Membrane-surface anchoring of charged diacylglycerol-lactones correlates with biological activities

Article abstract of DOI:10.1002/cbic.201000343

Synthetic diacylglycerol-lactones (DAG-lactones) are effective modulators of critical cellular signaling pathways, downstream of the lipophilic second messenger diacylglycerol, that activate a host of protein kinase C (PKC) isozymes and other nonkinase pr

Full text of DOI:10.1002/cbic.201000343

DOI: 10.1002/cbic.201000343
Membrane-Surface Anchoring of Charged Diacylglycerol-
Lactones Correlates with Biological Activities
Or Raifman,[a] Sofiya Kolusheva,[a] Said El Kazzouli,[b] Dina M. Sigano,[b] Noemi Kedei,[c]
Nancy E. Lewin,[c] Ruben Lopez-Nicolas,[d] Ana Ortiz-Espin,[d] Juan C. Gomez-Fernandez,[d]
Peter M. Blumberg,[c] Victor E. Marquez,*[b] Senena Corbalan-Garcia,*[d] and Raz Jelinek*[a]
Synthetic diacylglycerol-lactones (DAG-lactones) are effective
modulators of critical cellular signaling pathways, downstream
of the lipophilic second messenger diacylglycerol, that activate
a host of protein kinase C (PKC) isozymes and other nonkinase
proteins that share similar C1 membrane-targeting domains
with PKC. A fundamental determinant of the biological activity
of these amphiphilic molecules is the nature of their interac-
tions with cellular membranes. This study examines the biolog-
ical properties of charged DAG-lactones exhibiting different
alkyl groups attached to the heterocyclic nitrogen of an a-pyri-
dylalkylidene chain, and particularly the relationship between
membrane interactions of the substituted DAG-lactones and
their respective biological activities. Our results suggest that
bilayer interface localization of the N-alkyl chain in the R2 posi-
tion of the DAG-lactones inhibits translocation of PKC isoen-
zymes onto the cellular membrane. However, the orientation
of a branched alkyl chain at the bilayer surface facilitates PKC
binding and translocation. This investigation emphasizes that
bilayer localization of the aromatic side residues of positively
charged DAG-lactone derivatives play a central role in deter-
mining biological activity, and that this factor contributes to
the diversity of biological actions of these synthetic biomimetic
ligands.
Introduction
The lipophilic second messenger sn-1,2-diacylglycerol (DAG) is
released in situ from membrane phosphatidylinositol 4,5-bi-
sphosphate through the action of phospholipase C in response
to the occupancy of a wide range of G protein-coupled recep-
tors and receptor tyrosine kinases.[1] The other hydrolysis prod-
uct, inositol-1,4,5-triphosphate (IP3) triggers in turn the release
of calcium from intracellular stores. The released Ca2+ ions pro-
mote a weak and reversible association of the classical protein
kinase C (PKC) isoforms with the inner leaflet of the membrane,
after which the PKC penetrates further into the membrane by
its interaction with DAG. This process is accompanied by the
folding out of an N-terminal pseudosubstrate region, which
allows access of a myriad of substrates to the binding site of
the enzyme.
During the past several years we have developed a family of
conformationally constrained DAG analogues, known as DAG-
lactones, which were designed to overcome this spread in po-
tency between the natural product ligands and DAG.[6] The
generation of the prototypical DAG-lactone template (I) is con-
ceptually simple and involves the joining of the sn-2-O-acyl
moiety of DAG to the glycerol backbone with an additional
carbon atom to complete a five-membered ring (Scheme 1).
[a] O. Raifman, Dr. S. Kolusheva, Prof. R. Jelinek
Department of Chemistry, Ben Gurion University
Beer Sheva 84105 (Israel)
Fax: (+972)8647-2943
E-mail: razj@bgu.ac.il
[b] Dr. S. El Kazzouli, Dr. D. M. Sigano, Dr. V. E. Marquez
Laboratory of Medicinal Chemistry, Center for Cancer Research
National Cancer Institute at Frederick, National Institutes of Health
Frederick, MD 21702 (USA)
As a second messenger, DAG mediates the action of numer-
ous growth factors, hormones and cytokines by activating
members of the PKC family of enzymes, as well as several
other families of signaling proteins, for example, RasGRPs and
chimaerins, that share with PKC the C1 domain as a DAG-rec-
ognition motif. Many of these signaling pathways feature
prominently in the development and properties of cancer
cells[2,3] and, in consequence, PKC isozymes are being actively
pursued as therapeutic targets for cancer.[4] The majority of C1
binding ligands that are utilized are structurally rigid and com-
plex natural products, such as the prototypical phorbol esters
and the bryostatins.[5] These compounds bind to their C1 re-
ceptors with nanomolar affinities and are more than three
orders of magnitude more potent than the very flexible, natu-
ral DAG agonists.
Fax: (+1)301-846-6033
E-mail: marquezv@dc37a.nci.nih.gov
[c] Dr. N. Kedei, Dr. N. E. Lewin, Dr. P. M. Blumberg
Laboratory of Cancer Biology and Genetics
Center for Cancer Research, National Cancer Institute
Bethesda, MD 20892 (USA)
[d] R. Lopez-Nicolas, A. Ortiz-Espin, J. C. Gomez-Fernandez,
Prof. S. Corbalan-Garcia
Department of Biochemistry and Molecular Biology-A
Veterinary School, University of Murcia, 30100 Murcia (Spain)
Fax: (+34)968-364147
E-mail: senena@um.es
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cbic.201000343.
ChemBioChem 2010, 11, 2003 – 2009
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2003

V. E. Marquez, S. Corbalan-Garcia, R. Jelinek et al.
membranes, and this study aims to examine whether the pat-
tern of membrane association might account for the biological
differences. Application of biophysical techniques, including
fluorescence quenching[9] and differential scanning calorimetry
(DSC),[10] underscore the prominent contribution of the alkyl
residue in the aromatic unit (position R2) to the biological ac-
tivities of the DAG-lactone. In particular, our data point to a re-
lationship between bilayer surface localization and disruption
by the alkyl residues and binding/translocation of the PKC iso-
forms. Overall, the results expand our understanding of the
molecular parameters effecting PKC translocation to mem-
branes by synthetic DAG-lactones.
Results
Membrane translocation
Analysis of the translocation of three PKC isoforms induced by
DAG-lactones 1–5 showed pronounced differences among the
compounds (Figure 1 and Table 1). The translocation analysis
utilized MCF-7 cells that were transiently transfected with the
fluorescent constructs PKCa–EGFP (representative of classical
isoenzymes), PKCe–EGFP or PKCd–ECFP (representative of each
of the two subgroups of novel PKCs). Also presented in
Figure 1 and Table 1 is a comparison with the translocation in-
duced by phorbol 12-myristate 13-acetate (PMA, the standard
derivative used to characterize responses of PKC to phorbol
esters or other ligands targeted to the C1 domain).
Scheme 1. I) Structure of a DAG-lactone prototype with the embedded DAG
backbone shown with heavier lines. II) Structure of a prototype DAG-lactone
with an a-arylalkylidene chain. The lower panel shows the structures of com-
pounds 1–5.
Some of the most interesting DAG-lactone templates that we
have studied are those with an a-arylalkylidene chain (II).
These compounds seem to bind to PKC with the acyl chain
(R1) oriented toward the interior of the membrane and the a-
arylalkylidene chain directed to the surface of the C1 domain
adjacent to the lipid interface.[7]
Figure 1 shows the DAG-lactone 1 induced plasma mem-
brane localization of all three isoenzymes tested. The percent-
age of protein localized in the plasma membrane (Rmax) was
similar for the three isoenzymes (Table 1), although the mem-
brane distribution was heterogeneous and the three proteins
appeared localized in distinct areas of the plasma membrane.
In addition, Figure 1 indicates that 1 gave rise to localization of
PKCd–ECFP in vesicles distributed through the cytosol (Fig-
ure 1E) suggesting that the protein and/or lipid composition
of these vesicles influenced the targeting effect of the DAG-lac-
tone as compared to PMA in the case of this isoenzyme.
Because electrostatic interactions are important for the initial
membrane-binding process of PKC, particularly at the lipid-bi-
layer interface, we decided to exploit the role of electrostatic
attraction by converting the aryl group in II to a substituted
pyridine ring. We reasoned that the pyridine ring could bear
several alkyl groups of various sizes to generate a series of N-
alkylpyridinium chains that could bind to negatively charged
phospholipids or engage in cation–p interactions between the
positively charged p system and aromatic rings of certain
amino acids in the C1 domain, such as tryptophan, that ought
to be stabilized through van der Waals and p–p stacking inter-
actions.[8]
Table 1. Plasma membrane translocation parameters calculated for the
different PKC isoenzymes in MCF-7 cells stimulated with DAG-lactones.
The specific charged compounds that were designed (2–5,
Scheme 1) have different a-(4-N-alklypyridinium)alkylidene side
chains, which are derived from the parent, neutral DAG-lactone
1 by simple alkylation. Although the stereochemistry around
the double bond could be either E or Z, all of the compounds
with the exception of 5 favored the E stereochemistry. In all
the DAG-lactones (1–5) the most efficient, branched acyl chain
known to enhance membrane penetration was chosen,[6] and
at the other end of the molecule the a-(4-N-alkylpyridinium)al-
kylidene side chains were expected to interact near the surface
of the C1 domain adjacent to the lipid interface.
[b]
DAG-lactone
Isoenzyme
Rmax[a] [%]
t1/2 [s]
1
PKCa
PKCe
PKCd
PKCa
PKCe
PKCd
PKCa
PKCe
PKCd
PKCa
PKCe
PKCd
0.62Æ0.04
0.6Æ0.16
0.56Æ0.12
no effect
no effect
no effect
62Æ13
3.5Æ1.7
15Æ4.3
no effect
no effect
no effect
no effect
21Æ5
2–4
5
no effect
0.77Æ0.1
0.65Æ0.07
0.64Æ0.08
0.68Æ0.08
0.72Æ0.06
24Æ9
PMA
547Æ192
31Æ18
Here, we investigated the translocation properties, PKC bind-
ing, and lipid bilayer interactions of DAG-lactones 1–5
(Scheme 1). These ligands are shown to induce different pat-
terns and kinetic profiles for translocation of PKC isoforms to
47Æ22
[a] Maximal percentage of protein localized in the plasma membrane.
[b] Half-time of plasma membrane localization.
2004
www.chembiochem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2010, 11, 2003 – 2009

Anchoring of Charged Diacylglycerol-Lactones
(n-alkylpyridinium) chains of dif-
ferent lengths in the R2 position
of 1 completely eliminated the
translocation capabilities when
these derivatives were added to
MCF-7 cells transfected with the
different isoenzymes (Table 1).
This result highlights the biologi-
cal significance of the positive
charge generated by the pres-
ence of N-(n-alkylpyridinium) res-
idues, and the following experi-
mental analysis is designed to
elucidate the effects of the side
chains.
The DAG-lactone 5, which still
bears a positive charge, displays
a branched alkyl chain in the R2
position instead of a linear n-
alkyl chain. Remarkably, this
structural change facilitated
translocation of the novel PKCs
(PKCe–EGFP and PKCd–ECFP)
from the cytosol to the plasma
membrane at very similar rates
(Figure 2A–D and Table 1), al-
though the percentage of PKCe–
EGFP localized at the plasma
membrane was slightly higher
than
that
of
PKCd–ECFP
(Table 1). This DAG-lactone deriv-
ative also induced the localiza-
tion of PKCd–ECFP in vesicles
distributed through the cytosol
(Figure 2B). This effect was not
observed when the cells were
stimulated with PMA, suggesting
that the localization of the
enzyme in vesicles is directly re-
lated to effects of the DAG-lac-
tones.
Figure 1. Effect of DAG-lactone 1 on the plasma membrane localization of PKC isoenzymes. Confocal images of
MCF-7 cells expressing: A) PKCa–EGFP, C) PKCe–EGFP, or E) PKCd–ECFP stimulated with 40 mm DAG-lactone 1
(upper panels) or 40 mm PMA (lower panels). Time in seconds after stimulation is indicated in each micrograph.
The protein localization was measured by a line profile (pixel density) traced in each frame as indicated in the Ex-
perimental Section. The resulting net change in: B) PKCa–EGFP, D) PKCe–EGFP, or F) PKCd–ECFP plasma mem-
&
*
brane localization upon DAG-lactone ( ) and PMA ( ) stimulations is expressed as the (ImbÀIcyt)/Imb ratio (%) and is
represented versus time. The arrow indicates the stimulation time; Rmax and t1/2 parameters were calculated graph-
ically.[18] The profiles are representative of the results obtained in the cells analyzed (n=16 for both DAG-lactone
and PMA for PKCa–EGFP; n=12 and 10 for DAG-lactone and PMA, respectively, for PKCe–EGFP; n=13 and 12 for
DAG-lactone and PMA, respectively, for PKCd–ECFP). The scale bars correspond to 12 mm.
Dissociation constants
The binding potencies of the
DAG-lactones were evaluated in
vitro by competition of binding
Comparison of the half-times of plasma membrane localiza-
tion (t1/2, Table 1) demonstrates that 1 effected PKCe–EGFP
translocation to the plasma membrane significantly faster as
compared to PKCd–ECFP followed by PKCa–EGFP, suggesting
that PKCe–EGFP exhibits higher affinity for 1 compared to
PKCd–ECFP and PKCa–EGFP. Figure 1 and Table 1 further show
that 1 induced significantly faster translocation to the plasma
membrane than PMA with all PKC isoforms; this is consistent
with more rapid penetration. The additions of nonbranched N-
of [20-3H]phorbol 12,13-dibutyrate to PKC in the presence of
100 mgmLÀ1 phosphatidylserine (Table 2). In parallel with the
findings from the cellular translocation studies, DAG-lactone 5,
possessing the branched alkyl chain in the R2 position, showed
similar binding potency to the uncharged DAG-lactone 1. In
contrast, the three DAG-lactones 2–4, which possessed non-
branched N-(n-alkylpyridinium) chains together with a positive
charge, were approximately an order of magnitude less
potent.
ChemBioChem 2010, 11, 2003 – 2009
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
2005

V. E. Marquez, S. Corbalan-Garcia, R. Jelinek et al.
ing of NBD by water-dissolved
sodium dithionite following pre-
incubation of the vesicles with
the DAG-lactones, providing a
measure of bilayer interactions
of the compounds.[9]
Figure 3 demonstrates that in-
cubation of the NBD-PE–DMPC
vesicles with the DAG-lactones
yielded significant changes in
the rate of dithionite-induced
fluorescence quenching of the
bilayer-embedded dye. Impor-
tantly, while 1 did not effect the
fluorescence quenching rate of
NBD (compared to the control
vesicles), all the other charged
Figure 2. Effect of DAG-lactone 5 on the plasma membrane localization of PKC isoenzymes. Confocal images of
DAG-lactones examined yielded
faster quenching compared with
vesicles that were not preincu-
bated with the DAG-lactones
prior to addition of sodium di-
thionite. Faster quenching of the
MCF-7 cells expressing: A) PKCe–EGFP, or B) PKCd–ECFP stimulated with 40 mm DAG-lactone 5. Time in seconds
after stimulation is indicated in each micrograph. The scale bars correspond to 12 mm. C), D) The percentage of
protein localization was measured by a line profile (pixel density) traced in each frame as indicated in the Experi-
mental Section. The resulting net change in PKCe–EGFP or PKCd–ECFP plasma membrane localization upon DAG-
*
lactone ( ) stimulation is expressed as the (ImbÀIcyt)/Imb ratio (%) and is represented versus time. The arrow indi-
cates the stimulation time; Rmax and t1/2 parameters were calculated graphically.[18] The profiles are representative
of the results obtained in the cells analyzed (n=12 and 10 for PKCe–EGFP or PKCd–ECFP, respectively).
Table 2. Binding potencies of DAG-lactones for PKC determined in vitro
in the presence of 100 mgmLÀ1 phosphatidylserine.
Compound
Ki [nm][a]
Compound
Ki [nm][a]
DAG-lactone 1
DAG-lactone 3
DAG-lactone 5
88.5Æ3.3
DAG-lactone 2
DAG-lactone 4
1800Æ150
1210Æ41
1340Æ240
188Æ17
[a] Values are the meanÆSEM of triplicate experiments. Binding poten-
cies were determined by competition of [3H]PDBu binding and represent
the inhibitor dissociation constant, Ki.
Biophysical analyses
The membrane translocation and binding analyses summarized
in Figures 1 and 2 and Tables 1 and 2 point to significant differ-
ences in biological activities of the DAG-lactones 1–5. To evalu-
ate the relationships between the biological properties of the
compounds and their association with the cell membrane we
applied several biophysical techniques for investigating inter-
actions of 1–5 with lipid vesicles (Figures 3 and 4). In particular,
the biophysical experiments were designed to assess the con-
tributions of the positive charge and the alkyl residues in posi-
tion R2 to membrane interactions and localization of the com-
pounds.
Figure 3. Fluorescence quenching. Fluorescence intensities of NBD-PE dye
embedded within DMPC vesicles following preincubation with DAG-lactones
1–5. Initial fluorescence (at t=0, upon addition of sodium thionite) is de-
fined as 100%.
fluorescent dye is indicative of disruption of the bilayer head-
group region, which consequently results in greater exposure
of the dye to the water-soluble dithionite.[9]
The differences in the quenching rates of NBD among the
DAG-lactones 2–5, apparent in Figure 3, point to distinct
modes of bilayer binding by the compounds. Specifically, pre-
incubation of 4 with the NBD-PE–DMPC vesicles yielded pro-
nounced quenching significantly faster than all other com-
pounds (Figure 3). This result indicates substantial disruption
of the lipid bilayer surface by 4. Figure 3 also shows that 2 and
3 enhanced the dithionite-induced quenching rate of NBD, al-
though to a lesser extent compared to 4. These results are still
To probe the extent of the localization of DAG-lactones 1–5
at the bilayer/water interface we carried out fluorescence
quenching experiments utilizing DMPC vesicles into which the
fluorescent probe NBD-PE was incorporated[9] (Figure 3). The
NBD dye in lipid vesicles is embedded close to the bilayer in-
terface, thus providing a useful marker for surface interactions
of membrane-active compounds. The experiments summarized
in Figure 3 depicts the modulation of the fluorescence quench-
2006
www.chembiochem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2010, 11, 2003 – 2009

Anchoring of Charged Diacylglycerol-Lactones
In contrast to the minor modulation of the DMPC thermo-
gram by DAG-lactones 2–4, Figure 4 demonstrates that 1 and
5 give rise to significant spectral changes. In particular, the ap-
pearance of an extremely broad peak following incubation of
the DMPC vesicles with 1 demonstrates that the impact of this
DAG-lactone upon the cooperative properties of the bilayer
was substantial. The broad thermal transition induced by 1 in-
dicates that this derivative is closely incorporated within the
lipid bilayer, strongly modulating lipid interdigitation.
The DSC trace of 5 also displays a significant difference com-
pared to DAG-lactones 2–4 (Figure 4). Interestingly, the DSC
thermogram of 5 appears to show two populations of the mol-
ecule. One population, giving rise to a relatively narrow transi-
tion exhibiting Tm of 23.68C (Table 3), most likely corresponds
to the DAG-lactone in a bilayer interface orientation, similar to
the localization of 2–4. However, a second, broader peak (t1/2
=
Figure 4. DSC thermograms. Spectra were acquired following incubation of
DAG-lactones 1–5 with DMPC multilamellar vesicles.
3.38C) in the DSC thermogram exhibiting Tm of 21.68C most
likely corresponds to a subpopulation of 5 that is embedded
deeper within the lipid bilayer, consequently interfering with
the thermal transition of the phospholipids in a similar manner
to 1. This interpretation is also consistent with the NBD
quenching result of 5 (Figure 3), which indicates a low degree
of bilayer surface disruption by this ligand.
indicative of a relatively pronounced bilayer surface localization
of 2 and 3. DAG-lactone 5, on the other hand, moderately in-
creased the fluorescence quenching rate, suggesting that bi-
layer interface interactions of this DAG-lactone were not as sig-
nificant as those of 2, 3, and 4.
Figure 3 points to different degrees of localization and per-
turbation of the bilayer headgroup region. To further probe
the interactions of the DAG-lactones with lipid bilayers and
their effects on the bilayer properties we carried out DSC
measurements (Figure 4). DSC allows examination of the coop-
erative properties of lipid bilayers, and the modification of
these properties by interactions of membrane active mole-
cules.[10]
Discussion
This study is part of a comprehensive effort in our laboratories
aimed at elucidating the relationships between membrane an-
choring of synthetic DAG-lactones and the cellular activities of
these ligands.[12,13] The data presented here indicate that the
biological properties of DAG-lactones 1–5 are indeed closely
dependent upon the modes of membrane anchoring of the
molecules, particularly the extent of their binding and localiza-
tion at the bilayer/water interface.
The DSC thermograms in Figure 4 and corresponding spec-
tral parameters in Table 3 indicate that DAG-lactones 1–5 ex-
erted different effects on the lipid bilayers. Specifically, 2, 3,
and 4 gave rise to a very similar narrow Gaussian signal for
DMPC at around 228C (Figure 4). Furthermore, the DSC param-
eters (Tm and t1/2) corresponding to the vesicles incubated with
DAG-lactones 2–4, shown in Table 3, appear close to the
values derived from the thermogram of the control vesicles
(Table 3). These results are consistent with the fluorescence
quenching data in Figure 3, which points to prevalent bilayer–
surface interactions by 2–4; localization of the DAG-lactones in
the region of the lipid headgroups is indeed hardly expected
to effect the lipid bilayer thermal transitions, which depend
primarily on the dynamics of the phospholipids’ alkyl chains.[11]
Our results demonstrate that the parent DAG-lactone 1,
which does not carry a positive charge, inserts into the hydro-
phobic core of the lipid bilayer rather than accumulating at
the membrane surface. This characteristic most likely contrib-
utes to the effective translocation capabilities of this ligand in
the biological experiments. In contrast to 1, which efficiently
inserts into the membrane bilayer, DAG-lactones 2–4 appear
specifically localized at the bilayer/water interface (Figure 3
and 4). This feature is most likely associated with the extended
n-alkyl side residues in the R2 position, which do not shield the
positive amine, thereby retaining its electrostatic attraction to
the phosphate moieties of the phospholipid headgroups. The
pronounced bilayer surface interactions of DAG-lactones 2–4 is
the probable factor accounting for the absence of PKC translo-
cation with these ligands. These results fit with the determina-
tion of ligand binding affinities for PKC. Since the role of
ligand binding to the C1 domain of PKC is to facilitate the in-
sertion of the C1 domain into the lipid bilayer, driving translo-
cation, those ligands that do not insert into and disrupt the bi-
layer structure are less effective at promoting this insertion, al-
though other factors will also play a role.
Table 3. DSC parameters obtained from thermograms.
[a]
[a]
DAG-lactone
Tm
DH[b]
DAG-lactone
Tm
DH[b]
control[c]
2
4
24.1
23.4
23.7
3480
2690
2070
1
3
5
19.4
23.6
22.6
2600
2250
2200
[a] Maximum of DSC spectrum (weighted average, 8C). [b] Enthalpy
change (calmoleÀ1). [c] No addition.
DAG-lactone 5 was shown to induce translocation of PKCd
and PKCe, but not PKCa (Figure 1, Table 1). This observation
ChemBioChem 2010, 11, 2003 – 2009
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
2007

V. E. Marquez, S. Corbalan-Garcia, R. Jelinek et al.
might be related to the intermediate status of lipid bilayer in-
teractions of this ligand exhibiting lesser bilayer interface inter-
actions compared to 2–4, and partial insertion into the lipid bi-
layer (Figures 3 and 4). In particular, the DSC data in Figure 4
suggest that two populations of 5 exist in relation to mem-
brane interactions; deeper penetration of a subpopulation of 5
is most likely responsible for inducing membrane translocation
of the PKCs. The biophysical data indicate that the branched
alkyl chain and the Z stereochemistry of the double bond of 5
are the structural determinants promoting internalization of
the side residue beyond the bilayer surface through possible
shielding of the positive charge, thus facilitating the biological
functionality of the ligand.
position of DAG-lactone 5 compared to DAG-lactones 2–4,
these positively charged residues in the C1 domain of PKCd
and PKCe, which might initially contribute to the approach of
the compounds to the plasma membrane interface in a more
favorable orientation than PKCa, thus enabling the docking of
the lactone ring in a second step.
The divergent behavior of compounds 1–5 is particularly in-
triguing. Combinatorial libraries of DAG-lactones indicate that
marked differences in biological response could be generated
from modest structural variations in the hydrophobic domains
of the DAG-lactones.[7] The probable interpretation of those
observations was that the differential effects reflected changes
in the pattern of association of the various PKC isoforms and
other C1 domains containing effector proteins with membrane
microdomains. The present findings indicate that the combina-
tion of positive charge, which should be selective for negative-
ly charged membrane regions, together with appropriate varia-
tion in the hydrophobic moieties incorporated into the struc-
ture, can likewise have profound effects on the pattern of in-
teraction with membranes. Although not examined in the
present studies, such positively charged ligands should also be
influenced in their selectivity by differences in membrane po-
tential or local pH, whether with cells or cellular organelles,
such as mitochondria or lysosomes. We conclude that incorpo-
ration of charged moieties into DAG-lactones affords a promis-
ing strategy for generation of diversity within this class of
ligands.
Additional explanation for the differential effect on plasma
membrane translocation among the different PKC isoenzymes
might be also found in the small variations in the surface
charges exhibited by the C1 domains at the rim of the DAG-
lactone binding cleft. Figure 5 depicts an example in which sol-
vent-accessible surfaces have been modeled into the 3D struc-
tures of the C1B domains of PKCa and PKCd, respectively. It is
apparent in the model shown in Figure 5 that the surface and
the rim of the top region of the PKCa molecule, which in-
cludes the DAG-lactone interacting site, are more hydrophobic
than in PKCd. Note that the C1B domain of PKCd exhibits a
positive charge corresponding to Lys256, which is conserved in
PKCe (Arg267) but not in the C1B domain of PKCa (His128;
Figure 5). Taking also into account the deeper membrane dis-
Conclusions
In this study we examined the relationship between mem-
brane interactions and biological activities of charged DAG-lac-
tones exhibiting different alkyl groups attached to the hetero-
cyclic nitrogen of an a-pyridylalkylidene chain. The experimen-
tal data indicate that the extent of bilayer insertion of the non-
branched, N-(n-alkylpyridinium) chains of the positively
charged DAG-lactones intimately and significantly effect the
translocation capabilities of the molecules. We observed that
deeper insertion of the biomimetic DAG-lactone ligands into
the membrane is essential for facilitating recognition of the
ligands by PKC—a prerequisite for inducing translocation. In
contrast, anchoring of the DAG-lactones displaying positively
charged N-(n-alkylpyridinium) linear chains on the membrane
surface interferes with their recognition and accessibility to the
PKC enzymes, eliminating translocation. Overall, our analysis
underscores the significance of membrane anchoring and in-
sertion for the biological action of the DAG-lactone derivatives.
Experimental Section
Figure 5. Solvent accessible surface of the C1B domain of: A) PKCa, and
B) PKCd in the absence of ligands. The structures were calculated by using
DSVisualizer 2.0 with a probe radius of 1.4 ꢁ. Positively and negatively
charged regions are shown in dark and light gray, respectively, while the hy-
drophobic surface is represented in white. Amino acid residues of reference
have been labeled to help orientate the molecules. The molecules on the
left correspond to a front view with critical W or Y residues labeled. The mol-
ecules on the center correspond to the back view, and those in the right
correspond to a top view showing the DAG-lactone binding cleft.
Materials: 1,2-Dimyristoyl-sn-glycero-3-phosphocholine (DMPC)
was purchased from Avanti (Alabaster, AL, USA). Sodium dithionite
(Na2O4S2) and tris(hydroxymethyl)aminomethane (TRIZMA base
buffer, C4H11NO3) were purchased from Sigma. The fluorescent dye
N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)1,2-dihexadecanoyl-sn-glycero-3
phosphoethanolamine, triethylammonium salt (NBD-PE) was pur-
chased from Molecular Probes.
2008
www.chembiochem.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 2010, 11, 2003 – 2009

Anchoring of Charged Diacylglycerol-Lactones
Synthesis and compound characterization: Detailed description
of synthesis protocols and compound characterization are provid-
ed in the Supporting Information.
Tris base buffer, pH 11) to give a final concentration of 1 mm. The
decrease in fluorescence emission was recorded for 5 min at room
temperature by using 469 nm excitation and 560 nm emission on
an Edinburgh Co. FL920 spectrofluorimeter (Edinburgh Instru-
ments, UK). The fluorescence decay curves were calculated as a
percentage of the initial fluorescence measured before the addi-
tion of dithionite.
Construction of the expression plasmids: N-terminal fusions of
rat PKC with EGFP were generated by inserting cDNAs into the
multiple cloning site of the pEGFP-N3 (Clontech Laboratories)
mammalian expression vector as described previously.[14,15] The
cDNA encoding PKC–ECFP was generated as described.[16]
Differential scanning calorimetry: The multilamellar dispersion
was achieved by dissolving DMPC in chloroform/ethanol (1:1) and
drying in vacuo to constant weight. This was followed by addition
of deionized water (final concentration 2 mm). Glass beads were
then added and the sample shaken. DSC experiments were per-
formed on a VP-DSC calorimeter (MicroCal, USA). Distilled water
served as a blank. Selected quantities of DAG-lactones were added
and heating scans were run at a rate of 18CminÀ1. Data analysis
was performed by using Microcal Origin 6.0 software.
Cell culture and transfection: MCF-7 cells were grown in Dulbec-
co’s modified Eagle’s medium with fetal calf serum (10%). For con-
focal studies the cells were plated on glass coverslips and trans-
fected after 16–24 h with Lipofectamine-2000 (Invitrogen) by fol-
lowing the instructions provided by the manufacturer. The cells
were examined under the microscope 16 h after transfection. Cov-
erslips were washed with extracellular buffer HBS (3 mL; 120 mm
NaCl, 25 mm glucose, 5.5 mm KCl, 1.8 mm CaCl2, 1 mm MgCl2,
20 mm HEPES, pH 7.2). All added substances were dissolved or di-
luted in HBS. DAG-lactones and PMA were dissolved in DMSO and
diluted to the final concentration with extracellular buffer shortly
before the experiment. During the experiment, the cells were not
exposed to DMSO concentrations higher than 1%. All experiments
were carried out at 378C by using a Leica CTI controller 3700 incu-
bator. Experiments were performed independently on at least
three different occasions; recordings were obtained from 10–16
cells in each experiment.
Acknowledgements
This research was supported in part by the Intramural Research
Program of the National Institutes of Health, Center for Cancer
Research, National Cancer Institute. Grants from the Fundaciꢀn
Mꢁdica Mutua MadrileÇa-Spain, Fundaciꢀn Sꢁneca Spain 08700/
PI/08 and MICINN-Direcciꢀn General de Investigaciꢀn Spain
(BFU2008-01010).
Confocal imaging and data analysis of EGFP variants: Cells were
washed with HBS and analyzed by using a Leica TCS SP2 confocal
system with a Nikon HCX-PL-APO 63x/1.4-0.6 NA oil immersion ob-
jective. During imaging, cells were stimulated with 40 mm DAG-lac-
tones or PMA. Confocal images of EGFP constructs were obtained
by excitation with a laser Ar/ArKr at 488 nm and emission was col-
lected at wavelengths 505–550 nm. Confocal images of ECFP were
obtained by excitation with a blue laser diode at 405 nm and emis-
sion was recorded at wavelengths 470–490 nm. The time series
were analyzed by using the quantification profile tool included in
the Leica confocal software. An individual analysis of protein trans-
location for each cell was performed by tracing a line intensity pro-
file across the cell.[17] The relative increase in plasma membrane lo-
calization (R) of the enzyme for each time point was calculated by
using the ratio R=(ImbÀIcyt)/Imb, where Imb is the fluorescence inten-
sity at the plasma membrane and Icyt is the average cytosolic fluo-
rescence intensity. The Rmax is the maximal relative increase in
plasma membrane localization of the enzyme and t1/2 is the half-
time of translocation. Both parameters were calculated as de-
scribed previously.[18] Mean values are given Æstandard error devi-
ation (SEM).
Keywords: diacylglycerol
anchoring · membranes · protein kinases · vesicles
(DAG)-lactones
·
membrane
[1] S. G. Rhee, Annu. Rev. Biochem. 2001, 70, 281–312.
[2] E. M. Griner, M. G. Kazanietz, Nat. Rev. Cancer 2007, 7, 281–294.
[3] Y. Nishizuka, Science 1992, 258, 607–614.
[4] H. J. Mackay, C. J. Twelves, Nat. Rev. Cancer 2007, 7, 554–562.
[5] S. H. Choi, T. Hyman, P. M. Blumberg, Cancer Res. 2006, 66, 7261–7269.
[6] V. E. Marquez, P. M. Blumberg, Acc. Chem. Res. 2003, 36, 434–443.
[7] J. H. Kang, S. Benzaria, D. M. Sigano, N. E. Lewin, Y. Pu, M. L. Peach, P. M.
Blumberg, V. E. Marquez, J. Med. Chem. 2006, 49, 3185–3203.
[8] M. J. Rashkin, R. M. Hughes, N. T. Calloway, M. L. Waters, J. Am. Chem.
Soc. 2004, 126, 13320–13325.
[9] J. C. McIntyre, R. G. Sleight, Biochemistry 1991, 30, 11819–11827.
[10] R. B. Gennis, Biomembranes: Molecular Structure and Function, Springer,
New York, 1989.
[11] O. G. Mouritsen, Chem. Phys. Lipids 1991, 57, 179–194.
[12] L. Philosof-Mazor, R. Volinsky, M. J. Comin, N. E. Lewin, N. Kedei, P. M.
Blumberg, V. E. Marquez, R. Jelinek, Langmuir 2008, 24, 11043–11052.
[13] O. Raifman, S. Kolusheva, M. J. Comin, N. Kedei, N. E. Lewin, P. M. Blum-
berg, V. E. Marquez, R. Jelinek, FEBS J. 2010, 277, 233–243.
[14] M. Jose Lopez-Andreo, J. C. Gomez-Fernandez, S. Corbalan-Garcia, Mol.
Biol. Cell 2003, 14, 4885–4895.
Ki measurements: Binding of DAG-lactones to PKC was deter-
mined by competition of [20-3H]phorbol 12,13-dibutyrate (PDBu)
binding as described previously.[19] The assays were carried out at
378C in the presence of phosphatidylserine (100 mgmLÀ1). Values
represent the mean ÆSEM of triplicate independent assays, with
complete dose response curves determined in each assay.
[15] C. Marin-Vicente, J. C. Gomez-Fernandez, S. Corbalan-Garcia, Mol. Biol.
Cell 2005, 16, 2848–2861.
[16] S. Sꢂnchez-Bautista, S. Corbalꢂn-Garcꢃa, A. Pꢄrez-Lara, J. C. Gꢅmez-
Fernꢂndez, Biophys. J. 2009, 96, 3638–3647.
[17] T. Meyer, E. Oancea, Methods Enzymol. 2000, 327, 500–513.
[18] M. Guerrero-Valero, C. Ferrer-Orta, J. Querol-Audi, C. Marin-Vicente, I.
Fita, J. C. Gomez-Fernandez, N. Verdaguer, S. Corbalan-Garcia, Proc. Natl.
Acad. Sci. USA 2009, 106, 6603–6607.
[19] N. E. Lewin, P. M. Blumberg, N. J. Clifton, Methods Mol. Biol. 2003, 233,
129–156.
Fluorescence quenching: NBD-PE was added to the DMPC vesicles
at a molar ratio of 1:100 (probe/total phospholipids) and the lipids
were then dried together in vacuo prior to sonication. Samples
were prepared by mixing a selected quantity of DAG-lactones with
the vesicles (30 mL) containing the fluorescent probe and Tris base
buffer (30 mL; 50 mm, pH 8.2) followed by addition of distilled
water (total volume 1.5 mL). The quenching reaction was initiated
by adding sodium dithionite from a stock solution (0.6m in 50 mm
Received: June 11, 2010
Published online on August 16, 2010
ChemBioChem 2010, 11, 2003 – 2009
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chembiochem.org
2009