BIOCHEMISTRY: .C-IREAVBERG ET AL. 1161 RN.4 CODE WORDS AND PROTEIN SYNTHESIS, V-II. ON THE GENERAL NATURE OF THE RNA CODE BY M. KIRENBERG, P. LEDER, M. BERNFIELD, R. BRIMACOMBE, -.J. TRUPLX,* F. ROTIXAN~, AND C. V&AL NAI'IONAL HEART INSTITCTE, ?JITION.IL INSTITUTES OF EIEALTH, BETHESDA, YARYLAXD Commurkukd by Robert J. Hwbner, March 86, 1.965 Xucleotide sequences of RNA codons have been investigated recently by directing the binding of C"-.k&R,\T\ to ribosomes with trinucleotides of defined base sequence. The template activities of 19 trinucleotides$ have been described and nucleotide sequences have been suggested for RXA codons corresponding to 10 amino acids.L-6 In this report, the template activities of 26 additional trinu- cleotides are described and are related to the general nature of the RX-1 code. Matehis and Ydhod.s.4omponen~ of rsuctti: E. wii W3100 ribosomes and JRNA were prepared by modifications of methods described previously.~* Each P-aminoacyl~RNA was prepared in the presence of 19 CWamino acids. The assay for ribosomal bound P4.4-sRN.4 and components of reaction mixtures have been described.' The characteristics and amounts of labeled .4A+RXA not deacribed~ are shown in Table 1. Syn~~and charactiatia of oh.ipnuchot&s: hpC and UpC were. obtained from a T-l ribonuclease digest of RNA, and .Iph was prepared by chemical synthesis.% 10 ApC!, Ipc, CpA, CpG, GpA, and GpC were obtained from Gallard dchlessinger Corp., but required extensive puri- fication prior to use. GpGpLY was obtained by digesting poly UC with pancreatic RNaw A: treatment with alkaline phosphatare to remove terminal phosphate groups from the degradation products: and isolation by procedures similar to those dexribed for CpTpC.' The remaining trinucleotides were synthesized from the appropriate dinucleoside monophosphate, using either TABLE 1 RADIOACTIVE .~INOACYL-sRS.1 PREP.U~R.ITIONS* 0. 75 0 .36 0.85 0.15 0.47 0.70 0.57 1.03 0.42 1.00 3. -5. 10 2 9 3 3`2 .6 14.9 6.1 11.72 9.2 6.4 13.2 6 3 LR.6 6.5 o Other AA-sRN.1 zleta have been deacribed.~ t Ammo acids stated were labeled mth Cl& with the exception of HQryptophsn. and HQlycine. : E. coli ??o?? IS a Kl2 strain. .Iminosc I sRNA 3ynthetase preparations were from E. cdi ?????? ( We thank Dr. Charles Yenofaky for thu i$ coli swain and Dr. Ray Byrne for the HrGly-aRNA. A-kaRNA sod 100.000 X g supematant fractions were used for ths prspnrstion of II'-Cly-sRNA. 1162 BBXHEMISTRY: NIRBNBERG ET AL. PROC. N. A. S. TABLE 2 CE4BACTERIZ AlTOL OF -hU~UC!I.EO'ITDES Digestion (T-2 ribonuclea.~) Pmducte Base ratio 1.00~0.95/1.05 1 .o5j1.oo/O.95 1.00/0.95i'1.05 1.00/1.00/0.85 I 00/l .00/0.95 1.05'1.00~0.95 1.00;1.10/0.95 1.10/1.00j1.00 1.00/l .10/o 90 2.00/O 95 1 10.11.00~0 95 1.05/l .00/o. 80 0.95/l oo;`l. JO 0.90/l .00/l .I0 1.05/l OO/O.QO 1 .00/l .00/O 95 1.05/l 00/0.90 2.00/0.95 Loo/l 000 10 l.W~~.Q~jl.&J 1.00/0.90/1 .oo 0.95/1.09/1.10 1.00/0.95/1.05 A,PC,PA 2&G A:pC:pUc A,pG,pA A,PG,P~ A,pG,pV A,pU,pGd CPA p$Jg C:pA:pC C&,pX C,PG,PC* C,P~. ,pG G,pA49d g'$$d G:pC:pT l.a5/0.90/1 .oO 1.15/2.00 1.00/0.95/1 .Ml 1.00/0.90/1.10 I 00,`l OG/l 00 1.00/l 00~1.15 l.OO/1.oo'l.Ml 1.00/O. Q(l/l JO 1.05:2.W 1.00'1 .(KJ/l.OO 1.00/1.00/1.15 1.00/1.05!0.95 1 O$kl k,.;l.Oo 0.90/l OO/l.OO 0.95/l 00/l .I0 1.10/2.00 1 05/l .00/o. 90 1.2011 W/O !NI 0.85/1.0(1/1.00 0.85/l .00/l .05 0 95/2.00 1.00/1.15/0.QO 0.95/1.00/1.05 0.90/l 00/l. 15 1.10/1'.00/0.90 primer-requiring polynucleotide phoephorylase and s nucleoside 5'-pyropho~phate,~~ or a deriva- tive of bovine pancreatic rihonueleeee with 8 nucleoside 2',3'-cyche pho8pb.~&&~ (Table 2). The products were isolated by paper chromatography and electrophoresis, as previously described*-6 1% and the purity of each preparation wan assessed by two-dimensional chromatography of an aiiquot (2.0 A' units) on Whatman no. 40 paper. The first dimension (solvent -4) was n-pro- panel/ammonia/water, 55/10/S; the second dimension (solvent B) a-as 0.10 M sodium phos- phate, pEI 7.0, containing ammonium sulfate (0.4 gm/ml). Chain-length and base composition (Table 2) wee d&erm&dbyd&stkm~A"@tit~~f each trinucleotide with T-2 ribonuclease and 2.5 Am units with venom pha&odieste=, as previom+ described.~ 1% Rem?k and DisczcssioR.--In Table 3 are shown the &ects of 26 trinucleotides upon the binding to E. CO& ribosomes of 19 CY-AA-sRXA preparat,ions, each acylated with a different C1kmino acid (Cl'-Cys-sRXA not used). In addition, near the bottom of the table are shown the effects of 18 trinucleotides previously describedl+ upon the corresponding Cl'-k4aRKA (Cl'-Cys-sRXA and UpGpI? omitted). Many of these trinucleotides have not been isolated or synthesized previous!y. Several factors should be mentioned which may be useful in assessing the data. (a) It is often difficult to compare direci~y the response of one Cl'-A.4-sR?iA preparation to a template with that of another, for Kaji and Kajila have shown that VOL.53, 1965 BIOC'EEA4lSTRY: NIRE~VBERQ BT AL. 1103 both deacylated and acylated sRNA bind to ribosomes in response to polynucleotide templates. The extent of acylation of each C%4A-sRNA preparation must be considered (see Methods and Materials) as well aa the relative response of each Cl-M-sRNA to other trim&&ides. (b) A trinucleotide which stimulates the binding to riboeomes of one Cl'-A.4-sRNA gene&y decreases binding of other W-U-sRNA preparations. l (c) Background binding of C"-k4-sRNA to ribosomes appears to be a function of the sRX.4 species, the amount of sRNA added to a reaction, the proportion of sRNA acyh&d with a C19mino acid, and possibly tWsasp-uht;-of tempt&e RNA on the riboeumea or in the sRNA prepara- tions.`, `* * (d) Reactions contained limiting concentrations of ribosomes (as determined with ApApA, UpUpl;`, UpUpC, or GpUpU) and therefore were satu- rated with respect to these trinucleotides and CWAA-sRNB. -Most trinucleotides markedly stimulated the binding to ribosomes of only ooe O-A&RNA preparation ; however, a number of trinucleotides displayed lower template specificity for Cl'-AA-sRX.4. For example, ApCpU, SpCpC, and hpCpG stimulated CWThr-sRNA binding to ribosomes, but did not significantly stimulate the binding of 18 other Cl-AA-sRNA preparations. ApCpA also stimulated CWThr-sRNA binding. This trinucleotide also stimulated CP-Lys-sRNA binding. However, the template activity of ApCpA for Cl-Lys-sRXj-4 was only 10 per cent that of ApApA. The disparity between the temp&e activity of ApCpA and ApApA was even more apparent in reactions containing limiting concentrations of trinucleotidea (data not shown). Such considerations suggest that the sequences ApCpG, ApCpU, ApCpC, and ApCpA correspond to threonine Lwdons. It is possible that the template specificity of one synonym codon may differ from that of another; however, other alternatives, such as the possibiity that Cl'-Lys-sRSA may respond lo an impurity in the ApCp.4 preparation which we have been unable to detect, also must be considered. The data of Table 3 indicate that the sequence GpCpU corresponds to an R,\ia codon for alanine; CpCpA, CpCpU, and CpCpC correspond to proline (the tem- plate activity of CpCpA for Cl'-Pro-sR,\I'A was higher than that of pCpCpC (,cf. ref. 4) ; VpC'pG, CpCpU, and VpCpC correspond to serine (cf. ref. 4) : Gp-4pU and GpApC, U) aspartic acid; GpApA, to glutamic acid; Cp-IpE and CpApC, to histidine: CpApA and CpApG, to glutamine; CpGpC and CpGpA. to arginine; .\pUpG, to methionine; and CpCpG and UpVpG, to leucine (CpUpl: and CpUpC possibly serve as internal but qt. terminal Leu-codons'). It seems clear that GpCpI%erves aa a eodon for alanine, for this trinucleotide st,imulated only the binding of C"-alanine sRS.4 to ribosomes. This sequence is also in accord with predictions based upon amino acid replacement data. How- ever, the weaker response of C"--Ma-aRS-4 to .4pGpC, CpGpC, and UpGpC sug- gests that recognition of 2 out of 3 bases, the GpC portion only of the lat,ter tri- nucleotides, may permit C"-hla-sRSX binding. Similarly, Cl'-Glu-sRN,4 IF- spends best to GpApA, but also responds to a weaker extent to trinucleotides con- taining Gp.4, such as -4pGpA, CpGpA, and VpGpA. C"-Lys-YRXA responds best to ApApA but also recognizes .4pApG.S GpApA, ApCpA, CpApA, YpApA, and CpCpA. ,4dditional examples in Table 3 are readily apparent. These dat.a in- dicate that one trinucleotide sometimes LZM direct the attachment of a limited group of Cl'-AA-sRN,4 species to riboaomes. It is possibie that correct I+ 1164 BlOiXRkfISTR~`: XIREh-BERG ET AL. F'ROC. N. A. S. TABLE TEMPLATE SPECIFICITY OF TBIMJCLEOTIDES CM- A ~&&le@ of W- or H~AmincmcyLz$>A Bound to c1c "c:&Hr C'C .4apKH, Glu Gi"-NH* His Ilea -0 03 -0 02 -0 03 0.05 -0.22 -; ;; r 3:;" -0 04 -0 0 03 03 -0 05 G 0 07 0.01 0 0 01 0 33+ 0 19t (I -0.13 0.04 -0.23 -0.n" -0 34 z;-:s -0 13D 0.04 -0.4' 0 056 ti 0 62 %.02 -0 02 r$ :; 2 14 2.05 2.60 -0.02 0 -0 01 03 g.02 x 03 0 i 01 0 02 - 0 02 02 -0 (1 0 01 01 -0 0 04 01 0.01 -0 0 01 02 -0.13b -0.236 -0 a* -0 OS" -o.mb -0.12 -0 84" -0.01 23 -0.31 3.04 -0.02 -0.68 0.09 -0.55 -0.30 4X" C'C C'C Ala Am -0 16 -0.04 -0 15 -0 15 -0 01 -0 01 -0 15 -0.37 Tnouclaotidc g;: ~~ GpCpC CPCPA UpCpG GpApF zigi 0.71 -0.18 -0.15 0.06 -0.20 0 -0 01* -0.14 1.29 -0 o5* 07" -0.22 1.32 -0 -0 11 0 01 -; 8$ -0.31 -0 25 0 02" 0.01 -0 0: -G 01 -0.06 0 02 -0 03 0 04 -G 07 -0 03 0 52 0 26 -"o 0" G 02 -0 oi -0.03 G -0.08 -0.01 g.02 0.03 -0 02 0.03 0 :;:g -0.03 -0 01 -0 13 -o.oi -0 09 -0.41 -0 07 006 -0 076 0 t3h 0 01 -0 05 -0 06 0 10 -0 oz* -0.33 0 14 1 63 -0.20 1 42 0 28 -0 16 -0 12 -0.12 -0.01 0 01 -0 07 0 12 XE 0.02 0 04 -0 30 0 hl,Gpr APGPC ApGp.4 GpGpU CpGpC CPGPA gg: ApUpG CpCpG E.03 0.01 0 01 0.02 -0 01 0.05 0 07 -0" 02 0 19 0 04 -0 03 i -0 21 -0.07 -0 05 0 -0 02 -0.13 -0.01 -0 01 0.07 0.04 0.10 -0.03 -0 02 -0.03 0 zg ",: -0 02 c 0.02 0 12 0 14 -0.05 -0 13 -G 011 -0.11 -0.14 -0.11 -0 06 -0.01 0 21 . 0.02 0.04 0 21 -0 01 0.01 -0.14 0.11 1.65 1 16 00% 0.12 0.14" 2.a l.ld 1 19 ApApC 1.50 ADAPC . . The specificity of trinnaleotides in dim&rag the binding of C'C or H+minoacyl-#RNA to niammcs. ducibk st.im$+mn of M,-nRN 4 +d+ doe to the ddition of trinucleotidea am bold fact. Repro tamp+ activd~a of 18 tnnu&otadwe.e For compuimn. the mntauled tbtciompontaltrd ~outly deaeribed' -I are nhoarr it tht bottom of the table. aaCtions Y- qmf Aidho-+. the~mmunt of C'c-AAaRNA hated pryi- o$II~~ or& Table 1, and 0.150 A= units of trinuclwtide, aa spec&d. m . final volume of 50 ,,I. C"-Amph Hr cognition of 2 out of 3 bases in a trinucleofde, in or out of phase, or 2 bases in 1 trinucleotide and 1 base in an adjacent trinueleotide, often may su&e during protein synthesis. This striking phenomenon is often observed with trim&&ides containing 2 OT 3 purines. Since the stability of codon-ribosome-k4+RXA complexes may partially depend upon interactions between bases in codons and sRPiA, the afIinit,y of sRXA for a ribosome may be greater when each base in a codon is recognized correctly and in proper phase than when codon recognition is only partially correct. The a6vit.y of k& ApGpU and ApGpC in stimulat,ing binding of V-Ser~&-s~q -. . . . .--.__--. may indicate that these sequencers correspond to serine codons (in addition to UPCPU, UPCPC, and UpCpG). However, these assignments should be considered tentative. Although ApGpU and ApGpC have been proposed as asparagine eodon sequences,`" the data of Table 3 show that they do not aignifieantly afkct the binding of C'4-AsphZ&-sRXA under the conditions employed. We have previously reported that ApApU and ApApC stimulate binding of Asp- NH,sRI\`A with high specificjty.6 .4pGpA slightly stimulated the binding of both Cl'-Arg- and C14-Glu-sR?;k The sequence ApGpA was predicted for arginine on the basis of wdon sequence data obtained earlier and amino acid replacement data reported by Yanofsky and VOL. 53, 1965 BIOCHEMI,f3TRI-: .VIRE,VBERG ET .IL. 3 FOR Cl'- OR H3-h~~~~~~~~+EtRNA Rib;~~mee Du;;: .Idditio;,zf Trinucleotides* (TIC IL" I;:$ -0.09 -0.22 -0.2s o.o!a 0 1:: :x -0.03 -0.03 0 01 2% -0 05 -0.09 -0.18 -0.11 -0.09 -0.27 -8 E -0.07 -0.04 -0.38 0.30 0.79 . . 0.23 UPURJR ""6": CPUPC Lys 0.01 -0 03 IY' -0.10 0 11 0.04 -0. la -0.10 0 73 -0 07 -0 I3 0 10 - 0 03 a 10 -0 05 -0 IS -0 12 -0.01 -0.10 -0 16 -0.11 -8.11 -0 08 0.03 0.77 AP;.P; ApApG i&t 0 03 -b4 0.02 -0.18 0.03 -0.11 -0.09 -0. IO 0 1t.g -0 12 -0 02 -0.09 -0.02 18.:: -O.Oi -0.08 0 01 -0.17 -Y8 1.00 0.10 0.40 Pbe 0.02 ii.01 -0.07 -0.04 0.03 003 -X-f: -0.03 3. :: -~:lz -0 38 -0.17 -0.22 -8.E -0.34 -0.03 -0.04 -I: % -00.2 0.44 . . . . C'C Pro -0 02 -0.01 0.0.9 -0.01 -0.02 0.40 0.086 1x.g -0 076 -0.01 0.01 -0.01 -0.01 0.02 0 -0 04 -0.01 0.01 -0.07 -0.056 -0.12 0.09 0 00.03 Eb . . . l-. as 0.28 ""`I",7 PCPCPC 0.06 OPUPC CPCPC 0.15 CPCPrJ C'C ser -0 05 -ii E -0.15 -0 18 -0.08 1.09 0 01" o.oib -0.23 0 09 -0.03" -0 026 -0.09 0 02 0 : 3 0 03 -0.09 -0.12 -0 22 0.02 0 -0.0.5* 0.03 0.38 0.24* 1.27 UPC& UPCPC . C'C Thr 0 63 ts 0.75 0.08 -0 02 -0 04 -8.0": -0 08 -0.07 0.03 -0 04 -0.09 -0 03 -0.08 -0 02 -0.05 -0.11 -x.08 -d..OEl 0 01 -0.08 0.30 . . . T:pt -0.01 -8.E -0.03 0 0.03 0 02 0.04 0 02 -0 04 -0 01 -0 03 -0" 03 LO3 -0 04 -0.05 -0.05 -0.04 -8 O2 fEi -0 06 -0.03 0.30 . 1 it% C'L TY~ 0.03 0.03 0 01 0.03 0 oa 0.03 0 05 0 03 0 0.03 -a 2 0 01 0 03 -0 a3 0 01 0 02 0 a3 g.03 2-i 0 03 0.04 0.12 . . 0.111 ~PAPU 0.56 UPAPC . C'C Vd : ito1 0 0 0.03 g.03 -0 a4 0.02 0 03 -0 01 0 02 0.03 0 0.08 0.02 o.oa 0.03 -0.01 -0.02 0 02 -0.06 -0.10 0.05 0 22 l.s4 GPUPU assayed in IOO-rl reactiona; smountn of all components were doubled. o Background binding of C*Qmumaoyl-sRN.i M ribosomes in the absence of srinucleotidea is expressed in rum&a (ahown ~lesr the bottom of the table). All other values !A ,u+moles, were obtained by subtracting back- ground binding of C~~eminoacyl-aRNA From binding sabtsined upos addition of ti trinucleotide preparation. t sRNA may coot&n jome C'cAsp-aRN.1. co-workers.14 The recognition of one triplet by sRSd corresponding to several amino acids again suggests partial codon recognit,ion. Such observations should be considered in terms of in viva studies, particularly those related to extragenic suppression. Possible Base Seqjhen,ces of Nonsense Colons.--Since nonsense codons may per- form special functions in protein synthesis, we have been particularly interested in a small group of trinucieotides, UpAp;l, UpA\pG, CpGpAI, CpUpU, CpUpC, and .\pGpA. which either have little tqpplate activity. or have slight activity for two or more Cl"-.A-\-sRX-1. The possi&lity that CpUpU or Cpc'pC may serve as in- ternal, but not as terminal, codons for leucine has been discussed previously.4 The striking results of Sarabhai, Stretton, Brenner, and BoileL6 indicate that certain codons in "amber" mubants of T4 phage may correspond to an ammo acid in certain strains of E. coEi but may specify the terminus of a protein in other E. coli strains. Further analysis of mutant phages and amino acid replacement data have led Brenner and co-workers to suggest that CpApG or UphpA may 3pecify the end of a polypeptide chain in some E. coli strains and serine in an additional strain which contains a suppressor gene.17 Weigert and Garen also have found that mutations which lead to the formation of nonsense codons in the alkaline phos- phatase gene can be related to amino acid substitutions at sites corresponding to 1166 BIOCHEMZSTR F': ZCZREh-BERG ET .4 L. PROC. h'. A. s. nonsense codons only if the base composition of the nonsense codon is (V.4G) .Ir This codon also corresponds to serine in a st,rain containing an appropriate suppres- sor gene. We find that, the sequences UpApG and Up-4p,4 have almost no template activiry for Cl'-amino acids under the conditions employed (only very small st,imulation of Cl4-Asp-KHr and 0'-Lys-sRXA binding w&s observedj. These data are in full accord with the conclusions of Brenner and of Garen and their co-workers. In addition the sequences found for glutamine codons were CpApG and CpAp_4: the sequences found for serine codons were UpCpG. UpCpV, I:pCpC, ApGpT, and ApGpC (UpCpA also predicted 1. These seque%yes demonstrate st,ruet,ural rela- tionships between Terminator-, Glu-?;H T, and Ser-codons and suggest mechanisms for alternate codon rerognit,ion in different strains of E. c&i. Tlw GenercJ Nature of the Code.--Thus far, the template fun&one of 45 of i.he 64 trinucleotide sequences have been invest,igat,ed in this syst.em. -4 summary of the data and additional codon sequences which can be predict*ed from amino acid r(a- placement data reported for E. c&l4 and TMP mut&nts,lg. 2o are shown in Table i. Almost all of our earlier predictions were confirmed when the appropriate tri- nucleotide was t,ested 1-5 L . ?ierertheless. t,he summary shown in Table 4 should IIOI be thought of as an invariant codon dietionar>-, since it is clear that codon recogni- lion can be modified. Previous studies with randomly ordered pol.ynucleotides and cell-free protein synthesizing systems showed that synonym codons often differ in composition by only one bax2*~ ** This suggested that bases common to synonym codons occup? identical positions and, that either 2 out of 3 bases in a triplet sometim= may be recognized, or a base may be recognized correctly in 2 or more way~.*~, t4 On the o ?? ir possible that these asquencam we readable ioternrl-, but. nonrerdablr terming&. oodons. t UpA or rds. 1 5 A and UpApC may oorreapond to Terminator-. or Ber-codonc in different atrains of E. cc& (see text and 18). Svrn~~y and predictionr: The template activities of trinualeotidss in BOLDFACE have been atudled experi- men~y in this system. other ssguen- we pmdieted. Akhough trinucleotides src rrran$ed in pairs, one member of L pair mwy ???? greater tnmphtc wtivity than the other. .m not indi~hted. Estm~te of r&t& utmpletc e501enera Amino mad "P ment dam uvrd for theoe by RN02 in T k by Wattman md Rittman- redictions were obtained with E. coli by Y an&k-y." or we-r induced fitbold'. or Tsugita.~ VOL. aS.3, 196.5 BIOCREMZSTRY: AVIRENBERG ET AL. 1167 basis of such data Woese suggested a code in which A, C, G and U are independent,iy recognized at one position in the triplet, C = CT at a second position, and ;i = C and G = T-J at a third position.26 A modification was suggested by Eck in which I: = C and ,I = G at an unspecified position in a txiplet.~ We have previously shown that each member of a trinucleotide pair with 3'-terminal py-rimidines, such aa XpYpU and XpYpC, corresponds to the same amino acid,* 5 and each member of a codon pair with 3'-terminal purines, ApApA and ApApG, corresponds to lysine." Several generalizations can b made concerning the nature of the code. (u) Amino acids, which are structurally or metabolically related (such as synthesized in utio from a common precursor) often have similar RNA codons. Such relation- ships n-ould appear to reflect either the evolution of the code?4 or direct interactions between amino acids and bases in codons.n* s (6) Many codons may be recognized partially! or may contain alternate acceptable bayes at certain positions. (c) Recognition of the 3'-terminal base in a trinucleotide is most variable and fits several generalpatterns;C'=C;G=A;orG=A=U=C. (d)InthecaseofLeu- codons, U = C at the .5'-terminal position. (e) In most cases the apparent tem- plate activity of one member of a synon"ym codon set differs from that of another. These patterns apparently define the characteristics of several general recognition mechanisms. Although the molecular mechanisms which permit one codon to be distinguished from another are unknown, the pairing of bases in mRXA with bases in sRNA is an obvious one to consider. Enzymic modification of bases in certain doublet or triplet sequences in sRNA might affect codon recognition, as proposed by Ames and Hartman,= and also explain patterns of synonym codons sets. Inter- conversion of C and C (C * v) in an sRXA "anticodon" might result in a G = A pattern in mRSA codons and interconversion of .I and I (A e I) in a G = C pat- tern in mRSh codons. However, the possibility of altemat,e acceptable base pairing is raised by observations which indicate that one molecule of Phe-sRN,1 may recognize both VpUpU and UPUJ'PC.~ The nucleotide sequence found for an alanine codon (GpCpU found, GpCpC. GpCpX, and GpCpG predicted), together with the nucleotide sequence of an alanine sRNd isolated from yeast, reported by Holley et uL.,~ may provide clues to the recognition process. Two striking sequences in yeast Ma-sR-\i-1, IpGpCpSIeIprt, and dihydroUpCpGpGpdihydroU, each potentially comprising a single-stranded loop at the end of a hairpin-like double-st,randed segment, have been suggested as possible "ant'icodons".30 The Jirst sequence suggest,3 antipnrull~l XatsorKrick base pairing between -Ua-sRJ& and a GpCpC Ua-codon mRS.1 (CGI], GCC. The second sequence raises the possibility of parallel Watson-Crick pairing with an %a-codon of GpCpC, (CGG,:`GCC). Further work is necessary to determine whether one molecule of Ala-sR-\TA is capable of recognizing two or more alanine codons. It is a plesmre to thank Taysir Jaouni, Norma Zsbriskie, and Theresa Caryk for invaluable assistance. o Supported by American CanLer Society postdoctoral fellowship PF 201. t Supported by C-SPHS postdoctoral fellowship 6 F2 AM-17, lOWIll, and American Cancer Society PF 244. $ For brevity, trinucleoside diphosphates are referred to as trinucleotides. l Nirenberg, M., and P. Leder, science, 145, 1399 (19643. 1168 BIOCHEMISTRl`: A-IREA-BERG ET AL. PROC. x. A. S. * Leder, P., and M. XGrenberg. these PROCEEDINGS, 52,420 (1964). 1 Leder, P., and ht. W. Nirenberg: these PROCEEDINGS, 52, 13`21 il964j. ' Bernfield, M. R., and M. n`. Nirenberg. Sc&nxx, 147, 479 i 1965 j. 6 Trupin, J.. F. Rottman, R. Brimacombe. P. Leder. 31. Bernfield, and M. sirenberg, these PROCEEDINGS, 53, 807 `,1965;. 6 Sire&erg, M. W., in Meti in Enzymology. ed. S. P. Colowick and X. 0. Kapian il;ew York: Academic Press. 1964j, vol. 6. p. 17. 7 G-e&erg, M. W., J. H. Masthaei, and 0. VT. Jones, these PROCEEDINGS, 48, 104 (1962). 6 Pestka, S., R. E. lIarshall, and M. Y5-. sirenberg. these PROCEEDINGS. 53, 639 : 1965). 9 Lapidot. Y., and G. Khorana. .I .4m. ('hem Sot.. 85, 3R.T ! 1963, I( Chladek, S., and J. Smn, (`olLclzon C'ncch. C'hrm. Cpmm:ix.. 28, 1301 I963i. 11 Bemfield. 31. R.. and 31 M`. Kirenberg. .4bstra&i4Srh 5ar ionxl .\leel mg. .4meriran Chem- ical Sociel!. Chicago. Illinoi-. .\ugust 1!%4. I* Leder, P., ,11. F. smger, and 1:. L. C. Brimacwmbe. bubmitted tci Lirocko~ist~y. 13 Kaji. H., and -4. Kaji. the-ir PRVCEEDIKGS, 52, 1541 11964). I4 Tauofsky, C., in &nlh4sti and Stw&~~ oj Marwmolecukx, Cold Spring Harbor Symposia on Quantitative Biology. yol. 2& I 1963 L p. 5&l. 16 Marthaei. J. H.. H Klemkauf. and G. Schramm. .Ingw. C&m., 76, 71; 1,1964): r1ngetr. Chem. Intern. Ed. Engl.. 3, 590 (1964). I8 ,%rabhai, A. S.. A. 0. vi. Stretton , S. Brenner, and A. Bolle. .! aburc. 201, 13 i 1964,. ly Brenner, S., -4. 0. W. Stret.ton, and 6. Kaplan, .l`atuw, in press. 1s Weigert, M. C., and A. Garen, in press. le ~~ittmann, H. G., and B. Wit)tmann-Liebold. in S@.fhe& and Strudurt qf Manomoler&.s, &Id Spring Harbor Symposia on Quantitative Biology. vol. 25 ( 1963 j, p. 5SY. a Tsugita, -4., personal communication. *I Jones, 0. W., and M. W. %-e&erg, these PROCEEDINGS. 48, 2115 (196Z). 12 Xrenberg, M. W., and 0. W. Jones, in Symposium m ln.formational Ma.cromobcules, ed. H. Vogel, V. Bryson, and J. Lampen (New York: Academic Press, 1963), p. 451. f* Spryer, J. F., P. Lengyel, C. B&ho, A. J. Wahba, R. S. Gardner, and S. Ochoa, in Synthesti a7ui lStructure of Mb, Cold Spring Harbor Symposia on Quant,itative Biology, vol. 2-S (196.Q p. 559. 2` Nirenberg, M. W., 0. W'. Jones; P. Leder, B. F. C. Clark, W. S. Sly, and S. Pestka, in Synthesifi and Strudure 01 Macrmno e, Cold Spring Harbor Symposia OE Quantitative Biolog?-, vol. 28 (x%3), p. 549. fb WOf3X, C.. n`d.Ute, 194, 1114 i1962j. % Eck, R. Y., S&TUZ, 140, 4;: cl963j. P M-me, C. R., ICSP Rariccr o.f World ~&+ncc, 5,210 (1963). zs Ketiin, 1. B., in S@hesti and St&we oj McacmmoZecuk~, Cold Spring Harbor Symposia on Quantitative Biolm, vol. 2S (19633, p. 579. 29 Ames, B. K., and P. E. Hartman, in &mlnthxxis and Strutuw qf ManomolecLlks, Cold Spring Harbor Symposia on Quantitative Bioltre, vol 2C, I 1963). p. 349. a Hollev, R. W., J. .4pgar, G. il. Everett, J. T. Madison, 11. Narquisee, S. H. Merrill, J. R. Penswick, and A. Zamir, science, 147, 1462 (1965).