'The configurations a t the 3 positions of 3D-hydrosy-4-oxn-5or-cstrane and of 3a,17@-dihydrosy-4-oxa-5a-estrane were assigned frorn a study of the anomerizatior~s of these compounds in aqueous tetrahydrofuran. 3a-Chloro-4-oxa-5a-estrane solvolyzed in alkalirie rnethanol to give a mixture of 3a-and 3D-methoxy-4-oxa-5a-estrane, the former predominating. 'The differer~ce in the solvolytic behavior of 3a-chloro-4-osa-cia-steroids arid of a-glycosyl halides is discussed.T h e solvolysis of 3cr-cl~loro-4-oxa-5~-cholesta1~e in buffered or alltaline alcohols gives 3-allcosy derivatives 116th predoininant retention of configuration (1). The solvolysis of acylglycosyl halides, on the contrary, generally gives a product of inverted config~~ration (2) if participation by neighboring groups is avoided (3, 4) and if there is not excessive hindrance t o approach of the substituting molecule (5). A possible explanation for the behavior of the steroidal chloroether \\las considered t o be the shielding of the 0 face of ring X by the angular methyl group (I). Although the shielding of this ring is believed t o be less complete than the shielding of rings C and D (6), it has been invoked frequently to explain cases of preferential reaction of ring A on its a face; a recent exanlple is the h ydroboration of A'-, A2-, A3-, and A4-cl~olestenes (7).T o test this possibility, n-e have synthesized a cl~loroether ( I X , X = I-I) laclting the angular methyl group, and have studied the stereochenlistry of its solvolysis in methanol.
Synthetic RoutesThe synthetic route t o the desired cl~loroether I X (X = I-I) is outlined in Reaction Sche~ne 1. T h e three-stage preparation from estrone of the unsaturated ltetone I (X = I-I) is described in the literature (8). Ozonolysis of this ltetone gave the lceto acid I1 ( X = I-I), \vhich \\-as reduced \vith sodium borohydride t o the hydroxy acid I11 ( X = I-I, Y = OH). On acidification, the latter fornled the lactone IV (X = I-I). T h e a configuration a t the 5 position of the lactone \lras indicated by its nuclear magnetic resonance (n.m.r.) spectrunl (discusscd in ref . 9), and by the negative shift in molecular rotation (Table I) when it was hydrolyzed in allcaline solution (10, 11). I-Iydrogenation of the lactone over platinuill in acetic acid containing a small amount of perchloric acid gave the ether V (X = I-I), whereas reduction with lithium alu~niniunl hydride (cf. refs. 1 and 12) or diborane (cf. Dehydration of the heiniacetal with phosphorus oxychloride in pyridine yielded the dihydropyran VIII ( X = I-I), which in ether solution added hydrogen chloride (cf. refs. 1 and 14) to give the reactive chloroether I X ( X = I-I). This reaction is ltnolvn t o be reversible (1,15), and the sanlple of chloroether sent for analysis lost hydrogen chloride conlpletel y when allowed t o stand, and analyzed as the dihydropyran. I-Iomever, the