Functional transfer RNAs with modifications in the 3'-CCA end ...

Proc. Nati. Acad. Sci. USAVol. 91, pp. 10389-10393, October 1994Biochemistry

Functional transfer RNAs with modifications in the 3'-CCA end:Differential effects on aminoacylation and polypeptide synthesisMINGSONG LiU AND JACK HoROWITZ*Department of Biochemistry and Biophysics, Iowa State University, Ames, IA 50011

Communicated by Keith R. Yamamoto, July 14, 1994

ABSTRACT The trin de CCA sequence is present atthe 3' tennus of all mature tRNAs. Despite thishh degree ofconservation, we have been able to prepare in sitro tri s ofEscherichia coi tRNAV" wh altered 3' ternini that are readilyamnoacylated and can i n polypeptide syntesi. Re-placement of the 3'-termnal with either cytidine oruridine yields a tRNAvd viat that retains ahmost full ami-no qylation activity, having s y constants (V.1/K.) 40-50% that of wid-type tRNAVd. The tRNAVI variant with a3'-terinal e e c ab but b a poorsubstrate for valyl-tRNA sy se, largely as the result of adecrease in the catalytic coItant. End-grop a revealedthe absence of ie at the 3' end of the tRNAVd mutants andIdenifedi then ide expected from the sequence of the DNAtemplate as the pedi t 3'- nal residue; Valcytidinewas Wisted from the C76 mutant. Val-tRNAvYIwit 3'.CCG Is active In poly(U,G)-drected (Val, Phe) copoly-peptide sythsi, whereas the tIRNAVId mutants tn ing in3'-CCC and 3'-CCU, whih are readily -, are

tve. The differential effects of n ide susiluon onlaton and polpeptide synthesis suggest that the uni-

v conserved 3'-CCA end of tRNAs is monitored at two ormore steps In protei nthes that have different nule

s _ .~~~~~~~~~~

All known mature tRNAs have in common a 3' single-stranded CCA end that plays a vital role in tRNA function.The 3'-terminal adenosine interacts directly with the activesite of aminoacyl-tRNA synthetases and the cognate aminoacid becomes esterified to one of its free hydroxyl groups (forreview, see ref. 1). Chemical and nuclease protection studiesindicate that the 3'-CCA sequence also functions in latersteps of protein synthesis: in formation of a ternary complexbetween aminoacyl-tRNA, elongation factor Tu, and GTP (1)and in the binding oftRNA to 23S rRNA at the A, P, and Esites of the 50S ribosomal subunit during the elongation andtranslocation steps of the translation cycle (2-6).

Universal conservation ofthe 3'-CCA sequence suggests itis essential for the aminoacylation of tRNA and/or its par-ticipation in the later stages of protein synthesis. Previousstudies (1) of the effects on tRNA function by varying the3'-terminal adenosine relied largely on chemical modificationor enzymatic replacement of the adenosine. These showedthat the altered tRNAs are inactive in aminoacylation or areonly poorly aminoacylated. Substitutions for the penultimatecytidines have effects that vary with the tRNA studied (1).The scope of these investigations was limited, in part, bydifficulties in obtaining mutant tRNAs. However, the recentdevelopment ofsystems for the in vitro transcription oftRNAfrom recombinant plasmids or phagemids containing tRNAgenes (7) makes it possible to examine the contribution of the3'-CCA terminus to tRNA function in greater detail. As partof our studies of the role of the acceptor stem in the

recognition of Escherichia coli tRNAVan by valyl-tRNA syn-thetase (VRS), we have explored the functional importanceof each residue in the 3'-terminal CCA by systematicallyvarying its nucleotide sequence.

Modification of the 3' end of in vitro tRNA transcripts isdifficult because of the need to retain a restriction endonu-clease recognition site, usually that for BstNI (7), at thedownstream end of the tRNA gene to linearize the DNAtemplate for runoff transcription. We have circumventedlimitations on the 3' sequence of in vitro-transcribed RNA byinserting a Fok I site into a template-containing phagemid,downstream of the RNA gene, positioned to permit Fok I, aclass II restriction endonuclease that cleaves DNA at a siteremote from its recognition site, to linearize the templateDNA regardless of the RNA gene sequence (8).

In this report we describe the in vitro aminoacylation andpolypeptide synthetic activity ofa series oftRNAVal mutantswith modified 3'-CCA termini. In spite of the universalconservation of the 3'-CCA trinucleotide sequence, we haveobtained functional mutants at each position of the 3' end ofE. coli tRNAval, suggesting that no nucleotide in the CCAsequence is absolutely essential for protein synthesis. Thenucleotide substitutions affect aminoacylation and polypep-tide synthesis differently, implying that the 3'-CCA end ismonitored at two or more steps during protein synthesis andthat, at least for some tRNAs, these steps have differentnucleotide recognition specificities. A preliminary account ofthese results has appeared (9).

MATERIALS AND METHODSMaterials. L-[3H]Valine (30 Ci/mmol; 1 Ci = 37 GBq) was

purchased from Amersham and diluted to a specific activityof 5 Ci/mmol before use. Poly(U,G) (U/G, 3:1), pyruvatekinase, phosphoenolpyruvate, RNase T2, and yeast tRNAPhewere obtained from Sigma; T4 RNA ligase was a product ofNew England Biolabs. [y-32P]ATP (3000 Ci/mmol), boughtfrom New England Nuclear, was used to prepare 5'-[32P]pCpas described by England et al. (10). VRS was purified (11)from E. coli GRB276/pHOV1 and T7 RNA polymerase wasisolated from E. coli BL21/pAR1219 (12). tRNA-free E. coliS100 (13) and NH4Cl-washed E. coli ribosomes (14, 15) wereprepared as described. tRNAs were synthesized in vitro byT7 RNA polymerase (16) from tRNA genes inserted into thephagemid pFVAL119, which contains a Fok I site positionedto cleave the DNA template to produce tRNAs with thedesired 3' terminus (8). Transcribed tRNAs were purified byHPLC (8).

Site-Directed Mutagenesis. Mutations were introduced intothe cloned tRNAval gene by site-directed mutagenesis (17)using mutagenic oligonucleotides synthesized by the NucleicAcids Facility at Iowa State University. Mutants were se-lected by dideoxynucleotide DNA sequence analysis (18).

Abbreviation: VRS, valyl-tRNA synthetase.*To whom reprint requests should be addressed.

10389

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

10390 Biochemistry: Liu and Horowitz

Aminoacylation with Val. Initial rates of aminoacylationwith purified E. coli VRS were measured at 370C, in 60-IIreaction mixtures containing 100mM Hepes (pH 7.5), 15mMMgCl2, 10 mM KCl, 7 mM ATP, 1 mM dithiothreitol, 99 pM[3H]Val (5 Ci/mmol), and 4 pM tRNA (determined bymeasuring Val acceptance at high VRS concentration). Re-actions were initiated by addition of 1 nM VRS, and 10-,ulsamples were removed at the indicated times, spotted onWhatman 3MM paper, and processed as described by Bruceand Uhlenbeck (19).

Identification of3'-Terminal Nudeotides in tRNAv'I. tRNAswere labeled at the 3' end with 5'-[32P]pCp by phage T4 RNAligase (10). The labeled tRNAs were purified by electropho-resis on 15% polyacrylamide/8 M urea gels and eluted withbuffer containing 0.5 M ammonium acetate, 0.1% SDS,0.001% Triton X-100, and unlabeled carriertRNA (40 jg/ml).Complete RNase T2 hydrolyzates were prepared (20) and the32P-labeled 3Y-terminal nucleotides (N[32p]p) were separatedby two-dimensional chromatography on cellulose thin layerplates (20). The position of labeled nucleotides was deter-mined by autoradiography using Eastman Kodak X-OmatXAR-5 film.

Isolation of Cytidyl-Val. The tRNAval mutant with a 3'-CCC terminus (4 mg) was aminoacylated with [3H]Val to alevel of 41300 pmol perAA2w unit, as described earlier, exceptthat the concentration of VRS was much higher, 617 nM.After phenol/chloroform extraction and ethanol precipita-tion, any remaining free Val was removed by chromatogra-phy on a 30-ml Sephadex G-25 column equilibrated with 0.02M triethylammonium formate (pH 4.6). The aminoacyl-tRNAwas then treated with 50 pug of pancreatic RNase for 20 minat room temperature and the digest chromatographed on aDEAE-cellulose column (10 ml) was eluted with 0.02 Mtriethylammonium formate (pH 4.6) (21). Samples (0.5 ml)were collected, and the A260 and radioactivity were deter-mined. Peak fractions eluting between 6 and 8 ml werepooled, lyophilized to dryness, dissolved in 1.5 M triethyl-ammonium bicarbonate (pH 9.5), and incubated at 370C for 30min to hydrolyze the aminoacyl ester bond. The sample wasdried, dissolved in 0.5 ml of 0.4 M Tris HCl (pH 7.0), and itsUV absorption spectrum was compared with those of cyti-dine and adenosine.Poly(U,G)-Dlded (Val, Phe) Copolypeptide Synthesis.

Polypeptide synthesis was measured in 100-ip reaction mix-tures containing 50 mM Bicine (pH 7.5), 50 mM NH4Cl, 20mM magnesium acetate, 3 mM ATP, 1 mM GTP, 1 mMdithiothreitol, 4 units of pyruvate kinase, 10 mM phospho-enolpyruvate, 0.05 mM nonradioactive Val, 0.01 mM nonra-dioactive Phe, 40 ug oftRNA-free E. coli S100, 0.73 A260 unitof poly(U,G) (3:1), 4.2 A260 units of 1 M NH4Cl-washed E.coli ribosomes, 30 pmol of [3H]Val-tRNAval, and 96 pmol of

10

EQ / ~~~~~~~U76

0.0 0.5 1.0 1.5 2.0 2.5

Time, min

FIG. 1. Aminoacylation with valine of wild-type E. coli tRNA"a'and variants having altered 3' termimi. The time course of [3HIVal

incorporation into wild-type tRNAval (A76; open circles) and tRNA~~alvanants containing U76 (solid circles), C76 (open triangles), and G76

(solid triangles), was determined.

yeast tRNAPhC. Reactions were initiated by addition of ribo-

somes and incubation was at 370C for the specified times. The

reaction was stopped by adding 4 ml of 5% (wt/vol) trichlo-

roacetic acid and heating at 950C for 20 min. After cooling to

room temperature, the samples were collected on Millipore

filters and washed five times with 4 ml of 5%o trichloroacetic

acid. The filters were then dried and radioactivity was

measured.

RESULTS

Aminoacylatlon of tRNA~~a Mutants with Substitutions for

the 3'-Terminal Adenosine. Fig. 1 compares the time course

of in vitro aminoacylation of wild-type tRNAval and three

mutant transcripts with the usual 3'-terminal A76 replaced by

U76, C76, or G76. Both tRNAVaI variants with 3'-terminal

pyrimidines, U76 and C76, were readily aminoacylated by

purified VRS. Kinetic studies showed that the relative ami-

noacylation efficiency (Vma/Km) of the 3'-CCU and 3'-CCC

mutants was 40-50%o that of wild-type tRNAVaJ (Table 1); Km

remained essentially unchanged, whereas Van was some-

what decreased. In contrast, the G76 tRNAVaI variant, ter-

minating in 3'-CCG, was a poor substrate for the enzyme

(Fig. 1). Although the G76 mutant could be fully aminoac-

ylated to a level of 1100 pmol per A,.60 unit at high VRS

Table 1. Kinetic parameters for the aminoacylation of 3'-terminal E. coli tRNAval variantsVal, pmol per Km, Vn,".tmol Relative

Exp. RNA used A2,0 unit AM per min per mg Vmx/Km Vmx/Km1 Wild type 1350 1.4 5.0 3.6 (1.0)2 A76-. C76 1200 1.0 1.8 1.8 0.503 A76-+ U76 1580 1.9 2.7 1.4 0.394 A76 - G76 1100 0.78 0.017 0.022 0.00615 C74 +U74* 1690 1.2 1.4 1.2 0.336 C74- A74 1680 1.3 4.0 3.1 0.867 C74- G74 1340 0.7 2.7 3.9 1.18 C75 +U75* 1630 1.4 1.5 1.1 0.319 C75 -A75 1650 1.1 1.6 1.5 0.4210 C75 G75 1110 0.72 0.027 0.038 0.0111 C74C75 U74U75* 1130 2.5 0.049 0.020 0.0056

Aminoacylation kinetics were measured at 37C; RNA concentrations ranged from 0.5 to 4 pM. Kmand Vx values were determined by a least-squares fit of the double-reciprocal plot of the data.*Data are from ref. 8 and are included here for completeness.

Proc. Nad. Acad. Sci. USA 91 (1994)

Proc. Natl. Acad. Sci. USA 91 (1994) 10391

A EAp

.Cp

Gp Up

1D

t 2D

C

1D

t 2D

B : Ap

Cp

IG1D

t-E 2D

..1....s. Gc Ap

4flCp *Cp

u~~~~~pGt>9iGpo Gp Up

2D

FIG. 2. Identification of the 3'-terminal nucleotides of tRNAValvariants. The 3'-terminal nucleotides (Np) in complete RNase T2hydrolysates of tRNAs 3'-end-labeled with 5'-[32P]pCp by T4 RNAligase were separated by two-dimensional chromatography on cel-lulose thin-layer plates. (A) Native tRNAval (A76). (B-D) tRNAvalvariants having 3'-terminal C76, U76, and G76, respectively. Theposition of labeled nucleotides was determined by autoradiography;unlabeled nucleotide 3'-phosphate markers, visualized by UV shad-owing, are circled. D, dimension.

concentrations, its V./Km was 160 times lower than wildtype largely because of a 300-fold decrease in V. (Table 1).It seems that the identity of the nucleotide at the 3' terminushas little effect on the affinity ofVRS for tRNAval, but it does

modify k~,t, presumably by influencing the positioning of the3'-terminal nucleoside in the catalytic site of the synthetase.The bond linking Val to the tRNAval mutants is alkali

labile. Incubation of the aminoacylated tRNAs in 1.8 MTris HCl (pH 8.0) at 370C for 90 min, conditions routinelyused to remove amino acids esterified to tRNA (22), com-pletely deacylated the mutant tRNAs, indicating that theamino acid is bound to a ribose hydroxyl group.

Identification of the 3' Nucleotide in Mutant tRNAs. Wehave verified the presence ofthe expected 3' terminus in eachtRNAval mutant by 3'-end labeling with 5'-[32P]pCp and T4RNA ligase (11) and identifying the labeled 3' nucleotides(Np) in a complete RNase T2 digest of the ligation productsby two-dimensional TLC (20). The results (Fig. 2) confirmedthat each tRNA variant had predominantly the expected3'-terminal nucleotide; 3' adenosine was found only in wild-type tRNAval (Fig. 2A). As anticipated, because of theknown non-template-directed incorporation of one or moreextra nucleotides during in vitro tRNA transcription by 17RNA polymerase (7, 10, 23, 24), some 3'-terminal sequenceheterogeneity was evident in each tRNA transcript (Fig. 2).Seventy-five percent or more of the transcripts terminated inthe nucleotide expected from the sequence of the templateDNA; most of the remaining transcripts terminated in cyti-dine. Identical results were obtained when the tRNAs wereanalyzed before or after aminoacylation, effectively rulingout 3'-end turnover and incorporation of 3'-terminal adeno-sine during the aminoacylation reaction.

Isolation and Characterization of Val-Cytidine. More directevidence that Val is esterified to a 3'-terminal nucleosideother than adenosine was obtained by isolating and charac-terizing the aminoacyl-nucleoside from the mutant tRNAvaIterminating in 3'-CCC. After aminoacylation with [3H]Val,the tRNA was recovered by phenol/chloroform extractionand ethanol precipitation and was freed from residual unes-terified Val by gel filtration on Sephadex G-25. Hydrolysiswith RNase A released the 3'-terminal Val nucleoside, whichwas then isolated by column chromatography (Fig. 3) on

0.25

0.20

0.15

0.10

0.05

0.00

Elution volume, ml

75 FIG. 3. Isolation and characterizationE of Val-cytidine. The tRNAvaI variant- with a 3'-CCC terminus was aminoacyl-> ated with [3H]Val (1300 pmol per A260

unit) and hydrolyzed with pancreaticRNase. The digest was chromatographedon a DEAE-cellulose column (10 ml) andeluted with 0.02 M triethylammoniumformate (pH 4.6). Samples (0.5 ml) werecollected, and the A2w0 (solid circles) andradioactivity (open circles) were deter-mined. (Inset) Fractions containing ra-dioactivity were pooled and dried, andthe aminoacyl ester bond was hydro-lyzed. After being dried, the sample wasdissolved in 0.5 ml of0.4 M TrisHCl (pH7.0), and its UV absorption spectrum(open circles) was compared with thoseof cytidine (solid circles) and adenosine(triangles).

Biochemistry: Liu and Horowitz

10392 Biochemistry: Liu and Horowitz

DEAE-cellulose at pH 4.6, to stabilize the aminoacyl esterlinkage (21). This method readily separates positivelycharged nucleoside esters from weakly absorbed free nucle-osides and strongly bound negatively charged nucleotides(experiments not shown and ref. 21). All the radioactive Valwas eluted with the first UV-absorbing peak (Fig. 3), whichcontains the aminoacyl nucleoside. The second peak wasidentified as cytidine on the basis of its UV absorptionspectrum; nucleotides and oligonucleotides were eluted later(results not shown). Fractions containing [3H]Val werepooled and lyophilized to dryness, and the nucleoside com-ponent was identified as cytidine by its UV-absorption spec-trum at pH 7.0 (Fig. 3 Inset), after hydrolysis of the ami-noacyl ester link at pH 9.5. The calculated molar ratio ofVal/cytidine in the isolated aminoacyl cytidine is 0.9, basedon the known specific activity of the radioactive amino acidand the molar extinction coefficient of cytidine at 270 nm andpH 7.

Participation of 3'-Terminal tRNAvaI Variants in Polypep-tide Synthesis. To further examine the function oftRNAs withsubstitutions for the 3'-terminal adenosine, their activity inpolypeptide synthesis was tested in an in vitro system de-pendent on the addition of aminoacyl tRNA. AminoacylatedtRNA was used in these experiments to eliminate effects dueto differences in the aminoacylation rates of the differenttRNAs. This is especially important in experiments with theG76 mutant, which, although a poor substrate for VRS, canbe fully aminoacylated (Table 1). The random polynucleotidePoly(U,G) (U/G = 3:1) was found to be an effective mes-senger, provided Phe and tRNAPhle were also present. Fig. 4shows the time course of polypeptide synthesis and itsdependence on added poly(U,G). Additional control exper-iments (results not shown) indicated that only low levels ofVal were incorporated into polypeptide when tRNAPhC orribosomes are omitted. The results clearly establish that the3'-CCG mutant is an effective donor of Val for polypeptidesynthesis, there being little difference in the rate of incorpo-ration from wild-type tRNAval (Fig. 4). Since these reactionswere carried out in the presence of a large excess ofunlabeledfree Val, the [3H]Val incorporated must have come directlyfrom the added Val tRNAval. The two variants of tRNAvalwith 3'-terminal pyrimidines (U or C) were essentially inac-tive in polypeptide synthesis; incorporation ofVal from thesetRNAs was little above background levels (Fig. 4). All the

5

_ 4

E

0.

-

° 2

I

0l

0 20 40

Time, min60

FIG. 4. Poly(U,G)-directed (Val, Phe) copolypeptide synthesis.The time course of [3H]Val incorporation into polypeptide fromaminoacylated wild-type tRNAvaI (A76; open circles) and tRNAvalvariants containing U76 (solid circles), C76 (open triangles), and G76(solid triangles) was measured in a tRNA-dependent system.

aminoacylated tRNAs are stable under the conditions ofpolypeptide synthesis. Little or no hydrolysis of aminoacyltRNA occurs (results not shown), ruling out the possibilitythat the inability oftRNAVal mutants with 3'-terminal pyrim-idines to participate in polypeptide synthesis is due to a rapiddeacylation of the tRNAs.

Aminoacylation of tRNAvid Mutants with Substitutions forthe Conserved 3' Cytidines. The penultimate cytidines at theconserved 3'-CCA end oftRNAval were also not essential foraminoacylation of the tRNA. Substituting adenosine or gua-nosine for C74 yielded tRNAs that retained almost fullactivity (Table 1); aminoacylation of the U74 mutant wassomewhat reduced, but it remained a reasonably good sub-strate for VRS. More pronounced effects on the aminoacy-lation of tRNAVal were observed when C75 was replaced.Substitution of C75 with either U or A generated tRNAvalvariants with reduced but still significant activity, 2- to 3-foldlower than the wild-type tRNA. The G75 mutant was quite apoor substrate for VRS and the double mutant U74/U75 wasalmost inactive. Replacement of either C74 or C75 had littleeffect on the affinity of VRS for tRNAVal; Km values re-mained almost unchanged or even decreased somewhat. Asfor mutations at position 76, nucleotide changes at position 74or 75 affected primarily the catalytic step (VaJ,.)

DISCUSSIONTo our knowledge, these results provide the first descriptionof functional tRNAs having a nucleoside other than adeno-sine or an adenosine derivative at the 3' end. Evidently the3'-terminal adenosine is not an absolute requirement foraminoacylation of at least some tRNAs. Its replacement witheither cytidine or uridine yields E. coli tRNAval variants thatretain almost full aminoacylation activity. However, thevariant with a 3'-terminal guanosine, although fully charge-able, is a poor substrate for VRS. Some of the differences inaminoacylation activity of the 3'-end mutants may be due toeffects of sequence variation at the 3'-CCA terminus on thestructure and stability of the adjoining amino acid acceptorstem (25).When the acceptor end of tRNA binds to the synthetase,

its terminal nucleotide is properly positioned in the catalyticsite, presumably, through direct and water-mediated hydro-gen bonding between donor (d) and acceptor (a) groups on thenucleotide and appropriately situated amino acids at theactive site. The pattern of hydrogen-bonding acceptor anddonor sites at positions N1 and N6 of the pyrimidine ring ofadenosine is shown in Fig. 5 and can be designated as a d.Comparison with the structures of the other three nucle-obases reveals that cytidine and uridine, but not guanosine,resemble adenosine in having an a-d array of hydrogen bonddonor and acceptor groups. This commonality may accountfor the ability of cytidine and uridine to substitute for theusual A76 and suggests that the exocyclic amino group (N6)and the ring N1 of 3'-terminal adenosine are involved inhydrogen bonding at the catalytic site ofVRS. Such hydrogenbonding with the --OH ofa serine side chain in the active siteofaspartyl-tRNA synthetase has been observed in the crystalstructure of yeast tRNAASP complexed with its cognatesynthetase (26).Reuven and Deutscher (27) have reported that mutant SU3

genes coding for tRNATYrsu' with nucleotide substitutions forthe 3'-terminal adenosine can efficiently produce functionalsuppressor tRNAs in E. coli. Suppression by these mutanttRNAs, in contrast to that by normal tRNAs, required bothtRNA nucleotidyltransferase and exoribonuclease activities,indicating a need for 3'-end turnover, presumably to removethe incorrect 3' nucleotide and resynthesize the correct CCAterminus. Low levels of functional suppressors are, however,made even in cells lacking nucleotidyltransferase. This was

Proc. Nad. Acad Sci. USA 91 (1994)

Proc. Natl. Acad. Sci. USA 91 (1994) 10393

H, H a

N 0

N N1 N

N N H N N N

FR R

ADENOSINE GUANOSINE

da

0, N

'Ha

NN

N N

H R~~~~

URIDINE CYTIDINE

FIG. 5. Structures of purine and pyrimidine nucleosides showingthe hydrogen bond acceptor (a) and donor (d) sites used for basepairing. The shaded arrows indicate the aid hydrogen bond donor-acceptor pattern common to adenosine, cytidine, and uridine butabsent in guanosine.

ascribed to the presence of an additional enzyme in E. colicapable of repairing the 3' end of tRNA. The authors argueagainst the possibility that mutant tRNATYrsu' with substitu-tions for the 3'-terminal adenosine are capable of acceptingTyr and functioning as suppressors. Tyrosyl-tRNA synthetasemay differ from VRS in having a specific requirement for the3'-terminal adenosine. These differences are not related to theclassification of synthetases into two groups, class I and II (28,29), because VRS and tyrosyl-tRNA synthetases are bothclass I enzymes (30).The 3' end of elongator tRNAs also plays an active role in

the later steps of protein synthesis. Chemical and nucleaseprotection experiments have shown that the 3'-CCA se-quence interacts with the 23S rRNA of the large ribosomalsubunit (2-4). Functional E. coli tRNAVal mutants with GCAand ACA 3' termini were recently isolated (31), which whenbound to their cognate codon at the ribosomal P site affectedcodon recognition one or sometimes two codons downstreamof the Val codon. Furthermore, mutations at the conservedG2252 and G2253 of 23S rRNA, which may interact with the3'-CCA end of tRNA, also affect codon reading at theribosomal A site (32). Although the exact nature of theassociation of the 3'-CCA sequence with rRNA still needs tobe clarified, it has been suggested that base pairing betweenthe CCA sequence and a complementary sequence within thecentral loop of domain V of 23S rRNA is involved in tRNAbinding to the peptidyltransferase center (6). Our results onthe protein synthetic activity of the 3'-terminal mutants oftRNAval are consistent with this possibility. Mutations thatdisturb the base-pairing potential of the 3' end of tRNAv~d(A76 -) C76 or U76) inactivate the tRNA for polypeptidesynthesis. The G76 variant retains activity, possibly becauseof its ability to form a GNU wobble base pair with 23S rRNA.Alternatively, the ability oftRNAval mutants with 3'-terminalpurines to function in polypeptide synthesis, in contrast tothose with 3'-terminal pyrimidines, may indicate a role for theN7 of the purine ring in the interaction with rRNA.The universal conservation of the 3'-CCA end implies it

has a specific function. However, at least forE. coli tRNAVal,the 3'-CCA sequence is not an absolute requirement foraminoacylation orfor the later steps ofpolypeptide synthesis.Interestingly, the two 3'-terminal mutants of tRNAVaI thatcan be actively aminoacylated, 3'-CCU and 3'-CCC, are

unable to function in peptide-bond formation, whereas the

G76 mutant, which is a poor substrate for VRS, is active inpolypeptide synthesis. These results suggest that the 3'terminus oftRNA is proofread at two (or more) discrete stepsduring protein synthesis. At least for E. coli tRNAVal, thesecan be uncoupled and distinguished and have different nu-cleotide recognition specificities. Presumably, all requiredrecognition elements are present only in the 3'-CCA se-quence, accounting for its universal conservation at the 3'end of tRNA. It remains to be determined whether similarconsiderations apply to other tRNAs.

This research was supported by a grant from the National Insti-tutes of Health (GM 45546). This is Journal Paper J-15720 ofthe IowaAgriculture and Home Economics Experiment Station, Ames, LA,Project 2566.

1. Sprinzl, M. & Cramer, F. (1979) Prog. Nucleic Acid Res. Mol.Biol. 22, 1-69.

2. Moazed D. & Noller, H. F. (1991) Proc. NatI. Acad. Sci. USA88, 3725-3728.

3. Moazed, D. & Noller, H. F. (1989) Cell 57, 585-597.4. Lill, R., Lepier, A., Schwfigele, F., Sprinzl, M., Vogt, H. &

Wintermeyer, W. (1988) J. Mol. Biol. 203, 699-705.5. Tezuka, M. & Chladek, S. (1990) Biochemistry 29, 667-670.6. Lill, R., Robertson, J. M. & Wintermeyer, W. (1989) EMBOJ.

8, 3933-3938.7. Sampson, J. R. & Uhlenbeck, 0. C. (1988) Proc. NatI. Acad.

Sci. USA 85, 1033-1037.8. Liu, M. & Horowitz, J. (1993) BioTechniques 15, 264-266.9. Liu, M., Chu, W.-C. & Horowitz, J. (1994) FASEB J. 8, 1266

(abstr.).10. England, T. E., Bruce, A. G. & Uhlenbeck, 0. C. (1980)

Methods Enzymol. 65, 65-74.11. Chu, W.-C. & Horowitz, J. (1991) Biochemistry 30, 1655-1663.12. Grodberg, J. & Dunn, J. J. (1988) J. Bacteriol. 170, 1245-1253.13. Muench, K. H. & Berg, P. (1966) in Procedures in NucleicAcid

Research, eds. Cantoni, G. L. & Davies, D. R. (Harper andRow, New York), pp. 375-383.

14. Pestka, S. (1968) J. Biol. Chem. 243, 2810-2820.15. Erbe, R. W., Nau, M. M. & Leder, P. (1969) J. Mol. Biol. 39,

441-460.16. Chu, W.-C. & Horowitz, J. (1989) Nucleic Acids Res. 7,

7241-7252.17. Taylor, J. W., Ott, J. & Eckstein, F. (1985) Nucleic Acids Res.

13, 8765-8785.18. Sanger, F., Coulson, A. R., Barrell, B. G., Smith, A. J. H. &

Roe, B. A. (1980) J. Mol. Biol. 143, 161-178.19. Bruce, G. A. & Uhlenbeck, 0. C. (1982) Biochemistry 21,

855-861.20. Nishimura, S. (1979) in Transfer RNA: Structure, Properties

and Recognition, eds. Schimmel, P. R., S611, D. & Abelson, J.(Cold Spring Harbor Lab. Press, Plainview, NY), pp. 551-552.

21. Wolfenden, R., Rammler, D. H. & Lipmann, F. (1964) Bio-chemistry 3, 329-338.

22. Sarin, P. S. & Zamecnik, P. C. (1964) Biochim. Biophys. Acta91, 653-655.

23. Reyes, V. M. & Abelson, J. (1987) Anal. Biochem. 166,90-106.24. Samuelsson, T., Boren, T., Johansen, T.-I. & Lustig, F. (1988)

J. Biol. Chem. 263, 13692-13699.25. Limmer, S., Hofmann, H.-P., Ott, G. & Sprinzl, M. (1993)

Proc. Natl. Acad. Sci. USA 90, 6199-6202.26. Cavarelli, J., Eriani, G., Rees, B., Ruff, M., Boeglin, M.,

Mitschler, A. Martin, F., Gangloff, J., ThierTy, J.-C. & Moras,D. (1994) EMBO J. 13, 327-337.

27. Reuven, N. B. & Deutscher, M. P. (1993) Proc. NatI. Acad.Sci. USA 90, 4350-4353.

28. Eriani, G., Delarue, M., Poch, O., Gangloff, J. & Moras, D.(1990) Nature (London) 347, 203-206.

29. Cusack, S., Berthet-Colominas, C., Hirtlein, M., Nasser, N. &Leberman, R. (1990) Nature (London) 347, 249-255.

30. Cavarelli, J. & Moras, D. (1993) FASEB J. 7, 79-86.31. O'Connor, M., Willis, N. M., Bossi, L., Gesteland, R. F. &

Atkins, J. F. (1993) EMBO J. 12, 2559-2566.32. Gregory, S. T., Lieberman, K. R. & Dahlberg, A. E. (1994)

Nucleic Acids Res. 22, 279-284.

Biochemistry: Liu and Horowitz