Novel classes of heterocyclic compounds as adenosine antagonists were developed based on a template approach. Structure-affinity relationships revealed insights for extended knowledge of the receptor-ligand interaction. We replaced the bicyclic heterocyclic ring system of earlier described isoquinoline and quinazoline adenosine A(3) receptor ligands by several monocyclic rings and investigated the influence thereof on adenosine receptor affinity. The thiazole or thiadiazole derivatives seemed most promising, so we continued our investigations with these two classes of compounds. The large difference between a pyridine and isoquinoline ring in binding adenosine A(1) and A(3) receptors showed the importance of the second ring of the isoquinoline ligands. We prepared several N-[4-(2-pyridyl)thiazol-2-yl]benzamides, and these compounds showed adenosine affinities in the micromolar range. Most surprising in the series of the N-[4-(2-pyridyl)thiazol-2-yl]amides were the retained adenosine affinities by introduction of a cylopentanamide instead of the benzamide. A second series of compounds, the thiadiazolobenzamide series of compounds, revealed potent and selective adenosine receptor antagonists, especially N-(3-phenyl-1,2,4-thiadiazol-5-yl)-4-hydroxybenzamide (LUF5437, 8h) showing a K(i) value of 7 nM at the adenosine A(1) receptor and N-(3-phenyl-1,2,4-thiadiazol-5-yl)-4-methoxybenzamide (LUF5417, 8e) with a K(i) value of 82 nM at the adenosine A(3) receptor. 4-Hydroxybenzamide 8h is the most potent adenosine A(1) receptor antagonist of this new class of compounds. Structure--affinity relationships showed the existence of a steric restriction at the para-position of the benzamide ring for binding adenosine A(1) and A(3) receptors. The electronic nature of the 4-substituents played an important role in binding the adenosine A(3) receptor. Cis- and trans-4-substituted cyclohexyl derivatives were made next to the 4-substituted benzamide analogues. We used them to study the proposed specific interaction between the adenosine A(1) receptor and the 4-hydroxy group of this class of thiadiazolo compounds, as well as a suggested special role for the 4-methoxy group in binding the A(3) receptor. Both the adenosine A(1) and A(3) receptor slightly preferred the trans-analogues over the cis-analogues, while all compounds showed low affinities at the adenosine A(2A) receptor. Our investigations provided the potent and highly selective adenosine A(1) antagonist N-(3-phenyl-1,2,4-thiadiazol-5-yl)-trans-4-hydroxycyclohexanamide (VUF5472, 8m) showing a K(i) value of 20 nM. A third series of compounds was formed by urea analogues, N-substituted with thiazolo and thiadiazolo heterocycles. The SAR of this class of compounds was not commensurate with the SAR of the previously described quinazoline urea. On the basis of these findings we suggest the existence of a special interaction between adenosine receptors and a region of high electron density positioned between the thia(dia)zole ring and phenyl(pyridyl) ring. Molecular electrostatic potential contour plots showed that for this reason the ligands need either a thiadiazole ring instead of a thiazole or a 2-pyridyl group instead of a phenyl. The derived novel classes of antagonists will be useful for a better understanding of the molecular recognition at the adenosine receptors.