How nHDL grows into mature sHDL is beyond the frame of the present book but is worth
mentioning cursorily. A cytosolic enzyme, lecithin cholesterol acetyltransferase, transforms free
cholesterol into cholesteryl esters, which are highly hydrophobic and move to the core of the particle,
joined by equally hydrophobic triglycerides delivered by other proteins. As the particle grows,
phospholipid acyl chains tend to become exposed to its surface, which entails the recruitment of
additional molecules of apoA-I and/or other proteins. The end result of this process is a heterogeneous
collection of particles of various sizes, comprising variable proportions of protein and of bilayerforming and non-bilayer-forming lipids, illustrating, once more, the flexibility and adaptability of
apoA-I.
At the onset of their work (Carlson et al. 1997; Bayburt et al. 1998, 2002), S. G. Sligar and his
colleagues intended to use nHDL to stabilize in a water-soluble form well-defined patches of lipid
bilayer, whose surface would be used for in vitro experiments. It took some engineering of the apoA-I
sequence before well-defined lipid patches suitable for incorporating and studying guest MPs were
obtained (Denisov et al. 2004) and the efficiency, generality, and versatility of the approach became
apparent (Denisov and Sligar 2017).
3.3.2
The Formation and Structure of Nanodiscs
NDs are generally obtained following a protocol derived from that used to form reconstituted nHDL
(Jonas 1986; Jonas et al. 1990, 1991): an engineered version of apoA-I is mixed with lipids in detergent
solution in the proportion appropriate to the size of the discs one aims to produce, and the detergent is
removed. In the original procedure (Bayburt et al. 1998), the protein destined to stabilize the discs
(hereafter called “membrane scaffold protein,” MSP) was human apoA-I and the lipid DPPC
(diC 16:0 PC). The detergent was sodium cholate, which was removed by dialysis. A target MP, P450
reductase, was also included and ended up trapped within the NDs. Nowadays, NDs are systematically
prepared using specially engineered MSPs.
As reviewed in Viegas et al. (2016) and Denisov and Sligar (2017), all of these parameters can be
varied. Cholate can be replaced with almost any detergent, including OG, Triton X-100, decyl- or
DDM, diC 6 PC, CHAPS, and even sodium dodecylsulfate (SDS), an obvious caveat being that any MP
that one aims to capture should not denature in it or must be able to refold when the detergent is
eliminated. In practice, it is frequent that the mixture of MSPs and lipids solubilized in one detergent be
mixed with the guest MP solubilized in another so that one deals with a mixture of detergents. Various
protocols can be used for detergent removal. For detergents with a low CMC, like DDM, adsorption
onto Bio-Beads has to be preferred to dialysis, which is too slow. A vast variety of lipids and lipid
mixtures have been resorted to, including synthetic lipids such as DMPC, DPPC, and palmitoyloleylphosphatidylcholine (POPC, C 16:0 , C 18:1 PC), as well as mixtures comprising charged phospholipids
and lipid mixtures from natural sources, such as Escherichia coli polar or total lipids, egg PC, or
asolectin (soybean lipids). A key point is to carefully adjust the MSP/lipid ratio, so as to obtain
homogeneous preparations of NDs with the desired size. Detailed protocols are available from the
Sligar laboratory web site (http://sligarlab.life.uiuc.edu/nanodisc/protocols.html; see also Chap. 5,
§ 5.9.3, Protocol 5.3, about transferring a MP from amphipols to NDs).
ApoA-I itself does not produce monodisperse nHDL particles (Durbin and Jonas 1997; Li et al.
2004). Considerable efforts have therefore been invested to engineer it so as to obtain optimized MSPs
yielding homogeneous NDs of a variety of sizes and carrying a variety of tags (Fig. 3.7). Five types of
modifications have been experimented with. First, it was soon recognized that some of the N-terminal
amino acid residues do not take part in forming the ND belt, which led to the development of a series of
MSPs from which the first 11 or 22 residues of apoA-I have been deleted (e.g. MSP1D1 in Fig. 3.7)
3.3 Nanodiscs
107
mentioning cursorily. A cytosolic enzyme, lecithin cholesterol acetyltransferase, transforms free
cholesterol into cholesteryl esters, which are highly hydrophobic and move to the core of the particle,
joined by equally hydrophobic triglycerides delivered by other proteins. As the particle grows,
phospholipid acyl chains tend to become exposed to its surface, which entails the recruitment of
additional molecules of apoA-I and/or other proteins. The end result of this process is a heterogeneous
collection of particles of various sizes, comprising variable proportions of protein and of bilayerforming and non-bilayer-forming lipids, illustrating, once more, the flexibility and adaptability of
apoA-I.
At the onset of their work (Carlson et al. 1997; Bayburt et al. 1998, 2002), S. G. Sligar and his
colleagues intended to use nHDL to stabilize in a water-soluble form well-defined patches of lipid
bilayer, whose surface would be used for in vitro experiments. It took some engineering of the apoA-I
sequence before well-defined lipid patches suitable for incorporating and studying guest MPs were
obtained (Denisov et al. 2004) and the efficiency, generality, and versatility of the approach became
apparent (Denisov and Sligar 2017).
3.3.2
The Formation and Structure of Nanodiscs
NDs are generally obtained following a protocol derived from that used to form reconstituted nHDL
(Jonas 1986; Jonas et al. 1990, 1991): an engineered version of apoA-I is mixed with lipids in detergent
solution in the proportion appropriate to the size of the discs one aims to produce, and the detergent is
removed. In the original procedure (Bayburt et al. 1998), the protein destined to stabilize the discs
(hereafter called “membrane scaffold protein,” MSP) was human apoA-I and the lipid DPPC
(diC 16:0 PC). The detergent was sodium cholate, which was removed by dialysis. A target MP, P450
reductase, was also included and ended up trapped within the NDs. Nowadays, NDs are systematically
prepared using specially engineered MSPs.
As reviewed in Viegas et al. (2016) and Denisov and Sligar (2017), all of these parameters can be
varied. Cholate can be replaced with almost any detergent, including OG, Triton X-100, decyl- or
DDM, diC 6 PC, CHAPS, and even sodium dodecylsulfate (SDS), an obvious caveat being that any MP
that one aims to capture should not denature in it or must be able to refold when the detergent is
eliminated. In practice, it is frequent that the mixture of MSPs and lipids solubilized in one detergent be
mixed with the guest MP solubilized in another so that one deals with a mixture of detergents. Various
protocols can be used for detergent removal. For detergents with a low CMC, like DDM, adsorption
onto Bio-Beads has to be preferred to dialysis, which is too slow. A vast variety of lipids and lipid
mixtures have been resorted to, including synthetic lipids such as DMPC, DPPC, and palmitoyloleylphosphatidylcholine (POPC, C 16:0 , C 18:1 PC), as well as mixtures comprising charged phospholipids
and lipid mixtures from natural sources, such as Escherichia coli polar or total lipids, egg PC, or
asolectin (soybean lipids). A key point is to carefully adjust the MSP/lipid ratio, so as to obtain
homogeneous preparations of NDs with the desired size. Detailed protocols are available from the
Sligar laboratory web site (http://sligarlab.life.uiuc.edu/nanodisc/protocols.html; see also Chap. 5,
§ 5.9.3, Protocol 5.3, about transferring a MP from amphipols to NDs).
ApoA-I itself does not produce monodisperse nHDL particles (Durbin and Jonas 1997; Li et al.
2004). Considerable efforts have therefore been invested to engineer it so as to obtain optimized MSPs
yielding homogeneous NDs of a variety of sizes and carrying a variety of tags (Fig. 3.7). Five types of
modifications have been experimented with. First, it was soon recognized that some of the N-terminal
amino acid residues do not take part in forming the ND belt, which led to the development of a series of
MSPs from which the first 11 or 22 residues of apoA-I have been deleted (e.g. MSP1D1 in Fig. 3.7)
3.3 Nanodiscs
107