produced end up correctly folded (for reviews, see e.g. Grisshammer and Tate 1995; Maeda and
Schertler 2013; Goehring et al. 2014; Milić and Veprintsev 2015; Zorman et al. 2015).
Two alternatives are to express the target MP in vitro, in a cell-free expression (CFE) system, or
to target it to the cytosol of the cell, where it precipitates as nontoxic inclusion bodies (IBs). Either
approach can yield tens of mg of MPs, but coaxing these into adopting a functional structure remains a
major difficulty. In CFE, MPs can be provided with a surfactant into which they will insert and,
hopefully, fold during their synthesis, or they can be left to precipitate and be recovered later on by
dissolution with a detergent. The recovered MP is often nonfunctional. This approach will be discussed
in Chap. 7. In the second case, IBs are recovered by centrifugation, washed with a non-denaturing
detergent such as Triton X-100, and dissolved either with a chaotropic agent or a denaturing detergent,
typically urea for β-barrel MPs and sodium dodecyl sulfate (SDS) for α-helical ones. Obtaining tens or
hundreds of mg of MP in this way is relatively easy. Folding the recovered protein to a functional form
is nothing short of nightmarish. Many procedures have been described that lead from an unfolded to a
folded and functional MP, usually with limited yields (§ 6.2). Assisting the folding of MPs appears to
be one of the most successful and promising applications of amphipols (APols).
6.2
Context: Existing Approaches to Folding Membrane Proteins In Vitro
The first successful attempts at folding MPs in vitro were reported in the late 1970s/early 1980s. They
were largely a consequence of ongoing attempts at establishing the sequence of MPs by chemical
methods. Because DNA sequencing did not exist at the time, this required cutting chemically or
enzymatically the protein into short, overlapping peptides, sequencing them manually and putting
together the whole puzzle to obtain the complete sequence. For α-helical MPs, this raised particularly
redoubtable problems, because most transmembrane (TM) helices correspond to long stretches of
hydrophobic residues, which are highly insoluble in aqueous solutions and hard to handle. Organic
solvents had to be identified in which to separate and purify them. The technology thus developed
made it possible to obtain MPs in a fully unfolded state, as ascertained by spectroscopic methods.
Beyond sequencing, this work thus sets the stage for tackling the question of whether such unfolded
MPs could be brought back to a functional form.
The first MP refolding experiment on record appears to have been published by the group of Ulf
Henning in 1978. It was carried out on outer membrane protein A (OmpA, then called OmpII*) from
E. coli and took advantage of the fact that most β-barrel MPs migrate differently, upon
SDS-polyacrylamide gel electrophoresis (SDS-PAGE), depending on whether they have been heated
in SDS, and presumably unfolded, or kept at room temperature. Henning and coworkers showed that
addition of the outer membrane lipid lipopolysaccharide to the boiled species after it had been cooled
down caused it to migrate again at the same position as the unboiled species, strongly suggesting a
denaturation/renaturation phenomenon (Schweizer et al. 1978).
In 1981, the group of H. Gobind Khorana at MIT published a memorable paper in which they
demonstrated that bacteriorhodopsin (BR; see Chap. 1, § 1.6.1) could be refolded to a functional form
starting from a denatured form in SDS (Huang et al. 1981). When overproduced, BR, a light-driven proton
pump, accumulates in the plasma membrane of Halobacterium salinarum in the form of 2D protein/lipid
crystals, the so-called purple membrane. Its chromophore, retinal, which is covalently but loosely bound to
a lysine residue by a Schiff base, confers it a characteristic purple color (for an overview of BR structure
and function, see Chap. 1, § 1.6.1). When purple membrane is solubilized in SDS, BR denatures, and the
Schiff base hydrolyzes, releasing the retinal and bacterio-opsin (BO), the apoprotein. This causes an
absorbance peak shift from ~555 nm (dark-adapted BR) to ~382 nm (free retinal). The protocols developed
by Khorana and his coworkers are schematized in Fig. 6.1. In SDS, BO features a substantial amount of
334
6 Amphipol-Assisted Folding of Membrane Proteins
Schertler 2013; Goehring et al. 2014; Milić and Veprintsev 2015; Zorman et al. 2015).
Two alternatives are to express the target MP in vitro, in a cell-free expression (CFE) system, or
to target it to the cytosol of the cell, where it precipitates as nontoxic inclusion bodies (IBs). Either
approach can yield tens of mg of MPs, but coaxing these into adopting a functional structure remains a
major difficulty. In CFE, MPs can be provided with a surfactant into which they will insert and,
hopefully, fold during their synthesis, or they can be left to precipitate and be recovered later on by
dissolution with a detergent. The recovered MP is often nonfunctional. This approach will be discussed
in Chap. 7. In the second case, IBs are recovered by centrifugation, washed with a non-denaturing
detergent such as Triton X-100, and dissolved either with a chaotropic agent or a denaturing detergent,
typically urea for β-barrel MPs and sodium dodecyl sulfate (SDS) for α-helical ones. Obtaining tens or
hundreds of mg of MP in this way is relatively easy. Folding the recovered protein to a functional form
is nothing short of nightmarish. Many procedures have been described that lead from an unfolded to a
folded and functional MP, usually with limited yields (§ 6.2). Assisting the folding of MPs appears to
be one of the most successful and promising applications of amphipols (APols).
6.2
Context: Existing Approaches to Folding Membrane Proteins In Vitro
The first successful attempts at folding MPs in vitro were reported in the late 1970s/early 1980s. They
were largely a consequence of ongoing attempts at establishing the sequence of MPs by chemical
methods. Because DNA sequencing did not exist at the time, this required cutting chemically or
enzymatically the protein into short, overlapping peptides, sequencing them manually and putting
together the whole puzzle to obtain the complete sequence. For α-helical MPs, this raised particularly
redoubtable problems, because most transmembrane (TM) helices correspond to long stretches of
hydrophobic residues, which are highly insoluble in aqueous solutions and hard to handle. Organic
solvents had to be identified in which to separate and purify them. The technology thus developed
made it possible to obtain MPs in a fully unfolded state, as ascertained by spectroscopic methods.
Beyond sequencing, this work thus sets the stage for tackling the question of whether such unfolded
MPs could be brought back to a functional form.
The first MP refolding experiment on record appears to have been published by the group of Ulf
Henning in 1978. It was carried out on outer membrane protein A (OmpA, then called OmpII*) from
E. coli and took advantage of the fact that most β-barrel MPs migrate differently, upon
SDS-polyacrylamide gel electrophoresis (SDS-PAGE), depending on whether they have been heated
in SDS, and presumably unfolded, or kept at room temperature. Henning and coworkers showed that
addition of the outer membrane lipid lipopolysaccharide to the boiled species after it had been cooled
down caused it to migrate again at the same position as the unboiled species, strongly suggesting a
denaturation/renaturation phenomenon (Schweizer et al. 1978).
In 1981, the group of H. Gobind Khorana at MIT published a memorable paper in which they
demonstrated that bacteriorhodopsin (BR; see Chap. 1, § 1.6.1) could be refolded to a functional form
starting from a denatured form in SDS (Huang et al. 1981). When overproduced, BR, a light-driven proton
pump, accumulates in the plasma membrane of Halobacterium salinarum in the form of 2D protein/lipid
crystals, the so-called purple membrane. Its chromophore, retinal, which is covalently but loosely bound to
a lysine residue by a Schiff base, confers it a characteristic purple color (for an overview of BR structure
and function, see Chap. 1, § 1.6.1). When purple membrane is solubilized in SDS, BR denatures, and the
Schiff base hydrolyzes, releasing the retinal and bacterio-opsin (BO), the apoprotein. This causes an
absorbance peak shift from ~555 nm (dark-adapted BR) to ~382 nm (free retinal). The protocols developed
by Khorana and his coworkers are schematized in Fig. 6.1. In SDS, BO features a substantial amount of
334
6 Amphipol-Assisted Folding of Membrane Proteins