A series of pyrido- and pyrimidomorphinans (6a-h and 7a-g) were synthesized from naltrexone and evaluated for binding and biological activity at the opioid receptors. The unsubstituted pyridine 6a displayed high affinities at opioid delta, mu, and kappa receptors with K(i) values of 0.78, 1.5, and 8.8 nM, respectively. Compound 6a was devoid of agonist activity in the mouse vas deferens (MVD) and guinea pig ileum (GPI) preparations but was found to display moderate to weak antagonist activity in the MVD and GPI with K(e) values of 37 and 164 nM, respectively. The pyrimidomorphinans in general displayed lower binding potencies and delta receptor binding selectivities than their pyridine counterparts. Incorporation of aryl groups as putative delta address mimics on the pyrido- and pyrimidomorphinan framework gave ligands with significant differences in binding affinity and intrinsic activity. Attachment of a phenyl group at the 4'-position of 6a or the equivalent 6'-position of 7a led to dramatic reduction in binding potencies at all the three opioid receptors, indicating the existence of a somewhat similar steric constraint at the ligand binding sites of delta, mu, and kappa receptors. In contrast, the introduction of a phenyl group at the 5'-position of 6a did not cause any reduction in the binding affinity at the delta receptor. In comparison to the unsubstituted pyridine 6a, the 5'-phenylpyridine 6c showed improvements in mu/delta and kappa/delta binding selectivity ratios as well as in the delta antagonist potency in the MVD. Interestingly, introduction of a chlorine atom at the para position of the pendant 5'-phenyl group of 6c not only provided further improvements in delta antagonist potency in the MVD but also shifted the intrinsic activity profile of 6c from an antagonist to that of a mu agonist in the GPI. Compound 6d thus possesses the characteristics of a nonpeptide mu agonist/delta antagonist ligand with high affinity at the delta receptor (K(i) = 2.2 nM), high antagonist potency in the MVD (K(e) = 0.66 nM), and moderate agonist potency in the GPI (IC(50) = 163 nM). Antinociceptive evaluations in mice showed that intracerebroventricular (icv) injections of 6d produced a partial agonist effect in the 55 degrees C tail-flick assay and a full agonist effect in the acetic acid writhing assay (A(50) = 7.5 nmol). No signs of overt toxicity were observed with this compound in the dose ranges tested. Moreover, repeated icv injections of an A(90) dose did not induce any significant development of antinociceptive tolerance in the acetic acid writhing assay. The potent delta antagonist component of this mixed mu agonist/delta antagonist may be responsible for the diminished propensity to produce tolerance that this compound displays.

2,4-Diaminopteridines (21 compounds) and 2,4-diamino-5-methyl-5-deazapteridines (34 compounds) along with three 2,4-diamino-5-unsubstituted-5-deazapteridines and four 2,4-diaminoquinazolines, each with an aryl groups attached to the 6-position of the heterocyclic moiety through a two-atom bridge (either CH2NH, CH2N(CH3),CH2S, or CH2CH2), were synthesized and evaluated as inhibitors of the growth of Toxoplasma gondii in culture and as inhibitors of dihydrofolate reductase enzymes from T. gondii, Pneumocystis carinii, and rat liver. Exceptionally high levels of combined potency and selectivity as growth inhibitors of T. gondii and as inhibitors of the microbial enzymes relative to the mammalian enzyme were found among the 5-methyl-5-deazapteridines but not for the other heterocyclic types. Thirty of the 34 5-methyl-5-deaza compounds gave growth inhibition IC50 values lower than that of pyrimethamine (0.4 microM) with 14 compounds below 0.1 microM, values that compare favorably with those for piritrexim and trimetrexate (both near 0.02 microM). As inhibitors of T gondii DHFR, all but three of the 34 5-methyl-5-deaza compounds gave IC50 values in the order of magnitude with those of piritrexim (0.017 microM) and trimetrexate (0.010 microM), and 17 compounds of this group gave IC50 values versus P. carinii DHFR similarly comparable with those of piritrexim (0.031 microM) and trimetrexate (0.042 microM). Thirteen of these congeners gave both T. gondii growth inhibition and DHFR inhibition IC50 values of 0.10 microM or less, thus indicating facile penetration of the cell membrane. Eleven of these inhibitors of both T. gondii growth and DHFR have selectivity ratios (IC50 rat liver divided by IC50 T. gondii) of 5 or greater for the parasite DHFR. The highest selectivity ratio of nearly 100 belongs to the 5-methyl-5-deaza compound whose 6-substituent is CH2CH2C6H3(OCH3)2-2,5. This compound is over 10(3)-fold more selective for T. gondii DHFR than bridge homologue piritrexim (selectivity ratio 0.088), a compound now in clinical trials. The candidate with CH2NHC6H3(CH3)2-2,5 in the 6-position gave the highest P. carinii DHFR selectivity ratio of 4.0, which is about 60-fold more selective than trimetrexate (0.071) and 80-fold more selective than piritrexim (0.048) toward this enzyme. The 10 best compounds with respect to potency and selectivity includes six compounds bearing 2,5-disubstituted phenyl groups in the side chain (with little, if any, difference in effects of methyl, methoxy, or ethoxy), two side chains bearing 1-naphthyl groups, and two with 5,6,7,8-tetrahydro-1-naphthyl groups. Bridge groups represented in the 10 choice compounds are CH2NH, CH2N(CH3), CH2CH2, and CH2S. The high levels of both potency and selectivity among these agents suggest that in vivo studies now underway may lead to agents that could replace trimetrexate and piritrexim in treatment of toxoplasmosis and P. carinii pneumonia.