Identification of co-mutagenic chlorinated harmans in final effluent from a sewage treatment plant

Article abstract of DOI:10.1016/S1383-5718(01)00132-2

Harman (1-methyl-9H-pyrido[3,4-b]indole) reacted readily with sodium hypochlorite in an aqueous medium to give the mono-chlorinated derivatives, which reportedly have greater co-mutagenic activity than harman in the presence of o-toluidine toward Salmonella typhimurium TA 98 with S9 mix. Mono-chlorinated harmans were detected by concentration using blue rayon (BR) and GC/MS analysis in the final effluent from a sewage treatment plant in Shizuoka, Japan. The amounts adsorbed for 24h were 1-45ng/gBR for mono-chlorinated harman and 110-730ng/gBR for harman.

Full text of DOI:10.1016/S1383-5718(01)00132-2

Mutation Research 491 (2001) 65–70
Identification of co-mutagenic chlorinated harmans in
final effluent from a sewage treatment plant
Hitoshi Fukazawaa,b, Hidetsuru Matsushitab, Yoshiyasu Teraoa,∗
a
Institute for Environmental Sciences, University of Shizuoka, 52-1 Yada, Shizuoka 422-8526, Japan
b
Shizuoka Institute of Environment and Hygiene, 4-27-2 Kitaando, Shizuoka 420-8637, Japan
Received 17 October 2000; received in revised form 18 December 2000; accepted 18 December 2000
Abstract
Harman (1-methyl-9H-pyrido[3,4-b]indole) reacted readily with sodium hypochlorite in an aqueous medium to give the
mono-chlorinated derivatives, which reportedly have greater co-mutagenic activity than harman in the presence of o-toluidine
toward Salmonella typhimurium TA 98 with S9 mix. Mono-chlorinated harmans were detected by concentration using blue
rayon (BR) and GC/MS analysis in the final effluent from a sewage treatment plant in Shizuoka, Japan. The amounts adsorbed
for 24 h were 1–45 ng/g BR for mono-chlorinated harman and 110–730 ng/g BR for harman. © 2001 Elsevier Science B.V.
All rights reserved.
Keywords: Harman; Co-mutagenicity; Disinfection by-product; Blue rayon
1. Introduction
ted that such ␤-carbolines can be easily chlorinated
with sodium hypochlorite and the chlorinated prod-
ucts show 20 times more potent co-mutagenic activity
than the ␤-carbolines themselves in Ames test using S.
typhimurium TA98 in the presence of o-toluidine [10].
It is generally difficult to analyse water pollutants
because they are present in only minute concentra-
tions. Therefore, approaches to concentrate pollutants
in river water are very important. Frequently used
methods include extraction with organic solvents,
a variety of XAD resins, and Sep-pack cartridges.
In addition, blue cotton and blue rayon, which con-
sist of cotton and rayon covalently bonded to the
blue pigment copper phthalocyanine trisulfonate, can
specifically adsorb multicyclic planar compounds, as
reported by Hayatsu [11].
Harman (1-methyl-9H-pyrido[3,4-b]indole) and
norharman (9H-pyrido[3,4-b]indole) are present in
alcoholic beverages, like beer, wine, sake and whisky,
in cigarette smoke, and in fried meat and fish [1–5].
Although harman and norharman, ␤-carbolines, have
skeletons similar to those of mutagenic and carcino-
genic heterocyclic aromatic amines, it has been repo-
rted that they themselves are not mutagenic to
Salmonella typhimurium TA98, either with or without
S9 mix. However, they became mutagenic toward S.
typhimurium TA98 with S9 mix when incubated with
non-mutagenic aromatic amines, such as aniline or
o-toluidine. Accordingly, harman and norharman have
been called “co-mutagens” [6–9]. Recently, we repor-
In the present study, we attempted to isolate chlori-
nated harman that had been concentrated using blue
rayon from final effluent of a sewage treatment plant,
where harman was found to be present at higher level
∗ Corresponding author. Tel.: +81-54-264-5788;
fax: +81-54-264-5788.
E-mail address: terao@u-shizuoka-ken.ac.jp (Y. Terao).
1383-5718/01/$ – see front matter © 2001 Elsevier Science B.V. All rights reserved.
PII: S1383-5718(01)00132-2

66
H. Fukazawa et al. / Mutation Research 491 (2001) 65–70
and Hewlett Packard HP-5971A mass spectrometer.
The GC/MS conditions were as follows: column,
HP-5 trace analysis (5% diphenyl and 95% dimethy-
larylenesiloxane, coating film thickness, 0.1 ␮m;
0.25 mm i.d. × 30 m, Hewlett Packard Co.); injec-
tion mode, splitless; injection temperature, 250◦C;
column head pressure, 7 psi; column temperature,
70–250◦C; program rate, 10◦C/min; ionising volt-
age, 70 eV. MS measurement was performed in scan
mode for a qualitative analysis and in the selected ion
monitoring (SIM) mode for a quantitative analysis.
The monitored ions were the peaks at m/z 154, 181
and 182 for harman and m/z 181, 216 and 218 for
mono-chlorinated harmans.
Fig. 1. Chemical structures of harman and chloroharman.
than norharman. We describe here the identification of
three chlorinated harmans (Fig. 1) isolated by GC/MS
analysis.
2.3. Chlorination of harman with sodium
hypochlorite in aqueous medium
A solution of sodium hypochlorite was diluted with
pure water to 10 mg/l of available chlorine, and was
adjusted to pH 4 with 0.1 M hydrochloric acid. To
20 ml of this solution was added 400 ␮l of harman
solution (1 mg/ml acetone solution) and the mixture
was allowed to stand for 5 min at room temperature.
Sodium thiosulfate was added and the mixture was
adjusted to pH 8 with 0.2 M sodium hydroxide. The
reaction products were extracted twice with 2 ml of di-
chloromethane. The extract was dried over anhydrous
sodium sulfate and was subjected to GC/MS analysis.
3-Chloroharman that had been newly produced in
aqueous medium was isolated by a method similar
to that described previously [10] after a large-scale
reaction.
3-Chloro-1-methyl-9H-pyrido[3,4-b]indole (3-chlo-
roharman): mp 220–222◦C. HRMS m/z: 216.0429
(M+), calcd for C12H9ClN2, 216.0454. 1H-NMR
(CDCl3) δ: 2.87 (1H, s, CH3), 7.23 (1H, dt, J = 8.1,
1.0 Hz, 6-H), 7.50 (1H, d, J = 8.1 Hz, 8-H), 7.56
(1H, dt, J = 8.1, 1.0 Hz, 7-H), 7.80 (1H, s, 4-H),
8.05 (1H, d, J = 8.1 Hz, 5-H). 8.4 (1H, br, NH).
2. Experimental
2.1. Materials
Harman was purchased from Aldrich Chemical
Co. Inc. Both 6- and 8-chloroharmans were syn-
thesised by chlorination of harman and purified by
the same method as reported previously [10]. The
solvents used were of pesticide analysis grade and
sodium hypochlorite was purchased from Wako Pure
Chemicals Industry Ltd. Blue rayon was obtained
from Funakoshi Co. Ltd., Tokyo. The solid-phase
extraction disk SDB-XD was from Sumitomo 3 M,
Ltd. The silica gel column was prepared by plac-
ing 5 g of silica S-60 (Merck, Japan) in a glass tube
(10 mm i.d. × 300 mm). Water was treated with a
Milli-Q water purification system after distillation and
then purified by an activated carbon column. 1H-NMR
spectra of solutions in chloroform-d were taken with
a JEOL JNM-GSX270 (1H: 270 MHz) Fourier trans-
form spectrometer. The following abbreviations are
used: s = singlet, d = doublet, t = triplet, m =
multiplet, br = broad. Chemical shifts are shown in
ppm using tetramethylsilane as an internal standard.
2.4. Determination of harman in final effluent
from a sewage treatment plant by extraction with a
solid-phase disk
2.2. GC/MS apparatus and operating conditions
Gas chromatography/mass spectrometry (GC/MS):
Hewlett Packard HP-5890 II gas chromatograph
A solid-phase extraction disk (SDB-XD 47 mm Ø)
was washed with 5 ml of dichloromethane, 5 ml of

H. Fukazawa et al. / Mutation Research 491 (2001) 65–70
67
methanol and then 20 ml of pure water. Water samples
(3 l) were extracted with three disks. Harman ex-
tracted on the disks was eluted with dichloromethane
(4 ml, two times). The combined solution was dried
over anhydrous sodium sulfate, and concentrated on
a rotary evaporator to about 1 ml and further concen-
trated to 0.5 ml under a nitrogen stream. Harman was
determined by GC/MS analysis.
products. The MS spectra of peaks A, D and G are
shown in Fig. 2.
In the MS spectrum of peak G (Fig. 2), the molec-
ular ion peaks appear at m/z 216 and 218 with a
relative abundance of about 10:3, which suggests
that they are due to mono-chlorinated harman. The
MS spectra of peaks E and F also showed the same
molecular ion peaks at m/z 216 and 218 and similar
fragment patterns. These peaks may be due to isomers
of mono-chlorinated harman observed as peak G.
Peaks E and F were identified to be those of 8- and
6-chloroharman, respectively, by comparison with the
GC retention times of synthetic compounds described
previously [10]. Peak G was found to be due to the
third isomer of mono-chlorinated harman. The isomer
was isolated and determined to be 3-chloroharman by
1H-NMR analysis.
Peaks H and I showed similar MS spectra and both
had molecular ion peaks appeared at m/z 250, 252
and 254 (relative abundance = 10:6:1), suggesting
di-chlorinated harman. They were assumed to be the
positional isomers, but were not identified.
The MS spectrum of peak A shown in Fig. 2 indi-
cated that this compound is a mono-chlorinated com-
pound. However, this is not a harman derivative. The
structure was assumed to be a chlorinated derivative
of acetyl indole derived from harman. The MS spec-
tra of peaks B and C also suggested that they have
the same skeleton. Further structural determination
and the formation mechanism of these compounds
are under investigation.
2.5. Determination of harman and chlorinated
harmans in effluent from sewage treatment plant by
concentration with blue rayon
Three wire net bags of blue rayon (5 g per bag) were
hung in a final effluent from sewage treatment plant for
24 h. The three samples of blue rayon were combined
and washed five times with 1 l of pure water. Adsorbed
materials were extracted three times by shaking the
blue rayon in 300 ml of methanol/ammonia water
(50/1, v/v). The combined extracts were concentrated
to about 20 ml under reduced pressure. To this so-
lution was added 100 ml of pure water and 10 g of
sodium chloride, and the mixture was then extracted
three times with 30 ml of dichloromethane. The
dichloromethane layers were combined, dried over
anhydrous sodium sulfate, concentrated to about 1 ml,
and then applied to a silica gel column. After eluting
with 40 ml of hexane and 20 ml of 10% ethyl acetate
in hexane, the fractions eluted with 20 ml of ethyl
acetate and 30 ml of acetone were concentrated to
1 ml and then were subjected to GC/MS analysis.
3.2. Identification of harman and chlorinated har-
mans in final effluent from a sewage treatment plant
3. Results and discussion
3.1. Chlorination of harman with sodium
hypochlorite in an aqueous medium and GC/MS
analysis of the products
Three wire net bags that contained blue rayon (5 g
per bag) were hung in final effluent from a sewage
treatment plant in Shizuoka, Japan, for 24 h. Since
the extracts from final effluent contained various
compounds, they were purified using a silica gel col-
umn. The fractions including harman and chlorinated
harmans were concentrated and subjected to GC/MS
analysis. Their SIM chromatograms are shown in
Fig. 3. The fraction eluted with ethyl acetate con-
tained 3-chloroharman and that eluted with acetone
contained harman and 6- and 8-chloroharman.
Since preliminary experiments showed that the
chlorination of harman occurred rapidly with a low
concentration of sodium hypochlorite, the reactions
of harman (400 ␮g) were carried out by standing
for 5 min in 20 ml of sodium hypochlorite solutions
(chlorine, 10 mg/l). The GC/MS chromatogram of the
extract from the reaction mixture is shown in Fig. 2.
The peak of unreacted harman is D and there are
eight new peaks (A, B, C, E, F, G, H and I) for the
The quantitative results are summarised in Table 1.
Harman was detected in every measurement and

68
H. Fukazawa et al. / Mutation Research 491 (2001) 65–70
Fig. 2. Total ion chromatogram (TIC) and mass spectra: (1) TIC of the products formed by the reaction of harman with sodium hypochlorite
in aqueous medium. Mass spectra of (2) peak A, (3) peak D, (4) peak G.
ranged from 110 to 730 ng/g BR, and mono-chlorinated
harmans were identified and found in concentration
ranging from 1 to 45 ng/g BR when harman was
found at a relatively high concentration.
The ratios of mono-chlorinated harmans to harman,
which were calculated from the amounts adsorbed by
blue rayon, were 3.7–23.7% for 3-chloroharman,
1.1–5.1% for 6-chloroharman and 0.5–0.7% for
8-chloroharman. Based on the quantification of har-
man (27–75 ng/l) by extraction with a SDB-XD disk,
the amount of chlorinated harman was estimated
to be on the order of a few ng/l. This value is a
little less than that of mutagenic and carcinogenic
3-amino-1-methyl-5H-pyrido[4,3-b]indole (Trp-P-2)

H. Fukazawa et al. / Mutation Research 491 (2001) 65–70
69
Fig. 3. SIM chromatograms of the compounds absorbed with blue rayon in final effluent from a sewage treatment plant: (1) and (2) authentic
compounds, (3) harman in the acetone-solution eluate, (4) 3-chloroharman in the ethyl acetate-solution eluate, (5) 6- and 8-chloroharman
in the acetone-solution eluate.

70
H. Fukazawa et al. / Mutation Research 491 (2001) 65–70
Table 1
Adsorption amounts of harman and the chlorinated derivatives in final effluent from a sewage treatment planta
Date
Harman
6-Chloroharman
8-Chloroharman
3-Chloroharman
21 August
4 September
2 November
730
110
190
37
Traceb
2
5
27
Trace
45
Trace
1
a ng/g of blue rayon.
b Trace: <1 ng/g of blue rayon.
which was detected (on the order of 10 ng/l) in final
effluent from a sewage treatment plant [12].
[5] Y. Totsuka, H. Ushiyama, J. Ishihara, R. Sinha, S. Goto,
T. Sugimura, K. Wakabayashi, Quantification of the co-
mutagenic ␤-carbolines, norharman and harman, in cigarette
smoke condensates and cooked foods, Cancer Lett. 143 (1999)
139–143.
[6] M. Nagao, T. Yahagi, M. Honda, Y. Seino, T. Matsushima,
T. Sugimura, Demonstration of mutagenicity of aniline and
o-toluidine by norharman, Proc. Jpn. Acad. 53B (1977)
34–37.
[7] M. Nagao, T. Yahagi, T. Sugimura, Differences in effects of
norharman with various classes of chemical mutagens and
amounts of S9, Biochem. Biophys. Res. Commun. 83 (1978)
373–378.
[8] T. Sugimura, M. Nagao, K. Wakabayashi, Metabolic aspects
of the comutagenic action of norharman, in: R. Snyder, D.J.
Jollow, D.V. Parke, C.G. Gibson, J.J. Kocsis, C.M. Witmer
(Eds.), Biological Reactive Intermediates-II, Chemical
Mechanisms and Biological Effects Part B, Plenum Press,
New York, 1982, pp. 1011–1025.
[9] M. Mori, Y. Totsuka, K. Fukutome, T. Yoshida, T. Sugimura,
K. Wakabayashi, Formation of DNA adducts by the
co-mutagen norharman with aromatic amines, Carcinogenesis
17 (1996) 1499–1503.
[10] K. Nakano, K. Suyama, H. Fukazawa, M. Uchida, K.
Wakabayashi, T. Shiozawa, Y. Terao, Chlorination of harman
and norharman with sodium hypochlorite and co-mutagenicity
of the chlorinated products, Mutat. Res. 470 (2000) 141–146.
[11] H. Hayatsu, Cellulose bearing covalently linked copper
phthalocyanine trisulphonate as an adsorbent selective for
polycyclic compounds and its use in studies of environ-
mental mutagens and carcinogens, J. Chromatogr. 597 (1992)
37–56.
Co-mutagenic mono-chlorinated harmans were
detected along with harman in final effluent from a
sewage treatment plant. Although their concentra-
tions were not so high, chloroharmans have evidently
contaminated river water for a long time. With regard
to environmental pollutants, chloroharman should be
considered a toxic organohalogen compound, since it
seems likely that the chlorinated derivative exhibits
toxicity other than co-mutagenicity. In any case, the
occurrence of chloroharmans is likely due to disin-
fection with sodium hypochlorite in sewage treatment
plants.
References
[1] K. Wakabayashi, M. Nagao, H. Esumi, T. Sugimura, Food-
derived mutagens and carcinogens, Cancer Res. 52 (1992)
2092–2098.
[2] G.A. Gross, R.J. Turesky, L.B. Fay, W.G. Suillwell, P.L.
Skipper, S.R. Tannenbaum, Heterocyclic aromatic amine
formation in grilled bacon, beef and fish and in grill scrapings,
Carcinogenesis 14 (1993) 2313–2318.
[3] E.H. Poindexter Jr., R.O. Carpenter, The isolation of
harman and norharman from tobacco and cigarette smoke,
Phytochemistry 1 (1962) 215–221.
[4] J. Adachi, Y. Mizoi, T. Naito, S. Yamomoto, S. Fujiwara,
I. Ninomiya, Determination of ␤-carbolines in foodstuffs by
high-performance liquid chromatography-mass spectrometry,
J. Chromatogr. 538 (1991) 331–339.
[12] S. Mohri, I. Somiya, Y. Ono, Risk assessment of genotoxic
substances related to human diet in the water environment,
J. Jpn. Soc. Water Environ. 19 (1996) 847–854.