Structure-property relationships in the basicity and lipophilicity of arylalkylamine oxides

Article abstract of DOI:10.1002/(SICI)1522-2675(19991006)82:10<1630::AID-HLCA1630>3.0.CO;2-P

Homologous N,N-dimethyl-phenylalkylamine oxides and N,N-dimethyl- diphenylalkylamine oxides were prepared. Their basicity and lipophilicity (octan-1-ol/H₂O) were compared to those of the parent amines. In contrast to the amines, the basicity of all N,N-dimethyl-arylalkylamine oxides showed very limited pK(a) variations (range 4.65-5.01) with increasing chain length and number of Ph groups. The N-oxides in their neutral form had a log p(N) value lower by 2.77 ± 0.34 (n = 9) units than that of the parent amine. The log P(C) of the cationic N,N-dimethyl-diphenylalkylamines was lower than that of their neutral form, with a decrement diff(log P(N-C)) that increased from 3.25 to 4.21 in the homologous series. Unexpectedly, the decrement diff(log P(N-C)) for the N-oxides was much smaller than for the tertiary amines, being 0.23 for the aliphatic N,N-dimethyl-pentylamine oxide, 0.47 ± 0.13 for the phenylalkylamine oxides, and 0.80 ± 0.07 for the diphenylalkylamine oxides. In fact, the protonated N-oxides had log P(C) values that were quite comparable to those of the protonated parent amines. Because of the differences in basicity, the difference in distribution coefficients at physiological pH (log D^7.4) between a tertiary arylalkylamine and its N- oxide was 0.82 ± 0.66 (n = 9). The pharmacokinetic implication is that N- oxygenation may have a smaller effect on the urinary excretion of tertiary amines than usually assumed.

Full text of DOI:10.1002/(SICI)1522-2675(19991006)82:10<1630::AID-HLCA1630>3.0.CO;2-P

1630
Helvetica Chimica Acta ± Vol. 82 (1999)
Structure-Property Relationships in the Basicity and Lipophilicity of
Arylalkylamine Oxides
by Giulia Carona)b), Giuseppe Ermondia), Donatella Boschia), Pierre-Alain Carruptb), Roberta Frutteroa),
b
a
1
Bernard Testa ) )*, and Alberto Gasco )
a
Á
) Dipartimento di Scienza e Tecnologia del Farmaco, Universita di Torino, Via P. Giuria 9, I-10125 Torino
b
Â
Â
) Institut de Chimie Therapeutique, Section de Pharmacie, Universite de Lausanne, CH-1015 Lausanne
Homologous N,N-dimethyl-phenylalkylamine oxides and N,N-dimethyl-diphenylalkylamine oxides were
prepared. Their basicity and lipophilicity (octan-1-ol/H2O) were compared to those of the parent amines. In
contrast to the amines, the basicity of all N,N-dimethyl-arylalkylamine oxides showed very limited pKa
variations (range 4.65 ± 5.01) with increasing chain length and number of Ph groups. The N-oxides in their
neutral form had a log PN value lower by 2.77 Æ 0.34 (n ˆ 9) units than that of the parent amine. The log PC of the
cationic N,N-dimethyl-diphenylalkylamines was lower than that of their neutral form, with a decrement
diff(log PN C ) that increased from 3.25 to 4.21 in the homologous series. Unexpectedly, the decrement
diff(log PN C ) for the N-oxides was much smaller than for the tertiary amines, being 0.23 for the aliphatic N,N-
dimethyl-pentylamine oxide, 0.47 Æ 0.13 for the phenylalkylamine oxides, and 0.80 Æ 0.07 for the diphenylalkyl-
amine oxides. In fact, the protonated N-oxides had log PC values that were quite comparable to those of the
protonated parent amines. Because of the differences in basicity, the difference in distribution coefficients at
physiological pH (log D7.4) between a tertiary arylalkylamine and its N-oxide was 0.82 Æ 0.66 (n ˆ 9). The
pharmacokinetic implication is that N-oxygenation may have a smaller effect on the urinary excretion of tertiary
amines than usually assumed.
1. Introduction. ± Many reactions of metabolism facilitate the excretion of
xenobiotics by transforming them into more polar (less lipophilic) metabolites. Thus,
tertiary amines are an important group of xenobiotics (e.g., most alkaloids) against
which terrestrial animals had to evolve protection and elimination strategies.
N-Oxygenation is one of the major metabolic reactions of tertiary amines [1].
Depending on the physicochemical properties of the substrate, the reaction is catalyzed
by cytochrome P450 and/or flavin-containing monooxygenases. Tertiary amines
undergoing metabolic N-oxygenation are either strong bases (i.e., alkylamines,
arylalkylamines, and alicyclic amines) or moderate bases (i.e., heteroaromatic and
aromatic amines). The N-atom must also be sterically accessible, meaning, for
example, that N,N-dimethyl-alkylamines are far better substrates than N,N-diethyl-
alkylamines [1].
In pharmacokinetic terms, the N-oxygenation of drugs and other xenobiotics affects
their distribution and excretion in a manner that is highly dependent on the changes in
physicochemical properties produced by the metabolic reaction. It is recognized that
the N-oxygenation of tertiary amines is accompanied by a decrease in basicity and
lipophilicity [2]; however, the paucity of data available lacks systematic character [3].
In a recent study [4], we compared the lipophilicity of 4-substituted pyridines and their
1
)
Fax: ‡ 41216924525; e-mail: bernard.testa@ict.unil.ch

Helvetica Chimica Acta ± Vol. 82 (1999)
1631
N-oxides, and found that the decrement in lipophilicity was proportional to the partial
charge on the O-atom (H-bond acceptor basicity), itself inversely proportional to the
electron-withdrawing capacity of the 4-substituent.
Here, we extend this work to basic tertiary amines, namely N,N-dimethyl-
phenylalkylamines and N,N-dimethyl-diphenylalkylamines with varying chain lengths
(Fig. 1). Many drugs contain one of these two structural motifs. N,N-Dimethylpentyl-
amine (1) was included in the study to help assess the influence of the Ph ring(s).
Besides the lipophilicity of the neutral form, we also succeeded in measuring the
lipophilicity of the cationic form of several compounds. The data were then interpreted
in terms of structure-property relationships, revealing the complexity of the intra-
molecular effects influencing the basicity and pH-lipophilicity profile of the com-
pounds.
Fig. 1. Structures of the investigated compounds
2. Results and Discussion. ± 2.1. Basicity of Amines. The pKa values of the four
phenyl-alkylamines, 2 ± 5, varied between 8.80 and 9.90 and increased linearly (r2 ˆ
0.94, Eqn. 1) with increasing distance between the amino and the Ph group (Eqn. 1,
Table 1, and Fig. 2):
pKa ˆ 0.34(Æ 0.13) ´ nCH2 ‡ 8.55(Æ 0.36)
n ˆ 4; r2 ˆ 0.94; s ˆ 0.14; F ˆ 30
(1)
where n is the number of CH2 groups in the molecule. In this and the following
equations, 95% confidence limits are given in parentheses, n is the number of
compounds, r2 the square of the correlation coefficient, s the standard deviation, and F
is Fischerꢀs test.

1632
Helvetica Chimica Acta ± Vol. 82 (1999)
Fig. 2. Basicity (expressed as pKa) of amines and corresponding N-oxides as a function of chain length and
number of Ph groups
Very similar results have been reported for primary phenyl-alkylamines and
pyridyl-alkylamines [5]. The same trend was seen here for the four diphenyl-
alkylamines 6 ± 9 (Fig. 2), but the effect of the benzhydryl group was stronger than that
of the Ph group, with a range of 7.62 to 10.06 (Eqn. 2):
pKa ˆ 0.80(Æ 0.28) ´ n CH2 ‡ 7.04(Æ 0.79)
n ˆ 4; r2 ˆ 0.95; s ˆ 0.30; F ˆ 36
(2)
These intramolecular interactions are not novel and require no further discussion.
2.2. Basicity of N-Oxides. In sharp contrast to the tertiary amines 2 ± 9, the
corresponding N-oxides 2a ± 9a showed very limited pKa dependence (range 4.65 ±
5.01) on increasing chain length and number of Ph groups (Fig. 2). This is a clear
‡
indication that the strongly polar character of the N
O group and the resulting
basicity of the O-atom are not strongly influenced by the aromatic groups. Indeed, the
pKa of the alkylamine N-oxide 1a is 4.88, as compared to the average pKa of all nine N-
oxides (4.81 Æ 0.13).
Because of the large pKa differences between the tertiary amines, and the small pKa
differences between their N-oxides, the change in pKa resulting from N-oxygenation
(pKa(amine) pKa(oxide)) varies considerably. It follows that the pKa value of an
arylalkyl- or alkylamine N-oxide cannot be deduced from that of the parent amine, but
can be expected to be close to 4.8.
2.3. Lipophilicity of Neutral and Protonated Tertiary Amines. The octan-1-ol/H2O
log PN values (experimental log P of the neutral amines) increased almost linearly in
the two homologous series, the average increment per CH2 unit being 0.39 Æ 0.04 and
0.51 Æ 0.27 in the monophenyl and diphenyl series, respectively (Fig. 3,a). The
agreement with log P values calculated by the CLOGP algorithm [3] is good, with a
slope not significantly different from 1, and an intercept equal to 0 (Eqn. 3 and
Fig. 3,b):
log PN ˆ 1.23(Æ 0.28) ´ CLOGP 0.25(Æ 0.79)
n ˆ 8; r2 ˆ 0.94; s ˆ 0.28; F ˆ 91
(3)

Helvetica Chimica Acta ± Vol. 82 (1999)
1633
The phenyl-alkylamines in their cationic form were too hydrophilic for their log P
(indicated as log PC) to be measurable by potentiometry (log PC < 1). In contrast, the
cationic diphenyl-alkylamines were lipophilic enough to have a measurable log PC.
Here, the average increment per CH2 unit was 0.19 Æ 0.25 (Fig. 3,c).
Fig. 3. Lipophilicity plots: a) Variation in the log P value of neutral amines and corresponding N-oxides (log PN)
as a function of chain length and number of Ph groups. b) Linear relationships between calculated (CLOGP) and
experimental (log PN) data. Amines are represented by a full circle (.), N-oxides by an empty circle (*).
c) Variation in the log P of cationic amines and corresponding N-oxides (log PC) as a function of chain length and
number of Ph groups. d) Linear relationships between the log PN of amines and corresponding N-oxides
An informative parameter in structure-property relationships is the decrement in
lipophilicity resulting from protonation, designated as diff(log PN C). This decrement
was not constant but increased from 3.25 to 3.51 to 3.79 to 4.21 in the homologous series
of diphenyl-alkylamines. In fact, the decrement increased by 0.32 Æ 0.09 per CH2 unit.
This clearly indicates that the effect of an increase in chain length on the lipophilicity of
diphenyl-alkylamines is not felt similarly in the neutral and protonated amines.
2.4. Lipophilicity of Neutral N-Oxides. The log P values of the neutral N-oxides
increased linearly in the two homologuous series (monophenyl compounds: r2 ˆ 0.98,
increment ˆ 0.34 Æ 0.14; diphenyl compounds: r2 ˆ 0.94, increment ˆ 0.21 Æ 0.13)
(Fig. 3,a). The linear relationship between these experimental data and values
calculated by the CLOGP algorithm is shown in Fig. 3,b, and is described by Eqn. 4:
log PN ˆ 0.84(Æ 0.15) ´ CLOGP ‡ 0.09(Æ 0.13)
n ˆ 8; r2 ˆ 0.95; s ˆ 0.17; F ˆ 133
(4)

1634
Helvetica Chimica Acta ± Vol. 82 (1999)
A parameter of high relevance here is the decrement in lipophilicity resulting from
N-oxidation of the neutral amines (log PN(amine) log PN(oxide)) (Table 1). For
compounds 1a ± 7a, this decrement was 2.61 Æ 0.11 (n ˆ 7), but it became 2.77 Æ 0.34
(n ˆ 9), when the much larger decrement of compounds 8a and 9a was included. A plot
(Fig. 3,d) of log PN(amine) vs. log PN(oxide) indicates a linear trend (Eqn. 5) with
diaryl compounds being more dispersed than monophenyl compounds:
log PN (amine) ˆ 1.28(Æ 0.28) ´ log PN (oxide) ‡ 2.68(Æ 0.21)
n ˆ 8; r2 ˆ 0.93; s ˆ 0.27; F ˆ 97
(5)
Interestingly, a log PN(amine) log PN(oxide) value of only 1.91 had been found for
pyridine N-oxide [4]. This value expresses the less polar character of aromatic N-oxides
compared to aliphatic N-oxides, and it can be explained by resonance effects.
2.5. Lipophilicity of Protonated N-Oxides. All nine N-oxides were sufficiently
lipophilic when protonated to yield a measurable log PC value (Table 1). When
measuring the log P values of ions, one should not neglect the possible influence of
counterions. Such a phenomenon can best be seen when the log P value of an ion is
strongly influenced by the ionic strength m in the aqueous phase. To examine this effect,
we compared the log PN and the log PC values of the N-oxides 5a and 9a at m ˆ 0.15,
Table 1. Physicochemical Properties of Tertiary Phenyl-alkylamines and Diphenyl-alkylamines and Corresponding N-Oxides
a
pKa
)
pKa(amine)
pKa(oxide)
CLOGPb) log PNc
)
log PC d
)
diff(log PN C )e)
Amine N-oxide
log PN(amine)
log PN(oxide)f)
log PC(amine)
log PC(oxide)g)
1
1a
2
2a
3
3a
4
4a
5
5a
6
6a
7
7a
8
10.30
4.88 5.42
8.80
4.72 4.08
9.40
4.65 4.75
9.52
4.78 4.74
9.90
4.88 5.02
7.62
5.01 2.61
8.91
4.65 4.26
9.60
2.00
0.32
1.96
0.58
1.84
0.49
2.28
0.04
2.89
0.57
3.22
0.68
3.04
0.72
3.57
1.25
4.10
1.78
2.33
NDi)
1.98
NMh)
0.55
NM
0.94k)
0.23j)
0.37
0.37
0.50
2.45j)
2.55
2.66
2.77
2.69
2.55
2.57
3.21
3.46
0.57k)
2.38
0.28k)
2.73
0.04k)
3.15
0.46
3.37
0.82
3.58
1.01
NM
0.65k)
NM
0.54k)
NM
0.19
0.12
0.06
0.07
0.10
0.52
0.35
0.70
0.66
0.65
3.25
0.76
0.06
0.03
0.17
0.04
3.51
0.91
4.31
1.10
4.91
1.45
3.79
8a
9
9a
4.77 4.83
10.06
4.94 5.12
0.75
4.21
0.79
a) Determined by potentiometry, SD < 0.01; n ˆ 3. The average pKa of all nine N-oxides 4.81 Æ 0.13. b) Calculated with the
c
CLOGP algorithm [3]. ) log P of the neutral species determined by potentiometry, SD < 0.05; n ˆ 4. d) log P of the cationic
species determined by potentiometry, SD < 0.05; n ˆ 4. e) log P of the neutral form minus log P of the cationic form. This
f
decrement is 0.47 Æ 0.13 for the N-oxides 2a ± 5a, and 0.80 Æ 0.07 for the N-oxides 6a ± 9a. ) This decrement is 2.61 Æ 0.11 for
compounds 1a ± 7a (n ˆ 7). It is 2.77 Æ 0.34 (n ˆ 9) for compounds 1a ± 9a (n ˆ 9). g) The average difference is 0.06 Æ 0.08;
i
j
n ˆ 4. h) NM ˆ not measurable by potentiometry (log PC < 1). ) ND ˆ not determined. ) The CLOGP value was taken as
log PN. k) Determined by the shake-flask method and used to check the potentiometric result (see text for details).

Helvetica Chimica Acta ± Vol. 82 (1999)
1635
m ˆ 0.10, and m ˆ 0.05. As shown in Table 2, there were only minute fluctuations in pKa
and log P values. Also, the variations in log PC values did not reveal any trend and
remained within the experimental error. We conclude that ionic strength does not
detectably affect the properties of the compounds under the present conditions.
Table 2. Influence of Ionic Strength on pKa and log Pa)
0.05m KClb)
0.10m KClc)
0.15m KCld)
pKa
log PN
log PC
pKa
log PN
log PC
pKa
log PN
log PC
5a
9a
4.90
4.87
0.48
1.48
0.35
0.47
4.84
4.86
0.45
1.45
0.15
0.67
4.88
4.94
0.46
1.45
0.19
0.66
a) Determined by potentiometry, SD < 0.01; n ˆ 3. b) m ˆ 0.05. c) m ˆ 0.10. d) m ˆ 0.15.
As deduced from data in Table 1, the decrement in log P resulting from protonation,
diff(log PN C), was much smaller for the N-oxides than for the tertiary amines. Indeed,
this decrement was 0.23 for the aliphatic amine oxide 1a, 0.47 Æ 0.13 for the phenyl-
alkylamine oxides 2a to 5a, and 0.80 Æ 0.07 for the diphenyl-alkylamine oxides 6a ± 9a.
As a rule, the difference between the lipophilicity of neutral and cationic amines covers
a range from 3 to 4 in the octan-1-ol/H2O system [6]. Thus a normal value of
diff(log PN C ) (average value ca. 3.7) is observed for amines 6 ± 9, whereas the
lipophilicity of the N-oxides decreases only modestly upon protonation.
This latter finding is postulated to be due to a polarity effect, and more specifically
to the different manner in which protonation alters the polarity of amines and N-oxides.
To test this hypothesis, we used quantum mechanics at the ab initio 6-31G** level to
calculate ESP partial charges and N O interatomic distances in the model compounds
‡
Me3N, Me3N O , and their protonated forms. As seen in Table 3, neutral Me3N is
only slightly polar, but is transformed by protonation into a large cation rendered
highly hydrophilic by a well-distributed positive charge. This change differs markedly
‡
from that occurring with Me3N O (Table 3). Here, the N-oxide moiety is seen to be
strongly polarized and highly polar, and protonation has only limited effect on the
partial charges carried by the N- and O-atoms and on their interatomic distance. In
other words, protonation cannot increase strongly the hydrophilicity of an already
highly polar N-oxide.
Because of the vastly different effects of protonation on the log P of amines and the
corresponding N-oxides, the difference in lipophilicity between the protonated amines
Table 3. Charge Distribution and N O Interatomic Distance in Model Compounds, as Calculated at the 6-31G**
Level
Model compounds
Atomic charges
Distance N O [Š]
Me
N
O
H
Me3N
Me3N
Me3N
0.14
0.02
0.17
0.19
0.42
0.71
0.25
0.52
‡
O
H
0.77
0.54
1.3665a)
1.3839a)
‡
‡
0.25
0.47
Me3N OH
a)
The corresponding distance in Me2N OH is 1.3907 Š.

1636
Helvetica Chimica Acta ± Vol. 82 (1999)
and their protonated N-oxides (0.06 Æ 0.08; n ˆ 4) is negligible (Table 1). It even
appears that the protonated N-oxides 1a ± 5a may be more lipophilic than the
protonated form of their parent amines 1 ± 5. These observations point to the similarity
of protonated amines, neutral N-oxides and protonated N-oxides as far as their polar
intermolecular interactions are concerned.
2.6. Distribution Coefficients (Apparent Lipophilicity) of the Amines and their N-
Oxides. Distribution coefficients (expressed as log D) can be calculated for an amine at
any pH value knowing its pKa, log PN, and log PC values. log D Values often yield better
correlations with pharmacokinetic parameters than log P, presumably because the
influence of all ionization states is taken into account. To obtain some pharmacokineti-
cally relevant data from the results in Table 1, we calculated the log D values of the N-
oxides at pH 2.0, 5.0, and 7.4 (Table 4). At pH 7.4, the tertiary amines are mostly
protonated, whereas the N-oxides are mostly neutral. As a result, log D7.4(amine) value
is higher than log D7.4(oxide) value by 0.82 unit (n ˆ 9; SD ˆ 0.66) (Table 4), whereas
log PN(amine) is higher than log PN(oxide) by 2.77 units (n ˆ 9; SD ˆ 0.34) (Table 1).
This comparison indicates, in pharmacokinetic terms, that N-oxygenation may have a
smaller effect on the polarity and urinary excretion of tertiary amines than usually
assumed.
3. Conclusions and Perspectives. ± This study sheds light on some intramolecular
interactions influencing the basicity of arylalkylamine oxides, and especially their
lipophilicity in the neutral and protonated state.
Table 4. Distribution Coefficients of Amines and Corresponding N-Oxidesa)
log D2.0
log D5.0
log D7.4
log D7.4(amine) log D7.4(oxide)
b
b
1
1a
2
2a
3
3a
4
4a
5
5a
6
6a
7
7a
8
)
)
0.57
0.32
0.56
0.57
0.38
0.28
0.61
0.06
0.65
0.46
2.94
0.82
2.06
1.01
2.12
1.10
2.26
1.45
0.55
0.39
0.25
1.13
0.66
0.67
0.19
2.12
1.05
1.02
0.81
b
b
)
)
0.90
0.67
b
b
)
)
0.64
0.38
b
b
)
)
0.53
1.04
0.18
0.12
0.06
0.07
0.10
0.52
0.35
0.70
0.66
0.19
0.96
0.29
0.84
0.58
0.22
0.87
0.58
0.94
0.76
1.24
8a
9
9a
a) Calculated according to the following equation:
ꢀ
ꢁ
ꢀ
ꢁ
pH
a
10 pK
a
1
D ˆ PN
Á
‡ PC
Á
:
1 ‡ 10 pK
1 ‡ 10 pK
pH
pH
a
b) Not calculable due to the non-availability of a log PC value.

Helvetica Chimica Acta ± Vol. 82 (1999)
1637
The lack of variation in the pKa values of monophenyl-alkylamine N-oxides and
diphenyl-alkylamine oxides, as well as the weak log P differences between their neutral
and cationic forms, suggest that the ionizable and strongly polarized N-oxide moiety
has an overwhelming influence on these physicochemical properties. From a
pharmacokinetic perspective, the next step is to investigate the partitioning and
intermolecular interactions of N-oxides in anisotropic media such as micelles and
artificial membranes (liposomes). Work is in progress along these lines.
Experimental Part
General. Anal. grade octan-1-ol was purchased from Fluka Chemie (Buchs, CH). M.p.: Büchi 530
apparatus, uncorrected. 1H-NMR Spectra: Bruker AC-200, in (D6)DMSO, at 200 MHz. Elemental analyses:
within 0.4% unless indicated otherwise, obtained from REDOX (I-Cologno Monzese).
N,N-Dimethylamines 1 [7], 2 [8], 3 [8], 4 [8], 5 [8], 6 [9], 7 [10], 8 [11] were synthesized according to
published methods.
N,N-Dimethyl-4,4-diphenylbutylamine hydrochloride (9) was synthesized according to the procedure for 8
[11]: 0.25 g (11.0 mmol) of Na and 1.29 g (22.0 mmol) of K were vigorously stirred at reflux temp. under N2 in dry
THF (60 ml) until the Na/K alloy was formed (ca 4 h). To the cooled suspension, a soln. of diphenylmethyl
methyl ether [12] (2.30 g, 11.0 mmol) in dry THF (15 ml) was added dropwise at r.t. The mixture was stirred for
16 h, then cooled to 08, and a soln. of 3-chloro-N,N-dimethylpropylamine (2.67 g, 22.0 mmol) in dry THF
(10 ml) was added slowly. The mixture was further stirred for 3 h, the reaction was quenched by slow dropwise
addition of H2O (10 ml), THF was distilled under vacuum, and the residue was dissolved in Et2O. The product
was extracted with 4n HCl, the aq. phase was washed with Et2O, made alkaline with 6n NaOH, and extracted
with AcOEt. The org. phase was dried (MgSO4) and evaporated to afford the crude product, which was purified
by flash chromatography (FC) (silica gel; CH2Cl2/MeOH 9.5 :0.5) to obtain an oil (30%) which was
1
characterized as the hydrochloride. M.p. 145 ± 1488 (i-PrOH /Et2O). H-NMR ((D6)DMSO): 1.64 ± 1.56 (m, 2
H); 2.15 ± 2.03 (m, 2 H); 2.66 (s, 6 H); 3.07 (t, 2 H); 3.97 (t, 1 H); 7.15 ± 7.36 (m, 10 H); 10.65 (br., 1 H);
13C-NMR: 22.5; 31.7; 41.9; 50.1; 56.3; 126.3; 127.7; 128.5; 144.8. Anal. (C18H23N ´ HCl): C, H, N.
General Procedure for the Synthesis of the N,N-Dimethylphenyl-alkylamine N-Oxides 1a ± 5a. The amine
(50 mmol) in MeOH was treated dropwise under stirring at 08 with 30% H2O2 (150 mmol), and the resulting
mixture was stirred at r. t. until the reaction was complete (TLC detection; CH2Cl2/MeOH 9 :1). The solvent
was evaporated, the residue dissolved in H2O, and the resulting soln. was made alkaline with 37% NH4OH and
washed with AcOEt. The aq. phase was evaporated under vacuum and the residue dried in a desiccator under
vacuum over P4O10 and dissolved in MeOH. The soln. was saturated with gas. HCl, evaporated under vacuum,
and the residue was crystallized.
N,N-Dimethylpentylamine N-Oxide Hydrochloride (1a). Yield: 30%. Very hygroscopic oil that solidified in
1
the desiccator; crystallized from dry MeOH/Et2O. H-NMR ((D6)DMSO): 0.90 (t, 3 H); 1.29 ± 1.36 (m, 4 H);
1.70 ± 1.80 (m, 2 H); 3.46 (s, 6 H); 3.62 ± 3.70 (m, 2 H); 12.69 (br., 1 H). 13C-NMR: 13.83; 21.78; 21.89; 27.73;
55.25; 67.91. Anal. (C7H17NO ´ HCl ´ 1.75 H2O): C, H, N.
N,N-Dimethylbenzylamine N-Oxide Hydrochloride (2a). Yield: 60%. Hygroscopic white solid. M.p. 108 ±
1108 (dry MeOH/Et2O). 1H-NMR ((D6)DMSO): 3.17 (s, 6 H); 4.95 (s, 2 H); 7.49 ± 7.65 (m, 5 H); 12.90 (br., 1
H). 13C-NMR: 54.81; 70.18; 128.43; 128.66; 130.26; 133.06. As picrate, m.p. 157 ± 1598 ([13]: 158 ± 158.58).
N,N-Dimethyl-2-phenylethylamine N-Oxide Hydrochloride (3a). Yield: 75%. White solid. M.p. 156 ± 1588
(dry i-PrOH). 1H-NMR ((D6)DMSO): 3.15 (t, 2 H); 3.55 (s, 6 H); 3.96 (t, 2 H); 7.32 ± 7.37 (m, 5 H); 12.94 (br., 1
H). 13C-NMR: 28.49; 55.48; 66.33; 127.03; 128.79; 129.10; 136.20. Anal. (C10H15NO ´ HCl): C, H, N.
N,N-Dimethyl-3-phenylpropylamine N-Oxide Hydrochloride (4a). Yield: 45%. White solid. M.p. 90 ± 918
1
(dry i-PrOH/Et2O). H-NMR ((D6)DMSO): 2.11 ± 2.15 (m, 2 H); 2.66 (t, 2 H); 3.46 (s, 6 H); 3.64 ± 3.72 (m, 2
H); 7.22 ± 7.34 (m, 5 H); 12.71 (br., 1 H). 13C-NMR: 24.09; 31.57; 55.36; 67.77; 126.30; 128.39; 128.55; 140.45.
Anal. (C11H17NO ´ HCl ´ 0.25 H2O): C, H, N.
N,N-Dimethyl-4-phenylbutylamine N-Oxide Hydrochloride (5a). Yield: 50%. White solid. M.p. 112 ± 1148
(dry THF). 1H-NMR ((D6)DMSO): 1.57 ± 1.69 (m, 2 H); 1.76 ± 1.85 (m, 2 H); 2.65 (t, 2 H); 3.46 (s, 6 H); 3.67 ±
3.75 (m, 2 H); 7.20 ± 7.35 (m, 5 H); 12.70 (br., 1 H). 13C-NMR: 21.92; 27.66; 34.60; 55.33; 67.81; 126.01; 128.46;
141.66. Anal. (C12H19NO ´ HCl ´ 0.2 H2O): C, H, N.

1638
Helvetica Chimica Acta ± Vol. 82 (1999)
General Procedure for the Synthesis of the N,N-Dimethyl-w-diphenyl-alkylamine N-Oxides 6a ± 9a. A soln.
of 3-chloroperbenzoic acid (6.3 mmol) in dry toluene was added slowly to a stirred soln. of the amine (5 mmol)
in dry toluene (20 ml) kept at 08. The mixture was stirred at r. t. until all of the amine was oxidized (TLC
detection). Toluene was evaporated under vacuum, the residue was dissolved with Et2O, and the resulting soln.
was saturated with gas. HCl. The precipitated hydrochloride was filtered and crystallized.
N,N-Dimethyldiphenylmethylamine N-Oxide Hydrochloride (6a). TLC: petroleum ether/AcOEt 7:3.
Yield: 61%. White solid. M.p. 184 ± 186 8 (H2O). 1H-NMR ((D6)DMSO): 3.46 (s, 6 H); 6.43 (s, 1 H); 7.42 ± 7.68
(m, 10 H); 13.14 (s, 1 H). 13C-NMR: 55.78; 81.50; 128.02; 129.12; 129.72; 133.76. Anal. (C15H17NO ´ HCl):
C, H, N.
N,N-Dimethyl-2,2-diphenylethylamine N-Oxide Hydrochloride (7a). TLC: CH2Cl2/MeOH 7:3. Yield: 79%.
White solid. M.p. 125 ± 1298 (H2O). 1H-NMR ((D6)DMSO): 3.41 (s, 6 H); 4.55 (d, 2 H); 4.90 (t, 1 H); 7.19 ± 7.53
(m, 10 H); 12.91 (s, 1 H). 13C-NMR: 45.01; 56.43; 71.24; 127.00; 127.80; 126.95; 142.25. Anal. (C16H19NO ´ HCl ´
0.5 H2O): C, H, N.
N,N-Dimethyl-3,3-diphenylpropylamine N-Oxide Hydrochloride (8a). TLC: CH2Cl2/MeOH 7:3. Yield
86%. White solid. M.p. 158 ± 1608 (acetone). 1H- NMR ((D6)DMSO): 2.60 ± 2.67 (m, 2 H); 3.48 (s, 6 H); 3.46 ±
3.58 (m, 2 H); 4.05 (t, 1 H); 7.21 ± 7.42 (m, 10 H); 12.64 (s, 1 H). 13C-NMR: 27.52; 47.61; 55.44; 67.31; 126.60;
127.52; 128.68; 143.66. Anal. (C17H21NO ´ HCl): C, H, N.
N,N-Dimethyl-4,4-diphenylbutylamine N-Oxide Hydrochloride (9a). TLC: AcOEt/MeOH/37% NH4OH
8.5 :1:0.5. Yield: 74%. White solid. M.p. 1488 (i-PrOH). 1H-NMR ((D6)DMSO): 1.5 ± 1.7 (m, 2 H); 2.0 ± 2.1 (m,
2 H); 3.35, 3.36 (2s, 6 H); 3.66 ± 3.74 (m, 2 H); 4.02 (t, 1 H); 7.19 ± 7.37 (m, 10 H); 12.46 (s, 1 H). 13C-NMR: 21.02;
31.32; 49.97; 55.30; 67.75; 126.3; 127.66; 128.58; 144.68. Anal. (C18H23NO ´ HCl): C, H, N.
Potentiometric Determination of Protonation Constants. Potentiometric titrations were performed with the
PCA101 apparatus [14] (Sirius Analytical Instruments Ltd., Forrest Row, East Sussex, UK) as described in [15].
Determination of Lipophilicity Parameters. Potentiometry was the main technique used to determine the
pH-lipophilicity profiles of parent amines and N-oxides. The PCA101 and the new GlpKa instruments (Sirius
Analytical Instruments Ltd.) were used. At least four separate titrations of ca. 1 mm for each compound,
containing various volumes of octan-1-ol (from 1 ml of org. solvent/15 ml of H2O to 8 ml of org. solvent/8 ml of
H2O), were performed in the pH range 2 to 11.5. The titrations were carried out under Ar at 25.0 Æ 0.18 [14 ± 16].
Because the pH-metric technique cannot reliably determine partition coefficients of markedly hydrophilic
compounds, log P values below 0.5 were confirmed with the classical shake-flask procedure [17].
‡
Ab initio Quantum-Mechanical Calculations. The geometries of compounds Me3N, Me3N O ,
‡
‡
Me3N H, Me3N OH (Table 3) were optimized by ab initio calculations at the 6-31G** level with the
Spartan 5.0 software [18] with the standard convergence criteria. All calculations were run on a Silicon Graphics
Origin 2000 workstation. The ESP partial atomic charges were also calculated [19].
B.T. and P.A.C. are indebted to the Swiss National Science Foundation for support.
REFERENCES
[1] B. Testa, ꢁThe Metabolism of Drugs and Other Xenobiotics ± Biochemistry of Redox Reactionsꢀ, Academic
Press, London, 1995, Chap. 5.
[2] B. Testa, P. Jenner, ꢁDrug Metabolism: Chemical and Biochemical Aspectsꢀ, Marcel Dekker, New York,
1976, Sect. 1.1.1.6.
[3] Medchem95 Database, Daylight Chemical Information System, Inc., Irvine, California, 1995.
[4] G. Caron, P. A. Carrupt, B. Testa, G. Ermondi, A. Gasco, Pharm. Res. 1996, 13, 1186.
[5] J. M. Mayer, B. Testa, H. van de Waterbeemd, A. Bornand-Crausaz, Eur. J. Med. Chem. 1982, 17, 461.
[6] R. Fruttero, G. Caron, E. Fornatto, D. Boschi, G. Ermondi, A. Gasco, P. A. Carrupt, B. Testa, Pharm. Res.
1998, 15, 1407.
[7] S. H. Pine, F. G. Catto, F. G. Yamagishi, J. Org. Chem. 1970, 35, 3663.
[8] C. Z. Ding, X. Lu, K. Nishimura, R. B. Silverman, J. Med. Chem. 1993, 36, 1711.
[9] A. H. Wragg, T. S. Stevens, D. M. Ostle, J. Chem. Soc. 1958, 4057.
[10] H. E. Zaugg, B. W. Horrom, J. Am. Chem. Soc. 1953, 75, 292.
[11] U. Azzena, G. Melloni, E. Fenude, C. Fina, M. Marchetti, B. Sechi, Synth. Commun. 1989, 24, 291.
[12] C. R. Hauser, M. T. Tetenbaum, J. Org. Chem. 1958, 23, 233.
[13] J. P. Ferris, R. D. Gerwe, G. R. Gapski, J. Org. Chem. 1968, 33, 3493.
[14] A. Avdeef, Quant. Struct.-Act. Relat. 1992, 11, 510.

Helvetica Chimica Acta ± Vol. 82 (1999)
[15] A. Avdeef, J. Pharm. Sci. 1993, 82, 183.
1639
[16] A. Avdeef, in ꢁLipophilicity in Drug Action and Toxicologyꢀ, Eds. V. Pliska, B. Testa and H. van de
Waterbeemd, VCH Publishers, Weinheim, 1996, pp. 109 ± 139.
[17] J. C. Dearden, G. M. Bresnen, Quant. Struct.-Act. Relat. 1988, 7, 133.
[18] Spartan 5.0. Wavefunction, Inc., Irvine, California, 1997.
[19] S. R. Cox, D. E. Williams, J. Comput. Chem. 1981, 2, 304.
Received May 5, 1999