ATP-lipids - Protein anchor and energy source in two dimensions

Article abstract of DOI:10.1021/ja953937m

The ubiquitous function of ATP as energy equivalent in nature has resulted in a common folding pattern of ATP-binding proteins. Their binding pocket tolerates modifications of the adenine ring to some extend, whereas those of the triphosphate group strong

Full text of DOI:10.1021/ja953937m

5532
J. Am. Chem. Soc. 1996, 118, 5532-5543
ATP-LipidssProtein Anchor and Energy Source in Two
Dimensions
Lutz Schmitt† and Robert Tampe´*,†,‡
Contribution from the Lehrstuhl fu¨r Biophysik E22, Physik Department,
Technische UniVersita¨t Mu¨nchen, D-85748 Garching, Germany,
and Max-Planck-Institut fu¨r Biochemie, Am Klopferspitz 18a, D-82158 Martinsried, Germany
ReceiVed NoVember 27, 1995X
Abstract: The ubiquitous function of ATP as energy equivalent in nature has resulted in a common folding pattern
of ATP-binding proteins. Their binding pocket tolerates modifications of the adenine ring to some extend, whereas
those of the triphosphate group strongly affect the binding affinity. In consequence, immobilized C8- and N6-
modified ATP analogues are frequently used for affinity purification of ATPases or kinases. To combine this unique
recognition principle with the fascinating properties of self-assembly, we have synthesized a novel class of hydrolyzable
and nonhydrolyzable ATP-lipids where the nucleotides are covalently attached via C8- or N6-position of the adenine
ring to a synthetic lipid. These ATP-lipids were characterized by various enzyme assays in micellar solution, resulting
in ATPase and competition activities that are comparable to their free counterparts. The specific docking of actin
as a model of an ATP-binding protein to ATP-lipid monolayers was followed by film balance technique and
epifluorescence microscopy. Based on this specific interaction, actin-supported membranes were generated to study
shape transitions of vesicular systems. Due to the coupling of actin to ATP-lipid bilayers drastic changes in the
viscoelastic properties and shape transitions were observed by phase contrast microscopy. These results underline
the properties of these novel ATP-lipids as protein anchor or energy source in two dimensions. They can be applied
either to form phantom cells, actin-supported membranes or to orient and crystallize ATP-binding proteins at lipid
interfaces.
Introduction
proteins such as antibody fragments,13,14 and chelators which
are capable to bind histidine-tagged proteins.15,16 Some of these
functionalized lipids have been used successfully for two-
dimensional crystallization of proteins at lipid monolayers, such
as dinitrophenol for an IgG antibody,17 γ-phosphate linked
desoxy-ATP for ribonucleotide reductase,9 biotin for streptavi-
din,18 and novobiocin for DNA gyrase B subunit.4
Biotin-lipids are frequently used for the immobilization and
the self-organization of proteins at interfaces. To expand the
system to other proteins of interest, a streptavidin interface19
and a chemical biotinylation of the biomolecule20 are required
which often lacks side-specificity and stoichiometry. As a result,
the orientation of the bound molecule cannot be controlled. A
more recent approach combines genetic engineering with self-
assembly using a novel class of chelator lipids.15,16 This concept
opens up the possibility for the specific and reversible im-
mobilization and orientation of the vast variety of histidine-
tagged fusion proteins at chelator lipid interfaces.21-25
Biofunctionalization of interfaces can be achieved either in
an unspecific, the so-called physisorption, or in a specific way,
the so-called chemisorption.1,2 The unspecific immobilization,
which is based mainly on electrostatic and hydrophobic interac-
tion often affects the activity and function of the adsorbed
proteins. In contrast, the denaturation can be bypassed by a
specific immobilization and a defined orientation of the bio-
molecule at the interface. One prominent type of interfaces are
lipids, because of their fundamental properties such as self-
assembly, lateral organization, and variability. Moreover, nearly
every surface can be coated and functionalized by lipid films
using various transfer techniques. Examples for functional
moieties of modified lipids are dinitrophenol,3 novobiocin,4
progesterone,5 hydroxamates,6 cofactors like biotin,7 NAD+,8
or γ-phosphate linked desoxy-ATP,9 peptides,10 sugars,11,12
* Corresponding author: Robert Tampe´, Max-Planck-Institut fu¨r
Biochemie, Am Klopferspitz 18a, D-82158 Martinsried, Germany.
Phone: 0049-89-8578-2646. Fax: 0049-89-8578-2641. E-mail: tampe@
vms.biochem.mpg.de.
(9) Ribi, H. O.; Reichard, P.; Kornberg, R. D. Biochemistry 1987, 26,
7974-7979.
(10) Metzger, J. W.; Sawyer, W. H.; Wille, B.; Biessert, L.; Bessler, W.
G.; Jung, G. BBA 1993, 1149, 29-39.
† Technische Universita¨t Mu¨nchen.
‡ Max-Planck-Institut fu¨r Biochemie.
Keywords: actin, ATPases, ATP-binding, biofunctionalization, kinases,
membranes, molecular recognition, self-assembly, self-organization.
X Abstract published in AdVance ACS Abstracts, May 15, 1996.
(1) Ulman, A. An introduction to ultrathin organic films; Academic
Press: San Diego, CA, 1991.
(11) Haensler, J.; Schuber, F. BBA 1988, 946, 95-105.
(12) Spevak, W.; Nagy, J. O.; Charych, D. H.; Schaefer, M. E.; Gilbert,
J. H.; Bednarski, M. D. J. Am. Chem. Soc. 1993, 115, 1146-1147.
(13) Martin, F. J.; Wayne, L. H.; Papahadjopulus, F. Biochemistry 1981,
20, 4229-4238.
(2) Tampe´, R.; Dietrich, C.; Gritsch, S.; Elender, G.; Schmitt, L.
Nanofabrication and Biosystems: Integrating material science, engineering,
and biosystems; Cambridge University Press: Cambridge, 1996; in press.
(3) Weis, R. M.; Balakrishnan, K.; Smith, B. A.; McConnell, H. M. J.
Biol. Chem. 1982, 257, 6440-6445.
(14) Egger, M.; Heyn, S.-P.; Gaub, H. E. BBA 1992, 1104, 45-54.
(15) Schmitt, L.; Dietrich, C.; Tampe´, R. J. Am. Chem. Soc. 1994, 116,
8485-8491.
(16) Shnek, D. R.; Pack, D. W.; Sasaki, D. Y.; Arnold, F. H. Langmuir
1994, 10, 2382-2388.
(4) Lebeau, L.; Regnier, E.; Schultz, P.; Wang, J. C.; Mioskowski, C.;
Oudet, P. FEBS Lett. 1990, 267, 38-42.
(17) Uzgiris, E. E.; Kornberg, R. D. Nature 1983, 301, 125-129.
(18) Blankenburg, R.; Meller, P.; Ringsdorf, H.; Salesse, C. Biochemistry
1989, 28, 8214-8221.
(5) Lebeau, L.; Oudet, P.; Mioskowski, C. HelV. Chim. Acta 1991, 74,
1697-1706.
(19) Mueller, W.; Ringsdorf, H.; Rump, E.; Wildburg, G.; Zhang, X.;
Angermaier, L.; Knoll, W.; Liley, M.; Spinke, J. Science 1993, 262, 1707-
1708.
(6) Altenburger, J.-M.; Lebeau, L.; Mioskowski, C. HelV. Chim. Acta
1992, 75, 2538-2544.
(7) Bayer, E. A.; Rivnay, B.; Skutelesky, E. BBA 1979, 550, 464-473.
(8) Salord, J.; Tarnus, C.; Muller, C. D.; Schuber, F. BBA 1986, 886,
64-75.
(20) Savage, D.; Mattson, G.; Desai, S.; Nielander, G.; Morgensen, S.;
Conklin, E. AVidin-Biotin Chemistry: A Handbook, 2nd ed.; Pierce Chemical
Company: Rockford, IL, 1992.
S0002-7863(95)03937-0 CCC: $12.00 © 1996 American Chemical Society

ATP-Lipids
J. Am. Chem. Soc., Vol. 118, No. 24, 1996 5533
A more general concept would need a synthetic lipid
containing an affinity ligand, somehow ubiquitous in binding
to various proteins without additional modification or engineer-
ing steps. Because of its common use as energy source in
nature, ATP seems to be the most prominent candidate. As
known from affinity chromatography,26 ATP covalently linked
to the resin through its C8-27,28 or N6-position29,30 of the adenine
ring is favored for the purification of ATP-binding proteins,
while ribose- or phosphate-linked ATP-resins31 are less fre-
quently used. The favorite binding of ATPase and protein
kinases to C8- or N6-modified ATP-resin can be explained by
the related structure of their nucleotide binding pocket.32 While
the overall structure can vary from protein to protein, a loop is
apparent in all of them, representing a sequence finger print.
The sequences of this loop (Gly-X-X-Gly-X-Gly-Lys (Gly-loop)
or Y1-X-Y2-Asp-X-Gly-Y3-Y3x-X-Y4 (Yn: groups of three to
four particular residue types)) can be used to identify ATP-
binding proteins based on sequence comparison.33 The loop
forms an anion hole, into which the triphosphate group
complexed by divalent cations fits perfectly. Therefore, any
modification of the phosphate groups often decreases the
affinity, whereas modifications of the adenine ring can be
tolerated to some extend. Prominent examples for proteins
sharing this sequence motif are actin, F1-ATPase, H-ras-p21,
adenylate kinases, and heat shock proteins.
Here, we describe the synthesis and characterization of ATP-
lipids linked via the C8- or N6-position of the adenine ring to
the lipid. The interchangeable concept allows the synthesis not
only of either nonhydrolyzable or hydrolyzable ATP-lipids but
also of ADP- and AMP-lipids. While the hydrolyzable ATP-
lipids represent a self-organizing energy source in two dimen-
sions, the nonhydrolyzable derivative can act as ligand for
various ATP-binding proteins. To determine their biological
activity, the ATP-lipids were first characterized by enzyme
assays in micellar detergent solution. Using film balance
techniques with epifluorescence microscopy21,25 and phase
contrast microscopy, we studied the properties of these ATP-
lipids in lipid mono- and bilayers. Actin was chosen as model
to investigate the specific docking of an ATP-binding protein
to the ATP-lipid interface. Actin shares the basic principles of
ATP-binding and can be purified by C8-linked ATP affinity
chromatography. The strong interplay between coupling and
decoupling of the actin network and the membrane is essential
to maintain a high degree of softness or to induce shape
transitions in cell membranes. However, these effects are still
an enigma.34 Also, the mechanism of domain formation and
lateral organization is not understood. Therefore, the develop-
ment of membranes supported by an actin network will be one
approach to answer these questions (“phantom cells”). Due to
the specific enrichment and orientation, this concept might also
be applied for two-dimensional crystallization of ATP-binding
proteins at ATP-lipid monolayers.
Materials and Methods
Materials. The following chemicals were used: DODA, lead(IV)
acetate, ethylenediamine, and hexamethylenediamine (Fluka, Neu-Ulm,
Germany), TR-DPPE (Molecular Probes, Oregon, USA), DMPC
(Avanti Polar Lipids, Alabama, USA), Br-Ad, Cl-Ad, o-hydroxyphe-
nylenephosphochloridate, methylenediphosphoric acid, diphosphoric
acid tri-n-butylammonium salt, fluorescamine spray reagent, molyb-
denum blue spray reagent, and ornicol spray reagent (Aldrich, Stein-
heim, Germany). The enzyme assay kit for the C8-ATP analogues,
proteins and chemicals for the N6-ATP analogues assay, AMPPCP, and
Triton X100 were ordered from Sigma (Deisenhofen, Germany). NBD-
actin was prepared from rabbit muscle actin according to refs 35 and
36. TLC plates (Kieselgel 60 F254 and DIOL F254), Kieselgel 60 (40-
63 µm, 230-400 mesh), and LiChroPrep DIOL (40-63 µm, 60-230
mesh) were purchased from Merck (Darmstadt, Germany). All other
solvents were from Fluka (Neu-Ulm, Germany) and were reagent p.a.
grade or higher. Solvents were used without further purification unless
otherwise stated. Prior to use triethylamine was refluxed over calcium
hydride for 2 h and then destilled.
Lipid Characterization. Reactions were monitored by TLC (pre-
coated plates 0.25 mm, silica gel F254) and visualized by UV (aromatic
groups), fluorescamine (primary amino groups), ornicol (cis-diol system
in the ribose moiety), molybdenum blue (phosphate), and iodine
(organic compound). Solvent ratios are given in volume/volume. For
the solvent system chloroform/methanol/water/ammonia 65:25:2:2 an
aqueous solution of 25% ammonia was used. The concentration of
the Triton X100 solution is given in weight/volume. If necessary,
products were purified by silica gel chromatography (LiChroPrep DIOL
or silica gel 60). 1H-NMR (500 MHz or 400 MHz) were recorded on
a Bruker AM 500 or Bruker AM 400, respectively. 13C-NMR (100
MHZ) and 31P-NMR (161 MHz) were recorded on a Bruker AM 400.
Chemical shifts (δ) are given in ppm relative to solvent as internal
standard for 1H- and 13C-NMR and relative to 85% phosphoric acid as
external standard for 31P-NMR. MS (Finnigan, MAT, Forster City,
CA, USA) was performed in the negative and positive ion mode.
Phosphorus was determined according to ref 37. The given values are
the average of two independent measurements. UV spectra were
recorded on a Perkin-Elmer lambda 5 (Perkin-Elmer, U¨ berlingen,
Germany) after background correction and analyzed using the software
provided by Perkin-Elmer.
Enzyme Assays. The ATPase activity of the C8-modified nucle-
otides was determined by the GADH/PGK assay (Sigma). 3-Phos-
phoglycerate is converted to 1,3-bis-phosphoglycerate by the PGK
catalyzed reaction under ATP consumption. In the GADH-catalyzed
reaction 1,3-bis-phosphoglycerate reacts subsequently to glycerin-
aldehyde-3-phosphate by reduction of NAD+ to NADH. The activity
of the N6-modified nucleotides was analyzed by the HK/GPDH-coupled
assay.29 Glucose is converted to glucose-6-phosphate under consump-
tion of ATP by the HK-catalyzed reaction. The glucose-6-phosphate
is oxidized in the second step by the GPDH-catalyzed formation of
NADP+. Changes in A340 for both assays (increase or decrease) are
proportional to the level of ATP. Each assay was performed with 100
nmol of nucleotide (0.14 mM, 700 µL of assay volume) which was
estimated by the UV spectra using the molar extinction coefficients of
15.000 M-1 for ATP and AMPPCP,38 16.400 M-1 for C8-bromo-ATP,38
17.700 M-1 for C8-carboxyethylamino-, C8-aminoethylamino-ATP,28
and DODA-HM-C8-AMPPCP/ATP 14b,c, and 17.300 M-1 for N6-
carboxymethyl-ATP30 and DODA-HM-N6-AMPPCP/ATP 17b,c. Buff-
ers, substrates, and enzymes were used in concentration according to
(21) Dietrich, C.; Schmitt, L.; Tampe´, R. Proc. Nat. Acad. Sci. USA 1995,
92, 9014-9018.
(22) Gritsch, S.; Neumaier, K.; Schmitt, L.; Tampe´, R. Biosensors &
Bioelectronics 1995, 10, 805-812.
(23) Schmitt, L.; Bohanon, T. M.; Denzinger, S.; Ringsdorf, H.; Tampe´,
R. Angew. Chem. 1996, 108, 344-347.
(24) Gershon, P. D.; Khilko, S. J. Immun. Meth. 1995, 183, 65-76.
(25) Dietrich, C.; Boscheinen, O.; Scharf, K.-D.; Schmitt, L.; Tampe´,
R. Biochemistry 1996, 35, 1100-1104.
(26) Mosbach, K. AdV. Emzymol. 1978, 46, 205-278.
(27) Lee, C.-Y.; Lazarus, L. H.; Kabakoff, D. S.; Russell, P. J.; Laver,
M.; Kaplan, N. O. Arch. Biochem. Biophys. 1977, 178, 8-18.
(28) Trayer, I. P.; Trayer, H. R.; Small, D. P.; Bottomley, R. C. Biochem.
J. 1974, 139, 609-623.
(29) Guilford, H.; Larson, P.-O.; Mosbach, K. Chem. Scr. 1972, 2, 165-
170.
(30) Lindberg, M.; Mosbach, K. Eur. J. Biochem. 1975, 53, 481-486.
(31) Haystead, C. M. M.; Gregory, P.; Sturgill, T. W.; Haystead, T. A.
J. Eur. J. Biochem. 1993, 214, 459-467.
(35) Pardee, J. D.; Spudich, J. A. Methods Enzymol. 1982, 85, 164-
181.
(36) Detmer, P.; Weber, A.; Elzinga, M.; Stephens, R. E. J. Biol. Chem.
1981, 28, 99-105.
(37) Zhou, X.; Arthur, G. J. Lipid Res. 1992, 33, 1233.
(38) Ikehara, M.; Uesugi, S. Chem. Pharm. Bull. 1969, 17, 348-352.
(32) Schulz, G. E. Curr. Opin Struc. Biol. 1992, 2, 61-67.
(33) Mimura, C. S.; Holbrook, S. R.; Ames, G. F.-L. Proc. Natl. Acad.
Sci. U.S.A. 1991, 88, 84-88.
(34) Sackmann, E. FEBS Lett. 1994, 346, 3-16.

5534 J. Am. Chem. Soc., Vol. 118, No. 24, 1996
Schmitt and Tampe´
the manufacturer (Sigma) or to the literature.30 The synthesized ATP-
lipids were solubilized in 10% Triton X100 to analyze their specific
activity. During a period of 10 min, the changes of A340 were recorded
and normalized to the changes of the same concentration of ATP. The
relative ATPase activity is determined from three independent measure-
ments. The nonhydrolyzable ATP-lipids were characterized in com-
petition assays using for the GADH/PGK or HK/GPDH assay for the
C8- or N6-modified AMPPCP-lipid, respectively. Here, a constant ATP
concentration (100 nmol in 700 µL of volume) was assayed in the
presence of an increasing molar ratio of competitor (0.01-100) which
was solubilized in Triton X100. Changes of A340 were recorded over
a period of 1 min.
Film Balance Measurements. The film balance unit consists of
an epifluorescence microscope placed above a Langmuir trough (30
mm × 220 mm). The trough carries a subphase volume of 26 mL
containing buffer A (10 mM HEPES, pH 7.5, 140 mM NaCl, 1 mM
EDTA, 5 mM MgCl2, 0.15 mM DTT) as described.21 After spreading
of the lipid mixture, the lateral pressure was adjusted by the surface
area. NBD-actin was injected into the subphase via a small hole in
the trough without touching and disturbing the air/water interface at
21.5 mN/m surface pressure. The surface tension was measured with
a Wilhelmy system (accuracy ( 0.1 mN/m). For temperature control,
peltier elements were placed below the base-plate (accuracy ( 0.2 °C).
All experiments were performed at 20 °C at a surface pressure of 21.5
mN/m. Lipid monolayers were compressed with a rate of 3 Å2/
molecule × min. Compression, expansion, surface tension, and
temperature are computer-controlled. The lateral distribution of the
fluorescence-labeled lipids (TR-DPPE) and NBD-actin was imaged by
means of an epifluorescence microscope mounted on a x-y translation
stage. The fluorescence is detected with a SIT-camera (Hamamtsu,
Hamamatsu, Japan). By changing adapted filters, various fluorescence
dyes (NBD and TR) can be observed simultaneously.
Phase Contrast Microscopy. The setup consists of a modified,
commercially available Zeiss Axiovert 10 with a magnification of
40×.39 Images of the selected vesicle were taken with a charged
coupled device (CCD) camera unit (Hamamatsu, C 3077/C 2400) and
digitized as previously described.40 The image processing software is
based on the public domain program NIH Image. ATP-lipid containing
vesicles were prepared as follows: 30 µL (15 mg/mL) of a DMPC
solution (chloroform/methanol, 3/1) containing 1 mol% DODA-HM-
C8-AMPPCP (14b) or DODA-HM-C8-AMP (13b) were placed onto
the conducting sides of two ITO-deposited cover slides. After
evaporation of the solvent in vacuum, the cover slides were placed in
a specially developed chamber with the conducting sides facing each
other. Unilamellar giant vesicles were obtained after electroswelling
of the dried lipid film in buffer B (2 mM Tris, pH 7.5, 0.2 mM DTT,
0.2 mM MgCl2) in the presence of 7 µg/mL actin for 2 h at 26 °C (1V
amplitude and 10 Hz altering voltage).41 Prior to use, actin was dialysed
extensively against nucleotide-free buffer to remove any prebound
nucleotide. The absence of ATP was determined by the GADH/PGK
assay. Actin was desorbed from the outside of the vesicle by incubation
of the vesicle solution with 0.4 mM EDTA, 2 mM Tris, pH 7.5 at 26
°C for 30 min.
δ ) 6.0 (δ, 1H, H′-1), δ ) 8.1 (s, 1H, H-8). 13C-NMR (100 MHz,
DMSO): δ ) 25.2, 27.1 (CH3C), δ ) 61.5 (C′-5), δ ) 79.1 (C′-3), δ
) 81.7 (C′-2), δ ) 87.1 (C′-4), δ ) 90.9 (C′-1), δ ) 113.2 ((CH3)2C),
d ) 119.3 (C-5), d ) 126.3 (C-8), d ) 149.7 (C-6), d ) 152.8 (C-2),
d ) 155.1 (C-4). MS (FAB pos., C13H16N5O4Br): M + H+ ) 386.2
g/mol (79Br isotope) M + H+ ) 388.2 g/mol (81Br isotope), ratio:
100: 97.
Synthesis of 2′,3′-Isopropylidene-C8-bromoadenosine-5′-mono-
phosphate-(o-hydroxyphenylester) (Br-AMP-AC-S) 3. Br-Ad-AC
2 (2.22 g, 5.74 mmol) was dissolved in 60 mL of absolute dioxane
and stirred at room temperature for 3 h after the addition of 2.5 mL
(21.90 mmol) of 2,6-lutidine and 1.21 g (6.32 mmol) of o-phenylene-
phosphochloridate in 10 mL of absolute dioxane. The precipitated 2,6-
lutidine hydrochloride was filtered off, and the reaction mixture was
quenched by the addition of 60 mL of water. After adjusting the pH
to 12 the mixture was stirred for 30 min. The aqueous phase was
extracted with chloroform until no UV-absorption of the product at
263 nm was detected in the organic phase. After adjusting the pH to
1, the aqueous phase was again extracted with chloroform. The organic
phases were combined and solvent removed in vacuum. Yield: 2.34
g (4.19 mmol) 3; 73%. TLC: Rf (3) ) 0.8 in CHCl3/MeOH (1:1).
UV: λmax ) 263 nm in DMSO. 1H-NMR (400 MHz, DMSO): δ )
1.3 (s, 3H, (CH3)2C), δ ) 1.5 (s, 3H, (CH3)2C), δ ) 4.1 (m, 2H, H′-5,
H′′-5), δ ) 4.3 (m, 1H, H′-4), δ ) 5.1 (dd, 1H, H′-3), δ ) 5.6 (dd,
1H, H′-2), δ ) 6.0 (δ, 1H, H′-1), δ ) 6.6 (m, 1H, phenyl-H-4), δ )
6.8 (m, 1H, phenyl-H-5), δ ) 6.9 (m, 1H, phenyl-H-3), δ ) 7.0 (m,
1H, phenyl-H-6), δ ) 8.2 (s, 1H, H-8). 13C-NMR (100 MHz,
DMSO): δ ) 25.2 ((CH3)2C), δ ) 26.9 ((CH3)2C), δ ) 65.6 (C′-5),
δ ) 79.1 (C′-3), δ ) 81.3 (C′-2), δ ) 85.6 (C′-4), δ ) 90.7 (C′-1), δ
) 113.2 ((CH3)2C), δ ) 117.4 (phenyl C-4), d ) 118.8 (phenyl C-5),
d ) 119.2 (C-5), d ) 121.3 (phenyl C-3), d ) 124.6 (phenyl C-6), d
) 127.1 (C-8), d ) 139.6 (phenyl C-2), d ) 148.6 (phenyl C-1), d )
149.4 (C-6), d ) 150.7 (C-2), d ) 153.2 (C-4). 31P-NMR (161 MHz,
DMSO): d ) 5.8 (s). MS (FAB neg., C19H21N5O8PBr): M - H+
)
556.6 g/mol (79Br isotope) M - H+ ) 558.6 g/mol (81Br isotope),
ratio: 100: 97.
Synthesis of 2′,3′-Isopropylidene-C8-bromoadenosine-5′-mono-
phosphate (Br-AMP-AC) 4. Br-AMP-AC-S 3 (2.34 g, 4.19 mmol)
was suspended in 160 mL of absolute dioxane and stirred at room
temperature for 30 min after the addition of 2.61 g (5.87 mmol) of
lead(IV) acetate. The reaction mixture was evaporated to dryness and
subsequently dissolved in 110 mL of 10% aqueous triethylamine, pH
12.0. After 30 min the mixture was filtered and adjusted to pH 3.0.
The aqueous phase was extracted with chloroform until the organic
phase remained colorless, concentrated in vacuum, and lyophilized.
Yield: 1.75 g (3.77 mmol) 4; 90%. UV: λmax ) 263 nm in H2O.
1H-NMR (400 MHz, D2O): δ ) 1.3 (s, 3H, (CH3)2C ), δ ) 1.5 (s, 3H,
(CH3)2C), δ ) 3.6 (m, 2H, H′-5, H′′-5), δ ) 4.2 (m, 1H, H′-4), δ )
5.0 (1H, dd, H′-3), δ ) 5.7 (1H, dd, H′-2), δ ) 6.0 (d, 1H, H′-1), δ )
8.1 (s, 1H, H-8). 13C-NMR (100 MHZ, D2O): δ ) 25.18 ((CH3)2C ),
δ ) 26.99 ((CH3)2C ), δ ) 63.75 (C′-5), δ ) 81.72 (C′-3), δ ) 81.92
(C′-2), δ ) 85.64 (C′-4), δ ) 90.46 (C′-1), δ ) 113.34 ((CH3)2C), δ
) 119.13 (C-5), δ ) 126.13 (C-8), δ ) 149.84 (C-6), δ ) 152.96
(C-2), δ ) 154.92 (C-4). 31P-NMR (161 MHz, D2O): δ ) 3.5 (s).
MS (FAB neg., C13H17N5O7PBr): M - H+ ) 464.2 g/mol (79Br isotope)
M - H+ ) 466.2 g/mol (81Br isotope), ratio: 100: 97.
Synthesis of 2′,3′-Isopropylidene-C8-(aminohexylamino)adenos-
ine-5′-monophosphate (C8-HM-AMP-AC) 5b. Br-AMP-AC 4 (875
mg, 1.88 mmol) was dissolved in 75 mL of water and stirred at 130
°C for 2 h after the addition of 7.16 g (61.21 mmol) of hexamethyl-
enediamine. After cooling to room temperature, the solution was diluted
to 250 mL and purified using a linear gradient H2O-1 M AcOH (total
volume 2l) by anion exchange chromatography on Dowex AG1 × 8
(acetate form, 1 × 23 cm, flow rate 4 mL/min). Fractions containing
the product were pooled, concentrated in vacuum, and lyophilized.
Yield: 685 mg (1.37 mmol) 5b; 73%. UV: λmax ) 278 nm in H2O.
1H-NMR (400 MHz, D2O): δ ) 1.3 (m, 7H, CH3, Ad-NH(CH2)2-
CH2CH2(CH2)2NH3+), δ ) 1.5-1.6 (m, 5H, CH3, Ad-NH(CH2)4CH2-
CH2NH3+), δ ) 2.6 (m, 4H, Ad-NH-CH2CH2(CH2)2CH2CH2NH3+), δ
) 3.3 (m, 2H, Ad-NH-CH2(CH2)5NH3+), δ ) 3.5 (m, 2H, H′-5, H′′-
5), δ ) 4.3 (m, 1H, H′-4), δ ) 4.9 (dd, 1H, H′-3), δ ) 5.4 (1H, dd,
H′-2), δ ) 6.3 (δ, 1H, H′-1), δ ) 8.9 (s, 1H, H-8). 13C-NMR (100
MHZ, D2O): δ ) 25.50 (CH3), δ ) 27.24 (CH3), δ ) 27.92 (C′′′-4),
Experimental Section
Synthesis of ATP-Lipids. Synthesis of 2′,3′-Isopropylidene-C8-
bromoadenosine (Br-Ad-AC) 2. Br-Ad 1 (2.08 g, 6.01 mmol) was
suspended in 35 mL of acetone and stirred after the addition of 3.40 g
(18 mmol) of p-toluenesulfonic acid at room temperature until the
solution became clear. The solution was neutralized with 1 M Na2-
CO3, and the aqueous phase extracted with chloroform until no UV
absorption of the product at 263 nm was detected in the organic phase.
The combined organic phases were dried over anhydrous sodium sulfate
and solvent was removed in vacuum. Yield: 2.22 g (5.74 mmol) 2;
96%. TLC: Rf (2) ) 0.8 in CHCl3/MeOH (9:1). UV: λmax ) 263
nm in DMSO. 1H-NMR (400 MHz, DMSO): δ ) 1.3 (s, 3H,
(CH3)2C), δ ) 1.5 (s, 3H, (CH3)2C), δ ) 3.5 (m, 2H, H′-5, H′′-5), δ
) 4.2 (m, 1H, H′-4), δ ) 5.0 (dd, 1H, H′-3), δ ) 5.6 (dd, 1H, H′-2),
(39) Ka¨s, J.; Sackmann, E.; Podgornick, R.; Svetina, S.; Zeks, B. J. Phys.
II France 1993, 3, 631-645.
(40) Duwe, H. P.; Ka¨s, J.; Sackmann, E. J. Phys. France 1991, 51, 945-
962.
(41) Ha¨ckl, W. 1994 Diploma Thesis, Technical University Munich.

ATP-Lipids
J. Am. Chem. Soc., Vol. 118, No. 24, 1996 5535
Figure 1. Synthesis of C8- (left lane) and N6- (right lane) amino-modified 2′,3′-isopropylidene-adenosine-5′-monophosphates (AMP-AC). The
spacer length can vary from ethylene (n ) 1) to hexamethylene (n ) 3).
δ ) 28.25 (C′′′-5), δ ) 29.38 (C′′′-3), δ ) 30.84 (C′′′-6), δ ) 42.21
(C′′′-2), δ ) 45.15 (C′′′-1), δ ) 67.12 (C′-5), δ ) 82.89 (C′-3), δ )
83.87 (C′-2), δ ) 85.99 (C′-4), δ ) 91.08 (C′-1), δ ) 118.99
(C-5), δ ) 150.98 (C-8), δ ) 151.54 (C-6), δ ) 153.34 (C-2), δ )
154.97 (C-4). 31P-NMR (161 MHz, D2O): δ ) 2.5 (s). MS (FAB
neg., C19H32N7O7P): M - H+ ) 500.4 g/mol.

5536 J. Am. Chem. Soc., Vol. 118, No. 24, 1996
Schmitt and Tampe´
Figure 2. Synthesis of C8-modified nonhydrolyzable (X ) CH2, 14a,b) and hydrolyzable (X ) O, 14c) ATP-lipids. The spacer length can vary
from ethylene (n ) 1) to hexamethylene (n ) 3).
Synthesis of 2′,3′-Isopropylidene-C8-[((dioctadecyl)amino)succi-
nylaminohexylamino]adenosine-5′-monophosphate (DODA-HM-C8-
AMP-AC) 12b. DODA-Suc-NHS 11 (1.18 g, 1.63 mmol) was
dissolved in 30 mL of absolute dioxane by heating at 40 °C and stirred
at 70 °C over night after the addition of 1.17 mL (8.16 mmol) of
absolute triethylamine and 685 mg (1.37 mmol) of C8-HM-AMP-AC
5b. The reaction mixture was quenched by the addition of 20 mL of
1 M Na2CO3. The aqueous phase was extracted four times with
chloroform. The combined organic phases were dried over anhydrous
Na2SO4, and solvent was removed in vacuum. The product was purified
by silica gel chromatography with CHCl3/MeOH/H2O/NH3 (65:25:2:
2). After removing the solvent in vacuum the product was crystallized
from acetone. Yield: 1.03 g (0.93 mmol) 12b; 68%. TLC: Rf (12b)
) 0.6 in CHCl3/MeOH/H2O/NH3 (65:25:2:2). UV: λmax ) 278 nm in
CHCl3/MeOH (3:1). 1H-NMR (400 MHz, CDCl3/MeOD 3:1): δ )
0.75 (t, 6H, CH3), δ ) 1.15 (m, 60H, CH3(CH2)15CH2CH2N), δ ) 1.3
(m, 7H, (CH2)15CH2CH2N, (CH3)2C), δ ) 1.4 (m, 4H, NH(CH2)3(CH2)2-
CH2NH-Ad), δ ) 1.5 (m, 5H, NH(CH2)2CH2(CH2)3NH-Ad, (CH3)2C),
δ ) 2.3 (t, 2H, NCOCH2CH2CONH), δ ) 2.45 (t, 2H, NCOCH2CH2-
CONH), δ ) 3.0 (dm, 6H, CH2NCO, CONHCH2), δ ) 3.2-4.5 (m,
7H, not assignable), δ ) 5.0 (dd, 1H, H′-3), δ ) 5.2 (dd, 1H, H′-2), δ
) 5.8 (d, 1H, H′-1), δ ) 7.9 (s, 1H, H-2). 13C-NMR (100 MHz, CDCl3/

ATP-Lipids
J. Am. Chem. Soc., Vol. 118, No. 24, 1996 5537
Figure 3. Synthesis of N6-modified nonhydrolyzable (X ) CH2, 17a,b) and hydrolyzable (X ) O, 17c) ATP-lipids. The spacer length can vary
from ethylene (n ) 1) to hexamethylene (n ) 3).
MeOD 3:1): δ ) 14.28, 23.00, 26.58, 26.69, 27.25, 27.39, 28.03, 28.93,
29.23, 29.69, 30.01, 31.50, 32.26 (CH3(CH2)15CH2CH2N, HCH2(CH2)4-
CH2NH-Ad, COCH2CH2CO), δ ) 39.69 (CH2NCO), δ ) 43.14
(CONHCH2), δ ) 63.75 (C′-5), δ ) 70.28 (C′-3), δ ) 72.09 (C′-2),
δ ) 84.69 (C′-4), δ ) 88.21 (C′-1), δ ) 115.42 (C-5), δ ) 118.74
((CH3)2C), δ ) 143.42 (C-8), δ ) 147.17 (C-6), δ ) 149.17 (C-2), δ
) 152.85 (C-4), δ ) 172.46 (CO), δ ) 173.58 (CO). 31P-NMR
(161 MHz, CDCl3/MeOD 3:1): δ ) 4.25 (s). MS (FAB neg,
C59H109N8O9P): M - H+ ) 1103.9 g/mol.
Synthesis of C8-[((Dioctadecyl)amino)succinylaminohexylamino]-
adenosine-5′-monophosphate (DODA-HM-C8-AMP) 13b. DODA-
HM-C8-AMP-AC 12b (1.03 g, 0.93 mmol) was dissolved in 75 mL of
chloroform/methanol 1:1 and stirred at room temperature for 30 min
after the addition of 6.5 mL of concentrated sulfuric acid. The solution
was neutralized by the addition of 1 M Na2CO3, and the aqueous phase
was extracted four times with chloroform. The organic layers were
dried over anhydrous Na2SO4, solvent was removed in vacuum, and
the product was crystallized from acetone. Yield: 887 mg (0.84 mmol)

5538 J. Am. Chem. Soc., Vol. 118, No. 24, 1996
Schmitt and Tampe´
Table 1. Relative ATPase Activity of DODA-HM-C8-ATP 14c,
DODA-HM-N6-ATP 17c, and Various C8- and N6-Modified
Nucleotides
Table 2. IC50 Values for AMPPCP, DODA-HM-C8-AMPPCP
14b, and DODA-HM-N6-AMPPCP 17b Determined in a
Competition Assay
nucleotides
relative ATPase activity [au]
nucleotides
IC50 [molar excess]
ATPa,b
100%
AMPPCPa
1.3 ( 0.003
1.9 ( 0.3
C8-bromo-ATPa
36% ( 6%
65% ( 3%
47% ( 9%
56% ( 7%
58% ( 10%
53% ( 8%
DODA-HM-C8-AMPPCP 14ba
AMPPCPb
C8-aminoethylamino-ATPa
C8-carboxyethylamino-ATPa
DODA-HM-C8-ATP 14ca
N6-carboxymethyl-ATPb
DODA-HM-N6-ATP 17cb
1.4 ( 0.004
1.2 ( 0.3
DODA-HM-N6-AMPPCP 17bb
Experimental details are given in Materials and Methods: a GADH/
PGK assay. b HK/GPDH assay.
Experimental details are given in Materials and Methods: a GADH/
PGK assay. b HK/GPDH assay.
with those measured for various other C8-modified nucleotides
(ranging from 36% for C8-Br-ATP up to 65% for C8-
aminoethylamino-ATP) and N6-modified nucleotides (58% for
N6-carboxymethyl-ATP), indicating that the lipid-linked nucle-
otides are fully active and accessible for different ATP-
hydrolyzing enzymes.
The activity of the nonhydrolyzable ATP-lipids (DODA-HM-
C8-AMPPCP 14b and DODA-HM-N6-AMPPCP 17b) were
studied in a competition assay using nonhydrolyzable AMPPCP
as reference. The IC50 values of these nonhydrolyzable ATP-
lipids are summarized in Table 2. A 1.9- or 1.2-fold molar
excess of the C8- or N6-modified AMPPCP-lipid is needed for
50% inhibition of the ATPase activity. These IC50 values are
in agreement to those obtained for AMPPCP (1.3 and 1.4 for
both assays). In summary, the activities of the hydrolyzable as
well as nonhydrolyzable ATP-lipids are comparable to their free
counterparts in micellar solution.
Activity of ATP-Lipids in Self-Assembled Monolayers.
Next, we characterized the two-dimensional properties of the
synthesized ATP-lipids. The nonhydrolyzable ATP-lipid (DODA-
HM-C8-AMPPCP, 14b) was used as model compound and
spread from organic solution at the air-water interface. As
shown by area-pressure isotherms, the pure ATP-lipid monolayer
is stable during several cycles of expansion and compression
up to 20 mN/m. But, because of the electrostatic repulsion of
the highly negatively charged lipid head group, the pure
monolayer collapsed around 25 mN/m. Interestingly, the
isotherms are strongly dependent on the presence or absence
of Mg2+ in the subphase (data not shown).
To increase the stability of the ATP-lipid monolayer and to
avoid any steric hindrance during protein binding, the nonhy-
drolyzable ATP-lipid was diluted by the matrix lipid DMPC.
DMPC/DODA-HM-C8-AMPPCP (14b) (16 nmol) (9:1 mol:
mol, doped with 0.2 mol% TR-DPPE) was spread on subphase
buffer A. This monolayer is stable up to 30 mN/m. It can be
compressed and expanded several times without observing
changes of the area-pressure isotherms. To investigate the
specific docking of an ATP-binding protein to this functionalized
ATP-lipid layer, we chose actin as a model compound. It has
been speculated that polymerization and depolymerization of
actin in proximity to membranes is responsible for cell shape,
adhesion, and movement. Because of the high surface activity
of actin (about 20 mN/m), the ATP-lipid monolayer was
compressed to 21.5 mN/m. After equilibration over night, 1.6
nmol fluorescence-labeled actin (NBD-actin) was injected
yielding an equimolar protein/ATP-lipid ratio (55 pM). It is
important to note that actin is free of any prebound ATP as
determined by the GADH/PGK assay. As shown in Figure 4a,
the surface pressure increased about 1 mN/m within 40 min
after protein injection. In addition, we observed a significant
change in the phase behavior and fluidity of the ATP-lipid
monolayer by epifluorescence microscopy. Before protein
injection, a phase separation of the lipid mixture occurred around
20 mN/m leading to small, round-shaped domains which
aggregate to clusters. After protein injection, these clusters were
nearly completely redistributed forming isolated round-shaped
13b; 90%. TLC: Rf (13b) ) 0.4 in CHCl3/MeOH/H2O (65:25:4).
UV: λmax ) 278 nm in CHCl3/MeOH (9:1). Phosphorus: 1.2 ( 0.15
mol:mol adenine. 1H-NMR (400 MHz, CDCl3/MeOD 3:1): δ ) 0.75
(t, 6H, CH3), δ ) 1.15 (m, 60H, CH3(CH2)15CH2CH2N), δ ) 1.3 (m,
4H, (CH2)15CH2CH2N), δ ) 1.4 (m, 4H, NH(CH2)3(CH2)2CH2NH-Ad),
δ ) 1.5 (m, 2H, NH(CH2)2CH2(CH2)3NH-Ad), δ ) 2.3 (t, 2H,
NCOCH2CH2CONH), δ ) 2.45 (t, 2H, NCOCH2CH2CONH), δ ) 3.0
(dm, 6H, CH2NCO, CONHCH2), δ ) 3.2-4.5 (m, 7H, not assignable),
δ ) 5.0 (dd, 1H, H′-3), δ ) 5.2 (dd, 1H, H′-2), δ ) 5.8 (d, 1H, H′-1),
δ ) 7.9 (s, 1H, H-2). 13C-NMR (100 MHz, CDCl3/MeOD 3:1): δ )
8.84 (CH3), δ ) 14.28, 23.00, 26.58, 26.69, 27.25, 27.39, 28.03, 28.93,
29.23, 29.69, 30.01, 31.50, 32.26 (CH3(CH2)15CH2CH2N, NHCH2(CH2)4-
CH2NH-Ad, COCH2CH2CO), δ ) 39.69 (CH2NCO), δ ) 43.14
(CONHCH2), δ ) 63.75 (C′-5), δ ) 70.28 (C′-3), δ ) 72.09 (C′-2),
δ ) 84.69 (C′-4), δ ) 88.21 (C′-1), δ ) 115.42 (C-5), δ ) 143.42
(C-8), δ ) 147.17 (C-6), δ ) 149.17 (C-2), δ ) 152.85 (C-4), δ )
172.46 (CO), δ ) 173.58 (CO). 31P-NMR (161 MHz, CDCl3/MeOD3:
1): δ ) 4.25 (s). MS (FAB neg, C56H105N8O9P): M - H+ ) 1063.8
g/mol.
Synthesis of C8-[((Dioctadecyl)amino)succinylaminohexylamino]-
adenosine-5′-[â,γ-methylene-triphosphate] (DODA-HM-C8-AMP-
PCP) 14b. DODA-HM-C8-AMP 13b (118 mg, 0.11 mmol) was
dissolved in 2 mL of HMPTA and stirred over night at room
temperature after the addition of 95 µL (0.22 mmol) of tri-n-octylamine
and 35 µL (0.17 mmol) diphenyl chlorophosphate. To remove reactive
phosphorylation reagent the mixture was stirred for additional 2 h at
10-2 mbar at room temperature. After the addition of 40 mg (0.22
mmol) of methylenediphosphoric acid and 190 µL (0.44 mmol) tri-n-
octylamine, the mixture was stirred for 2 h at 60 °C. After quenching
the solution with 10 mL of chloroform and 10 mL of water, the aqueous
phase was extracted five times with chloroform. Combined organic
phases were dried over anhydrous Na2SO4 and solvent was removed
in vacuum. The product was purified by silica gel chromatography
(LiChroPrep DIOL) using a step gradient from chloroform to chloroform/
methanol (9:1) and finally chloroform/methanol/water (2:3:1). The
product was eluted with chloroform/methanol/water (2:3:1) and pooled,
and solvent was removed in vacuum. Yield: 43 mg (35 µmol) 14b;
32%. TLC: Rf (14b) ) 0.9 in CHCl3/MeOH/H2O (2:3:1) on silica
DIOL plates. UV: λmax ) 278 nm in CHCl3/MeOH (3:1). Phospho-
rus: 2.8 ( 0.2 mol:mol adenine.
Results
Activity of ATP-Lipids in Micellar Solution. Self-as-
sembled hydrolyzable and nonhydrolyzable ATP-lipids represent
an energy source or a competitor for various ATP-hydrolyzing
enzymes in two dimensions. Therefore, activities of the C8-
or N6-modified nucleotide analogues were analyzed by the
GADH/PGK assay or the HK/GPDH assay, respectively. To
avoid any steric hindrance due to self-organization, the ATP-
lipids were first characterized in micellar solution of Triton
X100. This nonionic detergent by itself does not influence both
assays up to concentration of 15% (data not shown). The
activities of the hydrolyzable ATP-lipids are summarized in
Table 1. DODA-HM-C8-ATP 14c carries an activity of 56%
( 7% and DODA-HM-N6-ATP 17c of 53% ( 8%. Note, that
the same concentration of ATP represents 100% activity.
Within the range of error, the data are in very good agreement

ATP-Lipids
J. Am. Chem. Soc., Vol. 118, No. 24, 1996 5539
domains. In addition, the monolayer became rigid and the
fluidity of the lipid monolayer was drastically reduced as
observed by photobleaching experiments.
an actin layer is attached inside and outside of the ATP-lipid
containing vesicle.
Since Mg2+ is essential for ATP-binding of ATPases and
kinases, the outer actin-network was removed by complexing
Mg2+ with EDTA. In this case, the membrane is only stabilized
by an actin layer at the inside of the vesicle. Due to this asym-
metric situation, the transitions between the different vesicle
states are observed at lower temperatures. Angular forms were
induced already at 26.3 °C (Figure 6b). In addition, these locally
restricted undulations resulted in a budding process directed only
to the outside (Figure 6c-e, indicated by arrows). The
irreversible transition of the free, round-shaped vesicles occurred
also at lower temperature (30.6 °C). Figure 6f summarizes
schematically the asymmetric situation where the actin layer
outside of the vesicle is removed by incubation with EDTA.
In order to exclude the possibility that this process is guided
by unspecific, electrostatic interactions of actin with the
negatively charged membrane interface, vesicles containing 1
mol% DODA-HM-C8-AMP (13b) were prepared under identi-
cal conditions. In this case, no influence of the protein onto
the viscoelastic properties of the bilayer were observed (data
not shown). Both shape transitions and thermal undulations
are comparable to an actin-free, pure DMPC vesicle.
The same experiment was carried out in the presence of 0.1
mM AMPPCP as competitor (Figure 4b). In contrast to the
first experiment, only a slight increase in the surface pressure
of 0.2 mN/m was detected after protein injection, indicating
that binding of actin can be blocked specifically by a nonhy-
drolyzable nucleotide. The slight increase of the surface
pressure can be explained either by an unspecific absorption of
NBD-actin or by an exchange reaction between free and lipid-
anchored nucleotide. In addition, we performed a control
experiment under identical conditions using an AMP-lipid
monolayer (Figure 4c). In the case of a DMPC/DODA-HM-
C8-AMP (13b) (9:1 mol:mol) monolayer, no interaction with
the ATP-binding protein was observed. In the last two control
experiments (Figure 4b,c), we could not detect any changes of
the fluorescence micrographs (Texas-Red and NBD) or of the
dynamic properties of the monolayer after protein injection. The
fluorescence micrographs of NBD-actin differ only if the protein
was injected in the absence of the competitor. While in this
case the NBD-actin was enriched at the ATP-lipid interface,
no fluorescence-labeled protein was detected in the presence
of the competitor (data not shown). Since docking of an ATP-
binding protein can be specifically blocked by AMPPCP and
since the charges of the AMPPCP-lipid complexed by Mg2+
and the AMP-lipid are identical, an electrostatic interaction can
be excluded.
Immobilization of Actin at ATP-Lipid Containing Vesicles.
The monolayer experiments demonstrate that actin is specifically
attached to the membrane by ATP-lipids as protein anchor.
Whereas the lipid monolayer represents an ideal model to study
recognition processes at the lipid interface, vesicles are com-
monly used as model for cell membranes. In this study, giant
unilamellar DMPC vesicles which are in the order of 10 µm
were prepared in the presence of monomeric actin (7 µg/mL)
by electroswelling. These vesicles were doped with 1 mol%
DODA-HM-C8-AMPPCP (14b) or DODA-HM-C8-AMP (13b),
respectively. Due to the specific interaction with the nonhy-
drolyzable ATP-lipid, the attached actin layers stabilize the
membrane. In consequence, drastic changes on the viscoelastic
properties of the membrane such as bending modes, thermal
undulations, and shape transitions were observed by phase
contrast microscopy.
Figure 5 summarizes the situation when actin is bound onto
both sides of the lipid vesicle. Increasing the temperature from
26 °C to 37 °C, these actin-stabilized vesicles run through three
distinct states: (i) Below 29.9 °C, spherical vesicles are observed
(Figure 5a,b). In contrast to actin-free vesicles which show a
very dynamic behavior in this temperature range, no thermal
undulations of the actin-stabilized vesicles were detected. In
this case, although the temperature is above the main phase
transition temperature of the matrix lipid DMPC (24.3 °C), the
membrane is very rigid due to the attached actin layer. (ii)
Increasing the temperature above 29.9 °C, the spherical vesicles
turn into edged forms with straight lines (Figure 5c-e, indicated
by arrows). Especially, the shape at 31.0 °C (Figure 5e) is
energetically completely disfavored for a pure lipid bilayer.
These stress-induced forms can only be explained by a protein
layer attached to both sides of the vesicle. (iii) Above 36.0
°C, round-shaped vesicle are recovered. But now, the vesicles
are very dynamic which is indicated by thermal undulations of
the bilayer (data not shown). This dynamic behavior is similar
to an actin-free DMPC vesicle and can be explained by an
irreversible thermal denaturation of the adsorbed actin layer.
After cooling and reheating no protein-induced stabilization was
observed. Figure 5f describes schematically the situation where
Discussion
As shown in Figure 1, the protected C8- and N6-amino
modified nucleotides C8-AE/HM-AMP-AC 5a,b and N6-AE/
HM-AMP-AC 10a,b were synthesized starting from Br-Ad 1
and Cl-Ad 6. Although the direct phosphorylation of the
primary 5′-hydroxyl group without protecting secondary alcohols
(2′, 3′-OH) is described,42-44 the method of Reese et al.45 using
o-phenylenephosphochloridate as phosphorylating agent in the
presence of the protected diol system was found superior in
yield, simplicity of purification, and cleavage of the phosphate
protecting groups in contrast to similar reactions using 2,2,2-
tribromoethylphosphoromorphilinochloridate46 or di(2-tert-bu-
tylphenyl)phosphochloridate.47 In the former reaction, the pro-
posed selective phosphorylation concerning 5′-OH selectivity
failed. The conversion of the nucleosides Br-Ad 1 and Cl-Ad
6 to the corresponding 2′,3′-isopropylidene derivatives Br-Ad-
AC 2 and Cl-Ad-AC 7 was achieved by standard protocols.48,49
The phosphorylation of both protected nucleosides Br-Ad-AC
2 and Cl-Ad-AC 7 was performed in the presence of a slight
excess of the phosphorylation reagent and of the sterically
hindered base 2,6-lutidine at room temperature. The phospho-
triester was quenched by water and stirring at pH 12 for several
minutes at room temperature. The phosphodiesters Br-AMP-
AC-S 3 and Cl-AMP-AC-S 8 were converted into the mo-
noesters Br-AMP-AC 4 and Cl-AMP-AC 9 by oxidative
removal of the o-hydroxyphenylester using lead(IV) acetate in
organic solution and fragmentation of the intermediate in
alkaline solution at room temperature.50 All products could
easily be isolated by extraction. In summary, the synthetic route
leads to the nucleotides in moderate yields without time-
consuming purification steps. It can easily be expanded to large
(42) Yoshikawa, M.; Kato, T.; Takenishi, T. Bull. Chem. Soc. Jap. 1969,
42, 3505-3508.
(43) Sowa, T.; Ouchi, S. Bull. Chem. Soc. Jpn. 1975, 48, 2084-2090.
(44) Imai, K.-I.; Fuji, S.; Takanohashi, K.; Furukawa, Y.; Masuda, T.;
Honjo, M. J. Org. Chem. 1969, 34, 1547-1550.
(45) Khwaja, T. A.; Reese, C. B.; Stewart, J. C. M. J. Chem. Soc. (C)
1970, 2092-2100.
(46) van Boom, J. H.; Crea, R.; Luyten, W. C.; Vink, A. B. Tetrahedron
Lett. 1975, 32, 2779-2782.
(47) Hes, J.; Mertes, M. P. J. Org. Chem. 1974, 25, 3767-3769.
(48) Greene, T. W. ProtectiVe Groups in Organic Chemistry; Wiley-
Interscience: New York, 1981.
(49) Hampton, A. J. Am. Chem. Soc. 1961, 83, 3640-3645.
(50) Warren, C. D.; Jeanloz, R. W. Biochemistry 1972, 11, 2565-2572.

5540 J. Am. Chem. Soc., Vol. 118, No. 24, 1996
Schmitt and Tampe´
methods exist for this purpose that include the activation with
dicyclohexylcarbodiimide,58 carbonyldiimidazolide,59 dibutyl
phosphinothioic bromide,60 and diphenylphosphochloridate.61,62
Because of the limited solubility of the AMP-lipids in DMF or
DMSO, the best results were obtained by the activation with
diphenylphosphochloridate.61 Furthermore, HMPTA instead of
DMF/dioxane or DMF/CHCl3 as solvent promotes the activation
of the AMP-lipids and their conversion to ATP-lipids which is
in line with other studies.63 Without the isolation of the
activated compound, the ATP-lipids were formed by reaction
with diphosphoric acid.
Figure 4. Time-dependent changes of the surface pressure of a DMPC/
DODA-HM-C8-AMPPCP 14b (9:1 mol:mol, doped with 0.2 mol% TR-
DPPE) monolayer after protein injection. Lipid (16 nmol) was spread
on buffer A. The monolayer was compressed to 21.5 mN/m and
equilibrated over night. NBD-actin (1.6 nmol) was injected into
subphase buffer A yielding a final concentration of 55 pM (arrow).
All experiments were carried out under identical experimental conditions
with the exception of: (a) protein injection in the absence of free
nucleotides in buffer A, (b) protein injection in the presence of 0.1
mM AMPPCP in buffer A, and (c) protein injection to a DMPC/DODA-
HM-C8-AMP 13b (9:1 mol:mol, doped with 0.2 mol% TR-DPPE) in
the absence of free nucleotides in buffer A.
In contrast to all AMP-lipids, which could be characterized
1
by H-, 13C-, and 31P-NMR, line-broadening (around 0.5 ppm
1
for H-NMR and around 5 ppm for 31P-NMR) was observed
for the triphosphate nucleotide derivative. Various ratios of
chloroform/methanol or addition of TFA did not improve the
NMR spectra. Therefore, the determination of phosphorus37
versus adenine (determined by UV-absorption) in combination
with the detected UV-maxima and TLC visualized with iodine
(alkyl chain), ornicol (ribose moiety), and molybdenum blue
(phosphorus) on silica DIOL was used to characterize the ATP-
lipids. Noteworthy, the ATP-lipids are hydrolyzed on non-
modified silica plates catalyzed by the acidic reactive surface.
Beside the described synthesis, we reconsidered various other
synthetic routes to prepare ATP-lipids. Following alternative
pathways, we synthesized various C8-functionalized ATP
analogues via C8-bromo-ATP38 starting from ATP. Bromine
was finally substituted by ethylenediamine27 or â-alanine.64 The
N6-modified ATP derivatives were generated by the reaction
of â-propiolactone65 or ethylenimine66 with ATP under acidic
conditions and subsequent Dimroth-rearrangement in basic
solution. The C8- and N6-modified nucleotides were condensed
with carboxy- or amino-functionalized lipids using active ester
chemistry in organic solution. In the case of amino-function-
alized ATP-derivatives, only low yields of ATP-lipids were
obtained, while in the case of carboxy-functionalized ATP-
derivatives no product formation was observed. Alternatively,
we tried to phosphorylate adenosine lipids in organic solution.
Br-Ad and Cl-Ad were converted to the corresponding C8- and
N6-modified nucleosides by substitution of the halogen atoms
with ethylene- and hexamethylenediamine, respectively and
condensed with DODA-Suc-NHS. As mentioned above, a
selective phosphorylation of the adenosine-lipids in organic
solution such as chloroform or methylenechloride without
protection of the 2′,3′-diol positions failed completely.42,44 After
protection, the phosphorylation by 2,2,2-tribromoethyl-phos-
phoromorphilinochloridate46 or di(2-tert-butylphenyl)phospho-
chloridate47 brought up problems due to the incomplete cleavage
of the phosphate protecting groups. In the case of phospho-
rylchloride42 side-reactions were observed.
scale preparations. In the last step the halogen-nucleotide-
monophosphates Br-AMP-AC 4 and Cl-AMP-AC 9 were
amino-functionalized by nucleophilic substitution leading to C8-
AE/HM-AMP-AC 5a,b28 and N6-AE/HM-AMP-AC 10a,b,30
respectively.
As known from biotin-lipids,51 DNA gyrase/lipid interac-
tion,52 or hapten-lipids53 the spacer length between lipid
backbone and ligand is a crucial point in protein-ligand
interaction and the formation of two-dimensional protein crystals
at functionalized lipid interfaces.53 In our approach this
variability is introduced in the amino-functionalization step.
Here, we chose only two possibilities, ethylenediamine and
hexamethylenediamine. The last one is widely used and often
the optimal choice as known from biotin-lipid systems54 or ATP-
affinity columns.26 But, in principle, any diamine including
polyethylene glycol derivatives can be used.
In the final step, the C8- and N6-modified nucleotides (C8-
AE/HM-AMP-AC 5a,b (Figure 2) and N6-AE/HM-AMP-AC
10a,b (Figure 3)) were coupled to the carboxy-activated lipid
DODA-Suc-NHS 11.15 Several methods are known in literature
for the cleavage of the acetal in the resulting DODA-AE/HM-
C8-AMP-AC 12a,b and DODA-AE/HM-N6-AMP-AC 15a,b,
respectively.55-57 All of these methods failed due to either
incomplete cleavage or side reactions. Without any side
reactions, the isopropylidene group was cleaved only by
concentrated sulfuric acid in chloroform/methanol (1:1).48
DODA-C8-AE/HM-AMP 13a,b and DODA-AE/HM-N6-AMP
1
16a,b were characterized by TLC, UV, H-, 13C-, 31P-NMR,
In summary, this elaborated route which can be transferred
in principle to other nucleotides and lipid derivatives opens up
a wide range of possibilities to synthesize not only ATP-lipids
but also ADP-lipids, AMP-lipids, and more important various
nonhydrolyzable triphosphates.
and MS. A separate activation step is necessary to convert these
AMP-lipids into ATP-lipids (DODA-AE/HM-C8-AMPPCP
14a,b, DODA-HM-C8-ATP 14c and DODA-AE/HM-N6-AMP-
PCP 17a,b, DODA-HM-N6-ATP 17c). Surprisingly, only a few
Because of the limited accessibility of functionalized lipids
in vesicles, we characterized the function of the ATP-lipids in
(51) Hoffmann, M.; Mu¨ller, W.; Ringsdorf, H.; Rourke, A. M.; Rump,
E.; Suci, P. A. Thin Solid Films 1992, 210, 780-783.
(52) Brisson, A.; Olofsson, A.; Ringler, P.; Schmutz, M.; Stoylova, S.
Biol. Cell 1994, 80, 221-228.
(58) Moffat, J. G. Can. J. Chem. 1964, 42, 599-604.
(59) Hoard, D. E.; Ott, D. G. J. Am. Chem. Soc. 1965, 87, 1785-1788.
(60) Hata, T.; Furusawa, K.; Sekine, M. J. Chem. Soc. 1975, 196-197.
(61) Michaelson, A. M. BBA 1964, 91, 1-13.
(53) Kimura, K.; Arata, Y.; Yasuda, T.; Kinosita, K.; Nakanishi, M. J.
Immunology 1990, 69, 323-328.
(54) Darst, S. A.; Ahlers, M.; Meller, P. H.; Kubalek, E. W.; Blankenburg,
H.; Ribi, H. O.; Ringsdorf, H.; Kornberg, R. D. Biophys. J. 1991, 59, 387-
396.
(55) Huet, F.; Lechevallier, A.; Pellet, M.; Conia, J. M. Synthesis 1978,
63-65.
(56) Tewson, T. J.; Welch, M. J. J. Org. Chem. 1978, 43, 1090-1092.
(57) Guindon, Y.; Yoakim, C.; Morton, H. E. J. Org. Chem. 1984, 49,
3912-3920.
(62) Cramer, F.; Wittmann, R. Chem. Ber. 1960, 94, 331-340.
(63) Ott, D. G.; Kerr, V. N.; Hansbury, E.; Hayes, F. N. Anal. Biochem.
1967, 21, 469-472.
(64) Chan, P. H.; Hassid, W. Z. Anal. Biochem. 1975, 64, 372-379.
(65) Zappelli, A.; Rossodivita, A.; Prosperi, G.; Pappa, R.; Re, L. Eur.
J. Biochem. 1976, 62, 211-215.
(66) Bu¨ckmann, F.; Wray, V. Biotech. Appl. Biochem. 1992, 15, 303-
310.

ATP-Lipids
J. Am. Chem. Soc., Vol. 118, No. 24, 1996 5541
Figure 5. Phase contrast micrographs of a giant unilamellar DMPC vesicle containing 1 mol% DODA-HM-C8-AMPPCP (14b) at various
temperatures. Vesicles were prepared in the presence of 7 µg/mL actin. Images were recorded and digitized as described in Materials and Methods.
Figure 6. Phase contrast micrographs of a giant unilamellar DMPC vesicle containing 1 mol% DODA-HM-C8-AMPPCP (14b) at various temperatures
after incubation with 0.4 mM EDTA, 2 mM Tris, pH 7.5 for 30 min. Images were recorded and digitized as described in Materials and Methods.
detergent solution where each ATP-lipid is highly diluted in
detergent micelles (Tables 1 and 2). While the aggregation
number of Triton X100 micelle is around 140,67 the nucleotide-
lipids are strongly diluted to one molecule per ten micelles in
both assays (0.14 mM ATP-lipid, 10% Triton X100). In the
case of the C8- and N6-modified ATP-lipids, the GADH/PGK
and HK/GPDH assay was used, respectively. Two assays were
needed because modifications of the adenine system affect the
conformation of the nucleotide.68,69 Thereby the affinity of these
(68) Evans, F. E.; Lee, C. Y.; Kapmeyer, H.; Kaplan, N. O. Bioorg.
Chem. 1978, 7, 57-67.
(67) Deutscher, M. P. Methods Enzymol. 1990, 182, 247.

5542 J. Am. Chem. Soc., Vol. 118, No. 24, 1996
Schmitt and Tampe´
Figure 7. Possible applications of the novel class of ATP-lipids (schematic illustration).
modified ATP-analogues to various ATPases and kinases is
changed. PGK represents one example for this sensitive
behavior.29 As shown in Table 1 modification of the adenine
ring of ATP reduces the biological activity of those C8-ATP
analogues. The reduction of the activity relative to ATP ranges
from 36% for Br-ATP,38 42% for C8-carboxyethylamino-ATP65
to 65% for C8-aminoethylamino-ATP.27 The value for DODA-
HM-C8-ATP 14c of 47% is in agreement with the values of
the C8-modified nucleotides. The activity of the solubilized
DODA-HM-N6-ATP (53%) is comparable with N6-carboxy-
methyl-ATP (58%).70 Next, the nonhydrolyzable ATP-lipids
(14b or 17b) were used as competitors in the corresponding
enzyme assays. A 1.9- or 1.2-fold molar excess of C8- or N6-
modified AMPPCP-lipid (14b or 17b) is needed for 50%
inhibition of the ATPase activity, respectively. These IC50
values fit to those obtained for AMPPCP indicating that the
lipid-anchored nucleotides are as active and accessible as free
AMPPCP.
In the second step, the binding behavior of actin to a
nonhydrolyzable ATP-lipid monolayer was investigated at the
air-water interface. As shown in Figure 4, the binding of NBD-
actin strongly depends on the presence or absence of nonhy-
drolyzable ATP in the subphase. No interaction of NBD-actin
with an AMP-lipid monolayer was found under identical
experimental conditions. Note, that the AMP-lipid and AMP-
PCP-lipid complexed by Mg2+ have the same net charge at
neutral pH. The slight increase of the surface pressure (<0.2
mN/m) is more likely due to an exchange reaction of actin
between free and lipid-bound nucleotides than an unspecific
interaction. In lipid mixtures which contain positively charged
lipids an electrostatic interaction of the negatively charged actin
with these monolayers induces a drastic increase of the surface
pressure (20 mN/m) below a so-called ”critical pressure”.71 This
interaction, which is explained by a penetration of actin into
the monolayer, depends on the amount of positively charged
lipids within the layer. In our system, an increase of only 1
mN/m was detected. The film balance measurements are
supported by fluorescence microscopy: While a diffuse fluo-
rescence of NBD-actin and drastic decrease in membrane fluidity
was observed after protein injection in the absence of competitor,
no changes were detected in its presence. The immobilization
of the ATP-binding proteins is mediated by the nucleotide head
group. Therefore, we can exclude an unspecific electrostatic
interaction.
Finally, the behavior of a lipid-protein composite generated
by the specific immobilization of actin to ATP-lipid containing
vesicles was studied by phase contrast microscopy.72 This
method have been successfully applied to determine thermal
undulations, shape transitions, and the viscoelastic properties
of vesicles.73 As shown in Figure 5, ATP-lipid containing
vesicles prepared in the presence of actin revealed a physical
behavior completely different to actin-free vesicles.39 The shape
of these thermally stressed vesicles can only be explained by a
fixation of the bilayer from both sides by an adsorbed protein
layer.
It has to be stressed that spontaneous polymerisation of actin
in solution can be excluded under these conditions. But, based
(71) Grimard, R.; Tancre´de, Giquaud, C. BBRC 1990, 190, 1017-1022.
(72) Cortese, J.; Schwab, B.; Frieden, C.; Elson, E. Proc. Nat. Acad.
Sci. U.S.A. 1989, 86, 5773-5777.
(73) Menger, F. M.; Gabrielson, K. D. Ang. Chem. 1995, 107, 2260-
2278.
(69) Sarma, R. H. L.; C.-H.; Evans, F. E.; Yathindra, N.; Sundaralingam,
M. J. Am. Chem. Soc. 1974, 96, 7337-7348.
(70) Sakamoto, H.; Nakamura, A.; Urabe, I.; Yamada, Y.; Okada, H. J.
Ferment. Technol. 1986, 64, 511-516.

ATP-Lipids
J. Am. Chem. Soc., Vol. 118, No. 24, 1996 5543
on the specific binding, the interfacial concentration of actin is
drastically increased, resulting in a two-dimensional concentra-
tion higher than the critical concentration needed for polymer-
ization.36 At this stage, polymerization can be further promoted
by a conformational change caused by the binding of actin to
the ATP-lipid. ATP-hydrolysis is only required for the depo-
lymerization step.74 Therefore, the specific immobilization can
result in additional protein-protein interactions. In a rough
estimation the ATP-lipid possess an area of around 50 Å2.
Consequently, actin must cover an area of around 5000 Å2 to
explain the membrane stabilization only by specifically bound
protein. Based on the X-ray structure of monomeric actin, it
requires an area of around 1500 Å2.75 Thus, the strong
interaction can be explained either by long-range interaction or
by the incorporation of additional monomeric actin not directly
bound to the ATP-lipid. It has to be determined whether the
formed protein-layer is somehow similar to F-actin. A detailed
analysis of the adsorbed protein layer and the physical state of
the vesicles is in progress.
Acknowledgment. We are indebted to Walter Ha¨ckl for
expert assistance with the vesicle experiments and Dr. E.
Sackmann for helpful and critical discussions concerning the
“flicker” experiments. The excellent help for providing NMR
and MS spectra of Dr. I. Zetl, Dr. J. Sonnenbichler, and Dr. W.
Scha¨fer (NMR and MS research units) at the Max-Planck-
Institut fu¨r Biochemie (Martinsried) are gratefully acknowl-
edged. We thank Karin Scharpf for the preparation of NBD-
actin, and Klaus Neumaier, Almuth Berisch, and Annette
Kloboucek for discussion and technical assistance concerning
film balance measurements. Fritz Bu¨ckmann (GBF, Braunsch-
weig) is gratefully acknowledged for the first steps in nucleotide
chemistry. This work was supported by the Deutsche Fors-
chungs Gemeinschaft (grant Ta 157/1).
Supporting Information Available: A detailed description,
Rf values, 1H-, 13C-, 31P-NMR, and MS data of all synthesized
compounds (23 pages). See any current masthead page for
ordering information.
Our interpretation of an ATP-lipid induced formation of an
actin layer is further supported by experiments in the presence
of EDTA (Figure 6). Due to its charge, EDTA is not able to
diffuse across a lipid bilayer. As a result, the adsorbed actin
layer at the outside of the vesicle is removed, whereas the inside
layer is not affected. In comparison to the case where the
vesicles are stabilized from both sides, the transition between
different vesicle shapes and the irreversible desorption of the
protein layer is shifted to lower temperatures (from 29.9 °C to
26.3 °C for stressed vesicles and from 36.0 °C to 30.6 °C for
free vesicles). In addition, the budding process which is only
directed to the outside of the vesicle demonstrates that the
stressed membranes are only stabilized from one side by the
actin layer (Figure 6c,d).
In this study, we have described the synthesis and charac-
terization of ATP-lipids. This universal, self-assembling ligand
allows the biofunctionalization of interfaces without biotinyla-
tion steps or genetic engineering techniques. Some potential
aspects of this novel class of ATP-lipids are summarized
schematically in Figure 7. As shown by film balance experi-
ments, actin interacts specifically with an ATP-lipid monolayer.
In eukaryotic cells, the actin network is anchored at the
membrane via specific proteins. The assembly and disassembly
of this polymer network are driven by ATP. The experiments
with ATP-lipid containing vesicles indicate that actin is trapped
at the interface by specific binding to nonhydrolyzable ATP-
lipid forming a protein network. Such polymer-supported mem-
branes should be a realistic model to mimic shape transitions
of membranes mediated by the attached cytoskeleton network
(“phantom cells”) or the mechanism controlling the softness of
cell membranes.34 In supported lipid layer, this biocompatible
network can preserve the essential dynamic properties of
reconstituted membrane proteins preventing protein interaction
with the solid support.76,77 Furthermore, based on their two-
dimensional characteristics, ATP-lipids can act as energy source
for ATP-dependent membrane proteins addressing the topology
of ATP-binding sites. Due to the universality of the ATP-ligand
in nature, a vast variety of ATP-binding proteins should be
specifically bound, enriched, and oriented at these ATP-lipid
interfaces that is one of the essential steps in two-dimensional
protein crystallization at functionalized lipid monolayers.
JA953937M
(78) Abbreviations: A: Absorption; abs.: absolute; ADP: adenosine-
5′-diphosphate; AcOH: acetic acid; AMP: adenosine-5′-monophosphate;
AMPPCP: â,γ-methylene-adenosine-5′-triphosphate; ATP: adenosine-5′-
triphosphate; Br-Ad: C8-bromo-adenosine; Br-Ad-AC: 2′,3′-isopropy-
lidene-C8-bromo-adenosine;
Br-AMP-AC-S:
2′,3′-isopropylidene-
C8-bromo-adenosine-5′-monophosphate-(o-hydroxy-phenylester);
Br-AMP-AC: 2′,3′-isopropylidene-C8-bromoadenosine-5′-monophosphate;
C8-AE-AMP-AC: 2′,3′-isopropylidene-C8-aminoethylaminoadenosine-5′-
monophosphate; C8-HM-AMP-AC: 2′,3′-isopropylidene-C8-aminohexyl-
aminoadenosine-5′-monophosphate; Cl-Ad: N6-chloro-purine-riboside;
Cl-Ad-AC: 2′,3′-isopropylidene-N6-chloro-purine-riboside; Cl-AMP-
AC-S: 2′,3′-isopropylidene-N6-chloro-purine-riboside-5′-monophosphate-
(o-hydroxy-phenylester); Cl-AMP-AC: 2′,3′-isopropylidene-N6-chloro-pu-
rine-riboside-5′-monophosphate; DMF: dimethylformamide; DMPC: 1,2-
dimyristoyl-sn-glycerol-phosphocholine; DMSO: dimethyl sulfoxide;
DODA-AE-C8-AMP-AC: 2′,3′-isopropylidene-C8-[((dioctadecyl)amino)-
succinylaminoethylamino]adenosine-5′-monophosphate; DODA-AE-C8-
AMP: C8-[((dioctadecyl)amino)succinylaminoethylamino]adenosine-5′-
monophosphate; DODA-AE-C8-AMPCP: C8-[((dioctadecyl)amino)succi-
nylaminoethylamino]adenosine-5′-[â,γ-methylene-triphosphate];
DODA-AE-C8-ATP: C8-[((dioctadecyl)amino)succinylaminoethylamino]-
adenosine-5′-triphosphate; DODA-AE-N6-AMP-AC: 2′,3′-isopropylidene-
N6-[((dioctadecyl)amino)succinylaminoethyl]adenosine-5′-mono-
phosphate; DODA-AE-N6-AMP:
N6-[((dioctadecyl)amino)succinyl-
aminoethyl]adenosine-5′-monophosphate; DODA-AE-N6-AMPCP: N6-
[((dioctadecyl)amino)succinyl-aminoethyl]-adenosine-5′-[â,γ-methylene-
triphosphate]; DODA-AE-N6-ATP: N6-[((dioctadecyl)amino)succinylami-
noethyl]adenosine-5′-triphosphate; DODA-HM-C8-AMP-AC: 2′,3′-iso-
propylidene-C8-[((dioctadecyl)amino)succinylaminohexylamino]adenosine5′-
monophosphate; DODA-HM-C8-AMP: C8-[((dioctadecyl)amino)succinyl-
aminohexylamino]-adenosine-5′-monophosphate;
AMPCP: C8-[((dioctadecyl)amino)succinylaminohexylamino]adenosine-
5′-[â,γ-methylenetriphosphate]; DODA-HM-C8-ATP: C8-[((dioctadecyl)-
amino)succinylaminohexylamino]adenosine-5′-triphosphate;
DODA-HM-C8-
DODA-
HM-N6-AMP-AC: 2′,3′-isopropylidene-N6-[((dioctadecyl)amino)succinyl-
aminohexyl]adenosine-5′-monophosphate; DODA-HM-N6-AMP: N6-[((dio-
ctadecyl)amino)succinylaminohexyl]adenosine-5′-monophosphate;
DODA-HM-N6-AMPCP: N6-[((dioctadecyl)amino)succinylaminohexyl]-
adenosine-5′-[â,γ-methylenetriphosphate]; DODA-HM-N6-ATP: N6-[((di-
octadecyl)amino)succinylaminohexyl]adenosine-5′-triphosphate;
DODA-Suc-NHS: N-[(hydroxysuccinimidyl)succinyl]dioctadecylamine;
DODA: dioctadecylamine; DTT: dithiothreitrol; EDTA: ethylenediamine-
tetraacetic acid; FAB: fast atomic bombardement; GADH: glycerin-
aldehyde-3-phosphate dehydrogenase (E.C. 1.2.1.12); GPDH: glucose-6-
phosphate dehydrogenase (E.C. 1.1.1.49); HEPES: N-(2-hydroxyethyl)-
piperazine-N′-2-ethanesulfonic acid; HMPT: hexamethylphosphoric triamide;
HK: hexokinase (E.C. 2.7.1.1); IC50: inhibitor concentration needed for
50% competition; ITO: indium-tin oxide; MS: mass spectrometry;
NADH: nicotinamide-adenine dinucleotide (reduced form); NADP: nico-
tinamide-adenine dinucleotide-3′-phosphate (oxidized form); N6-AE-AMP-
AC: 2′,3′-isopropylidene-N6-aminoethyladenosine-riboside-5′-monophos-
phate; N6-HME-AMP-AC: 2′,3′-isopropylidene-N6-aminohexyl-adenosine-
(74) Weisenberg, R. C.; Deery, W. J. Nature 1976, 263, 792-793.
(75) Kabsch, W.; Mannherz, H. G.; Suck, D.; Pai, E. F.; Holmes, K.
Nature 1990, 347, 37-44.
riboside-5′-monophosphate;
NBD:
7-nitrobenzo-2-oxo-1,3-diazole;
N6-HM-AMP-AC: 2′,3′-isopropylidene-N6-aminohexyl-adenosine-riboside-
5′-monophosphate; NMR: nuclear magnetic resonance; PGK:
phosphoglycerate kinase (E.C. 2.7.2.3); TFA: trifluoroacetic acid; TLC:
thin-layer chromatography; TR-DPPE: N-(TexasRed)-1,2-dipalmitoyl-sn-
glycerol-ethanolamine; UV: ultraviolet.
(76) Ku¨hner, M.; Tampe´, R. Sackmann, E. Biophys. J. 1994, 67, 217-
226.
(77) Dietrich, D.; Tampe´, R. BBA 1995, 1238, 183-191.