Báo cáo khoa học: Signaling events mediating activation of brain ethanolamine plasmalogen hydrolysis by ceramide pot

Signaling events mediating activation of brain ethanolamine
plasmalogen hydrolysis by ceramide
Eduardo Latorre
1
, M. Pilar Collado
1
, Inmaculada Ferna
´
ndez
1
, M. Dolores Aragone
´
s
1
and R. Edgardo Catala
´
n
2
1
Departamento de Bioquı
´
mica y Biologı
´
a Molecular I, Facultad de Quı
´
micas, Universidad Complutense de Madrid, Madrid, Spain;
2
Departamento de Biologı
´
a Molecular, Centro de Biologı
´

a Molecular ‘Severo Ochoa’, Universidad Auto
´
noma de Madrid,
Madrid, Spain
Ceramide is a lipid second messenger that acts on mul-
tiple-target enzymes, some of which are involved in other
signal-transduction systems. We have previously demon-
strated that endogenous ceramide modifies the metabolism
of brain ethanolamine plasmalogens. The mechanism
involved was studied. On the basis of measurements of
breakdown products, specific inhibitor effects, and previ-
ous findings, we suggest that a plasmalogen-selective
phospholipase A
2
is the ceramide target. Arachidonate-
rich pools of the diacylphosphatidylethanolamine subclass
were also affected by ceramide, but the most affected
were plasmalogens. Concomitantly with production of
free arachidonate, increased 1-O-arachidonoyl ceramide
formation was observed. Quinacrine (phospholipase A
2
inhibitor) and 1-O-octadecyl-2-O-methyl-rac-glycerol-3-
phosphocholine (CoA-independent transacylase inhibitor)
prevented all of these ceramide-elicited effects. Therefore,
phospholipase and transacylase activities are tightly cou-
pled. Okadaic acid (phosphatase 2A inhibitor) and
PD 98059 (mitogen-activated protein kinase inhibitor)
modified basal levels of ceramide and sphingomyelinase-
induced accumulation of ceramide, respectively. Therefore,
they provided no evidence to determine whether there is a

sensitive enzyme downstream of ceramide. The evidence
shows that there are serine-dependent and thiol-dependent
enzymes downstream of ceramide generation. Further-
more, experiments with Ac-DEVD-CMK (caspase-3 speci-
fic inhibitor) have led us to conclude that caspase-3 is
downstream of ceramide in activating the brain plasmalo-
gen-selective phospholipase A
2
.
Keywords: brain ethanolamine plasmalogens; caspase-3;
ceramide; phospholipase A
2
; plasmalogen-selective phos-
pholipase A
2
.
It is well known that messengers derived from sphingolipid
and glycerolipid, and their target enzymes, establish mul-
tiple relationships leading to the formation of complicated
networks for the effective transduction of signals. In the last
few years, the regulatory role of ceramide (Cer) generated by
the sphingomyelin cycle has received increasing attention. It
is known to activate multiple serine/threonine protein
kinases and protein phosphatases [1], leading to the tissue-
specific downstream regulation of several target enzymes,
some of which are involved in other lipid signaling
pathways. In this context, we have previously reported [2]
that exogenous sphingomyelinase (EC 3.1.4.12) treatment
brought about alterations in brain ethanolamine (Etn)
plasmalogen metabolism.

The role of plasmalogens as a source of second
messengers in lipid signal-transduction systems [3–5] and
as ubiquitous endogenous antioxidants [6] has been
investigated. Plasmalogens are phospholipids characterized
by the presence of a vinyl ether substituent at the sn-1
position of the glycerol backbone. They are especially
abundant in electrically active tissues, such as brain, where
most of them are Etn-phosphoglycerides. The latter have
the propensity to facilitate membrane fusion, strongly
suggesting their involvement in synaptic transmission [3].
In addition, Etn plasmalogens have been reported to be
involved in the vulnerability to oxidative stress associated
with aging and pathological conditions [6]. Evidence is
accumulating on age-related changes in the quantities [7]
and fatty acid profile of these phospholipids [7,8]. On the
other hand, significant and selective deficiencies in brain
Etn plasmalogens have been reported at the site of
neurodegeneration in Alzheimer’s disease [9], brain
peroxisomal disorders [8] and Down’s syndrome [10]. In
some instances, the decreased Etn plasmalogen levels are
accompanied by a marked increase in the concentration of
the degradation metabolites or their derivatives, such as
PEtn [11] or prostaglandins [3]. Therefore, the evidence
suggests that several phospholipase types may be involved
Correspondence to R. E. Catala
´
n, Departamento de Biologı
´
a
Molecular, Centro de Biologı

´
a Molecular ‘Severo Ochoa’,
Universidad Auto
´
noma de Madrid, E-28049 Madrid, Spain.
Fax: + 34 91 3974870, Tel.: + 34 91 3974869,
E-mail:
Abbreviations: BSS, balanced salt solution; Cer, ceramide; C
2
-Cer,
N-acetylsphingosine; Etn, ethanolamine; ET-18-OCH
3
,1-O-octa-
decyl-2-O-methyl-rac-glycerol-3-phosphocholine; MAPK,
mitogen-activated protein kinase; PLA
2
, phospholipase A
2
;
PtdEth, phosphatidylethanolamine.
Enzymes: phospholipase A
2
(EC 3.1.1.4); sphingomyelinase
(EC 3.1.4.12); CoA-independent transacylase (EC 2.3.1.147).
(Received 2 August 2002, revised 16 October 2002,
accepted 7 November 2002)
Eur. J. Biochem. 270, 36–46 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03356.x
in the metabolism of brain Etn plasmalogens in physio-
pathological states. The existence of a plasmalogen-select-
ive phospholipase A

2
(PLA
2
, EC 3.1.1.4) which selectively,
but not exclusively, acts on 1-alk-1¢-enyl-2-acyl-sn-glycero-
3-PEtn has been reported [4,5,12]. This enzyme has been
purified from bovine brain and shown to be specific for
neural tissues and distinct from other non-neuronal
plasmalogen-specific PLA
2
enzymes and brain PLA
2
enzymes [3,4]. There is evidence that Etn plasmalogen
degradation by a PLA
2
plays an important role in
neutrophil activation by agonists [13]. From the latter
study and others [14–16], the idea has emerged that Etn
plasmalogen hydrolysis may be coupled with the formation
of acyl Cers, eicosanoids and/or platelet-activating factor.
Thus, PLA
2
in the presence of a suitable acceptor molecule
possesses a dual enzymatic function, i.e. PLA
2
and CoA-
independent transacylase, generating: (a) free arachidonate,
which can be converted into eicosanoids [13,15], and (b) an
acyl derivative, mainly the arachidonoyl derivative [16]. On
the other hand, Etn plasmalogens can be resynthesized

from the lyso-plasmenylEtn released by a CoA-independ-
ent transacylase from 1-radyl-2-arachidonoylGroPCho,
generating lyso-platelet-activating factor derivatives, which
can lead to formation of platelet-activating factor by
transacetylation [15,16]. We would like to emphasize that
all this experimental evidence has been obtained in non-
neural cell-free systems or isolated cells.
Taking into account that the following have been
reported, (a) a brain plasmalogen-selective PLA
2
[4,5], (b)
abrainPLA
2
acting on Etn phosphoglyceride with trans-
acylase activity [16], and (c) a Cer-elicited decrease in brain
Etn plasmalogen levels concomitant with 1-O-acylCer
formation [2], the aim of this study was to clarify the
mechanism by which Cer regulates brain Etn plasmalogen
metabolism. First, we investigated the type of enzymatic
activities involved and, secondly, the involvement of poten-
tial downstream Cer target enzyme(s).
Materials and methods
Materials
Staphylococcus aureus sphingomyelinase [180 UÆ(mg pro-
tein)
)1
], N-acetylsphingosine (C
2
-Cer), Cer type III (from
brain sphingomyelin containing primarily stearic and ner-

vonic acids), phenylmethanesulfonyl fluoride, quinacrine
hydrochloride, ganglioside type II (from bovine brain
containing 15% N-acetylneuraminic acid) were purchased
from Sigma, St Louis, MO, USA. Bromoenol lactone was
from Alexis Biochemicals, La
¨
ufelfingen, Switzerland.
[1-
14
C]Arachidonic acid (55 mCiÆmmol
)1
)wasfrom
American Radiolabeled Chemicals Inc., St Louis, MO,
USA. [c-
32
P]ATP (3000 CiÆmmol
)1
)wasfromNuclear
Iberica, Madrid, Spain. [2-
14
C]Ethan-1-ol-2-amine hydro-
chloride (55 mCiÆmmol
)1
) was purchased from Amersham.
Escherichia coli diacylglycerol kinase (EC 2.7.1.107),
2¢-amino-3¢-methoxyflavone (PD 98059), okadaic acid,
Ac-DEVD-chloromethylketone (Ac-DEVD-CMK) and
a-iodocetamide were from Calbiochem, San Diego,
CA, USA. 1-O-Octadecyl-2-O-methyl-rac-glycerol-3-PCho
(ET-18-OCH

3
) was from Bachem AG, Budendorf, Switzer-
land. High-performance TLC plates were obtained from
Merck, Darmstadt, Germany. All other reagents were of the
highest analytical grade available. 1-O-AcylCer standard
was synthesized as described previously [2].
Tissue preparation and incubation of slices
Experiments were carried out with male Wistar rats (180–
200 g). The animals were maintained at 22–24 °Cand
given free access to standard laboratory diet and water
ad libitum. Rat care, handling and all the experimental
procedures were in accordance with internationally accep-
ted principles concerning the care and use of laboratory
animals. The rats were killed [2], and their brains were
removed. Pial vessels and white matter were carefully
discarded, and cerebral cortex was obtained. Slices
(dimensions: 350 · 350 lm) were prepared with a MacIl-
wain tissue chopper, as previously reported [17]. They were
equilibrated in a balanced salt solution (BSS): 135 m
M
NaCl, 4.5 m
M
KCl, 1.5 m
M
CaCl
2
,0.5m
M
MgCl
2

,
5.6 m
M
glucose, 10 m
M
Hepes, pH 7.4, equilibrated with
95% O
2
/5% CO
2
for 1 h. Aliquots (300 lL) of gravity-
packed slices were transferred to glass tubes containing
BSS and then sphingomyelinase (unless otherwise indica-
ted, the final concentration was 0.38 UÆmL
)1
, as described
previously [2]) dissolved in 50 m
M
phosphate buffer,
pH 7.4, with 50% (v/v) glycerol, or C
2
-Cer dissolved in
dimethyl sulfoxide (10–100 l
M
), or diluents alone were
added and the mixture incubated for 30 min at 37 °C[2].
In one set of experiments, slices were treated with 0.1 l
M
endothelin-1 for 30 min [18]. In experiments in which
different inhibitors were tested, the slices were preincubated

in their absence or presence before the addition of
sphingomyelinase or C
2
-Cer. When the inhibitors used
were dissolved in dimethyl sulfoxide or ethanol, the final
concentration of diluent was never higher than 1%. The
incubation mixtures were continuously gassed with 95%
O
2
/5% CO
2
. The incubations were stopped by removal of
the medium and replacement with 0.38 mL BSS containing
10 m
M
EDTA and 1 mL chloroform/methanol/13
M
HCl
(100 : 100 : 1, v/v/v). Lipids were immediately extracted as
described below.
As we used an inhibitor (ET-18-OCH
3
)withlow
diffusion through slices, some experiments with homogen-
ates were performed. Homogenates of cerebral cortex were
prepared in BSS equilibrated as described above. Previous
comparative experiments showed that cerebral slices and
homogenates exhibited the same responsiveness to the
sphingomyelinase treatment [2].
Experiments with labeled precursors

In some experiments, labeled precursors were used. Slices
from 8–10 brains were preincubated in the presence of
labeled precursors: 4 lCi (0.2 lCiÆmL
)1
)[1-
14
C]arachidonic
acid [2] or 50 lCi (2 lCiÆmL
)1
)[2-
14
C]ethan-1-ol-2-amine
hydrochloride [19] at 37 °C in BSS for 120 or 30 min,
respectively. The preincubations were continuously gassed
with 95% O
2
/5% CO
2
. Then, the incubation medium was
removed, and the slices were washed three times with cold
BSS. Aliquots of slices were taken for incubation with
sphingomyelinase or C
2
-Cer; incubations were stopped as
described above.
Ó FEBS 2003 Activated plasmalogen hydrolysis by Cer (Eur. J. Biochem. 270)37
Extraction of total lipids; separation of sphingolipids
Lipids were extracted as described previously [20]. The
organic phases were dried under a N
2

atmosphere, and total
lipids were weighed and dissolved in chloroform/methanol
(2 : 1, v/v). Lipids were separated by TLC. 1-O-AcylCer
was resolved by sequential 1D TLC in: (a) ethyl ether; (b)
chloroform/methanol/acetic acid/water (25 : 15 : 4 : 1.5,
v/v/v/v), and (c) chloroform/methanol/acetic acid
(65 : 2.5 : 4, v/v/v). The first solvent system was developed
through the plate, the second reached 7 cm from the bottom
of the plate, whereas the third reached 13 cm from the
bottom. This sequential TLC also resolves the nonesterified
fatty acid fraction and the Etn phospholipid subclasses, as
stated below. To determine Cer levels, one aliquot of total
lipid was subjected to alkaline hydrolysis in 0.1
M
metha-
nolic KOH at 37 °C for 1 h to remove glycerolipids, as
previously described [2]. Cer was resolved by sequential 1D
TLC using solvent systems (a) and (c) described above, but
the former reached 3 cm from the top of the plate, whereas
the latter was developed through the plate, as described
previously [2]. Lipid standards were cochromatographed
with samples. Lipids were visualized with iodine vapor, and
the bands of 1-O-acylCer and nonesterified fatty acids were
scraped from the plates to quantitate the radioactivity
incorporated by liquid scintillation. The bands correspond-
ing to Cers were scraped from the plates and extracted with
chloroform/methanol (4 : 5, v/v) and dried under a N
2
atmosphere for subsequent quantitation.
Analysis of the subclasses of Etn phospholipids

Etn plasmalogen levels were determined as described
previously [2].
In some experiments with [
14
C]arachidonic acid as
precursor, three further subclasses of Etn phospholipids
were separated, as previously described [21]. First, total Etn
phospholipids were obtained from the total lipids by
sequential 1D TLC as described above. After extraction
with chloroform/methanol (2 : 1, v/v), the dry residue was
incubated with 40 U phospholipase C per sample for 16 h.
The resulting diacylglycerols were extracted three times with
ether/hexane (1 : 1, v/v). Once the extracts had been dried,
acetylated derivatives were prepared by incubation for 3 h
in pyridine/acetic anhydride (1 : 5, v/v). The solution was
dried and extracted twice with ether/hexane (1 : 1, v/v). The
final dried residue was fractionated by TLC using sequen-
tially: (a) hexane/ether/methanol/acetic acid (90 : 20 : 3 : 2,
v/v/v/v), and (b) toluene as solvents [22]. Phospholipids were
visualized with iodine vapor and identified from the
respective standards and reported R
f
values. Once scraped
from the plate, the radioactivity in each fraction was
measured by liquid scintillation.
Radioenzymatic determination of Cer levels
Extracted Cer was phosphorylated in the presence of
diacylglycerol kinase, as described previously [2]. Cers were
solubilized and phosphorylated in the presence of 5 lgof
theenzymeand10m

M
[c-
32
P]ATP for 10 min. After
incubation, phosphorylated derivatives of Cer were extrac-
ted, fractionated by TLC, visualized by autoradiography
using Kodak X-Omat film and quantitated by liquid-
scintillation counting. Calibration curves were constructed
using known amounts of Cer.
Radioenzymatic determination of diacylglycerol mass
Aliquots of total lipids were phosphorylated in the presence
of diacylglycerol kinase, as described previously [23].
Aliquots of total lipids were evaporated under N
2
and the
dried lipids were solubilized and phosphorylated in the
presence of 5 lg enzyme and 10 m
M
[c-
32
P]ATP for 30 min.
Then, samples were spotted on silica gel TLC plates and
developed with chloroform/methanol/acetic acid/acetone/
water (40 : 13 : 12 : 15 : 8, v/v/v/v). Spots corresponding to
phosphatidic acid were visualized by autoradiography using
Kodak X-Omat film and quantitated by liquid-scintillation
counting. Calibration curves were constructed using known
quantities of 1-stearoyl-2-arachidonoylglycerol.
Analysis of water-soluble products of hydrolysis
of Etn phospholipids

In experiments with [
14
C]Etn, the upper phases from the
lipid extraction (see above) containing the water-soluble
metabolites were analyzed by TLC [24]. The upper phases
were lyophilized and the residue was then dissolved in 50%
ethanol, and Etn, PEtn and CDP-Etn tracers were added as
carriers. Water-soluble products were separated by TLC
using methanol/0.5% NaCl/NH
4
OH (50 : 50 : 5, v/v/v) as
solvent. Bands were detected with 1% ninhydrin in ethanol.
Spots were scraped from the plate and analyzed for
radioactivity counting.
Determination of 1-
O
-alkenyl-2-lysoGro
P
Etn
radioactivity
Aliquots of total lipids from experiments performed with
[
14
C]Etn were subjected to alkaline hydrolysis and separated
using TLC. The system used was chloroform/methanol/
acetic acid (65 : 25 : 4, v/v/v). After development, spots
were visualized with ninhydrin and identified from respect-
ive standards. Spots were scraped from the plates, and their
mass determined by measurement of phosphorus content
[25]. The radioactivity incorporated was quantitated by

liquid-scintillation counting.
Statistical analysis
Student’s t test was used for paired observations. P <0.05
was considered to be significant.
Results
Sphingomyelinase and C
2
-Cer affect brain
Etn plasmalogen metabolism
We have previously reported that Etn plasmalogen meta-
bolism is specifically affected by sphingomyelinase treatment
[2]. Here we first studied the effect of different concentrations
of sphingomyelinase on Etn plasmalogen and Cer levels
(Fig. 1). At a concentration of 0.38 UÆmL
)1
, sphingomye-
linase significantly (P < 0.05) decreased Etn plasmalogens
38 E. Latorre et al.(Eur. J. Biochem. 270) Ó FEBS 2003
to 65% (Fig. 1A). Concomitantly, a significant (P < 0.05)
increase in Cer levels was observed (100% over control
value), in agreement with our previous data [2]. Higher
sphingomyelinase concentrations further increased Cer
levels, but had no further effect on Etn plasmalogen
levels (Fig. 1A). A concentration of 0.19 UÆmL
)1
sphingo-
myelinase had a slight, but not significant, effect (data not
shown).
As many effects evoked by sphingomyelinase treatment
are mimicked by short-chain cell-permeable Cer analogs, we

also tested the effect of C
2
-Cer on Etn plasmalogen levels.
The concentration range of C
2
-Cer was chosen on the basis
of previous evidence [26,27]. A concentration of 50 l
M
was
the lowest capable of decreasing Etn plasmalogen levels by
55% of the control value (P < 0.05) (Fig. 1B). Higher
concentrations did not produce further variation in Etn
plasmalogen levels.
To determine the mechanism by which sphingomyelinase
and C
2
-Cer affect Etn phosphoglyceride metabolism, we
carried out labeling studies with [1-
14
C]arachidonic acid and
[1-
14
C]Etn (Table 1). Sphingomyelinase and C
2
-Cer both
significantly (P < 0.05) reduced labeling in the plasmalo-
gen fraction but scarcely affected that in the acid-resistant
fraction. Interestingly, the most noticeable result was the
low radioactivity from [1-
14

C]arachidonic found in the
plasmalogen fraction ( 10% of the control value) when
slices were treated with sphingomyelinase.
Experiments to separate the Etn phosphoglycerides into
their three subclasses, i.e. diacyl, alkylacyl and plasmalo-
gens, were also performed. In the light of the above data
(Table 1), we used [
14
C]arachidonate as the labeled precur-
sor. These results are presented in Table 2. Both diacyl and
plasmalogen fractions exhibited significantly (P < 0.05)
reduced radioactivity ( 30% of the control value) after
treatment with sphingomyelinase or C
2
-Cer.
Lipids were extracted [20] from one aliquot of incuba-
tion medium, and
14
C radioactivity was determined. This
provides a measure of activation of secretory PLA
2
.Results
in Table 3 show that extracellular [1-
14
C]arachidonate
release was not affected, but the cell-associated
14
C
Fig. 1. Dose–response relationship of sphingomyelinase-induced and
C

2
-ceramide-induced changes in brain Etn plasmalogen levels. (A)
Cerebral cortex slices were exposed to sphingomyelinase (SMase) for
30 min, and levels of Etn plasmalogens (PlsEtn; left axis; filled bars)
and ceramide (right axis; striped bars) were measured. (B) Cerebral
cortex slices were exposed to C
2
-ceramide for 30 min, and Etn plas-
malogens were measured. Data represent mean ± SE and are from
two experiments performed in triplicate. Values significantly different
from their respective controls are indicated: *P < 0.05.
Table 1. Variations in [
14
C]arachidonic acid-labeled and [
14
C]Etn-labeled Etn phospholipids evoked by sphingomyelinase and C
2
-Cer. Slices were
labeled with 0.2 lCiÆmL
)1
[
14
C]arachidonic acid for 120 min or 2 lCiÆmL
)1
[
14
C]Etn for 30 min. After removal of the labeled precursor, slices were
exposedto0.38UÆmL
)1
sphingomyelinase or two different C

2
-Cer concentrations for 30 min. Total lipids were split into two aliquots: one was
untreated, and the other was exposed to HCl fumes. Radioactivity in plasmalogen was obtained by subtracting the acid-resistant fraction from that
obtained in the total Etn phospholipids. Data are expressed as the percentage of radioactivity incorporated in each fraction with respect to that
incorporated in total lipid. They represent mean ± SD from one representative experiment of two experiments performed in quintuplicate.
ND, Not determined.
Treatment
Radioactivity incorporated into Etn phospholipids
[
14
C]Arachidonic acid [
14
C]Etn
Acid-resistant
fraction
Plasmalogen
fraction
Acid-resistant
fraction
Plasmalogen
fraction
Control 1.78 ± 0.32 0.62 ± 0.09 70.8 ± 1.7 9.20 ± 2.71
Sphingomyelinase 2.06 ± 0.10 0.06 ± 0.02* 67.3 ± 2.1 5.32 ± 1.50*
C
2
-Cer
100 l
M
1.65 ± 0.40 0.42 ± 0.04* 76.3 ± 3.1 4.62 ± 0.93*
50 l

M
1.62 ± 0.52 0.34 ± 0.06* ND ND
* P < 0.05 compared with respective control.
Ó FEBS 2003 Activated plasmalogen hydrolysis by Cer (Eur. J. Biochem. 270)39
radioactivity had increased by nearly 30% after treatment
with sphingomyelinase or C
2
-Cer.
Identification of the phospholipase type involved
in the Cer-elicited decrease in Etn plasmalogen levels
To determine the type of enzymatic activity involved, we
next measured the levels of the breakdown products
released by phospholipase type D or C, i.e. Etn, PEtn,
and diacylglycerol. In addition, an intermediary of their
biosynthesis, CDP-Etn, was measured (Fig. 2A,B). It is
evident that no significant alterations were elicited by
sphingomyelinase treatment. The involvement of a PLA
2
was tested by examining potential alterations in levels and
[
14
C]Etn radioactivity in the lyso form of Etn plasmalogens
evoked by sphingomyelinase or C
2
-Cer (Fig. 2C,D, respect-
ively). Surprisingly, no significant changes were found in the
presence of 0.38 UÆmL
)1
sphingomyelinase (Fig. 2C). How-
ever, in a dose–response study with C

2
-Cer as agonist, a
significant increase in the level of and radioactivity in
lyso-Etn plasmalogens could only be observed in the
presence of 100 l
M
C
2
-Cer (Fig. 2D).
Before definitely establishing whether a PLA
2
was the Cer
target, we next examined the effect of the widely used
nonspecific PLA
2
inhibitor quinacrine [4] on the sphingo-
myelinase-elicited effect (Fig. 3). Quinacrine alone (25 l
M
)
did not alter the
14
C radioactivity from [1-
14
C]arachidonic
acid in the Etn plasmalogens, but, in the presence of
sphingomyelinase, it not only prevented the decrease caused
by sphingomyelinase, but also evoked a significant
(P < 0.05) increase in the
14
C radioactivity found in Etn

plasmalogens.
This led us to hypothesize that the target enzyme for Cer
action may be the 39 kDa plasmalogen-selective PLA
2
described and characterized previously [3,5,12]. The enzyme
is specifically and markedly inhibited by sialic acid, glucos-
aminoglucans, gangliosides and sialoglycoproteins [3,5]. In
contrast, the brain 110 kDa cytosolic PLA
2
, acting prefer-
entially on PtdEtn, has been reported to be much less
sensitive to these inhibitory effects [3,5,12]. These differences
in behavior prompted us to test the effect of sphingomye-
linase on slices pretreated with a brain ganglioside mixture.
The ganglioside mixture did not itself evoke significant
variation in either
14
C radioactivity or levels of PtdEtn, but
did prevent the decrease in radioactivity in, and levels of,
Etn plasmalogens caused by sphingomyelinase (Fig. 3A,B).
We also tested the effect of bromoenol lactone, a specific
and potent inhibitor of myocardial Ca
2+
-independent
plasmalogen-specific PLA
2
[28] devoid of effect on the
brain plasmalogen-selective PLA
2
[4,9]. Pretreatment with

bromoenol lactone did not block the effect of sphingo-
myelinase on Etn plasmalogen levels (Fig. 3B). Therefore,
our results are in agreement with those reported for the
brain enzyme [4,9].
A first attempt was made to establish whether the
sphingomyelinase-sensitive PLA
2
acting on Etn plasmalo-
gens also shows CoA-independent transacylase activity. For
this, we used ET-18-OCH
3
, a specific inhibitor [29]. ET-18-
OCH
3
(25 l
M
) itself did not modify either
14
C radioactivity
in, or levels of, Etn plasmalogens (Fig. 3A,B, respectively).
However, when ET-18-OCH
3
was added before sphingo-
myelinase, the effect of sphingomyelinase on the Etn
plasmalogens was prevented (Fig. 3A,B). In agreement
with our previous report [2], we first observed a significant
(P < 0.05) sphingomyelinase-elicited increased production
of 1-O-[1-
14
C]acylCer (Table 4), which can be used as an

index of transacylase activity [16]. It is also evident that an
increase in the level of
14
C radioactivity in the nonesterified
fatty acid fraction was concomitantly evoked by sphingo-
myelinase. Interestingly, the ganglioside mixture (0.26 gÆL
)1
)
and ET-18-OCH
3
(25 l
M
) both completely prevented both
these sphingomyelinase-evoked effects.
Mechanism by which Cer decreases Etn plasmalogens
levels
Cer has been reported to activate okadaic acid-sensitive
protein phosphatase 2A. To test whether this protein
phosphatase is involved in the Cer effect, we treated brain
Table 2. Variations in [
14
C]arachidonic acid-labeled Etn phospholipid
subclasses evoked by sphingomyelinase and C
2
-Cer. Slices were labeled
with 0.2 lCiÆmL
)1
[
14
C]arachidonic acid for 120 min. After removal of

the labeled precursor, slices were exposed to 0.38 UÆmL
)1
sphingo-
myelinase or 100 l
M
C
2
-Cer for 30 min. Total Etn phospholipids were
hydrolyzed with phospholipase C. The resulting diacylglycerols were
extracted and the acetylated derivatives were prepared. After their
fractionation by TLC, the radioactivity in each was measured. Data
are expressed as radioactivity incorporated (d.p.m.) in each subclass
per mg of total lipids. They represent mean ± SD from one repre-
sentative experiment of two experiments performed in triplicate.
Treatment
Radioactivity incorporated into
Etn phospholipid subclasses
Alkenylacyl Alkylacyl Diacyl
Control 54.8 ± 7.2 49.3 ± 1.7 52.5 ± 7.2
Sphingomyelinase 38.8 ± 3.2* 43.8 ± 3.6 39.4 ± 5.1*
C
2
-Cer 33.1 ± 4.1* 48.9 ± 0.8 30.1 ± 4.1*
* P < 0.05 compared with their respective controls.
Table 3. Variations in the extracellular and cell-associated radioactivity
from [
14
C]arachidonic acid evoked by sphingomyelinase and C
2
-Cer.

Slices were labeled with 0.2 lCiÆmL
)1
[
14
C]arachidonic acid (AA) for
120 min. After removal of the labeled precursor, slices were exposed to
0.38 UÆmL
)1
sphingomyelinase or 100 l
M
C
2
-Cer for 30 min. Aliquots
(50 lL) from the incubation medium were taken for radioactivity
measurement. Tissue total lipids were extracted, dissolved, and aliqu-
ots (10 lL) were taken for radioactivity measurement. Extracellular
arachidonic acid is expressed as d.p.m. per aliquot and cell-associated
arachidonic acid as d.p.m. per mg total lipids. Data represent
mean ± SD from one representative experiment of two experiments
performed in triplicate.
Treatment
Radioactivity incorporated
Extracellular AA Cell-associated AA
Control 1030 ± 181 24266 ± 5488
Sphingomyelinase 892 ± 158 31061 ± 2575*
C
2
-Cer 942 ± 206 29611 ± 2455*
* P < 0.05 compared with their respective controls.
40 E. Latorre et al.(Eur. J. Biochem. 270) Ó FEBS 2003

slices with okadaic acid (2.5 and 25 n
M
) before sphingo-
myelinase treatment (Fig. 4). Okadaic acid alone produced
no change in Etn plasmalogen levels (Fig. 4A) but did
prevent the effect of sphingomyelinase treatment on Etn
plasmalogen levels.
Data on Cer levels are shown in Fig. 4B. Sphingomye-
linase increased the level of endogenous Cer by nearly 100%
(P < 0.05). Okadaic acid by itself did not alter Cer levels.
However, when slices were pretreated with okadaic acid, the
sphingomyelinase-elicited increase was prevented. There-
fore, okadaic acid was acting as a modulator of Cer
metabolism but not of the Cer-evoked effect.
On the other hand, Cer has been reported to induce
mitogen-activated protein kinase (MAPK) activity, which in
turn phosphorylates and activates cytosolic PLA
2
[1].
PD 98059 has been widely used as a specific inhibitor to
study whether p42/p44 MAPK is downstream of Cer
generation. Experiments with PD 98059 were therefore
performed (Fig. 5). At concentrations ranging from 10–
100 l
M
, PD 98059 significantly (P < 0.05) increased Cer
levels in a dose-dependent manner (Fig. 5B). Concomit-
antly, Etn plasmalogen levels decreased by about 75%,
in a dose-independent manner (Fig. 5A). Unexpectedly,
PD 98059 was able to prevent the sphingomyelinase-elicited

increase in Cer levels (Fig. 5B) and partially reverse the Cer-
evoked reduction in Etn plasmalogen levels (Fig. 5A).
It has been shown that specific protease activation is a
pivotal element in Cer-regulated processes. Thus, Cer acts
downstream of caspase-8 but upstream of caspase-3 [27]. In
addition, a serine proteolytic enzyme is also a Cer target
[30]. Therefore, iodoacetamide (as a thiol-specific inhibitor)
and phenylmethanesulfonyl fluoride (as a blocking agent of
serine enzymes) were tested (Fig. 6). Neither iodoacetamide
nor phenylmethanesulfonyl fluoride by themselves affected
basal Cer (Fig. 6B) or Etn plasmalogen (Fig. 6A) levels.
However, both inhibitors were able to prevent the Cer effect
on Etn plasmalogen levels (Fig. 6A) without modifying the
enhanced Cer levels (Fig. 6B).
In view of these results, we next explored whether
caspase-3 is involved in the regulation of plasmalogen-
selective PLA
2
. Experiments with the cell-permeable
caspase-3-specific tetrapeptide inhibitor Ac-DEVD-CMK
were performed. The Ac-DEVD-CMK concentration used
has been shown to inhibit apoptosis induced by 30 l
M
C
2
-Cer and caspase-3 activity [27]. The results obtained are
shown in Table 5. The caspase-3 inhibitor by itself did not
produce any effect, but partially prevented the sphingo-
myelinase-elicited decrease in Etn plasmalogen levels with-
out affecting Cer levels.

Etn plasmalogen hydrolysis can also be elicited
by endogenous agonists
We have previously reported that the neuropeptide endo-
thelin-1 is able to evoke Cer production in cerebral cortex
[18]. Therefore, we next hypothesized that Etn plasmalogen
hydrolysis may occur concomitantly with endogenous Cer
production evoked by a natural agonist. Table 6 shows that
treatment with 0.1 l
M
endothelin-1 for 30 min (conditions
under which maximum Cer production is evoked by
Fig. 2. Variations in breakdown products of Etn phospholipid evoked by sphingomyelinase (SMase) and C
2
-ceramide. (A) Sphingomyelinase-evoked
variations in [
14
C]Etn-labeled water-soluble metabolites; (B) sphingomyelinase-evoked variations in levels of total diacylglycerols; (C) sphingo-
myelinase-evoked variations in [
14
C]Etn radioactivity (left axis; open bars) and in levels of 1-O-alkenyl-2-lyso-GroPEtn (right axis; filled bars); (D)
C
2
-ceramide-evoked variation in [
14
C]Etn radioactivity (left axis; open bars) and in levels of 1-O-alkenyl-2-lyso-GroPEtn (right axis; filled bars).
Cerebral cortex slices were prelabeled with 2 lCiÆmL
)1
[
14
C]Etn for 30 min (A, C and D) and then exposed to either 0.38 UÆmL

)1
sphingomyelinase
or different C
2
-ceramide concentrations for 30 min. Levels of, and the radioactivity in, the metabolites were determined. Data represent mean ± SE
from two separate experiments performed in quintuplicate. Values significantly different from their respective controls are indicated: *P<0.05.
Ó FEBS 2003 Activated plasmalogen hydrolysis by Cer (Eur. J. Biochem. 270)41
endothelin-1) resulted in a significant (P < 0.05) decrease
( 35%) in Etn plasmalogen levels, concomitantly with an
increase of 60% in the Cer level.
Discussion
Involvement of the brain plasmalogen-selective PLA
2
in the Cer-elicited decrease in Etn plasmalogen levels
We have previously shown that sphingomyelinase decreases
the levels of brain Etn plasmalogens [2]. To rule out the
possibility that Etn plasmalogens are directly hydrolyzed by
sphingomyelinase, experiments with C
2
-Cer were per-
formed. The effect of sphingomyelinase on Etn plasmalogen
levels was mimicked by C
2
-Cer. Although several differen-
tial effects of sphingomyelinase and Cer analogs have been
described [26], we conclude that the decrease in Etn
plasmalogen levels is a response, at least in part, to
endogenous Cer accumulation.
The complete prevention of the Cer effect caused by
quinacrine and gangliosides, combined with the lack of

effect exhibited by bromoenol lactone, led us to think that
the enzyme involved is the 39 kDa plasmalogen-selective
PLA
2
[4,9,12]. That other Etn phospholipids besides
plasmalogens are affected is consistent with the specificity
shown by the brain 39 kDa plasmalogen-selective PLA
2
[12]. In addition, we also observed that there was no loss of
sphingomyelinase-elicited extracellular arachidonate or its
derivatives, which precludes the involvement of a secretory
PLA
2
.
Two additional findings are noteworthy. First, the arachi-
donate-rich pool of Etn plasmalogens is appreciably affected
by Cer (Tables 1 and 2). The docosahexaenoate-rich pool of
Etn plasmalogens is the other major pool of brain Etn
plasmalogens [7]andthereforeit wouldbeinterestingtostudy
it further. Secondly, plasmalogen hydrolysis by PLA
2
is
coupled with CoA-independenttransacylaseactivity, as these
processes are blocked in parallel by inhibitors of each
(gangliosides and ET-18-OCH
3
). This coupling has also been
observed in other PLA
2
enzymes acting on alkenylacylglyc-

erophospholipids of Etn, such as the 14 kDa PLA
2
present in
monocytes [31], or on diacylglycerophospholipids of Etn,
such as the 40 kDa PLA
2
of brain [16]. In fact, the latter
enzyme is a single polypeptide chain with a molecular mass of
 40 kDa, similar to that of the plasmalogen-selective PLA
2
described previously [12].
An interesting picture begins to emerge from the present
evidence. However, it is necessary to consider several facts.
Fig. 3. Effect of quinacrine (Q), ganglioside mixture (G), ET-18-OCH
3
(E) and bromoenol lactone (B) on sphingomyelinase (SMase)-induced
alterations in brain Etn plasmalogens (PlsEtn). (A) Radioactivity from
[
14
C]arachidonic acid in Etn plasmalogens is expressed as the per-
centage of radioactivity incorporated into these phospholipids with
respect to that incorporated into total lipids. (B) Levels of Etn plas-
malogens. Cerebral cortex slices were labeled with 0.2 lCiÆmL
)1
[
14
C]arachidonic acid for 120 min. After removal of the labeled pre-
cursor, slices were incubated with 250 l
M
quinacrine for 25 min,

0.26 gÆL
)1
ganglioside mixture for 2 min, or 10 l
M
bromoenol lactone
for 10 min. Cerebral cortex homogenates were labeled as described
above and exposed to 25 l
M
ET-18-OCH
3
for 2 min. They were then
treated with 0.38 UÆmL
)1
sphingomyelinase for 30 min. Respective
controls were performed by incubating slices or homogenates in the
presence of the respective solvents. Radioactivity and/or levels of Etn
plasmalogens were measured. Data represent mean ± SE and are
from two experiments performed in quintuplicate. Values significantly
different from the control are indicated: *P <0.05.
Table 4. Effect of plasmalogen-selective PLA
2
and CoA-independent
transacylase inhibitors on the formation of 1-O-acylCer and release of
free arachidonic acid evoked by sphingomyelinase. Slices were labeled
with 0.2 lCiÆmL
)1
[
14
C]arachidonic acid for 120 min. After removal of
the labeled precursor, slices were incubated with 0.26 gÆL

)1
ganglioside
mixture (G) for 2 min or with 25 l
M
ET-18-OCH
3
(E) for 2 min.
Then, 0.38 UÆmL
)1
sphingomyelinase was added for 30 min. Total
lipids were fractionated by TLC, and the radioactivity in 1-O-acylCer
and nonesterified fatty acid fractions was measured. Data are
expressed as radioactivity incorporated (d.p.m.) in each fraction per
mg total lipids. Data represent mean ± SD from one representative
experiment of two experiments performed in quintuplicate.
Treatment
Radioactivity incorporated
1-O-AcylCer Nonesterified fatty acid
Control 516 ± 107 9155 ± 609
Sphingomyelinase 732 ± 139* 13134 ± 2918*
G 597 ± 85 9209 ± 1609
G + sphingomyelinase 383 ± 200 6354 ± 503
E 686 ± 148 9884 ± 1177
E + sphingomyelinase 472 ± 95 8364 ± 807
* P < 0.05 compared with their respective controls.
42 E. Latorre et al.(Eur. J. Biochem. 270) Ó FEBS 2003
First, in the plasma membrane of eukaryotic cells, the
Etn-containing phospholipids reside in the inner leaflet
whereas sphingomyelin is located in the outer leaflet. There
is evidence that, during the early stages of apoptosis, this

asymmetric distribution is lost, resulting in exposure of
PtdEtn on the cell surface [32]. Thus, the potential activation
of a membrane-associated neutral sphingomyelinase by
apoptosis inducers would generate Cers that initiate a
cascade of events, including the hydrolysis of Etn plasma-
logens, which may be suitably positioned by a previous
traslocation event.
Secondly, Cer has been shown to be involved in oxidative
stress through the production of mitochondrial oxygen-free
radicals [33]. On the other hand, Etn plasmalogens are
antioxidant molecules that protect cells from oxidative stress
[6]. Cer-elicited hydrolysis of Etn plasmalogens could
produce an increase in susceptibility to oxidative agents,
leading to apoptosis. Thus, our data may indicate a new role
for Cer in apoptosis.
Caspase-3 is involved in the Cer-elicited decrease
of Etn plasmalogen levels
It has been previously reported that Cer does not affect
purified plasmalogen-selective PLA
2
[12]. Therefore, our
next experiments were designed to identify enzyme(s)
downstream of Cer capable of regulating plasmalogen PLA
2
.
We unexpectedly found that okadaic acid and PD 98059,
used as inhibitors of protein phosphatase and MAPK,
respectively, were able to modify the sphingomyelinase-
enhanced or basal endogenous Cer levels. Consistent with
this, complex modulation of Cer levels evoked by okadaic

acid has been reported [34]. Although we cannot rule out the
possibility that okadaic acid itself affects sphingomyelinase,
it is very likely that the metabolic fate of Cer is affected. As
the okadaic acid effect is evoked by concentrations as low as
2.5 n
M
, a protein phosphatase 2A may regulate Cer
metabolism.
Fig. 4. Effect of okadaic acid on basal and sphingomyelinase-altered Etn
plasmalogen levels and ceramide accumulation. (A) Etn plasmalogen
(PlsEtn) levels; (B) ceramide levels. Cerebral cortex slices were incu-
bated in the absence or presence of 2.5 or 25 n
M
okadaic acid (OKA)
for 10 min and then exposed to 0.38 UÆmL
)1
sphingomyelinase
(SMase) for 30 min. Etn plasmalogen and ceramide levels were
obtained from the same tissue sample. Data represent mean ± SE
from three experiments performed in quintuplicate. Values signifi-
cantly different from the control are indicated: *P <0.05.
Fig. 5. PD 98059-evoked effect on basal and sphingomyelinase-altered
Etn plasmalogens levels and ceramide accumulation. (A) Etn plasma-
logen (PlsEtn) levels; (B) ceramide levels. Cerebral cortical slices were
incubated in the absence or presence of several concentrations of
PD 98059 for 30 min and then exposed to 0.38 UÆmL
)1
sphingo-
myelinase (SMase) for 30 min. Etn plasmalogen and ceramide levels
were obtained from the same tissue sample. Data represent

mean ± SE from three experiments performed in quintuplicate. Val-
ues significantly different from control are indicated: *P <0.05.
Ó FEBS 2003 Activated plasmalogen hydrolysis by Cer (Eur. J. Biochem. 270)43
The effects of PD 98059 are even more complex, as the
inhibitor increased the basal levels of Cer but prevented the
sphingomyelinase-induced increase. This may be explained
in terms of some MAPK members being upstream and/or
downstream of Cer generation. As the decrease in Etn
plasmalogens was still observed in its presence, it is likely
that a Cer metabolite is also regulating the plasmalogen-
selective PLA
2
. Studies are currently being carried out to
clarify this point.
Despite the complex mechanism of action of PD 98059, it
is clear that the Cer-elicited activation of plasmalogen-
selective PLA
2
is not mediated by the activation of p42/p44
MAPK. This is a characteristic that is not shared by the
cytosolic PLA
2
from many cell types, including that from
rat cerebral cortex [35]. Nevertheless, the possibility of the
involvement of p38 MAPK remains open, as it is also a Cer
target [1], and the regulation of cytosolic PLA
2
by this
MAPK subfamily has been reported in other biological
systems [36].

The potential involvement of some proteolytic step in the
regulation of plasmalogen PLA
2
by Cer was also tested.
First, we studied the action of thiol protease and serine
protease inhibitors. Neither class of inhibitors was able to
modify both basal and sphingomyelinase-enhanced Cer
levels, but they did prevent the Cer-elicited lowering effect
on Etn plasmalogen levels. One possible explanation is that
there are proteases (or other enzymes) that contain serine or
cysteine in their active center downstream of Cer that
mediate the activation of the plasmalogen PLA
2
.This
hypothesis is supported by evidence on the regulation of
other types of PLA
2
by proteolytic cleavage phenomena
[37]. Alternatively, it is possible that serine and cysteine
residues are functionally important and/or are present in the
catalytic site of the plasmalogen-selective PLA
2
. Consistent
with this, it is well known that many esterases, including
PLA
2
, are sensitive to the action of iodoacetate and
phenylmethanesulfonyl fluoride. Furthermore, in studies
with the purified plasmalogen-selective PLA
2

, preliminary
evidence on its sensitivity to iodoacetate has been reported
[3]. In contrast, it has been reported that any serine residue is
essential for the transacylase reaction of the 40 kDa brain
Fig. 6. Preventive effect of iodoacetamide and phenylmethanesulfonyl
fluoride on the sphingomyelinase-elicited alterations in Etn plasmalogen
levels and ceramide accumulation. (A) Etn plasmalogen (PlsEtn) levels;
(B) ceramide levels. Cerebral cortex slices were pretreated with 10 m
M
iodoacetamide (I) or 2 m
M
phenylmethanesulfonyl fluoride (PMSF)
for 60 min and then exposed to 0.38 UÆmL
)1
sphingomyelinase
(SMase) for 30 min. Etn plasmalogen and ceramide levels were
obtained from the same sample. Data represent mean ± SE from two
experiments performed in quintuplicate. Values significantly different
from the control are indicated: *P < 0.05.
Table 5. Effect of caspase-3 inhibitor on Etn plasmalogen hydrolysis and
Cer accumulation evoked by sphingomyelinase. Cerebral cortex slices
were preincubated with 50 l
M
Ac-DEVD-CMK for 60 min, then
treated with 0.38 UÆmL
)1
sphingomyelinase for 30 min. Total lipids
were split into three aliquots: one was untreated, another was exposed
to HCl fumes, and the other was hydrolyzed by alkali. They were
fractionated by TLC, and the levels of Etn plasmalogens and Cer were

measured. Data are expressed as nmol each fraction per mg total lipids.
They represent mean ± SD from one representative experiment of
two experiments performed in quintuplicate.
Treatment
Lipid fraction level
Etn plasmalogens Cer
Control 179.6 ± 26.3 96.4 ± 19.0
Sphingomyelinase 64.7 ± 10.3* 202.4 ± 37.2*
Ac-DEVD-CMK 213.7 ± 44.5 145.7 ± 28.0
Ac-DEVD-CMK+
sphingomyelinase
132.9 ± 10.3* 227.3 ± 21.9*
* P < 0.05 compared with their respective controls.
Table 6. Effect of endothelin-1 (ET-1) on brain Etn plasmalogen and
Cer levels. Slices were incubated with 0.1 l
M
endothelin-1 for 30 min.
After total lipid extraction, Etn plasmalogen and Cer levels were
measured. Data are expressed as nmol per mg total lipids. They are
mean ± SD from one experiment performed in triplicate.
Treatment
Lipid fraction level
Etn plasmalogens Cer
Control 125.1 ± 31.1 96.4 ± 15.0
ET-1 77.5 ± 15.5* 154.2 ± 23.0*
* Significantly different (P < 0.05) from the control value.
44 E. Latorre et al.(Eur. J. Biochem. 270) Ó FEBS 2003
transacylase. These contradictory observations remain to be
clarified.
Of more interest was the fact that the caspase-3-like

protease-specific inhibitor Ac-DEVD-CMK could partially
abolish Cer-elicited Etn plasmalogen hydrolysis without
altering sphingomyelinase-elicited Cer accumulation. Sev-
eral PLA
2
enzymes are substrates for caspase-3, but,
depending on the type of PLA
2
, this cleavage leads to their
inactivation (in the case of the cytosolic PLA
2
,typeIV,
without arachidonate-phospholipid remodeling activity) or
activation (in the case of the Ca
2+
-independent PLA
2
,type
VI, with arachidonate-phospholipid remodeling activity)
[37]. As the brain plasmalogen-selective PLA
2
is Ca
2+
-
independent [4], and has been shown here to have CoA-
independent transacylase activity, its potential activation by
caspase-3 is consistent with the available evidence. Further-
more, it has been suggested that activation of hydrolysis of
Etn phospholipids by PLA
2

results from the covalent
modification of the enzyme [38].
In an attempt to establish the potential pathophysiolo-
gical significance of the present findings, we made a
preliminary study to determine whether a natural
Cer-generating agonist, such as the neuropeptide endothe-
lin-1 [18], can modify Etn plasmalogen levels in brain
tissue. The positive evidence obtained suggests that the
present findings can be extrapolated to in vivo conditions.
Further studies in this field are currently being performed
in our laboratory.
We may tentatively conclude that the findings reported
here are relevant to the knowledge of some processes in
Alzheimer’s disease and cerebral ischemia, despite the fact
that some aspects still remain unclear.
Activated caspase-3 has been in situ-immunodetected in
only a small subpopulation of hippocampal neurons, but
not in the cortex in patients with Alzheimer’s disease [39]. In
addition, amyloid beta peptide can activate caspase-3 and
induce neuronal apoptosis in vitro [39].Furthermore,ithas
been suggested [9] that stimulation of the Ca
2+
-independent
plasmalogen-selective PLA
2
may account for the decreased
levels of Etn plasmalogens found in the affected regions of
the brain in Alzheimer’s disease, such as the cerebral cortex.
It is feasible that an early and reversible activation of
caspase-3 by endogenous Cer may be sufficient to produce

an irreversible loss of Etn plasmalogens.
The picture in other pathological states is clearer. Brain
sections from patients with neuropathological evidence of
apoptosis secondary to stroke, seizure or trauma exhibit
neuronal-activated caspase-3 immunoreactivity [39]. Acti-
vation of a brain Ca
2+
-independent PLA
2
acting on PtdEtn
[38] and a decrease in Etn plasmalogen levels [6,40] have
been reported to occur in ischemia.
It is noteworthy that other Ca
2+
-independent PLA
2
enzymes acting on plasmalogens from non-neural sources
have been described [3]. Whether these are also regulated by
Cer remains an open question; the answer may provide
evidence for cross-talk phenomena between the sphingo-
myelin cycle and PLA
2
-mediated arachidonate metabolism
[30].
In summary, our data show, for the first time, that brain
Ethanolamine plasmalogen hydrolysis is regulated by the
endogenous level of Cer, and a caspase-3-like protease is a
downstream Cer effector.
Acknowledgements
We are indebted to Mrs M. V. Mora Gil and Mrs Belinda Benhamu´ .

This work was funded by grants from the DGICYT and the Fundacio
´
n
‘Ramo
´
n Areces’.
References
1. Liu, G.L., Kleine, L. & Hebert, R.L. (1999) Advances in the signal
transduction of ceramide and related sphingolipids. Crit. Rev.
Clin. Lab. Sci. 36, 511–573.
2. Latorre, E., Aragone
´
s, M.D., Ferna
´
ndez,I.&Catala
´
n, R.E.
(1999) Platelet-activating factor modulates brain sphingomyelin
metabolism. Eur. J. Biochem. 261, 1–9.
3. Farooqui, A.A., Yang, H C. & Horrocks, L.A. (1995) Plasma-
logens, phospholipases A
2
and signal transduction. Brain Res.
Rev. 21, 152–161.
4.Farooqui,A.A.,Yang,H C.,Rosenberger,T.A.&Horrocks,
L.A. (1997) Phospholipase A
2
and its role in brain tissue.
J. Neurochem. 69, 889–901.
5. Farooqui, A.A. & Horrocks, L.A. (2001) Plasmalogens, phos-

pholipases A
2
and docosahexaenoic acid turnover in brain tissue.
J. Mol. Neurosci. 16, 263–272.
6. Brosche, T. & Platt, D. (1998) The biological significance of
plasmalogens in defense against oxidative damage. Exp. Gerontol.
33, 363–369.
7. Favrelie
`
re, S., Stadelmann-Ingrand, S., Huguet, F., De Javel, D.,
Piriou, A., Tallineau, C. & Durand, G. (2000) Age-related changes
in ethanolamine glycerolipid fatty acid levels in rat frontal cortex
and hippocampus. Neurobiol. Aging 21, 653–660.
8. Martı
´
nez, M. & Mougan, I. (1999) Fatty acid composition of
brain glycerophospholipids in peroxisomal disorders. Lipids 34,
733–740.
9. Farooqui, A.A., Rapoport, S.I. & Horrocks, L.A. (1997b) Mem-
brane phospholipid alterations in Alzheimer’s disease: deficiency
of ethanolamine plasmalogens. Neurochem. Res. 22, 523–527.
10. Murphy, E.J., Schapiro, M.B., Rapoport, S.I. & Shetty, H.U.
(2000) Phospholipid composition and levels are altered in Down
syndrome brain. Brain Res. 867, 9–18.
11. Uchiyama-Tsuyuki, Y., Araki, H., Yae, T. & Otomo, S. (1994)
Changes in the extracellular concentration of aminoacids in the rat
striatum during transient focal cerebral ischemia. J. Neurochem.
62, 1074–1078.
12. Yang, H.C., Farooqui, A.A. & Horrocks, L.A. (1994) Effects of
glycosaminoglycans and glycosphingolipids on cytosolic phos-

pholipases A
2
from bovine brain. Biochem. J. 299, 91–95.
13. Nieto, M.L., Venable, M.E., Bauldry, S.A., Greene, D.G.,
Kennedy, M., Bass, D.A. & Wykle, R.L. (1991) Evidence that
hydrolysis of ethanolamine plasmalogens triggers synthesis of
platelet-activating factor via a transacylation reaction. J. Biol.
Chem. 266, 18699–18706.
14. Lee,T.,Ou,M.,Shinozaki,K.,Malone,B.&Snyder,F.(1996)
Biosynthesis of N-acetylsphingosine by platelet-activating factor:
sphingosine CoA-independent transacetylase in HL-60 cells.
J. Biol. Chem. 271, 209–217.
15. Svetlov,S.I.,Liu,H.,Chao,W.&Olson,M.S.(1997)Regulation
of platelet-activating factor (PAF) biosynthesis via coenzyme
A-independent transacylase in the macrophage cell line IC-21
stimulated with lipopolysaccharide. Biochim. Biophys. Acta 1346,
120–130.
16. Abe, A. & Shayman, J.A. (1998) Purification and characterization
of 1-O-Acylceramide synthase, a novel phospholipase A
2
with
transacylase activity. J. Biol. Chem. 273, 8467–8474.
17. Catala
´
n, R.E., Martı
´
nez, A.M., Aragone
´
s, M.D. & Miguel, B.G.
(1991) Selective time-dependent effects of insulin on brain phos-

phoinositide metabolism. Regul. Peptides 32, 289–296.
Ó FEBS 2003 Activated plasmalogen hydrolysis by Cer (Eur. J. Biochem. 270)45
18. Catala
´
n, R.E., Aragone
´
s, M.D., Martı
´
nez, A.M. & Ferna
´
ndez, I.
(1996) Involvement of sphingolipids in the endothelin-1 signal
transduction mechanism in rat brain. Neurosci. Lett. 220, 121–124.
19. Mozzi, R., Andreoli, V. & Buratta, S. (1997) Synthesis of etha-
nolamine phosphoglycerides in rat cerebral cortex subjected in vitro
to experimental hypoxia with and without hypocapnia. Neu-
rochem. Res. 22, 1223–1229.
20. Bligh, E.G. & Dyer, W.J. (1959) A rapid method of total lipid
extraction and purification. Can. J. Biochem. Physiol. 39, 911–917.
21. Surette, M.E., Winkler, J.D., Fonteh, A.N. & Chilton, F.H. (1996)
Relationship between arachidonate-phospholipid remodeling and
apoptosis. Biochemistry 35, 9187–9196.
22. Musial, A., Mandal, A., Coroneos, E. & Kester, M. (1995)
Interleukin-1 and endothelin stimulate distinct species of digly-
cerides that differentialy regulate protein kinase C in mesangial
cells. J. Biol. Chem. 270, 21632–21216.
23. Preiss,J.,Loomis,C.R.,Bishop,W.R.,Stein,R.,Niedel,J.E.
& Bell, R.M. (1986) Quantitative measurement of
sn-1,2-diacylglycerols present in platelets, hepatocytes and ras-
and sis-transformed normal rat kidney cells. J. Biol. Chem. 261,

8597–8600.
24. Kiss, Z. & Anderson, W.B. (1989) Phorbol ester stimulates the
hydrolysis of phosphatidylEtn in leukemic HL-60, NIH 3T3, and
baby hamster kidney cells. J. Biol. Chem. 264, 1483–1487.
25. Catala
´
n, R.E., Martı
´
nez, A.M. & Aragone
´
s, M.D. (1984) Evi-
dence for a role of somatostatin in lipid metabolism of liver and
adipose tissue. Regul. Peptides 8, 147–159.
26. Olivera, A., Romanowski, A., Sheela Rani, C.S. & Spiegel, S.
(1997) Differential effects of sphingomyelinase and cell-permeable
ceramide analogs on proliferation of Swiss 3T3 fibroblasts. Bio-
chim. Biophys. Acta 1348, 311–323.
27. Anjum, R., Ali, M., Begum, Z., Vanaja, J. & Kar. A. (1999)
Selective involvement of caspase-3 in ceramide induced apoptosis
in AK-5 tumor cells. FEBS Lett. 439, 81–84.
28. Zupan, L.A., Weiss, R.H., Hazen, S.L., Parnas, B.L., Aston,
K.W., Lennon, P.J., Getman, D.P. & Gross, R.W. (1993) Struc-
tural determinants of haloenol lactone-mediated suicide inhibition
of canine myocardial calcium-independent phospholipase A
2
.
J. Med. Chem. 36, 95–100.
29. Winkler, J.D., Eris, T., Sung, C M., Chabot-Fletcher, M., Mayer,
R.J.,Surette,M.E.&Chilton,F.H.(1996)Inhibitorsofcoenzyme
A-independent transacylase induce apoptosis in human HL-60

cells. J. Pharmacol. Exp. Ther. 279, 956–966.
30. Ballou, L.R., Laulederkind, S.J.F., Rosloniec, E.F. & Raghow, R.
(1996) Ceramide signalling and the immune response. Biochim.
Biophys. Acta 1301, 273–287.
31. Winkler, J.D., Bolognese, B.J., Roshak, A.K., Sung, C M. &
Marshall, L.A. (1997) Evidence that 85 kDa phospholipase A
2
is
not linked to CoA-independent transacylase-mediated production
of platelet-activating factor in human monocytes. Biochim. Bio-
phys. Acta 1346, 173–184.
32. Emoto, K., Toyama-Sorimachi, N., Karasuyama, H., Inoue, K. &
Umeda, M. (1997) Exposure of phosphatidylethanolamine on the
surface of apoptotic cells. Exp. Cell Res. 232, 430–434.
33. Rizzieri, K.E. & Hannun, Y.A. (1998) Sphingolipid metabolism,
apoptosis and resistance to cytotoxic agents: can we interfere?
Drug Resistance Updates 1, 359–376.
34. Spinedi, A., Di-Bartolomeo, S., Di-Sano, F., Rodolfo, C.,
Ambrosino, A. & Piacentini, M. (1999) Ceramide accumulation
precedes caspase-dependent apoptosis in CHP-100 neuroepithe-
lioma cells exposed to the protein phosphatase inhibitor okadaic
acid. Cell Death Differ. 6, 618–623.
35. Saluja, I., O’Regan, M.H., Song, D. & Phillis, J.W. (1999) Acti-
vation of cPLA
2
, PKC, and ERKs in the rat cerebral cortex during
ischemia/reperfusion. Neurochem. Res. 24, 669–677.
36. Husain, S. & Abdel-Latif, A.A. (1999) Endothelin-1 activates p38
mitogen-activated protein kinase and cytosolic phospholipase A
2

in cat iris sphincter smooth muscle cells. Biochem. J. 342, 87–96.
37. Atsumi, G., Murakami, M., Kojima, K., Hadano, A., Tajima, M.
& Kudo, I. (2000) Distinct roles of two intracellular phospho-
lipases A
2
in fatty acid release in the cell death pathway: proteo-
lytic fragment of type IVA cytosolic phospholipase A
2
alpha
inhibits stimulus-induced arachidonate release, whereas that of
type VI Ca
2+
-independent phospholipase A
2
augments sponta-
neous fatty acid release. J. Biol. Chem. 275, 18248–18258.
38. Edgar, A.D., Strosznajder, J. & Horrocks, L.A. (1982) Activation
of ethanolamine phospholipase A
2
in brain during ischemia.
J. Neurochem. 39, 1111–1116.
39. Selznick, L.A., Holtzman, D.M., Han, B.H., Gokden, M., Srini-
vasan, A.N., Johnson, E.M. & Roth, K.A. (1999) In situ
immunodetection of neuronal caspase-3 activation in Alzheimer
disease. J. Neuropathol. Exp. Neurol. 58, 1020–1026.
40. Viani, P., Zini, I., Cervato, G., Biagini, G., Agnati, L.F. &
Cestaro, B. (1995) Effect of endothelin-1 induced ischemia on
peroxidative damage and membrane properties in rat striatum
synaptosomes. Neurochem. Res. 20, 689–695.
46 E. Latorre et al.(Eur. J. Biochem. 270) Ó FEBS 2003