Box 5.2 (continued)
the complexes (Gohon et al. 2008), presumably because of electrostatic interactions between the
resin and the polyanionic APol, given that no such inconsistency is observed with non-ionic APols
(Sharma et al. 2012) (see main text).
In a comparative single-particle study of negatively stained plant aquaporin SoPIP2;1
tetramers complexed by various surfactants, it was concluded that the surfactant belt is thicker in
APol-trapped complexes than in detergent solution: 11 Æ 2 Å in OG, 13 Æ 2 Å in DDM, 15 Æ 2 Å
in LMNG, and 19 Æ 3 Å in A8-35 (Vahedi-Faridi et al. 2013; Fig. 5.19). However, the definition of
the envelope of the reconstructed image depended on an estimate of the amount of bound surfactant.
In the case of A8-35, this was taken to be ~220 kDa (taking the measured binding to BR as a basis
and assuming the mass of APol bound to be proportional to that of the protein), whereas, on the
basis of the data collected in Table 5.5 and Fig. 5.13, ~90 kDa would seem more likely given the
size of the TM region of aquaporins. The thickness of the belt deduced from this work is therefore
almost certain to be overestimated.
Based on the density of the dry surfactants, it is difficult to understand why MP/A8-35
complexes should be larger than MP/DDM ones, given that they generally comprise much less
surfactant and that the specific volumes of A8-35 and DDM are comparable (0.809 mLÁg
À1 for
A8-35 (Gohon et al. 2004), 0.81–0.837 mLÁg
À1 for DDM (le Maire et al. 2000)). However,
hydration must be taken into account. Water binding by DDM has been estimated, by either
equilibrium ultracentrifugation or SEC, to be in the range of 0.26–0.34 g water per g detergent
(de Vitry et al. 1991; Møller and le Maire 1993), that of A8-35, by a combination of biophysical
measurements, to be 1.22 Æ 0.45 g water per g polymer (Gohon et al. 2004). Combining the
extremes of each range, one concludes that the volume of hydrated DDM associated to 1 g of BR is
comprised between 3.9 and 4.7 cm
3
, that of hydrated A8-35 between 2.4 and 3.3 cm
3
. One would
therefore expect the BR-bound belt of A8-35 to be roughly between 0.5Â and 0.8Â as thick as the
DDM one.
Consistent with this view, a comparative NMR study of BR in DDM solution, DMPC
nanodiscs, or A8-35 concluded that the NMR spectra were comparable and all three systems were
adequate for solution NMR measurements (Etzkorn et al. 2013; cf. Chap. 10, Fig. 10.14). The work
included estimates of M and τ c values, which, at variance with SEC data, were consistent with BR/
A8-35 complexes being smaller than BR/DDM ones: M % 93 kDa in A8-35 vs. ~128 kDa in DDM,
τ c ¼ 31.7 vs. 41.4 ns. Because the BR/A8-35 preparation was partially aggregated, the data are
likely to actually overestimate the size of BR/A8-35 monomers. Indeed, on the basis of direct
measurements using [
3
H]A8-35, the mass of the monomeric complexes would be expected to be
~81 kDa (Gohon et al. 2008), assuming the lipid-free BR used by Etzkorn et al. (2013) to bind as
much APol as the lipoprotein complexes studied by Gohon et al. (2008), which is probably itself an
overestimate.
Everything considered, it seems therefore reasonable to conclude that the most reliable data
seem to point to MP/A8-35 complexes being smaller than MP/DDM ones.
The most detailed structural study to date of a MP extracted by SMA is that of the AcrB trimer
(Study 5.39 in Table 5.4). As mentioned above, this preparation contains relatively few lipids, in
insufficient number to form a full annulus (Postis et al. 2015). In keeping with these analyses, fitting the
crystallographic structure of AcrB into the low-resolution envelope obtained by single-particle analysis
of EM images of the negatively stained complex reveals some extra material surrounding the TM
region of the trimer, attributed to the polymer and lipids, but no lipid disc (Fig. 5.20).
Relatively little is known of the properties of MP/NVoy complexes. In Study 5.27 (Table 5.4), two
NVoy-trapped GPCRs, CRFR1 and CRFR2β, of respective molecular masses, ~47 and ~49 kDa, were
studied by SEC combined with ultraviolet (UV) absorbance, static light scattering, and refractive index
measurements (cf. Fig. 9.9 in Chap. 9). It was concluded that NVoy-trapped CRFR1 and CRFR2β
migrate as monomers, with experimental molecular masses of ~40 and ~48 kDa, respectively. Each of
them binds ~20 NVoy molecules, i.e. ~100 kDa of polymer (Klammt et al. 2011). This is about twice
more than the mass of A8-35 (~54 kDa) bound by BR, a protein that also features seven TM helices,
and comparable that of BR-bound NAPol, another glycosylated polymer (~97 kDa) (Table 5.5).
282
5 Formation and Properties of Membrane Protein/Amphipol Complexes
the complexes (Gohon et al. 2008), presumably because of electrostatic interactions between the
resin and the polyanionic APol, given that no such inconsistency is observed with non-ionic APols
(Sharma et al. 2012) (see main text).
In a comparative single-particle study of negatively stained plant aquaporin SoPIP2;1
tetramers complexed by various surfactants, it was concluded that the surfactant belt is thicker in
APol-trapped complexes than in detergent solution: 11 Æ 2 Å in OG, 13 Æ 2 Å in DDM, 15 Æ 2 Å
in LMNG, and 19 Æ 3 Å in A8-35 (Vahedi-Faridi et al. 2013; Fig. 5.19). However, the definition of
the envelope of the reconstructed image depended on an estimate of the amount of bound surfactant.
In the case of A8-35, this was taken to be ~220 kDa (taking the measured binding to BR as a basis
and assuming the mass of APol bound to be proportional to that of the protein), whereas, on the
basis of the data collected in Table 5.5 and Fig. 5.13, ~90 kDa would seem more likely given the
size of the TM region of aquaporins. The thickness of the belt deduced from this work is therefore
almost certain to be overestimated.
Based on the density of the dry surfactants, it is difficult to understand why MP/A8-35
complexes should be larger than MP/DDM ones, given that they generally comprise much less
surfactant and that the specific volumes of A8-35 and DDM are comparable (0.809 mLÁg
À1 for
A8-35 (Gohon et al. 2004), 0.81–0.837 mLÁg
À1 for DDM (le Maire et al. 2000)). However,
hydration must be taken into account. Water binding by DDM has been estimated, by either
equilibrium ultracentrifugation or SEC, to be in the range of 0.26–0.34 g water per g detergent
(de Vitry et al. 1991; Møller and le Maire 1993), that of A8-35, by a combination of biophysical
measurements, to be 1.22 Æ 0.45 g water per g polymer (Gohon et al. 2004). Combining the
extremes of each range, one concludes that the volume of hydrated DDM associated to 1 g of BR is
comprised between 3.9 and 4.7 cm
3
, that of hydrated A8-35 between 2.4 and 3.3 cm
3
. One would
therefore expect the BR-bound belt of A8-35 to be roughly between 0.5Â and 0.8Â as thick as the
DDM one.
Consistent with this view, a comparative NMR study of BR in DDM solution, DMPC
nanodiscs, or A8-35 concluded that the NMR spectra were comparable and all three systems were
adequate for solution NMR measurements (Etzkorn et al. 2013; cf. Chap. 10, Fig. 10.14). The work
included estimates of M and τ c values, which, at variance with SEC data, were consistent with BR/
A8-35 complexes being smaller than BR/DDM ones: M % 93 kDa in A8-35 vs. ~128 kDa in DDM,
τ c ¼ 31.7 vs. 41.4 ns. Because the BR/A8-35 preparation was partially aggregated, the data are
likely to actually overestimate the size of BR/A8-35 monomers. Indeed, on the basis of direct
measurements using [
3
H]A8-35, the mass of the monomeric complexes would be expected to be
~81 kDa (Gohon et al. 2008), assuming the lipid-free BR used by Etzkorn et al. (2013) to bind as
much APol as the lipoprotein complexes studied by Gohon et al. (2008), which is probably itself an
overestimate.
Everything considered, it seems therefore reasonable to conclude that the most reliable data
seem to point to MP/A8-35 complexes being smaller than MP/DDM ones.
The most detailed structural study to date of a MP extracted by SMA is that of the AcrB trimer
(Study 5.39 in Table 5.4). As mentioned above, this preparation contains relatively few lipids, in
insufficient number to form a full annulus (Postis et al. 2015). In keeping with these analyses, fitting the
crystallographic structure of AcrB into the low-resolution envelope obtained by single-particle analysis
of EM images of the negatively stained complex reveals some extra material surrounding the TM
region of the trimer, attributed to the polymer and lipids, but no lipid disc (Fig. 5.20).
Relatively little is known of the properties of MP/NVoy complexes. In Study 5.27 (Table 5.4), two
NVoy-trapped GPCRs, CRFR1 and CRFR2β, of respective molecular masses, ~47 and ~49 kDa, were
studied by SEC combined with ultraviolet (UV) absorbance, static light scattering, and refractive index
measurements (cf. Fig. 9.9 in Chap. 9). It was concluded that NVoy-trapped CRFR1 and CRFR2β
migrate as monomers, with experimental molecular masses of ~40 and ~48 kDa, respectively. Each of
them binds ~20 NVoy molecules, i.e. ~100 kDa of polymer (Klammt et al. 2011). This is about twice
more than the mass of A8-35 (~54 kDa) bound by BR, a protein that also features seven TM helices,
and comparable that of BR-bound NAPol, another glycosylated polymer (~97 kDa) (Table 5.5).
282
5 Formation and Properties of Membrane Protein/Amphipol Complexes