Biochimico et Biophysics Acta 922 (1987) 85-94

Elsevier

85

BBA 52661

Lipid metabolism in various regions of squid giant nerve fiber

Takashi Tanaka a, Hisao Yamaguchi ap*, Yasuo Kishimoto a and Robert M. Gould b

a John F. Kennedy Institute, Department of NeurolopJ Johns Hopkins University School of Medicine, Baltimore, MD

and b Department of Patholo%~caI Biochemistry, institute of Basic Research for Deuelopmental Disabiittieu, Staten Island: NY (U.S.A.)

(Received 15 May 1987)

Key words: Giant nerve fiber; Stellate ganglion; Lipid metabolism; (Squid giant axon)

The purpose of this investigation was to compare the incorporation of radioactivity from various precursors into lipids of different regions of squid giant nerve fiber systems including axoplasm, axon sheath, giant fiber lobes which contain stellate ganglion cell bodies, and the remaining ganglion including giant synapses. To identify the labeled lipids, stellate ganglia including giant fiber lobes and the remaining tissue were first incubated separately with [ ‘*C]glucose, [32P]phosphate, [‘*C]serine, [ ‘*C]acetate and 13H]myristate. The ra~oa~~vi~ from glucose, after conversion to glycerol and fatty acids, was incorporated into most lipids, including triacylglycerol, free fatty acids, cardiolipin, phosphatidylethanolamine, phosphatidylglycerol, phos- phatidylcholine, phosphati~linositol, phosphatidylserin~ sphingomyelin and eeramide 2-amin~thylphos- phate. The ra~oa~tivi~ from serine was largely incorporated into phosphatidylserine and, to a lesser extent, into other phospholipids, mainly as the base component. The sphingoid bases of ceramide and sphingomyelin were also significantly labeled. Saturated and monounsaturated and, to a lesser extent, polyunsaturated fatty acids of these lipids were synthesized from acetate, glucose and myristate. Among the major lipids, cholesterol was not labeled by any of the radioactive compounds used. Ganglion residues incorporated the most radioactivity in total lipids from either [“*C]glucose or (*4C]serine, followed by giant fiber lobes and then sheath. Axoplasm in~o~rated the least. Among various lipids, ph~phati~le~anola~ne with shorter saturated fatty acids and phosphatidylglycerol contained the most radioactivity from glucose in all regions. Axoplasm was chara&erized by a higher proportion of glucose radioactivity in ceramide, sp~ngomyelin and phosphatidyl~y~erol. Axoplasm and sheath contained a higher proportion of serine radioactivity than did the other two regions in ceramide. Essentially no radioactivity from [ ‘*C]galactose was incorporated in any region.

Introduction

The size of the squid giant axon has made the squid a desirable animal with which to study the physiological basis of nerve impulse propagation

[1,2]. The axon, formed from a syncytium of neu- rons in the stellate ganglion, is surrounded by a thin coat of Schwann cells. Recently, the giant axon and extruded axoplasm preparations have been used in studies of axonal transport. There is,

* Present Address: Department of Physiology, University of Correspondence: Y. Kishimoto, Kennedy Institute, 707 North Tokushima, School of Medicine, Tokushima 770, Japan. Broadway, Baltimore, MD 21205, U.S.A.

OHS-2760/87/SO3.50 0 1987 Elsevier Science Publishers B.V. (Biomedical Division)

86

I-, a paucity sf h ab5ut tne bio-

is of its fnmcticsw, we have corn-

87

tion according to Folch et al. [S]. The upper layer was treated with 200 mg of Dowex 5OW-X2 to remove alkali. The product was then analyzed by two-dimensional TLC on a plate coated with Avicel. Mixtures of phenol/water (100 : 38, w/v) and l-but~ol/propio~c acid/water (15 : 7 : 10, v/v) were used as developing solvents for the first and second direction, respectively.

Analysis of fatty acids. Radioactive fatty acids of individual glycerophospholipids were converted to methyl esters by mild alkaline methanolysis as previously described f9]. The methyl esters were purified on preparative TLC plates as described previously [lo], and separated according to the degree of unsaturation on a silver nitrate-impreg- nated silica gel G TLC plate (argentation-TLC). For impregnation, the plates were soaked for 3 min in 5% silver nitrate dissolved in methanol/ water (4 : 1) and then reactivated by heating at 110 o C for 1 h. After fatty acid methyl esters were spotted, the plate was developed with a mixture of hexane/ether (4: 6, v/v, to separate pentaenes from hexanes or 9 : 1, v/v to separate monoenes from saturated derivatives). Developed plates were exposed to X-ray film as described above. After autoradiography, the plates were sprayed with 8- anilino-1-naphthalenesulfonic acid. The radioac- tive bands, identified by co-migration with stan- dard fatty acid methyl esters which were included on the sides of the plate, were scraped off, and packed in a small column. The materials in each band were eluted with chloroform/methanol (2 : 1) and the radioactivity was counted after evapora- tion of the solvent.

Further fractionation of saturated methyl esters according to the chain length was carried out by reverse-phase HPLC as described by Aveldano et al. [IO]. The methyl esters eluted from the argenta- tion-TLC scraping were dissolved in a small volume of hexane/benzene (4 : 1) and transferred to a column containing 0.2 g of Unisil. Fatty acid methyl esters were then eluted with 5 ml of the same solvent. A model M45 pump (Waters Assoc.) and model SF 770 Spectroflow monitor (Kratos) were used with a Spherisorb C,, 5 p column obtained from Spectraphysics. After injection of the purified methyl esters, the HPLC column was eluted with 100% acetonitrile at a flow rate of 1 ml/min. The peaks were detected by monitoring

the absorbance at 210 nm. The retention time of each peak was compared with those of the stan- dard fatty acid methyl esters.

The polyunsaturated fatty acid methyl esters, separated by the above-mentioned argentation- TLC, were hydrogenate catalytically as described previously [ll]_ Because of the alkaline medium used in this procedure, the reaction product was acidified with HCl before extraction of the prod- ucts with hexanes. After evaporating the solvent, the product was esterified with diazomethane as described previously 1121 and reexamined with the argentation-TLC as described above.

Analysis of cholesterol. A small amount of labeled total lipids was mixed with 1.0 mg of authentic cholesterol and the mixture was frac- tionated by preparative TLC on silica gel G plates with ether/ benzene/ ethanol/ acetic acid (40 : 50: 2: 0.2) as a developing solvent. The band of cholesterol revealed by bromthymol blue spray was extracted with ether. The extract was bromi- nated as described by Grob and Winstein 1131. The product, after removal of solvents and excess bromine by nitrogen evaporation, was fractionated by preparative TLC similar to the system described above (except that the ratio of developing solvents was 35 : 55 : 2 : 0.2). Two bands were made visible by bromthymol blue spray; the major one corre- sponded to dibromochol~terol. The material recovered from the dibromocholesterol band was mixed with 21.6 mg of authentic dibromocholes- terol and the mixture was recrystallized five times from ethanol; specific activities of each batch of crystals as well as the material recovered from mother liquor were measured.

Results

Identification of radioactiuely labeled lipid-r To compare the metabo~sm of lipids in various

regions of the giant nerve fiber system, as described below, firm identification of labeled lipids was necessary. Therefore, several stellate ganglia, which included ceil bodies and synapses, from squid were incubated with several lipid precursors and labeled lipids, and their component fatty acids were identified.

Phospholipids. The ganglia were first incubated with [14C]glucose and the total lipids obtained

Fig. 1. TLC autoradiogram of tota lipids from squid ganglion tissue which was incubated with either [U-‘4CJgluco~e {AR), [W-‘4C]serine (B) or potassium [32PJphosphate (C). The spots made visible by exposure to I, vapor are circled by solid lines and

those sot visible but detected by radioactivity are circled by dotted lines. TG, triacylgiycerols; CE, ccrantide; CH, &oiesteroE; CL,

cardiolipin; WA, $ee fatty acids; PE, phosphatidylethanolami3le; UC, uncharactesized; PC, ~hos~~at~~y~c~o~i~e; PS, phospha\i-

89

glion phospholipid. Fourth, deacylated products of both phosphatidylglycerol preparations co- migrated with authentic glycerylphosphoglycerol.

In addition to the main radioactive lipids shown in Fig. 1A and C, we found that ceramide 2aminoethylphosphonate (a phosphonolipid) was also labeled from both [ l4 C]glucose and [ 32 Plphos- phate. The spot of ceramide 2-ethylaminophos- phonate overlapped with that of phosphatidyl- choline in the TLC system and became identifia- ble only after mild alkaline methanolysis of total lipids (data not shown). Its presence in ganglion lipids was verified by the following criteria. First, it was unaffected by mild alkaline methanolysis. Second, the radioactive compound co-migrated on two-dimensional TLC with the authentic ceramide 2-aminoethylphosphonate from a bivalve [14]. Judging from TLC autoradiogram of the mild alkali-treated radioactive lipids, the amount of radioactivity incorporated in the phospholipid spot was nearly equal to that of sphingomyelin. Al- though it was not detectable in the ganglion lipids labeled with [‘4C]glucose or [32P]phosphate, ceramide, the precursor of both ceramide 2- aminoethylphosphonate and sphingomyelin [15, 161, was relatively highly labeled when the ganglia were incubated with [‘4C]serine as shown in Fig. 1B.

Fatty acids. About 50% of the radioactivity in phosphatidylethanolamine which was derived from glucose was found in fatty acid moieties. Simi- larly, 30%, 40%, 40%, 15% and 20% of the radioac- tivity in phosphatidylglycerol, phosphatidylserine, phosphatidylcholine, phosphatidylinositol and cardiolipin, respectively, were found in the fatty acid moiety. In sphingomyelin and triacylglycerol about 30% and lo%, respectively, of the radioac- tivity was found in fatty acids. The labeling of fatty acyl moieties of phospholipids was further studied by incubating the squid ganglion with [14C]acetate and [3H]myristate. The total lipids obtained were. subjected to mild alkaline metha- nolysis and the fatty acid methyl esters were char- acterized by argentation-TLC followed by autora- diography. As shown in Table I, about 60% of the radioactivity in the [ l4 Clacetate-labeled total lipids was recovered in the methyl esters. Over half the radioactivity was present in saturated fatty acids, about one-third of the radioactivity was contained

TABLE I

SYNTHESIS OF FATTY ACIDS IN SQUID NERVE GAN-

GLION

Squid ganglia were incubated with a radioactive precursor.

Total lipids obtained were methanolyzed and the fatty acid

methyl esters obtained were fractionated by argentation-TLC.

The saturated methyl esters were further fractionated by re-

verse-phase HPLC.

Fractions [1-r4C]Acetate [9,10-‘HJMyristate

Total lipid dpm 274 900 4 145 400

Fatty acid methyl

esters dpm 163 300 2964000

Argentation TLC (W)

saturated 52.0 84.2

monoene 31.2 13.3

diene 5.2 0.1

polyene 11.3 0.3

Reverse-phase HPLC (5% of saturated)

12:o 0.2 0.1

14:o 3.5 71.3

16:0 46.7 25.0

17:o 1.2 0.1

l8:O 47.9 3.5 19:o 0.5 a

a Not detected.

in monounsaturated acids, and the rest was found in polyunsaturated acids. The radioactive material eluted from the spot of saturated methyl esters was further fractionated by reverse-phase HPLC. Most of this radioactivity was present in palmitic (47%) and stearic acids (48%). To confirm the association of small amounts of radioactivity with the polyenes, the material eluted from the spots of polyenes was hydrogenated and the product was reexamined with argentation-TLC. The autoradio- gram of the plate (not shown) indicated that all radioactivity was now included in the spot of saturated methyl esters.

When [9,10-3H]myristate was used as a pre- cursor, over 70% of radioactivity in total lipids was recovered as fatty acid methyl esters. Of the labeled fatty acids, about 85% were saturated fatty acids, including over 70% in unmodified my&ate and 25% in pahnitate. Most of the remaining radioactivity was found in stearate. About 15% of the radioactivity was incorporated to monoenoic fatty acids, but very little was detected in poly- enes.

2. Glucose (5 hj Total extracts (A) F&h lower layer (B)

3. Setine Totd extract (A) Folch lower layer (B)

4. Gdactose Total extracts (A) Folch &x=.w layer {ES)

3179 6.9

210 175 1999 2.406

4.1 2.9 6.6 3.8 9Q.2 4.5 95.3 4x

762

0.2 115

1317 1.4 329

3137 25.0 480 15.3

25% 563 22.2

total lipids in Folch lower layer increased 4.3-, 2.9-, 2.5- and 1.9-fold in axoplasm, sheath, giant fiber lobes and ganglion residue, respectively. Un- like glucose, [i4C]galactose was not incorporated into lipids (Expt. 4).

Extracted lipids from the [r4C]glucose labeled regions were separated by two-dimensional TLC and the radioactivity incorporated into the indi- vidual lipids was detected by autoradiography. The radioactive spots were extracted and the ra- dioactivity was measured (Table III). Phos- phatidylethanolamine was the main phospholipid formed from [14C]glucose in all giant fiber compo- nents except axoplasm. The sheath had the highest proportion of radioactivity in phosphatidyleth- anolamine, with giant fiber lobe and ganglion residue following close behind. Radioactive phos- phatidylethanolamine with mostly shorter and saturated chain fatty acids, and constituting only a small portion of the total molecular species [24] contained about two-thirds of the radioactivity incorporated into this lipid in all areas, including axoplasm.

squid nerves, contained a significant portion of lipid radioactivity in all giant fiber tissues. In fact, of the known labeled glycerophospholipids in axoplasm, phosphatidylglycerol contained the most radioactivity. Phosphatidylserine was a major curse-labeled lipid in giant fiber lobe, although not in other regions. Other ~ycerophospholipids, including phosphatidylcho~ne, phosphatidylinosi- to1 and cardiolipin, all contained moderate radio- activity in all tissue components.

An unusually high proportion of radioactivity was observed in sphingomyelin in axoplasm. Other regions incorporated far lower portions of glucose radioactivity into this lipid. Ceramide 2-aminoeth- ylphosphonate overlapped with phosphatidyl- choline with our two-dimensional TLC system. However, alkaline treatment of radioactive lipids from labeled giant fiber lobe showed that this phosphonolipid radioactivity was equal to that of sphingomyelin. Ceramide, the precursor of these two sphingolipids, although not visualized after TLC separation, contained significant radioactiv- ity in all giant fiber regions.

Phosphatidylglycerol, a minor phospho~pid of T~acylglycerol contained high radioactivity,

91

TABLE III

DISTRIBUTION OF RADIOACTIVITY AMONG VARIOUS LIPID CLASSES IN SQUID NERVES

Portions of total lipid extracts, after the Folch wash, were mixed with 200 pg of total lipids from squid brain and fractionated by

two-dimensional TLC. The plates were subjected to autoradiography and then exposed to iodine vapor. The radioactive and

nonradioactive spots were then scraped off and their radioactivities were counted. Results are presented as percentage of radioactivity

obtained in indi~d~ lipids.

Precursors

Triacylglycerol

Cholesterol Ceramide

Cardiolipin Free fatty acids

Shorter chain

Longer chain Phosphatidylglycerol

Phosphatidylchohue Phosphatidylseriue Phosphatidylinositol

Sphingomyehu Uncharacterized

Origin Other polar lipids

Axoplasm Sheath Giant fiber lobes Ganglion residues

Glc Ser GIG Ser Glc Ser GIC Ser

(W @) W (5%) W (W W (W

1.5 0.0 8.2 1.0 17.6 1.7 28.9 0.9

0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

1.0 15.7 1.0 24.9 0.5 4.8 0.2 6.7

0.6 0.0 1.7 0.1 2.0 0.2 1.4 0.0

0.3 2.3 3.4 2.2 0.8 1.4 0.0 0.2

4.8 0.5 16.1 5.5 14.7 7.8 8.3 6.1

10.2 0.0 32.5 0.0 22.6 0.0 19.2 0.0

4.8 0.0 12.9 0.0 7.3 0.0 7.2 0.0

1.8 1.3 3.5 2.5 5.4 1.6 3.6 3.3

1.1 38.1 0.5 16.0 10.4 59.1 0.4 57.3

1.2 1.6 3.1 2.1 2.9 2.5 3.5 2.0

5.3 2.8 1.4 1.2 1.3 2.1 0.4 0.5

4.2 2.5 0.0 1.4 0.2 I..5 10.5 1.2

72.3 13.8 28.0 14.7 21.3 9.4 23.4 6.8

2.8 21.6 2.7 31.0 1.7 9.5 3.3 16.5

92

especially in giant fiber lobe and ganglion reske. Axoplasm contained the lowest proportion of ra- dioactivity in this lipid, followed by sheath Cholesterol, one of the major lipids in the tissues, was not labeled in any region (see above). As described previously, free fatty acids, consistently detected in these tissues [24], were found always to contain radioactivity.

Several unidentified polar compounds (origin, uncharacterized and others in Fig. 1) contained a significant portion of radioactivity in all tissue components, especially axoplasm. After methanol- ysis, however, virtually no radioactivity was found in fatty acid fractions, indicating that the radioac- tivity is probably nonlipid.

[“4ClSerine incorporation. More [r4C]serine ra- dioactivity was incorporated into lipids by giant fiber lobe and ganglion residue than by other regions, The ratio of radioactivity incorporated into lipid to that present in tissue (B/A in Table II) was higher than that of [14C]glucose incorpora- tion in axoplasm, but lower in other components. In contrast to glucose, the [U-r4C]serine radioac- tivity was incorporated into lipids in ti highly selective manner (Table III). As might be ex- pected, phosphatidylserine was the most heavily labeled lipid in all regions of the giant fiber. Proportionally, giant fiber lobe had the most, fol- lowed by ganglion residue and then axoplasm. The sheath contained the smallest amount of ra- dioactivity in this lipid. 92% of the radioactivity was detected in unmodified serine. Ceramide, al- though a minor lipid of these tissues, contained a considerable portion of serine radioactivity, par- ticulariy in the sheath and axoplasm. Methanoly- sis followed by TLC examination indicated that over 90% of the radioactivity was in the sphingoid base. Small portions of radioactivity were found in other phospholipids as well as triacylglycerol. Less than 10% of the radioactivity in these lipids was present in the fatty acid moiety.

The giant axons with their surrounding Schwann cell-rich sheaths emanate from the stel- late ganglion, which can be readily separated into a cell body-rich giant fiber lobe and a more het- erogeneous residue containing giant synapses de-

rived from the paliial nerve. These component are physically and functionally tightly integ~ate6 and form initial segments of the giant axons, !?.7’j. To understand the metabolic relationship among these components, we have used the giant nerve fiber to study the incorporation of radioactive precursors, including I4 C-labeled glucose and serine, into lipids. Similar to the mammalian sys- tem, radioactive ghx~sc was iccorporated into a wide variety of lipids, whereas se&e was incoqm- rated into lipids more selectively. Considerable differences in the labeling patterns among @e giant fiber compartments were found, perhaps due to the functional and metabolic role of each eom- partment. To help characterize lipids which afe present in the nerve in very small amounts and to gain insight into the metabolic relationship among these lipids, incorporation of other radioactive lipid precursors, such as [r4C]acetate, [‘*C]my& state and [32P]phosphate into these lipids was also studied. It should be remembered, however, that the differences in the isotope u@take and endoge- nous dilution rendered the-absolute radioactivities difficult to compare.

Ceramide, phosphatidylethanolamine with shorter and saturated fatty acids and phobphati- dylglyeerol were present in such small amounts in these fiber regions &&at they remained undetecta- ble when TLC plates were stained for lipid. Nev- ertheless, they incorporated considerable radioac- tivity during the incubation. This observation in- dicated that these lipids are actively metabolized and may play an important role in lipid metabo lism and function in squid giant nerve fiber. High levels of glucose radioactivity were pmedt in phosphatidylglycerol and sphingomyelin in, axo- plasm. How the high metabolic turnover of these lipids may be linked to axoplasm function, nervti impulse propagation, and axonal transport war- rants further investigation.

The radioactivity in ceramide was highest. in sheath membranes (probably in Schknn cells) and in axonal membrane. This observation indi- cates a possible significance for sphingoli@s’in these membranes. The high degree of labeling ,of ceramide from radioactive Serine is in line with the fact that, as in mammalian .and bacterial systems 1151, the sphingoid base of ceramide in .s@tid nervous tissue is derived from serine.’ Since squid

93

nerve contains only a trace of glucocerebroside and rather large amounts of sphingomyelin and ceramide 2-aminoethylphosphonate, the ceramide must be used primarily for the synthesis of the latter two compounds. The highly labeled cera- mide in axoplasm possibly correlates with the similar high radioactivity of sphingomyelin in this region.

Cholesterol is ubiquitously distributed in all eukaryotic membranes and is believed to be indis- pensable in membrane structure and function [18]. Despite its importance and ubiquity, however, cholesterol is not synthesized in some species of invertebrates 119,201. For example, all arthropods examined do not synthesize cholesterol. As re- viewed by Teshima and Kanazawa [20], only about half of the examined Mollusca species, the phylum to which squid belongs, were shown to synthesize cholesterol. In this investigation, neither glucose nor acetate was incorporated into cholesterol by components of the giant fiber system. This indi- cates that squid probably obtains all of its cholesterol exogenously, a point to be considered for the aquaculture of this important edible marine animal.

De novo synthesis of saturated and monoun- saturated fatty acids from 14C-labeled glucose and acetate, as well as chain elongation of my&tic acid by squid giant nerve fiber was demonstrated in this investigation. Synthesis of polyunsaturated fatty acids in marine animals, especially fish, has been disputed, mainly because of the resemblance between the fatty acid compositions of many fish species and those of their respective foods [21]. In contrast to our previous finding that polyun- saturated fatty acids were absent in nerves of shrimp 1221 and earthworm [9], we found squid brain to be rich in these fatty acids (see preceding paper [24]). Since squid is the only one of these species known to eat fish we suspected that the squid nerve lacked the ability to synthesize poly- unsaturated fatty acids. In this investigation, how- ever, we have demonstrated unequiv~~ly that pol~nsaturat~ fatty acids are synthesized from acetate by squid steIlate ganglia. Since in the accompanying paper [24] Yamaguchi et al. showed that all polyunsaturated fatty acids in squid brain are C,, and C,, and are in n - 3 homologs, future studies are planned to determine whether squid

nerves, like mammalian tissues, require precursor polyenes, such as linolenic acid, for the synthesis of polyunsaturated fatty acids.

A portion of this investigation was performed in the laboratory of Dr. Nicholas Ingoglia at Woods Hole Marine Biological Laboratories. We are grateful for his support and helpful discus- sions. We also thank Drs. Taro Hori and Mutsumi Sugita for their generous gift of ceramide 2- a~n~thylphosphonate and Dr. Pamela Talalay and Janice White for assistance in preparation of this manuscript. This investigation was supported by grant BNS8314337 from the National Science Foundation and NS13559, NS13569 and NS13980 from the National Institutes of Health, U.S. Pub- lic Health Service.

References

1 Kuffler, SW., Nicholds, J.G. and Martin, A.R. (1984)

From Neuron to Brain, 2nd Edn., Sinauer Associates,

Sunderland 2 Baker, P.L. (1984) Squid Axon, Academic Press, New York

3 Gould, R.M., Pant, H., Gainer, H. and Tytell, M. (1983) J.

Neurochem. 40,1293-1299

4 Bnmetti, M. (1979) J. Neurochem. 32, 319-324

5 Folch, J., Lees, M. and Sloane-Stanley, G.H. (1957) J. Biol.

Chem. 226,497-509

6 Okamura, N., Stoskopf, M., Yamaguchi, H. and Kishimoto,

Y. (1985) J. Neurochem. 45, 1875-1879

7 Benson, A.A. and Maruo, B. (1958) Biochim. Biophys. Acta

27, 189-195

8 Kates, M. and Volcani, B.E. (1966) Biochim. Biophys. Acta

116,264278

9 Kishimoto, Y., Davies, W.E. and Radin, N.S. (1965) J.

Lipid Rex 6, 525-531

10 Aveldano, M.I., VanRoBins, M. and Horrocks, L.A. (1983)

J. Lipid Res. 24, 83-93

11 Kishimoto, Y. and Radin, N.S. (1963) J. Lipid Res. 4,

444-441 12 Kisbimoto, Y. and Hoshi, M. (1972) in Methods of Neuro-

chemistry, Vol. 3 (Fried, R., ed.), pp. 75-154, Marcel Dekker, New York

13 Grob, C.A. and Wins~~, S. (19.52) Helv. Chim. Acta 35,

782-802 14 Hori, T., Arakawa, I. and Sugita, M. (1967) 3. B&hem. 62,

67-70 15 Kisbimoto, Y. (1983) in The Enzymes Vol. 16 (Boyer, P.,

ed.), pp. 357-407, Academic Press, New York 16 Ho& T. and Sugita, M. (1984) in Biochemistry of Natural

C-P Compounds (Hori, T., Horiguchi, M. and Hayashi, A.,

94