scale of these experiments (see § 4.3.1.2.4), speaking against metastability. It does not seem that
the variable length of individual molecules be a critical factor in generating particle polydispersity,
since batches of A8-35 synthesized from precursor PAAs with either a broad (Ð % 3.1) or a very
narrow (Ð % 1.3) dispersity yielded particles with similar size distributions (C. Tribet and F. Giusti,
unpublished observations cited in Le Bon et al. 2014b). The hydrophobic/hydrophilic balance of
individual molecules and, possibly, the degree of randomness of octyl chain distribution are more
likely to determine the size of the resultant particles. This is an interesting observation, because it
suggests that, by restraining the molecule-to-molecule variability of the density and/or distribution of
octyl chains, narrower size distributions of APol particles and, as a consequence, of MP/APol
complexes could probably be achieved, an important factor for crystallization attempts. This hypothesis is consistent with SEC analyses showing that the particles formed by homotelomeric non-ionic
APols, which are comprised of molecules that vary in length but have a homogeneous composition,
have a much narrower size distribution than those formed by their heterotelomeric analogs, whose
composition and monomer distribution vary from molecule to molecule (see § 4.3.2).
Closer examination of the SEC data reveals some complications. First, there is evidence for some
polydisperse material eluting well after the main peak, beyond the position at which elutes horseradish
peroxidase, spreading almost to the total volume V t (Fig. 4.14A). This shallow, broad peak is observed
in all A8-35 preparations. It has not been studied in detail, but is suspected to arise from APol
molecules that are too small and/or not hydrophobic enough to assemble into particles. This region
also comprises the contribution from molecules that dissociate from the particles as they progress
through the column through APol-free buffer, but their concentration, equal to the CAC, is too low to
account for all of the absorbance measured and should be constant throughout.
Second, there is a very small peak eluting in the void volume V 0 of the column, pointing to the
presence of some very large objects. This population of particles is barely detectable in SEC, indicating
that they represent a very small fraction of the mass of the sample. It becomes, however, prominent in
radiation scattering experiments because, for a given molar concentration and composition, the
scattered intensity increases as the square of the molecular mass of the scatterers. It is, as noted
above, responsible for the peak at R S % 45 nm in DLS experiments (Fig. 4.14B) and for the upshot in
the SANS Guinier plots at very small angles (below 0.001 Å
À2 ; Fig. 4.14D). These objects are not
visible in AUC experiments (Fig. 4.14C) because they sediment too rapidly as compared to the small
particles. As shown by the SANS data, their proportion with respect to the small particles does not vary
with the concentration of the preparations (Fig. 4.14D), indicating the absence of an equilibrium. The
large objects, best characterized by DLS, can reach an average Stokes radius of up to ~60 nm and an
average molecular mass in excess of 1 MDa. Although conspicuous in scattering experiments, they
usually represent a very small mass fraction of the samples (~0.1% in the samples used in Gohon et al.
2006), and they do not associate with MPs (Gohon et al. 2008). They seldom interfere with
experiments but can, if their presence is a hindrance, be removed by either SEC fractionation or
ultracentrifugation.
As experience with A8-35 accumulated, a number of factors were found to compromise the
quasi-monodispersity of A8-35 particles and, correlatively, that of MP/A8-35 complexes:
• Artifactual covalent binding to the polymer, during the synthesis, of a hydrophobic contaminant can be a major source of problems. Some early batches of A8-35 exhibited complex SEC
elution patterns and curved Guinier plots (Gohon 2001; Gohon et al. 2006). This behavior
was characteristic of samples with a high degree (typically >6%) of artifactual grafting of
dicyclohexylurea (DCU), a by-product of the coupling reaction (see § 4.5, Protocol 4.1). The
broad size distribution found in such batches suggests that some APol molecules, presumably
because of a higher hydrophobicity and/or a nonrandom distribution of hydrophobic side
182
4 Chemical Structure, Synthesis, and Physical-Chemical Properties of Amphipols
the variable length of individual molecules be a critical factor in generating particle polydispersity,
since batches of A8-35 synthesized from precursor PAAs with either a broad (Ð % 3.1) or a very
narrow (Ð % 1.3) dispersity yielded particles with similar size distributions (C. Tribet and F. Giusti,
unpublished observations cited in Le Bon et al. 2014b). The hydrophobic/hydrophilic balance of
individual molecules and, possibly, the degree of randomness of octyl chain distribution are more
likely to determine the size of the resultant particles. This is an interesting observation, because it
suggests that, by restraining the molecule-to-molecule variability of the density and/or distribution of
octyl chains, narrower size distributions of APol particles and, as a consequence, of MP/APol
complexes could probably be achieved, an important factor for crystallization attempts. This hypothesis is consistent with SEC analyses showing that the particles formed by homotelomeric non-ionic
APols, which are comprised of molecules that vary in length but have a homogeneous composition,
have a much narrower size distribution than those formed by their heterotelomeric analogs, whose
composition and monomer distribution vary from molecule to molecule (see § 4.3.2).
Closer examination of the SEC data reveals some complications. First, there is evidence for some
polydisperse material eluting well after the main peak, beyond the position at which elutes horseradish
peroxidase, spreading almost to the total volume V t (Fig. 4.14A). This shallow, broad peak is observed
in all A8-35 preparations. It has not been studied in detail, but is suspected to arise from APol
molecules that are too small and/or not hydrophobic enough to assemble into particles. This region
also comprises the contribution from molecules that dissociate from the particles as they progress
through the column through APol-free buffer, but their concentration, equal to the CAC, is too low to
account for all of the absorbance measured and should be constant throughout.
Second, there is a very small peak eluting in the void volume V 0 of the column, pointing to the
presence of some very large objects. This population of particles is barely detectable in SEC, indicating
that they represent a very small fraction of the mass of the sample. It becomes, however, prominent in
radiation scattering experiments because, for a given molar concentration and composition, the
scattered intensity increases as the square of the molecular mass of the scatterers. It is, as noted
above, responsible for the peak at R S % 45 nm in DLS experiments (Fig. 4.14B) and for the upshot in
the SANS Guinier plots at very small angles (below 0.001 Å
À2 ; Fig. 4.14D). These objects are not
visible in AUC experiments (Fig. 4.14C) because they sediment too rapidly as compared to the small
particles. As shown by the SANS data, their proportion with respect to the small particles does not vary
with the concentration of the preparations (Fig. 4.14D), indicating the absence of an equilibrium. The
large objects, best characterized by DLS, can reach an average Stokes radius of up to ~60 nm and an
average molecular mass in excess of 1 MDa. Although conspicuous in scattering experiments, they
usually represent a very small mass fraction of the samples (~0.1% in the samples used in Gohon et al.
2006), and they do not associate with MPs (Gohon et al. 2008). They seldom interfere with
experiments but can, if their presence is a hindrance, be removed by either SEC fractionation or
ultracentrifugation.
As experience with A8-35 accumulated, a number of factors were found to compromise the
quasi-monodispersity of A8-35 particles and, correlatively, that of MP/A8-35 complexes:
• Artifactual covalent binding to the polymer, during the synthesis, of a hydrophobic contaminant can be a major source of problems. Some early batches of A8-35 exhibited complex SEC
elution patterns and curved Guinier plots (Gohon 2001; Gohon et al. 2006). This behavior
was characteristic of samples with a high degree (typically >6%) of artifactual grafting of
dicyclohexylurea (DCU), a by-product of the coupling reaction (see § 4.5, Protocol 4.1). The
broad size distribution found in such batches suggests that some APol molecules, presumably
because of a higher hydrophobicity and/or a nonrandom distribution of hydrophobic side
182
4 Chemical Structure, Synthesis, and Physical-Chemical Properties of Amphipols