resolution at the atomic level is a challenge in biocrystallography. MP structures represent less than 1% of the more than 100 000 structures deposited in the PDB. This deficit is caused by technically challenging steps: (i) large-scale MP production, (ii) the solubilization and purification of functionally stable MP complexes and (iii) the generation of crystals that diffract X-rays to high resolution. Without minimizing all of the previous essential biochemical steps, the crystallization of membrane proteins is a challenging task because amphiphilic molecules (namely surfactants) are ubiquitous in solution with membrane proteins, both bound to the protein and in micelle forms, which makes MP phase diagrams difficult to control (Nollert, 2005 ▸). Whereas well diffracting crystals can be obtained using the in meso crystallization method, i.e. by reconstitution in a lipidic environment, growing MP crystals directly in surfo, i.e. in micelles, remains the first and the most successful method used by structural biologists. Unfortunately, this method often results in poorly diffracting crystals. The difficulties in obtaining crystals that diffract X-rays to high resolution via the in surfo method can be attributed to different factors such as the instability of membrane proteins in detergent1 and the sparsity of polar residues present on the surface of MPs, resulting in an insufficient number of crystalline lattice contacts. To overcome these difficulties, a critical issue to consider is the selection of a detergent or other surfactant that will maintain the target MP in a stable, active and non-aggregated state. Stability is sometimes achieved with long-chain detergents, which form large micelles that are able to cover the hydrophobic transmembrane domain (Tate, 2010 ▸). For example, large micelle-forming detergents such as n-dodecyl-β-d-maltoside (DD-β-M) and nonaoxyethylene dodecyl ether (C12E9) are more likely to maintain an MP in solution by efficient shielding of the apolar surface of the MP. However, the large micellar size results in partial obstruction of the polar surface residues of the protein, thereby preventing the protein–protein interactions that are essential for crystal lattice formation. In contrast, small micelle detergents such as n-octyl-β-d-glucoside (β-OG) and n-octyl-β-d-maltoside (β-OM) leave more of the polar surface residues of the protein exposed to form the protein–protein contacts that are necessary for an ordered crystalline lattice. However, small micelle detergents may also inactivate MPs by intrusion of the detergent alkyl chain into the interior of the protein and/or by stripping away stabilizing lipids, cofactors or subunits. It is thus clear that detergent packing will significantly contribute to how complexes interact and self-assemble during phase transitions and how MPs are arranged within the crystalline lattice (Pebay-Peyroula et al., 1995 ▸; Penel et al., 1998 ▸). To overcome dissociative or destabilizing effects of detergents vis-à-vis MPs, different strategies have been developed over two decades to synthesize new surfactants (including in vitro synthesis; Park et al., 2007 ▸; Nehmé et al., 2010 ▸) to address the solubilization and purification (Duval-Terrié et al., 2003 ▸; Yu et al., 2000 ▸), trapping and stabilization (Chae et al., 2012 ▸; Zhang et al., 2011 ▸; Chae, Gotfryd et al., 2010 ▸) of membrane proteins. Most of these new surfactants are valuable for solubilizing and stabilizing MPs, while the number of effective surfactants for crystallization and X-ray crystallography is rather limited. Among these new surfactants, maltose-neopentyl glycol-3 (MNG-3) enabled the crystallization of cytochrome b 6 f to produce crystals that diffracted to a resolution similar to those obtained using the more usual DD-β-M (Chae, Rasmussen et al., 2010 ▸). More recently, three-dimensional crystals of different families of MPs were obtained using steroid-based facial amphiphiles either alone or mixed with detergents or lipids (Lee et al., 2013 ▸). However, crystallization successes are still rare because MP crystallization in surfactants is often considered to be a ‘black art’ in that the protocol involves screening several hundred potential conditions using a high-throughput methodology. While such screening strategies may sometimes lead to successful crystallization, the design of more efficient molecules requires understanding of the relationship between the amphiphile structure, the auto-assembly mechanism, the surfactant–protein interactions and the ability to promote availability of the lattice contacts required to crystallize MPs. In this context, we characterized the assembly properties of a new rigid surfactant, trans-propyl(bi)cyclohexyl-α-maltoside (PCC-α-M; Glycon), which was synthesized and described for MP stabilization and crystallization (Hovers et al., 2011 ▸). In terms of the shape and size of the micelles, we showed that whereas DD-β-M and PCC-α-M present quite similar ellipsoidal shapes and sizes, the aggregation number (N agg) is larger for PCC-α-M (N agg ≃ 165) than for DD-β-M (N agg ≃ 125) (Barret et al., 2013 ▸). Moreover, the study of intermicellar interactions via measurement of the second virial coefficient (A 2) for rational crystallization showed PCC-α-M to be more attractive than DD-β-M, inducing a lower cloud-point boundary for PCC-α-M than for DD-β-M. Although the solubility of RC-LH1-pufX in this new surfactant was found to be lower and crystallization was found to be easier (Barret et al., 2013 ▸), the diffraction quality was not increased sufficiently to make structure resolution possible, which is most likely to be the result of the detergent ‘belt’ surrounding the protein being too large (which is related to the higher aggregation number of PCC-α-M), therefore causing steric interference in the crystalline lattice. In an effort to improve MP crystallization and crystal packing, we synthesized new surfactants bearing a fluorinated segment on the hydrophobic chain. Fluorosurfactants have long been known to stabilize MPs without denaturation (Breyton et al., 2004 ▸), being less intrusive and less delipidating than hydrogenated surfactants. Indeed, fluorocarbons are not miscible with water nor with hydrocarbons and are more bulky and rigid than hydrogenated surfactants (Barthélémy et al., 2002 ▸). Therefore, to form small globular micelles (Polidori et al., 2006 ▸; Breyton et al., 2009 ▸) suitable for MP crystal growth, the length of the fluorinated segment must be adapted to the polar head size to reach an appropriate critical packing parameter (Israelachvili et al., 1977 ▸). Based on the structure and hydrophobicity of DD-β-M, while supposing the rule that 1 CF2 ≃ 1.5 CH2 (Shinoda et al., 1972 ▸; Ravey et al., 1988 ▸), we have synthesized two new fluorinated surfactants, F2H9-β-M and F4H5-β-M (i.e. nine or five hydrogenated C atoms and two or four fluorinated C atoms, respectively, at the end of the chain). The physico-chemical characteristics of micelle formation and their biochemical evaluation for the stabilization of membrane proteins have been thoroughly described by Polidori et al. (2015 ▸). In the present paper, we focus on the characterization of their micelle structures and intermicellar properties that may help in membrane-protein crystallization.

2. Materials and methods  

2.1. Solutions for experiments  

Dodecyl-β-maltoside (DD-β-M; <0.2% α anomer) and octyl fluorinated-β-maltoside (F6H2-β-M; <2% α anomer) were purchased from Anatrace. PCC-α-M was synthesized as described previously (Hovers et al., 2011 ▸). F2H9-β-M and F4H5-β-M were synthesized at IBMM/CBSA, Avignon as described elsewhere (Polidori et al., 2015 ▸). All surfactant stock solutions (between 50 and 100 mg ml−1 in water) were prepared 24 h prior to measurements in Milli-Q water filtered using 0.2 µM Millipore filters followed by centrifugation at 15 000 rev min−1 for 2 h prior to transfer to clean measurement cells. For surfactant phase-diagram determination, monodisperse polyethylene glycol 3350 [50%(w/v) solution in water] was purchased from Hampton Research.

2.2. Dynamic light scattering (DLS)  

Dynamic light-scattering experiments were performed at 20°C on a Zetasizer Nano-S model 1600 (Malvern Instruments, UK) equipped with an He–Ne laser (λ = 633 nm, 4.0 mW) at an angle of 173° (backscattering detection) in a 45 µl low-volume quartz cuvette. A total of five scans of 5 s duration were accumulated for each sample. The time-dependent correlation function G(τ) was measured for different concentrations of surfactants above their respective CMC in water using different percentages of PEG 3350 and was analysed either in cumulative mode if the distribution was monomodal [i.e. polydispersity index (PdI) < 20%] or in CONTIN mode if the distribution was multimodal (PdI > 20%) using the integrated Zetasizer software. The diffusion coefficient D t was plotted as a function of surfactant concentration in order to obtain (i) the hydrodynamic radius (R h) of the micelles from the intercept using the Stokes–Einstein equation D 0 = k B T/6πηR h, where k B is Boltzmann’s constant, T is the absolute temperature and η is the viscosity of the solvent, and (ii) the interaction parameter k D (Li et al., 2004 ▸) from the slope of D t = D 0[1 + k D(c − CMC)]. As for the second virial coefficient A 2 obtained from static light scattering or small-angle X-ray scattering, in the case where the micelle shape is independent of surfactant concentration, if k D is positive the interactions between micelles are repulsive and if it is negative they are attractive.

2.3. Small-angle X-ray scattering  

Micelle form factors of fluorosurfactants were characterized in water by small-angle X-ray scattering on the bioSAXS beamline ID14-eh3 and subsequently on BM29 (Pernot et al., 2013 ▸) at the European Synchrotron Radiation Facility, Grenoble, France. Surfactant scattering patterns were measured at 20°C at several concentrations ranging from 2.5 to 40 mg ml−1 in H2O. With a sample-to-detector distance of 2.425 m and an X-ray wavelength of 0.0931 nm, the achievable q-range was 0.05–4 nm−1. To prevent radiation damage during the scattering experiments, data were collected in ten successive 2 s frames and the solution flowed in the capillary during exposure. Averaged scattered intensities were water-scattering subtracted and normalized to the surfactant concentration. Forward scattering values (i.e. q→0) I(c, 0)/c and radii of gyration R g were evaluated using the Guinier approximation I(c, q) = I(0)exp(−q 2 R g 2/3) assuming that qR g < 1 at very small angles. The micelle molar mass and aggregation number of the surfactants were then calculated from the absolute forward intensity normalized to a reference of pure water (Orthaber et al., 2000 ▸). The aggregation number N agg was determined by dividing the micelle mass by that of the surfactant monomer,

2.3.

where N a is Avogadro’s number, r 0 is the classical electron radius (r 0 = 0.28179 × 10−12 cm e−1), Inline graphic is the measured surfactant specific volume (in cm3 g−1), M surf is the surfactant molar mass, ρsurf and ρ° are the scattering-length densities of the surfactant and water (in e− cm−3), respectively, and Inline graphic is the absolute scattering intensity of water, which is equal to 0.01632 cm−1 at 20°C.

The maximum particle dimension, D max, and the pair distribution function, P(r), were finally determined by inverse Fourier transformation using the program GNOM (Svergun, 1992 ▸) to evaluate the geometry of the micelles.

2.4. Density and partial specific volume measurements  

Surfactant specific volume (in cm3 g−1) was calculated from the precise measurement at 20°C (using an Anton-Paar DMA4500M density meter) of ρ and ρ°, which are the densities of surfactant solutions at different concentrations ranging from 0.2 to 10 mg ml−1 and of H2O, respectively. The surfactant stock solutions were prepared by precisely weighting both the surfactant and the solvent at about 10 mg ml−1. Surfactant solutions were obtained by successive dilutions from stock solutions. The specific volume was then obtained from the slope of the equation

2.4.

2.5. Cloud-point boundary  

Surfactant phase diagrams were determined by optical microscopy visualization (Carl Zeiss SteREO Discovery.V12 microscope). 10 µl droplets of a mixture of 2–10% PEG 3350 and 5–50 mg ml−1 surfactant in water were deposited in microbatch plates and immediately observed. For each PEG concentration, the surfactant concentration was increased by 1 mg ml−1 stepwise until a phase transition was observed.

3. Results and discussion  

Two new fluorinated surfactants, i.e. F2H9-β-M and F4H5-β-M, derived from the well known DD-β-M, bearing either a short perfluoroethyl or perfluorobutyl tip at the end of the aliphatic chain, are presented in Fig. 1 ▸. Their micellar properties in terms of CMC (Table 1 ▸), micellization behaviour, micelle homogeneity and biochemical evaluation for the stabilization of membrane proteins have been thoroughly described and compared with those of both fully hydrogenated DD-β-M and perfluorinated F6H2-β-M (Polidori et al., 2015 ▸). Here, we evaluate their interest for membrane-protein crystallization by describing their micellar assemblies and interactive properties. Since the pioneering work of Loll and coworkers (Hitscherich et al., 2000 ▸), it is now recognized that the behaviour of protein-free micelles, in particular the existence of attractive inter­actions between micelles and the location of the surfactant phase boundary in the phase diagram, is a good predictor of the crystallization of protein–surfactant complexes (PSCs; Berger et al., 2006 ▸; Hitscherich et al., 2001 ▸; Koszelak-Rosenblum et al., 2009 ▸; Loll et al., 2001 ▸). However, the growth of PSC crystals suitable for X-ray diffraction using the in surfo method depends not only on stable monodisperse PSCs but also on small homogeneous surfactant micelles that favour protein–protein contacts in the crystal lattice. We have therefore characterized the sizes, shapes and the interactions between micelles of the two novel fluorosurfactants by small-angle X-ray scattering and dynamic light scattering in water and in crystallization conditions that could be helpful for MP crystallization.

Figure 1.

Figure 1

Open in a new tab

Chemical structures of surfactants: DD-β-M (a), F2H9-β-M (b), F4H5-β-M (c) and F6H2-β-M (d).

Table 1. Surfactant parameters.

Surfactant (No. C/eq. C)† Formula Formula weight (Da) CMC‡ (mMgl1) Inline graphic (mlg1) surf (ecm3)
PCC--M (15C) C27H48O11 548.66 0.036/0.02 0.799 0.409
DD--M (12C) C24H46O11 510.60 0.17/0.08 0.819 0.400
F2H9--M (11C/12C) C23H39 F5O11 586.54 1.14/0.67 0.749 0.425
F4H5--M (9C/11C) C21H31F9O11 630.45 2.16/1.36 0.632 0.493
F6H2--M (8C/11C) C20H25F13O11 688.40 0.71/0.49 0.578 0.529
Open in a new tab
†

No. C corresponds to the number of carbons on the chain and eq. C corresponds to the number of C assuming that 1CF2 = 1.5CH2.

‡

From surface tensiometry (Polidori et al., 2015 ▸).

3.1. Fluorinated surfactant behaviour in water  

3.1.1. Dynamic light scattering  

Dynamic light scattering (DLS) is routinely used in crystallization to detect aggregates in solution, to evaluate the polydispersity of solutions and to determine the size of particles or macromolecules in solution. It can be also used to characterize weak interactions between particles (Li et al., 2004 ▸) prior to crystallization.

We have performed dynamic light-scattering experiments on the two new fluorosurfactants F2H9-β-M and F4H5-β-M in H2O at different concentrations and compared then with the hydrogenated DD-β-M and the perfluorinated F6H2-β-M. Fig. 2 ▸ shows examples of the correlation function (Fig. 2 ▸ a), the size distribution in intensity (Fig. 2 ▸ b) and the size distribution in volume (Fig. 2 ▸ c) for the three fluorosurfactants at about 20× CMC.

Figure 2.

Figure 2

Open in a new tab

Autocorrelation functions and size distributions of F2H9-β-M, F4H5-β-M and F6H6-β-M at 20× CMC in water.

Although large aggregates (>100 nm) appear in the size distribution in intensity for F2H9-β-M and F4H5-β-M, their contribution (intensity ∝ R 6) is negligible as seen from the distribution in volume (<1%). The hydrodynamic radius R h of the fluorosurfactant micelles was thus determined by extrapolation to zero concentration of D t = f(c − CMC) (Fig. 3 ▸) for the major distribution peak. We observe that R h increases as the length of the fluorinated segment increases (inserted table in Fig. 3 ▸), although the total number of C atoms (i.e. the chain length) decreases. This is the opposite of what is observed with alkyl chains (Meyer et al., 2015 ▸; Oliver et al., 2013 ▸), where the micelle size increases as the chain length increases.

Figure 3.

Figure 3

Open in a new tab

The diffusion coefficient as a function of concentration for the four surfactants in water.

The micelle size of fluoro­surfactants seems to be influenced by the length of the fluorinated segment. With a small segment, F2H9-β-M behaves as an alkyl chain with 11 C atoms. Micelles of F2H9-β-M are smaller than micelles of DD-β-M (H12-β-M), whereas for F4H5-β-M and F6H2-β-M the micelle size seems to be more influenced by the rigidity and bulkiness of the fluorinated segment than by the length of the chain. With a long fluorinated chain and a linear maltoside head, F6H2-β-M forms elongated micelles as expected from the packing parameter (Israelachvili et al., 1977 ▸) and as recently described (Frotscher et al., 2015 ▸). The only particular behaviour could be for F4H5-β-M, which is composed of half hydrogenated and half fluorinated C atoms and for which the immiscibility of fluorine and hydrogen could confer instability in micelle assembly. This behaviour has been thoroughly described elsewhere in terms of micellization thermodynamics (Polidori et al., 2015 ▸). However, for each surfactant the diffusion coefficient D t decreases as the surfactant concentration increases, which could be attributed either to attractive intermicellar inter­actions (k D < 0) or to an increase in micelle size. To discriminate between the two behaviours, SAXS experiments were performed on each surfactant in water as a function of concentration.

3.1.2. Small-angle X-ray scattering  

SAXS is a suitable tool to characterize the structure and form factors of macromolecules and particles in solution. To discriminate between interactions between micelles and the change in micelle size as observed by DLS, SAXS experiments were performed with each surfactant in water as a function of concentration. Fig. 4 ▸ depicts the SAXS patterns for F2H9-β-M, F4H5-β-M and F6H2-β-M, the forward intensities for determination of molar mass and aggregation number and the pair distribution function, P(r), as a function of surfactant concentration. The structural parameters (R g, D max and N agg) obtained from SAXS analysis are presented in Table 2 ▸.

Figure 4.

Figure 4

Open in a new tab

SAXS experiments in water. Left column, SAXS patterns; middle column, normalized forward intensities as a function of surfactant concentration; right column, pair distribution functions. Top row, F2H9-β-M; middle row, F4H5-β-M; bottom row, F6H2-β-M.

Table 2. Structural characteristics of surfactant micelles from SAXS and DLS.
  Data from SAXS Data from DLS
Surfactant Concentration (gl1) I(0)/c CMC (cm2g) R g (nm) D max (nm) L = (12R g/R c)1/2 (nm) N agg Concentration (gl1) D t (2s1) PdI
DD--M† 2.5 24.7 3.2 8.0   125 10 60.2 0.13
            20 61.9 0.09
            30 60.9 0.08
            40 59.9 0.11
            50 60.1 0.07
            70 56.3 0.17
            100 57.2 0.04
F2H9--M 2.5 21.41 2.57 8.9   60 20 73.6 0.60
5 22.93 2.54 7.3   65 30 64.7 0.70
10 23.22 2.50 7.6   65 40 62.8 0.66
20 23.84 2.48 6.5   67 50 57.9 0.55
40 20.53 2.37 8.2   58      
F4H5--M 2.5 31.75 6.17 17.2   38 12 47.1 0.89
  5 25.93 3.37 10.2   31 27 43.9 1.00
  10 33.10 3.22 10.8   40 41 42.3 1.00
  20 36.47 3.54 10.1   44 54 40.1 0.99
  40 43.43 3.79 10.9   52 85 37.7 0.95
              100 35.9 0.18
F6H2--M 1.5 728.82 12.38 NA 42.9/1.86 625 13 13.0 0.19
3.1 936.62 13.78 55 47.7/1.66 803 28 12.9 0.21
6.2 945.76 15.21 56 52.7/1.64 811 35 12.2 0.22
12.5 903.22 15.81 57 54.8/1.63 774 42 11.7 0.21
25 746.43 14.76 59 51.1/1.62 640 50 11.2 0.24
50 530.67 12.91 58.6 44.7/1.58 455 70 10.2 0.30
Open in a new tab
†

From Barret et al. (2013 ▸).

F2H9-β-M and F6H2-β-M present a stable shape at concentrations above 6–8× CMC, as seen over a large q-range (q > 0.5 nm−1). While F2H9-β-M seems to form small almost globular micelles with N agg ≃ 65 (i.e. molar mass ≃ 38 000 ± 1000 g mol−1), F6H2-β-M forms large elongated (rod-like) micelles from the log–log representation with an N agg of greater than 800 (i.e. molar mass ≃ 550 000 ± 5000 g mol−1). In contrast, F4H5-β-M exhibits an evolution in the SAXS curves probably owing to an evolution in micellization as observed by DLS. The molar masses of the micelles are found to be between 20 000 and 32 000 g mol−1. The maximum distances, D max, for the three fluorosurfactants obtained from the SAXS pair distribution function (Figs. 4 ▸ g, 4 ▸ h and 4 ▸ i) are in agreement with the values of the hydrodynamic radius obtained from DLS experiments. Both F2H9-β-M and F6H2-β-M present overall repulsive interactions, as seen from the decrease in normalized forward intensity as a function of concentration (Figs. 4 ▸ d and 4 ▸ f) when the micelle shape is stable. The second virial coefficients A 2 were characterized. We thus found that A 2 is +0.7 × 10−4 mol ml g−2 for F2H9-β-M and +0.08 × 10−4 mol ml g−2 for F6H2-β-M in water, with the decrease in the repulsive contribution of the second virial coefficient being owing to an increase in the micelle size (Tanford, 1961 ▸). The change observed in D t for F2H9-β-M and F6H2-β-M can therefore be attributed to interactions between micelles, as was observed for DD-β-M (Barret et al., 2013 ▸). The decrease in D t of F4H5-β-M may, in contrast, be correlated to the increase in micelle size. We calculated K D = k D/M for comparison with A 2 from SAXS. We found that K D is −0.62 mol ml g−2 for F2H9-β-M and −0.01 mol ml g−2 for F6H2-β-M. As previously reported in the literature (Harding & Johnson, 1985 ▸), the values of K D obtained by DLS are lower than the values of A 2 obtained from static measurements (SAXS or SLS) owing to hydrodynamic contributions. Despite this discrepancy, the inter­action parameters obtained from DLS experiments could be used to characterize the properties of surfactant micelles in solution conditions for MP crystallization.

3.1.3. Interactions by DLS in PEG solutions  

Interactions between micelles in solution conditions for membrane-protein crystallization have mainly been studied by measurement of the second virial coefficient (A 2 or B 22) using, for example, static light scattering or small-angle X-ray scattering (Barret et al., 2013 ▸; Hitscherich et al., 2000 ▸). Such methods are time-consuming and require calibration, proper solvent subtraction and concentration normalization. Without solvent subtraction or calibration to analyse the polydispersity and size distribution of particles in solution, dynamic light scattering is increasingly being used to characterize interaction parameters (Saluja et al., 2010 ▸; Shi et al., 2013 ▸; Yadav et al., 2011 ▸). In order to validate DLS measurements for MP crystallization diagnostics, experiments were initially performed in DD-β-M and PCC-α-M and compared with our previous SAXS studies (Barret et al., 2013 ▸). The variation of the diffusion coefficient ratio D t/D 0 is linear, since D t/D 0 = 1 + M K D(c − CMC), and decreases as the concentration of surfactant increases (Fig. 5 ▸ a), suggesting attractive interactions (K D < 0) between PCC-α-M micelles. The slope K D increases as the percentage of PEG 3350 increases owing to a depletion mechanism of the polymer (Asakura & Oosawa, 1958 ▸). The interaction parameters K D for DD-β-M and PCC-α-M are plotted as a function of the percentage of PEG 3350 and compared with the second virial coefficient A 2 as measured by SAXS (Fig. 5 ▸ b; Barret et al., 2013 ▸). The variation in K D follows the same trend as A 2 for the two surfactants, i.e. an increase in intermicellar attraction as the percentage of PEG increases, with the K D values being lower than the A 2 values for the two surfactants, as observed for FH-β-M in water.

Figure 5.

Figure 5

Open in a new tab

(a) DLS experiments on PCC-α-M as a function of the percentage of PEG 3350 for the determination of interaction parameters. (b) Comparison of K D = k D/M from DLS and A 2 from SAXS for DD-β-M and PCC-α-M.

3.2. Interactions for cloud-point boundary  

Membrane proteins are often observed to crystallize close to the detergent cloud-point boundary (Wiener & Snook, 2001 ▸), which corresponds to strong intermicellar attractions (Vivarès & Bonneté, 2004 ▸; Loll et al., 2001 ▸). With this in mind, we have characterized interactions between fluorinated micelles by determining the interaction parameter K D by DLS for the two fluorinated surfactants which present the most stable forms, i.e. F2H9-β-M and F6H2-β-M, as a function of concentration and compared them with that of DD-β-M. Fig. 6 ▸(a) shows the variation of K D as a function of the percentage of PEG 3350. As expected from the depletion effect, interactions between micelles become more attractive (K D < 0) as the polymer concentration increases, with the overall interactions being more attractive with F2H9-β-M than with F6H2-β-M or DD-β-M, in part owing to an increase in hard sphere potential and possible sugar-head hydration forces (Bauer et al., 2012 ▸) for F6H2-β-M. A similar trend is observed in the cloud-point boundary (Fig. 6 ▸ b) for the two surfactants compared with DD-β-M. The liquid–liquid phase transition is observed by optical microscopy at lower concentrations of surfactant and polymer for F2H9-β-M than for F6H2-β-M or DD-β-M. This result may be helpful for MP crystallization. Indeed, while a close correlation between solubility and second virial coefficient has long been demonstrated (Guo et al., 1999 ▸), the correlation of second virial coefficients between surfactant micelles on one hand and membrane protein–surfactant complexes on the other highlighted by Loll and coworkers (Hitscherich et al., 2000 ▸; Loll et al., 2001 ▸), as well as our recent results on free micelles and on the RC-LH1-pufX complex with PCC-α-M (Barret et al., 2013 ▸), suggest that a membrane protein purified in F2H9-β-M would have a lower solubility than one purified in DD-β-M.

Figure 6.

Figure 6

Open in a new tab

Interactions and cloud-point boundary of F2H9-β-M and F6H2-β-M compared with DD-β-M.

4. Conclusion  

Crystallization of membrane proteins in surfactant micelles suffers from a lack of fundamental studies to decipher the mechanisms, which could improve success in the growth of strongly diffracting crystals. Since the pioneering work of Loll and coworkers (Hitscherich et al., 2000 ▸), mechanistic studies have mainly dealt with traditional hydrogenated detergents. There is a lack of studies for some of the newer amphiphilic molecules that promote membrane-protein crystallization. In the present paper, we have assessed the potential of new fluorinated surfactants for MP crystallization by describing their structure and interactions in crystallizing solutions. It appears that a surfactant such as F2H9-β-M, which forms small stable micelles in an attractive regime in the presence of crystallizing agent while at the same time stabilizing MP, could be a good candidate for crystallization, as previously observed for PCC-α-M (Barret et al., 2013 ▸). However, comparison of these two surfactants, which form more attractive micelles than DD-β-M but with different aggregation numbers, raises an important question regarding the influence of the aggregation number on the amount of surfactant associated with the protein and its effect on crystal quality. Finally, the synthesis of a large amount of any new compound for MP stabilization and crystallization is not trivial. Since a thorough description of the surfactant micelle is a prerequisite for any surfactant exchange and membrane protein–surfactant complex crystallization, future studies in our laboratory will focus on large-scale synthesis of F2H9-β-M for MP crystallization trials and for comparison with other surfactants.

Acknowledgments

This study was supported by a PhD grant from the PACA region (Provence–Alpes–Cote d’Azur, France) via European Regional Development funds (FEDER). We are grateful to the European Synchrotron Radiation Facility for provision of synchrotron-radiation facilities and Adam Round for assistance in using beamlines ID14-eh3 and BM29.

Footnotes

1

Detergents are lipophilic surfactants that are able to solubilize lipid bilayers and extract MPs from membranes.

References

  1. Asakura, S. & Oosawa, F. (1958). J. Polym. Sci. 33, 183–192.
  2. Barret, L.-A., Barrot-Ivolot, C., Raynal, S., Jungas, C., Polidori, A. & Bonneté, F. (2013). J. Phys. Chem. B, 117, 8770–8781. [DOI] [PubMed]
  3. Barthélémy, P., Tomao, V., Selb, J., Chaudier, Y. & Pucci, B. (2002). Langmuir, 18, 2557–2563.
  4. Bauer, C., Bauduin, P., Girard, L., Diat, O. & Zemb, T. (2012). Colloids Surf. A Physicochem. Eng. Asp. 413, 92–100.
  5. Berger, B. W., Gendron, C. M., Lenhoff, A. M. & Kaler, E. W. (2006). Protein Sci. 15, 2682–2696. [DOI] [PMC free article] [PubMed]
  6. Breyton, C., Chabaud, E., Chaudier, Y., Pucci, B. & Popot, J.-L. (2004). FEBS Lett. 564, 312–318. [DOI] [PubMed]
  7. Breyton, C., Gabel, F., Abla, M., Pierre, Y., Lebaupain, F., Durand, G., Popot, J.-L., Ebel, C. & Pucci, B. (2009). Biophys. J. 97, 1077–1086. [DOI] [PMC free article] [PubMed]
  8. Chae, P. S., Gotfryd, K., Pacyna, J., Miercke, L. J. W., Rasmussen, S. G. F., Robbins, R. A., Rana, R. R., Loland, C. J., Kobilka, B., Stroud, R., Byrne, B., Gether, U. & Gellman, S. H. (2010). J. Am. Chem. Soc. 132, 16750–16752. [DOI] [PMC free article] [PubMed]
  9. Chae, P. S., Rasmussen, S. G. F. et al. (2010). Nature Methods, 7, 1003–1008. [DOI] [PMC free article] [PubMed]
  10. Chae, P. S., Rasmussen, S. G. F., Rana, R. R., Gotfryd, K., Kruse, A. C., Manglik, A., Cho, K. H., Nurva, S., Gether, U., Guan, L., Loland, C. J., Byrne, B., Kobilka, B. K. & Gellman, S. H. (2012). Chem. Eur. J. 18, 9485–9490. [DOI] [PMC free article] [PubMed]
  11. Duval-Terrié, C., Cosette, P., Molle, G., Muller, G. & Dé, E. (2003). Protein Sci. 12, 681–689. [DOI] [PMC free article] [PubMed]
  12. Frotscher, E., Danielczak, B., Vargas, C., Meister, A., Durand, G. & Keller, S. (2015). Angew. Chem. Int. Ed. 54, 5069–5073. [DOI] [PubMed]
  13. Guo, B., Kao, S., McDonald, H., Asanov, A., Combs, L. L. & Wilson, W. W. (1999). J. Cryst. Growth, 196, 424–433.
  14. Harding, S. E. & Johnson, P. (1985). Biochem. J. 231, 543–547. [DOI] [PMC free article] [PubMed]
  15. Hitscherich, C. J., Allaman, M., Wiencek, J., Kaplan, J. & Loll, P. J. (2000). Protein Sci. 9, 1559–1566. [DOI] [PMC free article] [PubMed]
  16. Hitscherich, C., Aseyev, V., Wiencek, J. & Loll, P. J. (2001). Acta Cryst. D57, 1020–1029. [DOI] [PubMed]
  17. Hovers, J. et al. (2011). Mol. Membr. Biol. 28, 171–181. [DOI] [PMC free article] [PubMed]
  18. Israelachvili, J. N., Mitchell, D. J. & Ninham, B. W. (1977). Biochim. Biophys. Acta, 470, 185–201. [DOI] [PubMed]
  19. Koszelak-Rosenblum, M., Krol, A., Mozumdar, N., Wunsch, K., Ferin, A., Cook, E., Veatch, C. K., Nagel, R., Luft, J. R., DeTitta, G. T. & Malkowski, M. G. (2009). Protein Sci. 18, 1828–1839. [DOI] [PMC free article] [PubMed]
  20. Lee, S. C., Bennett, B. C., Hong, W.-X., Fu, Y., Baker, K. A., Marcoux, J., Robinson, C. V., Ward, A. B., Halpert, J. R., Stevens, R. C., Stout, C. D., Yeager, M. J. & Zhang, Q. (2013). Proc. Natl Acad. Sci. USA, 110, E1203–E1211. [DOI] [PMC free article] [PubMed]
  21. Li, S., Xing, D. & Li, J. (2004). J. Biol. Phys. 30, 313–324. [DOI] [PMC free article] [PubMed]
  22. Loll, P. J., Allaman, M. & Wiencek, J. (2001). J. Cryst. Growth, 232, 432–438.
  23. Meyer, A., Dierks, K., Hussein, R., Brillet, K., Brognaro, H. & Betzel, C. (2015). Acta Cryst. F71, 75–81. [DOI] [PMC free article] [PubMed]
  24. Nehmé, R., Joubert, O., Bidet, M., Lacombe, B., Polidori, A., Pucci, B. & Mus-Veteau, I. (2010). Biochim. Biophys. Acta, 1798, 1100–1110. [DOI] [PubMed]
  25. Nollert, P. (2005). Prog. Biophys. Mol. Biol. 88, 339–357. [DOI] [PMC free article] [PubMed]
  26. Oliver, R. C., Lipfert, J., Fox, D. A., Lo, R. H., Doniach, S. & Columbus, L. (2013). PLoS One, 8, e62488. [DOI] [PMC free article] [PubMed]
  27. Orthaber, D., Bergmann, A. & Glatter, O. (2000). J. Appl. Cryst. 33, 218–225.
  28. Park, K.-H., Berrier, C., Lebaupain, F., Pucci, B., Popot, J.-L., Ghazi, A. & Zito, F. (2007). Biochem. J. 403, 183–187. [DOI] [PMC free article] [PubMed]
  29. Pebay-Peyroula, E., Garavito, R., Rosenbusch, J., Zulauf, M. & Timmins, P. (1995). Structure, 3, 1051–1059. [DOI] [PubMed]
  30. Penel, S., Pebay-Peyroula, E., Rosenbusch, J., Rummel, G., Schirmer, T. & Timmins, P. A. (1998). Biochimie, 80, 543–551. [DOI] [PubMed]
  31. Pernot, P. et al. (2013). J. Synchrotron Rad. 20, 660–664. [DOI] [PMC free article] [PubMed]
  32. Polidori, A., Presset, M., Lebaupain, F., Ameduri, B., Popot, J.-L., Breyton, C. & Pucci, B. (2006). Bioorg. Med. Chem. Lett. 16, 5827–5831. [DOI] [PubMed]
  33. Polidori, A., Raynal, S., Barret, L.-A., Dahani, M., Barrot-Ivolot, C., Jungas, C., Keller, S., Ebel, C., Breyton, C. & Bonneté, F. (2015). In preparation.
  34. Ravey, J.-C., Gherbi, A. & Stébé, M.-J. (1988). Prog. Colloid Polym. Sci. 76, 234–241.
  35. Saluja, A., Fesinmeyer, R. M., Hogan, S., Brems, D. N. & Gokarn, Y. R. (2010). Biophys. J. 99, 2657–2665. [DOI] [PMC free article] [PubMed]
  36. Shi, S., Uchida, M., Cheung, J., Antochshuk, V. & Shameem, M. (2013). Int. J. Biol. Macromol. 62, 487–493. [DOI] [PubMed]
  37. Shinoda, K., Hato, M. & Hayashi, T. (1972). J. Phys. Chem. 76, 909–914.
  38. Svergun, D. I. (1992). J. Appl. Cryst. 25, 495–503.
  39. Tanford, C. (1961). Physical Chemistry of Macromolecules. New York: Wiley.
  40. Tate, C. G. (2010). Methods Mol. Biol. 601, 187–203. [DOI] [PubMed]
  41. Vivarès, D. & Bonneté, F. (2004). J. Phys. Chem. B, 108, 6498–6507. [DOI] [PubMed]
  42. Wiener, M. C. & Snook, C. F. (2001). J. Cryst. Growth, 232, 426–431.
  43. Yadav, S., Scherer, T. M., Shire, S. J. & Kalonia, D. S. (2011). Anal. Biochem. 411, 292–296. [DOI] [PubMed]
  44. Yu, S. M., McQuade, D. T., Quinn, M. A., Hackenberger, C. P. R., Gellman, S. H., Krebs, M. P. & Polans, A. S. (2000). Protein Sci. 9, 2518–2527. [DOI] [PMC free article] [PubMed]
  45. Zhang, Q., Tao, H. & Hong, W.-X. (2011). Methods, 55, 318–323. [DOI] [PMC free article] [PubMed]

Articles from Acta Crystallographica. Section F, Structural Biology Communications are provided here courtesy of International Union of Crystallography

RESOURCES