The study was then extended to three other GPCRs: another leukotriene receptor, BLT2, the
serotonin receptor 5-HT 4(a) , and the cannabinoid receptor CB1 (Dahmane et al. 2009). In the presence
of asolectin, A8-35 improved the folding yields of 5-HT 4(a) previously observed by more than a factor
of two, namely from 20–25% in detergent/asolectin micelles (Banères et al. 2005) to ~30% in pure
A8-35 and ~60% in A8-35 + asolectin (Dahmane et al. 2009). For BLT2, which shares ~45% sequence
identity with BLT1, the yield improved from 3–4% in detergent/asolectin micelles to ~50% in pure
A8-35 and ~70% in A8-35 + asolectin (Dahmane et al. 2009). As regards CB1, which had not been
folded to any significant extent in detergent/lipid mixtures at the time of these experiments (but has
been folded to ~30% since; see Michalke et al. 2010), folding yields of ~30% were achieved in pure
A8-35 and of ~40% in A8-35/asolectin mixtures. Altogether, these data, obtained without any
extensive search for optimal folding conditions, suggest an interesting potential of APols as a generally
useful new tool for the folding of GPCRs. Indeed, more recently, another GPCR, the ghrelin GHSR-1a
receptor, has been successfully refolded in APols (Table 6.1, line 18, and § 6.3.1.1.3) (Banères et al.
2011; Bazzacco et al. 2012). Folding of a sixth GPCR, the type 2 arginine-vasopressin receptor, has
been reported in Banères et al. (2011), but experimental details have yet to be published.
Whereas in no case was a 100% yield reached, separation of active from inactive 5-HT 4(a) can be
achieved using a GR113808 affinity column (Banères et al. 2005), yielding ~96% active receptor
(Dahmane et al. 2009). In this context, it is worth noting that, whereas 3D crystallization of APoltrapped MPs remains a difficult challenge (see Chap. 11, § 11.3.1), A8-35-trapped BR has yielded
highly organized crystals (diffracting to <2-Å resolution) following direct transfer to lipidic
mesophases (Polovinkin et al. 2014) (Chap. 11, § 11.3.2). This observation has been recently
reproduced using a BR variant trapped in styrene-maleic acid copolymer (SMA) (Broecker et al.
2017). Crystallization of GPCRs in mesophases has proven remarkably successful (see e.g. Cherezov
et al. 2007; Rasmussen et al. 2011, and references therein), and it has been extended to a large variety
of MPs (for reviews, see Cherezov et al. 2006; Caffrey 2011; Cherezov 2011; Ishchenko et al. 2014).
Combining APol-assisted folding of GPCRs or other MPs with crystallization in mesophases, which
does not require very pure MPs, might therefore open very interesting perspectives for MP crystallization in general (see Chap. 11).
Box 6.2 Why Is There an Optimal Amphipol/Protein Ratio for Folding the BLT1
Receptor in Pure A8-35?
As shown in Fig. 6.7 (left), better yields are observed when folding the BLT1 receptor in pure
A8-35 at ~1:5 BLT1/A8-35 mass ratio than either with less (1:1) or with more (1:10 or 1:20) APol
present. How to account for this observation?
As regards the 1:1 w/w ratio, the most likely hypothesis is that the protein is not diluted
enough in the surfactant “phase” to prevent unproductive intermolecular interactions from forming
in the course of folding, preventing correct folding. Indeed, as shown in § 6.3.1.1.1, when BR is
folded at too low an APol/protein ratio, the yield drops (Fig. 6.6A), and part of the folded protein is
found in aggregated form (Fig. 6.6B).
As regards the drop of yield at high APol/BLT1 ratios, the explanation is less straightforward. If a stabilizing cofactor (prosthetic group, lipid, etc.) were diluted by the APol, one could
indeed expect a drop in yield at high APol/protein ratios. However, the system here comprises only
the protein, the APol, and, initially, dodecyl sulfate, the concentration of which in the aqueous
solution and in the APol particles during the initial stages of folding is set by the concentration of
free potassium ions. Dilution with further APol should not change the environment of the folding
protein. We can consider two types of interpretation:
• BLT1 is known to dimerize (Mesnier and Banères 2004). Upon SEC analysis of the
A8-35-folded BLT1, at least part of it does seem to migrate as a dimer rather than a
344
6 Amphipol-Assisted Folding of Membrane Proteins
serotonin receptor 5-HT 4(a) , and the cannabinoid receptor CB1 (Dahmane et al. 2009). In the presence
of asolectin, A8-35 improved the folding yields of 5-HT 4(a) previously observed by more than a factor
of two, namely from 20–25% in detergent/asolectin micelles (Banères et al. 2005) to ~30% in pure
A8-35 and ~60% in A8-35 + asolectin (Dahmane et al. 2009). For BLT2, which shares ~45% sequence
identity with BLT1, the yield improved from 3–4% in detergent/asolectin micelles to ~50% in pure
A8-35 and ~70% in A8-35 + asolectin (Dahmane et al. 2009). As regards CB1, which had not been
folded to any significant extent in detergent/lipid mixtures at the time of these experiments (but has
been folded to ~30% since; see Michalke et al. 2010), folding yields of ~30% were achieved in pure
A8-35 and of ~40% in A8-35/asolectin mixtures. Altogether, these data, obtained without any
extensive search for optimal folding conditions, suggest an interesting potential of APols as a generally
useful new tool for the folding of GPCRs. Indeed, more recently, another GPCR, the ghrelin GHSR-1a
receptor, has been successfully refolded in APols (Table 6.1, line 18, and § 6.3.1.1.3) (Banères et al.
2011; Bazzacco et al. 2012). Folding of a sixth GPCR, the type 2 arginine-vasopressin receptor, has
been reported in Banères et al. (2011), but experimental details have yet to be published.
Whereas in no case was a 100% yield reached, separation of active from inactive 5-HT 4(a) can be
achieved using a GR113808 affinity column (Banères et al. 2005), yielding ~96% active receptor
(Dahmane et al. 2009). In this context, it is worth noting that, whereas 3D crystallization of APoltrapped MPs remains a difficult challenge (see Chap. 11, § 11.3.1), A8-35-trapped BR has yielded
highly organized crystals (diffracting to <2-Å resolution) following direct transfer to lipidic
mesophases (Polovinkin et al. 2014) (Chap. 11, § 11.3.2). This observation has been recently
reproduced using a BR variant trapped in styrene-maleic acid copolymer (SMA) (Broecker et al.
2017). Crystallization of GPCRs in mesophases has proven remarkably successful (see e.g. Cherezov
et al. 2007; Rasmussen et al. 2011, and references therein), and it has been extended to a large variety
of MPs (for reviews, see Cherezov et al. 2006; Caffrey 2011; Cherezov 2011; Ishchenko et al. 2014).
Combining APol-assisted folding of GPCRs or other MPs with crystallization in mesophases, which
does not require very pure MPs, might therefore open very interesting perspectives for MP crystallization in general (see Chap. 11).
Box 6.2 Why Is There an Optimal Amphipol/Protein Ratio for Folding the BLT1
Receptor in Pure A8-35?
As shown in Fig. 6.7 (left), better yields are observed when folding the BLT1 receptor in pure
A8-35 at ~1:5 BLT1/A8-35 mass ratio than either with less (1:1) or with more (1:10 or 1:20) APol
present. How to account for this observation?
As regards the 1:1 w/w ratio, the most likely hypothesis is that the protein is not diluted
enough in the surfactant “phase” to prevent unproductive intermolecular interactions from forming
in the course of folding, preventing correct folding. Indeed, as shown in § 6.3.1.1.1, when BR is
folded at too low an APol/protein ratio, the yield drops (Fig. 6.6A), and part of the folded protein is
found in aggregated form (Fig. 6.6B).
As regards the drop of yield at high APol/BLT1 ratios, the explanation is less straightforward. If a stabilizing cofactor (prosthetic group, lipid, etc.) were diluted by the APol, one could
indeed expect a drop in yield at high APol/protein ratios. However, the system here comprises only
the protein, the APol, and, initially, dodecyl sulfate, the concentration of which in the aqueous
solution and in the APol particles during the initial stages of folding is set by the concentration of
free potassium ions. Dilution with further APol should not change the environment of the folding
protein. We can consider two types of interpretation:
• BLT1 is known to dimerize (Mesnier and Banères 2004). Upon SEC analysis of the
A8-35-folded BLT1, at least part of it does seem to migrate as a dimer rather than a
344
6 Amphipol-Assisted Folding of Membrane Proteins