Abstract:
Proteins are crucial and substantial for life. Their functions range from transporting
molecules, being responsible for muscle movement to structural functions, e.g., in
hair. Thereby, the form of the proteins strongly determines their function. Proteins
are commonly divided into four subgroups (globular, disordered, fibrous, and membrane
proteins) based on their structure with globular and membrane proteins being
the focus in this work. Generally, globular proteins are spherical, water-soluble proteins
and can show a rich phase behavior, including aggregation, liquid-liquid phase
separation (LLPS), reentrant condensation (RC), and crystallization upon adding
multivalent ions. Hence, understanding and being able to manipulate the phase behavior
of proteins (e.g., through the addition of multivalent salts) is vital for a wide
field of research. On the one hand, protein condensation is desired, since protein
crystals are the basis of most structural resolution of proteins, giving hints toward
their function. On the other hand, for example the formation of amyloid fibrils
within the human body may result in diseases such as Alzheimer.
This dissertation aims to fulfill two main goals: As a first step, characterizing the
respective protein-multivalent ion system, including determining the interactions
and phase behavior, and as a second step, determining the respective crystallization
pathway, i.e., the kinetics. In this regard, several theories exist that attempt to describe
and predict the crystallization process (classical and nonclassical). However,
these theories cannot be applied to every system, since no general crystallization
pathway exists and even similar systems can crystallize following different crystallization
pathways. In a classical process, the two parameters, structure and density,
change simultaneously during crystallization, whereas in nonclassical crystallization,
some kind of intermediate phase (such as the dense liquid phase after LLPS or clusters)
exists in between the initial solution and the final crystal. To elucidate the
protein phase behavior and crystallization pathway, different systems of negatively
charged globular proteins in the presence of multivalent cations were investigated.
Considerable parts of the present thesis are based on three own publications in peerreviewed
scientific journals.
In the beginning, the model systems containing the protein human serum albumin
(HSA) with the salts CeCl3 and YCl3 were studied in detail to gain an overview of
the phase behavior. Further, conditions inside and outside the region in the phase
diagram where LLPS occurs were investigated in order to determine the role of the
dense, protein-rich phase on the (nonclassical) crystallization pathway. A proposed
nucleation within the dense phase, as suggested by simulations and theory, was tested for its validity by the experiments performed for this system. In contrast
to the predictions, the data led to the surprising conclusion that nucleation takes
place within the dilute, protein-poor phase, and that the dense phase only acts as
a reservoir. During crystal growth within the dilute phase, this reservoir of dense
phase dissolves and facilitates further crystal growth.
Second, the two homologous proteins HSA and bovine serum albumin (BSA) were
studied with CeCl3 in solution and at a net negatively charged, hydrophilic surface
to investigate their phase and adsorption behavior. These are crucial experiments
to investigate the differences and similarities between these extremely similar proteins
as well as to determine the driving force of crystallization, since only HSA
was reported to crystallize in the presence of multivalent ions. While both proteins
showed a similar phase behavior including RC and LLPS, the LLPS binodal was
shifted toward much higher protein concentrations (cp) in the case of BSA. These
results indicate weaker intermolecular attractions for BSA, which was further supported
with small-angle X-ray scattering (SAXS) experiments. Consistent with the
literature, only HSA crystallized within the condensed regime in the respective phase
diagram. Since HSA is slightly more hydrophobic than BSA, this behavior was attributed
to additional hydrophobic interactions of HSA, which not only strengthen
the attractions, but also provide additional oriented protein-protein contacts due to
the hydrophobic patches, favoring crystallization. In turn, quartz-crystal microbalance
with dissipation (QCM-D) data show that at a hydrophilic surface, a higher
amount of the more hydrophilic BSA was adsorbed. Hence, even though the general
protein phase behavior is strongly determined by electrostatic interaction due to ion
binding and ion bridging, additional interactions such as hydrophobic interactions
cannot be neglected, in particular when it comes to crystallization.
Third, the model system β-lactoglobulin (BLG) with CdCl2 was studied in D2O by
real-time small-angle neutron scattering (SANS) and optical microscopy to shed light
on the role of a preordered, metastable intermediate phase (MIP) in a nonclassical
two-step crystallization process. With these techniques, it is possible to characterize
the MIP during crystallization. A broad SANS peak was visible directly after preparation,
indicating a correlation of the proteins within the clusters. Since the quantity
of MIP and the characteristic length scale within these clusters evolved during crystallization,
it was concluded that the proteins can still rearrange/reorient within the
MIP, but that preordered domains with a length scale slightly larger than typical
protein-protein distances within the crystal lattice serve as crystal precursors. In
contrast to the dense droplets of the HSA system in the first result part, this MIP
serves as a crystal precursor and due to its preordered nature, the system follows a
completely different nonclassical crystallization pathway.
Finally, a topic that is not necessarily connected to the phase behavior and nonclassical
crystallization of proteins was approached. Here, so-called nanodiscs, which
are composed of a lipid bilayer surrounded by a polymer belt, were studied. These nanodiscs are a promising tool to investigate membrane proteins in their native
state, including techniques amenable for soluble proteins as well as crystallization
and structural resolution of the membrane proteins. However, a complete characterization
of the behavior of the nanodiscs is still lacking. For this purpose, different
compositions of two polymers (styrene maleic acid (SMA) and diisobutylene maleic
acid (DIBMA)) with the same lipid dimyristoyl-phosphocholine (DMPC) were mixed
and characterized by dynamic light scattering (DLS) and SAXS. Using the two different
polymers resulted in different structural arrangements of the nanodiscs. In
addition, structural effects due to increasing polymer concentration and decreasing
temperature were investigated. These results enable the biophysical and chemical
investigation of membrane proteins in a much more controlled way.
Print-on-Demand