Anchor chain length alters the apparent mechanism of chloride channel function in SCMTR derivatives

Article abstract of DOI:10.1039/b211629b

Two membrane-anchored heptapeptides have been prepared and their pore-formation behavior in phospholipid bilayer membranes has been found to differ profoundly as a result only of alkyl chain length.

Full text of DOI:10.1039/b211629b

Anchor chain length alters the apparent mechanism of chloride channel
function in SCMTR derivatives
Paul H. Schlesinger,*a Natasha K. Djedovicˆ,b Riccardo Ferdani,b Jolanta Pajewska,b Robert Pajewskib
and George W. Gokel*b
a Department of Cell Biology and Physiology, Washington University School of Medicine, 660 S. Euclid
Ave., St. Louis, MO 63110, USA
b Program in Bioorganic Chemistry, Division of Bioorganic Chemistry, and Department of Molecular
Biology & Pharmacology, Washington University School of Medicine, 660 S. Euclid Ave., St. Louis, MO
63110, USA
Received (in Columbia, MO, USA) 2nd November 2002, Accepted 21st November 2002
First published as an Advance Article on the web 7th January 2003
Two membrane-anchored heptapeptides have been pre-
pared and their pore-formation behavior in phospholipid
bilayer membranes has been found to differ profoundly as a
result only of alkyl chain length.
CH2OCH2CO-GGG-OCH2Ph (76%, light yellow oil). As
previously reported,3 the PGGG fragment was added according
to the following sequence: (a) coupling TsOH·H2N-GGG-
OCH2Ph with Boc-proline to give Boc-PGGG-OCH2Ph, and
(b) Boc group removal to give ClH3N-PGGG-OCH2Ph.
Deprotection of (C10H21)2NCOCH2OCH2CO-GGG-OCH2Ph
gave the acid, which was coupled (EDCI, HOBt, Et3N, CH2Cl2,
0 °C?rt, 30 h) with ClNH3-PGGG-OCH2Ph to give 3 (60%,
white solid, mp 127–128 °C). Proton and 13C-NMR spectra
corresponded to the assigned structures.
Chloride ion is critical for cellular function but the mechanism
by which it is transported remains obscure.1 The recent solid
state structures of prokaryotic ClC Cl2 channels from Salmo-
nella typhimurium and Escherichia coli2 have dramatically
enhanced our understanding of structural relationships in these
elaborate proteins. The channels are dimeric; each nearly
autonomous monomer unit possesses 18 helical segments. The
inherent complexity of such a molecule is obvious. Surely,
compounds having far simpler structures must have functioned
as chloride channels, albeit less effectively, at an earlier stage of
evolution.
We recently reported the preparation of a synthetic mem-
brane-insertable, chloride transporter (‘SCMTR’).3,4 The com-
pound was designed to possess (1) a membrane anchoring unit,
(2) a heptapeptide that serves as headgroup, entry portal, and
selectivity filter, and (3) a connector for (1) and (2) that mimics
a phospholipid’s midpolar regime.5 We noted that among
modern chloride transporters, all ClC chloride protein channels
have, in their anion pathway, the conserved sequence
GKxGPxxH.6,7 Further, the presence of proline is known to
affect channel selectivity,8,9 and to induce a ‘kink’ or ‘bend’
(hinge-bend regime, GxxP) in a protein chain.10 ProlineAs
critical presence at the apex of C-peptideAs helix-loop-helix
motif11 additionally inspired its use in our anion channel
design.
We chose a symmetrical dialkylamine (R2N ~ ) to serve as the
membrane anchor in this family of compounds. For the
preparation of 2, dioctadecylamine and diglycolic anhydride
were heated (C6H5CH3, reflux, 48 h) to give (C18H37)2NCO-
CH2OCH2COOH, 1. When coupled to GGGPGGG-OCH2Ph
(in a sequence of steps previously reported), (C18H37)2NCO-
CH2OCH2CO-GGGPGGG-OCH2Ph, 2, was obtained.3 When
inserted into phospholipid liposomes, chloride was released in a
concentration dependent fashion. The analog of 2 in which Leu
replaced Pro was much less active. Remarkably, 2 showed
voltage dependent gating when assessed in bilayer clamp
experiments.3 We have now prepared an exact analog of 2 in
which the octadecyl chains are replaced by decyl groups to give
(C10H21)2NCOCH2OCH2CO-GGGPGGG-OCH2Ph, 3. This
novel compound transports Cl2 but exhibits a range of
properties quite different from those observed with 2. Specifi-
cally, the pore sizes of channels formed by 2 and 3 are the same
but the apparent ionophoretic mechanism differs.
We evaluated Cl2 release from unilamellar liposomes of
defined size and composition (150 ± 16 nm, 30% 1,2-dioleoyl-
sn-glycero-3-phosphate, 70% 1,2-dioleoyl-sn-glycero-3-phos-
phocholine). Concentration dependent chloride release was
observed with both 23 and 3 (data not shown). The apparently
greater Cl2 release activity of (C10H21)2NCOCH2OCH2CO-
GGGPGGG-OCH2Ph, 3, encouraged us to characterize the
pores further. This was done by using the ‘dextran block’
method, involving carboxyfluorescein (CF) release from lipo-
somes. In this experiment, transport of CF is attempted in the
presence of oligomerized sugars of various, known effective
diameters. Carboxyfluorescein self-quenches when trapped
within a vesicle but is readily detected by fluorescence when it
emerges from the liposome.12,13 When CF transport is blocked,
reduced fluorescence (corresponding to dye release) is ob-
served. Note that the diameter of the pore must be slightly larger
than the dextran used to block it.
Either 2 or 3 (100 mM) was inserted into vesicles, prepared as
above except that carboxyfluorescein (20 mM, 10 mM HEPES,
pH = 7.0) was trapped within the liposomes. The extravesicular
medium contained 100 mM KNO3 (10 mM HEPES, pH = 7.0).
The concentration dependent CF release data are shown in Fig.
1(a) and (b) for 2 and 3, respectively. Similar CF release rates
are observed for 2 at 25.3 mM and for 3 at 1.67 mM. This implies
that 3, which has shorter anchor chains, is ~ 15-fold more active
than is 2 under the conditions of this experiment.
The concentration dependent activities for 2 and 3 are
summarized in the log–log Hill14 plots (Fig. 1(c)). In these plots,
the slopes represent the molecularity of the limiting step. Both
plots have a slope of 2.0 ± 0.3 at 50% of the maximum rate of
pore activation. At lower concentration, the slope of 3 (curved
line) decreases to 0.89 ± 0.3 suggesting that dimer formation no
longer limits the rate of pore activation. The apparent pore sizes
The synthesis of 3 was undertaken as follows. Diglycolic
anhydride was heated with didecylamine in THF (reflux, 48 h).
After removal of the solvent and crystallization of the crude
product from Et2O, (C10H21)2NCOCH2OCH2COOH was ob-
tained (88%) as a white solid, mp 51–52 °C. This was coupled
to TsOH·H2N-GGG-OCH2Ph to give (C10H21)2NCO-
308
CHEM. COMMUN., 2003, 308–309
This journal is © The Royal Society of Chemistry 2003

Fig. 1 (a) Carboxyfluorescein release for SCMTR (2). (b) Carboxyfluorescein release for 3. (c) Hill plots for 2 and 3. (d) Apparent pore diameters for 2 and
3.
for 2 and 3 were obtained using size-specific block by dextran
polymers (Fig. 1(d)). The apparent pore activation is diminished
by dextrans of a size appropriate to block the pore.15 When the
blocker becomes too large, the pore activity is no longer
affected.16 Fig. 1(d) shows the near identity in sizes of the pores
formed by 2 and 3: 8.2 ± 0.44 Å. The formation of dimer pores
having nearly identical diameters supports a model in which the
dimerization step does not limit pore activation at low
concentrations of 2.
It occurred to us that these anchored heptapeptides might
form solution aggregates that fuse into liposomes affording
preformed pore structures. The aggregation behavior of both 2
and 3 were studied from ~ 2–60 mM by dynamic light
scattering. Compound 3 consistently formed 350 ± 70 nm
liposomes throughout this concentration range. Compound 2
formed 500 nm aggregates at < 30 mM but larger and more
complex aggregates were noted as concentration increased. In
the liposome assay, fluorescence dequenching is dominated by
the initial pore activation15 and is therefore sensitive to a shift
from a bi- to monomolecular mechanism.
We evaluated the slope conductances of 2 and 3 in planar
lipid bilayers under steady-state conditions of pore activity in
order to determine ion conduction characteristics (see Fig. 2). In
this experiment, 5.2 mM of 3 was added to the preformed bilayer
until the current stabilized; current was then determined as a
function of voltage. The same protocol was followed for 2,
except the compound was studied at both 23.6 and 89.7 mM. The
conductance of bis(C18) peptide 2 was 349 ± 69 pS under steady
state conditions (compared to 1.3 nS under ion gradient
conditions3). Shorter-chained 3 showed a conductance of 573 ±
13 pS (steady state). Both compounds displayed rectification at
voltages below Erev. This is in accord with the voltage
dependent gating behavior previously reported for 2.3 Less
curvature was observed in the data plot for 3 than for 2. Both
compounds produced stable pores of constant steady-state
conductance. The per-mole size of this conductance was
consistent with the concentration dependence of pore activation
in liposomes: 3 was 17-fold greater than 2. We conclude that at
least two mechanisms of pore activation occur. One involves
dimerization of monomeric, membrane-inserted, pore-forming
molecules. A second reflects the membrane insertion of a
preorganized pore unit. The latter occurs at a very low rate and
is observed only when the former pathway becomes slow.
The complex properties of these two simple molecules are
remarkable. Even more unexpected are the striking differences
in their insertion, transport, and gating properties. These
chemically simple channel-formers exhibit an astonishing
complexity in their membrane functions. They demonstrate
properties previously associated only with proteins and peptides
of substantial size. Their value as prototypes for channel
evolution is apparent from this astounding complexity.
We gratefully acknowledge support of this work by the NIH
(GM-36262) and by the Kilo Foundation.
Notes and references
1 C. Miller, Curr. Opin. Chem. Biol., 2000, 4, 148.
2 R. Dutzler, E. B. Campbell, M. Cadene, B. T. Chait and R. MacKinnon,
Nature, 2002, 415, 287.
3 P. H. Schlesinger, R. Ferdani, J. Liu, J. Pajewska, R. Pajewski, M. Saito,
H. Shabany and G. W. Gokel, J. Am. Chem. Soc., 2002, 124, 1848.
4 P. H. Schlesinger, R. Ferdani, R. Pajewski, J. Pajewska and G. W.
Gokel, Chem. Commun., 2002, 840.
5 G. W. Gokel, Chem. Comm., 2000, 1.
6 C. Fahlke, H. T. Yu, C. L. Beck, T. H. Rhodes and A. L. George Jr.,
Nature (London), 1997, 390, 529.
7 C. Fahlke, R. R. Desai, N. Gillani and A. L. George, J. Biol. Chem.,
2001, 276, 1759.
8 P. J. Corringer, S. Bertrand, J. Galzi, A. Devillers-Thiery, J.-P.
Changeux and D. Bertrandt, Neuron, 1999, 22, 831.
9 J. L. Galzi, A. Devillers-Thièry, N. Hussy, S. Bertrand, J. P. Changeux
and D. Bertrand, Nature, 1992, 359, 500.
10 C. J. Brandl and C. M. Deber, Proc. Natl. Acad. Sci. USA, 1986, 83,
917.
11 Y. Ido, A. Vindigni, K. Chang, L. Stramm, R. Chance, W. F. Heath, R.
D. DiMarchi, E. DiCera and J. R. Williamson, Science, 1997, 277,
563.
12 B. Hille, J. Gen. Physiol., 1968, 51, 199.
13 D. Levitt, Curr. Top. Membranes Transport, 1984, 21, 181.
14 I. Segel, Enzyme Kinetics. Behavior and Analysis of Rapid Equilibrium
and Steady-State Enzyme Systems, John Wiley & Sons, New York,
1975.
15 M. Saito, S. J. Korsmeyer and P. H. Schlesinger, Nat. Cell Biol., 2000,
553.
Fig. 2 Comparison of conductance in bilayers from insertion of 3 or 2. For
3 (C10, blue), the bilayer was exposed to 5.2 mM until the current stabilized;
current was determined as a function of voltage. The same protocol was
followed for 2, except the compound was at 23.6 or 89.7 mM. Each
experiment was repeated 3 times and error bars are shown.
16 N. Sakai, C. Ni, S. M. Bezrukov and S. Matile, Bioorg. Med. Chem.
Lett., 1998, 8, 2743.
CHEM. COMMUN., 2003, 308–309
309