Proc. NatL Acad. Sci. USAVol. 79, pp. 5813-5817, October 1982Biochemistry

An amber suppressor tRNA gene derived by site-specificmutagenesis: Cloning and function in mammalian cells

(Xenopus laevis tyrosine tRNA gene/mammalian cell suppressor/adenovirus 2-simian virus 40 hybrid/M13 cloning/in vitro suppression)

FRANK A. LASKI*, RAMAMOORTHY BELAGAJEt, UTTAM L. RAJBHANDARY*, AND PHILLIP A. SHARP**Center for Cancer Research and Department of Biology, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139; and tLilly Research Laboratories,Indianapolis, Indiana 46285

Communicated by S. E. Luria, June 14, 1982

ABSTRACT We describe the synthesis, cloning, expression,and in vivo function of a suppressor tRNA gene in mammaliancells. By using "primer-directed mutagenesis" on a Xenopus laevistyrosine tRNA gene cloned into the recombinant single-strandphage M13mp5, we have generated an amber suppressor tRNAgene that has a nucleotide change-GTA -* CTA-in the anti-codon sequence. The suppressor (Su) tRNA gene was introducedinto monkey kidney cells (CV-1) by using simian virus 40 (SV40)DNA as vector (SV40-tRNAT"` Su'). CV-1 cells infected with viruscontaining the mutant, but not the wild-type, tRNA gene producea functional amber suppressor tRNA as indicated by suppressionof amber mutations in co-infecting adenovirus serotype 2-SV40hybrids. Further evidence that suppression of these amber mu-tations is tRNA mediated was derived by isolation of total tRNAfrom CV-1 cells infected with the SV40-tRNATYr (Su') recombi-nant and its use in demonstration of read through of an ambercodon during in vitro translation of tobacco mosaic virus RNA inreticulocyte extracts. Interestingly, the amplification of an ambersuppressor gene in CV-1 cells does not interfere with SV40 pro-duction, suggesting that suppression of amber codons may not bevery deleterious to mammalian cell metabolism.

Isolation ofEscherichia coli cells with a wide variety ofnonsensesuppressor (Su) mutations was crucial for much ofthe early workon genetic analysis of chain-terminating mutations in bacterialand viral genomes (for review, see ref. 1). Suppressors of allthree nonsense codons, UAG (amber), UAA (ochre), and UGAhave been characterized in E. coli. Among eukaryotic organ-isms, tRNA-mediated suppression has been studied extensivelyonly in yeast. Quite recently, an amber suppressor has beenidentified in Caenorhabditis elegans and shown to involvechanges in a tRNA (ref. 2; N. Willis, R. Gesteland, K. Bolten,and B. Waterston, personal communication).The availability of mammalian cell lines with well-defined

tRNA suppressors would greatly facilitate genetic analysis ofnonsense mutants in mammalian cells and, in particular, animalviral genomes. As a step toward this objective, we report herethe generation of an amber suppressor derived from Xenopuslaevis tyrosine tRNA gene. We have introduced this amber sup-pressor tRNA gene into monkey kidney cells (CV-1) by usingsimian virus 40 (SV40) DNA as vector. We show that the tRNAgene is expressed and produces a functional tRNA that sup-presses-amber mutations both in vivo and in vitro. In particular,the in vivo activity of the suppressor tRNA gene is shown bythe fact that previous infection ofCV-1 cells with the SV40-tRNAsuppressor recombinant efficiently suppresses the amber non-sense mutation in the 30,000-dalton fusion protein of two mu-

tants of adenovirus serotype 2(Ad2)-SV40 hybrid (Ad2+NDl)isolated by Grodzicker et aL (3, 4).

MATERIALS AND METHODSConstruction of M13-tT (Su') by Site-Specific Mutagenesis.

The oligonucleotide primer C-A-C-C-T-A-G-A-G-T-C-C-T wassynthesized by using a modified phosphotriester method (ref.5 and references cited therein) and purified as a single peak byHPLC. A 5'-terminal phosphate was added with T4 polynu-cleotide kinase. Viral DNA of M13-tT was purified by alkalinesucrose sedimentation (6).

M13-tT virion DNA (9.0 jig) and the oligonucleotide primer(1.0 tkg) were mixed in 11 p1L of hybridization buffer (8 mMTrisHCl, pH 7.4/20 mM NaCl/4 mM MgCl2/0.4 mM 2-mer-captoethanol). After 1 hr on ice, 5 ,ul of the DNA solution wasadded to 8 ,ul of 30 mM Tris HCl,'pH 7.4/15 mM MgCl2/1.5mM 2-mercaptoethanol/1.245 mM dNTP/0.6 mM ATP. After10 min at 37°C, 2 pul of DNA polymerase I (Klenow fragment;New England BioLabs; 1,100 units/ml) and 1.5 ,u1 ofT4 ligase(New England BioLabs, 4 x 105 units/ml) were added. After2 hr at 370C, the reaction mixture was loaded onto an agarosegel containing ethidium bromide at 0.6 ug/ml. Covalentlyclosed circles were purified from the agarose gel and used fortransfection. The sequence of M13-tT (Su+) DNA was deter-mined by using the Maxam and Gilbert procedure (7).

Cloning of tRNAT), Genes (wt and Su') into SV40. M13-tTand M13-tT (Su') DNAs were cleaved with Kpn I [ in the SV40sequences at map unit (m.u.) 0.72] and EcoRI (in the M13mp5polylinker adjacent to the SV40 HindIll site at m.u. 0.86) andthe DNA fragments containing the tRNA gene were purifiedby gel electrophoresis. These fragments were cloned by usinga pBR322-SV40 recombinant (joined at the BamHI sites) thathad been cleaved with Kpn I (at SV40 m.u. 0.72) and EcoRI(at SV40 m.u. 1.0). The tRNA gene was thus placed in the lateregion of SV40 in the same polarity as late mRNA transcription.The resulting clones were designated pSV-tT-2 and pSV-tT-2(Su+) (see Fig. 3A).

Production of Recombinant SV40 Virus Stocks and Purifi-cation of DNA and RNA. E. coli plasmids and strains, SV40virus stocks, mammalian cell lines, and plasmid DNA prepa-rations for transfection ofCV-1 cells were as described (8). Virusstocks were prepared by excision of the viral DNAs frompBR322 sequences by digestion with the appropriate restrictionendonuclease, circularized by ligase, and used to transfect CV-1 cells by using DEAE-dextran (9). Viral infections and labeling

Abbreviations: SV40, simian virus 40; Ad2, adenovirus serotype 2; moi,multiplicity of infection; mu., map unit(s); Su, suppressor; bp, basepair(s); TMV, tobacco mosaic virus; wt, wild type; T antigen, SV40 largetumor antigen.

5813

The publication costs ofthis 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

Oct

ober

1, 2

020

Proc. Natl. Acad. Sci. USA 79 (1982)

conditions with 32P04 were as described (8). 32P-Labeled DNAand RNA were purified by Hirt extraction (10) followed by onephenol extraction, two phenol/chloroform extractions, and twoethanol precipitations.

Immunoprecipitations. Immunoprecipitations were by theStaphylococcus aureus method (11) after previous extensive in-cubation with nonimmune serum. Cell line 412 (12), which pro-duces a monoclonal antibody against the COOH-terminus ofSV40 large tumor antigen (T antigen), was a gift of E. G. Gur-ney, University of Utah.

RESULTSSite-Specific Mutagenesis of X. laevis tRNATYr Gene Cloned

into M13mp5. The sequence of a X. laevis gene for tRNATYr(13) is shown in Fig. 1. A 263-base-pair (bp) fragment containingthis gene has previously been cloned into the'late region ofSV40and its expression has been studied after infection of monkeycells (8). The X. laevis gene is efficiently transcribed and thetRNA is processed and modified.The basic scheme for site-specific mutagenesis (6, 15, 16) in-

volves the use of a single-stranded circular DNA (M134tT; see

a

3I9,b~8b8X;62b~~2-CHMnfI t c .

(1) (k) TC.G eA .

r G A AT

T CTCGG * - -

GT AGAG CG

G .

A-114b214

G t tG

A (®AGC H/inf IT

C G G C C |T

G C T GGT CCT T

GT AG

A TG *CG -CA* T CGAGC

C A-A CTT G-GT AG TA GTG

b-AGGACTGTAGGTG- template

OHTCJCTGAGATCCACp5 primer

1//nf I

FIG. 1. (a) Structure of M13-tT. M13-tT contains a 263-bp Hae II/Hha I fragment of X. laevis DNA flanked by SV40 DNA sequencescloned into the HindIII site of M13-mp5. The tRNATYr" gene segmentwas subcloned from the recombinant pSV-tT described by Laski et al.(8) into the M13-mp5 vector. M13-mp5 is a derivative of M13-mp2 (14)that has a HindIII/EcoRI polylinker cloned into the M13-mp2 EcoRIsite. Descriptions of the recombination sites are as follows: 1, theHindIII site of SV40 (m.u. 0.65) with the HindImI site of the M13-mp5polylinker; 2, the Hha I site of X. kaevis DNA with the Hha I site ofSV40 (m.u. 0.72); 3, the Hae II site of the X. laevis DNA was joined tothe Hae II site of SV40 (m.u. 0.82); 4, the HindIII site of SV40 (m.u.0.86) was joined to the Hindm site of the M13-mp5 polylinker. Thesequence of the coding region of the mature tRNATYr gene and its in-tervening sequence (13) as well as the HinfI sites that immediatelyflank the anticodon region are shown. (b) Sequence of the oligonu-cleotide primer. The oligonucleotide used for site-specific mutagenesisis complementary to the tRNAT gene except for one mismatch. In-corporation of the mutation results in an amber suppressing anticodon(CTA) and creates a HinfI site.

Fig. 1), carrying the X. laevis tRNATYr gene, as a template forthe synthesis of covalently closed circular duplex DNA in thepresence of an oligonucleotide as primer. The 13-base-long oli-gonucleotide primer used contains a mismatch in the anticodonregion of the tRNA gene (Fig. 1). To ensure specific primingby the oligonucleotide, the M13-tT DNA used was freed ofDNA fragments by alkaline sucrose gradient centrifugation.Conditions for specific priming were established by using 5'-32P-labeled oligonucleotide primer and analyzing the elongationproduct'by digestion with HindIII followed by gel electropho-resis. As expected, more than 50% of the radioactivitywas foundin the small 873-bp HindIII fragment that contains the tRNATYrgene (data not shown).The oligonucleotide primed product should contain both in-

complete and complete duplex DNAs. Because incomplete du-plex DNAs could contribute to high wt background, covalentlyclosed duplex DNAs were purified by electrophoresis in anagarose gel in the presence of ethidium bromide at 0.6 ug/mland then used to transfect E. coli. Forty independent phageplaques were screened for the desired site-specific mutation.Phage from each plaque were used to infect E. coli, replicativeform DNAs were isolated from the infected cells and analyzedby digestion with HinfI followed by electrophoresis on a 5%polyacrylamide gel. The desired mutation generates a newHinfI site within the tRNA gene (Fig. 1). Two clones containingthe new HinfI site were further analyzed and both were con-firmed to possess the desired mutation by DNA sequence anal-ysis (7). An autoradiograph of a sequencing gel covering theanticodon regions ofthe mutant [tRNATYr (Su+)] and wt tRNATYrgenes is shown in Fig. 2.

Cloning of the Mutant tRNA Gene in CV-1 Cells by UsingSV40. DNA segments from the M13 recombinants containingthe tRNA genes were excised and cloned into the late regionof SV40 DNA (Fig. 3a). Identical constructs were prepared withthe wt X. laevis tRNA gene and the tRNA (Su') gene. Virusstocks of the 'SV40 recombinants were prepared by co-transfec-tion of CV-1 cells with a SV40 rat preproinsulin recombinant,SV-rINS-7, that has a deletion/substitution in the early generegion of SV40 (Fig. 3a; unpublished results). Approximately1 wk after co-transfection, the CV-1 cells underwent a cyto-pathic degeneration. The resulting virus stocks were harvested

a b(q -L -$

G-c-A-G-

G-T...G-4G-;

L-C--

A-G-G-;A-G- I

FIG. 2. Maxam-Gilbert sequence analysis gel of the wt tRNATyrgene (a) and the tRNATlr (Su+) gene (b). Purified replicative formDNAof M13-tT and M13-tT(Su') were end labeled with 2P at the Bgl I siteby using polynucleotide kinase. This site is 37 bp from the anticodonregion of the genes. The anticodon triplets are bracketed.

5"Q14 Biochemistry: Laski et al.

I

3

Dow

nloa

ded

by g

uest

on

Oct

ober

1, 2

020

Proc. Natl. Acad. Sci. USA 79 (1982) 5815

b 1 2 3

A* -

A,_

B,C*-

a b c

SV-rtNS

.95SV40

SV-tT-2

F A* C E

,'i , ,' L AlOrL'.99 A .32 D 42 B ..5 C .6E .95

//

A' D B a-5 ~ . . _."1

FIG. 3. (a) Hind] restriction map of SV-rINS-7, SV40, and SV-tT-2. SV-tT-2 is identical to SV-tT-2 (Su+) except for a point mutation inthe anticodon region of the tRNATYr gene. Solid lines represent posi-tions of the Hindus sites in SV40 and m.u. values of the sites areshown. SV-rINS-7 has an insertion of sequences from rat preproinsulingene and a deletion of sequences from the early region ofSV40 creatingthe new restriction fragment A*. SV-tT-2 has an insertion/deletion inthe late region of SV40 creating the new restriction fragment C* anda deletion of late SV40 sequences (m.u. 0.86-1.0) creating the new re-striction fragment A'. pSV-tT-2 (Su+) and pSV-tT-2 have pBR322 re-combined into the BamHl (m.u. 0.14) site of SV-tT-2 (Su+) and SV-tT-2, respectively. The origin of replication (ori) and the direction of earlyand late transcription are shown. Restriction sites used in constructionof the recombinant: A, Hha I; o, Taq I; A, Bcl I; *, EcoRI. (b) Autora-diogram ofHindM digests of viral DNAs from SV40-infected (moi, 3)cells (lane 1), SV40-tT-2 (Su+)/SV-rINS-7-infected cells (lane 2), andSV-tT-2/SV-rINS-7-infected cells (lane 3). 32P-Labeled viral DNA wasdigested with Hindll and the fragments were separated by electro-phoresis on a 1.4% agarose gel. Bands are labeled according to a.

and labeled SV-tT-2/SV-rINS-7 and SV-tT-2 (Su')/SV-rINS-7.Amplification of the amber suppressor gene in CV-1 cells

does not interfere with SV40 reproduction (Fig. 3b). A 10-cmplate of CV-1 cells infected with 0.1 ml of either SV-tT-2/SV-rINS-7 or SV-tT-2 (Su')/SV-rINS-7 virus yielded comparableamounts of viral DNA 48 hr after infection [extracted by themethod of Hirt (10)]. Digestion of the 32P-labeled DNA withHindIII showed that each stock contained the two expectedcomplementing viral genomes with no detectable contaminat-ing wt SV40 DNA or variant DNA [Fig. 3b, lane 1 (SV40), lane2 (SV-tT-2 (Su+)/SV-rINS-7), and lane 3 (SV-tT-2/SV-rINS-7)].Note the absence of fragment A in lanes 2 and 3. In addition,the amount of labeled viral DNA extracted from recombinantinfected cells was comparable to that found 48 hr after infectionwith SV40 at a multiplicity ofinfection (moi) of 10 (Fig. 3b; com-pare the -intensities of common bands in lane 1 with those inlanes 2 and 3).

Expression of the X. laevis tRNAT)R Gene in CV-1 Cells.Expression of the tRNATyr(Su) gene in monkey cells was ex-amined by labeling SV-tT-2 (Su+)/SV-rINS-7-infected cellswith 32P04 46-52 hr after infection and analyzing total RNA foroverproduction of tRNAs. Cells infected with an equal volumeof SV-tT-2/SV-rINS-7 were analyzed in parallel. In both cases,the 32P-labeled RNA was resolved by electrophoresis in a 7.5%acrylamide/8.3 M urea gel. Fig. 4a shows that (i) the wt tRNATYrgene was efficiently expressed, yielding a prominent band mi-grating as expected for a 76-nucleotide-long tRNA (8), and (ii)the tRNAT~r(Su') gene was expressed at about 1/5 the level ofthe wt gene but also yielded a. prominent band migrating as

expected for a 76-nucleotide-long tRNA. Radioactive tRNAfrom these two bands was purified and analyzed. The RNaseT1 maps of the wt tRNATyr and the tRNATYr(Su+) are shown inFig. 4 b and c. As anticipated, an oligonucleotide A-C-U-C-f-A-m1Gp is found in the RNase T1 digest oftRNATYr(Su') (whichresults from the conversion of guanosine to cytidine in the an-ticodon) and the two olihonucleotides present in the wt tRNATyr(A-C-U-Gp and O-A-m Gp) are missing.

B&

x

#AmtGp

ACUGP *

4i

ACr*JO4An

FIG. 4. (a) Autoradiogram of [32P]RNA purified from CV-1 cellsand analyzed on a 7.5% polyacrylamide/8.3 M urea gel. Cells were in-fected with 0.1 ml of SV40 (10' plaque-forming units/ml), with 0.1 mlof the SV-tT-2(Su+)/SV-rINS-7 virus stock, or with 0.1 ml of the SV-tT-2/SV-rINS-7 virus stock. Arrow, location of mature tRNATYr. (b andc) RNase T1 maps of 32P-labeled tRNATYr purified from SV-tT-2/SV-rINS-7- and SV-tT-2(Su+)/SV-rINS-7-infected cells. RadioactivetRNATYr was eluted from bands indicated by the arrow in a, furtherpurified by electrophoresis in a 10% polyacrylamide/4 M urea gel, andanalyzed by RNase T1 mapping (8).

The reduced level of the tRNATYr from the Su+ mutant ismost likely due to inefficient splicing. This decrease in effi-ciency of splicing has also been observed in the splicing of invitro-transcribed precursors ofthe tRNATYr (Su') gene (unpub-lished results).

In Vivo Suppression of Amber Codons by Su+ tRNA inMammalian Cells. Grodzicker et aL (3, 4) have described theisolation and characterization of three nonsense mutants ofAd2-SV40 hybrid virus, Ad2+ND1. These mutants provide a

convenient means of testing the suppressor activity in vivo ofthe X. laevis tRNATYr(Su+). The hybrid virus Ad2+ND1 codesfor a 30,000-dalton fusion protein whose carboxyl terminus isderived from the SV40 T antigen. The presence of this regionofthe SV40T antigen gene allows Ad2+ND1 to grow in monkeycells (17). The three Ad2+ND1 mutants of Grodzicker et aL (3),which were isolated as host range mutants unable to grow inmonkey cells, are defective for synthesis in vivo of the 30,000-dalton protein but stimulate synthesis of shorter proteins. Twoof the Ad2+ND1 mutants, 140 and 162, were identified as am-ber mutants because addition of a yeast amber tRNAser sup-

pressor in an in vitro translation reaction permitted synthesisof the 30,000-dalton fusion protein (4). The other Ad2+ND1mutant, 71, was similarly identified as an ochre mutant.

CV-1 cells were infected with either SV-tT-2/SV-rINS-7 or

SV-tT-2 (Su+)/SV-rINS-7 or mock and incubated at 370C for 24hr to permit synthesis of viral encoded products, includingtRNAs. These cultures were then infected with either Ad2+ND1,Ad2+ND1-71, Ad2+ND1-140, or Ad2+ND1-162. Cells were

labeled with [3S]methionine and the SV40 T antigen-relatedproteins were immunoprecipitated by addition of a monoclonalantibody (12) that immunoprecipitates the 30,000-dalton fusionprotein. Gel analysis of the immunoprecipitated proteins isshown in Fig. 5. The 30,000-dalton fusion protein is immuno-precipitated from CV-1 cells infected with Ad2+ND1 (lane 1)but not from cells infected with the Ad2+ND1 mutants (lanes2, 3, and 4) nor from CV-1 cells previously infected with theSV40-X. laevis tRNATYr recombinant (SV-tT-2/SV-rINS-7) andthen infected with either Ad2+ND1 mutant (lanes 8, 10, and12). However, the amber mutants Ad2+ND1-140 and -162 weresuppressed in CV-1 cells previously infected with the SV40tRNATyr(Su) recombinant [SV-tT-2 (Su+)/SV-rINS-7; lanes 9and 11). That suppression is efficient is indicated by the fact that

a

Biochemistry: Laski et aL

Dow

nloa

ded

by g

uest

on

Oct

ober

1, 2

020

Proc. NatL Acad. Sci. USA 79 (1982)

su" Su Su SU Su+ Su sue SuN D 71 140 162 71 71 i40 140 162 162

1 2 3 4 5 6 7 8 9 10 T1 12

0.75 ug 3.75 atgI Ii -TPAVSu+ Su SV40 M -Su+ Su-SV4 M Su+

2 3 4 5 6 7 8 9 10

p160-

p 1l m..0s .u....-''

p30 -

FIG. 5. Electrophoretic analysis of immunoprecipitated proteinslabeled with [35Slmethionine in CV-1 cells after infection withAd2WND1, Ad2+NDI mutants, and SV40 recombinants on a 10% poly-acrylamide gel (18, 19).. Six-centimeter dishes of CV-1 cells were in-fected with 0.1 ml of SV-tT-2 (Su+)/SV-rINS-7 or 0.1 ml of SV-tT-2/SV-rINS-7 or mock infected. Twenty-four hours later, cells were fur-ther infected withAd2+NDl, Ad2+ND1-71 (an ochre mutant),Ad2WND1-140 (an ambermutant), or Ad2WND1-162 (an amber mutant} at a moiof approximately 50. Twenty-four hours later, cells were labeled with150 pCi of [35S]methionine (800-1200 Ci/mmol 1 Ci = 3.7 x 1010becquerels; New England Nuclear) in 1 ml (total vol) for 1 hr. A totalcell extract was made by lysis of the cell monolayer in 1 ml of RIPAbuffer (18). Proteins were then immunoprecipitated with a monoclonalantibody directed against the COOH terminus of SV40 T antigen.Lanes: 1, mock infected followed by Ad2+ND1; 2, mock infected fol-lowed by Ad2WND1-71; 3, mock infected followed by Ad2WND1-140; 4,mock infected followed by-Ad2WND1-162; 5, SV-tT-2 (Su+)/SV-rINS-7 followed by mock infected; 6, SV-tT-2/SV-rINS-7 followed by mockinfected; 7, SV-tT-2 (Su+)/SV-rINS-7 followed by Ad2+ND1-71; 8, SV-tT-2/SV-rINS-7 followed by Ad2+ND1-71; 9, SV-tT-2 (Su+)/SV-rINS-7 followed by Ad2WND1-140; 10, SV-tT-2/SV-rINS-7 followed byAd2+ND1-140; 11, SV-tT-2 (Sut)/SV-rINS-7 followed by Ad2+ND1-162; 12, SV-tT-2/SVrINS-7 followed by Ad2+ND1-162.

the amount of 30,000-dalton protein synthesized was equiva-lent to that found after Ad2'ND1 infection. Suppression is spe-cific for amber mutations because synthesis ofthe 30,000-daltonfusion protein was not observed after infection with the ochremutant, Ad2+ND1-71, (lane 7).

In Vitro Suppression 'of Amber Codons by Su' tRNA :inReticulocyte Extract. To show that the in vivo. suppression ofamber codons was tRNA mediated, tRNA-enriched fractionsextracted from CV-1 cells 48 hr after infection with either SV40or the SV40 tRNA recombinants were -tested for their potencyin promoting read through' of the amber termination signal intobacco mosaic virus (TMV) RNA. Pelham (20) has previouslyshown that translation of TMV RNA in reticulocyte' extractsprimarily yields a 110,000-dalton polypeptide. Addition of ayeast amber suppressor tRNA fraction to the extract promotesread through of termination to yield a 160,000-dalton polypep-tide. As shown in Fig. 6 (lanes 1 and 6), stimulation of synthesisof the same 160,000-dalton polypeptide was observed whentRNA fractions from CV-1 cells infected with'the SV40-tRNATYr(Su+) recombinant were added to the translation extract. Thiseffect was not observed with tRNA fractions from CV-1 cellsinfected with either SV40 or the SV40tRNATYr recombinant orwith mock-infected cells (Fig. .6).

'DISCUSSIONThis paper describes the use of site-specific mutagenesis to gen-erate an amber suppressor from a X. laevis tyrosine tRNA gene

FIG. 6. Electrophoretic. analysis of protein made in vitro by usingTMV RNA as template and in the presence of crude tRNA from re-combinant SV40-infected cells. Conditions for in vitro translation werea modification of that used by Pelham (20). Rabbit reticulocyte lysatewas prepared and treated with micrococcal nuclease as described byPeiham and Jackson (21). Crude tRNA was isolated 48 hr after infec-tion from SV-tT-2/SV-rINS7-,'SV-tT-2 (Su')/SV-r1NS-7-, SV4W-, andmock-infected cells. Cell RNA.was obtained by Hirt extraction (sameprocedure as for 32P-labeled RNA) and then crude tRNA was separatedfrom high molecular weight RNAs on a DEAE-Sephacel column (RNA'was loaded in 0.1 M NaCl/5( mM Tris HCl, pH 8.0;washed with 0.2M NaCl/50 mM Tris;HCl, pH 8.0; and eluted with 1.0 M NaCl/50mMTrisHCl, pH 8.0). All reaction mixtures contained 0.25 ug of TMVRNA (except that in lane 10), 10 'Ci of [35S]methionine, 7.9 ,ul of rabbitreticulocyte lysate, and '1 jl of rat liver tRNA (0,75 mg/ml). In ad-dition, reaction mixtures contained 0.75 tg (lanes 1-4) or 3.75 pg(lanes 6-10) of crude tRNA from cells infected with SV-tT-2 (Su')/SV-rINS-7 (lanes 1, 6, and 10), SV-tT-2/SV-rINS-7 (lanes 2 and 7),SV40 (lanes 3 and 8), or mock-infected-cells (lanes 4 and 9). Lane 5: nocrude tRNA. Incubation (total vol, 11 ,l) was for 60 min at 32°C.

and the cloning, expression, and function in vivo of this sup-pressor in mammalian cells. Recently, Temple et aL (22) havealso reported the isolation ofan amber suppressortRNA derivedfrom the human lysine tRNA gene and shown it to be active insuppression when injected into Xenopus oocytes.

Infection of CV-1 cells with SV40 recombinants containingthe Su' gene produces a state fully permissive for translationof amber codons. Co-infection of these cells with mutants ofAd2+ND1 containing amber mutations in a 30,000-dalton fusionprotein yielded levels of this protein comparable with thoseobserved after infection- with an equal moi of wt Ad2'ND1.Suppression is specific for amber mutations (an ochre mutantofAd2+ND1 was not suppressed under identical conditions) andis dependent on the single nucleotide change introduced.

It has been suggested that wild type tRNATYr which containsG instead of the modified nucleoside Q (queuosine) in the firstposition of the anticodon (23) can read amber codons. If so, itis unlikely that this kind of read through is an efficient processin mammalian cells. We have shown previously that the over-produced tRNATYr in cells infected with'the SV40-wt tRNATYrcontains G in place ofQ (8). The presence of high levels of thistRNATYr species does not suppress the Ad2'ND1 amber mu-tants (Fig. 5, lanes 10 and 12).Amber codons in mammalian cells may, however, be read

at a low frequency by certain endogenous tRNAs. An ambercodon is known to separate the gag and pol genes of Moloney

5%16 Biochemistry: Laski et.d

Dow

nloa

ded

by g

uest

on

Oct

ober

1, 2

020

Proc. Natl. Acad. Sci. USA 79 (1982) 5817

leukemia virus and a 1% read through could account for theamount of the 180,000-dalton gag-pol polypeptide synthesized(24). The nature of tRNAs responsible for such a low level ofread through is yet unknown. A number of amber mutants ofviral and cellular genes have been identified: the Ad2-SV40fusion protein gene ofAd2+ND1 (4), the thymidine kinase geneof herpes simplex virus (25), the A gene of SV40 (26), the hy-poxanthine/guanine phosphoribosyltransferase gene (M. Ca-pecchi, personal communication), and the human ,/3globin gene(27). A low-level read through of these amber codons would nothave been detected during the characterization ofmost of thesemutants.

As shown here (Fig. 3), amber suppressors are probably nothighly deleterious for the viability of cells in culture. This maynot be surprising because most mammalian genes terminate inUAA or UGA. By using the same procedures, it should be pos-sible to generate amber suppressors derived from other tRNAsas well as UAA and UGA suppressors. It will be interesting totest whether the UAA and UGA suppressors are more delete-rious to cells. The availability of a variety of nonsense suppres-sors in mammalian cells should permit isolation of many newnonsense mutants in viruses. Nonsense mutants are typicallytight mutations. The isolation of many tight viral mutantsshould, in turn, make complementation and recombination as-says ofanimal viruses more definitive. Finally, it should be pos-sible to treat various regions ofa viral DNA with mutagens and,by using cells carrying suppressor genes, identify all the essen-tial genes encoded in a segment of the viral DNA.

The most immediate objective in these studies is the creationof mammalian cells permissive for translation of nonsense mu-tations. Two approaches are possible: (i) as described here, atransient state can be established by previous infection of cellswith the SV40-Su+ recombinant and (ii) a permanent permis-sive state could be established by isolation of cell lines express-ing sufficient levels of the tRNATYr(Su+) from integrated DNAsequences. Given that the lytic cycle of SV40 extends over 4days and that many other more lytic viruses form plaques within24 hr, the first approach by itself should permit isolation of am-ber mutants of many lytic viruses. Obviously, it would be pref-erable to establish permanent permissive cell lines by integra-tion of the X. laevis tRNATYr (Su') gene into a chromosome.We congratulate Norman Davidson, our friend, colleague, and

teacher, on his 65th birthday. We are grateful to Terry Grodzicker forsupplying the Ad2+ND1 mutant viruses, which were crucial in thesestudies. We thank M. Horowitz, who supplied us with the recombinantSV-rINS-7. We thank Stuart Clarkson for generously supplying us withplasmid pt201 and Lynne Corboy, Lee Gehrke, David Roth, Ihor Lem-ischka, Jean Schwarzbauer, Connie Cepko, Andrew Fire, Mia Horo-

witz, Randy Kaufman, Birgit Alzner-DeWeerd, Regina Reilly, andMargarita Siafaca for their help and advice. This work was supportedby National Institutes of Health Grant GM17151 and American CancerSociety Grant NP114 to U. L. R., by National Institutes of Health GrantPOCA26717 and National Science Foundation Grant PCM7823230 toPA.S., and partially by Center for Cancer Biology, Massachusetts In-stitute of Technology, (Core) Grant POL-CA14051.

1. Steege, D. A. & Soll, D. G. (1979) in Biological Regulation andDevelopment Gene Expression, ed. Goldberger, R. F. (Plenum,New York), Vol. I, pp. 433-475.

2. Waterston, R. (1981) Genetics 97, 307-325.3. Grodzicker, T., Lewis, J. B. & Anderson, C. W. (1976) J. Virol.

19, 559-571.4. Gesteland, R. F., Wills, N., Lewis, J. B. & Grodzicker, T. (1977)

Proc. Nati Acad. Sci. USA 74, 4567-4571.5. Narang, S. A., Hsiung, H. M. & Brousseau, R. (1979) Methods

Enzymol 68, 90-98.6. Kudo, I., Leineweber, M. & RajBhandary, U. L. (1981) Proc.

Nati Acad. Sci. USA 78, 4753-4757.7. Maxam, A. M. & Gilbert, F. (1980) Methods Enzymol 65,

499-580.8. Laski, F. A., Alzner-DeWeerd, B., RajBhandary, U. L. & Sharp,

P. A. (1982) Nucleic Acids Res., in press.9. Sompayrac, L. M. & Danna, K. J. (1981) Proc. Natl Acad. Sci.

USA 78, 7575-7578.10. Hirt, B. (1967) J. Mol Biol. 26, 365-369.11. Kessler, S. W. (1975)J. Immunol. 115, 1617-1624.12. Gurney, E. G., Harrison, R. 0. & Fenno, J. (1980)J. Virol. 34,

752-763.13. Muller, F. & Clarkson, S. G. (1980) Cell 19, 345-353.14. Gronenborn, B. & Messing, J. (1978) Nature (London) 272,

375-377.15. Hutchison, C. A., Phillips, S., Edgell, M. H., Gillam, S.,

Jahnke, P. & Smith, M. (1978)J. Biol Chem. 253, 6551-6560.16. Razin, A., Hirose, T., Itakura, K. & Riggs, A. (1978) Proc. Natl.

Acad. Sci. USA 75, 4268-4270.17. Lewis, A. M., Jr., Levine, A. S., Crumpacker, C. S., Levin, M.

J., Samaha, R. J. & Henry, P. H. (1973)J. Virol. 11, 655-664.18. Cepko, C. L., Changelian, P. S. & Sharp, P. A. (1981) Virology

110, 385-401.19. Laemmli, U. K. (1970) Nature (London) 227, 680-685.20. Pelham, H. R. (1978) Nature (London) 272, 469-471.21. Pelham, H. R. & Jackson, R. J. (1976) Eur. J. Biochem. 67,

247-256.22. Temple, G. F., Dozy, A. M., Roy, K. L. & Kan, Y. W. (1982)

Nature (London) 296, 537-540.23. Bienz, M. & Kubli, E. (1981) Nature (London) 294, 188-190.24. Philipson, L., Andersson, P., Olshevsky, U., Weinberg, R., Bal-

timore, D. & Gesteland, R. (1978) Cell 13, 189-199.25. Cremer, K. J., Bodemer, M., Summers, W. P., Summers, W.

C. & Gesteland, R. F. (1979) Proc. Natl Acad. Sci. USA 76,430-434.

26. Rawlins, D. R. & Muzyczka, N. (1980)J. Virol 36, 611-616.27. Chang, J. C. & Kan, Y. W. (1979) Proc. Natl Acad. Sci. USA 76,

2886-2889.

Biochemistry: Laski et al.

Dow

nloa

ded

by g

uest

on

Oct

ober

1, 2

020