Radical relay chlorination selectivity

The selectivity of the radical relay chlorination is striking. In the case of the enone (13 Sdieme 18), and in related compounds with A-ring dienones, the m-iodobenzoate template at C-17 directs chlorination to C-9 and not to die preexisting functional groups of (13). The selective chlorination of C-9 seems to be quantitative, although in the first report the A "Lalkene (14) was isolated in only 77% yield. Later work has shown that the overall introduction of this double bond can have yields in the 90-93% range, and good yields for this reaction have also been reported from another laboratory. Template-directed radical relay chlorination on the a-face of steroids has also been successful in the a/b cu-coprostanol steroid series,and in the cholestanol series with iodophenyl templates linked by amide, ether, or sulfonate functions rather than carboxylic esters. ... 

Template-catalyzed remote chlorination reactions have also been examined in molecular complexes. In one early study (Scheme 24), ion pairing was used to hold a charged template near a chaig substrate. Selective catalyzed radical relay chlorination was observed, but the selectivity was not as good as has been seen when the template is covalently attached to the substrate. In more recent work better selectivity and some catalytic turnover has been observed. 

Catalytic turnover has also been seen in radical relay chlorinations in which the template is temporarily linked to the substrate in a mixed metal complex. The steroid phosphate (22) and catalytic ligand (23) both bind to zinc in a mixed complex, and the iodine atom of (23) directs chlorination of (22) at C-9 with reasonable selectivity (Scheme 25). Five or more turnovers are seen, when only 10% of the catalyst (23) is used. ... 

We also attached three steroid substrates to an iodophenyl template in 16 (Scheme 6-7) using a tris-silyl ether link, and saw that the template could direct selective radical-relay chlorination to all three tethered substrate species in the same work we reported a similar finding when a thiophene ring was the triply tethered template [49]. Thus the sulfur atom of thiophene can perform radical relay. In these triple catalytic functionalizations, the more reactive sulfuryl chloride was the preferred reagent. 

We have largely been describing reactions on steroid substrates, which are conforma-tionally rigid and permit selective functionalizations by appropriate tethered templates. However, when the templates are linked to flexible chains, the results can be used to learn about the conformational preferences of such flexible chains. In one study [56], we examined the positional selectivities of insertion reactions into flexible chains by attached benzophenone units, a process we had also examined earlier [31], and compared the results with those from the intramolecular chlorination of such flexible chains by attached aryliodine dichlorides. The results were complementary. In another study [57] we used long-chain alkyl esters of nicotinic acid in radical relay chlorination, and saw some interesting selectivities reflecting conformational preferences in these nominally flexible cases. 

The majw work to date on synthetic applications of remote functionalization has involved free radical chlorination. The earliest studies involved the direct attachment of aryliodine dichloride units to the steroid substrates, then intramolecular free radical chain chlorination in benzene or chlorobenzene solution (Scheme 14). Yields were only in the 50% region, but fairly good selecdvities were observed compound (6) afforded chiefly the 9-chloro derivative, while compound (7) produced the 14-chloro steroid. The yields and selectivities were considerably improved when it was realized that aromatic solvents promote intermolecular random processes by forming complexes with C1-, and when the radical relay method was developed. 

Limited studies have been done on template-directed chlorination on the -face of steroids. Compound (15 Scheme 19) was designed so that die template could curve around the angular C-18 methyl group and direct chlorination to C-20. Reaction with an excess of PhICh led to ca. 40% chlorination of C-20 with 25% unftmctionalized steroid. The 20-chloro steroid was converted in part to the which was ozonized to form the 17-acetyl steroid (15). A similar result was observed with the i-steroid derivative (17). The selectivities and yields are not yet up to those of other examples of the radical relay reaction. 

Diaiyl sulfide templates have also been used to direct chlorinations. The selectivities indicate that the chlorine atom is bound to the sulfur, but the yields are not as good as those with aryl iodide templates. The problem is that the sulfur gets oxidized under the reaction conditions. As expected, a thiophene ring is more stable to oxidation and its sulfur atom can still bind chlorine in a radical relay process." The best sulfur template so far examined is the thioxanthone system (Scheme 20). Thus with 3 equiv. PhICh compound (18) undergoes directed C-9 chlorination in 100% conversion, affording a 71% yield of the A iO-alkene sifter base treatment, along with some polar products from excessive chlorination. The thioxanthone template can be recovered unchanged. 

As we have described elsewhere [9], we were able to use this radical relay process to complete a successful synthesis of cortisone. With another appropriate template we were able to achieve a selective chlorination at C-17 of the steroid, which permitted us to remove the steroid side chain [10]. Recent work in Germany [11] has resulted in a somewhat easier sequence for converting the 17-chloro steroid derivative to a useful intermediate in which the side chain has been removed. It remains to be seen whether the steroid reactions by which we can produce corticosteroids or remove steroid side chains and generate useful intermediates actually lead to practical transformations of industrial interest. These reactions are under active investigation in several companies. 

---