Carcinogenesis vol.19 no.5 pp.

933937, 1998

Caffeine-derived N-nitroso compounds. V. Carcinogenicity of

mononitrosocaffeidine and dinitrosocaffeidine in bd-ix rats

Stan Ivankovic1, Jens Seibel2, Dymitr Komitowski3,

Bertold Spiegelhalder2, Rudolf Preussmann4 and
Maqsood Siddiqi4,5,6
1Division

of Perinatal Toxicology, 2Division of Toxicology and Cancer Risk

Factors, 3Division of Histodiagnostics, 4Division of Environmental
Carcinogens, German Cancer Research Center Heidelberg, Germany, and
5Chittaranjan National Cancer Institute, Calcutta, India
6To

whom correspondence should be addressed at: Chittaranjan National

Cancer Institute, 37, S.P.Mukherjee Road, Calcutta-700026, India
Email: cncinst@giascl01.vsnl.net.in

Mononitrosocaffeidine (MNC) and dinitrosocaffeidine

(DNC) are new N-nitroso compounds obtained from in vitro
nitrosation of caffeidine, a hydrolysis product of caffeine
present in a typically made and widely consumed tea from
Kashmir (India), a high incidence area of esophageal and
stomach cancer. The chemical synthesis, in vitro metabolic
studies and mutagenicity of the compounds has been
previously reported. DNC, a nitrosamide is highly mutagenic both with and without metabolic activation whereas
MNC, like several other aromatic asymmetric nitrosamines,
does not exhibit genotoxic or mutagenic properties. We now
report the results of the first carcinogenicity experiments on
chronic oral administration of these compounds in BD-IX
rats. The acute LD50 of MNC and DNC were about
1300 and 230 mg/kg b.w., respectively. Lung oedema and
gastrointestinal haemorrhages were the first symptoms of
intoxication observed after 2 days for both the compounds.
All three dose groups of MNC treated rats showed
localization of tumours in nasal cavity (93.9100% of
all malignant tumours). The tumours were histologically
diagnosed as neuroepitheliomas of the olfactory epithelium
(neuroblastoma of the bulbus olfactorii) and squamous cell
carcinoma of the nasal cavity in the ratio of 3:1. No
tumours of the nasal cavity were observed in the untreated
controls. DNC, in contrast, induced squamous cell
carcinoma of forestomach in 100% animals at low and
high doses, of which nearly half the tumours metastasized
predominantly into the peritoneum. No forestomach
tumours were seen in the untreated controls. The data
presented here clearly show the potential for induction of
malignant tumours and distinct organ-specificity by MNC
and DNC in rats, and support the postulate that a chronic
exposure to these compounds may provide a carcinogenic
risk for high incidence of gastrointestinal cancers in
Kashmir.
Introduction
A high incidence of both esophageal and stomach cancer has
been reported from Kashmir province in India (1,2). Our
analytical field studies in the area have sufficiently documented
a chronic human exposure to dietary N-nitroso compounds
*Abbreviations: MNC, mononitrosocaffeidine; DNC, dinitrosocaffeidine.
Oxford University Press

and their precursors (35), and biological monitoring studies

on healthy volunteers indicated a high potential for endogenous
nitrosation in the local population (6,7). We have earlier shown
that the method of preparation of a commonly consumed salted
tea in Kashmir (boiling of green tea leaves in presence of
sodium bicarbonate followed by the addition of sodium
chloride) leads to the formation of caffeidine and caffeidine
acid due to the alkaline hydrolysis of caffeine in tea. The
study also showed that caffeidine under acid-catalysed in vitro
nitrosation produces mononitrosocaffeidine (MNC*) and
dinitrosocaffeidine (DNC) besides other minor products (8).
MNC, an asymmetric N-nitrosamine and DNC, a nitrosamide
(for formulae see Figure 1) are new compounds whose chemical
synthesis, in vitro metabolic activation and mutagenic potential
has been reported (810). While DNC was found to induce
strong mutagenic response in different strains of Salmonella
typhimurium (Ames test) and exhibited a high potential for
inducing DNA single-strand breaks in primary rat hepatocytes,
MNC failed to show mutagenicity or genotoxicity in either
system (10). We now report the results on the carcinogenicity
of these compounds on chronic oral administration in BD IXrats using three dose levels for MNC and two for DNC.
Material and methods
Chemical synthesis of MNC and DNC
Since large amounts of the test compounds were needed for carcinogenicity
testing, it was necessary to modify and optimize the method reported earlier
for the synthesis of MNC and DNC (9). In the present method, caffeidine
nitrate 2 was obtained after alkaline hydrolysis of caffeine 1 (Figure 1). The
caffeidine nitrate 2 was directly used for the synthesis of MNC 3 without
further purification. Nitrosation of MNC was performed with NaNO2, acetic
anhydride and acetic acid (12), instead of using nitrosotetrafluoroborate (9).
No by-products, such as nitro derivatives were formed.
All reagents were of analytical grade. The synthesized compounds were
checked by 1H NMR, 13C NMR, IR and mass spectrometry. All obtained
analytical and spectroscopic values agreed with the prior published data (9).
Caffeidine nitrate 2 (N,1-dimethyl-4-(methylammonium-nitrate)-1H-imidazole5-carboxamide). A solution of sodium hydroxide (206 g, 5.15 mol) in water
(2.6 l) was added to caffeine 1 (500 g, 2.58 mol; Merck, Darmstadt, Germany)
The suspension was stirred for 48 h at room temperature resulting in colourless
solution. Caffeidine nitrate 2 was obtained as precipitate by the addition of
560 ml of nitric acid (65%) to the cooled reaction mixture. After an additional
stirring for an hour, the precipitate was obtained after suction and washed
with a small amount of cold water. The product was dried for 2 days at 60C
in vacuo, yielding 229 g (38.5%) of a white powder with decomposition
at 165C.
Mononitrosocaffeidine 3 (MNC) (N,1-dimethyl-4-(methylnitrosoamino)-1Himidazole-5-carboxamide). Solutions of sodium nitrite (110 g, 1.6 mol) in
160 ml of water and 100 ml of 18.5% hydrochloric acid were added
simultaneously to a cooled suspension of caffeidine nitrate 2 (184.8 g, 0.8 mol)
in 800 ml of water. A pH of 23 was maintained during the reaction. The
reaction mixture was stirred for 10 min at 5C after complete addition of
sodium nitrite. This was followed by addition of 600 ml of dichloromethane
to dissolve the formed MNC. Nitrosation was terminated by ammonium
sulphamate (30 g) added in small portions. After the separation of the CH2Cl2
phase, the aqueous reaction mixture was extracted with CH2Cl2 (43100 ml).
An additional amount of MNC was obtained by extraction with CH2Cl2
(33100 ml) after neutralization of the aqueous phase with sodium hydroxide
(40 ml, 5 N). The concentration of the combined organic layers (to 400 ml)

933

S.Ivankovic et al.

Fig. 1. Formation of DNC and MNC from cafeine and caffeidine.

afforded pure crystalline MNC 3 which was obtained after suction and yielded
121 g (78%).
Dinitrosocaffeidine 4 (DNC) (N,1-dimethyl-4-(methylnitrosoamino)-N-nitroso1H-imidazole-5-carboxamide). MNC 3 (39.5 g, 0.2 mol) was added to the
cooled solution (0C) of acetic acid (100 ml) and acetic anhydride (500 ml)
followed by the addition of sodium nitrite (138 g, 2 mol) in small portions
during 5 h at 0C. The temperature was then brought to room temperature
and the mixture was stirred for an additional 15 h, when a viscous solution
was obtained. To this was added a mixture of ice and water (600 ml) and the
excess of sodium nitrite was destroyed by addition of ammonium sulphamate
(30 g) in small portions. Extraction with diethylether (63200 ml) gave a
mixture of DNC and acetic anhydride. After drying the organic layer over
magnesium sulphate, the solution was evaporated in vacuo at 40C. Titration
of the resulting yellow oil with pentane (200 ml) gave a crude yellow powder.
This was dissolved in hot di-isopropylether (600 ml) and insoluble solids
were removed. On concentration of the solution to 150 ml and cooling at
0C, intense yellow crystals of dinitrosocaffeidine 4 were obtained, yielding
33 g (73%).
Determination of LD50
All animal experiments were performed according to the recommendations of
Good Laboratory Practice (GLP). Six male adult BD-IX rats (300 g, b.w.)
were used for each dose group. The dose factor was two. All animals were
starved overnight before gavage of single dose of test compounds. The signs
of toxicity and death were recorded during an observation period of 2 weeks.
The mean lethal dose (LD50) was calculated according to Burn (13).
Carcinogenicity experiments
One-hundred-day-old BD IX-rats (Tierzucht Hannover) were used for chronic
administration of the two test compounds. Animals were separated by sex,
and then randomly distributed into the experimental and control groups. All
animals were kept under conventional conditions and received Altromin pellet
food. The experimental design is summarized in Table I for MNC and DNC.
The required daily dose of MNC was dissolved in 25 ml drinking water
for males and 20 ml for females, respectively. This volume was mostly
consumed in the first half of the day; after ensuring that the vessels had been
emptied, tap water was given until the evening. Animals were then kept
without drinking water overnight. MNC was administered once daily for 5
days/week.

934

Dinitrosocaffeidine was administered by gavage twice per week. The

amount of dose was dissolved in a small volume of alcohol and then diluted
with tap water to an alcohol concentration of 20% (v/v) to obtain a clear
solution. All animals were starved overnight before gavage. Untreated controls
similarly received 20% alcohol twice a week. Treated animals and untreated
controls for DNC were observed for their life time. Moribund rats were killed
by ether anaesthesia and carefully autopsied, including brain and nervous
system. Tumours and other macroscopically visible lesions were registered
and examined histologically.

Results
The acute LD50 of MNC and DNC were about 1300 mg/kg
b.w. and 230 mg/kg b.w., respectively. Lung oedema and
gastrointestinal haemorrhages were the first symptoms of
intoxication observed after 2 days for both the compounds.
The final cause of death, which occurred ~1 week after the
treatment, in most cases was also lung oedema.
The detailed data from carcinogenicity experiments are
given in Table I. The oral administration of MNC was well
tolerated in all treated groups (3 mg, 5 mg and 15 mg/kg, five
times per week), and no signs of acute or sub-acute toxicity
were observed. The body weights in all three dose groups
were similar to those in the controls until the onset of tumour
formation. Similarly, the low dose group of DNC (6 mg/kg)
did not show any decrease in body weights compared to
controls indicating the absence of acute or sub-acute toxicity
at this dose. In the high-dose groups (60 mg/kg; twice weekly),
however, the treatment led to appreciable loss of weight and,
therefore, the dose was reduced to 30 mg/kg, twice weekly
after 2 weeks. The reduced dose was well tolerated and animals
gained body weight comparable to that of the controls. The
median administered total dose (D50) and the median survival
time (T50) are shown in Table I.

Caffeine-derived N-nitroso compounds

Table I. Experimental details of carcinogenicity testing of MNC and DNC

Compound

Initial no. of animals

N

Mononitrosocaffeidine (MNC)

Dinitrosocaffeidine (DNC)

Dose group

Treatmenta
(mg/kg b.w.)

Median total dose applied

D50 (mg/kg)

Median survival timeb

T50

534
406
(162500)
418
(307532)
346
(269454)

568
406
(188553)
364
(177498)
329
(211428)

683
222
(204295)
174
(133196)

685
274
(222327)
190
(154214)

37
17

29
24

Control
Low dose

0
3

825

825

16

29

Median dose

1355

1220

16

27

High dose

15

3330

3255

22
17

21
25

Control
Low dose

0
6

384

470

20

22

High dose

1500

1620

60 30

aFor

MNC: five times per week in drinking water. For DNC: twice a week by stomach tube. In the high dose group, the single dose of 60 mg/kg was reduced
to 30 mg/kg after 4 weeks (see text).
bNumber in parentheses: time of appearance of first and last tumour.

Table II. Carcinogenicity of mononitrosocaffeidine (MNC)

No. of
animals

Malignant tumours
no. (%)

Nasal cavity tumours

No. (%)
0

Other tumours

Neuroblastoma of
bulbus olfac.

Squamous cell carc.

of nasal cavity

Malignant
Adenocarcinoma of mamma, 2
Sarcoma of uterus, 2
Carcinoma of ovary, 1
Carcinoma of uterus, 1
Benign
Fibroadenoma of mamma, 11
s.c. fibroma, 1
Basophilic adenoma of pituitary, 2
Polyps of vagina or uterus, 2

Control

66

6 (9,1)

Low dose
(3 mg/kg/day)

42

3 (78.4)

31 (93.9)

23

Malignant
Adenocarcinoma of mamma, 1
Hepatocell. carcinoma, 1
Benign
Fibroadenoma of mamma, 1

Medium dose
(5 mg/kg/day)

45

29 (64.4)

28 (96.5)

20

Malignant
Adenocarcinoma of mamma, 1
Benign
Fibroadenoma of mamma, 2
s.c. Fibroma, 1
Papilloma of uterus, 1
Basophilic adenoma of pituitary gland, 1

High dose
(15 mg/kg/day)

43

32 (74.4)

32 (10)

24

Benign
Basophilic adenoma of pituitary gland, 1

As shown in Tables II and III, both MNC and DNC induced

malignant tumours in high to very high frequency. Since no
differences were observed with respect to tumour induction
between male and female animals, the data have been pooled.
All the three dose groups of MNC treated rats showed
localization of tumours in nasal cavity (93.9100% of all
malignant tumours) with the exception of one hepatocellular
carcinoma in the low-dose MNC group. Histologically, a
majority of tumours were of neurogenic origin and diagnosed
as neuro-epitheliomas of the olfactory epithelium (neuroblastoma of the bulbus olfactorii). Squamous cell carcinoma

of the nasal cavity were less frequently seen. Interestingly, the

ratio of neuro-epitheliomas and squamous cell carcinomas was
~3:1 in all three dose groups. No tumours of the nasal
cavity were observed in the untreated controls. The identified
malignant (9%) and benign tumours in control groups were of
known spontaneous pathology of the BD IX strain of rats, and
were similar in both controls.
As shown in Table III, DNC induced squamous cell
carcinoma of forestomach in 100% animals at both low and
high doses, of which nearly half metastasized predominantly
into the peritoneum. The severe displasia of the forestomach
935

S.Ivankovic et al.

Table III. Carcinogenicity of dinitrosocaffeidine (DNC)

Eff. animals
no.

Malignant tumours
no. (%)

Squamous cell
carcinomas of the
forestomach

Other tumours

Malignant
Adenocarcinoma of mamma, 3
Sarcoma of uterus, 1
s.c. sarcoma, 1
Benign
Fibroadenoma of mamma, 5
Chromophob. adenoma of pituitary gland, 2
s.c. fibroma, 2
Adenoma of adrenal gland, 1

Control

43

5 (8.6)

Low dose
(2 3 6 mg/kg/week)

42

42 (100)

42 (100)

Metastases of primary tumour in peritoneum, 20

Metastases of primary tumour in liver, 2

High dose
(2 3 30 mg/kg/week)

42

40 (92.5)

40 (10)

Metastases of primary tumour in peritoneum, 21

Benign
Papilloma of forestomach, 1

observed in the first dead animal could be considered as

representative of preneoplastic lesions. No forestomach
tumours were seen in the untreated controls.
Discussion
Although the epidemiological data suggest a multifactorial
aetiology of human cancers, a chronic exposure to exogenous
and endogenously produced N-nitroso compounds has been
strongly implicated in several cancers among high risk populations (14). There is now considerable evidence supporting the
causal role of tobacco-specific (TSNA), and betel-nut specific
nitroso compounds in lung, oral and upper respiratory tract
cancers (15). While no firm evidence is yet available, an
involvement of intragastrically formed direct alkylating
N-nitrosamides is long postulated to be the causal risk factor for
stomach cancer (16) and organ-specific, metabolic activation
dependent asymmetric N-nitrosamines are believed to be
involved in providing high risks for human esophageal
cancer (17).
The clinical and epidemiological data from Kashmir shows
a high incidence both for esophageal and stomach cancers
(1,2). This prompted us to conduct an analytical-field study in
the area (25). The study showed a widespread contamination
of preserved and stored food items with preformed known
volatile and non-volatile N-nitrosamines (3,4) besides a high
potential for endogenous formation of N-nitroso compounds
in the local inhabitants (6,7). While a trace level daily
exposure to nitroso compounds which are known to be rodent
carcinogens may increase the overall burden of mutagenic and
carcinogenic compounds, yet it may not be sufficient to account
for the high incidence of gastrointestinal cancers observed in
the local population in Kashmir. In our search for more specific
and high intake precursors of N-nitroso compounds in local
dietary make-up, we discovered that due to the peculiar method
of preparation of salted tea (i.e. addition of sodium bicarbonate),
the caffeine from tea leaves gives rise to caffeidine and
caffeidine acid due to alkaline hydrolysis. Further studies
suggested that caffeidine, due to its favourable chemistry of
nitrosation (9) will result in the formation of MNC and DNC
under acidic milieu of stomach in presence of high dietary
nitrate (5). This is bound to provide considerable exposure of
endogenously produced MNC and DNC among the local
936

inhabitants due to a liberal consumption (average: 4 cups/day/

adult) of salted tea in Kashmir (5). Estimates based on in vitro
analysis of caffeine show that the daily exposure of MNC and
DNC to salted tea drinkers in Kashmir may be ~150 gm
and 0.5 gm/adult (5,9). It therefore, became imperative to
investigate the carcinogenic potential of these compounds in
order to estimate risk extrapolation to humans. This study
presents the first carcinogenicity data on chronic administration
of mononitrosocaffeidine (MNC) and dinitrosocaffeidine
(DNC) in rats.
As shown in Table II, with the exception of one hepatocellular carcinoma, all tumours in three different dose groups were
localized in nasal cavity showing an extreme organ-specificity
in its carcinogenic action in rat. The cause of nasal cavity
being the preferred organ for carcinogenic action of MNC,
though unclear at the moment, may be related to the presence
of specific activating enzymes in the organ and high stability
of the compound under physiological conditions. MNC has
been shown to undergo metabolic de-methylation preferentially
at methylnitrosamino group leading to the putative formation
of imidazole diazonium ion capable of reacting with cellular
nucleophiles (11). A lack of any nasal cavity tumours in
control group further confirms the carcinogenic potential and
organ-specificity of MNC, and rules out its action in merely
enhancing the incidence of spontaneously induced tumours.
Primacy of nasal cavity tumours also indicate the systemic
action of MNC and shows that it can cause tumours in organs
other than those which directly come into contact or where it
may be formed from caffeidine as it may occur under human
situation. The target specificity of MNC may vary in different
species of animals and in humans depending on the levels and
specificity of enzymes responsible for its metabolic activation
and clearance of activated carcinogen. Change in target specificity of various N-nitrosamines in different animal species is
well documented (18). Although data on the median survival
time show a dose dependency in its carcinogenic action, it
appears that the lowest dose used in the study was close to
the dose required to produce optimal induction of tumours.
DNC, a highly mutagenic and genotoxic compound (10)
produced 100% tumours (squamous cell carcinomas of the
forestomach) at the site of its first contact under experimental
conditions, thereby showing exceptional organotropy in its

Caffeine-derived N-nitroso compounds

carcinogenic action. The target specificity of DNC is in

agreement with its direct methylating action and low stability
of the compound at physiological pH (9). The data also show
that the low dose regimen was in itself sufficient to produce
maximum number of tumours and hence dose response is
obscured (Table III). It is therefore necessary that future
experiments focus on low doses of DNC in order to obtain
a doseresponse relationship. An appreciable frequency of
metastases in peritoneum and liver shows high carcinogenic
potential, and capability of inducing fast growing tumours by
DNC in rats.
The results obtained in the study further support the possible
role of caffeine-derived nitroso compounds in providing high
risk of esophageal and stomach cancer in the local population
of Kashmir, where potential formation of such compounds has
been already contemplated due to the large consumption of
salted alkaline tea. The caffeine-derived N-nitroso compounds
(MNC and DNC), therefore, also acquire immense importance
as environmental carcinogens in other regions of the world
where alkaline conditions may be used for tea preparation
intentionally or due to the high alkalinity of natural water.

13. Burn,J.H. (1937) Biologische Auswertungsmethoden. Springer-Verlag,

Berlin.
14. Bartsch,H. and Montesano,R. (1984) Relevance of nitrosamines to human
cancer. Carcinogenesis, 5, 13811393.
15. Hecht,S.S. and Hoffmann,D. (1991) N-nitroso compounds and tobaccoinduced cancers in man. In ONeil,I.K., Chen,J.S. and Bartsch,H. (eds)
Relevance to Human Cancer of Nitroso Compounds, Tobacco and
Mycotoxins, IARC Scientific Publication No. 105. IARC, Lyon, pp. 5461.
16. Mirvish,S.S. (1983) The etiology of gastric cancer: intragastric formation
and other theories. J. Natl. Cancer Inst., 71, 631647.
17. Preussmann,R. (1987) Chemical factors in the pathogenesis of esophageal
carcinoma. In Wagner,G and Hi,Z.Y (eds) Cancer of Liver, Esophagus
and Nasopharynx. Springer-Verlag, Heidelberg, pp. 119123.
18. Preussmann,R. and Stewart,B.W. (1984) N-nitroso carcinogens. In
Searle,C.E. (ed.) Chemical Carcinogens, ACS Monograph No. 182, 2nd
edn. American Chemical Society, Washington, DC, pp. 643828.
Received on September 30, 1997; revised on December 1, 1997; accepted on
December 1, 1997

Acknowledgements
M.S. would like to thank the German Cancer Research Centre (DKFZ),
Heidelberg, for financial support during his stay at the Department of
Environmental Carcinogens, DKFZ, Heidelberg.

References
1. Siddiqi,M. and Preussmann,R. (1989) Esophageal cancer in Kashmir: an
assessment. J. Cancer Res. Clin. Oncol., 115, 111117.
2. Khuroo,M.S., Zargar,S.A., Mahajan,R. and Banday,M.A. (1992) High
incidence of oesophageal and gastric cancer in Kashmir in a population
with special personal and dietary habits. Gut, 33, 1115.
3. Siddiqi,M., Tricker,A.R. and Preussmann,R. (1988) The occurrence of
preformed N-nitroso compounds in food samples from a high risk area of
esophageal cancer in Kashmir, India. Cancer Lett., 39, 3743.
4. Kumar,R., Mende,P., Tricker,A.R., Siddiqi,M. and Preussmann,R. (1990)
N-nitroso compounds and their precursors in Brassica oleacea. Cancer
Lett., 54, 6165.
5. Siddiqi,M., Kumar,R., Fazili,Z., Spiegelhalder,B. and Preussmann,R.
(1992) Increased exposure to dietary amines and nitrate in a population
at high risk of oesophageal and gastric cancer in Kashmir, India.
Carcinogenesis, 13, 13311335.
6. Tricker,A.R., Kumar,R., Siddiqi,M., Khuroo,M.S. and Preussmann,R.
(1991) Endogenous formation of N-nitrosamines from piperazine and their
urinary excretion following antihelmintic treatment with piperazine citrate.
Carcinogenesis, 12, 15951599.
7. Kumar,R., Siddiqi,M., Fazili,Z., Wacker,C.D., Spiegelhalder,B. and
Preussmann,R. (1992) Effects of dietary nitrate on endogenous nitrosation
of piperazine in humans. Cancer Lett., 65, 139143.
8. Kumar,R., Mende,P., Wacker,C.D., Spiegelhalder,B., Preussmann,R. and
Siddiqi,M. (1992) Caffeine-derived N-nitroso compounds. I: Nitrosatable
precursors from caffeine and their potential relevance in the etiology of
esophageal and gastric cancers in Kashmir, India. Carcinogenesis, 13,
21792182
9. Kumar,R., Wacker,C.D., Mende,P., Spiegelhalder,B., Preussmann,R. and
Siddiqi,M. (1993) Caffeine-derived N-nitroso compounds. II. Synthesis
and characterization of nitrosation products from caffeidine and caffeidinic
acid. Chem. Res. Toxicol., 6, 5058.
10. Erdinger,L, Schmezer,P., Razdan,R., Kumar,R., Spiegelhalder,B.,
Preussmann,R. and Siddiqi,M. (1993) Caffeine-derived N-nitroso
compounds. III. Mutagenicity in S. typhimurium and in vitro induction of
DNA single-strand breaks in rat hepatocytes by mononitrosocaffeidine
and dinitrosocaffeidine. Mutat. Res., 292, 4149.
11. Razdan,R., Frei,E., Spiegelhalder,B. and Siddiqi,M. (1994) Caffeinederived N-nitroso compounds. IV. Kinetics of mononitrosocaffeidine
demethylation by rat liver microsomes. Cancer Lett., 79, 117122.
12. White,E.H. (1955) The chemistry of N-alkyl-N-nitrosamides. 1. Methods
of preparation. J. Am. Chem. Soc., 77, 60086010.

937