438 D- M. BROWN AND A. R. TODD 



which meets the above requirements and which depends on the known 

 lability towards alkali of esters, and in particular, phosphates of /3-keto and 

 /3-aldehydo alcohols. Such phosphates readily undergo an elimination reac- 

 tion under these conditions ;^^2 the reaction is illustrated by the conversion 

 of glyceraldehyde-3-phosphate to lactic acid under very mild alkaline and 

 pyruvaldehyde under acid conditions. ^^'-^^^ In model experiments,^^^ 

 adenosine-5 '-phosphate and adenosine-5' benzyl phosphate were oxidized 

 with periodic acid to give dialdehydes of formula XXV (Ad = adenine 

 residue, R = H or CH2C6H5) in which the phosphate residue is attached 

 in the /3-position to one of the aldehyde groups. In accordance with expecta- 

 tion, these products were extremely labile to alkali, the phosphoryl (or 

 benzylphosphoryl) group being rapidly removed at room temperature 

 even at pH 10.5; under these conditions the unoxidized nucleotides, as well 

 as ribonucleic acids,^^^^ are completely stable. Since adenosine-5' benzyl 

 phosphate is structurally analogous to a polynucleotide (in the latter the 

 benzyl group is replaced by a polynucleotide chain) the potentialities of 

 this observation are evident. If formula XXVI represents a polynucleotide 

 made up of n nucleotides, removal of the terminal phosphoryl groups by 

 means of a phosphomonoesterase would yield XXVII, which, containing 

 a free a-glycol system, would be oxidized by periodic acid to XXVIII. By 

 analogy with oxidized adenosine-5 '-phosphate, mild alkaline treatment at 

 pH 10.5 would be expected to remove the oxidized terminal nucleoside 

 residue (which could be identified by various methods) yielding XXIX, 

 which is a polynucleotide containing n — 1 residues but otherwise exactly 

 Uke the original XXVI ; on this product the whole process could be repeated. 



Although the method has not yet been applied to a large polynucleotide, 

 there is no reason to doubt its validity in such a case, especially as it has 

 been found applicable to dinucleotides where n = 2}^^ At the present time 

 this method represents the only reasoned approach to determination of 

 nucleotide sequence; if successful in its application it should also provide 

 definite information on the nature and extent of any branching which exists 

 in a given polynucleotide. 



It is of interest to note that the polynucleotide XXVI can be regarded as 

 a polymer in which the monomeric units are nucleoside-3 '-phosphates. It 

 would be equally reasonable (and it would not affect the previous discus- 

 sion) if it were represented as XXVIa, in which the monomeric units are 

 nucleoside-5 '-phosphates. To which of these two types the natural nucleic 

 acids belong is at present unknown. That apparently successful end-group 



132 R. p. Linstead, L. N. Owen, and R. F. Webb, /. Chem. Soc 1953, 1211. 



133 O. Meyerhof and K. Lohmann, Biochem. Z. 271, 89 (1934). 

 13* E. Baer and H. O. L. Fischer, J. Biol. Chem. 150, 223 (1943). 

 134a Cf. also C. A. Zittle, J. Franklin Inst. 242, 221 (1946). 



136 P. R. Whitfeld and R. Markham, Nature 171, 1151 (1953). 



