Oct. 20, 1962
COMMUNICATIONS TO THE EDITOR
3977
eration of zero-point energy differences for a C-H the two forces holding the proton in the transition stretching vibration of 2900 cm.-' is 7.8, and the state should be more nearly equal. Thus, the isolargest deuterium isotope effect reported in elec- tope effect with trimethoxybenzene should be trophilic aromatic substitution is 6.67.7 Thus, the stronger than those reported before. isotope effect weakening produced by the symmetriThis isotope effect on proton transfer to trical stretching vibration in the transition state of a methoxybenzene is twice the value which was used three-center reaction must be near its minimum to estimate that HzO is five times as strong an acid value in this case. as Dz0.2b Similar reasoning with the data for The 0-H stretching vibration in the solvated trimethoxybenzene gives a ten-fold difference in proton occurs a t 2900 cm.-1,8 and the value pre- acid strength between H20 and DzO. This undicted for the other isotope effect, klH20/klDa0, is 7.8 reasonably high isotope effect emphasizes the as well. This is considerably greater than the ob- danger inherent in basing conclusions on the asserved value of 2.93. But the transition states of sumption that isotope effects will have essentially the two steps in the exchange reaction are the same, constant values. and, if the observed valueof k2H/k2D is near its DEPARTMENT OF CHEMISTRY maximum value, the observed value of k~H20/klD~0ILLINOIS INSTITUTE OF TECHNOLOGY 16, ILLINOIS A . J . KRESGE must be near its maximum value also. This dis- CHICAGO UNIONCARBIDE RESEARCH INSTITUTE crepancy can be understood in terms of a predicted TARRYTOWN, NEW YORK Y. CHIANG secondary effect of the water molecules solvating RECEIVED JUNE 18, 1962 When a proton is transferred from its the p r o t ~ n . ~ solvent shell, the solvating water molecules revert to ordinary water. Since the 0-H stretching vibraCELLULOSE COLUMNS CONTAINING POLYRIBONUCLEOTIDES AND RIBONUCLEIC tion in liquid water is 3400 cm.-', this process is ACIDS accompanied by considerable bond-tightening. It has been estimated that the deuterium isotope Sir : Recent studies2, have shown that polydeoxyrieffect on this change is 0.7 per 0-H bond for an equilibrium process, 9a and the prediction has been bonucleotides can be covalently linked to cellulose made that this will reduce the kinetic isotope effect and then be employed as chromatographic adsoron proton transfer to a maximum value of about bents which selectively bind polynucleotides com3.6.9b The observed effect of 2.93 is in good agree- plementary to them in base sequence. The methods of synthesis utilized the glucosidic hydroxyl groups ment with this prediction. of T4 DNA2 and the terminal phosphate groups of TABLE I thymidine oligonucleotides.s This suggested to us that the free hydroxyl groups on ribose in polyRATESOF AROIIATIC HYDROGEN EXCHANGE BETWEEN 1,3,5ribonucleotides might be available for similar TRIMETHOXYBENZENE AND 0.050 M HCIOl AT 25' reactions. We have found that phosphocellulose Substrate Solvent lo* kz ( M - 1 m h - 1 ) No. of runs can indeed be linked to synthetic polyribonucleoTMB-t Hz0 3.722 A 0.030" 9 tides and made into columns capable of binding and TMB-d HzO 7.98 A . i o 0 7 desorbing polynucleotides. The specificity of adTMB-t Dz0 6.286 A .012" 5 sorption with respect to the nature of the bases, Error estimates are standard deviations of the mean salt concentration and temperature, is very similar values. to that for the formation of helical complexes in The difference between this approximately solution. Columns also have been prepared from maximum isotope effect on proton transfer to 1,3,5- natural ribonucleic acids. Use of these columns trimethoxybenzene and the other smaller isotope may constitute a chromatographic method for effects on slow proton transfer2 is understandable in isolating cellular components complementary in terms of the relative basicities of the various pro- base sequence to the RNAs. ton donors and acceptors.gb Isotope effects in We followed a procedure similar to Bautz and three-center reactions will be less than their Hall's2 adaptation of Khorana's carbodiimide reacmaximum value whenever the transition state is tion for forming phosphate-ester bonds between not truly symmetrical, that is, whenever the two acetylated phosphocellulose (Serva, 0.78 meq. P/g.) force constants governing the symmetrical stretch- and each of the listed polyribonucleotides: pylying vibration in the transition state are not equal.lO A, poly-C, poly-I, poly-U, bacteriophage virus I n the slow proton transfer reactions on which iso- RNA, E. coli transfer and ribosomal RNAs. tope effects have heretofore been reported, the pro- Nucleotide polymers and cellulose were dissolved in ton acceptor has usually been a strongly basic pyridine and reacted with dicyclohexyl carbodianion. In these transition states, therefore, the imide a t 115' for one hour. After the reaction proton is not bound with equal strength to the ac- product was isolated, it was chopped in a Waring ceptor molecule and the solvating water which it is blender a t 4 O , ground in a mortar, and washed exleaving. Trimethoxybenzene, on the other hand, tensively with neutral buffer a t 80' to liberate is a weaker base than these anions, and in this case pyridine and starting materials. After removal of (7) H.Zollinger, Rclu. Chim. Acta, 88, 1957,1617,1623 (1955). (8) M. Falk and P. A. Giguere, Can. J . Chem., 8 5 , 1185 (1959); C. G.Swain and R. F. Bader, Tetyahcdron, 10, 182 (1960). (9) (a) C. G. Swain and E. R. Thornton, J. A m . Chem. Soc., 88, 3884 (1961); (h) C. A. Bunton and V. J. Shiner, Jr., {bid., 8 8 , 42, 3207, 3214 (lY6l). (10) F. H Westheimer, Chcm. Reus.. 61, 265 (1961).
(1) Abbreviations used: poly-A or simply A, polyriboadenylic acid; poly-C or C, polycytidylic acid; poly-I or I, polyinosinic acid; poly-U or U, polyuridylic acid; tris, tris-(hydroxymethy1)-aminomethane. (2) E. K.F. Bautz and B. D. Hall, Proc. Nail. Acad. Sci. U.S., 48, 400 (19621. (3) P.T.Gilham, J . A m . Chcm. Soc., 84, 1311 (1962).
397s
COMUUNICATIOKS TO THE EDITOR
Vol. s4 TABLE I
BINDINGCAPACITY OF VARIOUSPOLYRIBOKUCLEOTIDECELLULOSE COLUMNS~
2 01 c
Cellulose column
z
w
LL LL
151
w
z H
I
3
i
8
1
_1
z
I
ol
0
nN 0 A
a
6
'
i 0 5
c
20
30
40
50
60 T 'C
70
SO
90
PI o
100
+
Fig. 1.-Melting curves for poly-A poly-U complex in solution and on column; the same polymer samples were used for both experiments; all samples were in 0.01 A1 tris buffer, p H 7.4, and 0.5 M NaCl: 0, optical density in solution for 1: 1 mixture of poly-A poly-U; 0, elution of polyU from poly-A column.
+
fine cellulose particles, the preparations were packed into chromatographic columns. By using CI4-labeled poly-A and transfer-RNA, we estimate that 10 to 30% of each polymer reacted with phosphocellulose; of this, 50 to 90% was discarded with the fine particles. For example, upon hydrolysis with ribonuclease, the poly-C cellulose column yielded 1.3 pmoles of cytidylic acid (0.4 mg.), thus confirming that little more than 1% of the original 30 mg. of polynucleotide was fixed in the chromatography column. This poly-C column specifically adsorbed 0.34 pmole of poly-I, presumably by forming a 1: 1hydrogen-bonded complex, indicating that the poly-C on the column was about 25% efficient in binding complementary polymers. The attachment of phosphocellulose to these lower molecular weight polyribonucleotides appears to proceed to a much smaller extent than its linkage to T4 DNA.2 This difference may reflect a greater accessibility of sidechain glucosidic hydroxyls than of the backbone ribose C'Z hydroxyl groups. We are investigating the possibility that there may be a component of mechanical entrapment in the cellulose-pyridine gel of the reaction mixture. This suggestion appears unlikely, however, since the small transfer-RNA molecules were retained to the same extent as the much larger poly-9 strands. Column specificity was determined by loading the column with solutions of various polynucleotides under complexing conditions until it was saturated. Elution was carried out under conditions designed to destroy nucleotide complexes (low salt or high temperature). Recovery of the test polymer was always quantitative upon elution with 0.001 M tris. Conditions for column loading and elution are listed in Table I, together with the quantities of
Amount of solution bound in mrmulesb Poly-A Poly-C Poly-I Poly-U
Poly-A 157
