pressure-increasing lipid components) promote or inhibit the formation of such
supercomplexes. Likewise, the composition of the lipid membrane and the membrane properties such as its curvature pressure most likely influence the folding of
the LHCII apoprotein and its assembly with pigments.
LHCII commends itself as a useful model for studying the influence of the lipid
membrane on the assembly and structural behavior of membrane proteins in general
because of its known structure, its availability in a recombinant form, and its selforganization, at least in detergent micelles. Moreover, the bound Chl molecules
serve as built-in fluorescence markers for monitoring the structural behavior of the
pigment–protein complex. To be able to correlate experimental observations of
aggregate formation with predictions from theory, recombinant LHCII has been
inserted in liposomes and assayed for complex–complex distances by intercomplex
FRET measurements, and for aggregate formation by quantifying aggregateinduced fluorescence quenching.
A multiscale simulation model to study the LHCII complex requires, as a first
step, model parameters for all components involved. As already mentioned above,
it will be neither possible nor useful to parameterize a CG model based on the actual
multicomponent (lipid bilayer/protein/pigments) system but one would rather
develop models for sensibly chosen subsystems. Although parameters for the
protein and the lipid bilayer can typically be found in many standard force fields,
a challenging first task is to obtain a reliable model for the pigments, irrespective of
the level of resolution. For many biological applications, the MARTINI CG force
field (described above) has become very popular and successful, in particular for
lipid bilayer and protein systems. To employ the MARTINI force field for simulations of the pigmented LHCII, a CG description and model parameters for the
pigment molecules needs to be added. We have developed a coarse-grained model
of the chlorophyll pigments (Chlb and Chla) that can be embedded into the existing
MARTINI force field to study the pigmented LHCII trimer in the future. To do this,
Chlb and Chla were parameterized in the presence of the lipid bilayer. This
reference system for parametrization was chosen for two reasons: most importantly,
the Chl–lipid interactions are highly relevant for the formation and behavior of the
LHCII protein–pigment complex in the lipid bilayer. About 50% of the pigment
molecules in the plant are bound to the light-harvesting complex, with 42 Chl
molecules per LHCII trimer. In vitro studies have shown that the folding of the
LHCII apoprotein and the pigment binding to the protein are tightly coupled
processes. In the LHCII monomer, many Chl pigments are situated in the outer
region of the protein, effectively forming an interface between protein and lipids.
Consequently, the Chl–lipid interactions are probably important for the assembly
and stability of the trimer. A second reason for choosing the Chl–lipid system as
reference for which the interactions between the MARTINI standard forcefield and
Chl can be tuned is that it is more tractable compared to the fully pigmented LHCII
membrane protein complex. The CG model for Chlb and Chla in the DPPC bilayer
was derived from a combination of a structure-based approach for bonded interaction potentials and a mixed structure-based and partitioning-based approach for
268
M. Deserno et al.
supercomplexes. Likewise, the composition of the lipid membrane and the membrane properties such as its curvature pressure most likely influence the folding of
the LHCII apoprotein and its assembly with pigments.
LHCII commends itself as a useful model for studying the influence of the lipid
membrane on the assembly and structural behavior of membrane proteins in general
because of its known structure, its availability in a recombinant form, and its selforganization, at least in detergent micelles. Moreover, the bound Chl molecules
serve as built-in fluorescence markers for monitoring the structural behavior of the
pigment–protein complex. To be able to correlate experimental observations of
aggregate formation with predictions from theory, recombinant LHCII has been
inserted in liposomes and assayed for complex–complex distances by intercomplex
FRET measurements, and for aggregate formation by quantifying aggregateinduced fluorescence quenching.
A multiscale simulation model to study the LHCII complex requires, as a first
step, model parameters for all components involved. As already mentioned above,
it will be neither possible nor useful to parameterize a CG model based on the actual
multicomponent (lipid bilayer/protein/pigments) system but one would rather
develop models for sensibly chosen subsystems. Although parameters for the
protein and the lipid bilayer can typically be found in many standard force fields,
a challenging first task is to obtain a reliable model for the pigments, irrespective of
the level of resolution. For many biological applications, the MARTINI CG force
field (described above) has become very popular and successful, in particular for
lipid bilayer and protein systems. To employ the MARTINI force field for simulations of the pigmented LHCII, a CG description and model parameters for the
pigment molecules needs to be added. We have developed a coarse-grained model
of the chlorophyll pigments (Chlb and Chla) that can be embedded into the existing
MARTINI force field to study the pigmented LHCII trimer in the future. To do this,
Chlb and Chla were parameterized in the presence of the lipid bilayer. This
reference system for parametrization was chosen for two reasons: most importantly,
the Chl–lipid interactions are highly relevant for the formation and behavior of the
LHCII protein–pigment complex in the lipid bilayer. About 50% of the pigment
molecules in the plant are bound to the light-harvesting complex, with 42 Chl
molecules per LHCII trimer. In vitro studies have shown that the folding of the
LHCII apoprotein and the pigment binding to the protein are tightly coupled
processes. In the LHCII monomer, many Chl pigments are situated in the outer
region of the protein, effectively forming an interface between protein and lipids.
Consequently, the Chl–lipid interactions are probably important for the assembly
and stability of the trimer. A second reason for choosing the Chl–lipid system as
reference for which the interactions between the MARTINI standard forcefield and
Chl can be tuned is that it is more tractable compared to the fully pigmented LHCII
membrane protein complex. The CG model for Chlb and Chla in the DPPC bilayer
was derived from a combination of a structure-based approach for bonded interaction potentials and a mixed structure-based and partitioning-based approach for
268
M. Deserno et al.