In view of the potent analgesia exhibited by the apparent structurally dissimilar morphine-6-O-glucuronide (M6G) and morphine-6-O-sulfate (M6S) conjugates of morphine, we have examined the effect of structural modification of M6S on analgesic activity, using the tail-flick test. Changes in the M6S structure were made that would affect the lipophilicity and polarity of the molecule. Subcutaneous (sc) and intracerebroventricular (ICV) administration of equimolar doses of morphine, M6S, 3-O-acetylmorphine-6-O-sulfate (M3A6S), 3-O-benzoylmorphine-6-O-sulfate (M3B6S), and 3-O-acetyl-N-methylmorphinium-6-O-sulfate (MM3A6S) were employed. M6S and M3A6S exhibited a longer duration of action and greater activity compared to morphine after SC and ICV administration. However, M3B6S and MM3A6S in doses equimolar to that of morphine were found to be inactive after both SC and ICV administration. In addition, M3A6S showed the highest potency in inhibiting electrically stimulated guinea pig ileum followed by M6S and M3B6S. Moreover, both M6S and M3A6S displayed a greater affinity than that of morphine to mu and kappa 3 receptor sites in guinea pig brain homogenate. In contrast, the nonanalgesic compounds M3B6S and MM3A6S showed weak receptor binding ability compared to morphine. These results indicate that lipophilicity alone is not a determinant of analgesic activity in these novel morphine derivatives. These modified effects of morphine by the conjugations at the 3- and 6-position, appear to be due to their altered interactions with opioid receptors.