JOURNAL OF BACTERIOLOGY, Feb. 1974, p. 351-359Copyright © 1974 American Society for Microbiology

Vol. 117, No. 2Printed in U.S.A.

Three Different Missense Suppressor Mutations Affecting thetRNAGGGGlY Species of Escherichia coliC. W. HILL, GABRIELE COMBRIATO, AND WILLIAM DOLPH

Department of Biological Chemistry, The Pennsylvania State University, Hershey Medical Center, Hershey,Pennsylvania 17033

Received for publication 5 October

The Escherichia coli suppressor mutation, supT, has been shown to cause aC a U substitution in the middle position of the tRNAGGGGlY anticodon. This isthe same tRNA species that is altered by the glyUsuAGA mutation studied pre-viously. This finding indicates that the supT mutant tRNA reads the glutamicacid codon, GAG. The supT suppressor has also been converted to a new sup-pressor, called glyUsuG AA, which will suppress the GAA mutation, trpA46. Thein vivo suppression efficiencies of each of these three missense suppressors hasbeen measured and are as follows: gIyUSUAGA, 3.6%; supT, 1.6%; andglYyUSUGAA,0.4%. Mistranslation by these mutant glycine tRNA species has no adverse af-fects on cell growth since cultures possessing the suppressors grow as fast as

cells without. The supT tRNA species can be observed as a peak in the profile ofglycyl-tRNA fractionated on a RPC-5 chromatographic column, indicating thatthe mutant tRNA can be aminoacylated with reasonable efficiency. This findingcontrasts with previous findings concerning the glyUsu AGA mutant tRNA whichis not significantly aminoacylated under the same conditions.

Escherichia coli has three distinct species ofglycine transfer ribonucleic acid (tRNA) whichcan be distinguished by their chromatographicmobilities and codon responses (6, 14). Studiesof an AGA specific missense suppressor muta-tion, glyUsuAGA, showed that the tRNAGG ;GC'Yspecies was coded by a single structural gene,glyU+, which is closely linked to the lysA locus.When these observations were made, we notedthat another suppressor mutation, supT, hadbeen shown by Eggertsson and Adelberg (9) tohave a map position very similar to that foundfor the glyUsuAGA mutation. This suggested thatthe supT and glyU loci might be identical andthat supT might also affect tRNA,;GG;Y. In thiswork we report that the supT mutation doesaffect the tRNAGGGGlY species, causing a C Utransition in the anticodon. The anticodon ofthe supT mutant tRNA is CUC, indicating thatit should read the glutamic acid codon, GAG.We also report the conversion of the supTsuppressor to a suppressor specific for the otherglutamic acid codon, GAA. This third missensesuppressor derived from tRNAGUGG'Y is calledglyUsuGAA. Including the frameshift suppressorisolated by Riddle and Roth (17) from a Sal-monella typhimurium culture, this brings tofour the number of different suppressors derivedfrom this one species of tRNA.

35'

MATERIALS AND METHODSBacteria. E. coli K-12 strains used are described in

Table 1. Standard genetic symbols and abbreviationshave been used (8, 21). Strains AB2077 and AB2551have been described previously (9), as have beenCH354 and CH465 (6). Strain CH548 was prepared bymating CH354 with the Hfr-12 strain AB2077, select-ing for Met+, Pur+ recombinants, and scoring for theQther markers. The lysA marker was introduced intoCH550 and CH571 by using trimethroprim selection(20) to isolate a thyA mutant of CH548 and then usingP1 transduction to co-transduce thyA + lysA into thisderivative. The final step in the preparation of CH550involved the introduction of the glyUsuAGA mutationusing a P1 lysate of CH465. The supT mutation inCH591 was isolated in this laboratory by selecting anIlv+ revertant of CH548 after nitrosoguanidine muta-genesis; the suppressor was then transduced intoCH571 to prepare CH591. The supT mutation soisolated is indistinguishable from supTlO by suppres-sor specificity, genetic linkage, or growth rates ofsuppressed mutants. CH726 was derived from CH571by first selecting a Ti phage-resistant deletion mu-tant (10) in which the deletion included the trpoperon; then trpA46 was introduced by using a P1lysate of a trpA46 strain supplied by Charles Ya-nofsky, selecting for ability to grow on indole withouttryptophan supplementation. The glySH and glySLalleles of the glycyl-tRNA synthetase gene were intro-duced into appropriate strains as described previously(14). Strains CH757 and CH760 are single lysogens fordefective lambda transducing phages. Isolation of

on July 3, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

HILL, COMBRIATO, AND DOLPH

TABLE 1. Genotypes oI E. coli strains

Straindesignation Genotype

AB2077 ilvD130 argHl purCl xyl-4 Hfr-12AB2551 ilvD132 thi-1 leu-6 lac-24 supTlO

Hfr-13CH354 trpA36 glySH metBCH465 trpA36 glyUsuAGA glySH argH thiCH548 trpA36 glySH xyl-4 ilvD130 argHlCH550 trpA36 IysA glyUsuAGA gLYSH xyl-4

ilvD130 argHlCH571 trpA36 lysA glySH xyl-4 ilvD130 argHiCH591 trpA36 lysA supT glySH xyl-4 ilvD130

argHlCH726 trpA46 lysA supT glySH xyl-4 ilvD130

argHlCH727 trpA461 LysA supT glySH xyl-4 ilvDI30

argHlCH729 trpA46 lysA glyUsuGAA glySH xyl-4

ilvD130 argHlCH735 trpA36 lysA gIyUsuAGA glySH xyl-4

ilvD130 argHlCH757 trpA36 lysA glySH xyl-4 ilvD130

argHl sui,11 (XcI857 S- d LysA+glyU+)

CH760 trpA461 lysA- supT glySH xyl-4ilvD130 argHl su,11- (AcI857 S- dIysA+ supT)

these defective lambda transducing phages utilizedthe lambda lysogen KS72 (18) which has the lambdaprophage integrated near lysA. The glYUSUAGA Sup-pressor was introduced into KS72, and isolation ofXdlysA gIyUsuAGA transducing phages was accordingto procedures classically used for Xdg isolation. OtherglyU alleles were then placed on the phage byrecombination. These procedures will be describedfully in a later publication. The supT alleles presentin CH760 are identical to that in CH591.Mating and mutagenesis procedures. Transduc-

tion with phage Plkc and Hfr mating were carried outas described previously (13). Agar plates used ingenetic procedures contained 1.5% agar, minimal-citrate medium (22), 0.5% glucose, and appropriatesupplementation of amino acids and other factors asnecessary. When introduction of a suppressor allelewas being selected directly, efficient recovery ofrecombinants or mutants was aided by supplementa-tion of a trace amount (0.1 g/ml) of the limitingamino acid. When ilvD130 supT cultures were grownwithout isoleucine and valine supplementation, leu-cine (50 Asg/ml) was always added to enhance thegrowth rate. Ultraviolet mutagenesis was carried outby spreading 108 cells on a selective plate and thenirradiating the plate for 10 s under a 15-W GeneralElectric germicidal lamp placed 46 cm from the plate.2-Aminopurine mutagenesis was accomplished byspreading 108 cells on a selective plate, followed by theaddition of a drop of 2-aminopurine solution (40mg/ml) to a sterile 12.7-mm filter disk placed on theplate; revertants appeared in a halo around the disk.

Procedures for tRNA isolation and sequencing.The low-phosphate medium used for growing culturesfor tRNA preparations has been described previously(12). When induction of the lysogenic strains was notdesired, growth was at 34 C throughout. To inducelysogenic cultures thermally, the cultures were ini-tially grown at 34 C until the optical density mea-sured at 590 nm reached 1.2; at this point an equalvolume of medium preheated to 50 C was added, andthe culture was shaken at 42 C for 15 min. Thetemperature was then reduced to 39 C, and the cellswere harvested 150 min after thermal induction.When ("'Piphosphate labeling was desired, up to 10mCi of carrier-free [32P]phosphate was added to 200ml of culture 60 min after thermal induction. Proce-dures for tRNA isolation and RPC-5 reversed-phasechromatography have been described previously (12).32P-labeled supTtRNA was purified in two ways. Thefirst involved RPC-5 column chromatography of gly-cyl-tRNA, followed by rechromatography of thepooled material from the first column after phenoxy-acetylation; this procedure was identical to the onedescribed previously for the purification of the wild-type tRNAGGG Gly (12). The second method utilized thetwo-dimensional acrylamide gel system described byIkemura and Dahlberg (15) as the first step in thepurification. In this case, tRNA was isolated fromCH760 cells which had been labeled with [32P]phos-phate between 60 and 150 min after thermal induc-tion. Upon fractionation by this two-dimensionalacrylamide gel procedure, supT tRNA could be iden-tified as a relatively discrete spot on the autoradio-gram. The appropriate area was eluted from the geland chromatographed on a RPC-5 column (0.9 by 45cm) with a 400-ml 0.5 to 0.6 M NaCl gradient in 0.01M MgClr0.01 M sodium acetate (pH 4.5)-0.005 M2-mercaptoethanol for elution. The supT tRNA waseluted as a sharp peak which was more than 90%isotopically pure as judged by the absence of con-taminating spots on the fingerprints. Most of theprocedures for sequencing are from the protocols ofBarrell (2), and all have been described before (12).U,-ribonuclease (U2-RNase) from Ustilago, which isspecific for the 3'-phosphodiester linkage of A residuesof T1 RNase fragments, was a gift of H. Okazaki of theSankyo Co. U2 RNase digestion of spot tx (seeResults) was with 10 sliters of 0.2 U/ml in 0.05 Msodium acetate-0.002 M ethylenediaminetetraacetate(pH 4.5) containing 0.1 mg of bovine serum albuminper ml for 4 h at 37 C. This prolonged incubation wasnecessary to obtain complete cleavage between the A-residues of the C-U-U-C-U-C-A-A-G- oligonucleotide(2). However, under these conditions theC-U-U-C-U-C-A- product was often cleaved further toproduce C-U-U-C-U-C-, and occasionally this second-ary product was obtained predominantly.

RESULTSGenetic relationship of the glyUsuAGA and

supT mutations. The similar linkage relation-ships of gIyUsuAGA (14) and supT (9) to the lysAlocus suggested that both mutations mightaffect the same locus and could possibly repre-

352 J. BAcTUOL.

on July 3, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

MISSENSE SUPPRESSOR MUTANTS

sent independent isolations of the same sup-pressor. The gIYUSUAGA and supT mutations arenot independent isolations of the same muta-tion, however, since the specificities of thesuppression are different, as shown by thefollowing observations. The gIyUsuAGA mutationwas isolated according to its ability to suppressthe trpA36 mutation (14) which has the AGAcodon at the mutant site (24). The supT muta-tion was isolated according to its specificity forthe ilvD130 mutation. Therefore if supT andglyUsu AGA are identical mutations, both trpA36ilvD130 supT and trpA36 ilvD130 glyUsuAGAstrains should be phenotypically Trp+, Ilv+.However, when such strains were constructed,the trpA36 ilvD130 g1YUSUAGA strain was Trp+,Ilv-, whereas the trpA36 ilvD130 supT strainwas Trp-, Ilv+, indicating that glyUsuAGA doesnot suppress ilvD130 and supT does not sup-press trpA36.Recombinational experiments (Table 2)

showed that, although these two suppressors arenot identical, they are very closely linked andquite possibly allelic. In these experiments, thegeneralized transducing phage P1 was used totransfer the lysA region from a lysA + su. donorto a trpA36 ilvD130 lysA suy recipient, selectingfor Lys + transductants and scoring for thesuppressor mutation present in the recombi-nant. As indicated in Table 2, almost all recom-binants possessed either the donor or recipientsuppressor mutation. None of the recombinantscarried both suppressors, and only 1 of a total of339 tested had neither. It is concluded thatthere is less than 1% recombination betweensupT and gIyUsuAGA. Since 1% recombinationoccurs between mutations separated by roughly160 base pairs (23), it is concluded that supT

TABLE 2. Recombination between glyUsu and supTa

RecombinationDonor Recipient gly. supT Both Nei-

U.U1supTBt ther

lysA+glyUsu lysA- supT 156 72 0 1lysA + supT lysA - glyUsu 28 82 0 0

aIl the first mating, strain CH465 was donor andCH591 was recipient; in the second experiment,AB2551 was donor and CH550 was recipient. Com-plete genotypes are listed in Table 1. P1 transductionwas performed as indicated in Materials andMethods. Lys+ transductants were selected on agarplates supplemented with glucose, arginine, trypto-phan, isoleucine, valine, and leucine. The presence ofthe glyUsu and supT alleles was determined bytesting the Trp and Ilv phenotypes, respectively.

b Suppressor allele of Lys+ transductant.

and glyUsuAGA are closely linked and possiblyallelic.

Effect of the supT mutation on tRNAGGG Gly.We have shown previously that the glYUSUAGAmutation leads to an alteration of tRNAGGGG1Yin both codon recognition (6) and interactionwith the glycyl-tRNA synthetase (5). If supT isallelic to glyUsuAGA, the tRNAGGGGlY speciesshould also be altered in supT cultures. Such analteration can be observed in tRNA prepara-tions fractionated on a RPC-5 reverse-phasechromatographic column. For these experi-ments and the primary sequence determinationexperiments to be described below, we em-ployed strains which are lysogenic for a defec-tive lambda transducing phage carrying thelysA glyU region of the chromosome. In theexperiment described in Fig. 1, [14C]glycyl-tRNA from strain CH757 (glyU+IXdglyU+) wasmixed with ['H]glycyl-tRNA from strain CH760(supT/XdsupT), and the mixture was fraction-ated on a RPC-5 column. Previous work (12) hasshown that the [14C]glycyl-tRNA in fractions 85to 91 is tRNAGGGGly From the results of Fig. 1, itis seen that this peak of tRNAGGG Gly is missingin the 9H-labeled preparation from the supTculture. The [14C]glycyl-tRNA material in thepeak centering in fraction 78 and that in thepeak centering in fraction 97 is tRNAGGA/GGIY(ref. 14 and personal communication from JohnCarbon). There is only one tRNAGGA/GGlY struc-

' 500 1

Ii

50 100 150 200

Fraction numb*rFIG. 1. Comparison of glycyt-tRNA profiles of

supT and wild-type cultures. The tRNA preparationswere aminoacylated with either [3H]- or [14C]glycine,mixed, and applied to the RPC-5 reversed-phasecolumn. Elution was with a 400-mI linear gradient of0.5 to 0.6 M NaCl in 0.01 M sodium acetate (pH4.5)-0.01 M MgCI2-0-005 M 2-mercaptoethanol.CH757 (glyU+) tRNA aminoacylated with [14C]gly-cine (0); CH760 (supT) tRNA aminoacylated with['HJglycine (0).

353VOL. 117, 1974

on July 3, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

HILL, COMBRIATO, AND DOLPH

tural gene (14) which produces the material inboth of these peaks. Fraction 78 tRNAGGA/GGlYpossibly differs from fraction 97 tRNAGGA/GGlyby a modification. The relevant point for thisstudy is that the [3H]glycyl-tRNA correspond-ing to the ["4C]glycyl-tRNA in fraction 97 isslightly skewed to the left in Fig. 1. Theexperiment to be described in Fig. 2 suggeststhat this lack of coincidence of the [3H]- and[I4C]glycyl-tRNA in fractions 94 to 100 is due tothe presence of supT specific tRNA in the3H-labeled preparation.

Induction of a prophage which carries atRNA structural gene should result in a largeamplification of that tRNA species (12, 19).Accordingly, strain CH760 was heat induced asdescribed in Materials and Methods, and thetRNA was isolated, aminoacylated with[i3H]glycine, and chromatographed on theRPC-5 column. [14C]glycyl-tRNA from CH760which had not been heat induced served as acontrol. The results of this experiment (Fig. 2)show that this procedure results in a largeincrease in glycine acceptor in the region (frac-tions 110 to 114) just behind the positionnormally occupied by wild-type tRNAGGG lyand just ahead of the position occupied by theslower component of tRNAGG;AGG'Y. From theseresults, we conclude that like the gIYUSUAGAmutation, the supT mutation causes an altera-tion in the tRNAGGGGlY species. However, theresults of Fig. 2 show that the effect of the supTmutation on tRNAGGGG lY is quite different fromthe effect of the glyUsuAGA mutation. Theg1YUSUAGA mutant tRNA has a greatly reduced

1000

1500

1000

500

50 100 150 200

Fractlon nu,nber

FIG. 2. Effect on the glycyl-tRNA profile of ther-mal induction of a supT transducing phage lysogen.Procedures were as indicated in Fig. 1. Transfer RNAisolated from an uninduced culture of CH760 andaminoacylated with ["4CJglycine (0); tRNA isolatedfrom a thermally induced culture of CH760 andaminoacylated with [3Hlglycine (-).

affinity for the glycyl-tRNA synthetase andcannot be aminoacylated sufficiently for it toappear as an observable peak in an experimentsuch as that shown in Fig. 2 (5). The supTmutation does not have such an extreme effecton the acceptor activity of the mutant tRNAsince the supT tRNA can be observed as asubstantial peak in Fig. 2.

Alteration of the tRNAGGGGY primary squence by supT. The complete primary se-quence of the E. coli tRNAGGGly species hasbeen determined recently (12). Given that thesupT mutation alters this tRNA species, it wasof obvious interest to determine the alterationin the primary sequence caused by the supTmutation. [32P]phosphate-labeled supT tRNAwas isolated from a thermally induced cultureof CH760 as indicated in Materials andMethods. Portions of the purified supT tRNAwere then digested with pancreatic RNase Aand with T,RNase, and the digests fractionatedby two-dimensional electrophoresis. The pan-creatic RNase fingerprint of the supT tRNAproved to be identical to the fingerprint ofwild-type tRNAGGGGIY. Identity of all of thepancreatic RNase fragments was confirmed bybase compositional analysis after NaOH diges-tion and by digestion with T, RNase.The T, RNase fingerprint of the supTtRNA,

however, showed a significant change from thewild-type species (Fig. 3). No material wasfound at the position normally occupied byfragment tll from the wild-type molecule, andthe position occupied by wild-type fragment tl2had approximately double the normal intensity.This material is designated tx in Fig. 3. Allother T, RNase fragments appeared in thenormal positions and were identical to thewild-type as judged by base composition andcharacterization by pancreatic RNase diges-tion. In addition, the modified bases in t8(D-A-G-) and tlO (T-I-C-G-) were present. Thechange in fragment tll was particularly inter-esting in that fragment tll contains the antico-don of the wild-type tRNA and has the sequenceC-U-U-C-C-C-A-A-G-, where the C-C-C se-quence is the anticodon. Fragment tl2 has thesequence C-U-C-U-A-U-A-C-G-. Digestion withpancreatic RNase of the spot tx produced molaryields of A-A-G-, A-U-, A-C- and G- in additionto U- and C- mononucleotides. The presence oftwo different G-containing fragments in thisdigest shows that spot tx is definitely a mixtureof two different T, RNase fragments. The A-U-,A-C-, and G- are normally found after digestionof the wild-type tl2 fragment, whereas A-A-G-is found in digests of wild-type tll. This resultsuggested that the supT mutation alters frag-

354 J. BACarmoL.

on July 3, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

MISSENSE SUPPRESSOR MUTANTS

ment tll such that it has a mobility identical tot12 in the fingerprinting procedure. It might benoted here that the base composition of wild-type tll is C-,, U-2, A-2, G-, whereas that of t12iS C-3, U-3, A2-, G-. Therefore, a C - Utransition in fragment tll would give it a basecomposition identical to t12, and such a changewould cause these fragments to have identicalelectrophoretic mobilities.

Further characterization depended on separa-tion of the apparently mutated tll fragmentfrom t12, a difficult task if these nonanucleo-tides have identical base compositions. Theapproach used was to digest the spot tx with U2RNase which is specific for the 3'-phosphodi-ester linkage of A residues of T, RNase frag-ments. Fractionation of the U2-RNase digest byelectrophoresis in 7% formic acid on di-ethylaminoethyl-cellulose produced two largefragments, y and z. Occasionally, fragment z wascleaved secondarily under the conditions usedfor U2 RNase digestion (see Materials andMethods). These fragments were characterizedby polynucleotide phosphorylase digestion (12),and by derivatization with the carbodiimidereagent followed by pancreatic RNase digestion(2) (Table 3). U2 RNase fragment y was shownto be C-U-C-U-A-, and is the product expectedafter U2 RNase digestion of wild-type fragmenttl2. Pancreatic RNase digestion of derivatizedfragment z produced A-, C-, U-C-, and U-U-C-(Table 3). Polynucleotide phosphorylase diges-tion of this fragment showed that C-U-U is thetrinucleoside diphosphate at the 5'-terminus ofthis large U2 RNase digestion product of tx.From this information, we conclude that thesequence of this oligonucleotide isC-U-U-C-U-C-A-, and that spot tx containsboth normal t12 and a new fragmentC-U-U-C-U-C-A-A-G-. This new fragment dif-fers from wild-type tll by a single C -- Utransition at the position representing the mid-dle base of the anticodon. The complete se-quence of the wild-type tRNAGGG ly and themutant supT tRNA is shown in Fig. 4. ThesupT tRNA has CUC as its anticodon, predict-ing that supT tRNA is specific for the glutamicacid codon, GAG.The finding that supT tRNA differs from the

wild type by a single C - U transition isconsistent with our observation that hydroxyla-mine, a mutagen known to cause GC - ATtransitions (11) is very effective in generatingsupT mutations.Conversion of supT to a GAA suppressor.

Yanofsky and co-workers (23, 24) deduced thatthe trpA46 mutation involves a gly (GGA) -b glu(GAA) transition at position 211 in the trpA

Cellulose Acetate pH 3 5 4

U.0

4

ILJ

CM

FIG. 3. Tracing of a fingerprint of a T, RNasedigest of 32P-labeled supT tRNA. Spots ti-t1O, tl3,and t14 are in positions identical to those found forwild-type tRNAGGGGY (12). The area occupied by tllin digests of wild-type tRNA is indicated by thebroken line. The position indicated by x is thatnormally occupied by t12, but is of approximatelytwice the normal intensity in the digest of the supTtRNA. B indicates the position of the blue dye markerxylene cyanole FF.

gene. Normally trpA46 cultures are only ob-served to revert to Trp+ by a mutation at thetrpA locus (24). In view of the above result thatthe supT mutant tRNA has a CUC anticodonand by inference reads the GAG codon, a trpA46supT double mutant should have two additionalpossibilities for reversion to Trp + besides simplestructural gene reversions. Both of these involvegenetic suppression and are illustrated in Fig. 5.First of all, the trpA46 mutant codon GAAcould be changed to GAG; in this case, the supTsuppressor tRNA would occasionally insert gly-cine at this codon, producing some active tryp-tophan A protein. Secondly, the anticodon ofthe supT tRNA could be changed to UUC,creating a new suppressor species that wouldread the trpA46 codon GAA as glycine, and alsoproducing some active tryptophan A protein.

AUUCCCUUCG 13UUCAAUGr ) ,

'II1

T*CG @

& UAG

& DAG

pG Q

pG 0B

@ AACG

G CCCG

G AG

Q CG

O CUCCAoN

355VOL. 117, 1974

on July 3, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

HILL, COMBRIATO, AND DOLPH

TABLE 3. Characterization of U2 RNase digestion products of spot tx

Snalte venom

U,RNase Electro- Alkaline hydrolysis" phosphodies-product | Treatment Product phoretic teram hy-procluct mobilitya drolysis"

C- A- U- -C -U

y None C-U-C-U-A- 0.58c 2.0 1 1.9y RNase A digestion of carbo- U-C- -0.55d 1 0.9

diimide derivative U-A- -0.35d 1 1.0y Polynucleotide phosphorylase C-U-C 1.4e 1 0.8 1 1.1

C-U 1.8e +z None C-U-U-C-U-C-A- 0.21Cz RNase A digestion of carbo- U-U-C- -0.84d 1 1.7

diimide derivative U-C- -0.55d 1 1.0z Polynucleotide phosphorylase C-U-U 1.4e 1 0.9 +

C-U 1.8e + +

aElectrophoretic mobilities are in relation to mobility of the blue dye marker xylene cyanol FF."After hydrolysis the products were separated by electrophoresis on paper at pH 3.5, and determined

quantitatively by liquid scintillation counting. Results are expressed as relative molar yields of nucleotide.c Electrophoresis was on diethylaminoethyl-cellulose in 7% formic acid.d Electrophoresis was on 3MM paper at pH 3.5.eElectrophoresis was on diethylaminoethyl-cellulose at pH 3.5.

A70HCCU

PG . C 70

sG. CC * G6s

15 * G.*C 60 UA A ,OAU UUCCC

CUUG .....*ego. GAGGGGAAC CA45 5o T

DA 25~G AUC

A * U 40

30C * 6U AU AC Ci"C

+ -TTrp Trp Trp

U A

G55

*C

'C C'FIG. 4. Cloverleaf model of tRNAGGG'GY indicating

the C _ U substitution caused by the supT mutationin the middle of the anticodon.

With this rationale in mind, a trpA46 lysAsupT ilvD130 strain, CH726, was constructed(Materials and Methods). CH726 is phenotypi-cally Trp-. This strain was mutagenized byvarious mutagens, and Trp+ revertants were

trpA codon GAG * GAA GAA

suppressor anticodon Cuc Cuc puUcFIG. 5. Reversion of a trpA46 supT culture to

Trp+. Mode I depicts the change of the trpA46 mutantcodon, GAA, to GAG such that it can be read asglycine by the supT suppressor tRNA. Mode IIdepicts the change of the supT tRNA anticodon fromCUC to UUC such that it will read the trpA46 mutantcodon GAA, as glycine. Neither of these modes oftrpA46 reversion is available to a trpA46 culturewhich does not possess the supT mutation.

selected and purified. Whether or not a glyU-derived suppressor was necessary for expressionof the Trp+ phenotype of these revertants couldbe tested by using a P1 lysate of a lysA+ glyU+strain to transduce the revertants to Lys+ andscoring for loss of the Trp+ phenotype amongthe transductants. If a lysA linked suppressorwas necessary for the revertant to be Trp+-, amajority of these Lys+ transductants wouldbecome simultaneously Trp-.When ultraviolet irradiation was used: for

mutagenesis, the majority of revertants vereextremely slowly growing without tryptopwnsupplementation, but grew normally with tryp-tophan supplementation. Four of five examples-of these slowly growing mutants were shown tohave a lysA-linked suppressor necessary for-theTrp+ phenotype, since 60 to 80% of the Lys+transductants become Trp- when transduced

UGG

J. BAcT=ioL.356

on July 3, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

MISSENSE SUPPRESSOR MUTANTS

with a lysA+ glyU+ donor. Restoration of supTto these Trp - derivatives by additional P1transduction did not result in restoration of theTrp+ phenotype. Therefore, it is concluded thatthese UV-induced Trp+ revertants contain a

new GAA-specific suppressor. Additional exper-

iments showed that this new suppressor was

specific for the trpA46 allele and would notsuppress trpA36 which has an AGA codon at themutant site. One of these new suppressedrevertants, CH729, was retained as the primarysource of this new suppressor, designatedglYUSUGAA.

Spontaneous or 2-aminopurine mutagenesisof the trpA46 supT strain, CH726, resulted inmany slowly growing Trp+ revertants whichcould also be shown to depend on suppressionfor the Trp+ phenotype. In this case, however,the Trp+ phenotype could be restored by thereintroduction of the supT suppressor to theLys+, Trp- transductants obtained after infec-tion with a P1 lysate of a lysA+ glyU+ donor. Inthese cases, it is concluded that a change at thetrpA locus had produced a trpA allele which issuppressible by the supT mutation. These mostlikely occurred by a change at the trpA46mutant codon such that it was changed fromGAA to GAG. One of these spontaneous revert-ants, CH727, was retained as the primarysource of this new trpA allele which we havedesignated as trpA461.Suppression efficiencies of. the glyU-

derived suppressors. We now have three differ-ent missense suppressors derived fromtRNAGG GGY. We also have available three trpAmutations each suppressible by a different one

of these suppressors; these are trpA36 (AGA),trpA46 (GAA), and trpA461 (GAG). The trpA36and trpA46 mutations are known to affect thesame glycine residue in the tryptophan A pro-tein (23) and, from the mode of its origin,trpA461 almost certainly affects this glycineresidue also. With these mutants available, it ispossible to examine suppression efficiencies ofeach of these suppressors. The experiments are

based on the observation by Brody and Ya-nofsky (4) that, whereas only active tryptophanA protein can participate in the conversion ofindole glycerol phosphate to indole (reaction 1),all tryptophan A protein (A-CRM) is active inthe conversion of indole to tryptophan (reaction2). Therefore, the quotient of activity in reac-tion 1 to activity in reaction 2 gives a measure ofthe suppression efficiency. The measurement ofthe suppression efficiencies for the three sup-pressors is given in Table 4. These results showthat glyUsuAGA is more efficient than supT

TABLE 4. Suppression efficiencies of glyUsuppressorsa

Effi-A A-CRM ciency

Genotype (U/mi) A(-CRM o

pression(%)

trpA36 glySH gIYUSUAGA ... 2.8 170 3.6trpA36 glYSL gIyUSUAGA * * * 2.4 438 1.2trpA461 glySH supT ...... 1.6 187 1.6trpA461 glySL supT ....... 1.5 643 0.4trpA46 glySH glYUSUGAA ... 2.0 823 0.4trpA . ................... 12.6 23.2 (100)

a Enzyme assays and the calculation of the efficien-cies of suppression are as described by Brody and Ya-nofsky (4).

which is in turn more efficient than glYUSUGAA-This finding is consistent with the observationthat when no tryptophan is supplied aglyUsuAGA-suppressed culture grows faster thanone suppressed by supT which in turn growsfaster than one suppressed by glYUSUGAA.Two different alleles for the glycyl-tRNA

synthetase have been observed to occur invarious E. coli K-12 cultures. One allele, glySH,produces a product more active than the otherallele, glySL (3, 16). Previously, it was observedthat the suppression efficiency of the glYUSUAGAmutation depended on which allele was present(14). The results in Table 4 show that theefficiency of supT also depends on the glySallele, supT glyS&`tailtures having a highersuppression efficiency than supT glYSL cul-tures. Dependence of the suppression efficiencyon the glyS allele indicates that the glycyl-tRNA synthetase is responsible for the amino-acylation of the suppressor tRNA in vivo. Whena P1 lysate of a glyUsuGAA culture was used toinfect a trpA46gIySL culture, no Trp+ transduc-tants were obtained, whereas the same P1 lysatecould transduce glyUsuGAA to a trpA46 glySHculture. The simplest explanation for this resultis that glyUsuGAA suppression is also dependenton the glyS allele and that when the glySL alleleis present suppression is so weak as to notproduce an observable Trp+ phenotype.None of the tRNAGGGGlY derived suppressors

has a detrimental effect on the cell, at least asfar as growth rates are concerned. Doublingtimes were measured for wild-type and mutantstrains growing in synthetic medium supple-mented with glucose, lysine, isoleucine, valine,leucine, arginine, and tryptophan, and werefound to be glyU+ (50 min), glyUsuAGA (52 min),supT (50 min), and glyUsuGAA (53 min).

357VOL. 117, 1974

on July 3, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

HILL, COMBRIATO, AND DOLPH

DISCUSSIONThe preceding has shown that the

tRNAGGGG 'ly species is affected by three differentmissense suppressor mutations. The gIYUSUAGAmutant tRNA reads AGA as glycine; evidencefor this is derived from in vitro experimentsutilizing a synthetic poly A-G message (14), aswell as from the specificity of the gIYUSUAGAsuppressor for a known AGA mutant (trpA36)(24). The demonstration in this report that thesupT mutation changes the anticodon oftRNAGGGG lY from CCC to CUC predicts that thesupT mutant tRNA should read the GAGglutamic acid codon as glycine. This supposi-tion is supported by the observation that thetrpA46 allele (GAA) can be readily converted toa supT suppressible allele after spontaneous or2-aminopurine mutagenesis. Evidence for thecodon specificity of the third suppressor muta-tion, gIYUSUGAA, is derived from its specificityfor a known GAA mutation, trpA46 (24). At thepresent time, it seems probable that the conver-sion of supT to gIYUSUGAA involves a secondchange in the anticodon of tRNAGGGGY, so thatthe gIYUSUGAA anticodon is UUC. We hope totest this hypothesis by isolating and sequencingthe gIYUSUGAA mutant tRNA.

According to the rules of wobble (7), a GAAsuppressor should also suppress GAG mutants,given that the amino acid inserted (in this caseglycine) is acceptable. Particularly, we wouldexpect glyUsuGAA to suppress the mutationsilvD130 and trpA461. We have constructed atrpA461 ilvD130 glyUsuGAA strain. This culturewill not produce observable colonies in 6 days ifisoleucine and valine supplementation is omit-ted from the plates and produces only tinycolonies in 6 days if tryptophan supplementa-tion is omitted. When tryptophan, isoleucine,and valine are all supplied, large colonies areproduced in 2 days. From these rather qualita-tive observations, it appears that gIyUsuGAA hasa very slight response to the GAG codon, butsubstantially less than for the GAA codon, sincetrpA46 glyUsuGAA cultures grow substantiallyfaster than the trpA461 gIyUsuGAA cultures ontryptophan-deficient plates. We conclude thatglyUsuGAA can read GAG by wobble, but thatthe resulting suppression is extremely weak (seebelow.)

All three of the glyU-derived suppressor mu-tations alter the chromatographic profile ofglycyl-tRNA. As reported previously (14), thegIYUSUAGA mutation leads to a disappearanceof the tRNAGGGGY species without the reappear-ance of a mutant species elsewhere in theprofile. The reason that the glyUsuAGA mutanttRNA is not seen anywhere in the profile is due

to the fact that it is aminoacylated at a muchreduced rate by the glycyl-tRNA synthetase (5).On the other hand, the supT mutant tRNA doesappear in the chromatographic profiles (frac-tions 110 to 114 of Fig. 2), indicating that themutant tRNA can be aminoacylated under theconditions used (a 10-fold excess of the glycyl-tRNA synthetase). The amount of supT tRNAapparent in the profiles is, however, less thanthe amount of wild-type tRNAGGGGlY producedin a wild-type strain. For example, the strainsCH757 and CH760 contain two copies of glyU+and supT genes, respectively. Yet as seen in Fig.1, a large peak of tRNAGGGG'y is present in theCH757 profile, but there is only enough supTtRNA in the CH760 profile to produce a slightskewing of one of the tRNAGGA/GGlY peaks. Thebasis for this quantitative difference is un-known. When glycyl-tRNA from a glYUSUGAAstrain is subjected to RPC-5 column chromatog-raphy (unpublished data), no material is ob-seved at the position normally occupied bytRNAGGGGlY, nor are any new peaks of glycyl-tRNA observed. There are several possible ex-planations for this observation. First thegiYUSUGAA tRNA species may coincide with oneof the other peaks of glycyl-tRNA and beobscured by it. A second possibility is that thepresence of the mutation may interfere with thebiosynthesis of the mutant tRNA. A thirdpossibility is that, like the glYUSUAGA mutanttRNA, the gIYUSUGAA mutant tRNA may beinefficiently aminoacylated under the condi-tions used and, therefore, not appear in theprofiles of glycyl-tRNA. Clearly, further ex-perimentation is necessary to clarify the effectof the glyUsuGAA mutation.The observation that glyUsuAGA suppression

is more efficient that supT suppression (Table4) is somewhat surprising since the degree ofaminoacylation of supT tRNA by glycyl-tRNAsynthetase in vitro is substantially higher thanthat observed for glyUsuAGA tRNA. However,suppression efficiencies of missense suppressorsshould depend on several factors other than theefficiency of aminoacylation of the suppressortRNA. Among these factors are the amount ofsuppressor tRNA made, the efficiency withwhich the suppressor tRNA associates with themessenger RNA-ribosome complex, and theamount of normal competing aminoacyl-tRNApresent which reads the suppressed codon. An-derson (1) has suggested that there is a verylimited amount of tRNAAGAArg in E. coli andthat the supply of arginyl-tRNAAGAArg mightcontrol the rate of protein synthesis in cistronscontaining the AGA codon. According to Ander-son's argument, when arginine must be inserted

358 J. BACEIOL.

on July 3, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

MISSENSE SUPPRESSOR MUTANTS

at maximal rate, one of the other five argininecodons would be used. If this is true, thelimitation on the supply of tRNAAGA Arg mighthelp explain the relatively high efficiency ofglyUsAGA suppression. It might be pointed outthat another AGA suppressor, glyTsUAGA, is ableto achieve 30% suppression of the trpA36 muta-tion (14). On the other hand, the relativelyinefficient suppression of trpA461 by supT,despite the observable acceptor activity of supTtRNA, might be related to the fact that GAG isone of only two glutamic acid codons. Sincethere are only two glutamic acid codons, onemight expect the cell to translate GAG withhigh efficiency, and this normal capacity wouldtend to lower the supT suppression efficiency.Mutation of the glyU tRNA species has no

apparent detrimental effect on the cell, sinceall three of the suppressor mutants studiedhere grow as fast as the glyU+ parent when themedium is supplemented to the point thatsuppression is not required for growth. Thisfinding indicates that neither the loss of thenormal GGG translating capacity oftRNAGGGG ly nor the mistranslation of othermRNA molecules caused by the suppressors isdetrimental. Presumably the tRNAGGA GGy spe-cies remaining is able to translate both GGAand GGG with sufficient efficiency. The non-essentiality of tRNAGGG Gly contrasts with theeffects of tRNAGGA,G GlYloss of tRNAGGA,GGlyby mutation deprives the cell of GGA translat-ing capacity, resulting in severe, detrimentaleffects (14).

Future experiments have as their aim thedetermination of the actual change caused inthe tRNAGGG Gly species by the glyUsu A GA andgIyUsuGAA mutations and the effect of each ofthe three suppressor mutations on the tRNAbiosynthesis, aminoacylation, and binding tomessenger RNA-ribosome complexes.

ACKNOWLEDGMENTSWe thank Howard Douds for his participation in certain

phases of this work. We are indebted to Barbara Bachmanand Charles Yanofsky for supplying some of the bacterialcultures utilized.

This work was supported by Public Health Service re-search grant GM-16329 from the National Institute of Gen-eral Medical Sciences. W.D. gratefully acknowledges supportby Public Health Service research support grant S01-RR05680.

LITERATURE CITED1. Anderson, W. F. 1969. The effect of tRNA concentration

on the rate of protein synthesis. Proc. Nat. Acad. Sci.U.S.A. 62:566-573.

2. Barrell, B. G. 1971. Fractionation and sequence analysisof radioactive nucleotides, p. 751-779. In G. L. Cantoniand D. R. Davies (ed.), Procedures in nucleic acid re-search, vol. 2. Harper and Row, New York.

3. Bock, A., and F. C. Neidhardt. 1966. Location of thestructural gene for glycyl ribonucleic acid synthetase

by means of a strain of Escherichia coli possessing anunusual enzyme. Z. Vererbungsl. 98:187-192.

4. Brody, S., and C. Yanofsky. 1963. Suppressor genealteration of protein primary structure. Proc. Nat.Acad. Sci. U.S.A. 50:9-16.

5. Carbon, J., C. Squires, and C. W. Hill. 1969. Geneticallyaltered tRNA GY subspecies in E. coli. Cold SpringHarbor Symp. Quant. Biol. 34:505-512.

6. Carbon, J., C. Squires, and C. W. Hill. 1970. Glycinetransfer RNA of Escherichia coli. II. Impaired GGA-recognition in strains containing a genetically alteredtransfer RNA; reversal by a secondary suppressormutation. J. Mol. Biol. 52:571-584.

7. Crick, F. H. C. 1966. Codon-anticodon pairing: thewobble hypothesis. J. Mol. Biol. 19:548-555.

8. Demerec, M., E. A. Adelberg, A. J. Clark, and P. E.Hartman. 1966. A proposal for a uniform nomenclaturein bacterial genetics. Genetics 54:61-76.

9. Eggertsson, G., and E. A. Adelberg. 1965. Map positionsand specificities of suppressor mutations in Escherichiacoli K-12. Genetics 52:319-340.

10. Franklin, N. C., W. F. Dove, and C. Yanofsky. 1965. Thelinear insertion of a prophage into the chromosome ofE. coli shown by deletion mapping. Biochem. Biophys.Res. Commun. 18:910-923.

11. Freese, E., E. Bautz, and E. B. Freese. 1961. Thechemical and mutagenic specificity of hydroxylamine.Proc. Nat. Acad. Sci. U.S.A. 47:845-855.

12. Hill, C. W., G. Combriato, W. Steinhart, D. L. Riddle,and J. Carbon. 1973. The nucleotide sequence of theGGG-specific glycine transfer RNA of Escherichia coliand of Salmonella typhimurium. J. Biol. Chem.248:4252-4262.

13. Hill, C. W., J. Foulds, L. Soll, and P. Berg. 1969.Instability of a missense suppressor resulting from aduplication of genetic material. J. Mol. Biol.39:563-581.

14. Hill, C. W., C. Squires, and J. Carbon. 1970. Glycinetransfer RNA of Escherichia coli. I. Structural genes fortwo glycine tRNA species. J. Mol. Biol. 52:557-569.

15. Ikemura, T., and J. E. Dahlberg. 1973. Small ribonucleicacids of Escherichia coli. I. Characterization by polyac-rylamide gel electrophoresis and fingerprint analysis.J. Biol. Chem. 248:5024-5032.

16. Ostrem, D. L., and P. Berg. 1970. Glycyl-tRNA synthe-tase: an oligomeric protein containing dissimilar sub-units. Proc. Nat. Acad. Sci. U.S.A. 67:1967-1974.

17. Riddle, D. L., and J. R. Roth. 1972. Frameshift suppres-sors. III. Effects of suppressor mutations on transferRNA. J. Mol. Biol. 66:495-506.

18. Shimada, K., R. A. Weisberg, and M. E. Gottesman.1972. Prophage lambda of unusual chromosomal loca-tions. I. Location of the secondary attachment sites andthe properties of the lysogens. J. Mol. Biol.63:483-503.

19. Squires, C., B. Konrad, J. Kirschbaum, and J. Carbon.1973. Three adjacent transfer RNA genes in Esche-richia coli. Proc. Nat. Acad. Sci. U.S.A. 70:438-441.

20. Stacey, K. A., and E. Simson. 1965. Improved method forthe isolation of thymine-requiring mutants of Esche-richia coli. J. Bacteriol. 90:554-555.

21. Taylor, A. L., and C. D. Trotter. 1972. Linkage map ofEscherichia coli strain K-12. Bacteriol. Rev.36:504-524.

22. Vogel, H. J., and D. M. Bonner. 1956. Acetyl-ornithinaseof Escherichia coli: partial purification and someproperties. J. Biol. Chem. 218:97-106.

23. Yanofsky, C., and V. Horn. 1972. Tryptophan synthetasea chain positions affected by mutations near the endsof the genetic map of trpA of Escherichia coli. J. Biol.Chem. 247:4494-4498.

24. Yanofsky, C., J. Ito, and V. Horn. 1966. Amino acidreplacements and the genetic code. Cold Spring HarborSymp. Quant. Biol. 31:151-162.

359VOL. 117, 1974

on July 3, 2020 by guesthttp://jb.asm

.org/D

ownloaded from