ESTIMATION 



hydroxide. They form crystalline methyl esters with m.ps. i6i to 

 162° C. and 163 to 164° C. respectively and [a]|^° = + 47° and + 39° 

 respectively in chloroform.^ According to du Vigneaud et al} the 

 methyl ester of biotin isolated from liver has m.p. 166 to 167° C. and 



Mf = + 57°. 



a- and ^-Biotin methyl esters are sparingly soluble in ether and 

 in light petroleum, somewhat more soluble in benzene or cyclohexane, 

 and soluble to the extent of about i % in methanol. In general, 

 halogenated hydrocarbons, alcohols and ketones are good solvents for 

 the esters, mesityl oxide, phorone, methyl heptenone and allylbenzene 

 being particularly useful for purposes of crystallisation.^ 



j8-Biotin is surface-active.^ 



References to Section 5 



1. F. Kogl and E. J. ten Ham, Naturwiss., 1943, 31, 208. 



2. V. du Vigneaud, K. Hofmann, D. B. Melville and P. Gyorgy, /. 



Biol. Chem., 1941, 140, 643. 



3. F. Kogl and L. Pons, Z. physiol. Chem,, 1941, 269, 61. 



4. V. R. Williams and H. B. Williams, /. Biol. Chem., 1949, 177, 745. 



6. STABILITY OF BIOTIN 



Biotin is inactivated by acid and alkali and by many a-amino acid 

 reagents, though not by ninhydrin.^ It is not destroyed by acylating 

 or alkylating agents or by carbonyl reagents. It is inactivated by 

 hydrogen peroxide and by rancid oils and fats with a high peroxide 

 content ; the rate of destruction is less in presence of a-tocopherol.^ 

 Since the product was active on yeast, but not on L. helveticus (see 

 page 452), it is most probably biotin sulphone. Biotin is also inac- 

 tivated by choline.^ 



References to Section 6 



1. G. B. Brown and V. du Vigneaud, /. Biol. Chem., 1941, 141, 85. 



2. P. L. Pavcek and G. M. Shull, ibid., 1942, 146, 351. 



3. J. S. Harrison and E. J. Miller, Analyst, 1949, 74, 463. 



7. ESTI2V1ATION OF BIOTIN 

 IVlicrobiological Methods 



The method first used for the estimation of biotin was a yeast 

 growth method,^' 2 which is not applicable to the examination of 

 turbid or highly coloured solutions, as the growth response is measured 

 turbidimetrically. Nevertheless, the method gives good results where 



421 



