Proc. Natl. Acad. Sci. USAVol. 80, pp. 6755-6759, November 1983Biochemistry

Anticodon loop size and sequence requirements for recognition offormylmethionine tRNA by methionyl-tRNA synthetase

(synthesis of mutant tRNAs in vitro/RNA-protein interactions/aminoacylation/T4 RNA ligase)

LADONNE H. SCHULMAN AND HEIKE PELKADepartment of Developmental Biology and Cancer, Division of Biology, Albert Einstein College of Medicine, Bronx, NY 10461

Communicated by Harry Eagle, July 21, 1983

ABSTRACT Previous work from our laboratory identifiedseveral specific sites in Escherichia coli tRNAfMet that are essen-tial for recognition of this tRNA by E. coli methionyl-tRNA syn-thetase (EC 6.1.1.10). Particularly strong evidence indicated a rolefor the nucleotide base at the wobble position of the anticodon inthe discrimination process. To further investigate the structuralrequirements for recognition in this region, we have synthesizeda series of tRNAfIet derivatives containing single base changes ineach position of the anticodon. In addition, derivatives containingpermuted sequences and larger and smaller anticodon loops havebeen prepared. The variant tRNAs have been enzymatically syn-thesized in vitro. The procedure involves excision of the normalanticodon, CAU, by limited digestion of intact tRNAMet with pan-creatic RNase. This step also removes two nucleotides from the3' CpCpA end. T4 RNA ligase is used to join oligonucleotides ofdefined length and sequence to the 5' half-molecule and subse-quently to link the 3' and modified 5' fragment to regenerate theanticodon loop. The final step of the synthesis involves repair ofthe 3' terminus with tRNA nucleotidyltransferase. The syntheticderivative containing the anticodon CAU is aminoacylated withthe same kinetics as intact tRNAfmet. Base substitutions in thewobble position reduce aminoacylation rates by at least five or-ders of magnitude. The rates of aminoacylation of derivativeshaving base substitutions in the other two positions of the anti-codon are 1/55 to 1/18,500 times normal. Nucleotides that havespecific functional groups in common with the normal anticodonbases are better tolerated at each of these positions than thosethat do not. A tRNAfMet variant having a six-membered loop con-taining only the CA sequence of the anticodon is aminoacylatedstill more slowly, and a derivative containing a five-membered loopis not measurably active. The normal loop size can be increasedby one nucleotide with a relatively small effect on the rate of ami-noacylation, indicating that the spatial arrangement of the nu-cleotides is less critical than their chemical nature. We concludefrom these data that recognition of tRNAf¶et requires highly spe-cific interactions of methionyl-tRNA synthetase with functionalgroups on the nucleotide bases of the anticodon sequence.

We have previously studied the effect of chemical modifica-tions at 25 different sites in Escherichia coli tRNAfMet on theability of the tRNA to be aminoacylated by E. coli methionyl-tRNA synthetase (EC 6.1.1.10) (1, 2). Most of these structuralalterations did not significantly impair the interaction of tRNAfmetwith Met-tRNA synthetase; however, modification of specificnucleotides in three structural regions drastically reduced me-thionine acceptance. These results focused our attention on theanticodon, the variable loop, and the acceptor stem of tRNAfMetfor more detailed analysis of the structural requirements forprotein-tRNA recognition. We have shown that the anticodonwobble base plays an essential role in this process (3, 4). In this

paper, we describe the results of a systematic examination ofthe effects of alterations in anticodon loop size and sequenceon recognition of tRNAMet by Met-tRNA synthetase.

MATERIALS AND METHODSMaterials. Nucleoside 3'-phosphates, nucleoside 5'-diphos-

phates, nucleoside 3',5'-bisphosphates, poly(A,C), and GpApCwere purchased from P-L Biochemicals. Nucleoside 5'-mono-phosphates, dinucleoside monophosphates, GpCpC, and GpCpUwere obtained from Sigma. [y-32P]ATP, [a-32P]ATP, and [ S]-methionine were purchased from Amersham. E. coli tRNAfmet(1.8 nmol/A260 unit), primer-dependent Micrococcus luteuspolynucleotide phosphorylase, calf intestinal alkaline phospha-tase, and nuclease P1 were obtained from Boehringer Mann-heim. RNases T1 and U2 were purchased from Calbiochem andPhy M and Bacillus cereus RNases were from P-L Biochemi-cals. E. coli Met-tRNA synthetase was purified from E. coli K-12 strain EM 20031 (5) and T4 RNA ligase was purified fromE. coli infected with T4 phage strain SP62, amN82 (6) as de-scribed. Purified rabbit liver tRNA nucleotidyltransferase wasa gift from M. Deutscher.

Synthesis of Oligonucleotides. The trinucleotides GpCpA andCpApU were synthesized by reaction of GpC with ADP andCpA with UDP, using polynucleotide phosphorylase as de-scribed by Thach and Doty (7). CpApGp was synthesized by asimilar reaction of CpA with GDP in the presence of RNase T1at 250 units/ml. CpAp was obtained from a digest (18 hr at37°C) of poly(A,C) (1:1) with RNase U2 (0.5 unit/mg of RNA)in 50 mM sodium acetate, pH 4.5, followed by incubation with0.125 M HCI at room temperature for 6 hr. The tetranucleo-tides GpCpApCp and GpCpUpAp were synthesized by addi-tion of pCp to GpCpA and pAp to GpCpU, using T4 RNA ligase(8). Treatment of the tetranucleotides with RNase T1 yieldedthe trinucleotides CpApCp and CpUpAp. The trinucleotidesCpApUp, CpCpUp, ApCpUp, and CpUpUp were similarlysynthesized by ligase-catalyzed addition of pUp to the corre-sponding trinucleoside diphosphates followed by cleavage ofthe resulting tetranucleotides with RNase T1. The tetranucleo-tide CpApUpAp was synthesized by addition of pAp to CpApU,using RNA ligase. All oligonucleotides were purified by columnchromatography and analyzed as described elsewhere (3). Oli-gonucleotides were phosphorylated at the 5' terminus by using[y-32P]ATP and PseT 1 polynucleotide kinase (9).

Synthesis of tRNAflet Containing Altered Anticodon LoopSequences. Half-molecule-sized fragments of tRNAM&et miss-ing the anticodon nucleotides and two nucleotides of the 3' ter-

Abbreviations: tRNATA'ut, tRNA containing the sequence CAU in theanticodon position that has been enzymatically synthesized in vitro fromhalf-molecule-sized fragments of Escherichia coli tRNA Met (other syn-thetic tRNAs are similarly indicated by the sequences in their anticodonloops); p*, 32P-labeled phosphate.

6755

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

Dow

nloa

ded

by g

uest

on

June

14,

202

1

6756 Biochemistry: Schulman and Pelka

minal CpCpA sequence were isolated after limited digestion ofthe tRNA with pancreatic RNase (3). Intact molecules contain-ing normal and altered anticodon loop sequences were synthe-sized by using RNA ligase and polynucleotide kinase,_ and the3'-terminal sequence was enzymatically replaced by using tRNAnucleotidyltransferase, as described before (3).

Aminoacylation of Modified tRNAfMet. Reaction mixturesfor assay of methionine acceptance (0.05-0.15 ml) contained0.01-0.2 ,uM tRNA, 20 mM imidazole HCI at pH 7.5, 0.1 mMEDTA, 2 mM ATP, 4mM Mg9l2, 150mM NH4Cl, bovine serumalbumin at 10 Ag/ml, and 17AM [3S]methionine (1,500-100,000cpm/pmol). Samples were equilibrated at 260C for 5 min andreactions were initiated by addition of purified E. coli Met-tRNAsynthetase. Incubation at 26°C was continued for various timesand aliquots (5-45 ,ul) were pipetted onto 2.5-cm Whatman 3MM filter disks, and the disks were added to cold 10% tri-chloroacetic acid containing I mM unlabeled methionine (10 mlper filter) and washed for 10 min;followed by four washes withcold 5% trichloroacetic acid and one wash with 95% (vol/vol)ethanol. The filters were dried for 20 min under a heat lamp

.and-placed in scintillation vials. with 5 ml of Econofluor (NewEngland Nuclear), and the radioactivities were measured in aliquid scintillation counter. Aliquots were taken in triplicate formeasurement of aminoacylation yields in 30-min incubations at26°C at various enzyme concentrations. Initial rates of ami-noacylation were measured under conditions iii which-methi-onine incorporation was linear with time and proportional toenzyme concentration. Aminoacylation kinetics were measuredat 0.04 nM enzyme for tRNAfMet and tRNACm' (aliquots takenat 1-min intervals), 0.4 nM for tRNAmtA (aliquots taken at 0.5-min intervals), 4 nM for tRNAfcmIu (aliquots taken at 0.5-minintervals), 40 nM for tRNAfmueu, tRNAftGt, and tRNAAJueA (ali-quots taken at 1-min intervals), and 40 nM for tRNOAect andtRNAfMAet (aliquots taken at 5-min intervals). Initial rates weremeasured at five different tRNA concentrations except for theCUA derivative (two). Relative initial rates of aminoacylationwere determined after calculation of the moles of methionineincorporated per mole of enzyme per minute for each tRNAderivative. Reaction mixtures containing enzyme but no tRNAor enzyme plus equivalent amounts of yeast tRNAPhe gave thesame blank values for methionine acceptance and were pre-pared in parallel with each assay. Derivatives that were inactiveat low Met-tRNA synthetase concentrations were tested for theirability to inhibit aminoacylation of tRNAfmet in reaction mix-tures (50 ,ul) that contained 0.03 MM tRNA et, 0.06-0.3 MtRNAfMet derivative, and 40pM enzyme. Incubations were at26°C and 9-,u1 aliquots were withdrawn at 1-min intervals formeasurement of methionine acceptance. Parallel incubation mix-tures containing enzyme but no tRNA or enzyme plus inhib-itor but no unmodified tRNAfMet gave the same blank values.

RESULTSSynthesis and Structural Characterization of tRNAMet De-

rivatives. Derivatives of tRNAfet containing alterations in an-ticodon loop size and sequence were synthesized by a proce-dure analogous to that described previously (3). Limited digestionof tRNAfmet with pancreatic RNase was used to generate half-molecule-sized fragments missing the anticodon nucleotides andthe two terminal nucleotides of the 3' CpCpA sequence. Thedephosphorylated 5' half-molecule was annealed with the 3'half-molecule -containing a3'-phosphate group. 5'-32P-Labeledoligonucleotides of different length and sequence were joinedto the 3'-OH group of the 5' fragmentby using T4 RNA ligase.The extended 5' fragment and the 3' fragment were dephos-phorylated at the 3' termini-and phosphorylated at the 5' ter-mini with polynucleotide kinase and [y-32P]ATP. The anticodon

loop was joined by incubation of the annealed complex withRNA ligase and the 3'-terminal CpCpA sequence was enzy-matically repaired by using tRNA nucleotidyltransferase in thepresence of unlabeled CTE and [a-32P]ATP. The final productswere isolated by polyacrylamide gel electrophoresis in 7 M urea.All of the tRNA derivatives migrated in the position expectedon the basis of the length of the oligonucleotide inserted in theanticodon loop except for tRNAMAeGt. Several different prepa-rations of this derivative migrated more slowly on denaturinggels than the products prepared by other trinucleotide inser-tions. Low specific activity 32p labels were incorporated at the5' and 3' termini and at the sites of joining in the anticodonloop, facilitating calculation of yields and product purity at eachstage of the synthesis. Results were similar to those obtainedpreviously for synthesis of tRNA!let derivatives containing basesubstitutions in the wobble position (3). Anticodon loop closureand CpCpA repair were essentially quantitative in all cases. Theoverall recovery of the desired tRNA derivatives was limited bythe incomplete addition of oligonucleotides to the 5' half-mol-ecule. The desired adducts were obtained in yields of 15-25%;however, partial degradation of the initial products by reversereactions of RNA ligase (10) necessitated fractionation of thereaction mixtures on denaturing polyacrylamide gels and re-duced the yield of desired product by as much as 50% in somecases.

Each tRNA derivative was digested with T1 RNase and theresulting'32P-labeled oligonucleotides were separated by chro-matography on DEAE-ceHuloserin the presence of 7 M urea.All samples yielded the expected 5' and 3' oligonucleotidesP*CpGp and CpApApCpCp*A plus an additional oligonucleo-tide derived from the anticodon loop region. The size of thisoligonucleotide was determined by chromatography with un-labeled oligonucleotide markers. The anticodon sequence ofnormal tRNAfmet is found in an 11-residue oligonucleotide,CmpUpCpApUpApApCpCpCpG, after digestion with T1 RNase.The corresponding synthetic derivative yielded the labeledproduct CmpUp*CpApUp*ApApCpCpCpG, which migratedwith the unlabeled oligonucleotide marker and yielded CmpUp*and ApUp* in a 1:1 ratio after further digestion with pancreaticRNase. Derivatives synthesized by insertion of CCU, CUU, ACU,CUA, and CAC in the anticodon loop also gave 32P-labeled un-decanucleotides on digestion with T1 RNase, while insertion ofCAUA and CACA yielded dodecanucleotides, insertion of CAyielded a decanucleotide, and insertion of C yielded a nonanu-cleotide. In each case, further digestion of the isolated oligo-nucleotide with pancreatic RNase gave the expected 32P-la-beled products. The tRNAfMet derivative synthesized by insertionofCAG yielded the T1 RNase product CmpUp*CpApGp*, whichgave CrpUp* and ApGp* in a 0.9:1.0 ratio after digestion withpancreatic RNase. On the basis of incorporated radioactivity,the products are 70-90% pure. On the basis of the ratio of 32Pto A2w, specific activities of 800-1,000 pmol/A2m unit are ob-tained. As noted previously (3), similar specific activities areobtained for equivalent amounts of intact tRNAfmet after elec-trophoresis and isolation from polyacrylamide gels due to con-tamination of small samples with nondialyzable, ethanol-pre-cipitable UV-absorbing material from the gels. The 32p labelincorporated during the synthesis has therefore been used tocalculate the actual concentration- of each product for amino-acylation studies.

As a further check on the synthetic procedure, tRNA deriv-atives were labeled at the 5' termini. by using polynucleotidekinase and high specific activity [y-32P]ATP for gel sequenceanalysis by the- method of Donis-Keller et al. (11). Sample au-toradiograms obtained from sequencing the products synthe-sized by insertion of the trinucleotides CAU, CAG, and ACU

Proc. Natl. Acad. Sci. USA 80 (1983)

Dow

nloa

ded

by g

uest

on

June

14,

202

1

Proc. Natl. Acad. Sci. USA 80 (1983) 6757

tRNA fMetRNCAU tRNAMetRNCAG tRNAACUG A A+U U+C G A A+UU+C G A A+U U+C

umIO G70G64 G__063 0700

___ iG64G53 loGom G53

G5i2 _ 553 ^52 _ _49 _ G524of '1-49

045 _- _

04242* G175 _({042G42

.~JS3.G 32'G29

!L,7

G31030G29

G26G26

[AI--S LeviG3iG30

_ G29

_ G26

FIG. 1. RNA sequencing gels of 5'-32P-labeled tRNAft deriva-tives. Lanes G, partial RNase T1 digest; lanes A, partial RNase U2 di-gest; lanes A+U, partial RNase Phy M digest; lanes U+C, partial B.cereus RNase digest. The region of each sequence corresponding to theanticodon nucleotides is enclosed in a box. The numbering of nucleo-tides in tRNAfht is based on the system adopted in ref. 12. The bandcorresponding to C35 in tRNAfc§ is fainit due to much more rapidcleavage at U36.

in the anticodon loop are shown in Fig. 1. The sequencing gelsconfirm that the structure of each synthetic tRNA derivativecorresponds to the predicted structure. All of the data indicatethat the CAG derivative has the same sequence as tRNAMetexcept for the U--G base substitution in the anticodon, andwe, therefore, conclude that the abnormal migration of this de-rivative on polyacrylamide gels is due. to an altered confor-mation under the partial denaturing conditions of the electro-phoresis.

Aminoacylation of tRNAfmet Derivatives. The kinetic pa-rameters for aminoacylation of tRNATAU were compared withthose obtained with intact tRNAfmet and a control sample thatwas isolated from a denaturing polyacrylamide gel in parallelwith the synthesized tRNAs. Intact tRNAfmet was aminoacyl-ated with a Km of 0.8 A.M and a VmX, of 3 jumol/min per mg.Different preparations of control tRNANet isolated from thegel and tRNAAgu had the same Km but a 20-40% lower Vm,.We conclude that a contaminant derived from the gel has a smalleffect on the rate of aminoacylationlby Met-tRNA synthetase;however, different preparations of the synthesized tRNAs gavevery similar results, indicating that the effect is relatively con-stant for different samples.

Aminoacylation of the synthesized tRNA derivatives con-taining anticodon loop modifications was examined at a varietyof Met-tRNA synthetase concentrations (Fig. 2). Under con-

Table 1. Aminoacylation of tRNAfmet derivatives containing basesubstitutions in the anticodon

Synthesized Moles of methionine acceptedanticodon per mole of tRNA in 30 mintsequence* 0.4 nM 4 nM 40 nMCAUCAUA (+1)CCUCUUCUACAGCACCA(-1)CACA (+1)C(-2)ACU

1.00.320.130.01

6758 Biochemistry: Schulman and Pelka

Table 2. Initial rates of aminoacylation of tRNAfMt derivativesat 40 nM tRNA

tRNA*tRNAfMettRNAfMet (gel)tCAUCAUACCUCUUCUACAGCACCACACUUAUAAUGAU

Moles Met-tRNA permole Met-tRNA

synthetase per min28.4522.8022.151.59

4.0 x 10-12.6 x 10-22.0 x 10-21.7 x 10-21.2 x 10-30.5 x 10-3106

* The oligonucleotide inserted in the anticodon loop of synthesizedtRNAfMet derivatives is indicated.

t Control sample isolated from a denaturing polyacrylamide gel in par-allel with the synthesized tRNAfmet derivatives.

t Only one sample of this derivative has been assayed.

ative order of aminoacylation activity as seen for the purifiedtRNAfmet derivatives was found. Heat treatment of isolatedproducts in the presence of Mg2" under conditions used to re-nature other tRNAs also failed to increase the levels of me-thionine acceptor activity.The apparent rates of aminoacylation shown in Table 2 have

not been corrected for any effect of anticodon base substitu-tions on the rate of enzymatic deacylation by Met-tRNA syn-thetase. Alterations in the relative rates of the forward and re-verse reactions catalyzed by aminoacyl-tRNA synthetases arenormally characterized by observation of a low plateau level ofproduct, the final value of which is dependent on the initialenzyme concentration (13). Aminoacylation of the anticodon-substituted tRNAfmet derivatives does not reach a prematureplateau value, however, but continues to show a slow increasein the amount of product formed with time in prolonged in-cubations (Fig. 2). This suggests that the major effect of thesesubstitutions is to reduce the rate of the forward reaction ratherthan to increase the rate of enzymatic deacylation. The relativerates shown in Table 2 are therefore believed to accurately re-flect the relative effects of alterations at different positions ofthe anticodon on aminoacylation, although the absolute forwardrates may be underestimated.The most defective tRNA derivatives were examined for their

ability to compete with normal tRNAfmet during aminoacyla-tion under conditions of limiting enzyme concentration. A 10-fold excess of wobble base substituted tRNAs, tRNAfm'! ortRNAfCMet, or a 5-fold excess of tRNAfCAkC or tRNAfAG had nodetectable effect on the rate of aminoacylation of control tRNAsamples. Due to a shortage of the synthesized tRNAs, higherconcentrations were not tested.

DISCUSSION

The data presented in this paper and in previous publications(3, 4) indicate a crucial role for the anticodon nucleotides oftRNAelet for recognition by Met-tRNA synthetase. Aminoacyl-ation is reduced to below levels of experimental detection bybase substitutions in the wobble position, and lesser, althoughstill dramatic, effects result from structural changes at the other

two positions of the anticodon. Base substitutions could alterinteraction of tRNAfmet with Met-tRNA synthetase by elimi-nation of essential ligands, by steric or electrostatic interfer-ence with binding, or by introducing conformational changesthat alter access of the enzyme to its normal binding sites onthe tRNA. While we cannot entirely exclude the possibility thatthe synthesized tRNA derivatives have a conformation signif-icantly different from that of native tRNA let, this seems anunlikely explanation for the wide range of effects of anticodonbase substitutions on aminoacylation rates. In addition, we havepreviously examined the conformation of a similar tRNAfmetderivative containing uridine in the wobble position by high-resolution NMR spectroscopy. Chemical deamination of thewobble cytidine to uridine eliminated methionine acceptor ac-tivity but produced no detectable loss of the secondary or ter-tiary N.H hydrogen bonds found in the native structure (2).The available data also indicate that small changes in the ge-

ometry of the anticodon loop, such as those resulting from py-rimidine-purine substitutions or changes in the inherent basestacking properties of different nucleotides, cannot account forthe large effects of anticodon base substitutions on aminoacyl-ation rates. Enlargement of the anticodon loop by one nucleo-tide, which would be expected to cause a similar small shift inthe position of anticodon nucleotides, has a relatively small ef-fect on the aminoacylation rate. On the other hand, tRNAfmethaving the permuted anticodon sequence ACU, with a normalpurine-to-pyrimidine ratio and anticodon loop size, has no de-tectable activity. These results indicate that while the spatialarrangement of the loop plays some role, the major factor inrecognition of tRNAfmet by Met-tRNA synthetase involves in-teraction of the enzyme with specific functional groups on an-ticodon nucleotides. The most stringent requirement is for acytidine base in the wobble position. The inability of other basesto substitute even weakly for cytidine suggests that there maybe a more extensive interaction of the enzyme with this nu-cleotide than with other anticodon bases, possibly involving si-multaneous binding to the ribophosphate backbone and to func-tional groups on the pyrimidine ring.The magnitude of the effect of sequence changes at the oth-

er two positions of the anticodon depends on the structure ofthe substituted base, those nucleotides having specific function-al groups in common with the normal nucleotides being tolerat-ed better than those that do not. Thus, cytidine inserted inthe middle position of the anticodon contains the structure-N=C(NH2)-, found at positions 1 and 6 of the normaladenosine nucleotide, and yields a tRNAfmet derivative that isamino-acylated at a rate 15-fold higher than that of the cor-responding uridine derivative lacking the basic ring nitrogenand exocyclic amino group. Similarly, guanosine contains the-NH-CO- structure found at the 3 and 4 positions of theuridine ring and is a 14-fold more effective substitute for uri-dine at the 3' end of the anticodon than is cytidine. The greaterthan 104-fold reduction in the rate of aminoacylation of thetRNAf9A'c derivative suggests that there may be a negative ef-fect of cytidine at this site in addition to the loss of some pos-itive interaction(s) with the normal uridine nucleotide.Few quantitative data on the relationship between anticodon

sequences and aminoacylation activity are available for othertRNAs. Bruce and Uhlenbeck have carried out a systematic studyof the effects of base substitutions at each position of the an-ticodon on aminoacylation of yeast tRNAPhe (14, 15). A 10-foldincrease in Km resulted from alterations in the wobble base, andsmaller effects were produced by changes at other anticodonpositions. These results indicate that no specific anticodon baseis essential for recognition of tRNAPhe by its cognate aminoacyl-tRNA synthetase; however, there may be a cumulative effect

Proc. Natl. Acad. Sci. USA 80 (1983)

Dow

nloa

ded

by g

uest

on

June

14,

202

1

Proc. Natl. Acad. Sci. USA 80 (1983) 6759

on binding of interaction of Phe-tRNA synthetase with a num-ber of sites in this and other regions. On the other hand, largeeffects on aminoacylation rates have been shown to result fromsingle base changes in the anticodon sequences of beef tRNATrP,yeast tRNAVaI, E. coli tRNAGIY, and E. coli tRNA"'9 (16-19),and a C-to-U change in the middle position of the anticodon ofE. coli tRNATrP increases the rate of mischarging of this tRNAwith glutamine by five to six orders of magnitude (20). The ef-fect of the same anticodon alteration on aminoacylation oftRNATrP by the cognate Trp-tRNA synthetase is significant butmuch smaller, a 59-fold increase in Km and 6.7-fold reductionin Vmax being observed (20). The available data therefore in-dicate a wide range of effects of anticodon base changes on ratesof acylation and misacylation by aminoacyl-tRNA synthetases,with E. coli Met-tRNA synthetase showing stringent structuralrequirements for recognition in this region. The lack of a cy-tidine base in the wobble position of a noncognate tRNA wouldbe sufficient to exclude its aminoacylation by Met-tRNA syn-thetase. Thus, it is possible that only a small number of otherunique sites are required to ensure complete accuracy in sub-strate selection by this enzyme. Chemical modification and en-zymatic excision data (1) have revealed that alterations in manydifferent regions of the structure of tRNAfmet have little or noeffect on methionine acceptor activity; however, modificationof a specific G residue in the variable loop or on the 3' side ofthe acceptor stem drastically reduces the rate of aminoacylation(21, 22). Small changes in Km or Vm., also result from structuralalterations at other specific sites in the acceptor stem region (1,2, 23).

The mechanism by which wobble base substitutions reduceaminoacylation rates of tRNAfmet by five or six orders of mag-nitude remains unclear. It seems improbable that the smallnumber of hydrogen bonds that could uniquely be made withcytidine could provide a sufficient difference in binding to ac-count for the observed rate differences or that interference byother nucleotides in the wobble position could be sufficient toexclude binding. It is more plausible that interaction of Met-tRNA synthetase with this cytidine leads to a rearrangement ofthe complex that increases the number of binding contacts ina cooperative fashion or that induces a long-range conforma-tional change that greatly enhances the rate of the catalytic step.Evidence for such an anticodon-dependent rearrangement hasbeen reported for other tRNA-synthetase complexes. Excisionof the Y base from the anticodon loop of yeast tRNA"he has beenshown to result in an altered interaction of the 3' terminus ofthe tRNA with Phe-tRNA synthetase (24), and tRNAT`P has beenshown to induce a conformational change in beef Trp-tRNAsynthetase that is dependent on the sequence of the anticodonloop (25).

The structural requirements for interaction of Met-tRNAsynthetase with the anticodon bases of its tRNA substrates ap-pear to involve highly localized sites on each nucleotide base,because the enzyme is relatively unaffected by significantstructural alterations only a few atoms removed from the ap-parent contact sites. For example, saturation of the 5,6 doublebond of the uridine at the 3' end of the anticodon with bisulfite

ion has only a small effect on the rate of aminoacylation oftRNAfmet (5) and the presence or absence of an acetyl moietyon the exocyclic amino group of the wobble base has no effecton the interaction of Met-tRNA synthetase with tRNAmet (26).Construction of tRNA derivatives containing base analogs moreclosely related to the natural nucleotide should assist in dif-ferentiating between positive and negative interactions of Met-tRNA synthetase with specific bases at each site of the anti-codon and allow further resolution of the number and locationof essential functional groups in this region.

This work was supported by American Cancer Society Grant NP-19.Partial salary support for L. H.S. was provided by National Cancer In-stitute Grant P30 CA1333 0-10.

1. Schulman, L. H. & Pelka, H. (1977) Biochemistry 16, 42564265.2. Schulman, L. H. (1979) in Transfer RNA: Structure, Properties

and Recognition, eds. Schimmel, P. R., S611, D. & Abelson, J. N.(Cold Spring Harbor Laboratory, Cold Spring Harbor, NY), pp.311-324.

3. Schulman, L. H., Pelka, H. & Susani, M. (1983) Nucleic Acids Res.11, 1439-1455.

4. Schulman, L. H. & Goddard, J. P. (1973)J. Biol. Chem. 248, 1341-1345.

5. Schulman, L. H. & Pelka, H. (1977)J. Biol. Chem. 252, 814-819.6. Higgins, N. P., Geballe, A. P., Snopek, T. J., Sugino, A. & Coz-

zarelli, N. R. (1977) Nucleic Acids Res. 4, 3175-3186.7. Thach, R. E. & Doty, P. (1965) Science 147, 1310-1311.8. England, T. E. & Uhlenbeck, 0. C. (1978) Biochemistry 17, 2069-

2076.9. Cameron, V., Soltis, D. & Uhlenbeck, 0. C. (1978) Nucleic Acids

Res. 5, 825-833.10. Krug, M. & Uhlenbeck, 0. C. (1982) Biochemistry 21, 1858-1864.11. Donis-Keller, H., Maxam, A. M. & Gilbert, W (1977) Nucleic Acids

Res. 4, 2527-2538.12. Gauss, D. H., Gruiter, F. & Sprinzl, M. (1979) in Transfer RNA:

Structure, Properties, and Recognition, eds. Schimmel, P. R., S611,D. & Abelson, J. N. (Cold Spring Harbor Laboratory, Cold SpringHarbor, NY), pp. 518-519.

13. Bonnet, J. & Ebel, J. P. (1972) Eur. J. Biochem. 31, 335-344.14. Bruce, A. G. & Uhlenbeck, 0. C. (1982) Biochemistry 21, 3921-

3926.15. Bruce, A. G. & Uhlenbeck, 0. C. (1982) Biochemistry 21, 855-

861.16. Scheinker, V. Sh., Beresten, S. F., Mashkova, T. D., Mazo, A.

M. & Kisselev, L. L. (1981) FEBS Lett. 132, 349-352.17. Chambers, R. W., Aoyasi, S., Furukawa, Y., Zawadzka, H. &

Bhanot, 0. S. (1973) J. Biol. Chem. 248, 5549-5551.18. Squires, C. & Carbon, J. (1971) Nature (London) New Biol. 233,

274-277.19. Chakraburtty, K. (1975) Nucleic Acids Res. 2, 1793-1804.20. Yarus, M., Knowlton, R. & Soll, L. (1977) in NucleicAcid-Protein

Recognition, ed. Vogel, H. J. (Academic, New York), pp. 391-408.21. Schulman, L. H. (1971)J. Mol. Biol. 58, 117-131.22. Schulman, L. H. (1972) Proc. Natl. Acad. Sci. USA 69, 3594-3597.23. Uemura, H., Imai, M., Ohtsuka, E., Ikehara, M. & S611, D. (1982)

Nucleic Acids Res. 10, 6531-6539.24. Lefevre, J. F., Bacha, H., Renaud, M., Ehrlich, R., Gangloff, J.,

von der Haar, F. & Remy, P. (1981) Eur. J. Biochem. 117, 439-447.

25. Beresten, S. F., Sheinker, V. Sh., Bolotina, I. A., Nurbekov, M.K., Mashkova, T. D., Avdonina, T. A. & Kiselev, L. L. (1981) Mol.Biol. (Engl. Trans.) 15, 805-815.

26. Stem, L. & Schulman, L. H. (1977)J. Biol. Chem. 252, 6403-6408.

Biochemistry: Schulman and Pelka

Dow

nloa

ded

by g

uest

on

June

14,

202

1