21.05.2014 Views

Lipophilicity and membrane interactions of cationic-amphiphilic ...

Lipophilicity and membrane interactions of cationic-amphiphilic ...

Lipophilicity and membrane interactions of cationic-amphiphilic ...

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

European Journal <strong>of</strong> Pharmaceutical Sciences 14 (2001) 167–175<br />

www.elsevier.nl/locate/ejps<br />

<strong>Lipophilicity</strong> <strong>and</strong> <strong>membrane</strong> <strong>interactions</strong> <strong>of</strong> <strong>cationic</strong>-<strong>amphiphilic</strong><br />

compounds: syntheses <strong>and</strong> structure–property relationships<br />

Christian D.P. Klein<br />

*, Gerald F. Tabeteh , Alice V. Laguna , Ulrike Holzgrabe , Klaus Mohr<br />

a, b b c b<br />

a<br />

Pharmaceutical Chemistry, ETH Zurich, ¨ Winterthurerstr. 190, CH-8057 Zurich, ¨ Switzerl<strong>and</strong><br />

b<br />

Department <strong>of</strong> Pharmacology <strong>and</strong> Toxicology, Pharmaceutical Institute, University <strong>of</strong> Bonn, An der Immenburg 4, D-53121 Bonn, Germany<br />

c<br />

Institute <strong>of</strong> Pharmacy <strong>and</strong> Food Chemistry, University <strong>of</strong> Wurzburg, ¨ Am Hubl<strong>and</strong>, D-97074 Wurzburg, ¨ Germany<br />

Received 22 December 2000; received in revised form 18 June 2001; accepted 18 June 2001<br />

Abstract<br />

This study was performed to elucidate the relationship between steric factors, lipophilicity, <strong>and</strong> the potency <strong>of</strong> <strong>cationic</strong>-<strong>amphiphilic</strong><br />

compounds to displace calcium ions from phosphatidylserine monolayers. The latter property is considered to be a substance/<br />

phospholipid affinity measure. A series <strong>of</strong> <strong>cationic</strong>-<strong>amphiphilic</strong> 3-phenyl-N,N-dimethylpropylamine derivatives with systematic structural<br />

variations was synthesized. <strong>Lipophilicity</strong> values were determined by chromatographic (RP-HPLC, log D<br />

7.4), shake-flask (log P), <strong>and</strong><br />

theoretical (CLOGP) techniques. The potency <strong>of</strong> the compounds to displace calcium ions from phosphatidylserine monolayers was<br />

45 21<br />

determined using a radiotracer technique, employing the isotope Ca . The experimental lipophilicity values <strong>of</strong> several isomeric<br />

biphenyl- <strong>and</strong> diphenyl-congeners differ more than could be expected from the CLOGP-calculations <strong>and</strong> show a good correlation to the<br />

calculated molecular surface areas. Although the affinity <strong>of</strong> the substances to the phospholipid monolayer tends to increase with<br />

lipophilicity, no general interrelation between the two properties could be found. Surprisingly, the assay system (a phospholipid<br />

monolayer) was quite sensitive towards small steric changes at the ‘lig<strong>and</strong>’ molecules. Stereochemical factors have a considerable<br />

influence on the interaction <strong>of</strong> solutes with phospholipid <strong>membrane</strong>s. It must be questioned whether lipophilicity measures alone, without<br />

taking other molecular features into account, can meaningfully be used to explain or predict the influence <strong>of</strong> solutes on <strong>membrane</strong>-related<br />

processes <strong>and</strong> properties. © 2001 Elsevier Science B.V. All rights reserved.<br />

Keywords: Phosphatidylserine; Solute-<strong>membrane</strong> <strong>interactions</strong>; <strong>Lipophilicity</strong>; Log P; Cationic-<strong>amphiphilic</strong>; Phospholipid monolayer<br />

1. Introduction <strong>interactions</strong>. If <strong>cationic</strong>-<strong>amphiphilic</strong> substances (see, for<br />

example, the general formula shown in Table 1) are added<br />

The interaction <strong>of</strong> drugs with biological <strong>membrane</strong>s is <strong>of</strong> to this system, the lipophilic parts <strong>of</strong> the molecules are<br />

major importance for their pharmacokinetic <strong>and</strong> pharmaco- incorporated into the hydrocarbon chain region <strong>of</strong> the<br />

dynamic properties. The essential structural feature <strong>of</strong> phospholipid monolayer. The <strong>cationic</strong> parts <strong>of</strong> the combio<strong>membrane</strong>s<br />

is the phospholipid bilayer, <strong>and</strong> artificial pounds are located within the phospholipid headgroup<br />

phospholipid preparations are <strong>of</strong>ten used as model systems region or at the phospholipid–water interface <strong>and</strong> displace<br />

for the probing <strong>of</strong> drug/bio<strong>membrane</strong> <strong>interactions</strong>. We calcium ions from their ionic ‘binding sites’. This process<br />

have established an assay system in which the adsorption is depicted in Fig. 1. The potency <strong>of</strong> <strong>cationic</strong> <strong>amphiphilic</strong><br />

21<br />

<strong>of</strong> Ca -ions to a phosphatidylserine monolayer, formed at compounds to displace calcium ions depends on their<br />

an air-buffer interface, can be measured by means <strong>of</strong> a affinity to the phosphatidylserine monolayer (Hauser <strong>and</strong><br />

radiotracer method (e.g., Girke et al., 1989). The addition Dawson, 1968; Scheufler et al., 1990; Vogelgesang <strong>and</strong><br />

<strong>of</strong> <strong>cationic</strong>-<strong>amphiphilic</strong> substances to the buffer solution Scheufler, 1990). It is usually expressed as the concen-<br />

21<br />

leads to a decrease in adsorbed Ca . The displacement <strong>of</strong><br />

21<br />

tration which leads to a 50% decrease in adsorbed Ca ,<br />

21<br />

Ca is due to an ion-exchanging effect: The negatively i.e. the IC50<br />

value.<br />

charged phosphatidylserine headgroups bind cations from We would like to present a dataset in which a common,<br />

the aqueous buffer, for example calcium ions, via ionic <strong>cationic</strong>-<strong>amphiphilic</strong> structure (N,N-dimethyl-3-phenylpropylamine,<br />

cpd. 1 <strong>of</strong> Table 1) has been systematically<br />

*Corresponding author. Tel.: 141-1-635-6072; fax: 141-1-635-6884. extended. In particular, we intended to clarify the relation-<br />

E-mail address: cklein@pharma.ethz.ch (C.D.P. Klein).<br />

ships between lipophilicity, steric factors, <strong>and</strong> the affinity<br />

0928-0987/01/$ – see front matter © 2001 Elsevier Science B.V. All rights reserved.<br />

PII: S0928-0987(01)00170-1


168 C.D.P. Klein et al. / European Journal <strong>of</strong> Pharmaceutical Sciences 14 (2001) 167 –175<br />

Table 1<br />

Structures, log P <strong>and</strong> log D7.4 values, <strong>and</strong> 2log IC50<br />

values <strong>of</strong> the dataset <strong>of</strong> <strong>cationic</strong>-<strong>amphiphilic</strong> compounds. All compounds were assayed as<br />

hydrochlorides<br />

a<br />

Cpd. log D7.4 log P CLOGP 2 log IC50<br />

(RP-HPLC, pH 7.4) (shake-flask, pH 12.5) (IC in mol/l)<br />

c c c<br />

50<br />

b<br />

1 R1–55H 1.50 n.d. 2.49 3.00<br />

2 R 5CH 2.00 n.d. 2.99 3.25<br />

1 3<br />

3 R 5C H 2.45 n.d. 3.52 3.45<br />

1 2 5<br />

4 R 5n-C H 2.88 n.d. 4.05 4.31<br />

1 3 7<br />

5 R 5i-C H 2.74 n.d. 3.92 3.60<br />

1 3 7<br />

6 R 5(C H ) CH 3.67 n.d. 4.98 4.92<br />

1 2 5 2<br />

7 R 5Phenyl 3.15 6.08 4.38 5.12<br />

1<br />

8 R 5Phenyl 3.02 6.20 4.38 4.70<br />

2<br />

9 R 5Phenyl 2.71 5.53 4.08 4.23<br />

3<br />

10 R 5Phenyl 2.34 5.40 3.93 3.93<br />

4<br />

11 R 5Phenyl 2.26 5.22 3.93 3.42<br />

5<br />

12 R 5OCH 1.48 n.d. 2.41 2.76<br />

1 3<br />

13 R 54-OCH –phenyl, 2.33 n.d. 3.85 3.79<br />

1 3<br />

R 5Phenyl<br />

4<br />

14 R 54-OH–phenyl, 1.06 n.d. 3.26 3.87<br />

1<br />

R 5Phenyl<br />

4<br />

15 R154-OCH 3–phenyl, 2.15 n.d. 3.85 3.29<br />

R55Phenyl<br />

16 3.19 n.d. 4.07 3.95<br />

17 2.13 n.d. 3.10 3.98<br />

a<br />

unless otherwise stated, all R5H.<br />

b<br />

n.d.: not determined.<br />

c<br />

n53–5 experiments, st<strong>and</strong>ard errors for the given values are smaller than 0.15 log units.<br />

Fig. 1. Cationic-<strong>amphiphilic</strong> compounds displace calcium ions from their ‘binding sites’ at the aqueous side <strong>of</strong> the phosphatidylserine monolayer.<br />

Phosphatidylserine <strong>of</strong> natural origin, which was used in the experiments, carries variable fatty acid chains <strong>and</strong> not necessarily palmitoyl chains as shown<br />

here.


C.D.P. Klein et al. / European Journal <strong>of</strong> Pharmaceutical Sciences 14 (2001) 167 –175 169<br />

to phosphatidylserine monolayers. The dataset shown here compounds. The CLOGP values were calculated for the<br />

is partially identical with a dataset given previously (Klein non-ionized species <strong>of</strong> the compounds.<br />

et al., 1999; Klein <strong>and</strong> Hopfinger, 1998). We would like to<br />

point out that all compounds which were also part <strong>of</strong> the<br />

previous dataset have been re-assayed, which accounts for 2.3. Experimental determination <strong>of</strong> partition coefficients<br />

minor differences between the former values <strong>and</strong> the ones<br />

presented here. It was necessary to re-determine these<br />

values because the employed phospholipid (phosphatidyl- 2.3.1. Shake-flask octanol/water partitioning<br />

serine) is <strong>of</strong> biological origin (bovine brain) <strong>and</strong> therefore Shake-flask octanol–water partition coefficients (Fujita<br />

the IC50<br />

values have a certain, though small, batch-to- et al., 1964) were determined for several isomeric bibatch<br />

variability. All values given here were determined phenyl- <strong>and</strong> diphenyl-propylamines (7–11). In order to<br />

using the same batch <strong>of</strong> phosphatidylserine.<br />

eliminate the ionized, <strong>cationic</strong> species <strong>of</strong> the compounds,<br />

For the previous dataset that also included other pharma- the partitioning experiments were carried out using a<br />

cological endpoints (antiarrhythmic activity <strong>and</strong> depression strongly basic (pH 12.5) aqueous buffer. We have previ<strong>of</strong><br />

a liposome phase transition temperature), we have tried ously measured pKa<br />

values <strong>of</strong> 9.5 to 10 for several close<br />

to develop Quantitative Structure–Activity Relationships analogs (Klein et al., 1999). Therefore, it can be assumed<br />

using intramolecular <strong>and</strong> intermolecular descriptors. that the <strong>cationic</strong> species was present in a concentration <strong>of</strong><br />

Whereas we succeeded in developing good QSAR models less than 1%.<br />

for the antiarrhythmic activity, the biophysical properties 1-Octanol <strong>of</strong> the best commercially available quality<br />

were more difficult to explain in a quantitative manner. was used for the shake-flask octanol/water partitioning<br />

The major drawback <strong>of</strong> the previous dataset was its experiments. The aqueous buffer consisted <strong>of</strong> trisodium<br />

fragmentation–the<br />

45 21<br />

Ca -displacing potency, for exam- phosphate trihydrate (0.2 M, 76.02 g/l) <strong>and</strong> potassium<br />

ple, was measured for only a few compounds <strong>of</strong> a hydrogen phosphate (0.1 M, 17.42 g/l), dissolved in<br />

congeneric series. In contrast, the present dataset allows a distilled water. The calculated pH <strong>of</strong> this buffer was 12.6,<br />

systematic study <strong>of</strong> steric <strong>and</strong> lipophilic effects.<br />

<strong>and</strong> a pH <strong>of</strong> 12.5 was measured. Buffer <strong>and</strong> octanol were<br />

saturated with each other <strong>and</strong> filtered before use. A diode<br />

array spectrophotometer (HP 8452A) with a 1 cm-cell was<br />

2. Material <strong>and</strong> methods used for the absorption measurements <strong>of</strong> the aqueous phase<br />

after partitioning, <strong>and</strong> the spectral data was stored <strong>and</strong><br />

2.1. Displacement <strong>of</strong><br />

45 21<br />

Ca from phospholipid evaluated using a computer (HP89532A UV–Vis s<strong>of</strong>tmonolayers<br />

ware). All experiments were run in triplicate <strong>and</strong> at room<br />

temperature (258C).<br />

A detailed description <strong>of</strong> the experimental procedure Plastic vials proved to be unsuitable for the partitioning<br />

®<br />

was given by Girke et al. (1989). In short, a teflon experiments, probably because the high lipophilicity <strong>of</strong> the<br />

planchette <strong>of</strong> 4.5 cm diameter was filled with 5 ml <strong>of</strong> compounds caused their adsorption at the container sur-<br />

®<br />

buffer (0.01 mM CaCl2<br />

supplemented with trace amounts face. Instead, glass vials with teflon seals were used.<br />

45<br />

<strong>of</strong> CaCl<br />

2, 5 mM NaCl, 2 mM TES, 2 mM histidine; pH All test compounds were dissolved (as hydrochloride<br />

7.5). Radioactivity was detected by a Geiger-Muller ¨ count- salts) in distilled water at a concentration <strong>of</strong> 2 mg/ml to<br />

ing tube (Frieseke & Hoepfner, Erlangen, Germany) give stock solutions. The amounts <strong>of</strong> sample, buffer <strong>and</strong><br />

positioned above the teflon dish. 3 nmol <strong>of</strong> phospha- octanol were chosen so that absorptions <strong>of</strong> 0.35 to 0.6 were<br />

tidylserine (from bovine brain; purity 98–99%, Sigma reached after partitioning. In case <strong>of</strong> the very lipophilic<br />

Chemical) dissolved in 1 ml <strong>of</strong> chlor<strong>of</strong>orm were carefully compounds 7 <strong>and</strong> 8, an absorption <strong>of</strong> only 0.17 <strong>and</strong> 0.25,<br />

applied to the surface <strong>of</strong> the buffer to form a phospholipid respectively, was measured. A typical partitioning experi-<br />

45 21<br />

monolayer. Adsorption <strong>of</strong> Ca to the phospholipid was ment was performed with 700 ml <strong>of</strong> sample stock solution<br />

indicated by a tw<strong>of</strong>old elevation <strong>of</strong> the counting rate under (corresponding to 1.4 mg <strong>of</strong> sample), 700 ml <strong>of</strong> buffer <strong>and</strong><br />

control conditions. Test compounds were dissolved in the 20 ml <strong>of</strong> octanol. All substances were pipetted into the<br />

buffer at the appropriate concentrations. An experiment vials, the octanol was added by microliter syringe, the vials<br />

included measurement <strong>of</strong> a range <strong>of</strong> concentrations in were shaked in a rotating shaking machine at 120 rpm for<br />

order to obtain a concentration/effect-curve for the dis- 60 min, centrifuged at 5000 rpm for 5 min to afford<br />

placement <strong>of</strong><br />

45 21<br />

Ca by a given test compound. Experi- complete phase separation, <strong>and</strong> the octanol was removed.<br />

ments were repeated three to five times.<br />

The absorption <strong>of</strong> the aqueous phase was determined. In<br />

most cases, absorptions were measured at the maximum <strong>of</strong><br />

2.2. Calculation <strong>of</strong> partition coefficients the B-b<strong>and</strong>. The concentration in the octanol layer was<br />

determined by difference. Calibration was done in exactly<br />

The ‘CLOGP for Windows V. 4.0.0’ program (Biobyte the same manner as the partitioning, except that no octanol<br />

Corp.) was used to calculate CLOGP values for the target <strong>and</strong> smaller amounts <strong>of</strong> sample were used.


170 C.D.P. Klein et al. / European Journal <strong>of</strong> Pharmaceutical Sciences 14 (2001) 167 –175<br />

2.3.2. RP-HPLC recorded using KBr tablets (Perkin-Elmer spectrometer,<br />

The distribution coefficients (log D<br />

7.4) <strong>of</strong> all target model 298). Combustion analyses (C, H, N) for the target<br />

compounds were determined by RP-chromatography with compounds gave results within60.4% <strong>of</strong> the calculated<br />

methanol/aqueous buffer (60/40 <strong>and</strong> 30/70) mobile values. All starting compounds were used as received<br />

phases (cf. Kaliszan (1990) for a review <strong>of</strong> log P de- (Merck, Fluka, Sigma-Aldrich). An overview <strong>of</strong> the syntermination<br />

by HPLC). The chromatographic systems were thetic pathways is given in Fig. 2.<br />

calibrated with solutes for which an experimentally de-<br />

13<br />

C NMR spectral data will be provided upon request.<br />

termined octanol/water partition-coefficient was available 1–5, 7, 10, 12: Synthesized as described previously<br />

(Hansch et al., 1995) <strong>and</strong> which are not charged at pH 7.4. (Klein et al., 1999).<br />

The capacity factors <strong>of</strong> the reference substances were 6, N,N-Dimethyl-3-(4-[3-pentyl]-phenyl)-propylamine<br />

correlated against the experimental octanol/water log P hydrochloride: 4-(3-Pentyl)-acetophenone (6a) was synvalues,<br />

<strong>and</strong> the correlation equation, thus obtained, was thesized from propiophenone in a two-step synthesis<br />

used to calculate lipophilicity values (now log D, because according to Cram (1952). Purified by distillation in vacuo.<br />

D 1<br />

the test compounds are ionized at pH 7.4) for the test 6a: bp 150–1538C [22 hPa], n20 1.5171, H NMR (CDCl<br />

3,<br />

2 6<br />

compounds. d ): 7.88 (d, 2 H, J58.4 Hz, H , H aro), 7.20 (d, 2 H,<br />

3 5<br />

The RP-HPLC system for the determination <strong>of</strong> log D7.4 J58.3 Hz, H , H aro), 2.55 (s, 3 H, COCH<br />

3), 2.38 (m, 1<br />

21<br />

values consisted <strong>of</strong> a Kontron Instruments HPLC pump H), 1.45–1.75 (m, 4 H), 0.73 (m, 6 H); IR [cm ] (film):<br />

‘420’, a Rheodyne injection valve with 50 ml sample loop, 2960, 1680, 1605, 1270. 6a (0.01 mol, 1.9 g) was<br />

a GromSil ODS-2 5 mm ‘fully endcapped’ 12534 mm converted to the corresponding chloropropeniminium salt<br />

column, <strong>and</strong> a Perkin Elmer LC-480 Autoscan diode array (3-chloro-3-(4-[3-pentyl]-phenyl) - prop - 2 - ene - 1 - ylidenedetector.<br />

The chromatographic data were stored <strong>and</strong> evalu- N,N-dimethyliminium perchlorate, 6b) in 60% yield by the<br />

ated on a PC (LabControl LC-DESplus s<strong>of</strong>tware). A previously described method (Klein et al. (1999), Method<br />

1<br />

methanol/buffer (60/40 or 30/70) mobile phase was used B). 6b: yellow crystals, mp 1608C. H NMR (CD3CN, d ):<br />

at a flow rate <strong>of</strong> 2 ml/min. The buffer was a pH 7.4 8.73 (d, 1 H, J510 Hz, CH–CH5N), 7.96 (d, 2 H, J58.6<br />

phosphate buffer, obtained by mixing a 0.2 M potassium Hz, H aro), 7.39 (d, 2 H, J58.5 Hz, H aro), 7.31 (d, 1 H,<br />

dihydrogen phosphate solution (1000 ml) <strong>and</strong> 0.1 M J510 Hz, CH–CH5N), 3.69 (s, 3 H, E–CH5N(CH )),<br />

3 2<br />

sodium hydroxide (1573.6 ml). N,N-Dimethylhexylamin 3.57 (s, 3 H, Z–CH5N(CH ) ), 2.45–2.55 (m, 1 H,<br />

3 2<br />

(0.02%) was added to the mobile phase to shield unreacted CH(CH2CH 3) 2), 1.50–1.80 (m, 4 H, CH(CH2CH 3) 2),<br />

21<br />

silanol groups <strong>and</strong> minimize peak tailing (Gill et al., 0.75 (m, 6 H, CH(CH CH ) ); IR [cm ]: 2940, 1640,<br />

2 3 2<br />

1982). The column <strong>and</strong> the mobile phase were thermo- 1075. 6b was hydrogenated (H<br />

2: 300–400 kPa) in methastated<br />

at 308C (water bath, Haake Thermostat D1).<br />

nol solution over Pd/C (10%) catalyst, followed by the<br />

Aromatic compounds with known octanol/water parti- usual workup procedures, to give the desired 6 in quantita-<br />

1<br />

tion coefficients (Hansch et al., 1995) were used for the tive yield. 6: White crystals, mp 1388C. H NMR (CD3OD,<br />

calibration <strong>of</strong> the system (benzene, chlorobenzene, toluene, d ): 7.15 (d, 2 H, J57.1 Hz, H aro), 7.09 (d, 2 H, J57.1<br />

ethylbenzene, cumene, phenylethanol, biphenyl, anth- Hz, H aro), 3.14 (m, 2 H, CH<br />

2–N(CH 3) 2), 2.87 (s, 6 H,<br />

3<br />

racene, N,N-dimethylaniline). All analytes were dissolved N(CH ) ), 2.69 (m, 2 H, C H ), 2.25–2.30 (m, 1 H,<br />

3 2 2<br />

2<br />

2 3 2 2<br />

in methanol at a concentration <strong>of</strong> 300 mg/ml. Experiments CH(CH CH ) ), 2.00–2.10 (m, 2 H, C H ), 1.49–1.70<br />

were run in triplicate, <strong>and</strong> peak maxima were determined (m, 4 H, CH(CH2CH 3) 2), 0.75 (m, 6 H, CH(CH2CH 3)).<br />

2<br />

21<br />

at the respective absorption maximum. Capacity ratios IR [cm ]: 2950, 1460.<br />

were determined as k95(t 2t )/t , where t is the average 8, 3-(39-Biphenylyl)-N,N-dimethylpropylamine hydror<br />

0 0 r<br />

retention time <strong>of</strong> the analyte <strong>and</strong> t0<br />

is the ‘retention’ time chloride: 3-Phenylbenzaldehyde (8a) (Shoppee, 1933) was<br />

<strong>of</strong> a non-retained compound (thiourea). A linear regression synthesized in a 4-step synthesis from 3-nitroaniline. 3-<br />

analysis was performed for the log k9/log P data <strong>of</strong> the Nitroaniline was converted to 3-nitrobiphenyl following<br />

reference compounds, <strong>and</strong> the regression equations were the procedure described by Elks et al. (1940). This nitro<br />

used to calculate the log D7.4<br />

<strong>of</strong> the compounds. The compound was reduced to give 3-aminobiphenyl, which,<br />

equations were: (MeOH/H2O 60/40) log D7.452.04 log upon diazotation <strong>and</strong> reaction with potassium iodide,<br />

2<br />

k911.37, n59, R 50.99; (MeOH/H O 30/70) log D 5 afforded 3-iodobiphenyl (Campaigne <strong>and</strong> Reid, 1946).<br />

2<br />

2 7.4<br />

2.21 log k9 20.59, n54, R 50.98. Conversion to the Grignard reagent <strong>and</strong> reaction with<br />

N-formylpiperidine (following the procedure <strong>of</strong> Olah <strong>and</strong><br />

2.4. Syntheses <strong>and</strong> analytical data Arvanaghi, 1981) gave 8a in 60% yield. Purification by<br />

column chromatography (silica gel, petrol ether/ethyl<br />

1 D 1<br />

20 3<br />

NMR spectra were recorded at 299.956 MHz ( H) or acetate 90/10). 8a: n 1.6277; H NMR (CDCl , d ):<br />

13<br />

75.443 MHz ( C) on a Varian XL 300 FT-NMR spec- 10.10 (s, 1 H, Ar–CHO), 7.38–8.12 (m, 9 H, H aro); IR<br />

trometer, with the solvent proton/carbon signals being<br />

21<br />

[cm ] (film): 1690 b, 1590, 1180, 750. To a stirred<br />

used as internal st<strong>and</strong>ard for the calibration <strong>of</strong> the fre- suspension <strong>of</strong> N,N-dimethylaminoethyltriphenylphosquency<br />

scale. Unless otherwise stated, all IR spectra were phonium bromide (0.00321 mol, 1.34 g) (prepared accord-


C.D.P. Klein et al. / European Journal <strong>of</strong> Pharmaceutical Sciences 14 (2001) 167 –175 171<br />

Fig. 2. Overview <strong>of</strong> synthetic pathways.<br />

ing to Marxer <strong>and</strong> Leutert, 1978) in THF (15 ml), an give the desired 3-(39-biphenylyl)-N,N-dimethylpropylequimolar<br />

solution <strong>of</strong> 1.4 M n-butyllithium in n-hexane amine hydrochloride (8) in quantitative yield. 8: mp<br />

1<br />

was added under nitrogen at 08C. After 30 min a solution 1458C; H NMR (CDCl<br />

3, d ): 7.63–7.11 (m, 9 H, H aro),<br />

1<br />

<strong>of</strong> 8a (0.00321 mol, 0.584 g) in THF (7 ml) was slowly 3.02–2.96 (t, 2 H, J57.5 Hz, C H), 2.79 (t, 2 H, J57.7<br />

3<br />

added, <strong>and</strong> the reaction mixture was stirred overnight at Hz, C H), 2.77 (s, 6 H, N(CH<br />

3) 2), 2.21–2.27 (m (b), 2 H,<br />

2 21<br />

608C. The solution was acidified with 3 N hydrochloric C H); IR [cm ]: 2490, 1480, 760.<br />

acid (10 ml) <strong>and</strong> the tetrahydr<strong>of</strong>uran was removed in 9, 3-(29-Biphenylyl)-N,N-dimethylpropylamine hydrovacuo.<br />

After extraction with toluene (2310 ml) the water chloride: 2-Aminobiphenyl was converted to 2-<br />

phase was made alkaline with 2 N sodium hydroxide <strong>and</strong> iodobiphenyl following the procedure given by Gilman et<br />

extracted with diethyl ether (3320 ml). The N,N-di- al. (1929). From 2-iodobiphenyl, 2-phenylbenzaldehyde<br />

methyl-39-phenylcinnamylamine base was obtained by (Zaheer <strong>and</strong> Faseeh, 1944) (9a) was prepared in 80% yield<br />

stripping the dried (K2CO 3) extracts <strong>of</strong> solvent in vacuo, using the method <strong>of</strong> Olah <strong>and</strong> Arvanaghi (1981). Purified<br />

D<br />

<strong>and</strong> converted to the hydrochloride by the addition <strong>of</strong> by distillation in vacuo. 9a: bp 1378C (7 hPa), n20<br />

1.6210;<br />

1<br />

ethereal hydrochloric acid. The hydrochloride salt (8b, H NMR (CDCl<br />

3, d ): 10.00 (d, 1 H, J50.9 Hz, carbonyl–<br />

yield: 50%) was recrystallized from ethanol/diethyl ether. H), 8.07 (d, 1 H, J57.9 Hz, H aro), 7.0–7.7 (m, 9 H, H<br />

1 21<br />

8b: mp1648C. H NMR (CDCl<br />

3, d ): 7.61–7.26 (m, 9 H, aro); IR [cm ] (film): 3040, 2820, 1680. This compound<br />

3<br />

J515.9 Hz, H aro), 6.82 (d, 1 H, C H), 6.50 (dt, 1 H, was converted to N,N-dimethyl-29-phenylcinnamylamine<br />

2 1<br />

J515.9, 7.3 Hz, C H), 3.79 (d, 2 H, J57.3 Hz, C H), hydrochloride (9b) by the procedure given above for 8a<br />

21 1<br />

2.82 (s, 6 H, N(CH<br />

3) 2); IR [cm ]: 2680, 2080, 1480, 970, (yield: 30%). 9b: mp.1908C. H NMR (CDCl<br />

3, d ): 7.45–<br />

760. 8b was hydrogenated as described above for 6b, to 7.28 (m, 8 H, H aro), 7.14 (d, 1 H, J57.6 Hz, H aro), 6.93


172 C.D.P. Klein et al. / European Journal <strong>of</strong> Pharmaceutical Sciences 14 (2001) 167 –175<br />

3<br />

(d, 1 H, J511.5 Hz, C H), 6.00 (dt, 1 H, J511.5, 7.1 was obtained in 47% yield with respect to 13a. The<br />

2 1<br />

Hz,C H), 3.60 (d, 2 H, J57.0 Hz, C H), 2.61 (s, 6 H, compound is strongly hygroscopic; therefore, no meaning-<br />

21 1<br />

N(CH<br />

3) 2); IR [cm ]: 2640, 1465, 735. Catalytic hydro- ful melting point could be determined. 13: H NMR (D2O,<br />

2 6<br />

genation (see 6b) afforded to desired 3-(29-biphenylyl)- d ): 7.40 (m, 5 H, H aro), 7.33 (d, 2 H, J58.6 Hz, H , H<br />

N,N-dimethylpropylamine hydrochloride (9) in quantita-<br />

3 5<br />

aro), 6.95 (d, 2 H, J58.6 Hz, H , H aro), 4.00 (t, 1 H,<br />

1 3<br />

tive yield. 9: mp 1158C. H NMR (CDCl31 CD3OD, d ): C H), 3.76 (s, 3 H, Ar–OCH<br />

3), 3.00 (m, 2 H, CH<br />

2–<br />

1<br />

7.16–7.42 (m, 9 H, H aro), 2.70–2.80 (m, 2 H, C H), 2.59 N(CH<br />

3) 2), 2.88 (s, 6 H, N(CH<br />

3) 2), 2.38 (m, 2 H, CH–<br />

3 21<br />

(s, 6 H, N(CH<br />

3) 2), 2.56 (m, 2 H, C H), 1.84 (m, 2 H, CH<br />

2–CH 2–N); IR [cm ]: 1610, 1250, 1030.<br />

2 21<br />

C H); IR [cm ]: 2650 b, 1460, 750.<br />

14, N,N-Dimethyl-3-(4-hydroxyphenyl)-3-phenylpropyl-<br />

11, N,N-Dimethyl-2,3-diphenyl-propylamine hydrochloin<br />

amine hydrochloride: A solution <strong>of</strong> 13 (3.16 g, 0.01 mol)<br />

ride (Wawzonek <strong>and</strong> Nagler, 1955): 2,3-Diphenylreflux<br />

concentrated hydrobromic acid (40 ml) was heated to<br />

propionitrile was prepared as described by Abe <strong>and</strong><br />

for 2 h. After cooling, water (100 ml) was added<br />

Miyano (1972). 0.02 mol (4.15 g) <strong>of</strong> this compound was <strong>and</strong> the resulting mixture was alkalized with conc. amdissolved<br />

in ethanol (20 ml). A 33% solution <strong>of</strong> dithree<br />

monia to give pH 9. The resulting solution was extracted<br />

methylamine in ethanol (0.3 mol, 40 g) <strong>and</strong> 10% palladium<br />

times with 75 ml <strong>of</strong> diethyl ether. The combined<br />

on barium sulfate (2 g) were added. The reaction mixture organic layers were washed twice with water, dried over<br />

was then shaked under hydrogen (300–400 kPa) until the K2CO 3, <strong>and</strong> about 120 ml <strong>of</strong> the solvent were evaporated.<br />

theoretical uptake <strong>of</strong> hydrogen had occurred. The catalyst Addition <strong>of</strong> ethereal hydrochloric acid gave the crude<br />

was removed by filtration <strong>and</strong> the filtrate evaporated to phenol hydrochloride, which was recrystallized from ace-<br />

dryness. The residue was dissolved in water (50 ml) <strong>and</strong><br />

1<br />

tone/diethyl ether. Yield: 70%. 14: mp 2128C. H NMR<br />

brought to pH 11 by adding 2 N sodium hydroxide.<br />

Subsequent extraction with diethyl ether (3350 ml),<br />

(CD3OD, d ): 7.28 (m, 5 H, CH aro), 7.13 (d, 2 H, J58.5<br />

Hz, H<br />

2<br />

, H<br />

6<br />

aro), 6.73 (d, 2 H, J58.6 Hz, H<br />

3<br />

, H<br />

5<br />

aro), 3.93<br />

drying <strong>of</strong> the pooled extracts (K2CO 3), <strong>and</strong> evaporation <strong>of</strong> (t, 1 H, CH), 3.00 (m, 2 H, CH<br />

2–N(CH 3) 2), 2.82 (s, 6 H,<br />

21<br />

the solvent under vacuum yielded the crude base, which N(CH<br />

3) 2), 2.46 (m, 2 H, CH–CH<br />

2–CH 2); IR [cm ]:<br />

was separated from the primary amine by-product by 2960, 1515, 1260, 1220, 705.<br />

flash-chromatograpy (silica gel, mobile phase: ethyl aceylamine<br />

15, N,N-Dimethyl-2-phenyl-3-(4-methoxyphenyl)-prop-<br />

tate/petroleum ether/conc. ammonia 900/500/10). The<br />

hydrochloride (Kindler et al., 1948): 2-Phenyl-3-<br />

N,N-dimethyl-2,3-diphenyl-propylamine hydrochloride (4-methoxyphenyl)-propionitrile (0.02 mol, 4.75 g), syn-<br />

(11) was prepared in the usual manner by precipitation thesized according to the procedure <strong>of</strong> Abe <strong>and</strong> Miyano<br />

with ethereal hydrochloric acid <strong>and</strong> was recrystallized from (1972), was treated with H<br />

2<br />

/Pd–C/dimethylamine as<br />

1<br />

acetone/diethyl ether. Yield: 70%. 11: mp 1698C; H described above for cpd. 11. The same workup procedure<br />

NMR (CDCl<br />

3, d ): 6.98–7.29 (m, 10 H, H aro), 3.5–3.6 as for cpd. 11 was employed, giving a yield <strong>of</strong> 85%<br />

2 1<br />

(m, 1 H, C H), 3.25–3.45 (m, 2 H, C H), 3.07 (dd, 1 H, N,N-dimethyl-2-phenyl-3-(4-methoxyphenyl)-propylamine<br />

3<br />

1<br />

J56.7, 13.6 Hz, C H), 2.89 (dd, 1 H, J58.3, 13.6 Hz, hydrochloride (15). 15: mp 1428C. H NMR (CD3OD, d ):<br />

3<br />

3 5<br />

C H), 2.60 (d, 3 H, J54.9 Hz, N(CH<br />

3) 2), 2.49 (d, 3 H, 7.15–7.30 (m, 5 H, H aro), 6.91 (d, 2 H, J58.7 Hz, H , H<br />

21<br />

2 6<br />

J54.9 Hz, N(CH<br />

3) 2); IR [cm ]: 2450, 1440, 690. aro), 6.69 (d, 2 H, J58.7 Hz, H , H aro), 3.70 (s, 3 H,<br />

2 1<br />

13, N,N-Dimethyl-3-(4-methoxyphenyl)-3-phenylpropyl- Ar–OCH<br />

3), 3.50 (m, 1 H, C H), 3.35 (m, 2 H, C H), 3.00<br />

3 3<br />

amine hydrochloride: 4-Methoxyacetophenone was treated (dd, 1 H, C H), 2.85 (dd, 1 H, C H), 2.56 b (s, 6 H,<br />

21<br />

with phenylmagnesiumbromide, affording 1-(4-methox- N(CH<br />

3) 2); IR [cm ]: 3490, 1505, 1240.<br />

yphenyl)-1-phenylethanol (Novak <strong>and</strong> Protiva, 1959) 16, N,N-Dimethyl-3-(2-tetralyl)-propylamine hydrochlo-<br />

(13a). 0.1 Mol <strong>of</strong> the carbinol 13a were dissolved in ride: 2-Acetyltetralin (Smith <strong>and</strong> Lo, 1948) (16a) was<br />

anhydrous N,N-dimethylformamide (0.65 mol, 47.5 g, 50.0 synthesized from tetralin using the method described by<br />

ml). The mixture was cooled in an ice bath, <strong>and</strong> phosphormethoxynaphthalene.<br />

Arsenijevic et al. (1988) for the synthesis <strong>of</strong> 2-acetyl-6-<br />

us oxychloride (0.2 mol, 30 g, 17.9 ml) was added<br />

16a (0.0115 mol, 2 g) was converted<br />

dropwise, while the internal temperature was kept below to the corresponding chloropropeniminium salt (3-chloro-<br />

108C. Stirring was continued for 24 h at ambient temperaperchlorate,<br />

3-(2-tetralyl)-prop-2-ene-1-ylidene-N,N-dimethyliminium<br />

ture. Afterwards, the mixture was poured into an ice-cold<br />

16b) in 55% yield by the previously described<br />

methanol/water (3/2) mixture, <strong>and</strong> 70% perchloric acid method (Klein et al. (1999), Method B). 16b: mp 1658C.<br />

(25 ml) was added to precipitate the perchlorate salt. This 1 H NMR (CD3CN, d ): 8.69 (d, 1 H, J510.3 Hz, CH–<br />

caused the precipitation <strong>of</strong> a yellowish, semi-solid mass, CH5N), 7.70 (m, 2 H, H aro), 7.25 (m, 2 H, CH–CH5N/<br />

which was removed manually. The crude 3-(4-methoxy- H aro), 3.68 (s, 3 H, E–CH5N(CH<br />

3) 2), 3.56 (s, 3 H,<br />

phenyl) - 3 - phenylprop - 2 - ene - 1 - ylidene-N,N - dimethyl- Z–CH5N(CH<br />

3) 2), 2.83 (s (b), 4 H, Ar(CH<br />

2) 2(CH 2) 2),<br />

21<br />

iminium perchlorate (13b) could be hydrogenated without 1.80 (m, 4 H, Ar(CH<br />

2) 2(CH 2) 2); IR [cm ]: 2920, 1650,<br />

further purification (see description for 6b). The crude 1570, 1090. Catalytic hydrogenation <strong>of</strong> 16b (see cpd. 6a)<br />

product base was distilled in vacuo. N,N-Dimethyl-3-(4- afforded N,N-dimethyl-3-(2-tetralyl)-propylamine hydromethoxyphenyl)-3-phenylpropylamine<br />

hydrochloride (13)<br />

1<br />

chloride (16) in quantitative yield. 16: mp 1728C. H NMR


C.D.P. Klein et al. / European Journal <strong>of</strong> Pharmaceutical Sciences 14 (2001) 167 –175 173<br />

(D2O, d ): 7.01–7.07 (m, 3 H, H aro), 3.13–3.18 (m, 2 H, lipophilicity <strong>and</strong> devised a new method for the calculation<br />

CH<br />

2–N(CH 3) 2), 2.92 (s, 6 H, N(CH<br />

3) 2), 2.65–2.74 (m, 6 <strong>of</strong> log P values. We built analogs 7–11 with the di-<br />

H, Ar(CH<br />

2)(CH));CH 2 2 2 2–CH 2–CH 2–N), 1.99–2.09 (m, methylaminopropyl side chain in all-trans conformation,<br />

2 H, CH<br />

2–CH 2–CH 2–N), 1.76 (m, 4 H, performed a geometry optimization using the AM1<br />

21<br />

Ar(CH<br />

2 )<br />

2 (CH<br />

2 )<br />

2 ); IR [cm ]: 2910, 1250, 790.<br />

semiempirical hamiltonian (MOPAC 6.0 (Stewart, 1990)),<br />

17: Synthesized according to Marxer <strong>and</strong> Leutert determined their Connolly surface areas (probe radius 1.4<br />

(1978). Å) (Cerius2, v.3.0) <strong>and</strong> correlated it to their log D7.4<br />

(RP-HPLC) values. The results are shown in Fig. 3.<br />

Although the statistical significance <strong>of</strong> a dataset with n55<br />

is limited, we think that there is a clear dependency<br />

3. Results <strong>and</strong> discussion between surface area <strong>and</strong> log P or log D<br />

7.4. Bulky<br />

compounds, such as 10 <strong>and</strong> 11, expose less <strong>of</strong> their<br />

All experimental results are given in Table 1, along with aromatic, lipophilic surface areas than the elongated comthe<br />

calculated log P values. Overall, the CLOGP program pounds (7, 8). This is reflected by higher log P/log D 7.4<br />

performed quite well in predicting the relative lipo- values in the latter case. An attempt to correlate the surface<br />

philicities <strong>of</strong> the compounds in the dataset.<br />

area <strong>of</strong> the other compounds with their log D values failed.<br />

Initially, we did not intend to determine any log P values This could be due to the different number <strong>of</strong> aromatic rings<br />

by the shake-flask method, <strong>and</strong> wanted to rely solely on the in the compounds.<br />

RP-HPLC method. However, the chromatographic log D7.4<br />

We would now like to focus on the interaction <strong>of</strong> the<br />

values for the isomeric compounds 7–11 were more analogs with phosphatidylserine monolayers. Over all,<br />

different than we expected. Whereas the calculated log P there certainly is a dependency between the lipophilicity<br />

values (CLOGP) vary from 3.93 to 4.38, the chromato- <strong>and</strong> the calcium-displacing potency <strong>of</strong> the compounds (see<br />

graphic method gave log D7.4<br />

values from 2.26 to 3.15. It Fig. 4). However, a more detailed consideration <strong>of</strong> the data<br />

should be noted that molecular modeling programs which in Table 1 gives a picture that is less clear.<br />

calculate log P values by means <strong>of</strong> an atom-based ap- The calcium-displacing potency, expressed as 2log<br />

proach predicted an even smaller difference in lipo- IC<br />

50, <strong>of</strong> the isomers 7–11 increases with lipophilicity, <strong>and</strong><br />

2<br />

philicities than CLOGP. The majority <strong>of</strong> commercial one finds a good, linear correlation (r 0.94; log D7.4<br />

from<br />

modeling programs uses atom-based approaches for the RP-HPLC) for the two properties — for this data-subset.<br />

calculation <strong>of</strong> log P values. Of course, some steric in- If, however, other compounds are considered, the relationfluence<br />

could be expected, but we began to doubt the ship between lipophilicity <strong>and</strong> <strong>membrane</strong> interaction bevalidity<br />

<strong>of</strong> the HPLC method <strong>and</strong> decided to perform comes much more vague. For example, compounds 7 <strong>and</strong><br />

‘classical’ shake-flask experiments for cpds. 7–11. Be- 16 are practically iso-lipophilic, whereas their potency to<br />

21<br />

cause the shake-flask technique can, due to experimental displace Ca -ions from phosphatidylserine monolayers<br />

problems (formation <strong>of</strong> micelles, accumulation <strong>of</strong> the differs by more than an order <strong>of</strong> magnitude. The same<br />

compounds at the octanol–water interface), not be per- observation, although less pronounced, can be made for<br />

fomed with <strong>amphiphilic</strong> compounds (5surfactants), it was compounds 9 <strong>and</strong> 5. In both cases, the bi-aromatic analogs<br />

necessary to use a strongly basic, aqueous buffer, in which (7, 9) are distinctly more active than the iso-lipophilic<br />

the ionized, <strong>amphiphilic</strong> species is practically eliminated.<br />

Therefore, a phosphate/hydrogen phosphate buffer was<br />

used for the partitioning experiments. Not surprisingly,<br />

compounds 7–11 are three orders <strong>of</strong> magnitude more<br />

lipophilic (as determined by the shake-flask method) at pH<br />

12.5 than at pH 7.4 (RP-HPLC). Nevertheless, the correlation<br />

between shake-flask- <strong>and</strong> RP-HPLC-values is good,<br />

<strong>and</strong> both methods give values for cpds. 7–11 that differ<br />

(‘within’ each method) by about one order <strong>of</strong> magnitude.<br />

What is the reason for the large variation <strong>of</strong> the log P<br />

values for cpds. 7–11? It appeared obvious to determine<br />

the molecular surface areas <strong>of</strong> these isomers <strong>and</strong> see if<br />

there was any interrelation with the log P values. Correlations<br />

between log P <strong>and</strong> surface area are not a new<br />

concept. Yalkowski <strong>and</strong> Valvani (1976) used the solventaccessible<br />

surface area <strong>of</strong> hydrocarbons to calculate their<br />

lipophilicity. These studies were later extended by Camilleri<br />

<strong>and</strong> colleagues (1988), who correlated the Connolly Fig. 3. Correlation <strong>of</strong> Connolly surface area versus log P for compounds<br />

(1983) surface areas <strong>of</strong> benzene derivatives to their 7–11. A probe radius <strong>of</strong> 1.4 A˚<br />

was used.


174 C.D.P. Klein et al. / European Journal <strong>of</strong> Pharmaceutical Sciences 14 (2001) 167 –175<br />

45 21<br />

Fig. 4. Correlation plot <strong>of</strong> 2log IC<br />

50<br />

(displacement <strong>of</strong> Ca from<br />

phosphatidylserine monolayers) versus log D<br />

7.4<br />

<strong>of</strong> all compounds.<br />

for the remaining dataset becomes 0.75. This observation<br />

shows once more that small structural changes can have a<br />

large influence on the interaction <strong>of</strong> solutes with phospholipid<br />

<strong>membrane</strong>s. The mechanistic background for the<br />

strong activity <strong>of</strong> compound 14 can only be hypothesized.<br />

One possible explanation could be that the phenolic moiety<br />

<strong>of</strong> compound 14 represents an additional ‘binding partner’<br />

for the monolayer/water interface, thereby increasing the<br />

compound’s affinity.<br />

The main conclusion that can be drawn from this work<br />

is that the interaction <strong>of</strong> phospholipid assemblies with<br />

solute molecules can have a remarkably high steric selectivity.<br />

The effect observed here, i.e. displacement <strong>of</strong><br />

45 21<br />

Ca , reflects binding <strong>of</strong> the <strong>cationic</strong> model <strong>amphiphilic</strong><br />

compounds. The octanol/water partition coefficient (log P)<br />

<strong>of</strong> the solutes might be a useful parameter for the prediction<br />

or explanation <strong>of</strong> general trends <strong>of</strong> binding affinity.<br />

However, other factors appear to be involved <strong>and</strong> minor<br />

steric variations may have a pronounced impact on the<br />

adsorption, absorption, <strong>and</strong> penetration behavior <strong>of</strong> small<br />

molecules.<br />

alkyl-aromatic compounds (5, 16) This is probably due to<br />

the rigidity <strong>of</strong> the biphenyl moieties, which aids their<br />

‘intercalation’ between the fatty acid chains <strong>of</strong> the phospholipids.<br />

The four compounds 2, 11, 15, <strong>and</strong> 17 have similar<br />

lipophilicities (log D : 2.00–2.26). The calcium-displac-<br />

Acknowledgements<br />

7.4<br />

ing potency <strong>of</strong> cpds. 2, 11, <strong>and</strong> 15 is about the same (2log<br />

D. Heber <strong>and</strong> M. Klingmuller ¨<br />

IC around 3.3), whereas 17 is distinctly more active<br />

<strong>of</strong> the University <strong>of</strong> Kiel,<br />

50<br />

(2log IC : 3.98). Furthermore, the calcium-displacing<br />

Pharmaceutical Institute, kindly provided cpds. 2, 3, <strong>and</strong> 4.<br />

50<br />

potency does not increase with lipophilicity. 17, the most<br />

C.K. appreciates financial support by the Konrad-<br />

active compound in this sub-group, is less lipophilic than<br />

Adenauer-Foundation.<br />

11 <strong>and</strong> 15. Compound 2, carrying a methyl group in the<br />

4-position, is nearly iso-lipophilic with cpd. 17 that has a<br />

chlorine in the 4-position. Another structural difference References<br />

between 2 <strong>and</strong> 17 is that 17 has an unsaturated side chain.<br />

Here, one might guess that the increased activity is due to Abe, N., Miyano, S., 1972. C-alkylation <strong>of</strong> active methylene compounds<br />

the fact that the side chain <strong>of</strong> 17 is more rigid, therefore by means <strong>of</strong> alcohols. VII. Synthesis <strong>of</strong> a-substituted phenylacetonitriles<br />

from a-phenylacetoacetonitrile. J. Org. Chem. 37, 526–528.<br />

facilitating the insertion <strong>of</strong> the compound into the phospholipid<br />

monolayer.<br />

Arsenijevic, L., Arsenijevic, V., Horeau, A., Jacques, J., 1988. Acetyl-6-<br />

methoxynaphthalene. In: Nol<strong>and</strong>, W.E. (Ed.). Org. Synth. Coll, Vol. VI.<br />

A very interesting observation was made with com- Wiley & Sons, New York, pp. 34–36.<br />

pounds 4 <strong>and</strong> 5, 4 carrying an n-propyl <strong>and</strong> 5 carrying an Camilleri, P., Watts, S.A., Boraston, S.A., 1988. A surface area approach<br />

iso-propyl group at the 4-position. These two geometric to determination <strong>of</strong> partition coefficients. J. Chem. Soc. Perkin Trans.<br />

isomers, whose log D values are 2.88 <strong>and</strong> 2.74, have 2, 1699–1707.<br />

7.4<br />

2log IC values <strong>of</strong> 4.31 <strong>and</strong> 3.60, respectively. This Campaigne, E., Reid, B.M., 1946. Some substituted acetophenones. J.<br />

50<br />

21<br />

Am. Chem. Soc 68, 1663.<br />

means that their potency to displace Ca -ions from Cerius2, V.3.0 User Guide. San Diego: Molecular Simulations Inc., 1997.<br />

phosphatidylserine monolayers differs by a factor <strong>of</strong> about CLOGP for Windows, release 4.0. BioByte Corp., Claremont, USA,<br />

5. We think that it is quite interesting to find such a degree 1999.<br />

<strong>of</strong> sensitivity to small steric changes in an assay system (a Connolly, M.L., 1983. Solvent-accessible surfaces <strong>of</strong> proteins <strong>and</strong> nucleic<br />

phospholipid monolayer) which is generally expected to be acids. Science 221, 709–713.<br />

Cram, D.J., 1952. Studies in stereochemistry. VIII. Molecular rearrangenot<br />

very dem<strong>and</strong>ing with respect to the geometries <strong>of</strong> its ments during lithium aluminum hydride reductions in the 3-phenyl-2-<br />

interaction partners. pentanol <strong>and</strong> 2-phenyl-3-pentanol systems. J. Am. Chem. Soc. 74,<br />

The only phenolic compound in this dataset (14) shows 2152–2159.<br />

the largest deviation from the log D /IC correlation Elks, J., Haworth, J.W., Hey, D.H., 1940. Union <strong>of</strong> aryl nuclei. Part V. A<br />

7.4 50<br />

curve. Its calcium-displacing activity is much higher than modification <strong>of</strong> the Gomberg reaction. J. Chem. Soc., 1284–1286.<br />

Fujita, T., Iwasa, J., Hansch, C., 1964. A new substituent constant p,<br />

one would expect from its relatively low log D<br />

7.4. Com- derived from partition coefficients. J. Am. Chem. Soc. 86, 5175–5180.<br />

pound 14 is certainly the main outlier in this dataset. If this Gill, R., Alex<strong>and</strong>er, S.P., M<strong>of</strong>fat, A.C., 1982. Comparison <strong>of</strong> amine<br />

2<br />

compound is not included in the correlation analysis, the r modifiers used to reduce peak tailing <strong>of</strong> 2-phenylethylamine drugs in


C.D.P. Klein et al. / European Journal <strong>of</strong> Pharmaceutical Sciences 14 (2001) 167 –175 175<br />

reversed-phase high-performance liquid chromatography. J. Chroma- Novak, L., Protiva, M., 1959. Antihistamin-substanzen XLVII. In p- und<br />

togr. 247, 39–45.<br />

m-stellung substituierte mephenhydramin-derivate. Coll. Czech. Chem.<br />

Gilman, H., Kirby, J.E., Kinney, C.R., 1929. An unusual type <strong>of</strong> 1,4- Commun. 24, 3966–3977.<br />

addition. J. Am. Chem. Soc. 51, 2260–2261.<br />

Olah, G.A., Arvanaghi, M., 1981. Aldehyde durch formylierung von<br />

Girke, S., Mohr, K., Schrape, S., 1989. Comparison between the activities Grignard- und organolithium-reagentien mit N-formylpiperidin.<br />

<strong>of</strong> <strong>cationic</strong> <strong>amphiphilic</strong> drugs to affect phospholipid-<strong>membrane</strong>s <strong>and</strong> to Angew. Chem. 93, 925–926.<br />

depress cardiac function. Biochem. Pharmacol. 38, 2487–2496.<br />

Scheufler, E., Vogelgesang, R., Wilffert, B., Pegram, B.L., Hunter, J.B.,<br />

Hansch, C., Leo, A., Hoekman, D., 1995. Exploring QSAR: Hydrophobic, Wermelskirchen, D., Peters, T., 1990. Uptake <strong>of</strong> cat<strong>amphiphilic</strong> drugs<br />

electronic, <strong>and</strong> steric constants. ACS Pr<strong>of</strong>essional Reference Book, into erythrocytes <strong>and</strong> muscular tissue correlates to <strong>membrane</strong> enrich-<br />

ACS, Washington, D.C.<br />

45<br />

ment <strong>and</strong> to Ca displacement from phosphatidylserine monolayers. J.<br />

Hauser, H., Dawson, R.M.C., 1968. The displacement <strong>of</strong> calcium ions Pharmacol. Exp. Ther 252, 333–338.<br />

from phospholipid monolayers by pharmacologically active <strong>and</strong> other Shoppee, C.W., 1933. Symmetrical triad prototropic systems. Part IX. J.<br />

organic bases. Biochem. J. 109, 909–916. Chem. Soc., 37–45.<br />

Kaliszan, R., 1990. HPLC methods <strong>and</strong> procedures <strong>of</strong> hydrophobicity Smith, L.I., Lo, C.P., 1948. The Jacobsen reaction: 6,7-dialkyltetralins. J.<br />

determination. Quant. Struct.-Act. Relat. 9, 83–87. Am. Chem. Soc. 70, 2209–2212.<br />

Kindler, K., Hedemann, B., Scharfe, ¨ E., 1948. Studien uber ¨ den mecha- Stewart, J.J.P., 1990. Mopac 6.0 release notes. Frank J. Seiler Research<br />

nismus chemischer reaktionen. X. Phenyl- und cyclohexyl-alkylamine Laboratory, United States Air Force Academy.<br />

durch hydrierung. Justus Liebigs Ann. Chem. 560, 215.<br />

Vogelgesang, R., Scheufler, E., 1990. R 56865 <strong>and</strong> flunarizine displace<br />

Klein, C., Hopfinger, A.J., 1998. Pharmacological activity <strong>and</strong> <strong>membrane</strong><br />

21<br />

Ca from phosphatidyserine monolayers in a stoichiometric manner.<br />

<strong>interactions</strong> <strong>of</strong> antiarrhythmics: 4D-QSAR/QSPR analysis. Pharm. Eur. J. Pharmacol. 188, 17–22.<br />

Res. 15, 303–311.<br />

Wawzonek, S., Nagler, R.C., 1955. The addition <strong>of</strong> dialkylamino- <strong>and</strong><br />

Klein, C.D.P., Klingmuller, ¨ M., Schellinski, M., L<strong>and</strong>mann, S., Hauschild, alkylaminomagnesium halides to a-phenylcinnamonitrile. J. Am.<br />

S., Heber, D., Mohr, K., Hopfinger, A.J., 1999. Synthesis, pharmaco- Chem. Soc. 77, 1796–1798.<br />

logical <strong>and</strong> biophysical characterization, <strong>and</strong> <strong>membrane</strong>-interaction Yalkowski, S.H., Valvani, S.C., 1976. Partition coefficients <strong>and</strong> surface<br />

QSAR analysis <strong>of</strong> <strong>cationic</strong>-<strong>amphiphilic</strong> model compounds. J. Med. areas <strong>of</strong> some alkylbenzenes. J. Med. Chem. 19, 727–728.<br />

Chem. 42, 3874–3888, published erratum appeared in J. Med. Chem. Zaheer, S.H., Faseeh, S.A., 1944. O-substituted biphenyls. II. J. Indian<br />

42, 5288 (1999). Chem. Soc. 21, 381–382.<br />

Marxer, A., Leutert, T., 1978. Die aminoalkylierung von carbonylverbindungen<br />

mit hilfe der Wittig-reaktion. Helv. Chim. Acta 61, 1709–<br />

1720.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!