Intra- and intermembrane pairwise molecular recognition between synthetic hydrogen-bonding phospholipids

Article abstract of DOI:10.1021/ja806954u

Multivalency and preorganization are fundamental aspects of molecular recognition at the lipid membrane-water interface and can render weak monomeric binding interactions selective and robust; this concept is important throughout biology, biotechnology, and materials science. Though hydrogen bonding is typically weakened in water, intramembrane hydrogen bonding between native lipids has been well-studied and is thought to contribute to lipid bioactivity and membrane function. We hypothesized that avidity and preorganization effects at the lipid-water interface could overcome solvent competition and allow for selective hydrogen-bond recognition between small, unstructured components. We have found that electrostatically identical vesicular membranes composed of cyanuric acid and melamine functionalized phospholipids 1 and 2 undergo selective apposition, fusion and adhesion in suspension and on solid support, indicating that their well-known low-dielectric hydrogen bonding properties translate effectively to the lipid-water interface. This work is notable and of general interest given the few detailed studies of aqueous phase hydrogen-bonding systems; we have extensively characterized this system, gaining structural, functional, and thermodynamic data. Furthermore, we have found that the designed lipid-lipid headgroup interactions result in dramatic alteration of the lipid phase morphology, providing insight into the coupling of molecular interactions with assembly state. As such, this work contributes to our understanding of fundamental phenomena such as molecular recognition at the lipid-water interface membrane chemistry and further illustrates the general possibility of designing selective hydrogen-bonding adhesive interactions from simple starting materials at other polar-apolar interfaces; this could have numerous materials and biotechnological applications. Copyright

Full text of DOI:10.1021/ja806954u

Published on Web 10/14/2008
Intra- and Intermembrane Pairwise Molecular Recognition between Synthetic
Hydrogen-Bonding Phospholipids
Mingming Ma,† Angel Paredes,‡ and Dennis Bong*,†
Department of Chemistry, The Ohio State UniVersity, 100 West 18th AVenue, Columbus, Ohio 43210, and
Department of Pathology and Laboratory Medicine, UniVersity of Texas Health Science Center,
Houston, Texas 77030
Received September 2, 2008; E-mail: bong@chem.osu.edu
Multivalency and preorganization are fundamental aspects of
molecular recognition at the lipid membrane-water interface and
can render weak monomeric binding interactions selective and
robust; this concept is important throughout biology, biotechnology,
and materials science.1 We previously described how small
molecule recognition between membranes is sufficient to selectively
merge vesicular membranes;2,3 recent reports have indicated similar
membrane chemistry directed by nucleic acid recognition.4 These
systems all utilize hydrogen bond (H-bond) donor-acceptor pattern
recognition between complex natural products or large biomolecules
to guide membrane chemistry, which prompted our examination
of yet simpler hydrogen bonding systems to study multivalent
surface recognition in water. Intramembrane hydrogen bonding
between native lipids has been well-documented and is thought to
contribute to lipid bioactivity and membrane function.5 We
Figure 1. (A) Synthetic lipids 1 and 2, headgroup functionalized with
cyanuric acid and melamine (boxed), respectively. (B) Possible hydrogen-
bonding patterns between lipids in intra- and intermembrane contexts. Dotted
lines indicate possible hydrogen bonds.
hypothesized that avidity and preorganization effects at the
lipid-water interface6 could overcome solvent competition and
allow for selective hydrogen-bond recognition between small,
unconstrained components. Lipid-mediated hydrogen-bonding in-
teractions have been previously reported in bilayer,7 micellar,8
homovesicular,9 and weakly binding heterovesicular contexts;10
other heterovesicular recognition systems have been reported that
are driven by metal-complexation11,12 and electrostatic13 interac-
tions. We have found that electrostatically identical vesicular
membranes composed of cyanuric acid and melamine functionalized
phospholipids 1 and 2 undergo selective heterovesicular apposition,
fusion, and adhesion in suspension and on solid support.
Mixtures of cyanuric acid (CA) and melamine (M) and deriva-
tives are known to form hydrogen-bonded rosette, tape, and
extended structures with 1:1 stoichiometry in organic solvents and
solid state.14 While these interactions have not been experimentally
demonstrated in aqueous solvent, it has been calculated that CA
and M derivatives should interact strongly.15 We found that CA-
and M-derivatized phospholipids 1 and 2 (Figure 1) could be
dispersed in aqueous buffer to yield vesicles composed purely of 1
or 2. This permitted careful examination of lipid assembly properties
and dependence on membrane composition. Monosubstituted CA
and M derivatives like 1 and 2 are known to form rosette or
extended tape structures with six H-bonds per unit; this could also
happen at the lipid-water interface. Interaction between lipids 1
and 2 was first probed in suspensions containing pyrene-function-
alized lipid (Pyr-PC). The excimer-monomer (E/M) fluorescence
intensity ratio of Pyr-PC is indicative of lipid mobility;16 vesicles
composed of only 1, 2, or fluid-phase unsaturated lipid 1-palmitoyl-
2-oleoyl-phosphatidyl-choline (POPC) all had similar E/M ratios
with a linear temperature dependence, as expected for fluid,
monophasic membranes. However, the 1:1 mixture of 1 and 2
Figure 2. Intramembrane lipid recognition. (Left) Temperature dependence
of pyrene PC excimer/monomer emission intensity in suspensions of 1 (red),
2 (blue), POPC (green), and a 1:1 mixture of 1 and 2 (black). (Right) DSC
upscan (- -) and downscan (—) traces of a suspension of a 1:1 mixture of
1 and 2; 1 and 2 upscans (- -, · · ·) and downscans (-, —) are all flat traces.
behaved like a more rigid gel phase with a shallow phase transition
at 22 °C and a sharper cooperative melting at 65 °C (Figure 2).
Differential scanning calorimetry (DSC) revealed exothermic and
endothermic peaks on the upscan at 20 and 75 °C, respectively;
these transitions shifted17 on the downscan to 65 °C (exothermic)
and 22 °C (endothermic). Together, the DSC and the pyrene data
clearly indicate a temperature dependent phase transition that
governs intramembrane lipid mobility, consistent with hydrogen
bonding.
Surface plasmon resonance (SPR) was used to study intermem-
brane binding and reaction by flowing large unilamellar vesicles
(LUVs) of 2 over surface bound 1 LUVs. The on-rate for adsorption
was ∼40 M-1 s-1, though the off-rate at various solution
concentrations was effectively zero, preventing calculation of an
affinity constant. Desorption is often undetectable in SPR studies
because of avidity and rebinding effects,18 but this could also be
† The Ohio State University.
‡ University of Texas Health Science Center.
9
14456 J. AM. CHEM. SOC. 2008, 130, 14456–14458
10.1021/ja806954u CCC: $40.75
2008 American Chemical Society

C O M M U N I C A T I O N S
Table 1. Thermodynamics of Intermembrane Recognitiona
temp (°C)
∆H (kcal/mol)
Ka (M-1
)
∆S (cal/K·mole)
∆G (kcal/mol)
10
25
40
50
-12.4 ( 0.06 1.37 × 106 -15.5 ( 0.25 -7.96 ( 0.04
-7.04 ( 0.01 1.48 × 106 +4.60 ( 0.03 -8.42 ( 0.002
-6.92 ( 0.07 6.73 × 104
0.0 ( 0.21 -6.92 ( 0.01
a Data fitting for Ka with a 1:1 binding model produced errors of
6.4%, 0.29%, and 1.47% for 10, 25, and 40 °C, respectively. No heat
was detected at 50 °C.
Figure 3. Intermembrane lipid recognition. (Left) Binding of surface
anchored 1 with 2 at varying concentration as labeled. (Right) Refractive
unit change upon deposition of 1 (s) and 2 (---) onto the bare SPR L1
chip followed by surface saturation with POPC and alternating treatment
(numbered) with 2, 1, 2, 1, 2 (s) or 1, 2, 1, 2, 1 (---).
mixed membrane would dilute surface ligand presentation and
diminish reactivity.2,12
Interestingly, the DLS and FRET data showed seemingly
opposing temperature trends. The aggregation rate observed by DLS
was much faster at 10 °C than at 25 °C, a result we attribute to
enhanced hydrogen-bonding interactions at lower temperatures.
However, the rate of lipid mixing was much faster at 25 than at 10
°C. This would suggest that although membrane apposition (driven
by headgroup interaction) is more efficient at lower temperature,
there is a thermal threshold for lipid mixing (lipid phase change)
that exists between 10 and 25 °C.
Isothermal titration calorimetry (ITC) of the 1 + 2 vesicle reaction
revealed a strongly exothermic 1:1 lipid interaction. Titration of freely
soluble CA and M derivatives did not result in detectable heat flow
(Supporting Information), indicating the strong effect of multivalent
membrane presentation. Notably, the heat of interaction between 1
and 2 is context-dependent; lipid compositions that aggregate but do
not mix have a significantly weaker negative enthalpy. The reaction
becomes less exothermic with increasing temperature, with ∆H ) -12
kcal/mol at 10 °C, ∆H ) -7 kcal/mol at 25 °C, and nondetectable
heat at 50 °C (Figure 4), consistent with DLS data that indicates loss
of aggregation at 50 °C (Supporting Information). If six hydrogen
bonds are formed on each CA and M, the upper limit set for enthalpic
contribution from each hydrogen bond at the lipid-water interface is
1-2 kcal/mol, depending on temperature. The loss of vesicle recogni-
tion with higher temperature may be ascribed to a “melting” of the
intermembrane lipid interactions, which occurs at a lower temperature
than in the intramembrane context. Though the vesicle-vesicle
interaction is a complex multivalent surface-surface adhesion coupled
with lipid mixing and phase transition, the ITC binding curves fit
remarkably well to a 1:1 binding model as a first approximation. Simple
1:1 fitting of the isotherms revealed that the ∆Grxn at 10 °C is roughly
the same as at 25 °C, -8 kcal/mol (Table 1). By this crude analysis,
the reaction of 1 and 2 LUVs exhibits significant enthalpy-entropy
compensation between these two temperatures, with enthalpy-driven
reaction at 10 °C and both enthalpy and entropy change favorable at
25 °C. We speculated that the temperature-dependent change in ∆H
and ∆S contributions signified a lipid phase-transition during membrane
fusion from the lamellar phase to a more disordered morphology.
Reaction stoichiometry was suggestive of complete fusion of the
membranes at both temperatures, pointing toward the formation of a
less-ordered nonlamellar phase at higher temperatures. Furthermore,
∆S at 40 °C is close to zero while the ∆H was similar to that found
at 25 °C, indicating that the decreased driving force can be attributed
completely to the less-favorable entropy change. This may be explained
if the suspensions of 1 and 2 undergo a phase transition at a higher
temperature (close to 40 °C), whereas the lipid mixture has a transition
around 25 °C. Consistent with the formation of less reactive lipid
phases of 1 and 2 at 40 °C, the calculated interaction stoichiometry at
40 °C was found to be 0.66 instead of 1:1.
Figure 4. Intermembrane lipid recognition. (Left) Lipid mixing rates when
LUVs containing 1.5% NBD-PE and 1.5% Rh-DHPE are reacted with
unlabeled LUVs. The remainder of the reacting membrane compositions
are 1 + 2 (O), 30% 1 in DPPC + 30% 2 in DPPC (0) and 50% 1 in POPC
+ 50% 2 in POPC (]). (Right) ITC at 10 (O), 25 (0), 40 (4), and 50 °C
(+) of injection of 1 LUVs into 2 LUVs. Markers indicate data points and
solid lines are fits to 1:1 binding model.
due to irreversible fusion of suspension LUVs to the surface bound
membrane (Figure 3). Indeed, treatment with 1 primed the surface
for binding 2, and vice versa, though alternating cycles of treatment
resulted in decaying signal intensity (Figure 3). This ruled out
“layer-by-layer” deposition and supports the fusion of tethered
vesicles with suspended vesicles to produce a lipid mixed product
that has diminished but residual affinity to both 1 and 2. Importantly,
the vesicles of 1 and 2 did not self-aggregate and deposit
continuously on the SPR chip, nor did they react with neutral POPC
membranes or POPG/POPC mixed membranes, but did react
selectively and strongly with each other.
Selective intermembrane bonding between lipids 1 and 2 was
also evident by dynamic light scattering (DLS); mixing 1 LUVs
with LUVs of 2 resulted in a rapid and irreversible increase in
particle size. Both phospholipid LUV suspensions had the same
ꢀ-potential of -11 mV in PBS and thus bind each other despite
electrostatic repulsion. Vesicle interaction was abrogated in low
salt buffer, as a result of increased repulsion between the -33 mV
ꢀ-potential surfaces under these conditions. Membrane apposition
of 1 and 2 is accompanied by rapid lipid mixing (Figure 4) as judged
by fluorophore dilution; similar to our previous findings, these
events depend markedly on membrane composition. Vesicles
prepared with POPC and 1 or 2 neither aggregated nor fused with
each other at any mole ratio; however, replacement of POPC with
dipalmitoyl phosphatidylcholine (DPPC) resulted in LUVs that
aggregated (20% 1 and 2) and LUVs that both aggregated and fused
(30% 1 and 2). These results likely derive from phase-separation
of saturated DPPC from the unsaturated lipids 1 and 2 due to
packing considerations; a single phase is expected with POPC.
Phase-separation of DPPC from 1 and 2 would produce subdomains
with a surface density of reactive lipids similar to a vesicle
composed entirely of the interacting lipids, while a single domain
Structural insight into the process of vesicle interaction and
hypothesized changes in lipid phase morphology was sought through
cryo-electron microscopy (cryo-EM), which indicated that LUVs
9
J. AM. CHEM. SOC. VOL. 130, NO. 44, 2008 14457

C O M M U N I C A T I O N S
tion. The system presented replaces the covalent sugar-phosphate
scaffold of DNA with a noncovalently assembled phospholipid surface
that also permits partial hydrophobic burial of the hydrogen-bonding
groups upon heteromeric membrane apposition, similar to base-stacking
and pairing. This finding underscores the presence of discriminating
molecular recognition between two membrane surfaces which serves
to optimize H-bond donor-acceptor interactions in both inter and
intramembrane contexts and to minimize repulsive lone-pair interac-
tions found in homomeric lipid hydrogen-bonding.
Finally, this study illustrates the general possibility of designing
selective hydrogen-bonding adhesive interactions from simple
starting materials at other polar-apolar interfaces;23 this could have
a numerous materials and biotechnological applications.1,24
Acknowledgment. This work was supported in part by the
University of Texas (A.P.), The Ohio State University, and an NSF-
CAREER award to D.B. We thank the Structural Biology Imaging
Center of University of Texas Health Science Center at Houston
and the Ohio Bioproducts Innovation Center for support of the
instrumentation used in this study.
Figure 5. (Top left) Cryo-EM of 1 LUVs, (top right) 2 LUVs, (lower left)
1 + 2 premixed 1:1, (lower right) product of 1 LUVs reacted with 2 LUVs;
round objects are gold beads. Scale bars (100 nm) in left panels apply to
each row. All samples were prepared at room temperature.
Supporting Information Available: Synthetic and experimental
procedures, lipid NMR and MS characterization, additional DLS, DSC,
ITC, and cryo-EM data. This material is available free of charge via
the Internet at http://pubs.acs.org.
of 1 and 2 appear morphologically similar and indeed are mostly
unilamellar, dissociated vesicles that can be found as aggregates
as well (Figure 5), consistent with DLS measurements. However,
the structures derived from a 1:1 mixture of the two lipids, as well
as the fusion product, are strikingly distinct from the pure vesicles
and mechanistically suggestive. While the pure lipids produce
vesicular structures that are mostly unilamellar, both the “premixed”
lipids and the products of vesicle reaction appear to be in the
hexagonal phase. Clear signatures of the hexagonal phase (presum-
ably HII) are observed as regular striations and lipidic structures
consistent with interlamellar attachment sites (Figure 5).19,20 Some
products of vesicle fusion are a combination of dense multilamellar
vesicles that contain HII phases within, while others appeared to
be vesicles in the process of fusing (Supporting Information).
Formation of a hexagonal phase is expected with strong lipid-lipid
hydrogen bonding at the lipid-water interface as H-bonding between
headgroups results in desolvation of the surface, which in turn
permits high surface curvature and close membrane apposition
through thinning of the repulsive hydration layer.5 Dehydration of
the membrane and formation of the HII phase should be entropically
favorable,19 consistent with the ITC results. The lamellar (LR) to
HII transition of the fused or mixed product of 1 and 2 is likely the
low-temperature broad transition21 seen at 22 °C in the DSC and
pyrene melt experiments. This could also explain the paradoxical
observation of seemingly decreased aggregation and faster lipid
mixing at 25 versus 10 °C. Fusion at 10 °C results in larger vesicles
while fusion at 25 °C produces dense HII phase particles that scatter
light similarly to the starting vesicles (Figure 5). Thus, apposition
has not decreased at higher temperature, but simply become less
detectable by DLS. This is consistent with the reaction thermody-
namics, which indicate equally facile interaction at both tempera-
tures. Moreover, though not detected by DSC, it is possible that
vesicles of 1 and 2 transition from the LR to HII phase around 40
°C, leading to the entropy-neutral reaction found by ITC. The
structural and functional similarities between the 1 + 2 premixed
suspensions and reacted 1 + 2 LUVs clearly demonstrate the notion
that designed H-bond donor-acceptor interactions at the lipid-water
interface can drive both intramembrane and intermembrane chemistry.
Though there are relatively few reports of designed, non-native
aqueous phase hydrogen-bond recognition,6,22 this is commonly
observed in native systems, most prominently in nucleic acid recogni-
References
(1) Mammen, M.; Choi, S.-K.; Whitesides, G. M. Angew. Chem., Int. Ed. 1998,
37, 2754–2794.
(2) Gong, Y.; Ma, M.; Luo, Y.; Bong, D. J. Am. Chem. Soc. 2008, 130, 6196–
6205.
(3) Gong, Y.; Luo, Y.; Bong, D. J. Am. Chem. Soc. 2006, 128, 14430–14431.
(4) Chan, Y.-H. M.; van Lengerich, B.; Boxer, S. G. Biointerphases 2008, 3,
17–21. Stengel, G.; Zahn, R.; Hook, F. J. Am. Chem. Soc. 2007, 129, 9584–
9585.
(5) Boggs, J. M. Biochim. Biophys. Acta 1987, 906, 353–404.
(6) Ariga, K.; Kunitake, T. Acc. Chem. Res. 1998, 31, 371–378.
(7) Kawasaki, T.; Tokuhiro, M.; Kimizuka, N.; Kunitake, T. J. Am. Chem.
Soc. 2001, 123, 6792–6800. Kimizuka, N.; Kawasaki, T.; Hirata, K.;
Kunitake, T. J. Am. Chem. Soc. 1998, 120, 4094–4104.
(8) Nowick, J. S.; Cao, T.; Noronha, G. J. Am. Chem. Soc. 1994, 116, 3285–
3289. Nowick, J. S.; Chen, J. S. J. Am. Chem. Soc. 1992, 114, 1107–1108.
(9) Menger, F. M.; Zhang, H. J. Am. Chem. Soc. 2006, 128, 1414–1415.
(10) Paleos, C. M.; Tsiourvas, D. J. Mol. Recognition 2006, 19, 60–67. Berti,
D.; Baglioni, P.; Bonaccio, S.; Barsacchi-Bo, G.; Luisi, P. L. J. Phys. Chem.
B 1998, 102, 303–308.
(11) Lim, C. W.; Crespo-Biel, O.; Stuart, M. C. A.; Reinhoudt, D. N.; Huskens,
H.; Ravoo, B. J. Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 6986–6991.
Waggoner, T. A.; Last, J. A.; Kotula, P. G.; Sasaki, D. Y. J. Am. Chem.
Soc. 2001, 123, 496–497.
(12) Mart, R. J.; Liem, K. P.; Wang, X.; Webb, S. J. J. Am. Chem. Soc. 2006,
128, 14462–14463.
(13) Marchi-Artzner, V.; Gulik-Krzywicki, T.; Guedeau-Boudeville, M.-A.;
Gosse, C.; Sanderson, J. M.; Dedieu, J.-C.; Lehn, J.-M. ChemPhysChem
2001, 2, 367–376.
(14) Fyfe, M. C. T.; Stoddart, J. F. Acc. Chem. Res. 1997, 30, 393–401.
(15) ten Cate, M. G. J.; Huskens, J.; Crego-Calama, M.; Reinhoudt, D. N.
Chem.sEur. J. 2004, 10, 3632–3639.
(16) Galla, H. J.; Hartmann, W. Chem. Phys. Lipids 1980, 27, 199–219.
(17) Sorrells, J. L.; Menger, F. M. J. Am. Chem. Soc. 2008, 11.
(18) Nieba, L.; Krebber, A.; Plockthun, A, Anal. Biochem. 1996, 234, 155–165.
(19) Epand, R. M. Methods Mol. Biol. 2007, 400, 15–26.
(20) Frederik, P. M.; Burger, K. N. J.; Stuart, M. C. A.; Verkleij, A. J. Biochim.
Biophys. Acta 1991, 1062, 133–141.
(21) Wenk, M. R.; Seelig, J. Biochim. Biophys. Acta 1998, 1372, 227–36.
(22) Oshovsky, G. V.; Reinhoudt, D. N.; Verboom, W. Angew. Chem., Int. Ed.
2007, 46, 2366–2393. Brunsveld, L.; Vekemans, J. A.; Hirschberg, J. H.;
Sijbesma, R. P.; Meijer, E. W. Proc. Natl. Acad. Sci. U.S.A. 2002, 99,
4977–4982. Brunsveld, L.; Folmer, B. J. B.; Meijer, E. W.; Sijbesma, R. P.
Chem. ReV. 2001, 101, 4071–4098. Hirschberg, J. H.; Brunsveld, L.; Ramzi,
A.; Vekemans, J. A.; Sijbesma, R. P.; Meijer, E. W. Nature 2000, 407,
167–170. Kato, Y.; Conn, M. M.; Rebek, J., Jr Proc. Natl. Acad. Sci. U.S.A.
1995, 92, 1208–1212. Ghadiri, M. R.; Granja, J. R.; Milligan, R. A.; McRee,
D. E.; Khazanovich, N. Nature 1993, 366, 324–327.
(23) Joh, N. H.; Min, A.; Faham, S.; Whitelegge, J. P.; Yang, D.; Woods, V. L.;
Bowie, J. U. Nature 2008, 453, 1266–1270. Grigoryan, G.; DeGrado, W. F.
Nat. Chem. Biol. 2008, 4, 393–394.
(24) Madueno, R.; Raisanen, M. T.; Silien, C.; Buck, M. Nature 2008, 454,
618–621. Bong, D. T.; Clark, T. D.; Granja, J. R.; Ghadiri, M. R. Angew.
Chem., Int. Ed. 2001, 40, 988–1011.
JA806954U
9
14458 J. AM. CHEM. SOC. VOL. 130, NO. 44, 2008