Platelet-activating factor and its structural analogues in the earthworm Eisenia foetida

f.~>..,~;.

;rT~

~t!ll

ELSEVIER

et Biophysica A~ta Biochimica et Biophysica Acta 1258 (1995) 19-26

Platelet-activating factor and its structural analogues in the earthworm Eisenia foetida Takayuki Sugiura a,*, Atsushi Yamashita a, Naomi Kudo a, Teruo Fukuda a, Tatsuya Miyamoto a, Neng-neng Cheng a, Seishi Kishimoto a, Keizo Waku a, Tamotsu Tanaka b,1, Hiroaki Tsukatani b, Akira Tokumura b a Facul~' of Pharmaceutical Sciences, Teikyo University, Sagamiko, Kanagawa 199-01, Japan b Facul~' of Pharmaceutical Sciences, The UniversiO" of Tokushima, Shomachi, Tokushima, Tokushima 770. Japan

Received 2 February 1995; revised 23 March 1995; accepted 5 April 1995

Abstract

The earthworm Eisenia foetida was shown to contain large amounts of ether-containing phospholipids such as alkylacylglycerophosphocholine (61.3% of choline glycerophospholipids) and alkenylacylglycerophosphoethanolamine (66.0% of ethanolamine glycerophospholipids). We also found a substantial amount of ether-containing PAF-like lipid in this animal, its level being increased after the animal is injured. We showed evidence that this PAF-like lipid consists of PAF and PAF analogues containing short chain fatty acids other than acetic acid. Notably, a propionic acid-containing species but not PAF itself, is the most predominant species in this animal. We also confirmed that the earthworms contain enzyme activities involved in the synthesis of PAF and short chain fatty acid-containing PAF analogues. Interestingly, the acetyltransferase activity in earthworms is resistant to high concentrations of the substrate lysophospholipid. Thus, both the structure of the PAF-like lipid and the properties of the enzymes involved in PAF-like lipid metabolism in the earthworms are somewhat different from those in mammalian tissues. Keywords: Platelet-activating factor; GC-MS; Acetyltransferase; Choline-phosphotransferase; (Lower animals)

1. Introduction

Platelet-activating factor (PAF) was first described as a chemical mediator of anaphylaxis by Benveniste et al. [1] in 1972. Its chemical structure was proposed to be a novel type of alkyl ether-linked phospholipid, 1-O-alkyl-2acetyl-sn-glycero-3-phosphocholine (alkylacetyl-GPC) in 1979 [2-4] and was demonstrated to be so later [5]. A number of studies to date have demonstrated that PAF is capable of stimulating not only platelets but also polymorphonuclear leukocytes, macrophages, endothelial cells, smooth muscle and various other tissues and cells in vitro and that PAF exerts profound effects such as rapid decreases in blood pressure and severe gastric ulcer when

Abbreviations: PAF, platelet-activating factor; CGP, choline glycerophospholipids; EGP, ethanolamine glycerophospholipids; GPC, snglycero-3-phosphocholine. * Corresponding author. Fax: + 81 0426 85 1345. ~Present address: Faculty of Engineering, Fukuyama University, Fukuyama 729-02, Japan. 0005-2760/95/$09.50 © 1995 Elsevier Science B.V. All rights reserved SSDI 0 0 0 5 - 2 7 6 0 ( 9 5 ) 0 0 0 9 0 - 9

administered to experimental animals [6-8]. PAF is now regarded as an important lipid mediator in various allergic reactions, shock and inflammation in mammals [6-8]. A unique characteristic of PAF is that it possesses an alkyl ether moiety at the 1-position of the glycerol backbone. The biological activity of the acyl analogue of PAF (l-acyl PAF) is usually very low [2], suggesting that the presence of the alkyl ether residue is crucially important in inducing various biological responses. Interestingly, white blood cells such as polymorphonuclear leukocytes and macrophages contain large amounts of alkyl ether-containing choline glycerophospholipids (CGP) which are usually nearly absent in other mammalian tissues [9]. The high levels of aikylacyl(long chain)-GPC in these inflammatory cells seem to be favorable in producing large quantities of PAF, which is by far more potent compared with the 1-acyl analogue, upon stimulation, because the deacylation-acetylation pathway plays a central role in the biosynthesis of PAF in this type of cells [10]. Recently, we found that very large amounts of alkylacyl-GPC are also present in various species of lower

20

T. Sugiura et al. / Biochimica et Biophysica Acta 1258 (1995) 19-26

animals other than insects [ 11,12]. The levels of the alkylacyl subclass in CGP in various species of invertebrates resembled those in white blood cells in mammals. This finding prompted us to examine whether PAF is present in these lower animals. In fact, we confirmed that considerable amounts of the PAF-like lipid are present in various multicellular invertebrates [11,12], although the levels in insects are generally very low. Despite the abundance of PAF-like lipid in several species of lower animals, its role in these animals still remains unclear. Furthermore, not much is known about the mechanism or regulation of the biosynthesis of PAF in these lower animals. In addition, little information is available concerning the exact chemical structure of the PAF-like lipid in these lower animals. It should be of value to determine whether the PAF-like lipid observed in these lower animals is PAF itself. In the present study, we investigated these points precisely using the earthworm Eisenia foetida. We found that these earthworms contain substantial amounts of the PAFlike lipid, the level of which increased dramatically after injury. Gas chromatography-mass spectrometry (GC-MS) analysis revealed that the bioactive materials are a mixture of PAF and PAF analogues such as 2-propionyl PAF. Enzyme activities involved in the metabolism of PAF and PAF analogues were also examined.

2. Materials and methods 2.1. Chemicals

CDP-[methyl-J4C]choline (55 mCi/mmol), l-O[hexadecyl- 1'2'- 3H]PAF (60 Ci/mmol) and 1-O-hexadecyl-2-[3H]acetyl PAF (10 Ci/mmol) were obtained from DuPont-New England Nuclear (Boston, MA). l-O[Hexadecyl-l'2'-3H]lysoPAF was prepared from 1-O[hexadecyl-l'2'-3H]PAF by brief treatment with sodium methoxide followed by purification by TLC [13]. [l~4C]Acetyl-CoA (60 mCi/mmol) was obtained from ICN Radiochemicals (Irvine, CA). 1-O-Hexadecyl PAF was purchased from Bachem (Bubendorf, Switzerland). l-OHexadecyl lysoPAF and 1-O-hexadecyl-2-acetyl-snglycerol were from Nova Biochem (Laufelfingen, Switzerland). Acetyl-CoA, propionyl-CoA, butyryl-CoA, and essentially fatty acid-free bovine serum albumin (BSA) were from Sigma (St. Louis, MO). CDP-choline was from Yamasa Shoyu (Chiba, Japan). The specific PAF antagonists CV6209 and TCV309 were generous gifts from Takeda Chem. Ind. (Osaka, Japan). 2.2. Animals E. foetida were obtained at a local fishing center and were maintained in containers with soil in a temperaturecontrolled room (20-22°C). The earthworms were placed in containers filled with wet cotton gauze 48 h prior to the

experiments to empty their digestive ducts. The cotton gauze was changed several times, and the earthworms were washed with water to thoroughly remove the excrement. Rabbits (female, Japan white) were obtained from Sankyo Labo Service (Tokyo, Japan). 2.3. Lipid analysis

The earthworms were killed in chloroform/methanol (1:2, v/v), immediately cut into pieces with scissors and then homogenized in a Polytron homogenizer. Total lipids were extracted by the method of Bligh and Dyer [14]. Dibutyryl hydroxytoluene (BHT) was added to avoid lipid oxidation (0.01%, w/v). Individual phospholipids were separated by 2-dimensional TLC developed first with chloroform/methanol/28% NHaOH (65:35:5, v / v ) and secondly with c h l o r o f o r m / a c e t o n e / m e t h a n o l / a c e t i c acid/water (5:2:1:1.3:0.5, v/v). Lipid spots were visualized under UV light after spraying with primuline. The phosphorus content was estimated according to the method of Rouser et al. [15]. To examine the proportions of alkenylacyl, alkylacyl and diacyl subclasses, choline and ethanolamine glycerophospholipids (CGP and EGP) were extracted from the silica-gel by the method of Bligh and Dyer. CGP and EGP were then hydrolyzed by phospholipase C (Bacillus cereus) and the resultant diradylglycerols were acetylated with acetic anhydride and pyridine. Three types of 1,2-diradyl-3-acetylglycerols (alkenylacyl, alkylacyl and diacyl subclasses) were separated by TLC developed first with petroleum ether/diethyl ether/acetic acid (90:10:1, v / v ) and then with toluene as described previously [16]. The fatty acyl moieties of 1,2-diradyl-3acetylglycerols were converted to fatty acid methyl esters by treatment with 0.5 M methanolic sodium methoxide and were analyzed by gas-liquid chromatography (GLC) using heptadecanoic acid (17:0) methyl ester as an internal standard [ 16]. 2.4. PAF-like lipid in the earthworms

The earthworms were placed in small beakers in a temperature-controlled room (20-22°C) and were pricked 30 times with a stainless steel needle. After 20 min-24 h, 19 ml of chloroform/methanol (1:2, v / v ) was added to the individual beakers, and the earthworms were cut into pieces with scissors. Total lipids were extracted according to the method of Bligh and Dyer, and fractionated by TLC developed with chloroform/methanol/water (65:35:6, v/v). The fraction corresponding to authentic PAF (between CGP and lysoCGP) was scraped off, and the PAFlike lipid was extracted from the silica-gel by the method of Bligh and Dyer. PAF-like lipid was assayed by measuring its ability to aggregate washed rabbit platelets in an aggregometer (Niko Hematracer, PAT-2M, Tokyo, Japan) using 16:0 PAF as a standard as described previously [ 17].

T. Sugiura et al. / Biochimica et Biophysica Acta 1258 (1995) 19-26

2.5. Preparation of subcellular fractions Earthworms and rabbit spleens were homogenized with a Potter-Elvehjem homogenizer in 25 mM Tris-HClbuffered (pH 7.4) 0.25 M sucrose containing 1 mM EDTA. The homogenate was first centrifuged at 700 X g for 10 rain, and the obtained supernatant was centrifuged twice at 7000 X g for 10 min each time. The resultant supernatant was taken and further centrifuged at 105 000 X g for 60 min. The pellet was resuspended in the same buffer and centrifuged again at 105 000 X g for 60 min. The washed pellet was resuspended in the same buffer. Each fraction (700 X g supernatant, 7000 X g pellet, 105 000 X g pellet and 105 000 X g supernatant) was stored at -80°C. Protein content was estimated according to the method of Lowry et al. [18].

2.6. Enzyme assays LysoPAF:acetyl-CoA acetyltransferase activity was measured as described previously [13]. Briefly, subcellular fractions (0.5 or 0.2 mg protein) were incubated with 20 /xM lysoPAF and 100/zM [ ~4C]acetyl-CoA (120 000 dpm) in 0.5 ml of 100 mM Tris-HCl buffer (pH 7.4) containing 2 mM CaC12 at 22°C for 5 rain. The enzyme reaction was terminated by adding chloroform/methanol (1:2, v/v). Total lipids were extracted by the method of Bligh and Dyer and fractionated by TLC developed with chloroform/methanol/water (65:35:6, v/v). The individual lipid spots were scraped off into counting vials and the radioactivities were estimated. Acetyltransferase activity was calculated from the conversion rate of [14C]acetyl-CoA to [laC]PAF. In some experiments, [3H]lysoPAF and unlabeled acetyl-CoA or other fatty acyl-CoA were employed instead of lysoPAF and [14C]acetyl-CoA. In this case, the conversion rates of [3H]lysoPAF to [3H]PAF or short chain fatty acid-containing [3H]PAF analogues were calculated. The l-O-alkyl-2-acetyl-sn-glycerol:CDP-choline cholinephosphotransferase activity was measured according to the method of Woodard et al. [19] with slight modifications. Subcellular fractions (0.5 mg protein) were incubated with 200 /zM 1-O-alkyl-2-acetylglycerol and 100 /zM CDP-[14C]choline (20000 dpm) in 0.5 ml of 100 mM Tris-HCl buffer (pH 7.4) containing 10 mM MgC12, 0.5 mg of BSA and 0.5 mM EGTA at 22°C for 5 min. The reaction was stopped by adding chloroform/methanol (1:2, v/v). Total lipids were extracted by the method of Bligh and Dyer, and fractionated by TLC developed with chloroform/methanol/water (65:35:6, v/v). The fraction corresponding to PAF was scraped from the plate into a counting vial, and the radioactivity was estimated. The enzyme activity was calculated from the conversion rate of CDP[ 14C]choline to [ 14C]PAF. The PAF acetylhydrolase activity was measured using 2-O-[3H]acetyl PAF as a substrate [20]. Subcellular frac-

21

tions (0.5 mg protein) were incubated with 20 /xM 2[3H]acetyl PAF (400000 dpm) dispersed in 0.5 ml of 100 mM Tris-HC1 buffer (Branson Sonifier, setting 2, 15 s) containing 2 mM EDTA at 22°C for 5 rain. The enzyme reaction was stopped by adding chloroform/methanol (1:2, v/v), and the total lipids were extracted by the method of Bligh and Dyer. The upper layer of the Bligh and Dyer extraction mixture was further washed twice with chloroform and the radioactivity retained in the upper layer was estimated. The acetylhydrolase activity was calculated from the radioactivity of [3H]acetate released into the upper layer and that of [3H]acetyl PAF added as the substrate.

2.7. GC-MS analysis of PAF-like lipid PAF-like lipid was separated by TLC developed with chloroform/methanol/water (65:35:6, v / v ) and further purified by TLC developed with methanol/water (2:1, v/v). The purified PAF-like lipid was hydrolyzed by phospholipase C (Bacillus cereus). The resultant diradylglycerols were treated with tert-butyldimethylchlorosilane (Sin-Etsu Chemical, Tokyo, Japan), imidazole and dimethylformamide. The resultant tert-butyldimethylsilyl (tBDMS) derivative was purified by TLC developed with hexane:diethyl ether (9:1, v/v). The tBDMS derivative of the PAF-like lipid was subjected to GC-MS analysis using a Hitachi M-80B mass spectrometer coupled with a gas chromatograph equipped with a fused silica capillary column (J and W Scientific, DB-I, 30 m × 0.242 mm I.D., 0.25 /zM-thickness) as described previously [21,22]. Ion peaks at m / z 415, 425,429, 443, 457, 471 and 485 were monitored.

3. Results The earthworm E. foetida was found to contain phospholipids at a level of 260/z g P / g body weight (the mean of three determinations). Table 1 shows the phospholipid composition. CGP and EGP were the predominant phosTable 1 Phospholipid composition of the earthworm E. foetida Phospholipids

%

Choline glycerophospholipids Ethanolamine glycerophospholipids Inositol glycerophospholipids Serine glycerophospholipids Cardiolipin Phosphatidic acid Lyso choline glycerophospholipids Lyso ethanolamine glycerophospholipids X Others

53.5 _+0.8 24.7 + 1.8 4.6 __+0.4 3.3 _+0.4 2.1 + 0.1 0.5 + 0.2 2.3 + 0.2 3.1 + 0.6 3.1+0.5 2.8 _+0.6

(n = 3) Lipids were extracted from the earthworms and fractionated by 2-dimensional TLC. Lipid phosphorus was estimated according to the method of Rouser et al. as described in Section 2.

T. Sugiura et al. / Biochimica et Biophysica Acta 1258 (1995) 19-26

22

Table 2 Alkenylacyl, alkylacyl and diacyl subclasses of choline and ethanolamine glycerophospholipid fractions of the earthworms E. foetida

Alkenylacyl Alkylacyl Diacyl

CGP

EGP

3.3 ± 1.2 61.4 ± 6.0 35.3 ± 6.2

66.1 -t-4.5 8.2 5:3.5 25.7 ± 5,4 (n = 4)

A "0 O.

3.0

0 J= 13. (n 0 r" O. m 0

2.0

E 0

E

3

CGP and EGP were hydrolyzed by phospholipase C and the resultant diradylglycerols were treated with acetic anhydride and pyridine. 1,2-Diradyl-3-acetylglycerols were separated by TLC developed first with petroleum ether/diethyl ether/acetic acid (90:10:1, v / v ) and then with toluene. The amounts of alkenylacyl, alkylacyl and diacyl subclasses were estimated from their fatty acyl quantities using 17:0 methyl ester as an internal standard as described in Section 2.

'~.

1.o .......................................

,lg .-7 ,lk

< 0,0

0

3

6

,

,

'

12

1B

24

'

Time ( h )

pholipid classes, together accounting for about 80% of total phospholipids. Relatively high levels of lysoCGP and lysoEGP were also found. We also observed that substantial amounts of inositol glycerophospholipids (IGP), serine glycerophospholipids (SGP) and cardiolipin, whereas we could not detect sphingomyelin in this species. In addition to these well-known phospholipid classes, E. foetida contained an unidentified phosphorus-positive polar lipid (X) which migrated between lysoCGP and the origin on 2-dimensional TLC plates. Further chemical elucidation of this unusual lipid fraction was not accomplished in the present study, although several investigators have recently reported the occurrence of novel glycophospholipids in Annelida [23,24]. The subclass composition of CGP and EGP is shown in Table 2. The most prominent observation was that a very large portion of CGP was accounted for by the alkylacyl subclass. The proportion of the diacyl counterpart in CGP is about 3 / 5 of that of the alkylacyl subclass. On the other hand, a large portion of EGP was accounted for by the alkenylacyl subclass rather than by the alkylacyl counterpart. The levels of alkenylacyl subclass in CGP and the alkylacyl subclass in EGP were relatively low. From the results shown in Table 1 and Table 2, it is apparent that alkylacyl-GPC, a potential precursor of PAF in mammalian inflammatory cells, is the most abundant phospholipid (34% of the total phospholipids) in this species. We next examined whether PAF is present in E. foetida.

Fig. 1. Changes in the levels of PAF-like lipid in the earthworms E.

foetida after pricking. Earthworms were pricked 30 times with a stainless steel needle, and after the indicated periods, were killed in chloroform/methanol and cut into pieces with scissors. Total lipids were separated by TLC and the fraction corresponding to authentic PAF was scraped off. PAF-like lipid was assayed by measuring its ability to aggregate washed rabbit platelets using 16:0 PAF as a standard as described in Section 2. The amounts of PAF-like lipid were expressed as the amounts of 16:0 PAF. (Q), pricked; (O), control. The values are the m e a n ± S.D. from 5 determinations.

These earthworms contained significant amounts of a PAF-like lipid as shown by bioassay using washed rabbit platelets. The aggregation induced by the sample was completely abolished by pretreatment of the platelets with the specific PAF antagonists CV6209 and TCV309 (data not shown). The level of PAF-like lipid in the earthworms was 630 nmol (estimated as C16:0 PAF)/mol phospholipid (9.0 pmol (as C16:0 P A F ) / g body weight). This value, however, was somewhat variable dependent on the room temperature: thus, we fixed the temperature at 2022°C for the in vivo experiments. Interestingly, the amounts of the PAF-like lipid were increased markedly after pricking the worms with a needle, reaching a peak at around 3 h and decreasing gradually thereafter (Fig. 1). These observations indicate that the PAF-like lipid is actively synthesized and metabolized in the earthworms upon physical stimuli such as pricking. Next, we explored the enzyme activities, involved in the

Table 3 Distribution of the enzyme activities involved in PAF metabolism among subcellular fractions

700 × g supernatant 7000 × g precipitate 105000 × g precipitate 105000 × g supernatant

LysoPAF:acetyl-CoA acetyltransferase ( p m o l / m i n per mg protein)

AlkylacetylglycerohCDP-choline cholinephosphotransferase ( p m o l / m i n per mg protein)

PAF acetylhydrolase

13 _+ 1 18 ±_ 1 56 +_ 1 0.4 ± 0.4

46 ± 4 29 _ 6 136 + 7 4 5:3

489 ± 6 247 ± 17 100 ± 2 1184 ± 58

( p m o l / m i n per mg protein)

(n = 3) Subcellular fractions were prepared from the earthworms and enzyme activities were estimated as described in Section 2.

T. Sugiura et al. / Biochimica et Biophysica Acta 1258 (1995) 19-26 1.5

c

D,. 01

E

1.0

m o

E 'lO

EL 0

0..

ii •

,w

0

,

5

10

15

T i m e (rain)

Fig, 2. Formation of PAF in earthworm homogenate induced by the addition of substrates. Earthworm homogenate (700X g supernatant, 0.5 mg protein) was incubated in 100 mM Tris-HC1 buffer containing 100 /zM acetyl-CoA and 2 mM CaCI 2 (pH 7.4) ( 0 ) or 100/zM CDP-choline, 0.5 mM EGTA and l0 mM MgCI 2 (pH 8.0) ( O ) at 22°C for the indicated periods. Total lipids were extracted and fractionated. The amounts of PAF (expressed as the amounts of 16:0 PAF) were estimated by bioassay using washed rabbit platelets. The values are the mean + S.D. from three determinations.

synthesis and catabolism of PAF. We found that the earthworms actually contained several PAF-metabolizing enzyme activities, and Table 3 shows the distribution of these enzyme activities in the subcellular fractions. The activities o f the P A F - s y n t h e s i z i n g enzymes lysoPAF:acetyl-CoA acetyltransferase and 1-O-alkyl-2acetyl-sn-glycerol:CDP-choline cholinephosphotransferase were both found mainly in the 105 000 X g pellet fraction (microsomal fraction), with negligible activities in the

e.-

400.

0,. 0') p,

3OO

>

0

E

O. v

105000 X g supernatant fraction (cytosolic fraction). On the other hand, the activity of the PAF-degrading enzyme PAF acetylhydrolase was mainly distributed in the 105 000 x g supernatant fraction. It appears, therefore, that the synthesis and the catabolism of PAF take place, in large part, at different loci within the cell. The activities of the above two PAF-synthesizing enzymes were further compared using the homogenate and endogenous lipid substrates. As demonstrated in Fig. 2, the addition of CDP-choline did not trigger the formation of measurable amounts of PAF. On the other hand, we found that the supplementation of acetyl-CoA in the homogenate per se induced time-and dose-dependent PAF formation. These observations strongly suggest that lysoPAF:acetylCoA acetyltransferase, rather than l-O-alkyl-2-acetyl-snglycerol:CDP-choline cholinephosphotransferase, is of physiological importance in the synthesis of PAF in this animal. Fig. 3 shows the effects of varying concentrations of lysoPAF on the activity of lysoPAF:acetyl-CoA acetyltransferase in the microsomal fractions obtained from earthworms and from rabbit spleen. As depicted in the right panel (b), the highest enzyme activity was observed at 20 /xM (22°C) or 30 /zM (37°C) lysoPAF in the rabbit spleen microsomal fraction. Interestingly, the enzyme activity was markedly reduced in the presence of higher concentrations of lysoPAF; indeed, the activity was almost undetectable when 120 /zM lysoPAF was employed. In contrast to rabbit spleen microsomal fraction, the enzyme activity in the earthworm microsomal fraction was less sensitive to high concentrations of lysoPAF (Fig. 3a). In fact, the highest activity was noted at 120 /xM lysoPAF when assayed at 22°C, suggesting that the nature of

l

(a)

20O

23

(b)

oi t

>

go

100

¢o

tj

,¢

Ot 0

.

, 40

. 8O

Lyso-PAF ~ M )

120

0

40

80

120

Lyso-PAF (pM)

Fig. 3. Effects of varying concentrations of lysoPAF on the activity of lysoPAF:acetyI-CoA acetyhransferase in earthworm microsomal fraction (a) and rabbit spleen microsomal fraction (b). Earthworm microsomal fraction (0.2 mg protein) and rabbit spleen microsomal fraction (0.2 mg protein) were incubated with I 0 0 / z M []4C]acetyl-CoA (120000 dpm), 2 mM CaC12 and various concentrations of lysoPAF at 22°C ( O ) or 37°C ( 0 ) for 5 min. Total lipids were extracted and fractionated by TLC. The radioactivity appearing in the PAF fraction was estimated and the enzyme activity was calculated. The data are the mean + S.D. from three determinations.

24

7". Sugiura et al. / Biochimica et Biophysica Acta 1258 (1995) 19-26 [M-571+ .I

. . . . . . . . .

, . . . . . . . . .

I . . . . . . .

,.i

. . . . . . . . .

l . . . . . . . . .

~ . . . . . . . . .

i

. . . . . . . . .

i

. . . . . . . . .

r.

300

"

200

'

100

"

t.

o 0. oi

L 471 0

~

]

~

-

E

• 457

"o (u

~ 8

~

E 0

. 443

"o D. m o

• 429

,< rl

0

•

0

,

200

.

,

•

400

,

.

600

r

•

BOO

,

.

1000

1200

-415 10

12

14

Short chain acyI-CoA (pM)

16

Time (min)

Fig. 4. GC-MS analysis of PAF-like lipid obtained from the earthworm E. foetida. Purified PAF-like lipid was hydrolyzed by phospholipase C. The resultant diradylglycerols were converted to tBDMS derivatives and analyzed by GC-MS as described in Section 2. The data are representative of separate three samples.

lysoPAF:acetyl-CoA acetyltransferase in earthworms is different from that in mammalian tissue. We then examined the structure of the PAF-like lipid by GC-MS analysis. Fig. 4 shows a typical ion chromatogram of the tBDMS derivative of PAF-like lipid from the earthworm at m/z 415, 429, 443, 457 and 471. From comparison of their retention times and mass spectra with those of standards, peaks 1-8 were identified as due to [M-57] ÷ of the tBDMS derivative, which was separated into 8 molecular species based on the combination of the fatty chain at the l-and 2-positions: C 16:0(alkyl)/C2:0(acyl), C 16:0(alkyl)/C 3:0(acyl), C 16:0(alkyl)/C 4:0(acyl), C 17:0(alkyl)/C3:0(acyl), C 18:0(alkyi)/C2:0(acyl), C16:0(alkyl)/C5:0(acyl), C18:0(alkyl)/C3:0(acyl) and C18:0(alkyl)/C4:0(acyl) (Table 4). Electron impact spectra at a top of peaks 1-8 consisted of a diagnostic ion produced by the rearrangement of the short-chain acyl and dimethylsilanol moieties; this can be seen at m/z 117, 131, 145 and 159 in the mass spectra of acetate, propionate, butyrate and valerate-containing species, respectively. An additional diagnostic ion, [alkyl + 130] ÷, was observed at m/z 355, 369 and 382 in the mass spectra of 16:0, 17:0 and 18:0-alkyl species, respectively. Interestingly, molecular species containing short chain fatty acids other than acetic acid (C2:0) such as propionic acid (C3:0) were highly abundant in this species. In fact, propionic acid (C3:0)-containing species accounted for 54% of the total molecular species, whereas acetic acid (C2:0)-containing species were minor components (12%). In our capillary GC, there was good separation between tBDMS derivatives of PAF-like lipids and their 1-acyl type analogues with the same molecular weight. For instance,

Fig. 5. Enzymatic formation of PAF-like lipid by earthworm microsomal fraction. Earthworm microsomal fraction (0.5 mg protein) was incubated in 100 mM Tris-HCl buffer (pH 7.4) containing 20 p,M [3H]IysoPAF, 2 mM CaCI 2 and various concentrations of acetyl-CoA (O), propionyl-CoA ( • ) and butyryl-CoA ( • ) at 22°C for 5 min. The enzyme reaction was stopped by adding chloroform/methanol (1:2, v/v). Total lipids were extracted and fractionated by TLC as described in Section 2. The data are the mean + S.D. from three determinations.

tBDMS derivative of 1-O-hexadecyl-2-propionyl-snglycerol eluted faster than that of 1-palmitoyl-2-acetyl-snglycerol. We detected no substantial amounts ( < 1%) of l-acyl type analogues of the PAF-like lipid in this animal. Thus, the molecular species profile of the PAF-like lipid in the earthworm was found to have several characteristic features. We further examined whether the earthworm microsomal fraction could synthesize PAF analogues such as 2-propionyl PAF and 2-butyryl PAF shown to be present in this animal. As demonstrated in Fig. 5, the earthworm microsomal fraction was capable of synthesizing 2-propionyl PAF and 2-butyryl PAF in addition to PAF from lysoPAF in the presence of propionyl-CoA, butyryl-CoA and acetyl-CoA, respectively. The formation of PAF and PAF analogues was increased with incubation period (data not shown) and with substrate concentration. The apparent

Table 4 GC-MS analysis of PAF-like lipid in the earthworm E. foetida Peak

Molecular species (sn-1/sn-2)

%

1 2 3 4 5 6 7 8

16:0(alkyl)/2:0(acyl) 16:0(alkyl)/3:0(acyl) 16:0(alkyl)/4:0(acyl) 17:0(alkyl)/3:O(acyl) 18:0(alkyl)/2:O(acyl) 16:0(alkyl)/5:O(acyl) 18:0(alkyl)/3:0(acyl) 18:0(alkyl)/4:O(acyl)

4 14 10 10 8 13 30 11

PAF-like lipid was purified and analyzed by GC-MS as described in Section 2. The peak numbers correspond to those in Fig. 4.

T. Sugiura et al. / Biochimica et Biophysica Acta 1258 (1995) 19-26

K m values for acetyl-CoA, propionyl-CoA and butyrylCoA were 541, 437 and 110 /zM, respectively, and the Vmax values were 442, 193 and 123 pmol/min per mg protein, respectively. The presence of Ca 2+ was essential for the maximal enzyme activity in each case (data not shown).

4. Discussion

Ether phospholipids are lipid molecules widely distributed throughout the animal kingdom; however, their biological functions have not been elucidated fully. As shown in Table 2, E. foetida contain large amounts of two types of ether phospholipids: alkylacyl-GPC and alkenylacyI-GPE. Similar results have already been observed for various species of invertebrates [11,12], although the levels differ between species. The value for the earthworms appears to be among the highest. High levels of alkylacylGPC in these lower animals are particularly noteworthy, because the levels of alkylacyl-GPC in mammalian tissues other than several cell types such as white blood cells are known to be generally very low [9]. Various species of insects have also been shown to contain negligible amounts of alkylacyl-GPC [12]. Thus, the accumulation of alkylacyl-GPC is possibly a characteristic feature of rather primitive lower animals, though there are some exceptional cases. Considering that alkylacyl-GPC is regarded as a stored precursor form of PAF in mammalian white blood cells, we examined the PAF levels in these lower animals. We found that the PAF-like lipid also appears to be widely distributed throughout the animal kingdom [12]. Several investigators have also reported the occurrence of PAF-like lipid in primitive eukaryotes such as the protozoa Tetrahymena pyr~fi)rmis [25] and the yeast Saccharomyses cere~'isiae [26]. Nakayama et al. [26] demonstrated that the PAF-like lipid in yeast consists of several molecular species of PAF and 1-acyl PAF. However, it still remains unclear whether the PAF-like lipid present in lower animals is PAF itself or PAF analogue(s). Furthermore, it is not known whether PAF-metabolizing enzymes are different from those of mammalian tissues. Here, we found that the PAF-like lipid obtained from the earthworms consists of PAF and several PAF analogues. Surprisingly, propionic acid-containing species, but not PAF itself, are the major constituents of the PAF-like lipid in the earthworms (Fig. 4 and Table 4). This is the first report demonstrating the occurrence of PAF analogues such as 2-propionyl PAF in lower animals. The abundance of 2-propionyl PAF and 2-butyryl PAF in the PAF-like lipid fraction of the earthworms is clearly distinct from the results for the PAF-like lipid obtained from mammalian tissues and cells such as ionophorestimulated human polymorphonuclear leukocytes [27] or bovine brain [21,28], where the PAF-like lipid consists exclusively of acetic acid-containing species, i.e., PAF

25

itself. Ninio and co-workers [27] confirmed that polymorphonuclear leukocytes do not produce 2-propionyl PAF even when stimulated in the presence of propionic acid. Thus, 2-propionyl PAF as well as other short chain fatty acid-containing PAF analogues are apparently minor components of PAF-like lipid, if any, in mammalian tissues and do not seem to be of physiological significance compared with PAF, although the biological activity of 2-propionyl PAF is almost comparable to that of PAF (about 70%) [2]. In contrast to the 1-alkyl species, bovine brain contains various species of short chain fatty acid-containing l-acyl PAF analogues [21,28]. However, the physiological significance of short chain fatty acid-containing l-acyl analogues of PAF still remains ambiguous, because these lipid molecules are known to possess only weak PAF-like activity [29]. The mechanism underlying the accumulation of short chain fatty acid-containing PAF analogues such as 2-propionyl PAF is not yet fully understood. There are two possible synthetic pathways: one is the enzymatic acylation of lysoPAF with short chain fatty acyl-CoA, and the other is the peroxidation and subsequent nonenzymatic cleavage of polyunsaturated fatty acyl residues of alkylacyl-GPC [22]. The 2-propionyl PAF analogue can theoretically be derived from 2 2 : 6 ( n - 3)-containing species. In fact, Tanaka et al. [22] demonstrated that l-hexadecyl-2docosahexaenoyl-GPC generates 2-propionyl PAF and 2succinyl PAF after undergoing lipid peroxidation. However, we could not detect appreciable amounts of 22:6(n 3) in E. foetida in this study. Therefore, it seems likely that various short chain fatty acid-containing PAF analogues including 2-propionyl PAF present in the earthworms are formed in large part through the acylation of lysoPAF with short chain fatty acyl-CoA. This was further supported by the observations that earthworms are enriched in lysoPAF (Yamashita et al., unpublished data), and that the earthworm microsomal fraction contains acetyltransferase activity catalyzing the transfer of short chain fatty acyl moieties from respective CoA esters to acceptor lysoPAF (Fig. 5). Thus, the availabilities of acetyl-CoA and propionyl-CoA as well as other short chain fatty acyl-CoA are important factors determining the fatty acyl profiles of the PAF-like lipid in this animal. The question remains as to how propionic acid and other short chain fatty acids are supplied in this species. There are several possible routes. Propionic acid is known to be produced through rumen fermentation in ruminants, and several types of microorganisms may also be involved in earthworms. It is reasonable to assume that earthworms have symbiotic relationships with various microorganisms in the soil. Alternatively, propionic acid may be derived from branched chain amino acids and branched chain or odd-numbered straight chain fatty acids. In any case, further studies are required to determine the origin of propionic acid and other short chain fatty acids in this animal. One of the striking observations in this study was that

26

T. Sugiura et al. / Biochimica et Biophysica Acta 1258 (1995) 19-26

the acetyltransferase activity in the earthworms was resistant to high concentrations of lysoPAF such as 120 /zM, in marked contrast to that in mammalian tissue (Fig. 3). We confirmed that the acetyltransferase activity shown here can be distinguished from long chain fatty acylCoA:IysoPAF acyltransferase activity on several counts such as the requirement of Ca 2+ for the maximal activity (data not shown). On the other hand, the acetyitransferase activities in rat spleen [30], murine peritoneal macrophages [31], rabbit alveolar macrophages [13] and rabbit spleen (Fig. 3) were shown to be reduced dramatically by high concentrations of the substrate lysoPAF ( > 40 ~M). Several investigators have demonstrated that acetyltransferase in mammalian tissues and cells is sensitive to treatment with other types of detergents [30,31]. It is plausible, therefore, that acetyltransferase in mammalian tissues and that in earthworms are considerably different in their stabilities toward detergents. Such a characteristic feature of earthworm acetyltransferase should be of value in studying the biochemical properties of this enzyme, especially of solubilized and purified enzyme protein. Finally, the role of the PAF-like lipid in the earthworms should be determined. It is unlikely that these animals contain PAF-like lipid without any biological function. Here, we showed evidence that earthworms contain enzyme activities catalyzing the formation of PAF and PAF analogues (Table 3 and Fig. 5), and it appears that the PAF-like lipid is continuously produced and metabolized in living animals so that its level in vivo can remain constant. On the other hand, the level of PAF-like lipid increased markedly under pathological conditions such as after injury. These observations suggest that PAF-like lipid has some physiological function, although to date its exact roles are still uncertain. PAF is known to be involved in diverse pathophysiological events in various mammalian tissues and cells such as cardiovascular, respiratory, renal, immune, nervous, reproductive, and digestive systems [68]. Therefore, it seems possible that PAF-like lipid also exerts similar effects (if not all) through putative specific receptor site(s) even in this lower animal. Alternatively, PAF-like lipid may play some unknown biological roles in this type of lower animal. Further detailed studies will clarify the physiological significance of PAF and PAF analogues in lower animals and other primitive eukaryotes. Such studies should be of value for the better understanding of the roles of PAF, an evolutionarily preserved molecule, in mammals.

Acknowledgements This study was supported in part by a Grant-in Aid (No. 03808019) from the Ministry of Education, Science and Culture of Japan.

References [1] Benveniste, J., Henson, P.M. and Cochrane, C.G. (1972) J. Exp. Med, 136, 1356-1377. [2] Demopoulos, C.A., Pinckard, R.N. and Hanahan, D.J. (1979)J. Biol. Chem. 254, 9355-9358. [3] Blank, M.L., Snyder, F., Byers, L.W., Brooks, B. and Muirhead, E.E. (1979) Biochem. Biophys. Res. Commun. 90, 1194-1200. [4] Benveniste, J., Tence, M., Varenne, P., Bidault, J., Boullet, C. and Polonsky, J. (1979) C.R. Acad. Sci. 289D, 1037-1040. [5] Hanahan, D.J., Demopoulos, C.A., Liehr, J. and Pinckard, R.N. (1980) J. Biol. Chem. 255, 5514-5516. [6] Braquet, P., Touqui, L., Shen, T.Y. and Vargaftig, B.B. (1987) Pharmacol. Rev. 39, 97-145. [7] Barnes, P.J., Page, C.P. and Henson, P.M, (Eds.) (1989) PlateletActivating Factor and Human Dieseases, Blackwell, London. [8] Prescott, S.M., Zimmerman, G.A. and Mclntyre, T.M. (1990) J. Biol. Chem. 265, 17381-17384. [9] Sugiura, T. and Waku, K. (1987) in Platelet-Activating Factor and Related Lipid Mediators (Snyder, F,, ed.), pp. 55-85, Plenum Press, New York. [10] Lee, T-C. (1987) in Platelet-Activating Factor and Related Lipid Mediators (Snyder, F., ed.), pp. 115-133, Plenum Press, NY. [11] Sugiura, T., Ojima, T., Fukuda, T., Satouchi, K., Saito, K. and Waku, K. (199l) J. Lipid Res. 32, 1795-1803. [12] Sugiura, T., Fukuda, T., Miyamoto, T. and Waku, K. (1992) Biochim. Biophys. Acta 1126, 298-308. [13] Sugiura, T., Ojima-Uchiyama, A., Masuzawa, Y., Fujita, M., Nakagawa, Y. and Waku, K. (199l) Lipids 26, 974-978. [14] Bligh, E.G. and Dyer, W.J. (1959) Can. J. Biochem. Physiol. 37, 911-917. [15] Rouser, G., Siakotos, A.N. and Fleischer, S. (1965) Lipids 1, 85-86. [16] Sugiura, T., Sekiguchi, N., Nakajima, M., Nakagawa, Y. and Waku, K. (1983) Lipids 18, 125-129. [17] Sugiura, T., Fukuda, T., Masuzawa, Y. and Waku, K. (1990) Biochim. Biophys. Acta 1047, 223-232. [18] Lowry, O.H., Rosebrough, N.J., Farr, A.L. and Randall, R.J. (1951) J. Biol, Chem. 193, 265-275. [19] Woodard, D.S., Lee, T-C. and Snyder, F. (1987) J. Biol. Chem. 262, 2520-2527. [20] Blank, M.L., Lee, T-C,, Fitzgerald, V. and Snyder, F. (1981) J. Biol. Chem. 256, 175-178. [21] Tokumura, A., Takauchi, K., Asai, T., Kamiyasu, K., Ogawa, T. and Tsukatani, H. (1989) J. Lipid Res. 30, 219-224. [22] Tanaka, T., Minarnino, H., Unezaki, S., Tsukatani, H. and Tokumura, A. (1993) Biochim. Biophys. Acta 1166, 264-274. [23] Sugita, M., Fujii, H., lnagaki, F., Suzuki, M., Hayata, C. and Hori. T. (1992) J. Biol. Chem. 267, 22595-22598. [24] Noda, N., Tanaka, R., Miyahara, K. and Kawasaki, T. (1993) Biochim. Biophys. Acta 1169, 30-38. [25] Lekka, M., Tselepis, A.D. and Tsoukatos, D. (1986) FEBS Lett. 208, 52-55. [26] Nakayama, R., Kumagai, H. and Saito, K. (1994) Biochim. Biophys. Acta 1199, 137-142. [27] Bossant, M.J., Farinott, R,, de Maack, F., Mahuzier, G., Benveniste, J. and Ninio, E. (1989) Lipids 24, 121-124. [28] Tokumura, A., Kamiyasu, K., Takauchi, K. and Tsukatani, H. (1987) Biochem. Biophys. Res. Commun. 145, 415-425. [29] Tanaka, T., limori, M., Tsukatani, H. and Tokumura, A. (1994) Biochim. Biophys. Acta 1210, 202-208. [30] Wykle, R.L., Malone, R. and Snyder, F. (1980) J. Biol. Chem. 255, 10256-10260. [31] Ninio, E., Mencia-Huerta, LM., Heymans, F. and Benveniste, J. (1982) Biochim. Biophys. Acta 710, 23-31.