In this study, we investigated the hypothesis of agonist-directed trafficking of receptor signaling for the alpha(2A)-adrenergic receptor (alpha(2A)-AR). alpha(2A)-ARs couple to both G(s) and G(i) to stimulate or inhibit adenylyl cyclase activity. Chinese hamster ovary-K1 cell lines expressing the porcine alpha(2A)-AR at high (alpha(2A)-H) and low (alpha(2A)-L) levels were used to estimate the relative efficacies (R.e.s) of a series of agonists for the G(s) and G(i) pathways. G(s)-mediated responses were measured after pertussis toxin treatment to inactivate G(i) in alpha(2A)-H, whereas G(i) responses were measured in alpha(2A)-L, where G(s) responses were absent. The full agonist UK-14,304 showed a large receptor reserve for G(i) responses in alpha(2A)-H but little receptor reserve for G(s) responses in alpha(2A)-H or for G(i) responses in alpha(2A)-L. With the exception of l-isoproterenol (ISO), all agonists showed similar R.e.s at the alpha(2A)-AR for G(s) and G(i) responses, with rank orders of R.e.s as follows: l-epinephrine = l-norepinephrine = UK-14,304 > p-aminoclonidine > or = BHT-920 > or = BHT-933 > clonidine = p-iodoclonidine > or = xylazine > or = guanabenz. Interestingly, ISO had the highest efficacy at the alpha(2A)-AR for activating G(s) versus G(i) (9-fold higher); however, it had low potency for both. By several criteria, the ISO response was mediated by the alpha(2A)-AR, supporting the hypothesis of agonist-directed trafficking of receptor signaling or agonist-specific G protein selectivity. In contrast, the apparent G(i) pathway selectivity of oxymetazoline appears to be mediated by an endogenous serotonergic receptor. It is intriguing that a classic beta-AR agonist that activates G(s) through beta(2)-ARs also appears to produce a G(s)-selective conformation of the G(i)-coupled alpha(2A)-AR.