Thompson, Ann. N.Y. Acad. Scd. 119, 347(1964).

44.- , private communication of work tobe published.

45. E. L. Fireman, Smithsonian Astrophys. Os.Contr. to Astrophys. 11, 373 (1967).

46. C. M. Merrihue, Ann. N.Y. Acad. Sd. 119,351 (1964).

47. D. Tinles, Smithsonian Astrophys. Obs. Contr.to Astrophys. 11, 399 (1967).

48. , Science 153, 981 (1966).49. R. H. Bieri, private communication.50. K. I. Mayne, Geochim. Cosmochim. Acta 9,

174 (1956).51. D. Tilles, Science 148, 1085 (1965).52. J. H. Reynolds and G. Turner, J. Geophys.

Res. 69, 3263 (1964).53. P. Signer and H. E. Suess, In Earth Science

and Meteorites (North-Holland, Amsterdam,1963), p. 241; H. E. Suess, H. Wanke, F.Wlotzka, Geochim. Cosmochim. Acta 28,595 (1964).

54. G. K. Wehner, C. Kenknight, D. L. Rosen-burg, Plantetary Space Sci. 11, 885 (1963);E. Anders and E. Mazor, Geochim. Cos-mochim. Acta, in press; D. Tilles, Trans.Am. Geophys. Union 48, 157 (1967).

55. H. Pettersson and H. Rotschi, Geochim.Cosmochim. Acta 2, 81 (1952).

56. B. Bonatti and G. R. Nayudu, Am. J. Sci.263, 17 (1965); E. Bonatti, in Researches In

Geochemistry, P. Abelson, Ed. (Wiley, NewYork, 1967), voL 2.

57. E. D. Goldberg, J. Geol. 62, 249 (1954).58. J. Brocas and B. Picciotto, J. Geophys. Res.

72, 2229 (1967).59. J. L. Barker, Jr., private communication.

This method was also suggested by P. A.Baedecker and W. D. Ehman, Geochim.Cosmochim. Acta 29, 329 (1965).

60. G. S. Hawkins, Ann. Review Astron. Astro-phys. 2, 149 (1964).

61. C. S. Nilsson and R. B. Southworth, privatecommunication.

62. R. E. McCrosky and H. Boeschenstein, Jr.,Smithsonian Astrophys. Obs. Spec. Rept.No. 173 (1965).

63. H. Brown, Jr., J. Geophys. Res. 65, 1679(1960).

64. E. Shoemaker and C. S. Lowery, Meteoritics3, 123 (1967).

65. M. F. Ingham, Monthly Notices Roy. Astron.Soc. 122, 157 (1961); R. H. Giese, SpaceSci. Reviews 1, 589 (1962-1963); J. L. Wein-berg, Ann. Astrophys. 27, 718 (1964); R. D.Wolstencroft and L. J. Rose, Astrophys. J.147, 271 (1967).

66. R. B. Southworth, Ann. N.Y. Acad. Sci. 119,54 (1964); M. I. S. Belton, Proc. ZodiacalLight and Interplanetary Medium Cont., Hon-olulu (1967).

67. G. S. Hawkins, Astron. J. 65, 318 (1960).68. E. Anders, Icarus 4, 399 (1965).

69. F. L. Whipple, Smithsonian Astrophys. Obs.Contr. to Astrophys. 7, 239 (1963).

70. N. B. Richter, The Nature of Comets(Methuen, London, 1963).

71. F. L. Whipple and E. L. Fireman, Nature183, 1351 (1959).

72. M. F. Comerford, Geochim. Cosmochim.Acta, in press.

73. F. L. Whipple, Proc. Roy. Soc. London Ser.A 296, 304 (1966).

74. D. Tilles, Icarus, in press.75. T. A. Mutch, Earth and Plan. Sci. Letters

1, 325 (1966).76. While this paper was being revised, three

reviews of parts, of the subject appeared [4,73, and J. F. Vedder, Space Sci. Reviews 6,365-414 (1966)]. We would particularly liketo thank J. R. Arnold and N. Bhandari fortheir interest and advice, and R. B. South-worth, R. E. McCrosky, A. F. Cook, II,U. B. Marvin, G. S. Hawkins, and R. H.McCorkell for helpful comments on themanuscript. The recent dust-collecting ex-pedition at Barbados was financed by ScienceResearch Council, U.K. D.W.P. acknowledgessupport from NASA (grant NSG-321). Wewould also like to thank J. L. Barker, Jr.,G. H. Megrue, R. H. Bieri, E. M. Shoe-maker, and S. 0. Thompson for providingdata in advance of publication. This reviewincludes results published or availablethrough the end of April 1967.

Poisonous Principles of Mushroomsof the Genus Amanita

Four-carbon amines acting on the central nervous

system and cell-destroying cyclic peptides are produced.

Theodor Wieland

T-he true poisonous mushrooms are

of the genus Amanita. The best knownof them, widely spread throughout theworld, except for the tropic zones, isundoubtedly the fly agaric, Amanitamuscaria (L. ex Fr.) Pers. ex Gray;its toxicity, however, is generally over-

estimated. Much more capable of in-citing fatal poisoning is the white mush-room A. verna (Bull. ex Fr.) Pers. ex

Vitt. sensu Fr. 1821 non al. (A. virosaLam. ex Secr.). Amanita tenuifolia (1),and A. bisporigera Atk. (2) are prob-ably closely related varieties. In Cen-tral Europe, the green death-cap A.phalloides (Vaill. ex Fr.) Secr., thegrine Knollenbldtterpilz, is noted forits toxicity. The more frequently oc-

curring yellow mushroom A. citrina(Schaeff.) Gray [A. mappa (Batsch ex

Lasch) Qu6l.] definitely contains no

peptidic toxins-unlike A. phalloides inthis respect-but it does containbufotenine (5-hydroxy-N-dimethyltrypt-

946

amine) in relatively high concentra-tion (3) and some other indole amines(4). It is of some interest that bufotenineoccurs also in several other plants,and that other components of poison-ous amanitas have been found also inmushrooms of different Galerina spe-cies (5). I shall outline the present stateof knowledge of the components ofA. muscaria and give a summary ofthe chemistry and toxicology of thepoisonous peptides of A. phalloides.

Amanita muscaria (6), the fly agaric,grows from July until the end of au-

tumn, preferably under fir or birchtrees, as an egg-shaped white cap. With-in 1 to 2 days the cap spreads andbursts open the outer shell, the sur-

face then becoming brilliant red withthe residue which remains as whitespots regularly distributed over it. Thewhite veil protecting the lamellae inthe young mushroom now hangs at thestem as a cuticular ring. The stem,

which grows up to 25 centimeters, hasa cup at its bottom. The cap mayreach 25 centimeters in diameter; theweight of an average mushroom is 60to 70 grams. Different subspecies ofvarying colors, like the native NorthAmerican one that is orange yellow,have been described.The name "fly" agaric was derived

from an insect-killing property of itsextracts; however, this characteristic isinconspicuous compared to modern in-secticides and can only be observed un-der special conditions. Wasson (7) in-terprets the "flies" as a symbol for thedemonic power of the mushroom,which has been used in Siberia as ahallucinogen because of its psychotrop-ic principles. The symptoms of intox-ication from the use of A. muscaria arevery complex and resemble those ofdrugs that act on the central nervoussystem, but with the addition of asso-ciated peripheral phenomena attributa-ble to muscarine.

Problem of Muscarine

The search for the toxic principleof Amanita muscaria started over 100years ago and led Schmiedeberg andKoppe in 1869 to a substance that ex-cites the parasympathetic nervous sys-tem. This substance they named "mus-carin" (8, 9). The sensation causedamong pharmacologists by the first

The author is director of the chemistry de-partment at the Max-Planck-Institute for MedicalResearch, Heidelberg, German Federal Republic.

SCIENCE, VOL. 159

on

Aug

ust 2

7, 2

014

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

on

Aug

ust 2

7, 2

014

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

on

Aug

ust 2

7, 2

014

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

on

Aug

ust 2

7, 2

014

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

on

Aug

ust 2

7, 2

014

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

on

Aug

ust 2

7, 2

014

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

on

Aug

ust 2

7, 2

014

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

totally impure drug substance was sur-passed only by the later discovery ofthe Vagtisstoff, acetylcholine (No. 5),by Loewi (10). Incidentally, acetylcho-line was also isolated from A. muscariaby Kogl et al. in 1957 (9). Attemptsto prepare muscarine (No. 1) in thepure state and to establish its structuralformula were successfully completedin 1954, with the crystallization ofmuscarine chloride and the determina-tion of its composition as C9H2002-N+C1- by Eugster and Waser (7); in1957 the same team and the Dutchgroup of Kogl and Jellinek (11) settledthe structure of the molecule mainlyby the use of x-ray crystallography.Accordingly, muscarine (compound 1)is an oxoheterocyclic quaternary salt, 2-methyl-3-hydroxy-5-trimethyl-ammoni-ummethyltetrahydrofuran chloride withS-configuration at position-2, R- atposition-3, and S- at position-5 of thering.The total synthesis of compound 1,

which succeeded immediately, disclosedthe whole area. In view of the threecenters of asymmetry, we anticipatefour diastereoisomeric racemates, com-pounds 1 to 4, that is, eight antipodes;nearly all have been obtained. I do notintend to enumerate here all the syn-thetic approaches; the first syntheses of(±)-muscarine and of its diastereo-isomers (12) may stand as examples.From the furan derivative (No. 7), ob-tained by condensation of ethylaceto-acetate with glucose, through a reactionsequence (shown at lower right), theimportant (+)-normuscarone (No. 8)was prepared. From compound 8, notonly normuscarine and, by quaterniza-tion, (+)-muscarine were available, butalso with the use of various reduc-ing agents the tertiary compounds ofthe salt compounds 2, 3, and 4 be-came accessible. The racemic muscarinewas resolved into the antipodes withthe (+)- and (-)-ditoluoyltartrates.The enantiomorphs of (±)-muscarine

differ greatly in potency, which is al-most exclusively associated with the(+)-isomer. This alkaloid exhibits ahighly specific activity at postganglionicparasympathetic effector sites. Its ef-fect as measured by the decrease ofblood pressure of anesthetized cats(minimum dose of the magnitude of10-8 gram per kilogram of bodyweight) or by the reduction of ampli-tude and rate of the beat of the per-fused frog heart, is nearly equal to thatof acetylcholine, the (-)-isomer beingabout 1000 times less active. The samerelation holds for (+)-muscarine, as1 MARCH 1968

+ ~~OH

H3C 3 C27N(CH3)3 H39HC~H~

HO

(I)

L(+)-Muscarine

H3C

HO NCH2-N(CH3)3

(4)Allomuscarine

(2)

Epimuscarine

H3C-Ce

(5)Acetylcholine

compared with the diastereomeric race-mates. No muscarine isomers have anysignificant nicotine-like action on skele-tal muscles. The closely related (+)-and (-)-muscarone iodides (6) haveeven greater activity in the blood-pres-sure test than acetylcholine, and theyexhibit strong nicotine-like activity onthe isolated rectus abdominalis of thefrog. Certain similarities of the struc-tures of acetylcholine (No. 5) andparticularly of ketone (No. 6) point toanalogous mechanism of physiologicalaction.

Further Search for Active Principles

To judge from results of numerouspharmacological experiments and inview of the extremely low concentra-tion of muscarine in A. muscaria(0.0002 percent of the fresh mush-room), it must be concluded that thefamous alkaloid cannot possibly be the

C2H502CNls Pb304 C2

CH3 0 [CH(OH)]3-CH20H

(3)

Epi-allomuscarine

H3CO CH2-N(C)3

(6)Muscarone

poisonous principle. Other mushrooms,particularly of the genus Inocybe, con-tain much larger amounts of muscarine;for example, Inocybe patouillardi (Bres.)s. lateraria Ricken contains 0.037 per-cent. These mushrooms cause muscarinetoxicoses which are distinctly differentfrom the toxic state produced by A.muscaria. In the latter case there aresymptoms of hallucinogenic and cen-trally acting substances. The 5-hydroxy-indole alkaloid, bufotenine, seemed toappear in paper chromatograms in mylaboratory (3), an observation not con-firmed by others; but other unidenti-fied indole bases are said to occurin A. muscaria. This is of great interestin connection with the psychotropic 4-hydroxyindole derivatives psilocybinand psylocin from the Mexican teon-anacatl (13) (Psilocybe, sp. Strophariacubensis). In view of the fact thatatropine and scopolamine in Amanitamuscaria of South African and Polishorigin was not also found in Dutch

ZH.502C + dimethylformamiEde.CH3 ° CHO + formic acid

(7) O.C2H502C N3C

1711 hydrazine .1,CH3 ° CH2-N(CH3)2 then HN02 CH3 ° CH2-N(CH3)2

Curtius degradation ain C6H5CH20H

C6H5CH20CONH2N HCI, 100°C

CH3 CH2-N(CH3)2

0 -LiAl H4__iAIH

H J§ CH2-N(CH3)2 H2/ NiC H Al(i-prop)3 _

(8)

Pt + H2H

CH °CN(CH3)2

.nor - (I)nor -(2)nor-(3) and nor-(4)

947

and German specimens (6), it is ap-parent that the investigation of thetoxic drugs has not yet come to asatisfactory end.

Chemically rather simple, centrallyactive crystalline substances, muscimol(No. 9), the fly killing ibotenic acid(No. 10), and the related muscazone(No. 11) have been discovered in thefly agaric and in other mushrooms.Compounds 9 and 10 strengthen thenarcotic power without being narcotics.Nevertheless, muscimol (No. 9), whicheasily results from ibotenic acid (No.10) by decarboxylation, may be con-sidered as an essential principle ofAmanita muscaria, merely on accountof the high concentration of 0.03 to0.1 percent of compound 10 in freshtissue.

¾V-.C2.H32~ HO .A *~OvCH2-NH3 X 05aCH-NH3

cO!(9)

OC-CH2I 1 2nH

7 3CO

(12)

(10)

Ihv

Oc'- H-N~H3o=4`N3L - H

H Co°

(11)

The isoxazole derivatives No. 9 andNo. 10 were isolated in 1960 andshortly afterward by Eugster and hiscolleagues and by other research groupsindependently. These derivatives wereisolated from several mushrooms, andthey were provided with several names.Consensus in nomenclature has nowbeen achieved (14). Muscimol (No. 9)stands for the former Pyroibotenslureand for agarine similarly obtained fromAmanita muscaria by English chemists(15). Ibotenic acid (No. 10) has alsobeen isolated from A. strobiliformis(Vitt.) Quel.-which is the equivalentof A. solitaria (Fr.) Quel.-and fromA. pantherina. Tricholomic acid (No.12), a compound with insecticidal prop-erties obtained from the mushroomTricholoma muscarium Kawamura(16), is the erythro-2,3-dihydro deriva-tive of ibotenic acid (No. 10).

HOy.j

CH2CL

HO

IBJtHCN(15)

Br

HO

-CH12CN 15CHCNib>e (16) NH2

Muscimol (9) --C02 Ibotenic acid (10) -h Muscazone(II)

948

One sequence out of several synthesesis represented (17) by the reaction ofcompound No. 13, 3-hydroxyl-5-chloro-methylisoxazol, with ammonia to givemuscimol (No. 9) in good yield. Thesame compound with cyanide forms thecyanomethylisoxazole (No. 14), whosea-position of the side chain is bromi-nated with N-bromosuccinimide in thelight. The bromonitrile (No. 15) isconverted to the aminonitrile (No. 16)with ammonia, and mild hydrolysisof the amino nitrile leads to ibotenicacid (No. 10); decarboxylation of com-pound 10 gives compound 9. An in-teresting photoreaction-comparable toa Lossen rearrangement of hydroxamicacid derivatives-converts the isoxazolsystem of compound 10 into the 2-(3H)-oxazolone ring of muscazone (No.17). Tricholomic acid (No. 12) and itsthreo-isomer have been synthesizedfrom erythro- and threo-3-hydroxyglu-tamic acid .(18). Both tricholomic acid(No. 12) and ibotenic acid (No. 10)were discovered in Japan as a resultof research conducted to elucidate thefly-killing property of the correspond-ing mushrooms. However, an acciden-tal finding, that these substances havegreat tastiness, their flavors being about20 times more intense than that ofsodium glutamate, was reported by theJapanese investigators. This effect andan additional synergism in taste withnucleotide seasonings make the newcompounds interesting with regard totheir possible use in food formulation.

Amanita phailoides and Related Species

The deadly poisonous, green mush-room Amanita phalloides (Vaill. ex Fr.)Secr., is sometimes called the "deathcap"; it grows in Central Europe fromJuly until the end of October and isassociated with deciduous trees; it growspreferably under beeches in loose foreststhat are not too dry. It develops froman egg-shaped white state within1 to 2 days to a height of about 10to 15 centimeters. The slightly vaultedcap reaches diameters up to 12 centi-meters; it is smooth, more or less deepolive-green, and often patterned withdarker, radially extending, filamentousstreaks. The lamellae are white, thestem, which sometimes shows palegreenish cross stripes, bears at itsupper part a big white cuff and endsin a large tuber which is surroundedby a leafed sheath. The toxic mush-room has no specific smell or taste,

in contrast to the related nontoxicA. citrina (Schaeff.) Gray. This yellowmushroom is easily discernible. by itsodor of raw potato. The white toxicvariety, A. verna (Bull. ex Fr.) Pers.ex Vitt., appears during the summer, indeciduous forests; in Europe its oc-currence is not common. It is the"destroying angel" or "deadly agaric"of the North American continent. Thewhite A. virosa (Lam. ex Secr.) is nowconsidered identical to A. verna. Itcontains the same toxic peptides thatA. phalloides does. The other varietiesin North America (1, 2), A. tenuifoliaand A. bisporigera, which are extremelyrich in toxins, have not yet been ob-served in Europe according to myknowledge.The white or green Amanita causes

more than 95 percent of fatal mush-room poisoning. These mushrooms aremistakenly confused with the deliciouschampignon Tricholoma equestre (thewhite forms with Agaricus campestris),which appears in similar places, butwhich bears from the beginning faintlypink, and later darker red, lamellae

Start

0.1-

.2-

.3-

.4-

.6-

.7-

.8-

.9D_- IT,Cf

Phallacidin O(2,5)

Phallisin0(2,5)

Phalloidin Q'I(21,.0)

Phalloin o-(1.5) k-Phallin BO(10)

8-Amanitin )

P -Amanitin(0.4)

Amanin (0;5)c Amanitin C

(1.0)

a -Amaniin(0,35) _

Y.Amanitin Q(0.2)

Amanullin O(not toxic)

Fig. 1. Descending paper chromatogramin a butanone, acetone, water system(30: 3: 5) of the identified ingredientsof Amanita phalloides. For clarity, thephallotoxins (left) are arranged separatelyfrom amatoxins (right). Size of spots is arough indication of the relative amounts;figure in parentheses is the miniimum lethaldose per kilogram of body weight of thewhite mouse.

SCIENCE, VOL. 159

.5

and has a typical anise-like smell. Thehigh percentage of fatalities occurs fromAmanita phalloides and related mush-rooms because the first symptoms ofintoxication-vomiting and diarrhea,which are not caused by the lethaltoxins-do not become apparent untilseveral hours after ingestion, and dur-ing this time the deadly toxins have al-ready reached the liver. We know withcertainty that the slow-acting amatoxinscause death by destroying the livercells. Their fatal action, however, startsshortly after the toxins reach the organand is irreversible.The exploration of the poisonous

substances of A. phalloides started atthe beginning of the last century. Asurvey was made in 1959 (19) and morerecently (20). In 1937, F. Lynen andU. Wieland (21) in the Munich Chemi-cal Laboratories of the Bavarian Statesucceeded in isolating a toxic materialin crystalline form which they called"phalloidin." Four years later, H. Wie-land and R. Hallermayer (22) obtainedan even more toxic component andnamed it "amanitin." Both these toxinsare representatives of two families ofpoisonous peptides. My collaboratorsand I have resolved amanitin into aneutral (a-amanitin) and an acidic (38-amanitin) toxin; and added to the poi-sonous substances already discovered-y-amanitin, 8-amanitin, c-amanitin, andamanin. This group is called the familyof amatoxins. The phallotoxins includephalloidin, phalloin, phallacidin, phal-lisin, and phallin B. The chromatogramof a variety of substances, includingthe nontoxic amanullin, are schematical-ly represented in Fig. 1.

Paper chromatography has provedvery useful for analysis of crude mush-room extracts and of the various frac-tions obtained during isolation work.After numerous experiments, a mixtureof butanone, acetone, and water(30: 3 : 5 by volume) was found toallow separation of nearly all of thephytotoxins (Fig. 1). The solvent mix-ture has been modified by addition ofsome n-butanol by Block et al. (23).In thin-layer analysis on silica gel G,mixtures of butanone and methanol(1: 1) or of n-butanol, acetic acid, andwater (4: 1: 1) were successful (2).The amatoxins (except amanin) givean immediate violet color with cinnamicaldehyde in an atmosphere of hydro-chloric acid gas; the phallotoxins andamanin slowly give a weaker bluecolor. With diazotized sulfanilic acid(Pauly reagent), the first group forms1 MARCH 1968

an intense red compound, whereas thephallotoxins give only a weak yellowcolor. The much greater sensitivity ofthe amanitins to color reagents maybe the reason why the spot of phal-loidin has not been observed in mush-room analyses of other laboratories.Here a column chromatographic separa-tion combined with ultraviolet spec-troscopy of the eluent seems indis-pensable (24). In the poisonous speciesthere are about 10 milligrams of phal-loidin, 8 milligrams of a-amanitin, 5milligrams of /8-amanitin, and 0.5 milli-grams of y-amanitin per 100 grams offresh tissue (corresponding to about 5grams of dry weight). Similar valuesfor a- and /3-amanitin have been re-ported by Tyler et al. for A. phalloidesand A. verna, but a higher content(up to 5 milligrams per gram of drymaterial) in A. bisporigera (2). Sincethe lethal dose of the amanitins islower than 0.1 milligram per kilogramof body weight for human beings, itis conceivable that the toxin contentof one mushroom weighing 50 grams(about 7 milligrams of amanitins) maybe sufficient to kill a man.

In cases of poisoning, a prominentfeature is always the long period oflatency between ingestion and the ap-pearance of the first symptoms. Usual-ly not until 10 to 24 hours have elapseddo abrupt violent emesis and diarrheabegin, and the illness sometimes culmi-nates in rapid death, with cholera-likemanifestations. These gastrointestinalsigns have nothing to do with the toxicpeptides. If this phase is overcome, atransient and false remission takes placeand gradually more pronounced signsof injury to parenchymatous organsappear, chiefly the liver. A hemolyticagent which is precipitated by organicsolvents and is very labile on heating,was obtained from Amanita phalloidesas early as 1891 by Kobert and wascalled "phallin." This substance is pre-sumably also destroyed in the gastroin-testinal tract and therefore is scarcelyresponsible for lethal poisoning.

Chemistry of the Toxic Peptides

Chemically, the toxins have somecharacteristics in common. All of themare cyclopeptides (with a molecularweight of about 1000) composed ofonly a few amino acids, some of whichdo not occur in proteins. Each cyclo-peptide contains a gamma-hydroxylatedamino acid and a sulfur atom derived

0.9

:~~~~1 ~ ~~-0.5

250 300Wavelength (mp)

Fig. 2. Qualitative ultraviolet absorptionspectra in water of amanin (curve 1) andphalloidin (curve 2).

from a cysteine residue which is con-nected to the indol nucleus of atryptophan side chain, thus dividingthe cycle into two rings. In the ultra-violet range, they exhibit intensive ab-sorption maxima near 300 nanometers(Figs. 2 and 3). One of the differencesbetween the two families lies in therapidity and mode of their toxic action.The phallotoxins act quickly and, athigher doses, death in animals oc-curs within 1 to 2 hours. In contrast,the action of the amatoxins is delayed

Fig. 3. Qualitative ultraviolet spectra inwater of a-amanitin in neutral solution(curve 1), and after addition of a fewdrops of very diluted NaOH (curve 2).

949

H H H2CR1I II

H3C-C co NHNH-CH-CO-NH-C--CH--C-CH2R2IHC~~~ ~ ~~IO

2

NH HCC/S H gHH2C

.-0 HII HC-R3

4i-CO---RS HN-CO-CH-NH--CO

H-C-OH

(17) R

& R% R& P-R4 Ra Phalloidin OH H CH3 CHR OHb Phalloin H H CHa CHo OHc Phallisin OH OH CHR CHR OHd Phallicidin OH H CH(CH8)2 CO2H OHe Phallin B (tentatively) H H CH2CJM5 CH8 H

so that with very high doses it is notpossible to reduce the time required forlethal action to less than 15 hours.The relative toxicity is, however, justthe opposite, a-amanitin being 10 to20 times more toxic than phalloidinand therefore the major poisonous con-stituent of deadly amanita.

Phallotoxins

The group of phallotoxins consistsof at least five members (Fig. 1), whosecommon cyclic heptapeptide skeleton isshown in formula 17.

After these toxins are heated with6N hydrochloric acid, six differentamino acids are liberated, and in theprocess the thioether bridge is splitinto L-/3-oxindolyl-3-alanine (No. 18)and L-cysteine (No. 19) ("nucleophilic"hydrolysis at the a-carbon).The characteristic ultraviolet spec-

trum of phallotoxins (Fig. 2) is dueto the thioether of tryptophan. Com-pounds of this type can be synthesizedby coupling /3-substituted indoles with

NH-CH-CO H2N-CH-CO2HH2C H2C

-CO-CH 1100,6 hrs.HN (18)

(17) Raney-Ni H02C-CH-CH2-SHNH-CH-CO

H2C +} ~~~~~~~(19)LH2C1;CO-CH-CH3

HN(20)

950

sulfenyl chlorides. The 2-thioether ob-tained from /8-indolylacetic acid andmethylsulfenyl chloride, CH3SCI, isidentical to phalloidin in the spectrum(curve 2) of Fig. 2. Curve 1 representsamanin, which is also a derivative oftryptophan, but-as a sulfoxide-it be-longs to the family of amatoxins.When treated with Raney-nickel in

boiling methanol, the thioether bridgeof the phallotoxins is opened hydro-genolytically, thus giving nontoxicdethio compounds (No. 20). The onlyamino acids common to all of the fivephallotoxins are L-alanine and the cou-pling product, called tryptathionine, of L-cysteine and L-tryptophan. Other aminoacids present in the different toxinsare: L-allohydroxyproline (No. 21) innearly all of them (No. 17, a to d)except in phallin B (17e) which con-tains instead proline; D-threonine (No.

H02C OH

(21)

C02H

H-C-NH2I

HO-C-H'I

CH3(22)

C02H

H-C-NH2

H-C-OH

C02H(23)

HOH2C HOH2C

H3C NH CL- HOH2 ° H CL-(24) *(26)

H3C(25)

22) in (No. 17, a-c and e); L-valine,and D-erythrohydroxyaspartic acid (No.23) in phallacidin (No. 1 7d); andphenylalanine in phallin B (No. 17e).

In all toxic amanita peptides, includ-ing the amatoxins, a y-hydroxylatedamino acid occurs. The y-hydroxy groupin the side chain causes a selectiveand mild splitting of the peptide bond(50 to 80 percent trifluoroacetic acidat 20°C, 2 hours) by its neighboringgroup effect through y-lactone forma-tion. The seco- compounds (No. 27) ob-tained by such a partial hydrolysis aretotally nontoxic like the dethio com-pounds (No. 20).From a hydrolyzate of phalloidin

(17a) the y-lactone-hydrochloride of onediastereoisomer of y,8-dihydroxyleu-cine has been isolated, to which weascribed the erythro structure (No. 24).This does not permit a conclusion tothe native configuration; during hydrol-ysis of the peptide, an inversion at they-carbon might have occurred. Underthe same conditions, phalloin (17b)yields the lactone of y-hydroxyleucine(25). This compound, among otheraminolactones, was also discovered re-cently in hydrolyzates of gelatin (25).In phallisin (No. 17c) a trihydroxylatedamino acid, y,a,8'-trihydroxyleucine-lactone (No. 26) could be disclosed asa result of periodate oxidation: asparticacid was formed from the open acidof compound 27 by an excessive glycol-splitting reaction (26).So it appears that all phallotoxins

contain -y-hydroxyamino acids derivedfrom L-leucine. As to the function ofthe -y-hydroxy group as a prerequisiteof toxicity, several experiments havebeen conducted to remove or to sub-stitute this group in the intact molecule.The following sequence of reactionsgave the decision: phalloidin (No. 17a)was oxidized by periodate to give thestill toxic "ketophalloidin" (No. 28). Thetoxic properties of compound 28 canstill be referred to a hydroxyl group incompound 29, which may be formed

IC¢H3CH2-COH

P symbol forbicyctus

-NH-CH-C~RNH-CH-CHR 28) reduction (30) jRaney-NiIA:: C -OH CH3 CH-ICQH X H2-l-OH a)C CH2QH2

(17 I NH2 H

(17) (27) (29) (31)

SCIENCE, VOL. 159

RI,H3C\ ICH-CH2R2

CHHN-CH-CO-NH-CH-CO-NH-CH2-CO

CHO/3

O.S R4 H-C-CH

H K III IC H

~~~~CH2I I I

OC-CH-NH-CO-CH-NH-CO-CH2-NH

H2C-COR3C32)

RL R2 %R8 R4a ca-Amanitin OH OH NH2 OHb 8-Amanitin OH OH OH OHc y-Amanitin H OH NH2 OHd Amanin OH OH OH He Amanullin H H NH2 OH

by enzymatic hydrogenation of the car-bonyl of structure 28 in the liver.Therefore, a dithiolane (No. 30) was

prepared from compound 28 (27) whichwas again toxic. Finally, the sulfuratoms of compound 30 could be sub-stituted by hydrogen without attackingthe thioether bridge by means ofRaney-nickel catalyst (28). The result-ing norphalloin (No. 32) proved as toxicas phalloidin, thus eliminating the pres-ence of a y-hydroxyl group as a re-

quirement for toxicity of the phallo-toxins.

Amatoxins

The group of amatoxins has at leastsix members, whose common cyclicoctapeptide skeleton is shown in formu-la No. 32. The minor compounds 8-and E-amanitins have been isolated ina pure state, but their exact structurehas not yet been elaborated. To thisgroup also belong the "outsiders"amanin (No. 32d) and amanullin (No.32e). The former differs from,-amani-tin (No. 32b) only in its lacking of a

phenolic hydroxyl 'group in position-6of the indole nucleus. Therefore, thecolor reactions of amanin resemblethese of the phallotoxins, but the toxi-cological (slow) action is that of theamatoxins.

a-Amanitin (No. 32a) is the amideof the carboxylic acid f8-amanitin (No.32b). During acid hydrolysis the aminoacids not concerned in the thioetherbridge are liberated in the originalstate: glycine (2 moles), L-asparticacid, L-hydroxyproline, L-isoleucine1 MARCH 1968

and, varying among the members, L--y-hydroxyisoleucine as lactone (No.33) from y-amanitin (No. 32c) andy,8-dihydroxyisoleucine (No. 34) froma- or ,B-amanitin (No. 32, a and b)and from amanin (No. 32d). Thus thehydroxylated amino acids of amatoxinsare derived from isoleucine.

Amanullin (No. 32e) is totally non-toxic. It contains no y-hydroxylated sidechain, 'but rather a second molecule ofisoleucine (28a). In contrast to thephallotoxins, the presence of a .y-hydroxygroup is apparently a prerequisite fortoxic activity. The identification of thenatural -y-hydroxyisoleucine with one ofthe eight possible diastereoisomeric anti-podes as compound 33 has been madevery recently chiefly by nuclear mag-netic resonance spectroscopy (29). It issupposed that the corresponding y-, 8-dihydroxyamino acid of a-amanitin hasan analogous structure (No. 34). Like-wise the phallotoxins (No. 27), also

H3C CH3 NH3CL- HOH2CCH3kH3 Cl-

(33) (34)

NH-CH-CO H2N-CH-CO2HH 2

1 H H

Ghro d. productsCO-CH-NH ,

(32) 1 Raney-NI H ICH-H-SO2HNH-CH-CO NH2 (37)H 2 I disprop.

HO2C-CH--S03HCH3 1H NH2 (38)

CO-CH-NH (+ cystine)(35)

nontoxic secoamatoxins, are formed bytreatment with trifluoroacetic acid atroom temperature. Similarly, the dethiocompounds (No. 35) obtained withRaney-nickel in the amanitin series arealso nontoxic.The chromophoric part of the ama-

toxins until very recently was supposedto consist of a similar 2-thioether ofan indole, the sole difference from phal-lotoxins being a hydroxyl in the 6-position. This phenolic group causesa shift of the ultraviolet maximum tolonger wavelengths (302 mu) as com-

pared with phallotoxins, and an addi-tional bathochromic shift after the addi-tion of a little alkali as a consequenceof phenate ion formation (Fig. 3).From oxidation and reduction experi-ments with model thioethers, H. Faul-stich arrived at a sulfoxide structureof the amatoxins (30).

Accordingly, on acid hydrolysis thethioether bridge is broken in a waydifferent from that in phallotoxins. Herea predominantly "electrophilic" hydroly-sis is observed, one which results in6-hydroxytryptophan (No. 36) andcysteinesulfinic acid (No. 37). Thelatter compound manifests itself ascysteic acid (No. 38), a product ofdisproportionation. The tryptophan de-rivative (No. 36) is very labile to acidsand can be detected only in traces.Amanin (No. 32d), which has an

unsubstituted tryptophan moiety likephalloidin, is also a sulfoxide and ishydrolyzed in an analogous way, andtryptophan and cysteic acid are formed.

On the Mechanism of Toxic Action

The first point of attack of both thetoxins is the liver (31). There are, how-ever, considerable differences in toxici-ties and in modes of action. Phalloidinhas a marked affinity to the microsomalfraction of the liver cell. This has beenshown by perfusing an isolated rat liverwith radioactive toxin, and homogeniz-ing, and fractionating the homogenateby centrifugation. Only the microsomefraction retained a constant activity af-ter several washings (32). This resultcorresponds well to that of von derDecken et al. (33), who found that in-corporation of amino acids into theprotein of the liver microsome fractionwas inhibited when the animals werepoisoned with phalloidin 1 to 2 hoursbefore.

Here, and in its other effects on livermetabolism, phalloidin acts only if ithas been administered to the intact

951

animal. This fact led the Swedish group(33), and recently Fiume (34), to theidea that phalloidin is not toxic initself, but is converted into a toxiccompound by drug-metabolizing en-zymes of the liver. Newborn rats en-dure highly toxic doses of the sub-stance without being killed. Apparent.ly a toxic effect is absent as long asthe system of drug-metabolizing en-zymes is very rudimentary. If theseenzymes are damaged by carbon tetra.chloride or other liver poisons (35),toxicity of phalloidin is reduced, too.In spite of all of these observations,there has been no consistent evidencefor a toxification in liver up to now.

Further insight into the mechanismof the toxic action was obtained byelectron microscopy (36). This methodconfirmed that only the cells of liverare changed by phalloidin. The earliestultrastructural changes-observable 15minutes after injection of phalloidin-are dilatation of the endoplasmic reticu-lum, which occasionally forms verylarge vacuoles, swelling of mitochon-dria, and deposition of fat droplets.The alterations of the reticulum offera plausible explanation of the reducedprotein synthesis in the liver of intoxi-cated animals (33) and are also con-sistent with the observation of a pre-ferred adsorption of radioactive toxinby the microsomal elements of liver(32).The mechanism of the action of the

amanitins is distinctly different fromthat of the phallotoxins. As alreadymentioned, the evidence of intoxica-tion appears considerably later. Herethe nuclei are primarily affected, asrecent electron micrographs of Fiumeand Laschi (36) demonstrate. Withoutany damage in the endoplasmic reticu-lum, the nucleoli of liver cells beginto fragment 15 hours after a-amanitinis given to rats. Cytoplasmatic lesionsfollow the nuclear ones and presumablyare a consequence of the nuclear dam-age. The cytotoxic effects of the amani-tins are not restricted to the liver cells;cytological changes are also observed inkidney. Fiume and Stirpe (37) have ob-served that, in mouse liver nuclei, theRNA content decreases progressivelyduring the first 24 hours of an intoxica-tion of the animals with c-amanitin.Since the toxin prevents neither the

incorporation of 3H-uridine and 3H-thymidine into the nucleic acids of hu-man tumor cells of the KB-Eagle line,nor the replication of viruses, nor thereproduction of bacteria, it is suggestedthat RNA synthesis is not impaired, atleast in these systems (38). Mouse livercells, however, showed a clear responseto a-amanitin in their reduced abilityto incorporate 14C-orotic acid intoRNA, and an effect in vitro has beenfound after all by the Italian patholo-gists. a-Amanitin added to normalmouse liver nuclei inhibited the RNApolymerase reaction activated by man-ganous ion and ammonium sulfate(39) by about 80 percent at 10-7molar concentration (40).

Accordingly, as in the case of phal-loidin and also with the amanitins, aninsight into the mechanism of actionon a molecular basis has not yet beenachieved. One might expect that thiswould also be of general interest incell physiology.Mushrooms of the genus A manita

contain numerous biologically activesubstances. Muscarine, exciting theparasympathetic nervous system, hasbeen isolated from Amanita muscaria(fly agaric), its chemical structure hav-ing been established and synthesizedmainly by Swiss workers. It is pre-sumably not the poisonous principle.Research in Swiss, Japanese, and Britishlaboratories has led to the discoveryof fly-killing and centrally active sub-stances of the isoxazole type like"muscimol," "ibotenic acid," and "tri-cholomic acid." The green, deadly,agaric Amanita phalloides and its whiterelatives contain two groups of lethalsulfur-containing cyclopeptides, the"phallotoxins" and the "amatoxins,"whose structures have been determined.Among other unusual amino acids,they contain leucine and isoleucine resi-dues wich are hydroxylated in -y- and8-positions. The phallotoxins act on theendoplasmatic reticulum of liver cells,whereas the nucleus is the target ofthe cytopathogenic action of the ama-toxins.

References and Notes

1. S. S. Block, R. L. Stephens, W. A. Murrill,J. Agr. Food Chem. 3, 584 (1955).

2. V. E. Tyler, Jr., R. G. Benedict, L. R.Brady, 3. E. Robbers, J. Pharm. Sci. 55,590 (1966).

3. Th. Wieland, W. Motzel, H. Merz, Ann.Chem. 581, 10 (1953).

4. V. E. Tyler, Jr. and D. Groger, Planta Med.12, 397 (1964).

5. For a short specification of toxic Galerina,see R. G. Benedict, V. E. Tyler, Jr., L. R.Brady, L. J. Weber, J. Bacteriol. 91, 1380(1966).

6. C. H. Eugster, "Uber den Fliegenpilx," Publ.Naturforsch. Ges. ZiJrich, 31 December 1966.

7. V. R. Wasson and R. G. Wasson, Mush-rooms, Russia and History (Pantheon, NewYork, 1957), vols. 1 and 2.

8. C. H. Eugster, Advances Org. Chem. 2, 427(1960).

9. S. Wilkinson, Quart. Rev. (London) 15, 153(1961).

10. 0. Loewi and A. Nawratil, Arch. Ges.Physiol. 214, 678, 689 (1926).

11. F. Kogl, C. A. Salemink, H. Schouten, F. Jel-linek, Rec. Trav. Chim. 76, 109, 112 (1957);F. Jellinek, Acta Cryst. 10, 277 (1957).

12. C. H. Eugster, Helv. Chim. Acta 40, 2462(1957); , F. Hifliger, R. Denss, E.Girod, ibid. 41, 205, 583, 705 (1958).

13. A. Hofmann, Indian J. Pharm. 25, 245(1963).

14. C. H. Eugster and T. Takemoto, Helv. Chim.Acta 50, 126 (1967).

15. K. Bowden, A. C. Drysdale, G. A. Mogey,Nature 206, 1359 (1965); Tetrahedron Lett.1965, 727 (1965).

16. T. Takemoto and T. Nakajima, YakugakuZasshi 84, 1183, 1186 (1964); Chem. Abstr.62, 8121c (1965).

17. A. R. Gagneux, F. HMfiger, C. H. Eugster,Tetrahedron Lett. 1965, 2077 (1965); A. R.Gagneux, F. Hifliger, R. Meier, C. H.Eugster, ibid., p. 2081; H. Goth, A. R.Gagneux, C. H. Eugster, H. Schmid, Helv.Chim. Acta 50, 137 (1967).

18. H. Iwasaki, T. Kamiya, 0. Oka, J. Ueyanagi,Chem. Pharm. Bull. (Tokyo) 13, 753 (1965).

19. Th. Wieland and 0. Wieland, Pharmacol.Rev. 11, 87 (1959).

20. Th. Wieland, Fortschr. Chem. Org. Naturstoffe25, 214 (1967); Pure Appl. Chem. 9, 145(1964).

21. F. Lynen and U. Wieland, Ann. Chem. 533,93 (1938).

22. H. Wieland and R. Hallermayer, Ibid. 548,1 (1941).

23. S. S. Block, R. L. Stephens, A. Barreto, W.A. Murrill, Science 121, 505 (1955).

24. Th. Wieland, H. Schiefer, U. Gebert, Na-turwissenschaften 53, 39 (1966).

25. Th. Wieland and J. Dolling, ibid., p. 526.26. U. Gebert, H. Boehringer, Th. Wieland,

Ann. Chem. 705, 227 (1967).27. Th. Wieland and D. Rehbinder, Ibid. 670,

149 (1963).28. Th. Wieland and R. Jeck, Ibid., in press.28a. Th. Wieland and A. Buku, ibid., in press.29. P. Pfaender and M. Hasan, private communi-

cation.30. H. Faulstich, Th. Wieland, Ch. Jochum, Ann.

Chem., in press.31. For a summary, see 0. Wieland, Clin. Chem.

2, 323 (1965).32. D. Rehbinder, G. Loffier, 0. Wieland, Th.

Wieland, Z. Physiol. Chem. 331, 132 (1963).33. A. von der Decken, H. Low, T. Hultin,

Biochem. Z. 332, 503 (1960).34. L. Fiume, Lancet 1965-I, 1284 (1965).35. G. L. Floersheim, Biochem. Pharmacol. 15,

1589 (1966).36. L. Fiume and R. Laschi, Sperimentale 115,

288 (1965).37. L. Fiume and F. Stirpe, Blochim. Blophys.

Acta 123, 643 (1966).38. L. Fiume, M. La Placa, M. Portolani, Speri-

mentale 116, 15 (1966).39. J. R. Tata and C. C. Widnell, Blochem. J. 98,

604 (1966).40. F. Stirpe and L. Fiume, Ibid. 103, 67P

(1967).41. I thank my numerous collaborators who en.

abled me to do the Amantia work; Dr. W.Konz, Chemische Fabrik C. H. Boeh-ringer Sohn, Ingelheim, for invaluable assist-ance in preparing the raw materials; andProfessor C. H. Eugster for revision of thepart of the manuscript concerning A. mus-caria.

952 SCIENCE, VOL. 159