the carbon film are more evenly dispersed in APol-trapped samples, so that collecting a large
number of well-separated images becomes easier. In Study 12.3 (Flötenmeyer et al. 2007), this
effect was tentatively attributed to the surface tension of the air/water interface being possibly
higher for the APol-trapped sample than for the detergent-solubilized one. A high surface tension,
such as that between pure water and air, makes it energetically costly for the film to form a
meniscus with a small radius of curvature, whereas the adsorption of surfactants at the interface
lowers this energetic cost. As a result, thinning of the film is easier in detergent solution, as actually
observed (cf. Fig. 12.9E, F), and the large particles of mitochondrial Complex I that were the
target of Study 12.3 are more likely to be drained toward the meniscus that forms at the contact
between the liquid film and the supporting carbon in detergent solution than after trapping with
APols.
Later studies showed that, at equilibrium, the surface tension of A8-35 solutions is in fact not
much different from that of detergent solutions (Giusti et al. 2012; see Chap. 4, § 4.3.1.2.4).
However, kinetic effects may come into play. In a typical cryo-EM experiment, (i) the free APol is
separated from the MP/APol complexes by SEC, and (ii) the excess solution deposited on the grid
is blotted a few seconds before freezing (Dobro et al. 2010). Blotting must remove most of the
surfactant layer adsorbed at the air/water interface, bringing the surface tension close to that of pure
water. In detergent solutions, which are above the CMC, the interfacial monolayers can reform
quickly due to the rapid diffusion of monomeric detergent molecules, and the surface tension may
drop again before the film is frozen. In solutions of MP/APol complexes, where the concentration
of free APol is very low, the monolayers will reform due to the migration to the surface of
occasional free APol particles and of MP/APol complexes. This process, however, is necessarily
slower than the adsorption of small detergent molecules, and it seems possible that, at the time of
freezing, the surface tension is still higher, opposing the formation of a sharply bent meniscus and
the thinning of the film. Note, however, that such a mechanism falls short of explaining why
dynactin is more evenly spread in the presence of A8-35 than in the absence of any surfactant
(Study 12.20).
These processes deserve a systematic examination, because understanding them would likely
be of importance in achieving a better control of the distribution of the particles across the aqueous
film, particularly for large complexes.
(iii) Orientation. A preferential orientation of the particles with respect to the plane of the film can be
an advantage or a disadvantage depending on the reconstruction procedure used. In the conical
tilt reconstruction method, which relies on imaging the same field of particles at two different
angles, it is an advantage. When the samples are not tilted and reconstruction is based on
reorienting with respect to one another averaged 2D projections covering all possible angles, it
becomes a handicap. Little information is available about the way MP/APol complexes interact
with carbon films. In Study 12.13, it was noted that the OmpF filaments that form when OmpF/
A8-35 complexes are depleted of free APol tend to orient with the TM region of BR facing the
carbon film (Arunmanee et al. 2014). In Study 12.9, on the contrary, it seems that replacing DDM
with A8-35 results in a preferential, although not systematic, orientation of the particle with their
fourfold symmetry axis normal to the film, that is with the APol layer away from the carbon
surface (Liao et al. 2013; compare, above, images and averages in DDM, Fig. 12.27A, B, with
those of APol-trapped complexes, Fig. 12.27C, D).
12.3 The Use of Amphipols for Membrane Protein Electron Microscopy Studies
575
number of well-separated images becomes easier. In Study 12.3 (Flötenmeyer et al. 2007), this
effect was tentatively attributed to the surface tension of the air/water interface being possibly
higher for the APol-trapped sample than for the detergent-solubilized one. A high surface tension,
such as that between pure water and air, makes it energetically costly for the film to form a
meniscus with a small radius of curvature, whereas the adsorption of surfactants at the interface
lowers this energetic cost. As a result, thinning of the film is easier in detergent solution, as actually
observed (cf. Fig. 12.9E, F), and the large particles of mitochondrial Complex I that were the
target of Study 12.3 are more likely to be drained toward the meniscus that forms at the contact
between the liquid film and the supporting carbon in detergent solution than after trapping with
APols.
Later studies showed that, at equilibrium, the surface tension of A8-35 solutions is in fact not
much different from that of detergent solutions (Giusti et al. 2012; see Chap. 4, § 4.3.1.2.4).
However, kinetic effects may come into play. In a typical cryo-EM experiment, (i) the free APol is
separated from the MP/APol complexes by SEC, and (ii) the excess solution deposited on the grid
is blotted a few seconds before freezing (Dobro et al. 2010). Blotting must remove most of the
surfactant layer adsorbed at the air/water interface, bringing the surface tension close to that of pure
water. In detergent solutions, which are above the CMC, the interfacial monolayers can reform
quickly due to the rapid diffusion of monomeric detergent molecules, and the surface tension may
drop again before the film is frozen. In solutions of MP/APol complexes, where the concentration
of free APol is very low, the monolayers will reform due to the migration to the surface of
occasional free APol particles and of MP/APol complexes. This process, however, is necessarily
slower than the adsorption of small detergent molecules, and it seems possible that, at the time of
freezing, the surface tension is still higher, opposing the formation of a sharply bent meniscus and
the thinning of the film. Note, however, that such a mechanism falls short of explaining why
dynactin is more evenly spread in the presence of A8-35 than in the absence of any surfactant
(Study 12.20).
These processes deserve a systematic examination, because understanding them would likely
be of importance in achieving a better control of the distribution of the particles across the aqueous
film, particularly for large complexes.
(iii) Orientation. A preferential orientation of the particles with respect to the plane of the film can be
an advantage or a disadvantage depending on the reconstruction procedure used. In the conical
tilt reconstruction method, which relies on imaging the same field of particles at two different
angles, it is an advantage. When the samples are not tilted and reconstruction is based on
reorienting with respect to one another averaged 2D projections covering all possible angles, it
becomes a handicap. Little information is available about the way MP/APol complexes interact
with carbon films. In Study 12.13, it was noted that the OmpF filaments that form when OmpF/
A8-35 complexes are depleted of free APol tend to orient with the TM region of BR facing the
carbon film (Arunmanee et al. 2014). In Study 12.9, on the contrary, it seems that replacing DDM
with A8-35 results in a preferential, although not systematic, orientation of the particle with their
fourfold symmetry axis normal to the film, that is with the APol layer away from the carbon
surface (Liao et al. 2013; compare, above, images and averages in DDM, Fig. 12.27A, B, with
those of APol-trapped complexes, Fig. 12.27C, D).
12.3 The Use of Amphipols for Membrane Protein Electron Microscopy Studies
575