Façade detergents as bicelle rim-forming agents for solution NMR spectroscopy

2.1 Materials

Façade-EM, Façade-EPC, CHAPS, lauryl sarcosinate (LS) and 1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC) were the products of Avanti Polar Lipids Inc, Alabaster, USA. Stable isotope derivatives (D₂O, ammonium chloride) were purchased from the Cambridge Isotope Laboratory, Tewksberry, USA. Structures of all used lipids and detergents are shown on Figure 1.

Figure 1:

Structures of bicelle rim-forming agents, used in the current study: CHAPS, Façade-EPC, and Façade-EM.

2.2 Protein expression and purification

Construct corresponding to the transmembrane and cytoplasmic domains of rat p75NTR (p75-ΔECD, residues 245–425) with N-terminal polyhistidine tag and trombin cleavage site was expressed in Escherichia coli culture and purified as described in Ref. [10] using metal-chelating and size exclusion chromatography LS solution. Protein samples were all uniformly 15N-labeled.

2.3 Sample preparation

To prepare the protein-free bicelles, ~3 mg of DMPC in dry powder were dissolved by the necessary amount of 10% stock solution of Façade-EM or Façade-EPC and then diluted by the 30 mm NaPi buffer, pH 6.0, to the final volume of 200 μl. All samples were freezed in the liquid nitrogen and heated to 40°C in ultrasonic bath for 3–4 times prior to the NMR experiments. To assemble the bicelles, containing the p75-ΔECD, DMPC was added to 5 mg of protein in LS at the lipid/protein ratio 150. Then, the detergent was removed from the solution by six subsequent cycles of 100-fold dialysis against the 20 mm pH 7.0 sodium phosphate buffer until the formation of liposomes, each cycle of dialysis took 24 h. Necessary amounts of 10% stock solution of CHAPS, Façade-EM, or Façade-EPC were added to the solution to form q=0.7–1.5 bicelles, followed by 15 min of ultrasonification at 40°C, and sample was concentrated to 450 μl using the 30 kDa Amicon centrifugal filtering units (Sigma-Aldrich, St. Louis, USA). To vary the q ratio, the solution was added to the necessary amount of lipid/detergent in dry powder, vortexed, and put into the ultrasonic bath for 5–6 min at 40°C.

2.4 NMR experiments

All NMR spectra were recorded on Bruker Avance 700 and Avance III 600 spectrometers with triple-resonance TCI cryoprobes (Bruker Biospin, Rheinstetten, Germany). Diffusion coefficients of lipids were measured by ¹H NMR as described in Ref. [25] using double BPLED experiment [35]. The maximal gradient pulse strength was equal to 52 G/cm, diffusion delays were typically 100–300 ms long. Experiments were run at 15°C, 27°C, and 40°C, for some samples the temperature dependence in the range 10°C–50°C was recorded. Thick-wall NMR tubes (Norell Inc, Landsville, USA) were utilized to avoid the convection at high temperatures. At least 30 min interval was used after the change of the temperature. The diffusion coefficients of lipids were measured using signals at 0.85, 1.25, and 3.25 ppm. Diffusion coefficients of Façade-EPC were measured using signals at 0.72, 0.93, 0.99, 3.27, and 3.28 ppm. Diffusion coefficients of Façade-EM were measured based on the signals at 0.72, 0.93, and 0.99 ppm. BEST-TROSY (transverse relaxation optimized spectroscopy) spectra [36] were recorded at several temperatures (usually 20°C and 30°C) for each studied q in DMPC/Façade-EPC, DMPC/Façade-EM, and DMPC/CHAPS bicelles to estimate the folding of the intracellular domain of p75-ΔECD.

To interpret the obtained diffusion coefficients we utilized the formalism that was developed and tested in our previous work [25], including the correction for the obstructed diffusion of large particles [37] and solvent [38] and taking into account the discoidal shape of bicelles [39].

To interpret the obtained radii we use the cylindrical model of DMPC/Façade bicelles [25]:

where R is the radius of a bicelle, r_⊥ is the thickness of the rim, q^* is the molar ratio, lipid over detergent, which is corrected to take into account that a fraction of the detergent is soluble in the monomeric form; and λ is the ratio of detergent volume over the DMPC volume.

3.1 Façade-EM behaves similarly to CHAPS in isotropic bicelles

Out of five commercially available Façade detergents we selected Façade-EM and Façade-EPC for the purpose of our study. Façade-EPC is the only substance that contains zwitterionic phosphocholine moieties (Figure 1), which is similar to the headgroups of lipids used in bicelles. Additionally, this detergent has the lowest molecular weight, 797 Da. Low molecular weight implies that bicelles of Façade-EPC contain more lipids than bicelles of the same mass, formed with the heavier rim-forming surfactant. Façade-EM is one of the four detergents that are modified with maltose or glucose residues. It is characterized by one of the lowest molecular weight (1115 Da, 2 maltose residues), but is more bulky and amphipathic than Façade-TEG (996 Da, 3 glucose residues) and, therefore, seems to less likely enter the bilayer region of bicelles.

As the first step, we characterized the DMPC/Façade-EM and DMPC/Façade-EPC bicelles, using ¹H NMR diffusion approach, that was published in our previous study of CHAPS and CHAPSO-based bicelles [25]. Radii of bicelles were measured as a function of q^* (the molar ratio, lipid over detergent, which is corrected to take into account that a fraction of the detergent is soluble in the monomeric form) and ambient temperature. Like in the previous work, we performed our measurements at three temperatures: 15°C, 27°C, and 40°C to sample the conditions below, above, and close to the phase transition temperature of DMPC in bilayers (24°C). If any temperature-dependent effects on the bicelle size were detected, data at several additional temperature points was obtained. As Façades are the derivatives of cholic acid, we compared the obtained data to the behavior of CHAPS-formed bicelles. CHAPS is structurally very similar to Façades and was previously shown to be the optimal conventional rim-forming detergent for small isotropic bicelles that are used in solution NMR spectroscopy [25].

Analysis of the data obtained for DMPC/Façade-EM bicelles reveals that as a rim-forming agent Façade-EM behaves in many aspects similarly to CHAPS and CHAPSO. Appearance of the signal from the lipid methyl group in ¹H spectra (narrow splitted triplet) is the same, indicating the homogeneous environment of lipid fatty chains (Figure 2E). Bicelles composed of Façade-EM are greater than DMPC/CHAPS particles at low q^*. At q^*=0.6 the radii of DMPC/Façade-EM complexes were 0.4 nm greater than the radii of DMPC/CHAPS bicelles, which can be explained by the thicker rim formed by the more bulky Façade molecules. However, the R(q^*) dependence of DMPC/Façade-EM bicelles is less steep than the one of DMPC/CHAPS, and at q^* higher than 1.3–1.4 bicelles of Façade-EM become smaller than bicelles formed by CHAPS (Figure 2A). The fit of the data obtained at 15°C to the Eq. (1) resulted in following parameters: r_⊥=1.7±0.1 nm and λ=3.1±0.4. Thus, Façade-EM forms the rim 0.6 nm thicker than CHAPS and occupies the volume that is ca. two times greater, which is in agreement with the differences in the detergent structures and molecular weights [25]. Like DMPC/CHAPS, DMPC/Façade-EM particles undergo temperature-dependent size variation. The effect is shifted to higher q^*–R(T) dependence of DMPC/Façade-EM at q^*=1.36 is almost equivalent to the dependence observed in DMPC/CHAPS solution at q^*=1.28 (Figure 2C). It is also noteworthy that the growth of bicelles starts at higher temperatures in bicelles formed by Façade-EM: radii of bicelles are the same at 15°C and 27°C at q^*=1.6, while radii of DMPC/CHAPS particles are remarkably different at these two temperatures even at lower q^* (1.4). Finally, this effect is less pronounced: q^*=1.6 DMPC/Façade-EM bicelles grow up by 15%–20% at 40°C, while the size of q^*=1.5 DMPC/CHAPS bicelles is increased twofold. The quantity of free Façade-EM is extremely low in bicelles [34] (it does not exceed 0.1 mm), therefore, we did not manage to exactly measure it. On the other hand, we can state that the effect of free Façade-EM on q^* is negligible at almost any reasonable concentration of bicelles. To sum up, we can conclude that the use of Façade-EM instead of CHAPS allows obtaining the bicelles with thicker rim, and, therefore, smaller bilayer region; however, such bicelles reveal the weaker temperature-dependent effects.

Figure 2:

(A) R(q^*) dependence for the DMPC/CHAPS (red Ref. [25]) and DMPC/Façade-EM (blue) bicelles at 15 (triangles), 27 (squares), and 40°C (circles). Fit of the data at 15°C to the Eq. (1) is shown by blue solid line for Façade-EM and by dashed red line for CHAPS. (B) R(q^*) dependence for DMPC/CHAPS (red Ref. [25]) and DMPC/Façade-EPC (blue) bicelles at 15 (triangles), 27 (squares), and 40°C (circles). Fit of the data at 15°C and 27°C to the Eq. (1) are shown by blue dashed and solid lines, respectively. (C, D) R(T) dependence is plotted for q^*=1.0 (blue triangles) and q^*=1.28 (magenta triangles) DMPC/CHAPS bicelles [25], q^*=1.15 (C, brown squares) and q^*=1.36 (C, red squares) DMPC/Façade-EM bicelles, q^*=1.1 (D, brown squares) and q^*=1.6 (D, red squares) DMPC/Façade-EPC bicelles. R(T) dependencies were measured at almost all q^* studied within the current work, and the reported q^* were selected as representative for the different kinds of R(T) dependencies. (E) Fragments of ¹H NMR spectra of q^*=1.36 DMPC/Façade-EM, q^*=1.15 DMPC/CHAPS, and q^*=2.15 DMPC/Façade-EPC bicelles. Fragments contain the signal of the DMPC methyl group at 0.87 ppm and were recorded at 40°C.

3.2 Façade-EPC forms isotropic bicelles which do not grow in size upon heating

In contrast, bicelles formed by Façade-EPC behaved differently than the solutions of CHAPS and Façade-EM. First, DMPC/Façade-EPC bicelles are always much smaller than particles observed in the DMPC/CHAPS and DMPC/Façade-EM mixtures at the same q^* (Figure 2B): at q^*=2.2°C and 15°C they were as small as 3.9 nm, which corresponds to q^*=1.5 DMPC/CHAPS bicelles at the same temperature. Signal from the lipid methyl group in 1H spectra is broad and is split into several unequal components, indicating the heterogeneous environment of lipid fatty chains and possible mixing between the lipids and detergents (Figure 2E). Fit of the data at 27°C to Eq. (1) results in parameters which are rather strange: r_⊥=1.22±0.03 nm and λ=2.3±0.1. That is, while the rim of DMPC/Façade-EPC particles is almost of the same thickness as the rim of DMPC/CHAPS bicelles (1.1 nm), Façade-EPC is packed much more loosely on the rim and occupies the volume which is 1.7 times greater than the volume of CHAPS. On the other hand, at 15°C the Façade-EPC-based bicelles are described by different parameters: r_⊥=1.6±0.1 nm and λ=3.9±0.4. Thus, below the lipid phase transition Façade -EPC molecules are packed much less tightly in the rim of bicelles, which makes it slightly thicker. At 15°C parameters of the Façade-EPC-based bicelles are very much alike the parameters of Façade-EM ones. However, we need to point out that the model (1) does not describe properly the behavior of bicelles if the mixing between the lipid and detergent is present, which is very likely taking into account our considerations that are given below. Façade-EPC/DMPC solutions also demonstrate a very peculiar temperature-dependent behavior (Figure 2D). At q^* below 2.0 such bicelles do not grow but rather get smaller upon heating. The growth of bicelles is observed only at very high q^* (3.3) and large size (~5 nm) and is weakly expressed: radii of bicelles are increased by ca. 25% at 40°C. The reduced size of the DMPC/Façade-EPC particles explains the reported ability of Façade-EPC to dissolve the greater amount of lipids than CHAPS [34]. Façade-EPC bicelles have smaller radii at the same q^* than CHAPS or CHAPSO-based particles. Therefore, the transition from the soluble bicellar phase to other phases, which are characterized by the presence of extremely large and insoluble particles, such as perforated bilayers, could be expected to occur at higher q^* for Façade-EPC.

In case of Façade-EPC we also managed to measure the concentration of detergent in the monomeric state using the diffusion coefficients of lipids and detergents [25], which was not possible in case of Façade-EM: due to the extremely low CMC the diffusion of lipids and detergents were indistinguishable within the experimental error. The contents of monomeric Façade-EPC gradually decreased from 1.4 mm at q^*=1 to 1.1 mm at q^*=3.3. The concentration of monomeric Façade-EPC was also dependent on the ambient temperature, similarly to CHAPS and other rim-forming detergents (Figure S1, [25]). Below 20°C, the concentration of free Façade-EPC begins to grow with the decrease of temperature, and may reach 3.0 mm at 10°C and q^*=1.0. The pronounced temperature dependence of the monomeric Façade-EPC explained the differences in q^* for the radii measured at different temperatures (Figure 2). Summarizing the paragraph, the use of Façade-EPC allows obtaining the bicelles with relatively thin rim which do not grow with temperature until their size exceeds 4 nm (the maximal size applicable for solution NMR spectroscopy).

3.3 Façade-EM supports the native and soluble state of the intracellular domain of a model membrane protein

To test the ability of Façades to support the native folding of water-soluble domains we inserted the p75-ΔECD protein, which contains both the transmembrane (TM) and intracellular domains of p75NTR to the bicelles formed by Façade-EM, Façade-EPC, and CHAPS at pH 7.0. To study the folding and solubility of the domain we used the ¹H, ¹⁵N-HSQC (heteronuclear single quantum correlation) spectra, which may be considered as “fingerprints” of the protein spatial structure. The globular “death domain” (DD) of p75NTR is very sensitive to the contents of the membrane mimetic. DD is connected to the TM domain by a very flexible linker and yields the perfect HSQC spectra inside LPNs [10], which are identical to the spectrum of isolated domain in aqueous solution. However, the presence of harsh detergents, such as dodecylphosphocholine or DHPC (in DMPC/DHPC bicelles), results in the improper folding of the domain and altered peak positions in NMR spectra (Figure 3). In contrast, in DMPC/CHAPS bicelles the folding of the DD is native, cross-peaks of the DD are at the same positions in HSQC spectra, but are much less intense (Figure 4). The intensities of the DD signals in NMR spectra are dependent on both q^* and temperature – best spectra are obtained in large bicelles (q^*>1.5) and at low temperature (20°C) (Figure 5). Most likely such behavior is caused by the interaction between the DD and CHAPS in the rim of slowly tumbling bicelles. If the domain is adsorbed on the rim of the bicelle, the intensity of the signals will be decreased due to the transverse relaxation, which is enhanced either because of the slow Brownian tumbling of bicelles or transitions between the soluble and absorbed state of the domain. This hypothesis is supported by the fact that only the signals of the flexible linker region between the TM domain and DD are observed in NMR spectra of p75-ΔECD in pure CHAPS (Figure 3). In larger bicelles the probability of the interaction between the bicelle rim and DD is statistically reduced, which explains the improvement of NMR spectra at higher q^*. To quantify this effect we measured the ratio between the intensities of several signals corresponding to the DD of the protein and the intensity of the signal of G334, which is located in the flexible linker region of p75-ΔECD (Figure 5). The intensity of G334 does not depend on the membrane mimetic and is a simple measure of the protein concentration, and the ratio is a measure of the spectrum quality and solubility of the DD.

Figure 3:

¹H,¹⁵N-TROSY (transverse relaxation optimized spectroscopy) spectra of p75-ΔECD in DMPC/MSP1 LPNs [10], DMPC/Façade-EM q=1.6 bicelles, dodecylphosphocholine (DPC) [10], and CHAPS. Spectra were acquired at pH 7.0, 30°C.

Figure 4:

¹H, ¹⁵N-TROSY spectra of p75-ΔECD in DMPC/CHAPS (left column), DMPC/Façade-EM (middle column), and DMPC/Façade-EPC (right column) at various q^*. All spectra were recorded at 20°C, pH 7.0 and were processed and visualized identically. Signals that were used to assess the quality of the spectrum of p75NTR DD are indicated.

Figure 5:

Quality of NMR spectra of the p75NTR DD in various bicelles. The ratios of the intensities of signals, corresponding to G365 and S379 from the DD of p75-ΔECD, over the intensity of signal, corresponding to G334 from the flexible linker, connecting the DD and TM domain of the protein, are shown for MSPD1/DMPC LPNs [10], DMPC/CHAPS, DMPC/Façade-EM, and DMPC/Façade-EPC bicelles formed at various q^*. Data were obtained at 20°C (filled bars) and 30°C (empty bars).

The results of the experiments with Façade-EM revealed that this detergent retains the folded and soluble state of the p75NTR DD in a wide range of q^*. Even at q^*=0.75 at 20°C the spectrum of the DD in DMPC/Façade-EM bicelles was better than in any other bicellar mixture and the quality of spectrum at q^*=1.6 was very close to the quality observed in LPNs. At higher temperatures worse spectra were obtained; however, the temperature-based effects were not much pronounced at q^* above 1 for the majority of DD cross-peaks (Figure 5). It is noteworthy that the spectra of p75-ΔECD in Façade-EM/DMPC are almost identical to the spectra in LPNs [10] (Figure 5), implying the same structure and mobility of both death and chopper domains of the receptor. On the other hand, use of Façade-EPC instead of CHAPS did not result in any improvement of the DD spectrum. We studied the quality of NMR spectra of p75-ΔECD over a wide range of q^*:1.8–3.9 of DMPC/Façade-EPC mixtures. These bicelles are smaller than ones of CHAPS and Façade-EM, and we have to ascertain that spectra would not improve dramatically, when the size of bicelles becomes comparable to the size of MSP1-based LPNs. The spectra of p75-ΔECDin DMPC/Façade-EPC, obtained at all q^* in the range 1.8–3.9, were of a very poor quality, the majority of signals were broad and some signals were absent. At temperatures higher than 30°C additional signals were present in the spectra of p75NTR, indicating that the detergent causes partial unfolding of the receptor’s DD (Figure S2). Therefore, we conclude that the Façade-EM in bicelles is the mildest detergent and allows obtaining the globular domain of a model membrane protein in a folded and soluble state, while Façade-EPC and CHAPS tend to interact with the domain, which results in worse quality of NMR spectra.

3.4 The nature of temperature-induced effects

In the present work, we tested the Façade-EM and Façade-EPC as bicelle rim-forming agents in an attempt to avoid or abate the temperature-induced growth of bicelles and either denaturation of soluble and globular domains of large membrane proteins or their adsorption on bicelle surface. While we succeeded in achieving the second goal, the temperature-induced size variation was still persistent in new types of bicelles; however, the effect seems to be reduced dramatically in comparison to the effect observed for the DMPC/CHAPS species. DMPC/Façade-EM particles grew with temperature at large q^*, while DMPC/Façade-EPC bicelles became smaller at high temperatures at low q^* and larger at higher q^*. This is at most interesting taking into account the very similar structures of Façade-EM, Façade-EPC, and CHAPS. All three are facial amphiphiles – “flat” molecules with hydrophobic and hydrophilic faces. The difference between the three lies in the architecture of the polar face of the detergent – CHAPS has the least polar surface formed by three hydroxyl groups, while both Façades have bulky and highly polar substitutes at these positions. Thinking about the origin of the temperature-induced effects on the size of isotropic bicelles one could suggest three explanations. First, the size of bicelles can be increased or decreased upon heating due to the mixing between the lipid and detergent in the planar region or rim of the particle. State when lipids and detergents are separated is unfavorable from the point of entropy; therefore, the extent of mixing will always be greater at higher temperature. Such mixing was shown to take place in large anisotropic DMPC/DHPC and DMPC/CHAPSO bicelles [24], [31] and can explain the behavior of low-q^* DMPC/Façade-EPC mixtures. If the lipid is mixed with Façade-EPC in the rim of the bicelles, then their size will be decreased, if the magnitude of mixing is elevated. Possible mixing in the rim of DMPC/Façade-EPC particles can explain some of the reported findings: small size of DMPC/Façade-EPC bicelles, distorted appearance of the lipid signals in ¹H NMR spectra, strange parameters of the Eq. (1) that describe the R(q^*) dependence for these bicelles, and inability of Façade-EPC to support the folded and soluble state of p75NTR DD. It is likely that the actual thickness of the rim in Façade-EPC bicelles is greater than calculated 1.2 or 1.6 (at 15°C) nm and corresponds to the length of DMPC molecule. This unique behavior of Façade-EPC can be caused by the presence of the phosphocholine groups in the detergent, which makes it miscible with lipids. However, while the physical reasons of such changes are not clear, it could happen that lipids in the planar region of DMPC/Façade-EPC particles experience a phase transition, which indeed results in the altered packing of Façade-EPC in the rim of the bicelles as revealed by the fitting of the experimental data to model (1), which, in turn, leads to the altered R(q^*) dependencies.

On the other hand, the observed growth of large isotropic bicelles upon heating cannot be easily explained by the mixing between the lipids and detergents in the planar area of bicelles. First, the question is raised why the temperature-induced growth of bicelles is not observed in low-q^* mixtures. And, second, effect of the same magnitude is observed for CHAPS and Façade-EM if we consider the size of the lipid bilayer, not the size of the bicelles. We can estimate the radius of the bilayer region by subtracting the rim thickness r_⊥, obtained as a parameter of the Eq. (1), from the bicelle radius R. Assuming that the rim thickness is independent of the temperature, the radius of bicelle bilayer region is increased in q^*=1.6 DMPC/Façade-EM from 2.5 to 3.6 nm upon heating from 27°C to 40°C. Similarly, in q^*=1.3 DMPC/CHAPS mixture, the radius of lipid bilayer is increased from 2.6 to 3.6 nm. This is strange, as Façade-EM contains bulky maltose chains, and it is difficult to imagine how such a molecule can enter the lipid bilayer in the same way and amount as CHAPS. Nevertheless, we can presume that in large bicelles above the phase transition, the surface of bilayer is “mosaic” with hydrophobic spots on the hydrophilic field [40], and facial amphiphiles can exit the rim to shield such spots from the aqueous environment. This process will be accompanied by the growth of bicelles due to the increase of the apparent q^* ratio. Alternatively, the growth of bicelles can be explained by the changes in lipid packing. The average volume of lipid molecule in the bilayer is the key parameter of Eq. (1), and it is known to increase with the ambient temperature [41]. In model bilayer systems the average volume per molecule of lipid is increased by 10%–15% upon phase transition and is additionally increased by 5% upon heating from 30°C to 50°C. Analysis of the data on Façade-EM bicelles reveals that the volume of DMPC needs to be increased by 60% to explain the observed growth of bicelles upon heating from 15°C to 40°C (Figure 2A), which seems scarcely possible, and further heating would require even a greater change in the average volume per lipid molecule. The effect that is observed in large DMPC/CHAPS bicelles is much more pronounced and requires a dramatic increase in the volume per lipid molecule. Therefore, while it may happen that bicelles lack the lateral pressure that determines the lipid density in liposomes or other bilayer systems and the packing of lipids in bicelles is distorted much stronger upon heating, it seems extremely unlikely. Third, the temperature-induced growth of bicelles may be explained by the change of the detergent packing in the rim – increased rim thickness or decreased average volume per molecule of the detergent. Summarizing, we show that temperature-induced growth of bicelles is a common feature of all bicelles which have a relatively large patch of lipid bilayer and does not depend on the shape and nature and polarity of the hydrophilic face of a detergent. It is obvious that the phenomenon requires a certain amount of lipids in the bilayer to occur and is somehow related to the phase transition of lipids.

3.5 Applicability of Façades for the structural studies

The initial goal of the work was to develop the novel membrane mimetics for NMR studies. With this regard, it is necessary to compare the parameters of Façade-formed and conventional bicelles that determine their applicability for structural studies with solution NMR and other techniques. One of the relevant parameters is the number of lipids in the bilayer region. Sometimes it is necessary to maintain a certain size of the bilayer to allow the interaction between the juxtamembrane regions or soluble domains of membrane proteins with the membrane. Neither Façade-EM nor Façade-EPC provide any advantages with this respect in comparison to the conventional detergents: rim of Façade-EM is 0.6 nm thicker than ones of CHAPS and DHPC [24], [25] and rim of Façade-EPC is a mixture of detergent and lipid and is likely to have thickness comparable to the length of the lipid molecule. Therefore, given the size of bicelles is the same, DMPC/Façade mixtures will be characterized by the lower size of the bilayer area. On the other hand, Façade-EM is a very mild detergent and demonstrates the lowest propensity to unfold the extramembrane domains of large membrane proteins or interact with them. Therefore, if the protein of interest contains the soluble domain that is susceptible to the presence of detergents, Façade-EM-based bicelles can be considered as an alternative to LPNs. Besides, as we have already mentioned in the introduction, bicelles grow in size upon dilution due to the detergent solubility in the monomeric state [27]. In our recent work, we suggested an approach to control the size of bicelles upon dilution; however, it does not work well, when the concentration of bicelles is much lower than the concentration of free detergent [25]. Façade-EM is characterized by very low CMC (0.2 mm [34]), which is 40 times lower than CMC of CHAPS. Such a small CMC allows the precise size control of Façade-EM-based bicelles even at extremely low concentrations. Thus, Façade-EM bicelles need to be considered as an alternative to other membrane mimetics in case the low lipid-to-protein ratio and low concentration of proteins are required simultaneously, e.g. in Förster resonance energy transfer or electron spin resonance studies. As for Façade-EPC, we need to accept that it does not provide any advantages and should not to be considered as a bicelle rim-forming agent.