You are on page 1of 28

This document contains text automatically extracted from a PDF or image file.

Formatting may have

been lost and not all text may have been recognized.
To remove this note, right-click and select "Delete table".
Archives of Biochemistry and Biophysics
Vol. 370, No. 2, October 15, pp. 300–307, 1999
Article ID abbi.1999.1410, available online at on

Biosynthesis and Inactivation of N-Arachidonoylethanolamine

(Anandamide) and N-Docosahexaenoylethanolamine
in Bovine Retina
T. Bisogno,*,1 I. Delton-Vandenbroucke,† A. Milone,*,1 M. Lagarde,† and V. Di Marzo*,1,2
*Istituto per la Chimica di Molecole di Interesse Biologico, CNR, Via Toiano 6, 80072 Arco Felice, Napoli, Italy;
and †Institut National de la Sante´ et de la Recherche Me´dicale, U352 Biochimie and Pharmacologie,
INSA de Lyon, 20, avenue A. Einstein, 69621 Villeurbanne (Cedex), France
Received June 10, 1999, and in revised form July 26, 1999
N-Arachidonoylethanolamine (anandamide; AEA)
and 2-arachidonoylglycerol (2-AG), the two proposed
endogenous agonists of cannabinoid receptors, and
the putative AEA biosynthetic precursor, N-arachido-
noylphosphatidylethanolamine (NArPE), were identi-
fied in bovine retina by means of gas chromatography–
electron impact mass spectrometry (GC-EIMS). This
technique also allowed us to identify N-docosahex-
anoylethanolamine (DHEA) and 2-docosahexanoylg-
lycerol (2-DHG), two derivatives of docosahexaenoic
acid (DHA), one of the most abundant fatty acids es-
terified in retina phospholipids and necessary for op-
timal retinal function. N-Docosahexaenoylphosphati-
dylethanolamine (NDHPE), the potential biosynthetic
precursor for DHEA, was also found. The fatty acid
composition of the sn-1 and sn-2 positions of bovine
retina’s most abundant phospholipid classes, also de-
termined here, were in agreement with a phospholip-
id-dependent mechanism for 2-AG, 2-DHG, AEA, and
DHEA biosynthesis, as very high levels of polyunsatu-
rated fatty acids, including DHA, were found on the
sn-2 position of phosphatidylcholine (PC) and -etha-
nolamine (PE), and measurable amounts of di-docosa-
hexanoyl-PC and -PE, two potential biosynthetic pre-
cursors of NDHPE, were detected. Accordingly, we
found that isolated particulate fractions from bovine
retina could release AEA and DHEA in a time-depen-
dent fashion. Finally, a fatty acid amide hydrolase
(FAAH)-like activity with subcellular distribution and
pH dependency similar to those reported for the brain
enzyme was also detected in bovine retina. This activ-
ity was inhibited by FAAH inhibitors, phenylmethyl-
sulfonyl fluoride and arachidonoyltrifluoromethyl-
ketone, and appeared to recognize DHEA with a lower
efficiency than AEA. These data indicate that AEA and
its congeners may play a physiological role in the
mammalian eye.
© 1999 Academic Press
Key Words: cannabinoids; anandamide; 2-arachido-
noylglycerol; bovine retina; docosahexaenoic acid;
Two receptor subtypes for 9-tetrahydrocannabinol
(THC)3, the active constituent of Cannabis sativa, have
been identified and named CB
and CB
. While the
former protein is expressed in both nervous and non-
nervous tissues, the CB
receptor subtype is mostly
confined to cells of the immune system (1, 2). Anand-
amide (N-arachidonoylethanolamine; AEA) and
2-arachidonoylglycerol (2-AG) were isolated from cen-
tral and peripheral tissues (3–5) and suggested to act
as endogenous cannabinoid receptor ligands (hereafter
referred to as “endocannabinoids”). Pathways for AEA
and 2-AG biosynthesis and inactivation by intact cells
have been identified (as reviewed in (6)). It was sug-
1 Affiliated with the National Institute for the Chemistry of Bio-
logical Systems, CNR.
2 To whom correspondence should be addressed. E-mail:
VDM@TRINC.ICMIB.NA.CNR.IT. Fax: 39-081-8041770.
3 Abbreviations used: THC, 9-tetrahydrocannabinol; AEA,
N-arachidonoylethanolamine; 2-AG, 2-arachidonoylglycerol; NAE,
N-acylethenolamine; NArPE, N-arachidonoylphosphatidyletha-
nolaminde; PE, phosphatidylethanolamine; AA, arachidonic acid;
PC, phosphatidylcholine; FAAH, fatty acid amide hydrolase; IOP,
intraocular blood pressure; DHA, docosahexaenoic acid; NDHPE,
N-docosahexanoyl-PE; DHEA, N-docosahexanoylethanolamine;
2-DHG, 2-docosahexanoylglycerol; NP–HPLC, normal phase–high
pressure liquid chromatography; GC, gas chromatography; EIMS,
electron impact mass spectrometric; PBS, phosphate-buffered saline;
DGBZ, diglycerobenzoate; PMSF, phenylmethylsulfonyl fluoride;
ATFMK, arachidonoyltrifluoromethylketone.
300 0003-9861/99 $30.00
Copyright © 1999 by Academic Press
All rights of reproduction in any form reserved.
gested that AEA, analogous to other N-acylethanol-
amines (NAEs), is released from a preformed mem-
brane phospholipid, N-arachidonoylphosphatidyletha-
nolamine (NArPE) through the action of a microsomal
phospholipase D (7). NArPE, in turn, was shown to
derive, by analogy with other N-acyl-PEs, from the
N-arachidonoylation of phosphatidylethanolamine
(PE), catalyzed by a Ca2 -dependent microsomal
transacylase using arachidonic acid (AA) derived from
the sn-1 position of phosphoglycerides (8–11). An al-
ternative AEA biosynthetic pathway, through the en-
ergy-free condensation between AA and ethanolamine,
was also suggested to occur in some reproductive tis-
sues (11). 2-AG biosynthesis depends on sn-2–arachi-
donate-containing phospholipid precursors such as
phosphatidylcholine (PC), phosphatidylinositol, and
phosphatidic acid (9, 12, 13). An enzyme catalyzing the
hydrolysis of AEA to AA and ethanolamine was re-
cently characterized and named “fatty acid amide hy-
drolase” (FAAH) (14). This enzyme catalyzes the hy-
drolysis also of other fatty acid amides and esters,
including 2-AG (see (15) for a review). AEA exhibits
numerous THC-like properties (recently reviewed in
(16, 17), including an inhibitory effect on intraocular
blood pressure (IOP) in normotensive rabbits (18–21).
A series of chiral -substituted AEA derivatives were
synthesized in order to minimize the degradation and
increase the effect of the endocannabinoid on IOP.
These compounds, unlike AEA, produced hypotensive
effects that were not preceded by an initial elevation of
IOP (22), thus suggesting that the arachidonate pro-
duced from AEA hydrolysis could be responsible for
this undesired side-effect. Evidence both in favor and
against the involvement of cannabinoid receptors in
the IOP-lowering effects of cannabinoids have been
reported. High levels of CB
mRNA were found in the
ciliary body of the eye (23), but one study failed to
demonstrate any potent effect for the synthetic canna-
binoid receptor agonist, WIN-55,212-2, or for the met-
abolically stable AEA congener, (R)-( -methanand-
amide (24). In another study (25), however, unilateral
administration of AEA, its metabolically stable R- -
isopropyl analogue, or the synthetic cannabinoid CP-
55,940 did reduce the IOP. Only the effect of the latter
two compounds was prevented by pretreatment of the
animals with the CB
receptor antagonist SR141716A,
suggesting that AEA hypotensive action could be me-
diated also by other mechanisms, possibly due to AEA
hydrolysis to AA. In fact, AEA hydrolase activities
have been identified in porcine ocular tissues (26), thus
prompting a physiological role of AEA, and possibly
2-AG, in the eye. Among the ocular tissues analyzed,
the retina showed the highest specific enzymatic activ-
ity. In this tissue, an endogenous cannabinoid tone
leading to inhibition of dopamine release was recently
suggested based on the finding that a CB


facilitates, whereas a synthetic cannabinoid agonist

inhibits, the release of the neurotransmitter from this
tissue (27). However, despite the increasing evidence
for a possible physiological role of AEA in the retina, no
study has addressed so far the question of whether this
or the other endocannabinoid, 2-AG, is synthesized by
this tissue.
The retina contains a high level of docosahexaenoic
acid (22:6 n 3; DHA), a polyunsaturated fatty acid
that is essential for optimal retinal function as shown
by the observation that low levels of DHA were found
in red blood cells from patients with retinis pigmentosa
(28, 29) or in the plasma of dogs affected by progressive
rod–cone degeneration (30). The fatty acids composi-
tion of retina phosphoglycerides has been reported for
rat and bovine species (31, 32), and DHA was shown to
be among the most abundant fatty acids in all phos-
pholipids, with high levels being found in PC and PE.
Furthermore, unusually high levels of di-docosahex-
anoyl-PC and -PE species, i.e., two of the potential
fatty acid donors for the formation of a putative N-
docosahexanoyl-PE (NDHPE) and, subsequently, of N-
docosahexanoylethanolamine (DHEA), have been re-
ported for rat retina (31). Starting from this back-
ground, the objective of the present study was to
provide an answers to the following two open ques-
tions: (i) Are 2-AG, AEA, and AEA putative biosyn-
thetic precursor, NArPE, produced and metabolized by
the retina? (ii) In view of the high levels of phospho-
lipid-bound DHA found in this tissue, is it possible that
retinas also produce and metabolize 2-docosahexanoyl-
glycerol (2-DHG), DHEA, and DHEA putative precur-
sor, NDHPE?
DHEA was synthesised by mixing equimolar amounts of docosa-
hexaenoyl-chloride (obtained by reacting docosahexaenoic acid
(Sigma) with an excess of oxalyl chloride) and ethanolamine for 20
min at 4°C in anhydrous dichloromethane. The mixture was then
brought to dryness under a flow of N
and purified by normal-phase
high-pressure liquid chromatography (NP-HPLC) carried out with a
semipreparative column (Spherisorb S5W) eluted with a 40-min
linear gradient from 90 to 80% of n-hexane in 2-propanol (flow rate
2 ml/min). Under these conditions DHEA was eluted after 26 min.
NP-HPLC fractions containing DHEA were pooled, and the solvent
was evaporated under a flow of N
. Gas chromatography–electron
impact mass spectrometric (GC-EIMS) analysis of the tris-methyl-
silyl-ether, carried out as previously described (33) on a HP-5989B
quadrupole mass analyzer equipped with an electron impact source
operating at 70 eV and 250°C, confirmed the chemical structure of
the compound, which was eluted from the GC column after 20.5 min.
The GC column was a capillary fused silica column (Optima-1-MS,
0.25 m, Macherey–Nagel; length, 30 m; i.d., 0.25 mm) and was
eluted with a programmed temperature gradient from 200 to 300°C
(5°C/min; flow rate, 1 ml/min; injector and detector temperature,
260°C). Freshly enucleated bovine eyes were obtained from Lyon–
Corbas slaughterhouse. The retinas were removed and extracted
with chloroform/methanol (2/1, by vol). d
-2-AG and d
-AEA (5 nmol),
synthesized from d
-AA and ethanolamine or glycerol (3, 34), were
FIG. 1. Characterization of N-docosahexaenoylethanolamine and 2-docosahexaenoylglycerol in bovine retina. (A) Total ion
EIMS spectrum of synthetic DHEA (1 nmol), (B) selected ion monitoring EIMS spectrum of native DHEA, and (C) total ion-
monitoring EIMS
spectrum of native 2-DHG from bovine retina.

BOO 7000 6000 5000
4000 300 2000 1000
7000 3

134 A
150 200 250 300 350 400 B
175 428
10000 J
1000": 6
50 100 150 200 250 300 350 400 450 S00
added to extracting mixture. The lipids were purified by silica gel
column chromatography and NP-HPLC as described (34). NP-HPLC
fractions with the same retention time of monoacylglycerols (17–22
min) and NAEs (24–28 min) were liophylized and trimethylsilylated
as previously described (33). The samples were analyzed by GC-
EIMS under conditions previously shown to allow the separation of
the components of each family of lipids (34). In order to detect AEA
and DHEA, the EIMS was run in the selected ion-monitoring mode
to improve sensitivity. For other NAEs, 2-AG and 2-DHG, which
were more abundant, the EIMS was run in the total ion current
mode. The amounts of AEA and 2-AG were calculated from the area
ratio between the peaks at m/z 404 and 412 (loss of methyl group
from undeuterated and d
-AEA) and m/z 507 and 515 (loss of
methyl group from undeuterated and d
-2-AG), respectively. The
amounts of endogenous DHEA and 2-DHG were calculated by using
calibration curves obtained with external synthetic standards of
DHEA and 2-AG, respectively. NArPE and NDHPE were identified
as the corresponding NAEs released from the digestion of NArPE-
like chromatographic fractions with S. chromofuscus PLD (Sigma,
UK) (33). For analysis of phosphoglyceride fatty acid composition,
the retinas were removed from freshly enucleated bovine eyes and
placed in cold PBS, minced into small pieces, and homogenized in a
Dounce homogenizer in the presence of butylated hydroxytoluene (50
M) to prevent oxidation. After extraction with chloroform/methanol
(2/1, by vol), phospholipids were separated by two-step TLC by using
the developing system chloroform/methanol (80/8, by vol) followed by
the developing system chloroform/methanol/methylamine (60/20/5,
by vol). Phospholipids were visualized by spraying with 0.02% 2 ,7 -
dichlorofluorescein in 95% aqueous ethanol and extracted from the
silica gel with chloroform/methanol/water (5/5/1, by vol). To analyze
their fatty acid composition, phospholipids were transmethylated for
90 min at 100°C with 5% concentrated sulfuric acid in methanol in
vials sealed under N
to prevent oxidation. The resulting fatty acid
methyl esters were analyzed by GC using a Supelco SP2380 capillary
column (0.2 m, 30 m 0.25 mm) with helium as a carrier gas and
identified by comparison with commercial standards. PE and PC
were converted to their respective diradylglycerols by digestion with
phospholipase C (35). Diradylglycerols were derivatized with 3,5-
dinitrobenzoylchloride in anhydrous pyridine and separated into
diacylglycerobenzoates (DGBZ) and ether-linked glycerobenzoates
by TLC on silica gel using a solvent system of toluene/hexane/diethyl
ether (50/45/5, by vol) (36). The DGBZ were extracted from the silica
gel with hexane and fractionated into molecular species by HPLC on
a Superspher RP-18 column (5 mm 25 cm 4.6 mm) using
acetonitrile/isopropanol (90/10, v/v) at a flow rate of 1 ml/min. Mo-
lecular species were detected at 230 nm and identified by comparison
with standards or by GC analysis of the fatty acid methyl esters. The
former method allowed us to identify and quantify the fatty acids on
the sn-1 (or sn-3) and sn-2 position. Experiments on NAE biosyn-
thesis were performed by using a procedure similar to that previ-
ously described for rat testes (37). Aliquots (3.2 mg of total proteins)
of pooled particulate fractions from a 280,000g centrifugation of
retina homogenates were resuspended in 1 ml Tris–HCl 50 mM, pH
7.4, and incubated at 37°C for different intervals of time (0, 15, or 30
min). Control incubations were carried out with boiled membranes or
with aliquots of 280,000g supernatants. After incubation the mix-
tures were extracted with chloroform/methanol (2/1, by vol) contain-
ing 5 nmol of d
-AEA and d
-palmitoyl-, d
-stearoyl-, d
-oleoyl-, and
-linoleoylethanolamides. The organic extracts were purified and
analyzed as described above. The amounts of NAEs were calculated
by the isotope-dilution GC-EIMS methods described previously (33,
37). Finally, fatty acid amide hydrolase assays were performed as
described previously (38). Bovine retinas were homogenized in 50
mM Tris–HCl, pH 7.4, and centrifuged sequentially at 1500g (10
min), 80,000g (30 min), and 280,000g (30 min). Pellets from the three
centrifugation steps and the supernatant from the last centrifuga-
tion were incubated in 50 mM Tris, pH 9, at 37°C for 30 min in


Fatty Acid Composition of Bovine
Retina Phosphatidylcholine (PC)
and Phosphatidylethanolamine (PE)
Fatty acid PC PE
14:0 0.73 0.18 0.10 0.01
16:0 DMA 0.05 0.02 0.62 0.10
16:0 42.36 3.86 9.26 0.84
16:1 n-9 0.98 0.07 0.23 0.03
16:1 n-7 0.41 0.05 0.12 0.03
18:0 DMA 0.13 0.07 3.46 0.58
18:0 18.07 1.30 33.28 2.19
18:1 n-9 16.08 1.55 6.64 0.59
18:1 n-7 3.20 0.43 1.50 0.29
18:2 n-6 1.10 0.38 0.85 0.30
20:0 0.16 0.03 0.59 0.10
18:3 n-3 0.08 0.02 0.05 0.01
20:2 n-6 0.31 0.03 0.13 0.04
20:3 n-6 0.68 0.16 0.65 0.06
20:4 n-6 4.37 1.13 9.48 1.17
20:5 n-3 0.20 0.09 0.17 0.05
22:4 n-6 0.43 0.16 2.08 0.46
22:5 n-6 0.36 0.32 1.33 0.01
22:5 n-3 0.59 0.12 1.74 0.46
22:6 n-3 10.50 1.59 27.85 2.32
Note. Phospholipids were transmethylated and analyzed by GC.
Data are in mol% SD, n 4.

presence of 50,000 cpm (12.5 M) of [14C]AEA synthesized from

arachidonoylchloride and [14C]ethanolamine as described (3). Incu-
bations were also carried out in buffers at different pH values (38) or
in the presence of arachidonoyltrifluoromethylketone (ATFMK, Bi-
omol), phenylmethylsulfonyl fluoride (PMSF, Sigma) or AEA and
DHEA, synthesised as described above.
In the present study we investigated the occurrence,
in bovine retinas, of the two endocannabinoids AEA
and 2-AG together with their (i) DHA homologues, (ii)
possible phospholipid precursors, and (iii) biosynthetic
and hydrolytic enzymes.
The EIMS spectrum of a synthetic standard of
DHEA and of a bovine retina lipid component with the
same chromatographic behavior in HPLC (r
26 min)
and GC-EIMS (r
20.42 min) analyses are shown in
Figs. 1A and 1B. Synthetic DHEA displayed a typical
EI fragmentation pattern, with a molecular ion at
m/z 443, and main fragment ions at m/z 428 (loss
of a methyl group), m/z 284 and 159 ( -cleavage
followed by rearrangement), m/z 268 and 175 ( -
cleavage), and m/z 116 (cleavage between C1 and
C2 of the ethanolamine moiety). The native component
exhibited all the selected ion signals at m/z 443,
428, 175, and 116 (Fig. 1B). These data indicate that
DHEA is a component of lipid extracts from bovine
retina. The amount of DHEA obtained from the area of
the GC peaks compared to those of the external stan-
Molecular Species Analysis of Bovine
Retina Phosphatidylcholine (PC)
and Phosphatidylethanolamine (PE)
sn-1–sn-2 PC PE
22:6 n-3–22:6 n-3 1.31 0.23 4.16 0.39
Dipolyunsaturated 1.51 0.18 6.70 1.05
18:1–22:6 n-3 1.22 0.05 4.35 0.10
16:0–22:6 n-3 8.05 0.43 14.03 0.17
18:1–20:4 n-6 0.47 0.20 0.86 0.13
16:0–20:4 n-6 2.93 0.60 1.29 0.01
16:0–22:5 n-6 0.83 0.29 0.83 0.28
18:0–22:6 n-3 15.19 1.79 47.83 1.30
18:0–20:4 n-6 4.18 0.31 9.35 1.55
18:0–22:5 n-6 1.11 0.56 2.05 0.45
16:0–18:1 25.66 0.36 2.16 0.76
16:0–16:0 20.52 0.22 2.61 0.33
18:0–18:1 7.24 0.53 1.76 0.56
16:0–18:0 7.64 0.26 0.48 0.15
Others 2.22 0.05 1.54 0.05
Note. Phospholipids were converted into diradylglycerols by treat-
ment with phospholipase C, derivatized with 3,5-dinitrobenzoyl-
chloride, and analyzed by HPLC. Data are in mol% SD, n 3.
Since AA is the most abundant polyunsaturated fatty acid after DHA
in retina phospholipids (see Table I and (31, 32)), the presence of
sn-1 arachidonoyl-PC and -PE in bovine retina can be assumed as
being very likely (the HPLC system used for the separation of dira-
dylglycerol does not allow us to identify these molecular species).
dard was 39.9 19.0 pmol/g (n 2). We also found
measurable amounts of AEA, as detected from the
presence of a GC peak (r
18.00 min) with ion signals
at m/z 419, 404, 175, and 116, typical of the tris-
methyl-silyl-ether of AEA. The amount of AEA was
64.0 9.6 pmol/g (n 2). Finally, the tris-methyl-
silyl-ethers of other typical NAEs were also found,
including the C16:0 (r
13.78 min), C18:0 (r
min), C18:1 (r
16.30 min), and C18:2 (r
min) homologues, as well as the more unusual C22:4
congener (r
20.75 min). These compounds were
Biosynthesis of N-Acylethanolamines (NAEs) in Bovine Retina Membranes
0 min 15 min 30 min
N-Palmitoylethanolamide (C
) 52.4 2.6 60.1 8.2 70.8 11.1
N-Stearoylethanolamide (C
) ND 172.9 31.2 330.9 30.0
N-Oleoylethanolamide (C
18:1,n 9
) 159.6 31.2 165.9 7.8 188.9 16.7
N-Linoleoylethanolamide (C
18:2,n 6
) 12.2 5.6 31.3 4.1 35.2 1.8
N-arachidonoylethanolamide (C
) ND 288.8 20.3 774.3 58.4
N-Docosahexaenoylethanolamide (C
) 92.0 15.2 342.0 142.0 873.4 247.8
Note. NAE release from 280,000g isolated pellets incubated at 37°C in 50 mM Tris–HCl for increasing periods of time. After the
the organic extracts of each incubate were purified by HPLC, trimethylsilylated, and analyzed by isotope-dilution GC-EIMS.
Data are
expressed as fmol mg protein 1 SE, n 3. ND, not detectable.


identified on the basis of their typical fragmentation

patterns in total ion-current EIMS analyses (data not
shown), but could not be quantified in this experiment
due to the absence of the appropriate deuterated stan-
dards in the lipid extraction solvent. An estimate of the
relative amounts in the retina of these NAE species,
however, can be obtained from the experiments carried
out with membrane preparations (see below and Table
III). When the silica column fraction eluted with chlo-
roform/methanol (1/1, by vol) was digested with S.
chromofuscus phospholipase D, it released NAEs that,
once derivatized and analyzed by GC-EIMS, showed
the presence of AEA and DHEA (51.2 31.1 and
26.6 18.1 pmol/g, respectively, n 2) as well as of
the C16:0, C18:0, C18:1, C18:2, and C22:4 NAEs. These
data indicate for the first time the presence in bovine
retina of the NAE family of lipids, including AEA and
DHEA, and of their putative biosynthetic precursors
according to the “phospholipid” pathway (6), i.e., the
corresponding N-acylPEs. Of the several tissues where
the presence of NAEs has been reported so far (re-
viewed in (6, 11)), only rat brain was shown previously
to contain DHEA—in amounts slightly higher than
AEA—as well as the corresponding N-acyl-PE,
NDHPE (8). GC-EIMS also revealed that bovine retina
lipid extracts contained 2-AG (data not shown) and its
docosahexaenoyl homologue, 2-DHG (Fig. 1C). Typical
signals of the EIMS spectrum of the bis-tris-methyl-
silyl-ether of the latter metabolite were the molecular
ion at m/z 546, the peak at m/z 531 (loss of a
methyl group), and the peak at m/z 456 (loss of one
moiety of tris-methyl-silyl-alcohol). The EIMS spec-
trum of the bis-tris-methyl-silyl-ether of 2-AG has been
previously reported (12) and exhibited the same frag-
mentation pattern described above for 2-DHG, with
typical ions at m/z 522, 507, and 432. The amounts
of 2-AG and 2-DHG were 1.63 0.31 and 0.34 0.11
nmol/g, respectively (n 2). It should be pointed out
that the GC-EIMS method used here allows to separate
FIG. 2. Fatty acid amide hydrolase (FAAH)-like enzymatic activity
in bovine retina. FAAH activity was measured by quantifying
[14C]ethanolamine released from [14C]AEA hydrolysis. (A) Subcellu-
lar localization of [14C]AEA-hydrolyzing activity in bovine retina
homogenate. (B) pH dependency of the [14C]AEA-hydrolyzing activity
from bovine retina 280,000g pellets, expressed as a percentage of the


2-monoacylglycerols from their 1(3)-isomers, which are

usually eluted from the column half a minute later.
Occasionally, small amounts of 1(3)-AG and 1(3)-DHG
were found to accompany the corresponding 2-isomers.
The finding of the ethanolamides, N-acyl-PEs and
2-glycerol esters of AA and DHA in the retina, al-
though described here for the first time, is not surpris-
ing if one looks at the fatty acid composition of the two
most abundant phosphoglyceride classes in this tissue,
i.e., PC and PE (Tables I and II). In both phospholipids
we found, in agreement with previous studies (for ex-
ample (31, 32)), very high levels of DHA and AA,
whereas other polyunsaturated fatty acids were less
abundant. These two fatty acids are esterified at both
the sn-2 and, to a lesser extent, sn-1 positions of PE
and PC, thus indicating that the potential phospholipid
precursors for either the 2-glycerol ester or the N-acyl-
PE, respectively, of AA and DHA are available in bo-
vine retina. The hypothesis that AEA and DHEA could
be biosynthesized in the retina like in nervous tissue
(7–10), i.e., directly from their N-acyl-PE precursors, is
in agreement with the observation that total mem-
brane preparations from this tissue—which suppos-
edly contain both these precursors and the phospho-
lipase D enzyme necessary for their conversion into
NAEs (6, 11)—released increasing amounts of AEA
and DHEA when incubated at 37°C in a physiological
buffer (Table III). Also the other NAEs found here in
lipid retina extracts were produced under these condi-
tions, although in lower amounts, whereas no NAE
was released from heat-inactivated membrane prepa-
rations or from aliquots of the 280,000g supernatant
incubated under the same conditions (data not shown).
However, under the conditions used here the activation
of membrane-bound phospholipase A
and D enzymes
that could catalyze the release of free AA and ethanol-
amine from phospholipids and PE, respectively, cannot
be ruled out. Therefore, although we did not add exog-
enous arachidonate or DHA and ethanolamine to the
incubation mixtures, it is still possible that AEA and
DHEA are also produced from the condensation of the
corresponding fatty acids with ethanolamine, as it has
been suggested for AA in some reproductive tissues
(11, 37). Further studies will be required in order to
investigate further the biosynthetic mechanism for
AEA and DHEA in the retina. Interestingly, however,

activity at pH 10 (maximal activity). (C) Effect of various substances

on [14C]AEA hydrolysis by bovine retina 280,000g pellets. The effect
is expressed as a percentage of the activity in the absence of inhib-
itors. Data are means SE from three separate experiments carried
out in duplicate. The asterisk indicates values statistically different
from controls (P 0.05, as determined by the unpaired Student’s t
test). AEA, N-arachidonoylethanolamine; DHEA, N-docosa-
hexaenoylethanolamine; PMSF, phenylmethylsulfonyl fluoride; AT-
FMK, arachidonoyltrifluoromethylketone.

% Maximal activity % Mßximal activity'

the N-acyl composition of NAEs produced from mem-
brane incubations seems to reflect the fatty acid com-
position of PE rather than PC (Table I), thus suggest-
ing that PE may act as the source of the N-fatty acids
of NAEs. Within the PE class, the fatty acid composi-
tion on the sn-2 position rather than the sn-1 position
reflected more closely the composition of NAEs re-
leased after 30-min incubations, thus suggesting that,
under these conditions, the remodeling of phospholip-
ids may occur or that NAEs are produced in part
through the “condensation” pathway (11, 37). Indeed,
synthesis of AEA through this pathway, very probably
through reversal of the action of AEA hydrolase, was
shown to occur in porcine retina when using a very
high concentration (250 mM) of ethanolamine (26).
Porcine ocular tissues, including the retina, were
recently shown to express a membrane-bound AEA
hydrolase whose pH dependency, sensitivity to inhibi-
tors, and specificity for polyunsaturated NAEs other
than AEA was not determined (26). As shown in Fig.
2A, also bovine retina homogenates contain a
[14C]AEA-hydrolyzing activity mostly associated with
particulate fractions (e.g., pellets from 80,000g and
280,000g centrifugations). The enzyme activity in
these fractions was lower than that reported for por-
cine retina, possibly also because we did not use a
saturating concentration of the substrate. However,
bovine retina AEA hydrolase displayed optimal activ-
ity at pH 10 (Fig. 2B), very similar to that previously
observed with FAAH from mammalian brain and rat
liver (15). Furthermore, the enzymatic hydrolysis of
[14C]AEA by microsomal membranes was not only
counteracted by 100 M AEA and 200 M PMSF, but
also by 100 M ATFMK, a more specific inhibitor of
FAAH (Fig. 2C). DHEA (100 M) also significantly
inhibited [14C]AEA hydrolysis, although to a lesser ex-
tent than that observed with 100 M AEA. This latter
finding suggests that the enzyme recognizes as sub-
strate also the ethanolamide of docosahexaenoic acid,
albeit with a lower efficiency. This may indicate that
the half-life of DHEA in bovine retina is relatively
longer than that of AEA, thus possibly compensating
for the lower activity of DHEA at CB
receptors (39).
However, the apparent K
and V
values for the
hydrolysis of DHEA should be calculated by using the
appropriate labeled compound before drawing any con-
clusion on the efficiency with which this NAE species is
recognized by the bovine retina enzyme.
In conclusion, the data reported herein have shown
for the first time that the retina contains AEA and
DHEA and their putative direct biosynthetic precur-
sors, NArPE and NDHPE, as well as precursors for
these latter phospholipids, i.e., diarachidonoyl- and di-
docosahexaenoyl-PC and -PE. We showed that isolated
membranes from bovine retina release enzymatically
AEA and DHEA and contain a FAAH-like activity.


Finally, this tissue also contains 2-AG and 2-DHG, as

well as some of the potential phospholipid precursors
for these compounds. The finding, in bovine retina, of
the endocannabinoids: (i) provides biochemical
grounds to the previously reported pharmacological
evidence of an endocannabinoid tone controlling dopa-
mine release in this tissue (27), and (ii) suggests that
AEA and 2-AG may play a role as local hypotensive
agents in the eye. Future studies will be required in
order to assess not only if the levels of AEA and 2-AG
change with the onset of pathological conditions, but
also to understand the role in the eye of DHEA and
2-DHG. In fact, both these metabolites are weak ago-
nists at the CB
cannabinoid receptors (39, 40), and it
is unlikely that they exert a hypotensive action, unless
it is conclusively proven that AEA activity on IOP is
also due to noncannabinoid receptor-mediated effects.
DHEA and 2-DHG might also act as local inhibitors of
AEA and 2-AG hydrolysis, thereby enhancing the local
effects of the two endocannabinoids, as previously sug-
gested for other NAEs and monoacylglycerols (17, 41).
In any event, the findings reported in the present study
should spark further research efforts aimed at devel-
oping new AEA-derived hypotensive drugs for the cure
of glaucoma.
1. Matsuda, L. A., Lolait, S. J., Brownstein, M. J., Young, A. C., and
Bonner, T. I. (1990) Nature 346, 561–564.
2. Munro, S., Thomas, K. L., and Abu-Shaar, M. (1993) Nature 365,
3. Devane, W. A., Hanus, L., Breuer, A., Pertwee, R. G., Stevenson,
L. A., Griffin, G., Gibson, D., Mandelbaum, A., Etinger, A., and
Mecoulam, R. (1992) Science 258, 1946–1949.
4. Mecoulam, R., Ben-Shabat, S., Hanus, L., Ligumsky, M., Kamin-
ski, N. E., Schatz, A. R., Gopher, A., Almong, S., Martin, B. R.,
Compton, D. R., Pertwee, R. G., Griffin, G., Bayewitch, M., Barg,
J., and Vogel, Z. (1995) Biochem. Pharmacol. 50, 83–90.
5. Sugiura, T., Kondo, S., Sukagawa, A., Nakane, S., Shinoda, A.,
Itoh, K., Yamashita, A., and Waku, K. (1995) Biochem. Biophys.
Res. Commun. 215, 89–97.
6. Di Marzo, V., Melck, D., Bisogno, T., and De Petrocellis, L. (1998)
Trends Neurosci. 21, 521–528.
7. Di Marzo, V., Fontana, A., Cadas, H., Schinelli, S., Cimino, G.,
Schwartz, J-C., Piomelli, D. (1994) Nature 372, 686–691.
8. Sugiura, T., Kondo, S., Sukagawa, A., Tonegawa, T., Nakane, S.,
Yamashita, A., Ishima, Y., and Waku, K. (1996) Eur. J. Biochem.
240, 53–62.
9. Di Marzo, V., De Petrocellis, L., Sugiura, T., and Waku, K. (1996)
Biochem. Biophys. Res. Commun. 227, 281–288.
10. Cadas, H., di Tomaso, E., and Piomelli, D. (1997) J. Neurosci. 17,
11. Schmid, H. H. O., Schmid, P. C., and Natarajan, V. (1996) Chem.
Phys. Lipids 80, 133–142.
12. Stella, N., Schweitzer, P., and Piomelli, D. (1997) Nature 388,
13. Bisogno, T., Melck, D., De Petrocellis, L., and Di Marzo, V. (1999)
J. Neurochem. 72, 2113–2119.
14. Cravatt, B. F., Giang, D. K., Mayfield, S. P., Boger, D. L., Lerner,
R. A., and Gilula, N. B. (1996) Nature 384, 83–87.
15. Ueda, N., Goparaju, S. K., Katayama, K., Kurahashi, Y., Suzuki,
H., and Yamamoto, S. (1998) J. Med. Invest. 45, 27–36.
16. Di Marzo, V. (1998) Biochim. Biophys. Acta 1392, 153–175.
17. Mechoulam, R., Fride, E., and Di Marzo, V. (1998) Eur. J. Phar-
macol. 359, 1–18.
18. Merritt, J. C., Mckinnon, S., Armstrong, J. R., Hatem, G., and
Reid, L. A. (1980) Ann. Ophthalmol. 12, 947–950.
19. Cooler, P., and Gregg, J. M. (1977) South. Med. J. 70, 951–954.
20. Pate, D. W., Jarvinen, K., Urtti, A., Jarho, P., and Jarvinen, T.
(1995) Curr. Eye Res. 14, 791–797.
21. Mikawa, Y., Matsuda, S., Kanagawa, T., Tajika, T., Ueda, N.,
and Mimura, Y. (1997) Jpn. J. Ophthalmol. 41, 217–220.
22. Pate, D. W., Jarvinen, K., Urtti, A., Jarho, P., Mahadevan, T.,
and Jarvinen, T. (1997) Pharm. Res. 14, 1738–1743.
23. Porcella, A., Casellas, P., Gessa, G. L., and Pani, L. (1998) Brain
Res. Mol. Brain Res. 58, 240–245.
24. Hodges, L. C., Reggio, P. H., and Green, K. (1997) Ophthalmic
Res. 29, 1–5.
25. Pate, D. W., Jarvinen, K., Urtti, A., Mahadevan, T., and Jarvi-
nen, T. (1998) Life Sci. 63, 2181–2188.
26. Matsuda, S., Kanemitsu, N., Nakamura, A., Mimura, Y., Ueda,
N., Kurahashi, Y., and Yamamoto, S. (1997) Exp. Eye Res. 64,
27. Schlicker, E., Timm, J., and Gothert, M. (1996) Naunyn-
Schmiedeberg’s Arch. Pharmacol. 354, 791–795.
28. Hoffman, D. R., and Birch, D. G. (1995) Invest. Ophthalmol. Vis.
Sci. 36, 1009–1018.


29. Schaefer, E. J., Robins, S. J., Patton, G. M., Sandberg, M. A.,

Weigel-Di Franco, C. A., Rosner, B., and Berson, E. L. (1995) J.
Lipid. Res. 36, 1427–1433.
30. Aguirre, G. D., Acland, G. M., Maude, M. B., and Anderson, R. E.
(1997) Invest. Ophthalmol. Vis. Sci. 38, 2387–2407.
31. Stinson, A. M., Wiegand, R. D., and Anderson, R. E. (1991) Exp.
Eye Res. 52, 213–218.
32. Aveldano, M. A., and Bazan, N. G. (1983) J. Lipid Res. 24,
33. Bisogno, T., Berrendero, F., Ambrosino, G., Cebeira, M., Ramos,
J. A., Fernandez-Ruiz, J. J., and Di Marzo, V. (1999) Biochem.
Biophys. Res. Commun. 256, 377–380.
34. Bisogno, T., Sepe, N., Melck, D., Maurelli, S., De Petrocellis, L.,
and Di Marzo, V. (1997) Biochem. J. 322, 671–677.
35. Louie, K., Wiegand, R. D., and Anderson, R. E. (1988) Biochem-
istry 27, 9014–9020.
36. Croset, M., Bayon, Y., and Lagarde, M. (1992) Biochem. J. 281,
37. Schmid, P. C., Schwindenbammer, D., Krebsbach, R. J., and
Schmid, H. H. O. (1998) Chem. Phys. Lipids 92, 27–35.
38. Maurelli, S., Bisogno, T., De Petrocellis, L., Di Luccia, A.,
Marino, G., and Di Marzo, V. (1995) FEBS Lett. 377, 82–
39. Felder, C. C., Briley, E. M., Axelrod, J., Simpson, J. T., Mackie,
K., and Devane, W. A. (1993) Proc. Natl. Acad. Sci. USA 90,
40. Sugiura, T., Kodaka, T., Nakane, S., Miyashita, T., Kondo, S.,
Suhara, Y., Takayama, H., Waku, K., Seki, C., Baba, N., and
Ishima, Y. (1999) J. Biol. Chem. 274, 2794–2801.
41. Ben-Shabat, S., Fride, E., Sheskin, T., Tamiri, T., Rhee, M.-H.,
Vogel, Z., Bisogno, T., De Petrocellis, L., Di Marzo, V., and
Mechoulam, R. (1998) Eur. J. Pharmacol. 353, 23–31.