SEPARATION BY PAPER CHROMATOGRAPHY 253 



glycol with butanol permits addition of more water and gives higher Rp values, with- 

 out markedly altering the order of separation of the bases.'' Addition of ethanol gives 

 higher Rf values with some loss of resolution, ^^ Dioxane-* and 2-methoxyethanol 

 (methyl cellosolve)^^ have been added to butanol-water mixtures, and also lead to 

 increased Rf values. Water-saturated n-butanol has been mixed with acetic acid, 

 with ethyl acetate and morpholine, and with methyl glycol and morpholine with 

 some success in separating deoxj^ribosides.** 



Amyl alcohol saturated with water gives inconveniently low Rf values for most 

 substances. 2^ Isopropanol-water-ammonia (Table I, solvent e) has been found a use- 

 ful mixture by Hershey, et al. :" it resolves 5-hydroxj'methylcytosine from the other 

 bases, but adenine and uracil run together, and the bases are less well separated from 

 their ribosides than in butanol. Mixtures of tetrahydrofurfuryl alcohol with propanol 

 and amyl alcohol buffered at different pH's have also been tried, ^* with some success 

 in separation of nucleotides, but for separation of the bases and nucleosides they 

 appear to be inferior to simple butanol solvents. Several other mixtures of alcohols 

 with NH3 and HCl have also been tested on a limited range of substances, ^^ and the 

 Rf values of orotic acid^" and of uric acid and its riboside*' in several solvents have 

 been recorded. 



Solvent systems based on collidine and quinoline instead of alcohols have been 

 tested by Vischer and Chargaff.'' In these (e.g.. Table I, solvent g) the bases migrate 

 in a different order, xanthine and hypoxanthine running more rapidly and cytosine 

 much more slowly. However, the strong absorption of ultraviolet light by these sol- 

 vents is a serious drawback. 



Isobutyric acid mi.xed with water and ammonia, as tested bj' Lofgren,*^ distrib- 

 utes the bases and nucleosides in a different order from other solvents, as shown in 

 Table I (solvent h), although Rf values are grouped undesirably close together. 

 (Compare also Tamm et al.^^'^) The effect is partially retained, with better spread of 

 Rf values, in a solvent containing u-butanol (75 ml.), isobutyric acid (37.5 ml.), 

 water (25 ml.), and ammonia (2.5 ml. of 25% soln.). A mixture containing piperidine, 

 tried by the same author, gave rather poor separations. 



A limitation of all these neutral or weakly basic or acidic solvent systems 

 for quantitative analysis of nucleic acids is their low capacity for guanine. 

 Because of its insolubility, guanine in amounts of more than a few micro- 

 grams tends to form "tails" or double spots, or to remain partly at the ori- 

 gin. This difficulty may be avoided by use of solvents containing relatively 

 high concentrations of hydrochloric acid. Such a system was first used by 

 Smith and Markham^' for separation of the purine bases and pyrimidine 

 nucleotides obtained by mild acid hydrolysis of PNA (Table II, solvent a). 

 A mixture containing isopropanol and hydrochloric acid (Table I, solvent 

 /) was subsequently developed for separation of the bases from DNA.^^ 



" S. G. Laland, W. G. Overend, and M. Webb, J. Chem. Soc. 1952, 3224. 



" W. S. MacNutt, Biochem. J. 50, 384 (1952). 



" A. D. Hershey, J. Dixon, and M. Chase, J. Gen. Physwl. 36, 777 (1953). 



" D. C. Carpenter, Anal. Chem. 24, 1203 (1952). 



" B. Bheemeswar and M. Sreenivasaya, Current Sci. (India) 20, 61 (1951). 



«" E. Leone and E. Scala, Boll. soc. ital. biol. sper. 26, 1223 (1950). 



" E. Leone and D. Guerritore, Boll. soc. ital. biol. sper. 26, 609 (1950) . 



62 N. Lofgren, Acta Chem. Scand. 6, 1030 (1952). 



