On the Coding of Genetic Information M.W. NIRENBERO,~. W. JONES, P. LEDER, B.F.C. CLARK, W. S. SLY, AND S.PESTKA National Institutes of Health, Bethesda, Maryland The process of expressing genetic information by aligning amino acids in proper sequence during protein synthesis usually requires the RNA polym- erase-catalyzed synthesis of a strand of RNA complementary to DNA. Recent experiments re- ported at this Symposium, and elsewhere, suggest that M-RNA (messenger RNA) synthesized in vivo is complementary to only one of the two strands of DNA (Robison and Guild, 1963; Marmur, et al., 1963; Spiegelman, 1963; Wood and Berg, 1963). The M-RNA becomes bound to ribosomes, perhaps forming a polysomal aggregate and the amino acids may be carried to these sites and ordered in correct sequence by specific transfer RNA species. It is possible that M-RNA code words are read from a fixed point by nucleotide sequences in transfer RNA complementary to those in M-RNA code words. Thus, coding errors during protein synthesis may be minimized by the requirement for correct recog- nition at three successive steps; that is at the DNA- M-RNA, M-RNA-transfer RNA (or other inter- mediate), and amino acid-transfer RNA-activating enzymelevels. Little is known about the mechanisms which impart specificity at the last two steps. SOME FACTORS INFLUENCING THE MESSENGER EFFICIENCY OF SYNTHETIC POLYNUCLEOTIDES The effects of base composition, catalytic ability, molecular weight, and secondary structure upon the messenger activity of synthetic polynucleotides will be considered at this time. The messenger activity of a synthetic polynucleo- tide may be related to its molecular weight. In E. coli extracts, poly U containing more than 100 uridylic acid residues per chain has greater template activity than smaller chains (Matthaei et al., 1962), but oligo A fractions containing as few as 9-10 adenylic acid residues per chain have been found recently by Jones et al. (1963) to direct polylysine synthesis. Also, in yeast extracts, oligo U of average chain length 11 actively directs phenylalanine incorporation (Marcus et al., 1963). Although cer- tain oligonucleotides may be degraded by nucleases more rapidly than polynucleotides and thus appear to be less efficient as templates for protein synthesis, RNA chain-length must be considered when com- paring template activities of different RNA fractions. Also, secondary structure of RNA greatly in- fluences its messenger activity. When poly U is mixed with poly A, double- and triple-stranded helices are formed which are completely inactive in directing polyphenylalanine synthesis (Nirenberg and Matthaei, 1961). Oligo A also forms helices with poly U, and the extent of inhibition of polyphenyl- alanine synthesis can be correlated with oligo A chain-length and oligo A-poly U helix stability (Nirenberg et al., 1963). In addition, Singer et al. (1963), have investigated a series of copolymers containing varying amounts of U and G and have found that guanine-rich polymers containing a high degree of ordered secondary structure (perhaps due to G-G interactions) also are inactive as templates for protein synthesis. These results suggest that RNA with a high proportion of helical structure may have little template activity for protein synthesis. Recent experiments have shown that poly U-poly A helices do not bind to ribosomes, and for this reason may be unable to direct protein synthesis (Cukier and Nirenberg, unpublished results). It is possible also that small, localized areas of ordered structure may serve as periods in protein synthesis. It is difficult to compare directly the messenger efficiencies of different polynucleotide preparations because the efficiency is modified by molecular size and secondary structure. However, if the average chain length and secondary structure of different RNA preparations are assumed to be approximately equal, the data of Table 1 suggest that nucleotide content may not influence greatly the overall tem- plate efficiency of M-RNA. Poly U, poly UC, poly ACG, and poly UACG contain 1, 8, 27, and 64 triplets respectively, and the preparations of these polynucleotides shown in Table 1 have been found to direct 1, 4, 9, and 18 amino acids, respectively, into protein. The essential point is that approxi- mately the same total quantity of amino acids were directed into protein by each polynucleotide. Al- though these data must be interpreted with care because the same factor may not limit the incor- poration rate of each amino acid, they suggest that the polynucleotide preparations may have approxi- mately equal template efficiencies and that most nucleotide sequences may be able to code for amino 649 550 NIRENBERG, JONES, LEDER, CLARK, SLY, AND PESTKA TABLE 1. TEMPLATE ACTIVITIES OB 1, 2, 3, AX'D 4 BASE POLYNUCLEOTIDES Polynucleotide Base ratio Moles-per cent Possible triplets Cl"-Amino acids directed into protein U UC ACG UACG 100 .- 56 - :z 32 13 iFi - - 46 5 - - 22 25 PAE PI?& LY62' IL!& PRO LEU ALA MET LYS SER ARG CYSH ALA PRO SER VAL ARG THR GLY THR GLU-NH, TRY GLU-NH, ASP-NH, TYR ASP-NH, HIS PHE HIS PRO LEU SER Total C14-Amino acid incorporation (mpmoles) 3.21 2.91 2.22 5.09 See text for details. Code word nucleotide sequences are arbitrary. acids. Although nonsense sequences may exist, thus far none have been demonstrated definitively. Polynucleotides containing all base combinations now have been used to direct protein synthesis in E. coli extracts. A qualitative summary of these data is presented in Table 2. Only those polynucleo- tides containing the minimum bases necessary to direct an amino a.cid into protein are shown. For example, phenylalanine is directed into protein by poly U and other U containing polymers; however, since other bases are not required, phenylalanine is listed only under poly U. Poly U, poly A, and poly C direct phenylalanine, lysine, and proline, respectively, into protein. Polylysine synthesized in E. coli extracts under the direction of poly A has been found to contain 3-15 lysine residues per chain (Jones, Yaron, Sober, Heppel and Nirenberg, unpublished results). No messenger activity has been demonstrated for poly G (Matthaei et al., 1962), but the highly ordered structure of poly G might mask template activity. However, a poly- nucleotide composed only of hypoxanthine (poly I) with less secondary structure than poly G, still has not been found to direct amino acids into protein. Since hypoxanthine can replace G in RNA code words, the 2-amino group of G does not appear to be essential for coding amino acids (Basilio et al., 1962; Nirenberg and Jones, unpublished data). Each polynucleotide composed of 2 different bases has 8 triplets, but no polymer has been found to direct more than 6 different amino acids into protein.-Poly UC is unique in that it codes for only 4 amino acids, even though all UC triplets appear to function as code words (see Coding Ratio Section). It is important to note also that polynucleotides containing only two different bases direct with great specificity almost all amino acids into protein. These findings undoubtedly reflect basic molecular characteristics of both the recognition process and the general nature of the code. THE CODING RATIO A series of poly AC and poly UC preparations with different proportions of bases were synthesized and their activities in stimulating cell-free amino acid incorporation into protein were determined. As shown in Fig. 1, poly AC directs the incor- poration into protein ofproline, histidine, threonine, TABLE 2. SUMMARY OF CODING DATA Poly Poly U A C G PHE LYS PRO - UA UC UG AC AG CG UAG TYR LEU ILEU ASP-NH, LEU SER LEU HIS ARG VAL ASP-NH, GLU CYSH GLU-NH, TRY THR GLU-NH, ASP * ARC ALA SER THRt MET ASP Only those polynucleotides containing the minimal number of bases necessary to stimulate an amino acid into protein are shown. Amino acids coded by homopolynucleotides are not listed again under randomly-ordered polynucleotides. * Predicted t Reported by Wahba et al. (1963). CODING OF GENETIC INFORMATION 551 0 IO 20 30 40 50 60 70 80 90 TIME [MINUTES) FICWRE 1. The rate of Ci*-amino acid incorporation into protein directed by poly AC (base ratio = A, 47% and C, 53%). Reaction mixture components are described in the legend of Table 3. Incubations were stopped at the times indicated by the addition of 3.0 ml of 10% TCA at 3". The samples were heated at 90" for 20 min, chilled and then filtered through Millipore filters and washed with 5% TCA at 3'. Radio- activity measurements were performed in a thin-window gas flow Nuclear Chicago Corp. counter with a counting efficiency of 23%. Each point represents the ppmoles of CY-amino acid incorporated into protein due to the addition of poly AC. asparagine, glutamine, and lysine at linear rates for 15-20 min. Reactions were terminated after 10 min of incubation, while the rates of incorpora- tion were still linear. In Table 3 is presented an example of the data obtained for each of the five poly AC preparations tested. The theoretical proportions of the four doublet and eight triplet permutations expected in randomly-ordered poly AC (containing by analysis, 47% A and 53% C) are shown in the first and second columns respectively. In the third and fourth columns are shown the ,u,umoles of each C14-amino acid directed into protein by this polymer. A total of 1,685 ,upmoles of amino acids were directed into protein, and the relative propor- tions of each amino acid incorporated, in per cent, are shown in the last column. The 4 doublet permutations do not contain enough specific information to code for the 6 amino acids incorporated, whereas the information con- tent of the 8 triplet words is adequate. The per cent incorporation of lysine, asparagine, glutamine, and histidine agrees well with triplet code word fre- quencies,, but not with doublet frequencies. If all triplets were read, some amino acids would respond to 2 or more code words, for 6 amino acids would then be coded by 8 words. In such cases, the sum of the triplet frequencies would have to be com- pared with the corresponding amino acid incorpora- tion data. For example, if CBA and CCA both coded for one amino acid, the sum of their frequencies is 24.9%, which cannot be distinguished from the frequency of the doublet CA (also 24.9%). There- fore, this experimental approach may allow deter- mination of the coding ratio for some, but not all, amino acids. Analysis of a series of polynucleotides with varying base-ratios permits comparisons to be made with greater accuracy. The expected statistical relationship between code word frequency and poly- nucleotide base-ratio are presented graphically in TABLE 3. COMPARISON BETWEEN AD~INO ACIDS INCORPORATED AND RNA CODE WORD FREQUENCIES Theoretical code word frequency in poly AC containing 47% A and 53% C Doublets Triplets Cia-Amino acids incorporated into protein Frequency of ,apMoles CY-Amino acids Cl*-Amino acid incorporated incorporated per cent AA 22.1 AAA 10.4 AC 24.9 AAC 11.7 CA 24.9 ACA 11.7 cc 28.1 CAA 11.7 Total 100.0 CCA 13.2 ACC 13.2 CAC 13.2 ccc 14.9 Total 100.0 Lysine Asparagine Glutamine Threonine Histidine Proline per cent 183 10.8 192 11.6 157 9.3 444 26.3 159 550 392:: Total 1685 Total 100.0 Each reaction mixture contained the following components in a final volume of 0.25 ml: 0.1 M Tris, pH 7.8; 0.01 M magnesium acetate; 0.05 M KCl; 6 x 10es M mercaptoethanol; 1 x 10-s M ATP; 5 x 10-s M potassium phosphoenol- pyruvate; 5 pg of crystalline phosphoenolpyruvate kinase (Calif. Corp. Biochem. Research); 0.8 x 1O-4 M W-amino acid; (approximately 30,000-150,000 cpm/reaction mixture); 3.2 x 1O-p M each of 19 CY-L amino acids minus the W-amino acid; 15 pg of polynucleotide when specified; and 1.1 mg E. coli preincubated S-30 protein (Nirenberg and Matthaei, 1961). Reaction mixtures were incubated at 37' for 10 min. Protein precipitation, washing, and counting were performed as described by Nirenberg and Matthaei (1961). The theoretical frequencies in per cent of doublets and triplets in polynucleotides were calculated as follows: The fre- quency of the triplet AAA in this poly AC preparation would be 0.47 x 0.47 x 0.47 x 100 = 10.4%. The doublet frequency for CA would be 0.47 x 0.53 x 100 = 24.9%. The ypmoles of each amino acid incorporated in the absence of polynucleotide were: lysine, 30; asparagine, 54; glutamine, 49; threonine, 40; histidine, 26; proline, 36. 552 NIRENBERG, JONES, LEDER, CLARK, SLY, AND PESTKA ------ Theoretical Frequency of RNA Codewords - Observed Frequency of AC Amino Acid hxorporotion Or - I~~~~~~ 1 0 -*IO 20 30 40 50 60 70 60 96 100 PERCENT C IN POLY AC FIGURE 2. Comparison of Cl*-histidine, CY4-threonine, Cl'-asparagine and C14-glutamine incorporation data with the theoretical frequencies of doublet and triplet code words in poly AC preparations. The solid lines represent the experimentally determined incorporation data calcu- lated as described in Table 1. The dotted lines represent the theoretical frequencies of RNA code words. AU assays were performed as described in Table 3. Figs. 2 and 3. Theoretical frequencies in per cent of doublet and triplet code words are shown on the ordinate and the base-ratio is shown on the abscissa. Nucleotide sequence is arbitrary, and each curve represents only one of the three possible sequence permutations. As noted before, the sum of the frequencies of the triplets AAC and ACC equals the frequency of the doublet AC. Thus the AC curve represents either the doublet AC, or the sum of the two triplets AAC plus ACC. Also shown are the observed @*-amino acid incorporation data. Each point represents a different poly AC preparation with the indicated base-ratio. As shown in Fig. 2, the observed incorporation of Cl*-histidine agrees well with the theoretical frequency of the triplet ACC and differs markedly from both the AAC triplet and AC doublet curves. The data also demonstrate that the observed incorporations of both Cl*-asparagine and Cl*-glutamine agree well with the frequencies of AAC triplets. In contrast, the incorporation of Cl*-threonine is similar to the expected frequencies of either the doublet AC, or the two triplets, AAC plus ACC. Therefore, threo- nine appears to be coded either by a doublet or by two triplets, and one cannot differentiate between these alternatives on the basis of these data. In Fig. 3 are presented the template activities of poly AC preparations for Cl*-proline and Cl*-lysine. The experimentally obtained incorporation data indicate that proline is coded either by the doublet CC or by the two triplets CCC and CCA. Cl*-lysine appears to be coded by the triplet AAA. W-AMINO ACID INCORPORATION DIRECTED BY POLYUC The data of Fig. 4 show that proline is directed into protein either by the doublet CC or by the sum 1 CC&CA,, ; /CL z 1\ ,\ 0 " 90 -`$, \\ - -- - - - -- - - g 80- \' Theoretical Frequency Theoretical Frequency \ `1 of RNA Codewords of RNA Codewords w \ \ E 70- \ ' - Observed frequency of - Observed frequency of \ ' p" \ ' Amino Acid lncorporofion Amino Acid lncorporofion - \ \ \ \ ' AA AA 60- \ `\ \ `\ AAAD: AAC `. ., Y 60 t `\ `, -1;; \ `\ AAAt AAC `. ., 50 t AAA>\\ "$ \ 1 1 IO- o- ' ' ' ' 0 10 20 30 40 50 60 70 80 90 100 PERCENT C IN POLY AC FIGURE 3. Comparison of UP-proline and Cl'-lysine incorporation data with the theoretical frequencies of doublet and triplet code words in poly AC preparations. The solid lines represent the experimentally determined incorporation data calculated as described in Table 1. All assays were performed as described in Table 3. CODING OF GENETIC INFORMATION go- 00- 70- 60- 50- 40- 30- 20- IO- ------- Theoretical Frequency of RNA Codewords 70 c `\\`\\ - - Observed Eequency of Observed Eequency of \ \ 60- \ `\ uu Amino Acid Incorporation Amino Acid Incorporation 50- 40- 30 - 20- IO- 0 0 0 IO IO 20 20 30 30 40 40 50 50 60 60 70 70 80 80 90 90 100 100 PERCENT C IN POLY UC PERCENT C IN POLY UC FIGURE 4. Comparison of C"-proline and C!r4-phenyl- alanine incorporation data with the theoretical frequencies of doublet and triplet code words in poly UC preparations. The solid lines represent the experimentally determined incorporation data calculated as described in Table 1. All assays were performed as described in Table 3. of the two triplets CCC and CCU. Cl*-phenylalanine appears to be coded either by the doublet W or by the two triplets UUU and UUC. It is important to note that if the code words corresponding to these amino acids are triplets, both CCC and CCU would code for proline and both UW and UUC would code for phenylalanine. In Fig. 5 are shown the poly UC-directed serine and leucine incorporation data. Both serine and leucine appear to be coded either by the doublet UC, or by the two triplets WC and UCC. Coding of serine or of leucine by one, rather than 2 triplets is not indicated. It is important to note that if serine is coded by triplets, one triplet would have to contain 2 U residues and the other 2 C residues. Triplet words for leucine also would contain either 2 U or 2 C residues. ,,' ,A `\ 0 '.._ \ /r ( , ( , ( , ( b_ 0 IO 20 30 40 50 60 70 60 SO 100 PERCENT C IN POLY UC FIQIJRE 5. Comparison of Cl*-serine and Cl'-leucine incorporation data with the theoretical frequencies of doublet and triplet code words in poly UC preparations. The solid lines represent the experimentally determined incorporation data calculated as described in Table 1. All assays were performed as described in Table 3. These experiments strongly suggest that histidine, asparagine, glutamine, and lysine are coded by triplet words and that the RNA code ca,nnot be composed only of doublets. Threonine, proline, phenylalanine, serine, and leucine were found to be coded either by multiple triplets or by doublets. These data are summarized in Table 4. A mixed doublet-triplet code cannot be excluded on the basis of the available data; however, a uniform code containing only triplets would appear more probable. THE CURRENT CODE WORD DICTIONARY Assuming for the present that all amino acids are coded by triplets, current approximations of RNA code words may be summarized as shown in Table 5. Nucleotide sequence is arbitrary. Fifty of the 64 possible triplets have been assigned. Almost all amino acids can be coded by polynucleotides containing 2 different bases. Since polynucleotides containing 3 bases direct protein synthesis as efficiently as polymers containing only 2 bases, it TABLE 4. SUMMARY OF CODINQ RATIO DATA C14-Amino acid Code Word* Triplet Doublet Histidine ACC - Asparagine CAA - Glutamine AAC - Lysine AAA - Threonine CCA + ACA or AC Proline CCC + CAC + CUC or CC Phenylalanine uuu f ucu or UU Serine cuu f ecu or cu Leucine uuc + ucc or UC * Nucleotide sequences are arbitrary. 554 NIRENBERG, JONES, LEDER. TABLE 5. SUMMARY OF RNA CODE WORDS Amino acid RNA code words* Alanine Arginine Asparagine Aspartic acid Cysteine Glutamic acid Glutamine Glycine Histidine Isoleucine Leucine Lysine Methionine Phenylalanine Proline Serine Threonine Tryptophan Tyrosine Valine CCC CGC ACA GUA UUG GAA AAC UGG ACC UAU UUG AAA UGA uuu ccc ucu CAC GGU AUU UGU UCGt AGA AUA GCAt GAUt AGA AGG ACUT UAA uuc AAU cuu ecu ucc CAA UGAt ACGt 3 CGAt GAAt GACt AGUt CGG ucc UUA CCA CCGt UCGt ACG * Arbitrary nucleotide sequence. t Probable. seems probable that most 3 base words are recog- nized. Tentative assignments are given for such words. It seems clear that most amino acids are coded by multiple words. Furthermore, multiple words corresponding to one amino acid often differ in base composition by only 1 nucleotide. These observa- tions also suggest that nucleotide sequences in multiple words often may be identical. A triplet code may be constructed wherein correct hydrogen bonding between 2 out of 3 nucleotide pairs may, in some cases, suffice for coding; or alternatively, a base at one position in the triplet sometimes may pair optionally and correctly with 2 or more bases. It should be noted that a triplet code of this type in some respects would bear a superficial resem- blance to a doublet code and would be in accord with all of the data available. The coding data obtained thus far clearly indicate that most nucleotide sequences can code for amino acids with great specificity. Weisblum et al. (1962) have reported that multiple species of leucine transfer RNA recognize different code words in synthetic polynucleotides; however, additional data presented at this symposium (not submitted for publication) by Benzer and by von Ehrenstein and Gonano suggest that code word specificity in directing leucine incorporation may be greater with synthetic polynucleotides than with natural M- RNA. It is important to emplhasize the possibility that randomly-ordered synthetic polynucleotides may test the cell's potential to recognize code words, and that the entire potential may not be u-tilized in vivo, except perhups during mutation. Thus M-RNA synthesized by a cell may not contain as many code words as randomly-ordered polynuoleotides. CLARK, SLY, AND PESTKA Several groups of amino acids are shown in Table 6 which either are synthesized in vivo from the same precursor, or have similar structures. For example, phenylalanine, tyrosine, and tryptophan are derived from shikimic acid, and isoleucine, valine, and leucine are synthesized from a-keto butyrate. RNA code words corresponding to these amino acids are shown also. Such comparisons suggest that a family of amino acids may recognize a family of code words whose members contain similar bases. Although not all amino acids fit this pattern, enough additional examples may be cited to warrant the. suggestion that such relationships reflect either the evolu- tionary development of the code, or the recognition of nucleotides in code words by amino acids. The latter has been proposed by Woese (1963) and also is discussed by Weinstein in this volume. OLIGODEOXYTHYMIDYLATE DIRECTED POLYLYSINE SYNTHESIS The chemical synthesis of oligodeoxynucleotides by the method of Khorana and his associates (1961, 1962) and the demonstration of an oligodeoxynu- cleotide-dependent synthesis of polyribonucleotides, catalyzed by RNA polymerase (Furth et al., 1961; Stevens, 1961; Chamberlain and Berg, 1962; Fala- schi et al., 1963), provided an opportunity to study their ability to stimulate cell-free amino acid incor- poration. Since poly A serves as a template for polylysine synthesis (Gardner et al., 1962), oligo dT (oligodeoxythymidylate) has been used to direct poly A, and subsequent polylysine synthesis, as follows: ( 1) ATP OWo dT RNA Polymerase tPoly A + PP (2) Lysine PZ!?? ___..____....____... f po]ylysine E. coli Extracts, etc. In addition, natural DNA and poly U have been shown to direct polylysine synthesis. Poly A was synthesized in RNA polymerase-oligo dT reaction mixtures (stage I) as described in the TABLE 6. RELATIONSHIP BETWEEN AMINO ACIDS OF SIMILAR METABOLIC ORIGIN OR STRUCTURE AND THEIR RNA CODE WORDS Aromatic amino acids Dicarboxylic amino Ileu, Val, Leu acids and amides family PHE UUU ASP-NH, AAU ILEU UUA uuc AAC UAA TYR UUA ASP AUG VAL UUG TRY UGG GLU AUG LEU UUG AAG UUA uuc GLU-NH, AUG ucc AAC CODING OF GENETIC INFORMATION 555 oy ' ' ' ' ' 0 I.0 2.0 3.0 4.0 50 0 20 40 60 so loo 120 rn~1 MOLES (pdT)IN OLIGO dT,s-Ir MINUTES FIGURE 6. Characteristics of C"-poly A synthesis in RNA polymerase (stage I) reaction mixtures. In the figure on the left, reaction mixtures were incubated at 37' for 15 min, then were deproteinized and washed with 5% TCA at 3". The symbols in the figure on the right represent the following: A,, -polymer; , o f1.2 mpmoles of base residues in ohgo dT,,-,,. Each stage I reaction mixture contained the following in a final volume of 0.125 ml: 4 x lo-* M Tris, pH 7.8; 4 x IO-`M MgCl,; 10e3 M MnCl,; 1.2 x 1O-2 M merceptoethanol; 1.6 x 10es M 3.C"-ATP, tetralithium salt (Schwarz BioReseerch, Inc.); and 20 ,ug E. coli RNA polymerase protein (20 units [Chamberlain and Berg, 19621). legend accompanying Fig. 6, and then components supporting amino acid incorporation into protein (stage II) as in the legend of Fig. 7, were added. After further incubation, incorporation of Cl*- lysine into polylysine was determined by precipi- tation with a TCA-tungstate solution (Gardner et al., 1962). The data of Fig. 6 show that CY4-AMP incorpora- tion was dependent upon the addition of oligo dT,,-,, (13-14 nucleotides per chain) to stage I reaction mixtures, and that C14-AMP incorporation was proportional to the amount of oligo dT added within the range of 2.4 or less m,umoles of nucleo- tide residues in oligo dT,,-,,. The average chain length .of the C14-product synthesized in the presence of oligo dT,,-,, was determined by deproteinizing the reaction mixtures, removing the C14-ATP by paper chromatography, hydrolyzing the Ci4-product in 6.3 N KOH and separating the nucleotides by paper chroma- tography. The radioactivity of adenosine, adeno- sine-3' (2')~5'-diphosphate and adenosine-3' (2')- monophosphate was found to be 339, 308, and 22,900 cpm, respectively. Thus, oligo dT,,-,, stimu- lated the synthesis of poly A of average chain- length 60-70 (PA) (adenylate) residues: These data confirm similar results obtained by Furth et al. (1961) and Falaschi et al. (1963). Falaschi et al. (1963) also demonstrated that oligo dT chains are not elongated by the addition of (PA) residues to the free 3'-hydroxyl ends of oligo dT chains, and have obtained evidence which suggests that oligodeoxynucleotides serve as tem- plates rather than primers. Although our RNA polymerase preparations were purified 100 to 150- fold (Chamberlain and Berg, 1962), we have detected unprimed nucleotide incorporation under other conditions, Further enzyme purification will be necessary to determine unequivocally whether oligodeoxynucleotides function only as templates in this system. CHARACTERISTICS OF THE C14-P~~~~~~~~~ SYNTHESIS After incubating stage I reaction mixtures at 37', stage II components were added as described in the legend accompanying Fig. 7. No increase in C14- lysine incorporation was found in the absence of oligo dT, whereas the addition of 1.2 m,umoles of (pdT) residues in oligo dT,s-,, stimulated C14-lysine incorporation at a linear and almost optimal rate for 30 min. In separate reaction mixtures, V4-AMP incorporation into poly A was determined at the end of the stage I incubation. No incorporation was observed in the absence of oligo dT, whereas the addition of 1.2 mpmoles of (pdT) in oligo dT directed the incorporation of 3.8 mpmoles of AMP loo0 I 01 ' ' ' ' ' ' ' ' ' ' a 0 IO 20 30 40 50 60 70 so 90 100 MINUTES FIGURE 7. Characteristics of oligo dTiS-,, directed syn- thesis of Cik-polylysine. The symbols represent the following: * , minus oligo dT; ). , plus 1.2 mpmoles (pdT) residues in oligo dT,,-,,; residues in oligo dTis-i(; 0 plus 2.4 mpmoles (pdT) o , plus 4.8 mpmoles (pdT) resi- dues in oligo dT,,-,,. Components of stage I reaction mixtures are as noted in the legend of Fig. 6. Stage II reaction mixtures contained, in 0.25 ml: 6 x 10-1 M Tris, pH 7.8, 2 x 1O-3 M MgCl,; 1.2 x low2 M magnesium acetate; 5 x lo-' M MnCl,; 1.2 x 10-Z M mercepto- ethanol; 2.8 x 1O-s M ATP; 5 x lo-* M KCl; 5 x 1O-3 M potassium phosphoenolpyruvate; 5 pg crystalline phospho- enolpyruvate kinase (Calif. Corp. Biochem. Research); 2 x lo-' M each of 19 L-amino acids; 2 x 10v4 M C"-L- lysine (Nuclear Chicago Corp.) with specific redioactivity of 4-3 mcuries/mmole; 20 pg RN4 polymerase protein (20 units [Chamberlain and Berg, 19621) and 1.1 mg E. coli extract protein (Nirenberg and Metthaei, 1961). Stage I reaction mixtures were incubated at 37" for 15 min before the addition of stage II components. NIRENBERG, JONES, LEDER, CLARK, SLY, AND PESTKA into poly A. Poly A, in turn, directed the incorpora- tion of 0.64 mpmoles of C14-lysine into polypeptide. The effect of oligo dT upon the individual incor- poration of 18 other C14-amino acids (minus cysteine) also was determined. Only C14-lysine was directed into protein, which demonstrated marked amino acid specificity. Oligo dT was found to be resistant to digestion with pancreatic DNase. DNase partially inhibited C14-lysine incorporation into protein, whereas puro- mycin and pancreatic RNase were strongly in- hibitory. As expected, no inhibition was observed in the presence of actinomycin D. It is important to note that if stage I incubation was not performed, oligo dT directed little or no C14-AMP incorporation into poly A in stage II reaction mixtures, and thus no stimulation of Ci4- lysine incorporation could be observed. RNA polymerase has been found to catalyze the synthesis of complementary RNA in the presence of either RNA or DNA templates (Weiss and Naka- moto, 1961; Weiss, 1963; Krakow and Ochoa, 1963). Poly A synthesized from poly U templates was found to stimulate C14-lysine incorporation into protein. In other experiments no stimulation of V4-lysine incorporation was observed when poly C rather than poly U was added to stage I reaction mixtures. RNA polymerase also has been shown (Chamberlain and Berg, 1962) to catalyze a DNA dependent synthesis of poly A from ATP (in the absence of UTP, GTP, and CTP). Under these con- ditions, poly A synthesized under the direction of calf thymus DNA stimulated C14-lysine incor- poration. Paper chromatography of polylysine, as described in the legend accompanying Fig. 8, permits separa- tion of lysine peptides of different chain lengths. Peptides containing approximately eleven or more lysine residues remain at the origin, whereas the mobilities of smaller peptides are as follows: lysine > di- > tri- > tetra- > penta- > hexa- > hepta- > octa- > nona- > deca-lysine. As shown in Fig. 8, most of the C14-product synthesized in the presence of oligo dTis-i4 remained at the origin after chromatography. Digestion with trypsin con- verted the C14-product almost quantitatively to peptides which migrated with free, di-, tri-, and tetra-lysine. In separate experiments the C14-product was eluted from the origin. An aliquot was hydrolyzed completely with HCl, and another, with trypsin. Both aliquots were chromatographed as before. One Ci4-spot having the characteristic mobility of free lysine was found following acid hydrolysis. After digestion with trypsin, CY4-products with the expected mobilities of free, di-, tri-, and tetra-lysine were found. In addition, the C14-polylysine which CHROMATOGRAPHIC POSITION OF LYSINE PEPTIDES P&V- LOICnl LISlNl OLICOPIPlIOES PEWIL- 1~If.k TRI- Ol- LKM 2400 - 2200 - 2000 - IS00 - 1600 - r `: ;' 1400 - 3 ,' 1200 - I f =, 1000 - s $ 600 - 600 - 400 - 200 - t o,J(J ] , , , , , ] , , , , , , 0 I 2 4 6 6 IO I2 14 16 INCHES FIGURE 8. Chromatographic analysis of the C"-polylysine synthesized in stage I plus stage II reaction mixtures under the direction of oligo dT,,-,, before and after tryptic digestion. The symbols represent the following: 0, minus oligo dT; 0, plus oligo dTiJ-il; A, plus oligo dT,,-1, followed by tryptic digestion. Reactions were carried out as described in the legends of Figs. 6 and 7. After incubating stage II reaction mixtures at 37" for 30 min, the reactions were terminated by the addition of 2 ml of cold 10% TC.4. The supernatant solution was extracted three times with equal volumes of ether, con- centrated in vacua, and the residue was compared chro- matographically with a partial tryptic digest of chemically synthesized Cl*-polylysine (YEDA Research and Develop- ment Co., Ltd., Rehovoth, Israel) on Whatman 3MM paper in a solvent similar to that described by Waley and Watson (1953) containing pyridineln-butyl alcohol/acetic acid/water = 6/9/3/7 v/v for 56 hr. remained at the origin was shown to contain carboxyl-terminal radioactivity by a hydrazinolysis method (Akabori et al., 1952). EFFECT OF MOLECULAR WEIGHT UPON THE ACTIVITY OF OLIGO dT Falaschi et al. (1963) demonstrated that oligo dT chains containing less than 4 residues did not stimulate poly A synthesis, but that longer chains were stimulatory. The activity was proportional to chain length until maximum stimulation was CODING OF GENETIC INFORMATION 557 TABLE 7. RELATION BETWEEN 01.1~0 dT CHAIN LENGTH .*ND ACTIVITY C14.AMP U4.Lysine incorporation incorporation Addition (mpmoles) (mpnoles) None 0.1 0.075 17 mpmoles oligo dT,-, 1.5 0.237 17 mpmoles oligo dT,-, 6.8 0.387 17 mpmoles oligo dTd_,, 13.0 0.406 The components of the reaction mixtures and conditions are described in the legends accompanying Figs. 6 and 7. reached with oligo dT fractions containing 14 or more (pdT) residues per chain. The data of Table 7 are in accord with their findings. Oligo dT fractions containing as few as 6-7 base residues per chain were of sufficient length to stimulate the synthesis of poly A with messenger activity for C!14-lysine incorporation. Almost optimal stimulation was obtained with oligo dT fractions containing 7-8 residues per chain. These and the previous experiments demonstrate that small chemically synthesized oligo dT fractions stimulate the synthesis of longer chains of poly A, which in turn direct the synthesis of polylysine containing 11 or more residues per chain. It is possible that defined oligodeoxynucleotides may be useful in the determination of nucleotide sequence and polarity of RNA code words, and also in the study of control mechanisms related to DNA- directed protein synthesis. REFERENCES AX~BORI, S., K. OHNO, and K. NARITA. 1952. On the hydrazinolysis of proteins and peptides. Bull. Chem. Sot. Japan, 25: 214-218. BASILIO, C., A. J. WAHBA, P. LENCYEL, J. F. SPEYER, and S. OCHOA. 1962. Synthetic polynucleotides and the amino acid code, V. Proc. Natl. Acad. Sci., 48: 613-616. CHAYBERLAIN, M. and P. BERG. 1962. DNA-directed synthesis of RNA by an enzyme from E. coli. Proc. Natl. Acad. Sci., 48: 81-94. FALASCHI, A., J. ADLER, and H. G. KHORANA. 1963. Chemically synthesized deoxypolynucleotides as tem- plates for RNA polymerase. J. Biol. Chem., 2.38: 3080- 3085. FURTH, J. J., J. HURWITZ. and M. GOLDMANN. 1961. The directing role of DNA in RNA synthesis. Biochem. Biophys. Res. Commun., 4: 362-367. GARDNER, R. S., A. J. WAHBA, C. BASILIO, R. S. MILLER, P. LENGYEL, and J. F. SPEYER. 1962. Synthetic polynucleotides and the amino acid code, VII. Proc. Natl. Acad. Sci., 48: 2087-2094. JONES, 0. W., E. E. TOWNSEND, H. A. SOBER, and L. A. HEPPEL. 1963. Effect of chain length on the template activity of polyribonucleotides. Biochem. (in press). KHOR~NA, H. G. and J. P. VIZSOLYI. 1961. Studies on polynucleotides. VIII. J. Am. Chem. Sot., 83: 675-685. KHORANA, H. G., J. P. VIZSOLYI, and R. K. RALPH. 1962. Studiesonpolynucleotides. XII. J. Am. Chem. Sot., 84: 414-418. KRAKOW, J. S. and S. OCHOA. 1963. RNA polymerase of Azotobacter &nelandil, I. Proc. Natl. Acad. Sci., 49: 88-94. MARCUS, L., R. K. BRETTHAUER, R. N. BOCK, and H. 0. H~LVORSON. 1963. The effect of poly U size on the incorporation of phenylalanine in the cell-free yeast system. Proc. Natl. Acad. Sci. (in press). MARMUR, J., D. M. GREENSPAN, E. PALECEK, F. M. KAHAN, J. LEVINE, and M. MAKDEL. 1963. Specificity of the Complementary RNA formed by B. suhlilis infected with Bacteriophage SP8. Cold Spring Harbor Symp. @ant. Biol. Vol. 28. MATTHAEI, J. H., 0. W. JONES, R. G. MARTIN, and M. W. NIRENBERG. 1962. Characteristics and compo- sition of RNA coding units. Proc. Natl. Acad. Sci., 48: 666-657. NIRENBERG. M. W. and J. H. M.&TTHIEI. l961. The dependence of cell-free protein synthesis in E. coli upon naturally occurring or synthetic polyribonucleo- tides. Proc. Natl. Acad. Sci., 47: 1588-1602. NIRENBERG, M. W., J. H. MATTHAEI. 0. W. JONES, R. G. MARTIN, and S. H. BARONDES. 1963. Approxi- mation of the genetic code via cell-free protein syw thesis directed by template RNA. Fed. Proc. 22: 55-61. ROBISON, M. and W. R. GUILD. 1963. Evidence for message reading from a unique strand of DNA. Fed. Proc., 22: 643. SINGER, M. F., 0. W. JONES, and M. W. NIRENBERG. 1963. The effect of secondary structure on the template activity of polyribonucleotides. Proc. Natl. Acad. Sci., 49: 392-399. SPIEGELMAN, S. and M. HAYASHI. 1963. The present status of the transfer of genetic information and its control. Cold Spring Harbor Symp. Quant. Biol. Vol. 28. STEVENS, A. 1961. Net formation of polyribonucleotides with base compositions analogous to DNA. J. Biol. Chem., 236: PC43-PC45. WALEY, S. G. and J. WATSON. 1953. The action of trypsin on polylyeine. Biochem. J., 55: 328-337. WAHBA, A. J., R. S. GARDNER, C. BASILIO. R. S. MILLER, J. F. SPEYER, and P. LEKGYEL. 1963. Synthetic polynucleotides and the amino acid code, VIII. Proc. Nat]. Aced. Sci., 49: 116-122. WEISS, S. B. and T. NAICAMOTO. 1961. On the participa- tion of DNA in RNA biosynthesis. Proc. Natl. Acad. Sci., 47: 694-697. WEISS, S. B. 1963. Properties of DNA-dependent syn- thesized RNA. pp. 61-66. In: Informational Macro- molecules. [eds.] H. J. Vogel, V. Bryson, J. 0. Lampen, Academic Press, New York. WEISBLUM, B., S. BENZER, and R. W. HOLLEY. 1962. A physical basis for degeneracy in the amino acid code. Proc. Natl. Acad. Sci., 48: 1449-1454. WOESE, C. 1963. The Genetic Code. I.C.S.U. Reviews (in press). WOOD, W. B. and P. BERG. 1963. Studies on the "messenger" activity of RNA synthesized with RNA polymerase. Cold Spring Harbor Symp. Quant. Biol. Vol. 28.