In a search for more selective A1 adenosine receptor agonists, N6-[(R)-(-)-1-methyl-2-phenethyl]-1-deazaadenosine (1-deaza-R-PIA, 3a), N6-cyclopentyl-1-deazaadenosine (1-deazaCPA, 3b), N6-cyclohexyl-1-deazaadenosine (1-deazaCHA, 3c), and the corresponding 2-chloro derivatives 2a-c were synthesized from 5,7-dichloro-3-beta-D-ribofuranosyl-3H-imidazo[4,5-b]pyridine. On the other hand, N-ethyl-1'-deoxy-1'-(1-deaza-6-amino-9H-purin-9-yl)-beta-D-ribofuranu ronamide (1-deazaNECA, 10) was prepared from 7-nitro-3-beta-D-ribofuranosyl-3H-imidazo[4,5-b]pyridine, in an attempt to find a more selective A2 agonist. The activity of all deaza analogues at adenosine receptors has been determined in adenylate cyclase and in radioligand binding studies. 1-DeazaNECA proved to be a nonselective agonist at both subtypes of the adenosine receptor. It is about 10-fold less active than NECA but clearly more active than the parent compound 1-deazaadenosine as an inhibitor of platelet aggregation and as a stimulator of cyclic AMP accumulation. The N6-substituted 1-deazaadenosines largely retain the A1 agonist activity of their parent compounds, but lose some of their A2 agonist activity. This results in A1-selective compounds, of which N6-cyclopentyl-2-chloro-1-deazaadenosine (1-deaza-2-Cl-CPA, 2b) was identified as the most selective agonist at A1 adenosine receptors so far known. The activity of all 1-deaza analogues confirms that the presence of the nitrogen atom at position 1 of the purine ring is not critical for A1 receptor mediated adenosine actions.