Transmission of binding information across lipid bilayers.

Abstract

A synthetic transmembrane receptor that is capable of transmitting binding information across a lipid bilayer membrane is reported. The binding event is based on aggregation of the receptor triggered by copper(II) complexation to ethylenediamine functionalities. By labelling the receptor with fluorescent dansyl groups, the copper(II) binding event could be monitored by measuring the extent of fluorescence quenching. Comparing the receptor with a control receptor lacking the transmembrane linkage revealed that the transmembrane receptor binds copper(II) ions more tightly than the non-spanning control receptor at low copper(II) concentrations. Since the intrinsic binding to copper(II) is the same for both receptors, this effect was attributed to synergy between the connected interior and exterior binding sides of the transmembrane receptor. Thus, this is the first reported artificial signalling event in which binding of a messenger on one side of the membrane leads to a cooperative binding event on the opposite side of the membrane, resembling biological signalling systems and helping us to get a better understanding of the requirements for more effective artificial signalling systems.

Full text

Page 1

DOI: 10.1002/chem.200601723
Transmission of Binding Information across Lipid Bilayers
Harmen P. Dijkstra,[a, b]Jordan J. Hutchinson,[a]Christopher A. Hunter,*[a]
Haiyuan Qin,[a]Salvador Tomas,[a]Simon J. Webb,[a]and Nicholas H. Williams*[a]
Introduction
Cell membranes constitute an important barrier between
cells? internal fluid and the external medium and regulate
many important biological events.[1]They are dynamic, non-
covalent, fluid assemblies and many biological membrane
processes, mostly regulated by membrane proteins and car-
bohydrates, depend on the fluidity of the membrane lipids.[2]
Cell membranes provide a unique platform to sense, re-
spond to, and transduce signals and information. In these
biological signalling processes, cooperativity is believed to
play a crucial role. For many ligand/receptor interactions,
however, the strength of cooperative binding affinities is still
not well understood. In particular, cooperative binding of
receptors constrained at the cell membrane surface is still
largely unexplored.[3]In this case, the receptors bound at a
membrane surface can only move in two dimensions, thus
the binding interactions within membranes are expected to
be thermodynamically more favourable than binding inter-
actions in bulk solution where movement occurs in three di-
mensions.[4]Also the polarity at the membrane interface is
significantly different from bulk solution, and this microen-
vironment can have a dramatic effect on binding interac-
tions.[5]Natural cell membrane receptors often fully span
lipid bilayers, for example, tyrosine kinase receptors and G-
protein coupled receptors, which have seven transmembrane
units.[6]These transmembrane receptors are crucial to func-
tions such as initiating signal transduction pathways. In
signal transduction, the messenger binds to the external
domain of the transmembrane receptor on the cell surface
thereby causing significant receptor reorganization and a
change in the transmembrane conformation. This process
consequently leads to aggregation of the interior receptor
sites, thereby stimulating an (often cooperative) intracellular
response that in turn triggers a signal cascade inside the
cell.[7]For example, bacteria chemotaxic receptors, a family
of transmembrane receptors, can detect tiny changes in the
concentration of specific chemicals. Recent studies show
that in these receptors, binding of ligands at the external sur-
Abstract: A synthetic transmembrane
receptor that is capable of transmitting
binding information across a lipid bi-
layer membrane is reported. The bind-
ing event is based on aggregation of
the receptor triggered by copper(II)
complexation to ethylenediamine func-
tionalities. By labelling the receptor
with fluorescent dansyl groups, the cop-
per(II) binding event could be moni-
tored by measuring the extent of fluo-
rescence quenching. Comparing the re-
ceptor with a control receptor lacking
the transmembrane linkage revealed
that the transmembrane receptor binds
copper(II) ions more tightly than the
non-spanning control receptor at low
copper(II) concentrations. Since the in-
trinsic binding to copper(II) is the
same for both receptors, this effect was
attributed to synergy between the con-
nected interior and exterior binding
sides of the transmembrane receptor.
Thus, this is the first reported artificial
signalling event in which binding of a
messenger on one side of the mem-
brane leads to a cooperative binding
event on the opposite side of the mem-
brane, resembling biological signalling
systems and helping us to get a better
understanding of the requirements for
more effective artificial signalling sys-
tems.
Keywords: cooperative
fluorescence
transmembranes · vesicles
binding
·
··
receptors
[a] Dr. H. P. Dijkstra, J. J. Hutchinson, Prof. C. A. Hunter, Dr. H. Qin,
Dr. S. Tomas, Dr. S. J. Webb, Dr. N. H. Williams
Centre for Chemical Biology
Krebs Institute for Biomolecular Science
Department of Chemistry
University of Sheffield, Sheffield, S37HF (UK)
Fax: (+ +44)114-222-9346
E-mail: C.Hunter@sheffield.ac.uk
N.H.Williams@sheffield.ac.uk
[b] Dr. H. P. Dijkstra
Current address: Organic Chemistry & Catalysis
Faculty of Science, Utrecht University, Padualaan 8
3584 CH, Utrecht (The Netherlands)
Chem. Eur. J. 2007, 13, 7215–7222? 2007 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim
7215
FULL PAPER

Page 2

face of the membrane can cause receptor aggregation and
such receptor clusters in turn affect the lateral packing inter-
actions of the vertical signalling pathway at the opposing
surface of the membrane.[8]As a consequence of the cluster-
ing, the chemotaxic receptors elicit changes in their internal
environment which leads to whole-cell movement. It is clear
that trimers of dimers play a key role in controlling this
transmembrane signalling rather than isolated dimeric re-
ceptors.[9]
An ongoing research theme in our group is the develop-
ment of synthetic membrane-bound receptors and the study
of their aggregation behaviour upon addition of a messen-
ger. Although these systems represent simplified models of
biological membrane receptors, they can give us valuable in-
sight into the mechanisms by which biological signalling pro-
cesses may take place. For this purpose, a cholesterol based
receptor 1 was synthesised containing a fluorescent dansyl
ethylenediamine head group (Scheme 1a).[10]Cholesterol in-
serts well into lipid bilayers and serves as a membrane
anchor, and when copper(II) was used as a messenger its co-
ordination to the ethylenediamine group led to the forma-
tion of receptor aggregates. The fluorescent properties of
the dansyl head group were used in this study as a probe to
monitor the aggregation process in situ as copper coordina-
tion to the membrane-bound receptors leads to a strong de-
crease of the fluorescent signal.[11]By constraining the mem-
brane-bound receptors to a membrane surface where they
can only move in two dimensions, the binding interactions
between the membrane-bound receptors were expected to
be very different from those of the receptors in solution.
The surprising assembly of a 4:1 receptor/copper(II) com-
plex was deduced and the observed binding constants to
copper(II) were significantly larger at the membrane surface
relative to binding by similar amines in bulk solution. The
affinity for copper(II) ions was strongly dependent on the
membrane concentration of the receptors 1 and higher ag-
gregates were strongly favoured with increased concentra-
tion of the receptors in the membrane. These observations
are due to the higher effective concentration of the ligands
achieved by constraining them to the membrane volume
and to medium effects at the interface, rather than entropic
benefits from pre-organising the receptors at the bilayer sur-
face. Lehn and co-workers reported a similar concentration
dependent binding behaviour for a membrane bound dike-
tone ligand in binding studies with Eu3+ +, obtaining larger
observed equilibrium constants at higher receptor load-
ings.[12]As the size of receptor-ligand clusters can have a sig-
nificant effect on the binding processes, these data suggest
that by changing the number of
receptors at the membrane sur-
face, cells can finely control
their biological responses to ex-
ternal changes.[13]
For receptor 1, the extra- and
intravesiclebinding
are separated with no synergy
between the two sides of the
lipid bilayer, implying that the
internal and external binding
events are the same and lack
transmembrane communication.
Therefore, to study how infor-
mation canbe
across lipid bilayers by binding
of a messenger to the receptor
on the outside of a cell leading
to an intravesicular receptor
change, we decided to investi-
gate a synthetic receptor that
spans the membrane. Such a
system resembles biological sig-
nalling events and can give us
valuable information on their
mechanisms. Here, we report a
novel transmembrane spanning
receptor 2 (Scheme 1b) based
on the dansyl ethylenediamine
cholesterol moiety and its ag-
gregation behaviour in vesicles
upon addition of Cu2+ +
(messengers). More specifically,
activities
transmitted
ions
Scheme 1. Representation of the binding interactions in receptor: messenger complexes of a) receptor 1 and
b) transmembrane receptor 2 in the lipid bilayer of vesicles.
www.chemeurj.org
? 2007 Wiley-VCH Verlag GmbH&Co. KGaA, WeinheimChem. Eur. J. 2007, 13, 7215–7222
7216

Page 3

we studied transmembrane cooperativity by utilising the
fluorescence properties of the dansyl functionalities of re-
ceptor 2 at both sides of the membrane and by comparing
the results with the control receptor 1 which lacks the trans-
membrane orientation.
Results and Discussion
The transmembrane receptor 2 is based on a synthetic trans-
membrane signalling system that we recently developed to
mimic the tyrosine kinase receptor.[14]The membrane-span-
ning molecule 2 consists of two cholesterol units that are
linked together via a rigid dialkyne bridge. The incorpora-
tion of cholesterol assures insertion of this molecule into the
lipid bilayers whereas the linear, rigid structure and the
polar head groups favour a transmembrane orientation of 2.
Notably, these receptors can potentially also adopt a U-
shaped conformation, meaning that both polar head groups
are situated on the same side of the membrane and are thus
not spanning the membrane. From the current study it is
very difficult to distinguish between both forms and there-
fore a certain percentage of U-shaped receptors cannot be
fully excluded. However, it has been shown several times by
others that introducing rigidity, for example, a dialkyne unit,
into the core of specific lipids force them almost exclusively
in a transmembrane orientation,[15]even more flexible bo-
laamphiphilic molecules have been shown to adopt a com-
pletely transmembrane conformation.[16]Furthermore, exact-
ly the same transmembrane unit has been reported recently
in another transmembrane spanning system[17]and the trans-
membrane signalling system we reported recently is based
on the same dialkyne–cholesterol scaffold as 2 and results
obtained with that system clearly suggested a transmem-
brane orientation.[14]
In support of this assumption, we observed that vesicles
prepared with the longer molecules have notably different
physical properties compared with those that contained only
a cholesterol anchor—they were far harder to extrude
through the polycarbonate filters that we used, to the extent
that on one occasion the force required caused mechanical
failure of the extrusion apparatus (bursting the syringe
barrel). The shorter embedded molecules did not exhibit
this behaviour, which is apparent even at low loadings
(0.5% of 2). We quantified this observation by measuring
the pressure required to pass 800 nm vesicles through
200 nm filter once. That is, the molecules were incorporated
into lipids, then extruded to form homogeneous unilamellar
800 nm vesicles. These were then passed through a 200 nm
filter using a syringe pump and monitoring the force exerted
on the barrel using a pressure traducer. This revealed that
the spanning molecules required five-fold higher pressure
(extruding 20 mm lipid at a flow rate of 2 mL per minute)
than vesicles containing the same fraction of cholesterol de-
rivative but where they are not linked by the dialkyne unit.
If the embedded molecules adopt a U-shape, then it is diffi-
cult to envisage that the physical properties should be great-
ly different. Therefore, we conclude that the majority of re-
ceptors 2 adopt a transmembrane orientation when included
in vesicular bilayers.
Following our earlier results with half-spanning mem-
brane receptor 1, we again chose dansyl ethylenediamine as
the sensing moiety and copper(II) ions as messenger mole-
cules. The dansyl group is a well-developed environmentally
sensitive fluorophore and coordination of copper(II) ions to
the ethylenediamine ligands quenches the dansyl fluores-
cence. From our previous studies we know that copper(II)
chloride freely crosses the membrane in the presence of 1 or
2,[10]therefore we know that, after equilibrating the solution,
binding of copper(II) ions to the dansyl head groups occurs
on both sides of membrane. The aggregation behaviour of 2
upon binding to copper(II) ions can then be effectively stud-
ied by monitoring the quenching of the dansyl fluorescence.
Non-membrane spanning receptor 1 serves as a useful con-
trol system and comparing the results obtained with 1 and 2
gives us direct information about the presence of synergy
(and thus cooperativity) between the interior and exterior
copper(II)-coordination events in the transmembrane span-
ning receptor 2.
The synthesis of transmembrane receptor 2 is depicted in
Scheme 2. It was synthesised from cholenic acid, a bifunc-
tional analogue of cholesterol, in four steps. The coupling of
cholenic acid with propargyl alcohol mediated by dicyclo-
hexylcarbodiimide, gave cholenic acid propargyl ester 3 in
near quantitative yield. Heating of the ester 3 with bromo-
ACHTUNGTRENNUNGacetyl chloride under reflux in dry THF afforded the bromo-
ACHTUNGTRENNUNGacetylated propargyl ester of cholenic acid 4. Subsequent
displacement of the bromide from 4 with dansyl ethylenedi-
ACHTUNGTRENNUNGamine 8, available through the condensation of dansyl chlo-
ride with ethylenediamine in dry acetonitrile in the presence
of sodium carbonate as a base, resulted in the formation of
5. A tail-to-tail dimerisation of 5 was achieved using Glaser–
Hay coupling conditions[18]with freshly prepared copper(I)
chloride[19]in the presence of N,N,N’,N’-tetramethylethyle-
nediamine. This coupling reaction had to be conducted in
air, allowing free diffusion of molecular oxygen into the re-
action. The excess copper(II) ions were removed from the
crude product by thoroughly washing with EDTA and water
several times. Purification by silica gel chromatography
yielded receptor 2 as a pale yellow solid.
Compound 1 was synthesized in two steps from cholester-
ol using previously reported methods.[10]Membrane-bound
receptor 1 was used here as a control for membrane-span-
ning receptor 2 in the membrane aggregation studies.
Preparation of vesicle solutions: In this study we used large
unilamellar egg yolk phosphatidylcholine (PC) vesicles with
diameters of 800 nm. Unilamellar PC vesicles containing re-
ceptors 1 and 2 in MES buffer (pH 6.0) were prepared by
extrusion of a mixture of the lipids and the appropriate re-
ceptor through 800 nm polycarbonate membranes. The total
concentration of dansyl head groups was always kept the
same; since 2 has twice as many dansyl head groups as 1, in
each comparison experiment the concentration of [1]=[2]?
Chem. Eur. J. 2007, 13, 7215–7222? 2007 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim
www.chemeurj.org
7217
FULL PAPER
Lipid Bilayers

Page 4

2. This allows us to make a direct comparison between the
aggregation behaviour of 1 and 2 upon addition of the mes-
senger, copper(II) chloride. Furthermore, we kept the bulk
concentration of the lipids constant at 2.0 mm,[20]but varied
the percentage of receptors in the vesicles. The next step
was monitoring the fluorescence quenching of each recep-
tor–vesicle solution after addition of copper(II) chloride and
allowing the reaction mixture to equilibrate for 30 minutes.
The latter was necessary to allow copper(II) chloride to dif-
fuse through the lipid bilayer assuring a homogeneous intra-
and extravesicular distribution of copper(II) ions.[10]
Control experiments: We initially performed some control
experiments to evaluate background processes that might
lead to fluorescence quenching which is not related to cop-
per(II) binding. Firstly, we checked for self-quenching of the
fluorescent signal, for example due to spontaneous receptor
aggregation, by monitoring the fluorescence of both recep-
tors embedded in vesicles at various concentrations without
the addition of CuCl2. The fluorescent signal of vesicles con-
taining receptor 1 and 2 and with dansyl head group load-
ings ranging from 1.00–7.50 mol% was measured by exciting
the samples at 337 nm. The data was plotted as emission
fluorescence intensity at 520 nm versus mol% of dansyl
head groups in the vesicles
(data not shown). As expected,
the fluorescence intensity for
receptor 1 increased linearly
with increasing receptor loading
in the vesicles. Since dansyl
groups dissolved in aqueous sol-
vents display very low intensity
emission compared with the in-
terfacial dansyl groups of mem-
brane-embedded lipid 1 this lin-
earity shows that lipid 1 was
fully incorporated into the vesi-
clesoverthis
range. However for transmem-
brane receptor 2, we observed
incompleteincorporation
high receptor
(3.75 mol%
2,
7.50 mol%dansyl
groups). Asthis
allow a fair comparison be-
tween the aggregation behav-
iour of 1 and 2 upon addition of
CuCl2, we decided only to use
low receptor
2.50 mol% and 1=5.00 mol%)
for the comparison studies. Sec-
ondly, we also investigated the
possibility of vesicle aggrega-
tion in solution as this would
also potentially affect the out-
come of the receptor aggrega-
tion experiments. Lehn and co-
concentration
at
loading
equalto
head
notwould
loadings(2=
workers when investigating the coordination behaviour of a
membrane-anchored bispyridine ligand with Co2+ +and Ni2+ +
ions have recently reported receptor-dependent vesicle ag-
gregation, induced by metal-receptor coordination.[21]To
rule out this possibility with our system, the fluorescence of
receptor 2 (0.50, 1.25 and 2.50 mol%) embedded vesicles at
different vesicle concentrations (0.1–2.0 mm lipid concentra-
tion) in MES buffer (pH 6.0) was investigated. We found no
shift in the maximum emission wavelength of the diluted
vesicles and also the emission fluorescence intensity at
520 nm vs. vesicle concentration gave a linear fit, showing
that there is no vesicle aggregation in this concentration
range. Finally, as it is known that cholesterol and related ste-
roids affect many biological processes[22]such as membrane
rigidity, we decided to investigate whether added cholesterol
in the membrane will significantly influence the aggregation
behaviour of our receptors. Therefore, we conducted a
number of experiments in which various amounts of choles-
terol (CL) were incorporated into vesicles (800 nm diame-
ter, 2.0 mm lipid) containing receptor 1. In these experi-
ments, the total concentration of cholesterol groups in the
membrane was kept constant (thus [CL] + + [1]=20.0 mm),
and only the [CL]/[1] ratio was varied. Comparing the fluo-
Scheme 2. Synthesis of transmembrane receptor 2. i) DMAP, DCC, CH2Cl2, RT, 48 h (90%); ii) Bromoacetyl
chloride, THF, reflux, 1.5 h (69%); iii) CH2Cl2, 08 8C!RT, 1 h (92%); iv. Na2CO3, MeCN, reflux, 20 h (49%);
v) i) CuCl, TMEDA, O2, CH2Cl2, RT, 3 d, ii) EDTA wash (45%).
www.chemeurj.org
? 2007 Wiley-VCH Verlag GmbH&Co. KGaA, WeinheimChem. Eur. J. 2007, 13, 7215–7222
7218
C. A. Hunter, N. H. Williams et al.

Page 5

rescence intensity of these experiments at 520 nm with par-
allel control experiments in which there was no added cho-
lesterol revealed that in the concentration range of [1 + +
CL]=7.50 mol% there was no influence of added cholester-
ol on the optical properties and thus on the aggregation be-
haviour of the receptors in the vesicles. Thus, the influence
of varying the cholesterol concentration was neglected in
the aggregation comparison studies of receptors 1 and 2.
Binding comparison studies: To study the aggregation be-
haviour of 1 and 2 in the lipid bilayer upon addition of a
CuIImessenger at various receptor loadings, different vesicu-
lar solutions were prepared containing 0.50, 1.25, and
2.50 mol% of embedded 2 and 1.00, 2.50, and 5.00 mol% of
embedded 1, respectively. Titration of a 10 mm copper(II)
chloride solution into these solutions and equilibrating the
resulting solution for 30 minutes resulted in a dramatic de-
crease in the fluorescent intensity from 400–650 nm with
each addition, as shown in Figure 1 for a typical experiment
with 1.25 mol% receptor 2 loading. Comparing the results
obtained with 1 and 2 indicated that transmembrane-span-
ning receptor 2 showed a similar response to its half-span-
ning analogue 1 upon CuIIcoordination to the ethylenedia-
mine dansyl head groups. It should be noted that at higher
loadings of receptor 2 (5.00 and 7.50 mol% dansyl groups),
reproducibility of the titration experiments became unrelia-
ble. Therefore, to determine whether there is synergy be-
tween the interior and exterior domains of receptor 2, we
decided to only compare the aggregation results obtained
for 1 and 2 at lower receptor loadings, that is, 1.00 and
2.50 mol% dansyl groups.
The results of the copper(II) titration experiments with
vesicles containing 1 and 2 at dansyl head group loadings of
1.0 and 2.5 mol% are summarised in Figure 2. In all compar-
ison experiments receptor 2 embedded vesicles showed a
more efficient fluorescence quenching upon addition of CuII
ions than those containing 1 (Figure 2, grey dots for 2 and
black dots for 1). For example, at 1.00 mol% dansyl head
group loading (Figure 2a) only 14 mm CuII-ions are required
to quench 50% of the initial fluorescence for 2 whereas in
the case of control receptor 1 a much higher concentration
of CuII-ions (30 mm) is needed for the same amount of
quenching. In addition, fluorescence quenching is more effi-
cient at higher receptor loading, that is, 2.5 mol% dansyl
head group loadings (Figure 2b), for both receptors (as can
be deduced from the steepness of the curves by comparing
Figure 2a and b), confirming that higher copper(II) binding
affinities are obtained with increasing receptor loadings.
This behaviour was reported earlier for 1[10]and again em-
phasises that both receptors behave similarly with respect to
copper(II) coordination.
Cooperative binding: The fact that receptor 2 has a trans-
membrane orientation means that receptor 2 can potentially
couple binding events across the lipid bilayer. Previously, we
have already determined that receptor 1 upon addition of
CuIIions can form 1:2 and 1:4 coordination complexes (Cu12
and Cu14, respectively) in the membrane depending on the
concentration of embedded 1.[10]In addition, it was found
that the observed first (Cu1 complex) and second binding
Figure 1. Typical fluorescent spectra for a titration of receptor 2 in phos-
pholipid vesicles with copper(II) ions. Aliquots of 10 mm CuCl2solution
were added to a vesicle solution (2.0 mm lipid concentration) containing
1.25 mol% receptor loading and the fluorescence spectrum was measured
after each addition. The arrow indicates the direction of the change upon
addition of copper(II) ions.
Figure 2. Titration curves showing the decrease in the fluorescence of
vesicle-bound receptor 2 (grey) and 1 (black) with a) 1.00 mol% and b)
2.50 mol% dansyl head group loading upon the addition of CuIIions.
Control experiments with added cholesterol in receptor 1 containing vesi-
cles are also shown for comparison but the data are practically identical.
Aliquots of 10 mm CuCl2solution were added to the appropriate solu-
tions and the fluorescence spectrum was measured after equilibration.
Chem. Eur. J. 2007, 13, 7215–7222 ? 2007 Wiley-VCH Verlag GmbH&Co. KGaA, Weinheim
www.chemeurj.org
7219
FULL PAPER
Lipid Bilayers