Enzymatic 9Methylation of

~Acetylserotonin to Melatonin

Abstract. An enzyme, hydroxyindole-O-
methyl transferase. that can transfer the
methyl group of S-adenosylmetbionine to
the hydroxy group of N-acetylserotonin to
form the hormone melatonin is described.
This enzyme, which is highly localized in
the pineal gland,  also 0-methylates
serotonin.

Recently Lerner and co-workers (I )
isolated a new hormone, melatonin (N-
acetyl&methoxytryptamine) , from the
pineal gland and peripheral nerves of
man, monkey, and cow. This compound
was found to lighten the color of frog
melanocytes and block the actions of
the melanocyte-stimulating and adreno-
corticotropic hormones (I ) . McIsaac
and Page have recently shown that

Table-l, Enzymatic 0-methylation of N-acetyb
serotonin to melatonin. The soluble supematant
fraction obtained from 16 mg of cow pineal
gland was incubated at 37oC with 0.1 pmole of
N-acetylserotonin, 100 #moles of phosphate
buffer (PH 8.0), and 0.1 cmole of S-adenosyl-
methionine. After 2 hours' incubation melatonin
was determined in the incubation mixture. (4).

system

Melatonin formed
(wmoles)

Complete system
S-adenosylmethionine
omitted

11


0

serotonin (5-hydroxytryptamine) is con-
verted to N-ace'tylsero&min in vivo (2).
We wish to report the isolatiofi of an
enzyme that forms melatonin by the O-
methylation of N-acetylserotonin.

Since melatonin was found to be
highly localized in the pineal gland (1) ,
this tissue was examined for the pres-
ence of an enzyme that could O-
methylate hydroxyindoles. Pineal glands
from cows (3) were homogenized with
ice-cold isotonic potassium chloride and
centrifuged at 78,000g. The resulting
soluble supernatant fraction was incu-
bated with N-acetylserotonin and S
adenosylmethionine at pH 8.0. After
a 2-hour incubation at 37oC the reac-
tion product was extracted from the
incubation mixture with chloroform and
the organic phase was washed with
water to remove residual substrate. The
chloroform extract was then evaporated
to dryness in a stream of warm air and
the residue taken up in 3N HCl. A
fluorescent metabolite was found to be
present in the acid extract with a
maximum fluorescent peak in 3N HCI
at 540 rnp upon activation at 310 rnp;
this is characteristic of 5-hydroxy- and
5methoxyindoles. This metabolite had
the same fluorescent spectrum, the same
RI values in butanol, acetic acid, and
water (100 : 35 : 70) (0.91) and in
N-propanol and 1N ammonia (5 : 1)
(0.891, and the same color reactions
and partition coefficient as authentic
melatonin. When S-adenosyhnethionine
was omitted from the incubation mix-
ture, no melatonin was formed (Table
1) . These observations demonstrate the
existence of an enzyme (hydroxyindole-
O-methyl transferase) that can transfer
the methyl group of Sadenosylmethi-

Hq-f!!-tt-l&&Ha + Sadenosylmethionine -

Fig. 1. Transfer of the methyl group of S-adenosylmethionine to the hydroxy group of
N-acetylserotonin.

1312

onine to the hydroxy group of N.
acetylserotonin. The reaction is shown
in F:g. 1.

HydroxyindoleO-methyl transferas,
has been purified about 20-fold from
beef pineal gland by heat treatment,
ammonium sulfate fractionation, and
adsorption and elution from alum&,
Cy gel (4). Unlike catechol-O-methyl
transferase (5), the enzyme has no
requirement for Mg++. It could not &
detected in liver and kidney of a num.
ber of mammalian species, but was
found in the pineal gland of the man.
key (4). The lack of the requirement
for Mg" and the unique localization
of hydroxyindole-O-methyl transfera%
indicates that it is different from
catechol-O-methyl transferase (5) and
the other known transferases (6, 7).
Incubation of serotonin with hy-
droxyindole-O-methyl transferase and
S-adenosylmethionine resulted in the
formation of a product having the
' characteristics of authentic 5-methoxy-
serotonin. However, the rate of O-
methylation of serotonin was only
one-tenth that of N-acetylserotonin
(4). This finding suggests that acetyla-
tion precedes 0-methylation in the
formation of melatonin as follows:

Serotoni-

    N-acetylserotoni- melatonin

From the results described in this
report and elsewhere, it is becoming
increasingly apparent that 0- and N-
methyltransferases (5-7) requiring S-
adenosylmethionine are playing key
roles in the biosynthesis and inactivation
of biologically active amines and their
derivatives (8).
             JULIUS AXELgoD

          HERBERT WEISSBACH
Laboratory of Clinical Science,
National Znstitute of Mental Health,
and Laboratory of Clinical BiochemistUp
National Heart Znstitutk,
Bethesda, Maryland

References and Notes

1. A. B. Lcmer, J. D. Case, k. Takahashk T. H'
Lee W Mm-i J Am. Chewi Sot 80, zSR!
(19b);`A. B.`&mcr, J. D. &e,`W. Mar"
M. R. Wright, Nature 183, 1821 (1959); A. "
Lemer, J. D. Case, R. V. Heinzelma% J. Am
Chsm. Sot. 81, 6084 (1959).
2. W. Mchac and I. EL Page, 1. 8101. Chr"'

-.

334-858 (1959).

3. We are greatly indebted to Dr. A. B. Le;$
for supplying US with. beef pineal glands
melatotdn.

4. A detaikd description of the assay, prop@'
                   ' dole-@
SpeciflcitY, and localization of hydroxyln
methvl transferasa will be published in a futur'

5. J"?!i%!!,%;ence 126 400 (1957); J. .4d
rod and R. Tomchick, i Biol. Chem. 233. "'
(1958).
6. N. Kirshner .and McC. Goodall, Blochi~l. "
Biophys. Acta 24, 658 (1957).
7. D. D. Brown, J. Axclmd, R. To@@"
Nature 1833, 680 (1959).
8. We thank Dr. A. B. timer, Dr. S. Uden fried
and Dr. B. Witkop for helpful di~C@~~'
during the course of the study.
21 Decembu 1959   SCIIJNCB, VOL. I"