Recently we reported [1,2] the synthesis and antibacterial activity of 3-deoxy-5-O-(4deoxymycaminosyl)tylonolide ( 1). This study demonstrated that removal of both the 3and 4'-hydroxy groups of 5-O-mycaminosyltylonolide greatly enhances the activity. To prepare a more-active derivative than 1, introduction of a methyl group at C-2 in the macrolactone ring of 1 has been undertaken. This modification was expected to give a derivative having longer duration in the blood through restriction of the approach of enzymes hydrolyzing the lactone ring (ring opening inactivates 1) by the presence of the methyl group. However, conventional methylation was expected to cause simultaneous methylation of the 2'-hydroxy and 3'-dimethylamino groups. In relation to this synthesis, we were also interested in preparing the 2'-O-methyl derivative 6, as 6 was expected to be a key probe for determining whether the T-esters of 1 are themselves active or not (that is, they show activity without or after hydrolysis). In erythromycin, a macrolide antibiotic similar to 1, the 2'-esters were inactive until hydrolyzed [3,4], but in human plasma, they were hydrolyzed enzymatically to restore the activity, the hydrolytic state being inverse-proportionally related to the lipophilicity of the 2'-acyl groups attached. This character suggests that erythromycin T-esters might be useful as prodrugs. In our compound 1, if the T-O-methyl derivative 6 has antibacterial activity (in vitro, for example) similar to that of 1, the free 2'-hydroxyl would not be judged essential for activity and thus the 2'-O-acyl derivatives (7-11), especially the simple 2'-O-acyl derivative 7, would predictably have antibacterial activity. However, a lack of activity in 6 would indicate the free 2'-OH group to be important, and that the 2'-O-acyl derivatives would show activity after hydrolysis.