A versatile framework for the design of ligand-dependent, transgene-specific transcription factors.

The ability to regulate transgene expression will be essential for the safety and efficacy of many gene therapies. Various ligand-dependent transcription factors, including steroid hormone receptors, have been modified to enable transgene-specific regulation. To minimize effects on cellular gene expression, chimeric steroid receptors have been constructed by replacing their native DNA binding domain (DBD) with a heterologous DBD, like that from the yeast transcription factor GAL4. This approach has limitations for human gene therapy, including the potential immunogenicity of the GAL4 domain and the inability to discriminate between different GAL4-linked transgenes in the same cell. To address this, we have constructed chimeric regulators containing the human estrogen receptor (ER) ligand binding domain (LBD) and a Cys(2)-His(2)-type zinc finger DBD. Cys(2)-His(2) zinc finger domains are common among human DNA binding proteins and can be engineered to selectively bind different DNA sequences. We demonstrate over 500-fold drug-dependent transgene induction with these chimeric regulators in vitro and the ability to regulate an adenovirus-delivered transgene in mice. Two chimeras containing different Cys(2)-His(2) domains displayed highly sequence-specific binding and regulation. Incorporating a point mutation in the ER LBD that ablates estrogen binding enables selective in vivo regulation with the clinically useful anti-estrogen tamoxifen. These Cys(2)-His(2)-ER LBD chimeras represent a versatile framework for creating transgene-specific regulators potentially useful for human gene therapy applications.

Biochemical and functional characterization of 1-benzyl substituted trimetoquinol affinity analogs on rat and human beta-adrenoceptors.

The site of interaction for the 1-(3',4',5'-trimethoxybenzyl) group of trimetoquinol (TMQ) with beta-adrenoceptors (beta-ARs) is important for the rational design of highly potent and beta3-AR-selective analogs. 1-Benzyl ring-substituted TMQ analogs were evaluated for binding affinities and biochemical activities (cyclic AMP accumulations) in Chinese hamster ovary (CHO) cells expressing the rat and human beta3-AR, and for functional activities on isolated rat tissues. Binding affinities (K1 approximately 0.055 to 1.5 microM) for the rat beta3-AR and potencies for adenylyl cyclase activation (K(act) approximately 0.43 to 2;5 nM) of the 3'-monoiodo or 3',5'-diiodo derivatives with 4'-isothiocyanato-, 4'-amino, 4'-acetamido, or 4'-alpha-haloacetamido substitutions were higher than those of (-)-isoproterenol, and comparable to those of BRL 37344 [(+/-)-(R*,R*-[4-[2-[[2-(3-chlorophenyl)-2-hydroxy-ethyl]amino]propyl]ph enoxy]-acetic acid sodium]. A similar rank order of binding affinities (K(i) approximately 0.11 to 2.5 microM) and potencies (K(act) approximately 0.45 to 9.5 nM) was obtained for TMQ analogs on the human beta3-AR. The 4'-acetamido and 4'-alpha-chloroacetamido analogs of 3',5'-diiodoTMQ were more potent than (-)-isoproterenol in rat atria (beta1-AR) and rat trachea (beta2-AR) and exhibited partial agonist activities, whereas full agonist activities were observed in rat esophageal smooth muscle (EC50 approximately 2-8 nM, beta3-AR). 4'-alpha-Chloroacetamido-3',5'-diiodoTMQ-mediated chronotropic responses in atria were sustained and resistant to washout. Further, the 4'-alpha-chloroacetamido and 4'-alpha-bromoacetamido analogs of 3',5'-diiodoTMQ demonstrated significant concentration-dependent irreversible binding to the rat beta3-AR. Reversible beta-AR agonists such as (-)-isoproterenol, BRL 37344, and 4'-acetamido-3',5'-diiodoTMQ or nucleophilic 1-amino acids (lysine, glutathione, cysteine) did not protect against this irreversible binding. Thus, the lipophilic 1-benzyl ring of TMQ analogs interacts with a hydrophobic region of the beta-AR that may represent an exo-site or an allosteric binding site.

Beta-adrenoceptor subtype activities of trimetoquinol derivatives: biochemical studies on human beta-adrenoceptors expressed in chinese hamster ovary cells.

The beta-adrenoceptor activities of trimetoquinol (TMQ) isomers and selected derivatives were evaluated on human beta-adrenoceptor subtypes expressed in Chinese hamster ovary cells. In cAMP accumulation assays, (-)-TMQ was 214-, 281-, and 776-fold more potent than (+)-TMQ at stimulating beta(1)-, beta(2)-, and beta(3)-adrenoceptor subtypes, respectively. In radioligand binding assays, (-)-TMQ exhibited 123-, 331-, and 5-fold greater affinity than (+)-TMQ for beta(1)-, beta(2)-, and beta(3)-adrenoceptor subtypes, respectively. (-)-TMQ and (+/-)-TMQ activated the human beta(3)-adrenoceptor with an 8.2- and 3.4-fold greater efficacy, respectively, than the reference beta-adrenoceptor agonist (-)-isoproterenol (efficacy = 1). The 3',5'-diiodo analogs of TMQ were partial agonists of the beta(2)-adrenoceptor relative to (-)-isoproterenol, and their potencies were 5- to 10-fold higher at the beta(3)-adrenoceptor as compared with beta(1)-adrenoceptors. Modification of the catechol (6,7-dihydroxy) nucleus, such as replacement of the 7-hydroxy group with a chloro group (7-chloroTMQ), ring fluorination (8-fluoro and 5,8-difluoro analogs), or preparation of bioisosteric tetrahydrothiazolopyridine (THP) derivatives of TMQ yielded compounds that displayed partial agonist activity (relative to (-)-isoproterenol) or were inactive at the beta(2)-adrenoceptor and exhibited beta(3)-adrenoceptor-selective stimulation compared with the beta(1)-adrenoceptor. Furthermore, the 3',5'-diiodo-4'-methoxybenzylTHP derivative of TMQ was 65-fold more potent than the corresponding 3',4',5'-trimethoxybenzylTHP at the human beta(3)-adrenoceptor. Our results indicate that 6, 7-dihydroxy-catechol-modified and 1-benzyl halogen-substituted derivatives of TMQ represent promising leads for the development of beta(3)-adrenoceptor-selective agonists.

Depolarization of membrane potential and the smooth muscle contraction by isothiocyanatobenzyl imidazoline in guinea-pig stomach circular muscle.

Isothiocyanatobenzyl imidazoline (IBI) produces characteristic slowly developing contraction of many smooth muscle preparations including the circular smooth muscle of the guinea-pig stomach. Changes in the membrane potential were recorded intracellularly, and the muscle contraction induced by IBI was investigated. IBI at 100 micromol/l slowly produced a sustained depolarization of the membrane with a maximum change of approximately 15 mV. This depolarization could not be blocked by 1-hyoscyamine, 100 nmol/l. An imidazoline analogue, oxymetazoline at 1 micromol/l, did not change the resting membrane potential as observed after IBI. Significant membrane depolarization after IBI still occurred in Ca2+-free medium. During IBI-induced depolarization, sudden reduction of Na+ to 30 mmol/l in the medium reduced the depolarization slightly. IBI-induced depolarization was additive with that produced by 20 mmol/l K+ in the medium. In the presence of tetraethylammonium chloride or levcromakalim or nifedipine, IBI continued to depolarize the membrane although functional pharmacological experiments showed that the contractile effects of IBI were significantly inhibited by 30 micromol/l levcromakalim and abolished by 100 nmol/l nifedipine. At 100 micromol/l phentolamine (reported by others as an inhibitor of ATP-sensitive potassium channels) completely blocked IBI-induced contraction. Phentolamine (30 micromol/l) blocked the contractile effects of IBI by 50%. On the other hand, S(-)-Bay K 8644, a voltage-dependent calcium channel activator, was additive with the contractile response of IBI. These results indicated that IBI produced membrane depolarization and contraction of the guinea-pig stomach circular muscle, by a mechanism not involving muscarinic receptors or alpha-adrenoceptors. Even though levcromakalim, an ATP-sensitive potassium channel opener, could not inhibit IBI-induced depolarization, the ATP-sensitive potassium channel and the voltage-dependent calcium channel may be intrinsically linked with the action of IBI.

Are presenilins intramembrane-cleaving proteases? Implications for the molecular mechanism of Alzheimer's disease.

The amyloid-beta protein (Abeta) is strongly implicated in the pathogenesis of Alzheimer's disease. The final step in the production of Abeta from the amyloid precursor protein (APP) is proteolysis by the unidentified gamma-secretases. This cleavage event is unusual in that it apparently occurs within the transmembrane region of the substrate. Studies with substrate-based inhibitors together with molecular modeling and mutagenesis of the gamma-secretase cleavage site of APP suggest that gamma-secretases are aspartyl proteases that catalyze a novel intramembranous proteolysis. This proteolysis requires the presenilins, proteins with eight transmembrane domains that are mutated in most cases of autosomal dominant familial Alzheimer's disease. Two conserved transmembrane aspartates in presenilins are essential for gamma-secretase activity, suggesting that presenilins themselves are gamma-secretases. Moreover, presenilins also mediate the apparently intramembranous cleavage of the Notch receptor, an event critical for Notch signaling and embryonic development. Thus, if presenilins are gamma-secretases, then they are also likely the proteases that cleave Notch within its transmembrane domain. Another protease, S2P, involved in the processing of the sterol regulatory element binding protein, is also a multipass integral membrane protein which cleaves within or very close to the transmembrane region of its substrate. Thus, presenilins and S2P appear to be members of a new type of polytopic protease with an intramembranous active site.

Activation of L-type calcium channel by tolazoline derivatives: role of isothiocyanate moiety.

Studies have investigated the pharmacologic mechanism of 2-(4'-isothiocyanatobenzyl) imidazoline (IBI) and analogs for interaction with imidazoline receptors (IRs), alpha-adrenergic receptors (alpha-ARs), and calcium channels in cardiovascular muscle systems. IBI differs from tolazoline by substitution of an electrophilic isothiocyanato (NCS) group. Unlike tolazoline, which is a partial alpha-AR agonist, IBI produced an irreversible, slow-onset, and sustained contraction of rat aorta with an median effective concentration (EC50) value of 5 microM, and a maximal contraction (116%) greater than that of phenylephrine (100%) and tolazoline (59%). The IBI-induced contractions were dependent on calcium channels and independent of alpha-ARs or IRs. Similarly, structure-activity relation studies in rat aortic smooth muscles on a series of synthesized IBI analogs indicated that NCS analogs, but not those without the NCS group, exhibited effects by a non-alpha-AR, non-IR, but a calcium channel-dependent mechanism. Thus the presence of an intact IBI ring in these analogs is not a requirement for these activities. Further, IBI inhibited dihydropyridine (DHP, [3H]PN 200-110 and [3H]Bay K 8644) binding to L-type calcium channels of T-tubule membranes in rabbit skeletal muscle. In contrast to nifedipine, IBI and NCS derivatives (nifedipine-NCS, naphazoline-NCS) only partially (50-88%) displaced specific binding of these radioligands. A single site of noncooperative interaction was observed for nifedipine (nH = 0.97), whereas tolazoline-NCS (IBI, nH = 1.46) and nifedipine-NCS (nH = 1.37) exhibited a positive cooperativity in binding to DHP sites. These receptor-binding data indicate that NCS derivatives bind to L-type calcium channels and interact allosterically with DHP-binding sites. Direct binding of the NCS group to specific nucleophilic protein sites of the calcium channel may be responsible for its activation and the subsequent contractile effects of IBI.

Medetomidine analogs as alpha 2-adrenergic ligands. 3. Synthesis and biological evaluation of a new series of medetomidine analogs and their potential binding interactions with alpha 2-adrenoceptors involving a "methyl pocket".

The synthesis and the biological evaluation of a new series of medetomidine analogs are reported. The substitution pattern at the phenyl ring of the tetralin analogs had a distinct influence on the alpha 2-adrenoceptor binding affinity. 4-Methylindan analog 6 was the most potent alpha 2-adrenoceptor binding ligand among these 4-substituted imidazoles, and its alpha 2-adrenoceptor selectivity was greater than the 5-methyl tetralin analog 4c. Ligand-pharmacophore and receptor modeling were combined to rationalize alpha 2-adrenoceptor binding data of the imidazole analogs in terms of ligand-receptor interactions. The structure-activity relationships that were apparent from this and previous studies were qualitatively rationalized by the binding site models of the alpha 2-adrenoceptor. The benzylic methyl group of medetomidine or the naphthyl analog 2a was superimposable with the alpha-methyl group of (-)-alpha-methylnorepinephrine and fit into the proposed "methyl pocket" of the alpha 2-adrenoceptor defined by the residues Leu110, Leu169, Phe391, and Thr395.

Iodinated analogs of trimetoquinol as highly potent and selective beta 2-adrenoceptor ligands.

A series of trimetoquinol (1, TMQ) analogs were designed and synthesized based on the lead compound 2, a diiodinated analog of trimetoquinol which exhibits improved selectivity for beta 2-versus beta 1-adrenoceptors (AR). To determine the influence of 1-benzyl substituents of trimetoquinol on beta 2-AR binding affinity and selectivity, we replaced and/or removed the 3'-, 4'-, and 5'-methoxy substituents of trimetoquinol. Replacement of the 4'-methoxy group of 2 with an amino (21c) or acetamido (15) moiety did not significantly alter beta 2-AR and thromboxane A2/prostaglandin H2 (TP) receptor affinity. Substitution with a 4'-hydroxy (18) or -iodo (21b) group did not significantly alter beta 2-AR affinity, but greatly reduced TP receptor affinity (380- and 1200-fold, respectively). Further, the beta 2-AR can accommodate larger substituents such as a benzamide at the 4'-position (26b). Other monoiodo derivatives (24, 26a) have similar or slightly lower affinity to both beta 2-AR and TP receptor compared to their diiodo analogs. Interestingly, removal of the 4'-substituent of 3',5'-diiodo analogs increased beta 2-AR affinity with little or no effect on beta 1-AR and TP binding. Thus, analog 21a displayed highly potent (pKi 9.52) and selective (beta 2/beta 1 = 600) binding affinity for beta 2-AR. On the other hand, trifluoromethyl substituents at the 3'- and 5'-positions (27) essentially abolished binding affinity at beta 2-AR and TP receptors. The differential binding effects of the aforementioned trimetoquinol modifications on the receptor systems may reflect differences in the binding pocket that interacts with the benzyl portion of trimetoquinol analogs. Thus, manipulation of the 1-benzyl moiety of trimetoquinol (1) has resulted in analogs that exhibit potent beta 2-AR binding affinity and significantly lower beta 1-AR and TP receptor affinities.

Influence of amine substituents on 5-HT2A versus 5-HT2C binding of phenylalkyl- and indolylalkylamines.

The effect of 15 different amine substituents on 5-HT2A and 5-HT2C serotonin receptor binding was investigated for two series of compounds (i.e., phenylalkylamine and indolylalkylamine derivatives). In general, amine substitution decreases receptor affinity; however, N-(4-bromobenzyl) substitution results in compounds that bind at 5-HT2A receptors with high affinity (Ki < 1 nM) and with > 100-fold selectivity. Although parallel structural modification in the two series result in parallel shifts in 5-HT2C binding, these same modifications alter 5-HT2A binding in a less consistent manner.

Antagonism of 1-(2,5-dimethoxy-4-methylphenyl)-2-aminopropane stimulus with a newly identified 5-HT2- versus 5-HT1C-selective antagonist.

DOM [i.e., 1-(2,5-dimethoxy-4-methylphenyl)-2-aminopropane] is a 5-HT1C/2 serotonin agonist that exerts stimulus control of behavior in animals. In order to determine if the discriminative stimulus effect of DOM is 5-HT1C- or 5-HT2-mediated, it would be informative to conduct tests of stimulus antagonism with a 5-HT1C- or 5-HT2-selective antagonist. To date, no such agents exist. Although the neuroleptic agent spiperone binds at D2 dopamine receptors and 5-HT1A serotonin receptors, (a) it displays about a 1000-fold selectivity for 5-HT2 versus 5-HT1C sites and (b) it has been used as a "5-HT2-selective" antagonist. Because spiperone is a behaviorally disruptive agent, it is not suitable for use in drug-discrimination studies. Using the spiperone molecule as a starting point, a limited structure-affinity investigation was conducted in order to identify a suitable antagonist with high affinity and selectivity for 5-HT2 receptors, and yet an antagonist that might lack the disruptive actions of spiperone. Various modifications of the spiperone molecule were examined, but most resulted in decreased 5-HT2 affinity or in loss of selectivity. One compound, 8-[3-(4-fluorophenoxy)propyl]-1-phenyl-1,3,8-triazaspiro[4.5]de can-4-on e (26), was shown to bind at 5-HT2 sites with high affinity (Ki = 2 nM) and > 2,000-fold selectivity versus 5-HT1C sites. In tests of stimulus antagonism using rats trained to discriminate 1 mg/kg of DOM from saline vehicle, 26 behaved as a potent antagonist (ED50 = 0.003 mg/kg) and lacked the disruptive effects associated with spiperone. As such, (a) it would appear that the DOM stimulus is primarily a 5-HT2-mediated, and not 5-HT1C-mediated, phenomenon, and (b) compound 26 may find application in other pharmacologic investigations where spiperone may not be a suitable antagonist.