Eur. J. Med. Chem.

22 (1987) 433—437 433

© Elsevier, Paris

Original paper

A quantum-chemical and experimental study of the hallucinogen

(+)-1-(2,5-dimethoxy-4-nitrophenyl)-2-aminopropane (DON)
Juan Sebastian GOMEZ-JERIA!, Bruce K. CASSELS? and Juan Carlos SAAVEDRA-AGUILAR?

\1Paboratoire de Chimie Physique, Université de Paris VI, 11, rue Pierre et Marie Curie, 75005 Paris, France, and
Departamento de Quimica, Facultad de Ciencias, Universidad de Chile, Casilla 653, Santiago, Chile,
2Laboratoire de Pharmacognosie, Université de Paris XT, rue J.-B.-Clément, 92296 Chatenay-Malabry Cedex, France, and
3 Departamento de Ciencias Neuroldégicas, Facultad de Medicina, Universidad de Chile, Santiago, Chile
(Received November 7, 1986, accepted June 5, 1987)

Summary — The electronic structure of 1-(2,5-dimethoxy-4-nitrophenyl)-2-aminopropane (DON) was calculated at

the CNDO/2 level, and the racemic compound was synthesized and found to be hallucinogenic at doses of 4 mg. DON
differs from its similarly active congeners in that a hydrophilic nitro group replaces lipophilic substituents at C-4 of the
benzene ring. The implications for the mechanism of serotonin receptor binding of these drugs are discussed.

Résumé -- Une étude de chimie quantique et expérimentale sur le (=-)(diméthoxy-2,5 nitro-4 phényl)-1 amino-2 propane
(DON) 4 propriétés hallucinogénes, La structure électronique du (diméthoxy-2,5 nitro-4 phényl)-1 amino-2 propane (DON)
a été calculée au niveau CNDO/2 et la substance racémique, testée, s'est avérée hallucinogéne a la dose de 4 mg. Le DON se
distingue de ses congénéres @ activité semblable par le remplacement des substituants lipophiles du C-4 du noyau benzénique
par un groupe nitro hydrophile. On discute la portée de ces faits sur le mécanisme de la liaison entre ces drogues et les récepteurs
a sérotonine.

frontier molecular orbitals / CNDO/2 / electrophilic superdelocalizability / atomic charge / lipophilicity / serotonin receptor affinity /
hallucinogenic potency / DON

Introduction prediction of PAA hallucinogenic activities [3], although

exceptions have been noted [4], and several different lines
One of us recently proposed’ a method relating receptor of evidence suggest that non-serotonin receptors are also
affinity to molecular—electronic structure (i.e., electronic probably involved in the psychopharmacology of PAA and
and steric effects) [1]. Its application to QSAR analyses IAA (indolealkylamines) [5]. These peripheral serotonin
of phenylalkylamines (PAA) [2] has led to significant receptors are not clearly related to binding sites in the
correlations for this class of psychotropic substances. In mammalian brain [6—9], but nevertheless, it can be shown
this paper, we report a quantum-chemical QSAR study on that pA, values determined using the rat stomach fundus
1-(2,5-dimethoxy-4-nitrophenyl)-2-aminopropane (DON), [3] are significantly correlated with affinities for central
based on an equation developed earlier for PAA [2], which serotonin binding sites [10].
indicates that this compound may be psychoactive at low In the case of the PAA, there has been some controversy
doses. Pharmacological testing of the target compound regarding the role of the C-4 substituent. It is well known
showed that it is indeed a potent hallucinogen. that 2,5-dimethoxy substitution is associated with high
hallucinogenic potency, and that introduction of a hydro-
phobic atom or group at C-4, such as bromine, methyl or
ethyl, enhances this activity, while also affecting the quali-
Models, Methods and Calculations tative nature of the drug-induced experience in humans
[11]. It has been speculated that the receptor possesses a
The equation mentioned above relates the variation of lipophilic site to accommodate this substituent [12]. The
relevant molecular—electronic structure factors to the varia- affinities of 1-(2,5-dimethoxy-4-X-phenyl)-2-aminopropanes
tion of the binding affinity (pA,) for serotonin receptors for the rat stomach fundus serotonin receptor are generally
in the rat stomach fundus [3]. This sensitive pharmaco- high (pA, ~ 7.0), but it has been suggested that the lipo-
logical preparation appears to be a useful model for the philicity of the C-4 substituent only plays a minor role
434
in these cases. [3].It seems important therefore to evaluate setting. The subjects were asked to describe their experiences
a broader range of 4-substituted PAA to clarify these points. at the end of each session,
In particular, a nitro group at this position, though hydro-
philic, could be expected to Fulfill the electronic and steric
requirements of the receptor.- Results and Discussion
(+)-1-(2,5-Dimethoxy-4- siitrophenyl)- -2--aminopropane (ra-
cemic DON) was first synthesized more than ten -years The relevant molecular—electronic structure indices of
ago and found to be equipotent with the psychoactive 4- DON are presented in Table I with the calculated and
methyl analogue (DOM) in an apparently non-specific rat experimental receptor affinities and the corresponding data
behavioral assay [13]. The fact that,the nitro group is. _for the 4-bromo, -methyl and--ethyl analogues. It is clear
distinctly hydrophilic (a = — 0.28 vs. 0.56 for methyl, that the 4-nitro compound has superdelocalizability and
0.86 for Br, and 1.02 for ethyl [14]) and that DON itself. charge indices which are, very close to those of the highly
has an octanol—water partition coefficient which is low’ potent DOB. Consequently, the introduction of these values
in relation to those of its strongly hallucinogenic 4-methyl, into the appropriate QSAR equation [2]:
-bromo, and -ethyl analogues DOM, DOB, and DOEt
(log P = 1.74 vs. 2.24, 2:54, and 2.76, respectively) and pA, = 14.8645 + 2.4328 SP + 5.5463 Qs + 8.3527 Qs
is not as effective as these on serotonin receptors in the (where S¥ is’'the electrophilic superdelocalizability at
sheep umbilical artery (log RBR = 0.67 vs. 1.00, 1.57, C-2 and QO, and Qg, are the net charges at C-5 and C-8,
and 1.59, respectively) [12], suggested that this compound respectively) predicts a receptor affinity for DON which
might not be a very potent hallucinogen in humans. Much is of the same order as that calculated for DOB and very
more recently, however, it has been shown that DON close to the experimental value determined for the more
produces stimulus generalization in rats trained
to: discri- active (R)-(—) isomer of DON. It should be stressed that
minate DOM from saline, arid that its affinity for the rat the experimental pA, values for DON are affected by
stomach fundus serotonin receptor lies in the same range particularly large standard deviations [3]. Also, the QSAR
as those of the highly potent hallucinogenic PAA - [3]. was developed using receptor affinities determined for
DON had not been included among the compounds used fourteen achiral phenylethylamines, four pairs of enantio-
to generate the QSAR for. PAA cited above [2]. It thus meric phenylisopropylamines and the less active (S)-(++)
seemed of interest to determine if this QSAR could predict . enantiomer of DOB (the pA, value for (R)-DOB is not
the experimentally determined pA, receptor affinity, and available, presumably due to its strong agonist properties
if the correlation between this in. vitro property and hallucino-
genic potency in humans [3] also held for. this substance.
4).
The electronic structure of DON was obtained using
Table I. Structure indices and calculated and experimental rat stomach
Molecular Orbital Theory at the CNDO/2 level [15], employ- fundus serotonin receptor binding affinities of 1-(2,5-dimethoxy-4-
ing the geometry described in our previous QSAR study [2J. "X-phenyl)-2-aminopropanes.
The electron density maps were obtained for the highest
occupied molecular orbital (HOMO) and the second highest xX SE Qs Qs pAz calc. . pAa exp. [3]
occupied molecular orbital (SHOMO) from the eigenvectors
of DON [16]. For the benefit of comparison, we made NOs — 3.8702 0.1941 0.1196 7.52 7.49 + 0.26
the same computations for 1-(4-bromo-2;,5-dimethoxyphenyl)- 7.07 + 0.47»
Br — 3.9632 0.1909 0.1194 7Al 7.35 + 0.08>
2-aminopropane (DOB), 1-(2,5-dimethoxy-4-methylphenyl)- oe ns 6.93
+ 0.08¢
2-aminopropane (DOM), and 1-(4-ethyl-2,5-dimethoxyphe- CHs “i+ 4.0431 °° 0.1457 0.1148 ~ 6.79 7.15 + 0.134
nyl)-2-aminopropane (DOEt). All the calculations — were a 6.41
+ 0.08¢
carried out for the N-protonated | forms. CaHs oe 3.9916. 0.1450 0.1190 6.95 7.18 + 0.09»
‘f

“( R)- (.
oF me oo »(),
. oe
Chemistry and Pharmacology “(S)- CH).

(£)}DON was synthesized as described previously [3].

Since the nitrate was found to be easier, to purify than the ~ The correlation found by Glennon, et al. [3] between the
hydrochloride, the former salt..was used, in the clinical pA, value in the rat stomach fundus and the total hallucino-
trials. (+)-DON nitrate was subjected to. preliminary genic dose in humans (THD), predicts a THD of 1.94 umol
screening by.eight normal volunteers familiar with the.effects of DON (0.5 mg of free base or 0.6 mg of the nitrate) from
of psychotropic substances, following a published procedure our calculated receptor affinity. Even assuming that the
[17]. Once the active dose level had been established, the contribution of the (S)-(+-)-isomer to this activity is negli-
number of subjects was increased, to a total of 16 (12 male, gible, these, values suggest that DON may be a potent
4 female, aged 21—51 years), with a total of 30 trials. .The hallucinogen, if its relatively poor lipid solubility. does not
‘double conscious’ technique of Alles et a/. and Shulgin ez al. prevent it from reaching the CNS. efficiently. Acute oral
[18—20] was used throughout. Some trials were conducted administration of a total dose of 2 mg of the racemic nitrate
individually and some in -group, sessions in: an informal (6.6 wmol) results, after a period of a little more than 1 h
 435

in an amphetamine-like stimulated state. When 3 mg involved in these studies, however, do not allow any definitive
(10 wmol) are taken, these symptoms are maintained but interpretation. It would therefore be of interest to determine
stomach cramps and anxiety appear, the volunteers declaring its potencies (and those of related compounds) by different
that the general feeling is suggestive of-the beginning of a routes of administration circumventing gastrointestinal
hallucinogenic experience. Ata total dose of 4.5 mg (15 wmol), absorption and the blood—brain_ barrier.
the malaise is minimal or absent, and a psychedelic (not Figs. 1—4 show the electron density maps for the HOMO
psychotomimetic) state ensues, with intense visual hallucina- and the SHOMO of DON and DOB. Inspection of these
tions, enhanced perception of color, slight hyperthermia figures reveals that the SHOMO of DON (Fig. 2) is very
and often a desire for physical activity. Visual and some- similar.to the HOMO of DOB (Fig. 3). Also, the HOMO
times auditive distortions persist for about 8h and amphet- of DON (Fig. 1) resembles the SHOMO of DOB (Fig. 4).
amine-like stimulation is usually present at least into the The HOMO. and SHOMO of DOB are separated by about
14th h, although some subjects also reported longer-lasting 0.9 eV in the CNDO/2 calculation, and the corresponding
psychedelic experiences. levels of DON are closer still. The photoelectronic specira
Our calculation of a pA,, similar to the experimental of these substances show that the two lowest ionization
value for (R\(—)-DON, using a theoretical correlation potentials differ by 0.98 eV for DOB [25] and 0.73 eV
equation (i.¢., containing no empirical lipophilicity para- for DON (R. Gleiter, B. K. Cassels and J. 8. Gomez-Jeria,
meters), supports. the suggestion that the lipophilicity of
the PAA C-4 substituent has little effect on receptor affinity
[3]. It must be kept in mind that the hydrophobic constant
z increases with group size and is correlated with molecular
weight and van der Waals volume [21]. It is therefore of
interest that empirical equations, developed using the sum
of either the van der Waals volumes or the molar refractions
of the ring substituents in PAA (the sum of the substituent
zm parameters gave no significant correlation [22]), can
predict the serotonin receptor affinity of DON extremely
well. It seems very likely that this receptor has a region
controlling only the size (and presumably the orientation)
of the substituent at PAA C-4, but not its lipophilicity,
and that the electronic effect of this atom or group is
paramount in determining the pA, value of the drug.
Regarding in vivo properties, it is worth noting that a
hydrophilic group at the C-4 of a PAA has no crucial
influence on the transport properties of the drug, which
reaches its central sites of action in effective concentrations
even when log P is as low as.1.74 (cf. 2.24 for DOM, 2.54
for DOB, and 2.76 for DOEt [12]). More striking in this
Fig...1. Electron density map of the HOMO of DON in the plane
context is the fact that mescaline should have any central Z=0.5 A (1073 e).
effect at all with its logP of 0.78 [12]. It should be remembered
that the correlation of pA, with THD. [3],.which we have
used to predict.the activity of DON, is based on a set of
data points which includes (+)-DOB and its. 4-unsubstituted
analogue. It has recently, been shown that the hallucinogenic
potencies of these two substances are much less than had
been: generally acknowledged [23, 24]. Furthermore, the
literature source [11] of the human pharmacological data
indicates that the potencies of (—)-and (+)-DOM. and of
(+)-DOEt are lower than the figures used: to derive the
correlation [3]: Therefore, any prediction of THD values
made in this way probably errs on the low side. The predicted
value for DON, from our calculated pA, receptor affinity,
is 1.9 umol. This suggests that, if the (S)-(--) isomer contri-
butes little to its activity, (+.)-DON nitrate should be psycho-
active at doses of 1.2 mg or slightly less, while our experi-
mental (acute, oral) THD. for (+)-DON nitrate. is about
4.5 mg. This discrepancy. probably reflects the above
mentioned source of error in obtaining the correlation
between pA, and THD [3], but it may also be largely
explained by the unfavorable partition constant of DON Fig. 2. Electron density map of the SHOMO of DON in the plane
or by more efficient metabolic disposal. The uncertainties Z=0.5 A (10° e).
436

unpublished results). It therefore seems possible that the

energy levels are actually reversed.
Figs. 5 and 6 show, respectively, the HOMO and SHOMO
of protonated serotonin. Both orbitals appear to be centered
in the aromatic system exhibiting z character. If one dis-
regards the C-4 substituent, the SHOMO of DON (Fig. 2)
and the HOMO of DOB (Fig. 3) are quite similar to the
HOMO of serotonin (Fig. 5). This resemblance could
explain, in a first approach, the high serotonin receptor
affinities of DON and DOB [3].
The HOMO of DON (Fig. 1) is centered on the oxygen
atoms.of the nitro group and has x character. The SHOMO
of DOB (Fig. 4) flanks the bromine atom, and corresponds
to a 3p, lone pair of electrons that might participate in a
charge transfer interaction. Thus, the high electron density
available at the 4-substituents of DON and DOB and,
Fig. 5. Electron density map of the HOMO of serotonin in the plane
Z=0.5 A (10-3 e).

Fig. 6. Electron density map of the SHOMO of serotonin in the plane

Z=0.5 A (10° e).
Fig. 3. Electron density map of the HOMO of DOB in the plane
Z=0.5 A (10-3 e).

presumably, of the very potent DOI [11], but not of DOM,

DOEt or in serotonin, might also favor receptor binding.
The main conclusions of this work are: 1) (4+)-DON is
a very potent hallucinogen when taken orally, and its
activity is well correlated with its pA, in the rat stomach
fundus preparation [3]. 2) The theoretical QSAR developed
for PAA [2] is a good predictor of the pA, of (R)(—)-DON
in the same system [3]. 3) Although the hydrophilicity of
the nitro group and the consequent low octanol—water
partition coefficient of DON do not seem to diminish its
affinity for the appropriate receptor(s) in the brain, they
may reduce its access to the CNS to an appreciable extent.
4) The similarity of the SHOMO and HOMO of DON
and DOB, respectively, to the HOMO of serotonin could be
implicated in their high serotonin receptor affinities. The
corresponding orbitals of DOM and DOEt (not shown)
also resemble those of serotonin in their electron density
distribution. 5) The 4-substituents of hallucinogenic PAA
appear to enhance receptor binding by modulating the
Fig. 4. Electron density map of the SHOMO of DOB in the plane electron distribution in the aromatic ring, primarily at
Z = 0.5 A (10-3 e). their point of attachment, and also by a direct interaction
 437

with a complementary area of the receptor surface, a situation Cohen M.L., Schenck K. W., Colbert W. & Wittenauer L.
which cannot occur with the 7-unsubstituted serotonin. (1985) J. Pharmacol. Exp. Ther. 232, 770
Cohen M. L. & Wittenauer L. (1985) J. Pharmacol. Exp. Ther.
6) The nitro group and the bromine atom of DON and DOB 233, 75
have high electron densities contributing to the occupied Clineschmidt B. V., Reiss D. R., Pettibone D. J. & Robinson J. L.
frontier molecular orbitals. These electron-rich regions (1985) J. Pharmacol. Exp. Ther. 235, 696
have no counterpart in serotonin, but they may nevertheless Cohen M.L. & Wittenauer L. (1986) Life Sci. 38, 1
10 Shannon M., Battaglia G., Glennon R. A. & Titeler M. (1984)
add to the pharmacological potency of DON and DOB. Eur. J. Pharmacol. 102, 23
To extend the results of this work in such a way as to i1 Shulgin A.T. (1978) in: Handbook of Psychopharmacology,
clarify the influence of the 4-substituent of PAA on both Vol. 11 (Iversen L. L., Iversen S. & Snyder S. H., eds.), Plenum
the receptor affinities and the pharmacologic activities of Press, New York, p. 243
12 Nichols D. A., Shulgin A. T. & Dyer D. C. (1977) Life Sci. 21,
hallucinogens belonging to this structural class, it will be 569
necessary to test more 1-(2,5-dimethoxy-4-X-phenyl)-2- 13 Coutts R.T. & Malicky J. L. (1973) Can. J. Chem. 51, 1402
aminopropane analogues with a broader range of X groups, 14 Hansch C., Leo A., Unger S. H., Kim K.H., Nikaitani D. &
especially those with hydrophilic characteristics. Lien E. J. (1973) J. Med. Chem. 16, 1207
15 Pople J. A. & Beveridge D.L. (1970) in: Molecular Orbital
Theory, McGraw Hill Book Co., New York, pp. 75—79
16 Gomez-Jeria J.S. & Venegas-Garrido R. (1983) in: MAPDEN:
Acknowledgements A FORTRAN IV Program for the Calculation and Drawing of
the Electronic Density Distribution of Molecular Orbitals from
This work received financial support from the University of Chile Semiempirical Calculations, Universidad de Chile, Santiago,
(DIB Project Q 2442) and from the Chilean National Research Fund Chile
(Project 1075). One of us (J.S.G.-J.) thanks Drs. J. Maruani and 17 Jacob P. Tl & Shulgin A. T. (1983) J. Med. Chem. 26, 746
A. Toro-Labbé for stimulating discussions. 18 Alles G. A. (1959) in: Neurepharmacology (Abramson H. A., ed.),
Madison Printing Co., Madison, N.J., pp. 128—152
19 Shulgin A. T., Sargent T. W. & Naranjo C. (1969) Nature (London)
221, 537
References 20 Shulgin A.T., Shulgin L.A. & Jacob P. Ill (1986) Methods
Find. Exp. Clin. Pharmacol. 8, 313
1 Gomez-Jeria J.$. (1983) Int. J. Quant. Chem. 23, 1969 21 Yang G.-z., Lien E. J. & Guo Z.-r. (1986) Quant. Struct.-Act.
2 Gémez-Jeria J. $8. & Morales-Lagos D. (1985) in: QSAR in Design Relatsh. Pharmacol. Chem. Biol, 5, 12
of Bioactive Compounds (Kuchar M., ed.), J. R. Prous, Barcelona, 22 Gupta S. P., Singh P. & Bindal M. C. (1982) Farmaco Ed. Sci.
p. 145 37, 566
3 Glennon R. A., Young R., Benington F. & Morin R. D. (1983) 23 Shulgin A. T. (1981) J. Psychoact. Drugs 13, 99
J. Med. Chem. 25, 1163 and references cited therein 24 Cassels B. K. & G6mez-Jeria J.S. (1985) J. Psychoact. Drugs
4 Nichols D. E., Shooler D., Yeung M.C., Oberlender R.A. & 17, 129
Zabik J. E. (1984) Life Sci. 35, 1343 25 Domelsmith L. N., Eaton T. A., Houk K. N., Anderson G. M. III,
5 Freedman D. X. (1986) Pharmacol. Biochem. Behav. 24, 407 Glennon R. A., Shulgin A. T., Castagnoli N. Jr. & Kollman P. A.
and references cited therein (1981) 4. Med. Chem, 24, 1414