As shown throughout these results on dual protein systems, the experimental
conditions for optimum cross-assembly are specific for each system because of the
required charge and size compensation. This could explain the earlier results
reported in 1990 that showed that Lf interacts and forms complexes with β-Lg
and BSA but not with α-La [141]. A specific optimum pH value favouring Lf/α-La
cross-assembly was probably missing in this work. We have recently confirmed
this assumption by showing that, once the conditions are optimised, LF interacts
with α-La as well as with β-Lg, as revealed by isothermal titration calorimetry
experiments (unpublished data). Indeed, we also observed self-assemblies into
microspheres of β-Lg/LF, α-La/LF and β-Lg/LYS binary mixtures at specific,
system-dependent pH values (unpublished data).
The visualisation of protein microspheres by confocal microscopy showed that,
for all the dual systems described above, the two proteins are perfectly co-localised
in the three dimensions of the microsphere [5, 134]. This is probably related to the
fact that the building blocks initiating the cross-assembly are hetero-oligomers
(dimers, trimers, tetramers) formed by the two proteins involved, as shown experimentally in the case of apo α-La/LYS [135, 136].
Kinetically speaking, even if the reaction of spontaneous interaction–assembly
between proteins is very fast, experimental evidence shows that microspheres are
not formed immediately after mixing the two proteins [142]. At a given total protein
concentration, the organisation into microspheres is a dynamic, kinetically controlled process. This is well illustrated by the work conducted by our group on α-La
and LYS using confocal microscopy [142]. As shown in Fig. 7, branched
aggregates or “clusters of nanospheres” are formed rapidly after protein mixing.
Then, the clusters of nanospheres progressively re-organize into well-defined
spherical particles of a few micrometres when protein concentration is in the submillimolar range. In these experimental conditions, the process takes about 20 min
to reach the final particle organisation.
The formation of spherical structures in mixtures of oppositely charged globular
proteins seems to be a generic process. Beyond the binary systems, we have shown
that it is also possible to form microspheres by mixing three different proteins,
provided the mixing is performed in suitable proportions. This is the case, for
example, for a mixture containing negatively charged Ova and positively charged
avidin and LYS. This offers the possibility to design these new supramolecular
structures in complex protein mixtures. Interestingly, Sugimoto et al. [143] were
able to form amyloid-like fibrils in dual globular protein system including native
LYS and pre-denatured Ova. The fibrils were obtained at pH 7.5 with an initial
molar ratio LYS:Ova of 3. However, the exact stoichiometry recovered in the
formed fibrils was not indicated. This would be of interest for comparison with
the protein stoichiometry of 2 recovered in the LYS/Ova microspheres (see above).
Furthermore, the authors identified the exact peptidic sequence of Ova that interacts
with LYS. The self-assembly of various macromolecules such as proteins in
structures with very specific geometries requires the development of highly specific
interactions combining kinetic and thermodynamic aspects. Different models exist
that describe the laws of oriented molecular assembly between various
92
S. Bouhallab and T. Croguennec
conditions for optimum cross-assembly are specific for each system because of the
required charge and size compensation. This could explain the earlier results
reported in 1990 that showed that Lf interacts and forms complexes with β-Lg
and BSA but not with α-La [141]. A specific optimum pH value favouring Lf/α-La
cross-assembly was probably missing in this work. We have recently confirmed
this assumption by showing that, once the conditions are optimised, LF interacts
with α-La as well as with β-Lg, as revealed by isothermal titration calorimetry
experiments (unpublished data). Indeed, we also observed self-assemblies into
microspheres of β-Lg/LF, α-La/LF and β-Lg/LYS binary mixtures at specific,
system-dependent pH values (unpublished data).
The visualisation of protein microspheres by confocal microscopy showed that,
for all the dual systems described above, the two proteins are perfectly co-localised
in the three dimensions of the microsphere [5, 134]. This is probably related to the
fact that the building blocks initiating the cross-assembly are hetero-oligomers
(dimers, trimers, tetramers) formed by the two proteins involved, as shown experimentally in the case of apo α-La/LYS [135, 136].
Kinetically speaking, even if the reaction of spontaneous interaction–assembly
between proteins is very fast, experimental evidence shows that microspheres are
not formed immediately after mixing the two proteins [142]. At a given total protein
concentration, the organisation into microspheres is a dynamic, kinetically controlled process. This is well illustrated by the work conducted by our group on α-La
and LYS using confocal microscopy [142]. As shown in Fig. 7, branched
aggregates or “clusters of nanospheres” are formed rapidly after protein mixing.
Then, the clusters of nanospheres progressively re-organize into well-defined
spherical particles of a few micrometres when protein concentration is in the submillimolar range. In these experimental conditions, the process takes about 20 min
to reach the final particle organisation.
The formation of spherical structures in mixtures of oppositely charged globular
proteins seems to be a generic process. Beyond the binary systems, we have shown
that it is also possible to form microspheres by mixing three different proteins,
provided the mixing is performed in suitable proportions. This is the case, for
example, for a mixture containing negatively charged Ova and positively charged
avidin and LYS. This offers the possibility to design these new supramolecular
structures in complex protein mixtures. Interestingly, Sugimoto et al. [143] were
able to form amyloid-like fibrils in dual globular protein system including native
LYS and pre-denatured Ova. The fibrils were obtained at pH 7.5 with an initial
molar ratio LYS:Ova of 3. However, the exact stoichiometry recovered in the
formed fibrils was not indicated. This would be of interest for comparison with
the protein stoichiometry of 2 recovered in the LYS/Ova microspheres (see above).
Furthermore, the authors identified the exact peptidic sequence of Ova that interacts
with LYS. The self-assembly of various macromolecules such as proteins in
structures with very specific geometries requires the development of highly specific
interactions combining kinetic and thermodynamic aspects. Different models exist
that describe the laws of oriented molecular assembly between various
92
S. Bouhallab and T. Croguennec