The Chemical Synthesis of Amino Acyl Adenylates* PauL Bere Department of Microbiology, Washington University School of Medicine, St. Louts, Missouri (Received for publication, March 3, 1958) The present paper deals with a description of the method of preparation and partial purification of several amino acid adenyl- ates. Two procedures for the synthesis of these compounds have previously been described, but these have certain disadvantages for general application. ‘The first of these methods (1) which leads to the synthesis of the leucyl, alanyl, and phenylalanyl adenylates in yields of about 10 per cent, involved the reaction of the amino acid acyl chloride with the silver salt of adenosine 5’-phosphate (A5P)!. In another method, Wieland é al. (2) used pi-valine thiophenol hydrochloride as the activated amino acid and they were able to effect a transfer of the valine moicty to ASP in vields of 10 to 20 per cent. Recently the usefulness of N, N’-dieyclohexylearbodiimide for the synthesis of nucleoside pyrophosphate derivatives was ele- gantly demonstrated by Khorana eé al. (3-5). Earlier, Zetzche and Fredrich (6) had used the earbodiimides for the synthesis of carboxylic acid anhydrides. It seemed, therefore, that the car- bodiimides might offer a useful reagent for coupling the amino acids to A5P by an acyl phosphate linkage. Soon after our studies were under way (7) Talbert and Huennekens (8) reported the synthesis of butyry] adenylate with DCC. The method to be described here involves the use of DCC to effect a condensa- tion of the carboxyl group of a free amino acid with the phosphate of adenylic acid in aqueous pyridine. Using this procedure the aclenylate derivatives of p- and L-methionine, L-phenylalanine, L-tryptophan, and L-serine have been prepared and purified. MATERIALS AND METHODS Crystalline A5P (free acid) was obtained from the Sigma Chemical Co. and the t-amino acids were products of the Cali- fornia Foundation for Biochemical Research or of Nutritional Biochemical Corp. ASP deaminase was prepared by the procedure for Preparation A of Kalckar (9) and dialyzed against 0.05 m potassium succinate buffer, pH 6.0, to remove ammonium sulfate. Hydroxylamine was freshly prepared by neutralizing a stock solution of 4 m hydroxylamine hydrochloride with 3.5 ~n NaQH to a pH of 6.5. The concentration of the amino acid adenylates was measured spectrophotometrically by conversion to the amino acid hydrox- amates. To 0.5 ml. of neutralized hydroxylamine (2 m) were added the amino acyl adenylate and water to a total volume of 1 ml. After three minutes 1 ml}. of a solution of acidic ferric chloride (10) was added and the mixture was shaken rapidly to * This work was supported by a research grant from the U. 8. Public Health Service. 1The ubbreviations used are: A5P, adenosine 5’-phosphate; ATP, adenosine triphosphate; DCC, dicyclohexylearbodiimide. remove gas bubbles, filtered, and the optical density at 540 mu was measured against a blank containing no amino acid adenyl- ate. The concentration was calculated with extinction cocfh- cients obtained with synthetic amino acid hydroxamates. otal A5P was determined with A5P deaminase (9) after pre- liminary hydrolysis of an aliquot of the amino acid adenylate at pH 10 for 5 minutes at room temperature. Free A5P (in the presence of amino acyl adenylate) was determined with a large amount of A5P deaminase to complete the reaction in 1 to 2 minutes, and thus minimize the slow liberation of A5P due to destruction of the amino acid adenylate. Ribose was determined by the Mejbaum method (11) with A5P as the standard. Methionine was determined by a modifi- cation of the method of McCarthy and Sullivan (12), and phos- phate was measured by the method of Fiske and SubbaRow (13). RESULTS DCC in aqueous pyridine brings about the formation of the substituted acyl phosphate derivative from an amino acid and A5P. With methionine, for example, the reaction proceeded to completion (Table I). The final value attained depended upon the amount of methionine or ASP employed and remained con- stant for at least 90 minutes. Whether this is due to the stability of the methiony! adenylate under these conditions or to the at- tainment of a steady state in which the rate of breakdown was equal to the rate of synthesis is not known. A detailed descrip- tion of the preparation and isolation of t-methionyl adenylate follows. Synthests of L-methionyl Adenylate—ui-Methionine (2.0 mmoles) and A5P (1.92 mmoles) were mixed with 3.2 ml. of cold water and 10.4 ml. of pyridine in a 250 ml. glass-stoppered flask. 8 N HC} (0.25 ml.) was added and the mixture was stirred in an ice bath with the aid of a magnetic stirrer. DCC (50 mmoles), dissolved in 12 ml. of pyridine, was added and the mixture was stirred vigorously. The formation of t-methiony! adenylate was determined on aliquots removed at various time intervals (see “Methods”). After 3 to 3.5 hours there was no further increase in methionyl adenylate formation. The value attained was usu- ally between 90 and 95 per cent of the theoretical maximum based on the amount of A5P used. The reaction was terminated and the crude methiony] adeny!- ate was precipiated by the addition of about 150 ml. of acetone chilled to —15°. After 45 seconds the precipitate was filtered rapidly with the aid of suction, washed with small portions of a mixture of acetone-alcohol (60:40) at 0°, then with ether (0°), and sucked almost dry on the filter. The material was then dried further at 3° overnight in vacue over P2Os and paraffin. The precipitation, washing, and air drying were completed in ap- proximately 8 minutes. The material obtained at this stage 608 September 1948 weighed 1.77 gm. and contained N,N“edieyelohexylurea, t-me- thionine, ASP, a trace of pyridine hydrochloride, and methionyl adenylate in about 40 to 50 per cent vield baged on the ASP consumed. Even with the precautions of working at low tem- peratures and rapid filtration, there seems to have been appreci- able breakdown of the amine acid adenylate. The crade material was evenly suspended in cold water at a coneentration of 100 mg. per ml. and Altered to remove the in- soluble dievelohexyl urea. ‘The residue was washed with 4 ml. of cold water and the wash and original filtrate were combined. The pH was adjusted to 3 to 4 with HCI and 85 ml. af cold ethanol (—15°} were added. After 1 hour at —15°, the pre- cipitate was filtered, washed with cold ethanol, and dried in vacuo at 3°, ‘Vhe vield was 670 mg. and this contained 0.6 mmole af methiony) adenylate with a purity of 46 per cent based on the optical density at 260 my. At this stage the methionyl adenylate was stable for periods of at least 2 months when kept dry and cold, The removal of A5P and further purification of the methiony! nilenylate was carried out as follows. Abowt 100 to 206 me. were dissolved in 2 to 4 ml. of cold water and the pH was adjusted to 4 to 4.3 with solid potassium bicarbonate. The solution was pissed through a Dowex 1 column (Cl form, 10 per cent cross- linked, 1 & 4 em. and the column wus then washed with 1 to 2m. of water. The wash and original liquid which had passed through the column were combined, adjusted to pH 4.5, and used as such. Under the shove conditions, free A5P was quanti- tatively adsorbed to the column, whereas 60 to 80 per cent of the methionvl adenylate appeared in. the efHuent (LE). The methionyl adenylate concentration decreased about 5 per cent per day when the material was kept frozen. Properties and Analysis of Methionyt Adenylate—Spectropho- Tasie I Siunthests of z-methionyl adenylate with DCC { ASP | i-Methionine i Methionine adenylute parnales wemeles minales 1.0 5.0 1.01 0.0 i 5.0 0.0 1.0 1.0 : 0.92 6.3) 0.75 0.30 0.3] 0.40 0.29 O31 ‘ O14 0.14 * Mensured as methionine ivdroxamate as «deseribed in **Meth- ods.** Tasie IL Analysts of L-methionyl adenylate i A3P Methianine Adenine? | Ribose Total P : i i i Pree Bound | Bound} “Fatal 1.00 08 0.97 0.03 0.99 Q.97 | LW 5 *'The values are expressed as ratigs based on adenine which was determined by the optical density at 260 my at pH 2 using an extinction coefficient of 15.1 < 10 ema? (14). t Determined as methionine hydroxamate as described in “Methods.” P. Berg 600 tometric examination of this solution revealed an ultraviolet ab- sorption spectrum which was essentially indistinguishable from that of free A5P. The ratios of the absorption at 280 to 260 my and 250-to 260 my at pH 2 were 0.22 and O84, respectively, Fria. 1. Paper electrophoretic separation of methionyl adenylate and ASP: The separation was carried out on Whatman 3MM paper using 0.02 st sodium citrate buffer, pH 3.1, and a voltage of 14 volts.perem. for 3 hours at 3°. A, erude u-methiony! adenylate exposed to 0.01 s KOH for 5 minutes at 25°; B, crude u-methionyl *, crude t-methionyi adenylate treated with 2 a0 hydroxylamine; C adenylate; D, ASP marker; and £, purified 1-methiony! adenylate. The spots were visualized and the photographs taken with an weraviolet lamp (2540 Ay, The small band sometimes found in crude meéthionyl adenylate preparations has not been identified but may be diadenosine 5’-pyrophosphate (15). Tapie LIT Synthesis of amina acyl adenylutes t i Spectral ratio Amina acyl eT haya Spectral! elias Fraction Yield eat A adenylate purity 286 muy 1230 my 260 mye: 26) sys % | OY u-Seryl ade- | Acetone precipitate 45 53 nylate Alcohol precipitate 32 4b Unadsorbed by Dowex | 25 OA | 0.22 | 0.85 i-Phenyl - Acetone precipitate §2 56 tlanylade- | Alechol precipitate 30 45 aylate Unadsorbed by Dowex 23 98 | 0.23 | 0.86 L-Trypto- Acetone precipitate 37 60 phanyl Alcohol precipitate 40 oF adenviate | Unadsorbed by Dowex 28 OTF | 0.39 | 0.86 p-Methiony] | Acetone precipitate adenylate | Atcohel precipitate Unadsorbed by Dowex ; 20 95 O.21 ) O84 "The fractions are the same as those described for the synthesis of L-methionyl adenylate. * "fhe purity is based on the ratio of the amount of amino acyl adenylate to total AdP. 610 for methiony] adenylate, compared to 0.22 and 0.84 for free ASP (14). Analyses for the various constituents of the purified methionyl adenylate (Table II) show reasonably good agreement between the AdP, total methionine, and bound methionine. Occasionally, certain preparations were contaminated with more free methi- onine than shown in Table II. This, however, rarely exceeded a value of 20 per cent free methionine. Paper electrophoresis studies with methionyl adenylate showed it to be slightly cationic at pH 3.1 and easily separable from ASP which migrates as an anion under these conditions (Fig. 1). Exposure of methionyl adenylate to neutral hydroxylamine or 0.01 x KOH for 5 minutes at room temperature resulted in the disappearance of the methionyl adenylate and formation of A5P. Synthesis of Other Amino Acyl Adenylates—The amino acyl adenylates of L-serine, L-phenylalanine, L-tryptophan, and p- methionine have been prepared with the use of the same pro- cedure already described for methionyl adenylate. The data fot recoveries, purity, and similar properties are summarized in Ta- ble IIT. DISCUSSION The alkyl carbodiimides have proved to be extremely usefu reagents for the synthesis of a number of compounds of biological interest. In addition to the nucleotide pyrophosphate deriva- tives (3-5), the unsymmetrical nucleotide pyrophosphate coen- zymes, such as cytidine diphosphocholine (16), diphosphopyridine nucleotide (17), flavinadenine dinucleotide (18), and, through the nucleoside 5’-phosphoramidate, uridine diphosphoglucose (19), have been prepared with DCC. More recently this reagent has been utilized for the synthesis of a deoxydinucleoside monophos- phate (20) and dideoxynucleotides (21). In the present case DCC has been employed to link an amino acid to A5P. The studies with L-methionyl adenylate indicate that the linkage is an anhydride between the amino acid carboxy] group and the phosphate of A5P (Fig. 2). The evidence for this conclusion is based on the following properties. The purified compound contains A5P and methionine in a 1:1 ratio. The absorption spectrum is identical to that of free A5P, indicating that the amino acid is not linked to the adenine group. At pH 3.1 the compound moves slowly as a cation on paper electro- phoresis and it is not retained by the strongly cationic adsorbent Dowex 1. The remaining uncertainty in the proof of structure is in the position of the amino acyl group. It could reside in an ester linkage on the 2’- or 3’-hydroxyl group of the ribose or as shown in Fig. 2 in an anhydride linkage with the 5’-phosphate group. De Moss e¢ al. (1) have used as evidence for a linkage with the 5’-phosphate group the inability of adenylic deaminase O Oo | 7 ° CH,;—S—(CH.)2—CH—C—O—P—O—CHy, Adenine | NH;* 0- H H H | H OH OH Fig. 2 Synthesis of Amino Acyl Adenylates Vol. 233, No. 3 to deaminate the amino acid adenylate derivative. However, it is not elear whether substitution in the 2’ or 3’ position on the ribose would likewise prevent the action of adenylic deaminase. The rapid and quantitative formation of the amino acid hy- droxamate in the presence of hydroxylamine at pH 6.5 would appear to argue in favor of the anhydride linkage. It has been pointed out recently (22), however, that amino acid esters also react with neutral hydroxylamine to form the hydroxamate. It should be emphasized however that Raacke (22) has pointed out that at pH 7 and below the rate of amino acid hydroxamate formation is very slow and usually incomplete. With the amino acyl adenylates the reaction is complete within a few minutes, a be- havior which is more characteristic of the anhydride. Moreover the enzymatic formation of ATP from the amino acyl derivatives and inorganic pyrophosphate is more easily reconciled with the formulation shown in Fig. 2. Similar evidence for the structure of t-leucyl adenylate has been presented by De Moss et al. (1). Although the detailed description for the preparation of only a few of the amino acid adenylates is presented here, preliminary experiments with leucine, valine, isoleucine, alanine, glycine, threonine, tyrosine, and arginine have demonstrated that these too are converted to the adenylate derivatives under conditions similar to those described above. With some of these amino acids the reaction proceeded more slowly and the final value reached was only 30 to 60 per cent of the theoretical maximum. Attempts to prepare the histidyl, glutamyl, and aspartyl deriva- tives of adenylie acid have been unsuccessful to date. The rea- sons for this are not clear but in the case of the dicarboxylic acids there is the possibility of internal cyclization to form 5 and 6 membered cyclic anhydrides which might be unstable in aqueous pyridine. With regard to histidine, it has been shown (23) that imidazole catalyzes a rapid breakdown of the acyl adenylate de- rivatives and it is conceivable that the imidazole group of histi- dine might promote the breakdown of a carboxy! activated histidine in the aqueous pyridine system. It does seem possible however, that with further work including the use of suitable - protected derivatives of the amino acids, all of the naturally oc- curring amino acids could be converted to the adeny] derivatives with DCC. SUMMARY The present paper describes the chemical synthesis of the methiony], seryl, phenylalanyl, and tryptophanyl adenylates from the free amino acids and adenosine 5’-phosphate in the presence of dicyclohexylcarbodiimide. These compounds have been obtained in relatively pure form in over-all yields ranging from 20 to 30 per cent by a combination of alcohol precipitation and treatment with Dowex 1 Cl- 10 per cent cross-linked resin. The propertics and analyses of L-methionyl adenylate indicate that the amino acid is linked to the phosphate group of adeno- sine 5’-phosphate in an acyl phosphate linkage. Acknowledgment-—I am deeply grateful to Dr. H. G. Khorana for many valuable suggestions in the use of dicyelohexyl-carbo- diimide and for extending to me the hospitality of his labora- tory at the British Research Council Laboratory, Vancouver, British Columbia. September 1958 P. Berg 611 REFERENCES 1. DeMoss, J. A., Genuru, S. M., anp Novenur, G. D., Proc. 13. Fiske, C. H., anp SupspaRow, Y., J. Biol. Chem., 66, 375 Natl. Acad. Set., 42, 325 (1956). (1925). 2. Wiptanp, T., NremMann, E., anp Pruniperer, G., Angew. 14. Beaven, G. H., Houmwar, E. R., anp Jounson, E. A., In Chem., 68, 305 (1956). E. Caarcarr anp J. N. Davipson (Editors), The nucleic 3. Kuorana, H.G., J. Am. Chem. Soc., 76, 3517 (1954). acids, Vol. 1, Academie Press, Inc., New York, 1955, p. 4. Haun, R. H., anp Kuorana, H. G., J. Am. Chem. Soc., 76, 493. 5056 (1954). 15. Kuorana, H.G., J. Am. Chem. Soc., 76, 3517 (1954). 5. CHamBers, R. W., anp Kworana, H.G., J. Am. Chem. Soc., 16. Kennepy, E. P., J. Biol. Chem. , 222, 185 (1956). 79, 3752 (1957). 17. Hugues, N. A., Kenner, G. W., Toop, A., J. Chem. Soc., 6. Zetzcuz, F., anp Frepricu, A., Ber. dewt. chem. Ges., 72, 3733 (1957). 1477 (1939). 18. Husnnesens, F. M., anp Kiucour, G. L., J. Am. Chem. Soc., 7. Bere, P., Federation Proc., 16, 152 (1957). 77, 6716 (1955). 8. Tarperr, P.'T., ann Huennexens, F.M.,J.Am. Chem. Soc., 19. Cuamprrs, R. W., Morrat, J. G., AND Kuorana, H. G., 78, 4671 (1956). J. Am. Chem. Soc., 19, 4240 (1957). 9. Katcxar, H. M., J. Biol. Chem., 167, 461 (1947). 20. Kaorana, H. G., Tener, G. M., Morrat, J. G., ano Pou, 10. Jones, M. I!., Buacs, 8., Fuynn, R. M., anp Lipmann, F., E. H., Chem. and Ind., 1523 (1956). Biochim. et Biophys. Acta, 12, 141 (1953). 21. Kuorana, H. G., Razzzett, W. E., Ginuam, P. T., TenEr, ll. Mesnaum, W., Z. Physiol. Chem., 258, 117 (1939). G. M., anp Pot, E. H., J. Am. Chem. Soc., 79, 1002 (1957). 12, Dusyorr, J. W., and Borsoox, H., J. Biol. Chem., 176, 789 22. Raacxs, I. D., Biochim. et Biophys. Acta, 27, 416 (1958). (1948). 23. Jencks, W. P., Biochim. et Biophys. Acta, 24, 227 (1957).