expression of a coronavirus ribosomal frameshift signal in … expression... · 2020. 2. 1. ·...

J. Mol. Biol. (1997) 270, 360±373

Expression of a Coronavirus Ribosomal FrameshiftSignal in Escherichia coli: Influence of tRNAAnticodon Modification on Frameshifting

Ian Brierley1*, Michayla R. Meredith1, Alison J. Bloys1

and Tord G. Hagervall2

1Division of VirologyDepartment of PathologyUniversity of CambridgeTennis Court Road, CambridgeCB2 1QP, UK2Department of MicrobiologyUniversity of UmeaÊS-90187, UmeaÊ, Sweden

Present address: M. R. Meredith,UK.

Abbreviations used: RSV, Rous shuman immunode®ciency virus; Hbronchitis virus; RRL, rabbit reticultransglycosylase; MMTV, mouse m

0022±2836/97/280360±14 $25.00/0/mb

Eukaryotic ribosomal frameshift signals generally contain two elements, aheptanucleotide slippery sequence (XXXYYYN) and an RNA secondarystructure, often an RNA pseudoknot, located downstream. Frameshiftingtakes place at the slippery sequence by simultaneous slippage of tworibosome-bound tRNAs. All of the tRNAs that are predicted to decodeframeshift sites in the ribosomal A-site (XXXYYYN) possess a hypermodi-®ed base in the anticodon-loop and it is conceivable that these modi®-cations play a role in the frameshift process. To test this, we expressedslippery sequence variants of the coronavirus IBV frameshift signal instrains of Escherichia coli unable to modify fully either tRNALys ortRNAAsn. At the slippery sequences UUUAAAC and UUUAAAU (under-lined codon decoded by tRNAAsn, anticodon 50 QUU 30), frameshiftingwas very inef®cient (2 to 3%) and in strains de®cient in the biosynthesisof Q base, was increased (AAU) or decreased (AAC) only two-fold. InE. coli, therefore, hypomodi®cation of tRNAAsn had little effect on frame-shifting. The situation with the ef®cient slippery sequences UUUAAAA(15%) and UUUAAAG (40%) (underlined codon decoded by tRNALys,anticodon 50 mnm5s2UUU 30) was more complex, since the wobble baseof tRNALys is modi®ed at two positions. Of four available mutants, onlytrmE (s2UUU) had a marked in¯uence on frameshifting, increasing theef®ciency of the process at the slippery sequence UUUAAAA. No effecton frameshifting was seen in trmC1 (cmnm5s2UUU) or trmC2(nm5s2UUU) strains and only a very small reduction (at UUUAAAG)was observed in an asuE (mnm5UUU) strain. The slipperiness of tRNALys,therefore, cannot be ascribed to a single modi®cation site on the base.However, the data support a role for the amino group of the mnm5 sub-stitution in shaping the anticodon structure. Whether these conclusionscan be extended to eukaryotic translation systems is uncertain. AlthoughE. coli ribosomes changed frame at the IBV signal (UUUAAAG) with anef®ciency similar to that measured in reticulocyte lysates (40%), therewere important qualitative differences. Frameshifting of prokaryotic ribo-somes was pseudoknot-independent (although secondary structuredependent) and appeared to require slippage of only a single tRNA.

# 1997 Academic Press Limited

Keywords: ribosomal frameshifting; tRNA anticodon modi®cation; RNApseudoknot; lysyl-tRNA; Q base
*Corresponding author
Institute of Virology, University of Glasgow, Church Street, Glasgow G11 5JR,

arcoma virus; ORF, open reading frame; Q, queuosine; Y, wyebutoxine; HIV,TLV, human T-cell leukaemia virus; BLV, bovine leukaemia virus; IBV, infectiousocyte lysate; IPTG, isopropyl-b,D-thiogalactopyranoside; TGT, tRNA guanineammary tumour virus; pfu, plaque-forming units.

971134 # 1997 Academic Press Limited

tRNA Anticodon Modi®cation and Frameshifting 361

Introduction

Several viruses use an ef®cient ÿ1 ribosomalframeshifting mechanism to control expression oftheir replicases. Frameshifts of this class were®rst described as the process by which the gag-pol polyprotein of the retrovirus Rous sarcomavirus (RSV) is expressed from the overlappinggag and pol open reading frames (ORFs: Jacks &Varmus, 1985). Related frameshift signals havesince been documented in an increasing numberof systems, including several other retroviruses, anumber of eukaryotic positive-strand RNAviruses, a double-stranded RNA virus of yeast,some plant RNA viruses and certain bacterio-phage (reviewed by Brierley, 1995). The phenom-enon is not restricted to viruses; frameshiftsignals of the ``retrovirus type'' have been de-scribed in a number of Escherichia coli insertionelements (reviewed by Chandler & Fayet, 1993;Farabaugh, 1996) and in a conventional cellulargene, the dna X gene of E. coli (Blinkowa &Walker, 1990; Flower & McHenry, 1990;Tsuchihashi & Kornberg, 1990; Tsuchihashi,1991). The mRNA signals that specify frameshift-ing appear to be composed of two essential el-ements; a heptanucleotide ``slippery'' sequence,where the ribosome changes reading frame, anda region of RNA secondary structure, often in theform of an RNA pseudoknot, located a fewnucleotides downstream (Jacks et al., 1988a;Brierley et al., 1989; ten Dam et al., 1990). Themolecular mechanism of the frameshift process isonly poorly understood, but work from severalgroups supports a model (Jacks et al., 1988a) inwhich the elongating ribosome pauses upon en-countering the region of mRNA secondary struc-ture, facilitating realignment of the slipperysequence-decoding tRNAs in the ÿ1 frame. Theheptanucleotide stretch that forms the slippery se-quence contains two homopolymeric triplets andconforms, in the vast majority of cases, to themotif XXXYYYN. Frameshifting at this sequenceis thought to occur by simultaneous slippage oftwo ribosome-bound tRNAs, presumably peptidyland aminoacyl tRNAs, which are translocatedfrom the zero (X XXY YYN) to the ÿ1 phase(XXX YYY: Jacks et al., 1988a). Following the slip,the tRNAs remain base-paired to the mRNA inat least two out of three anticodon positions.There is considerable experimental support forthis model, particularly from site-directed muta-genesis studies (Jacks et al., 1988a; Dinman et al.,1991; Dinman & Wickner, 1992; Brierley et al.,1992), sequencing of trans-frame proteins (Hiziet al., 1987; Jacks et al., 1988a,b; Weiss et al., 1989;Nam et al., 1993) and nucleotide sequence com-parisons (Jacks et al., 1988a; ten Dam et al., 1990).The protein sequencing studies indicate that theframeshift occurs at the second codon of the tan-dem slippery pair, i.e. at that codon decoded inthe ribosomal aminoacyl (A) site (XXXYYYN).The importance of the A-site tRNA in frameshift-

ing was also apparent from the mutagenesis stu-dies; point mutations affecting the A-site tRNAwere generally more inhibitory than those affect-ing the P-site tRNA.

A key question that has remained unanswered iswhether frameshifting at the slippery sequence ismediated by canonical tRNAs, or requires the par-ticipation of special ``shifty'' tRNAs, more prone toframeshift than their ``normal'' counterparts (Jackset al., 1988a). At naturally occurring frameshiftsites, of the codons that are decoded in the riboso-mal A-site prior to tRNA slippage (XXXYYYN),only ®ve are represented in eukaryotes, AAC,AAU, UUA, UUC and UUU, and two in prokar-yotes, AAA and AAG (Farabaugh, 1996). These co-dons are decoded by tRNAs with a highlymodi®ed base in the anticodon loop (see Hat®eldet al., 1992 and references therein). In tRNAAsn

(AAC, AAU), the wobble base is queuosine (Q), intRNAPhe (UUC, UUU), wyebutoxine (Y) is presentjust 30 of the anticodon, in tRNALys (AAA, AAG),the wobble base is 5-methylaminomethyl-2-thiouri-dine (mnm5s2U) (prokaryotes) and in tRNALeu

(UUA), 2-methyl-5-formylcytidine is present at thewobble position (Debarros et al., 1996). Hat®eldet al. (1992) have suggested that hypomodi®ed var-iants of these tRNAs may exist that function asspeci®c ``shifty'' tRNAs, since such variants willhave a considerably less bulky anticodon and bemore free to move around at the decoding site. In-direct support for this hypothesis comes from anexamination of the modi®cation status of the antic-odons of the aminoacyl-tRNAs that are requiredfor translation at and around the frameshift sites ofhuman immunode®ciency virus type 1 (HIV-1),human T-cell leukaemia virus type I (HTLV-1) andbovine leukaemia virus (BLV: Hat®eld et al., 1989).It was found that in HIV-1 infected cells, most ofthe tRNAPhe lacked Y and in HTLV-1 and BLV in-fected cells, most of the tRNAAsn lacked Q. How-ever, research from other groups has suggested analternative hypothesis, in which frameshifting ismediated by standard cellular tRNAs (Tsuchihashi,1991; Tsuchihashi & Brown, 1992; Brierley et al.,1992). These authors propose that frameshifting ata particular site depends, amongst other par-ameters, upon the strength of the interaction be-tween the slippery sequence codons and thetRNAs decoding it and that if this interaction isrelatively weak, then slippage is more likely tooccur. The strength of the interaction betweenmRNA and tRNA is likely to be in¯uenced con-siderably by the kind of base-pair that forms be-tween the 30 base of the codon and the 50 base ofthe anticodon (position 34) at the wobble position.Modi®cation of the anticodon wobble base of fra-meshift site-decoding tRNAs may well in¯uencethis interaction and hence the level of frameshiftingobserved. A prediction of the hypothesis is that theanticodon modi®cations present in tRNAs that de-code highly ef®cient slippery sequences reduce rec-ognition of the corresponding codons.

Figure 1. The basic plasmids used in this study, thepMM series, were prepared by subcloning 585 bp NheI-EcoRI fragments containing the IBV frameshift regionfrom plasmids pFS7, pFS8 or mutant derivatives (Brier-ley et al., 1989, 1991) into BamHI-cleaved, plasmidpET3xc (Studier et al., 1990). Both fragment(s) and vectorwere end-®lled using the Klenow fragment of DNApolymerase I prior to ligation with T4 DNA ligase. Theresulting plasmids contain the IBV ORF 1a/1b frame-shift signal (sequence information from base-pairs 12,286to 12,511; Boursnell et al., 1987) ¯anked by portions ofthe in¯uenza A/PR8/34 PB1 gene (sequence infor-mation from base-pairs 1140 to 1167 (50) and 1167 to1500 (30); Young et al., 1983) located downstream of, andin frame with, the ®rst 783 bp of coding sequence of thebacteriophage T7 gene 10. The ensemble is under thecontrol of the bacteriophage T7 promoter and a T7 tran-scription termination signal is present at the end of thecoding sequences. In the relevant T7-expressing bacteria(see Materials and Methods), the constructs are pre-dicted to express a 33 kDa non-frameshifted product,corresponding to ribosomes that terminate at the IBV 1astop codon and a 50 kDa frameshift product from ribo-somes that frameshift prior to encountering the stopcodon and continue to translate the 1b ORF in the ÿ1frame.

362 tRNA Anticodon Modi®cation and Frameshifting

Here, we have attempted to test these hypoth-eses by measuring directly the in¯uence of tRNAanticodon modi®cation on frameshifting. Ourapproach was to determine the ef®ciency of theframeshift signal of the coronavirus infectiousbronchitis virus (IBV; Brierley et al., 1987, 1989) inmutant strains of Escherichia coli unable to modifyfully either tRNALys or tRNAAsn. The IBV frame-shift signal, which is present at the overlap of the1a and 1b ORFs of the virus genomic RNA, is awell-characterised eukaryotic system (Brierley et al.,1991, 1992) that comprises the slippery sequenceUUUAAAC and a downstream RNA pseudoknot.We began by determining the components of theIBV signal required for ef®cient frameshifting inE. coli and then proceeded with the investigation ofthe role of tRNA anticodon modi®cation in frame-shifting. Four slippery sequence variants of the IBVsite (UUUAAAX, where X was A, C, G or U) weretested, focusing speci®cally on the A-site decodingtRNAs. The results obtained indicate that in E. colithere is little in¯uence of the tRNA modi®cationstatus on frameshifting. Hypomodi®cation oftRNAAsn had only a slight effect on frameshiftingand of the tRNALys mutants tested, only trmE (an-ticodon s2UUU) had a marked in¯uence, increasingthe ef®ciency of the process at the slipperysequence UUUAAAA.

Results

Sequence requirements for IBV frameshiftingin E. coli

The ef®ciency of frameshift signals of the IBVtype in the eukaryotic rabbit reticulocyte lysate(RRL) in vitro translation system has been shownto be in¯uenced by the nature of the slippery se-quence, the integrity of the downstream RNAstructure and the precise spacing between the twoelements (Brierley et al., 1991, 1992). We began bycon®rming that these requirements were main-tained in E. coli. Complementary DNAs (cDNAs)containing variants of the IBV signal were sub-cloned (see Materials and Methods) into the E. coliexpression vector pET3xc (Studier et al., 1990) tocreate the pMM series of plasmids (see Figure 1).pET3xc contains the ®rst 783 bp of coding sequenceof the bacteriophage T7 gene 10, ¯anked by a T7promoter and transcription termination signal. InE. coli BL21 cells, which contain an IPTG-inducibleT7 RNA polymerase, a 261 amino acid residue por-tion of gene 10 is expressed from pET3xc as anabundant, Triton-insoluble product that is rela-tively easy to purify (see Materials and Methods).In the pMM plasmids, the IBV cDNAs were clonedin frame with and downstream of the gene 10 se-quence of pET3xc at a unique BamHI site. In BL21cells, these constructs were predicted to express a33 kDa non-frameshifted product corresponding toribosomes that terminated at the IBV 1a stopcodon and a 50 kDa frameshift product from ribo-somes which frameshifted prior to encountering

the stop codon and continued to translate the 1bORF in the ÿ1 frame. The constructs tested areshown in Figure 2, the expressions in Figure 3 andthe results are summarised in Table 2. For thepET3xc plasmid, which does not contain an IBV in-sertion, only two major species were present, theN-terminal portion of gene 10, with an apparentmolecular mass of about 34 kDa and lysozyme,carried over from the puri®cation procedure (seeMaterials and Methods). For the pMM expressions,three species were seen; lysozyme, a 36 kDaspecies corresponding to the non-frameshifted pro-duct and for most constructs, a 50 kDa frameshiftproduct. The identity of the 36 kDa and 50 kDaproteins was con®rmed by demonstrating thatboth proteins reacted in Western blots with amonoclonal antibody directed against the N termi-nus of gene 10 (AMS Biotechnology Ltd, Europe;data not shown). We also con®rmed that the36 kDa and 50 kDa proteins were restricted to theinsoluble fraction of E. coli cell extracts; no solublenon-frameshifted or frameshifted protein was de-tected (data not shown).

Figure 2. Frameshift constructs tested in E. coli. A, Slippery sequence variants. In each construct, the slipperysequence is boxed. Nucleotides that differ from the wild-type sequence (pMM2) are indicated in bold. In pMM10 andpMM11, the distance between the slippery sequence and RNA pseudoknot was decreased (GGG deleted) or increased(UAC inserted) by 3 nt, respectively. In pMM6 and pMM5, the pseudoknot was deleted (as in construct pFS7.6;Brierley et al., 1989). B, Pseudoknot variants. Complementary and compensatory changes were created within thepseudoknot region. In this representation of the pseudoknot, the stems are arranged vertically and the loops areshown as thick lines. For each base-pair(s) studied, the two complementary changes (no base-pairing) and the com-pensatory change (base-pairing restored) are boxed and labelled with a mutant number. C. Stem±loop constructstested. Two constructs were tested that formed a stem±loop structure rather than a pseudoknot. Plasmid pMM23forms only stem 1 due to a deletion that removed the downstream pseudoknot-forming region (as in constructpFS7.6; Brierley et al., 1989). Plasmid pMM7 is a stem±loop construct in which the stem nucleotides are of the samelength and nucleotide composition as the stacked stems of the pseudoknot in pMM2 (as in construct pFS8.26; Brierleyet al., 1991).

Figure 3. A and B, Frameshiftassays in E. coli. Puri®ed proteinsexpressed from the relevant frame-shift plasmids were separated onSDS 15% polyacrylamide gels anddetected by staining with Coomas-sie brilliant blue R as described inMaterials and Methods. The non-frameshifted (stop) and frame-shifted (frameshift or fs) productsare indicated by arrows. Lysozyme(lyso) carried over from the puri®-cation procedure is indicated by anarrow. The relevant mutant num-ber is indicated at the bottom ofeach track. A number of abbrevi-ations are employed: MW, highmolecular mass protein size stan-dards (Amersham); UUUAAAX,variant slippery sequence where Xis A, C, G or U as indicated abovethe relevant track; spacer, the slip-pery sequence-pseudoknot spacingdistance was increased (3 nt) ordecreased (ÿ3 nt) as indicated; noPK, pseudoknot deletion mutantswith slippery sequence UUUAAAC(C) or UUUAAAG (G); stem 1 orstem 2 complementary (M) and

compensatory (pWT) mutations were analysed in blocks of six (6 bp change), three (3 bp change) or single base-pairs(1 bp change); ssGGG, slippery sequence (UUUAAAG) and downstream codon are indicated; UAAssUAG, slipperysequence ¯anked by termination codons; point, single point mutation in stem 2 of construct pMM4; only S1, constructcan form only stem 1; simple-stem, construct forming long stem±loop structure as detailed in Figure 2C.


Slippery sequence requirements

The generality of the simultaneous slippagemodel of frameshifting for sites expressed in E. coliis not fully established. It was important, therefore,to determine whether frameshifting at the IBV slip-pery sequence in E. coli deviated from the conven-tional simultaneous slippage mechanismascertained in RRL. We created a series of mu-tations at or around the IBV slippery sequence andtested for frameshifting by expression in E. coliBL21 (see Figure 3A). The wild-type IBV slipperysequence, UUUAAAC (pMM1), decoded bytRNAAsn (anticodon 50 QUU 30) stimulated onlylow-levels of frameshifting (2%), as didUUUAAAU (pMM8, 3%). In E. coli, therefore,tRNAAsn is considerably less slippery than it is ineukaryotic cells or the RRL (Brierley et al., 1989). Incontrast, tRNALys (anticodon 50 U8UU 30; where U8is mnm5s2U), which reads UUUAAAA (pMM9)and UUUAAAG (pMM2), was highly slippery;frameshifting at these sites occurring with greatef®ciency (15% and 40%, respectively). The slip-periness of E. coli tRNALys has been documentedand will be discussed in detail later. In E. coli thehierarchy of frameshifting for the seventh nucleo-tide of the slippery sequence (UUUAAAN) wasN G > A4U C. This is almost the reverse ofthe situation seen in RRL, where the hierarchy for

N is C > A U4G (Brierley et al., 1992), and isconsistent with earlier studies of frameshifting inE. coli (Weiss et al., 1989; Tsuchihashi & Brown,1992). The introduction of different termination co-dons (UGA, pMM24; UAG, pMM27; UAA,pMM28) or an alternative sense codon (UGG,pMM25) immediately downstream of the slipperysequence had little effect on frameshifting,although a discernible reduction was seen with theUAG (32%) and UGA (29%) terminators. By ¯ank-ing the slippery sequence with termination codons(pMM26), one immediately downstream (UGA) inthe zero phase and a second immediately upstream(UAA) in the ÿ1 phase, we were able to con®rmthat frameshifting takes place within theUUUAAAG stretch (Figure 3B). As with pMM27,we observed a slight reduction in frameshift ef®-ciency with this construct, which had UAG as thedownstream termination codon (see Discussion).We also introduced mutations within the slipperysequence. Unsurprisingly, the sequence UUUA-CAG (pMM30) was non-functional, since his mu-tation would prevent slippage of the A-sitedecoding tRNA. However a P-site mutation,UCUAAAG (pMM29) was fully competent in fra-meshifting, suggesting that in the case of the IBVsignal in E. coli, the process does not involve simul-taneous slippage of two tRNAs, but rather ÿ1 slip-page of a single tRNALys (from AAG to AAA).


RNA secondary structure requirements

Ef®cient frameshifting at the IBV signal, at leastin the RRL, depends upon the RNA pseudoknotstructure, which cannot be replaced functionally bya stable stem±loop structure of the same predictedsize and nucleotide composition as the stackedstems of the pseudoknot (Brierley et al., 1991). Weinvestigated whether the RNA secondary structurerequirements for frameshifting in E. coli were con-served by measuring the frameshift ef®ciency of aseries of pseudoknot mutants (see Figures 2 and 3).Complete removal of the pseudoknot dramaticallyreduced frameshifting, irrespective of whether theslippery sequence was UUUAAAC (pMM6) orUUUAAAG (pMM5). The pattern of frameshiftingobserved for mutations within the pseudoknot clo-sely paralleled that seen in the RRL, in that desta-bilization of either stem of the pseudoknot reducedframeshifting ef®ciency (pMM3, 12, 13, 15, 20, 21in stem 1; pMM4, 17, 18 in stem 2) and compensa-tory mutations predicted to restore the structure ingeneral increased frameshifting (pMM16, 22 instem 1; pMM19 in stem 2). Of the compensatorymutants in stem 1, two of the three analysed hadframeshift ef®ciencies considerably below that seenfor the wild-type structure (pMM14, 12% andpMM16, 20%). This is a phenomenon that has beenobserved in a number of studies of eukaryotic fra-meshifting (see ten Dam et al., 1995) and may re-¯ect a functional requirement for a particularpseudoknot conformation that is imprecisely repro-duced in some of the compensatory mutants. Theef®ciencies of frameshifting measured for the stem2 point mutations (pMM4, 15%; pMM17, 15% andpMM18, 9%) were considerably greater than thatseen previously in RRL, where frameshifting inthese mutants is reduced to about 5% (Brierleyet al., 1991). This supports the idea that ef®cientframeshifting in E. coli can be mediated by simplehairpin loop stimulators. Indeed, in construct

Table 1. Bacterial strains used in this work

Strain Genotype

XA105 ara � (lac-proB) nalA rif thi metB argEam supXA10B ara � (lac-proB) nalA argEam rif thi metB supTH48 As XA105 but fadL::Tn10TH49 AsTH48 but trmC2TH69 As TH48 but trmC1DEV1 thi1 rel1 spoT1 lacZ(UAG)DEV16 As DEV 1 but valR, trmETH78 As TH79 but trmETH79 ara � (lac-proB) nalA rif thi metB argEam valTH159 As TH160 but asuE107TH160 ara � (lac-proB) nalA rif thi metB argEam supSJ1502 araD139 � (argF-lacU139) thi1 deoC relA1 rSJ1505 As SJ1502 but tgt1K12�tgt � (tgt)TG1 supE, thi � (lac-proAB) F0 (traD36 proABBL21(DE3) hsdS gal (lcIts857 ind1 Sam7 nin5 lacUV5-T

a Unpublished results.b This strain was prepared in the laboratory of Professor Helga

Fahrstrasse 17, D-8520 Erlangen, Germany.

pMM23, which contains the potential for formationof only stem 1, frameshifting occurred ef®ciently(22%). Moreover, when the pseudoknot of pMM2was replaced by a large stem±loop structure of thesame predicted size and nucleotide composition asthe stacked stems of the pseudoknot (pMM7), fra-meshifting occurred at a high level in E. coli cells(32%), in contrast to the situation in RRL, wheresuch a structure promotes only low levels (1 to 2%)of frameshifting. This observation highlights an im-portant difference between prokaryotic andeukaryotic ribosomes in terms of their response toRNA stimulators associated with frameshift sites.A similarity that is maintained, however, is the ne-cessity for precise spacing between the stimulatorystructure and the slippery sequence. As in the RRL(Brierley et al., 1989), increasing (pMM11) or de-creasing (pMM10) the spacing distance by threenucleotides greatly reduced frameshifting at theIBV site in E. coli.

The role of tRNA anticodon modificationin frameshifting

To investigate the role of tRNA anticodonmodi®cation in the frameshift process, pairs ofplasmids with variations in the last nucleotide ofthe slippery sequence (UUUAAAN) were ex-pressed in E. coli strains de®cient in tRNA modi®-cation (see Table 1). In these experiments, T7 RNApolymerase was provided by infecting cells withbacteriophage l CE6 (see Materials and Methods),which contains the gene for T7 RNA polymeraseunder the control of the PL and PI promoters.Under these circumstances, we found that ex-pression levels were generally lower than thoseseen upon IPTG-induction of BL21 cells and werein¯uenced by the growth-rate of the cells (re¯ect-ing the ef®ciency of the l CE6 infection). However,the signal to noise ratios (in terms of expressedproteins to cellular background) were suf®ciently

Reference or source

G Miller et al. (1977)B Miller et al. (1977)

Hagervall & BjoÈ rk (1984)Hagervall & BjoÈ rk (1984)Hagervall & BjoÈ rk (1984)Elseviers et al. (1984)Elseviers et al. (1984)This work

R supB This workHagervall & McCloskeya

B fadR::Tn10 Hagervall & McCloskeya

psL150 Reuter et al. (1991)Reuter et al. (1991)Kerstenb

lacIq lacZ� M15) hsh� 5 Gibson et al. (1984)7 gene 1) Studier & Moffatt (1986)

Kersten, InstituÈ t fuÈ r Biochemie, UniversitaÈ t Erlangen-NuÈ rnberg,

Figure 4. Anticodon modi®cations studied. TheFigure shows the anticodon stem±loop nucleotides, thenucleoside modi®cations and the mutants tested forboth lysyl and asparaginyl-tRNAs.

Figure 5. Frameshift assays in modi®cation-de®cientE. coli hosts. Frameshift plasmids were expressed in avariety of E. coli strains (see Table 1), the puri®ed ex-pression products separated on SDS 15% polyacryl-amide gels and detected by staining with Coomassiebrilliant blue R as described in Materials and Methods.The non-frameshifted (stop) and frameshifted (frame-shift) products are indicated by arrows. Lysozyme car-ried over from the puri®cation procedure is indicated byan arrow (lyso). The E. coli strain employed is indicatedat the bottom of each track. The plasmids expressedwere pMM1 (UUUAAAC), pMM2 (UUUAAAG), pMM8(UUUAAAU) and pMM9 (UUUAAAA). The relevantgenotype is indicated above each track (wt, wild-typewith respect to the modi®cation genotype); Q; Q status,either absent (ÿ) or present (). A, Bacteria cultured inLB medium; B, bacteria cultured in a de®ned minimalmedium (see Materials and Methods), except asuE (LBmedium).


high that estimates of frameshifting ef®ciency werereliable and reproducible. We studied two antico-don modi®cations, the Q base of tRNAAsn and themnm5s2U base of tRNALys (see Figure 4).

The role of Q

Plasmids pMM1 (UUUAAAC) and pMM8(UUUAAAU) were expressed in a wild-type host(SJ1512) and in two strains de®cient in the bio-synthesis of Q due to a lack of functional tRNAguanine transglycosylase (TGT: SJ1515, K12 �tgt).In these strains, Q is not incorporated into tRNAand tRNAAsn has the anticodon 50 GUU 30. Ascan be seen in Figure 5A, the absence of Q hadonly a modest effect on frameshifting, resulting ina twofold increase (UUUAAAU) or decrease(UUUAAAC) in frameshift ef®ciency (see Table 2).It is clear therefore that in E. coli, hypomodi®cationof the anticodon of tRNAAsn does not lead to a dra-matic increase or decrease in frameshifting.

The role of mnm5s2U

Unlike tRNAAsn, the anticodon wobble base oftRNALys is modi®ed at two positions, at position 5,where hydrogen is replaced by a methylamino-methyl group and at position 2, where oxygen isreplaced by sulphur. Although a number of E. colimutants exist that are defective in the synthesisof mnm5s2U, a fully unmodi®ed strain is notavailable. Plasmids pMM9 (UUUAAAA) andpMM2 (UUUAAAG) were expressed in wild-typehosts (TG1, TH48, DEV 1, TH79, TH160) and in the

relevant defective strains listed in Table 1. Essen-tially, four mutants were examined; trmC1(Hagervall & Bjork, 1984), which has thehypermodi®cation 5-carboxymethylaminomethyl-2-thiouridine (cmnm5s2U; TH69); trmC2 (Hagervall& Bjork, 1984), containing the undermodi®ed 5-aminomethyl-2-thiouridine (nm5s2U; TH49); trmE(Elseviers et al., 1984), which possesses only the 2-thiouridine substitution (s2U; TH78, DEV 16) andasuE107, which contains only the 5-methylamino-methyl substitution (mnm5U; TH159). In thetrmC1, trmC2 and trmE strains, the presence of thesulphur atom was potentially problematic, since itcan be replaced by selenium, which is thought toconfer altered decoding properties (Wittwer &

Table 2. The ÿ1 frameshifting ef®ciencies from the analyses of Figures 3 and 5Construct Feature Ef®ciency (%)

A. Slip-site variants pMM9 UUUAAAA 15pMM1 UUUAAAC 2pMM2 UUUAAAG 40pMM8 UUUAAAU 3pMM29 UCUAAAG 40pMM30 UUUACAG 1

B. Flanking codons pMM28 UUAUUUAAAGUAA 40pMM27 UUAUUUAAAGUAG 32pMM24 UUAUUUAAAGUGA 29pMM25 UUAUUUAAAGUGG 41pMM26 UAAUUUAAAGUAG 33

C. Spacing variants pMM11 Spacer 3 nucleotides 8pMM10 Spacer ÿ3 nucleotides 2

D. PK mutations pMM6 Pseudoknot deletion 1pMM5 Pseudoknot deletion 1pMM23 Stem 2 deleted 22pMM7 Long stem-loop 32pMM12 Stem 1 3 bp change 8pMM13 Stem 1 3 bp change 10pMM14 Stem I pseudo-wt 12pMM15 Stem 1 1 bp change 10pMM3 Stem 1 1bp change 19pMM16 Stem 1 pseudo-wt 20pMM20 Stem 1 6 bp change 2pMM21 Stem 1 6 bp change 2pMM22 Stem 1 pseudo-wt 38pMM4 Stem 2 point mutation 15pMM17 Stem 2 1 bp change 15pMM18 Stem 2 1 bp change 9pMM19 Stem 2 pseudo-wt 37

E. Role of tRNA anticodon modi®cation

Slip sequence tested Host/modi®cation

UUUAAAC SJ1512/wt Q 2SJ1515/tgt1 Qÿ 1K12�tgt Qÿ 1

UUUAAAU SJ1512/wt Q 3SJ1515/tgt1 Qÿ 6K12�tgt Qÿ 5

UUUAAAA TG1/wt 15TH48/wt 15TH49/trmC1 16TH69/trmC2 15DEV1/wt 15DEV16/trmE 32TH79/wt 15TH78/trmE 32TH160/wt 15TH159/asuE 15

UUUAAAG TG1/wt 40TH48/wt 41TH49/trmC1 40TH69/trmC2 40DEV1/wt 40DEV16/trmE 48TH79/wt 40TH78/trmE 49TH160/wt 40TH159/asuE 33

Each value quoted represents the average of three to ®ve independent measurements, whichvaried by less than 5%; i.e. a measurement of 40% frameshift ef®ciency was between 38%and 42%. The abbreviations used are: PK, pseudoknot; wt, wild-type.



Ching, 1989). This was not a problem with theasuE strain; the level of mnm5Se2U has beenmeasured in an asuE strain and was not detected(Kramer & Ames, 1988). Previous studies, how-ever, have indicated that selenium is not detectablyincorporated into tRNA when bacteria are culturedin minimal medium (Wittwer & Stadtman, 1986).Although these authors have suggested that this isalso the case for LB medium, we decided tomeasure frameshifting in the trmC1, trmC2 andtrmE strains during culture in a de®ned minimalmedium prepared from chemicals of high purity inaddition to assays in standard LB medium. The re-sults of the analysis are shown in Figure 5 and aresummarised in Table 2. The signal to noise ratio inthe minimal medium assays was increased some-what as a consequence of the slower growth-rateof the strains. The asuE strain (Figure 5B) grewvery slowly in minimal medium and was onlytested in LB medium. The frameshift ef®cienciesmeasured for the various mutants were not in¯u-enced by the growth medium. No effect on frame-shifting was seen in the trmC1 or trmC2 strainswhen either the UUUAAAA or UUUAAAG-con-taining test plasmids were expressed. The most no-ticeable in¯uence was in the trmE strain, whereframeshifting was stimulated over twofold (from15% to 32%) at the UUUAAAA site and also in-creased at the UUUAAAG site, although to a lesserextent (from 40% to 48%). These increases wereseen in two independent trmE strains (TH78, DEV16). Frameshifting in the asuE background was un-altered at the UUUAAAA site but was reduced alittle at the UUUAAAG site (from 40% to 33%).

Discussion

Signals for frameshifting in E. coli

The generality of the simultaneous slippagemodel of frameshifting for sites expressed in E. coliis not fully established. Weiss et al. (1989) have ex-pressed a variant of the MMTV frameshift signal inE. coli and have shown that ribosomes respond toboth of the tandem slippery codons of the MMTVframeshift signal as predicted by the simultaneousslippage model. At the frameshift signal of theE. coli dnaX gene, mutagenesis of the slippery se-quence (A-AAA-AAG) has con®rmed that both ly-sine codons are required for ef®cient frameshifting(Tsuchihashi & Brown, 1992). Similarly, frameshift-ing at the G-T ORF overlap required to produce abacteriophage l tail assembly protein occurs by atwo tRNA slip (Levin et al., 1993). However, in theE. coli insertion element IS1, frameshifting isknown to occur by ÿ1 slippage of a single lysyltRNA at the sequence A-AAA (from the under-lined codon onto the overlapping AAA codon), de-spite the fact that the A4 stretch is embeddedwithin two potential and conventional slippery se-quences (U-UUA-AAA-AAC; Sekine & Ohtsubo,1992). An unexpected ®nding from our analysis of

the slippery sequence requirements for IBV frame-shifting in E. coli was the high ef®ciency of theUCUAAAG mutant (pMM29), since such a mutanthas only low activity in RRL (Brierley et al., 1989).The simplest explanation for this observation isthat frameshifting at the IBV site in E. coli occursby slippage of a single tRNA at the second homo-polymeric triplet of the slippery sequence(UUUAAAX) rather than by simultaneous slippageof two tRNAs. The P-site codon in the mutant,UCUAAAG, is probably decoded by a minortRNALeu isoacceptor with anticodon 50 UAG 30(Inokuchi & Yamao, 1995). If this tRNA were toslip into the ÿ1 reading frame (in accordance withthe simultaneous slippage model) it could formonly a single G-U base-pair with the ÿ1 framecodon. At present, therefore, we favour single-tRNA slippage. Whether tRNALys slips when inthe P or A-site of the ribosome is not know. Recentevidence supports the idea that frameshifting atthe HIV-1 slippery sequence (U-UUU-UUA) inE. coli occurs not by simultaneous slippage of Pand A-site-bound tRNAs, but when these tRNAsare in the ribosomal E and P-sites (Hors®eld et al.,1995). This hypothesis was proposed following thediscovery that the presence of a termination codonimmediately downstream of the U6A stretch re-duced frameshifting some ®ve- to tenfold in amanner that was independent of sequence contextand could be modulated by prokaryotic release fac-tor 2. These observations were consistent with thelast six nucleotides of the slippery sequence occu-pying the E and P-sites and the termination codonthe A-site prior to tRNA slippage. In the presentstudy, we did detect a slight reduction in frame-shift ef®ciency when the downstream terminatorsUAG and UGA were employed, raising the possi-bility that a fraction of the frameshift events moni-tored involve the E-site; but the data are mostconsistent with the slippery sequence occupyingthe standard P and A-sites. This conclusion is sup-ported by the fact that, as is also seen in RRL(Brierley et al., 1989, 1992), the spacing distance be-tween the slippery sequence and pseudoknot hadto be maintained at six nucleotides for ef®cient fra-meshifting to occur. Protein sequencing studiesand further mutagenesis experiments are in pro-gress in an attempt to improve our understandingof the precise mechanism of tRNA slippage at theIBV site in E. coli.

The requirements for downstream RNA second-ary structure were tested in constructs with themost ef®cient (UUUAAAG) slippery sequence,using a series of complementary and compensatorypseudoknot mutants. The response of prokaryoticribosomes to these mutants was generally similarto that seen with the eukaryotic ribosomes of theRRL in vitro translation system (Brierley et al.,1991). However, an important difference wasnoted; when the IBV pseudoknot was replaced bya stable stem±loop structure of the same predictedsize and nucleotide composition, frameshifting wasmaintained at a high level in E. coli cells, in con-


trast to the situation in RRL, where such a struc-ture promotes only low levels of frameshifting. Innaturally occurring frameshift sites in E. coli, the re-quirements for stimulatory RNA structures appearto be variable. At the G-T ORF overlap of bacterio-phage l, no stimulatory secondary structure is ap-parent (Levin et al., 1993). In contrast, the dnaXframeshift signal includes both a downstreamstem±loop (Tsuchihashi & Kornberg, 1990) and anupstream stimulatory element formed between aShine-Delgarno-like sequence (50 AGGGaG 30)located 10 nt upstream of the slippery sequenceand the 30-end of 16 S rRNA (30 UCCUcC 50:Larsen et al., 1994). Variation is also evident at theframeshift sites of bacterial insertion sequences(Chandler & Fayet, 1993). In IS3, a pseudoknot isrequired for ef®cient frameshifting (Sekine et al.,1994) and in IS150, stimulation is via a downstreamstem±loop that may form a pseudoknot (VoÈgeleet al., 1991). In contrast, although a number of po-tential RNA structures are predicted to formdownstream of the slippery sequence of IS1(Sekine & Ohtsubo, 1989; Escoubas et al., 1991),they do not appear to play a role in frameshifting(Sekine & Ohtsubo, 1992). Whatever the case, astimulatory structure is absolutely required for IBVframeshifting in E. coli; the UUUAAAG sequencealone was unable to stimulate detectable frame-shifting. This is perhaps unsurprising, since in theexpression constructs employed, no obvious Shine-Delgarno-like sequences are present upstream ofthe slippery sequence.

Influence of tRNA anticodon modificationon frameshifting

Our investigation of the role of tRNA anticodonmodi®cation in frameshifting was prompted by theexperiments reported by Hat®eld et al. (1989, 1992),who proposed that hypomodi®cation of the antico-dons of those tRNAs implicated in decoding fra-meshift sites may promote ef®cient frameshifting.We began by expressing our frameshift reporterconstructs in E. coli tgt mutants unable to bio-synthesize Q. In these cells, however, we detectedonly a modest in¯uence of hypomodi®ed tRNAAsn,with frameshifting reduced by about twofold onthe slippery sequence UUUAAAC and increasedby a similar magnitude at the UUUAAAU site. Sofor tRNAAsn, at least in E. coli, anticodon hypomo-di®cation per se is insuf®cient to promote highly ef-®cient frameshifting. The modest effects noted,however, are consistent with the view that thestrength of the pre-slippage codon-anticodon inter-action is important in frameshifting. The hypomo-di®ed variant of tRNAAsn with anticodon 50 GUU30 would be expected to pair more strongly withthe AAC codon than with AAU, and hence frame-shifting should decrease at the AAC codon and in-crease at AAU, as was observed. The low levels offrameshifting seen in E. coli at slippery sequencesdecoded by tRNAAsn, with or without Q, is in con-trast to the situation in higher eukaryotic cells,

where tRNAAsn is highly slippery (Brierley et al.,1992), despite possessing an identical anticodonloop (Chen & Roe, 1978). The altered frameshift ca-pacity may simply be a re¯ection of the eukaryotictranslational environment, but a role for Q ineukaryotic frameshifting cannot be ruled out.

The situation with UUUAAAA and UUUAAAG,decoded by tRNALys, is more complex, since thewobble base of this tRNA (mnm5s2U) is modi®edat two positions (see Figure 4) and an E. coli strainexpressing a fully unmodi®ed tRNA is unavailable.In the present study, of the four available mutantswith altered modi®cation of the wobble base oftRNALys, only trmE (wobble base is 2-thiouridine)had a marked in¯uence on frameshifting, increas-ing the ef®ciency of the process over twofold atUUUAAAA and to a lesser extent at UUUAAAG.The asuE mutant (wobble base mnm5U) showed asmall reduction in frameshifting at UUUAAAG.The increased ef®ciency seen in the trmE strains isconsistent with the idea that hypomodi®cation ofthe anticodon stimulates frameshifting (Hat®eldet al., 1992). However, this hypothesis cannot ex-plain the intrinsic slipperiness of the hypermodi-®ed ``wild-type'' tRNALys. The biological role (or atleast one of the roles) of the tRNALys modi®cationsis thought to be in the regulation of the confor-mational ¯exibility or rigidity of the base to ensurethe correct translation of codons during proteinsynthesis, particularly to prevent decoding of AACand AAU. Proton NMR studies of the confor-mational characteristics of modi®ed uridine bases(Yokoyama et al., 1985) indicate that the mnm5s2

modi®cation introduces conformational rigiditysuch that the ribose exclusively adopts a C30-endoform, which favours recognition of adenosine, al-lows weak recognition of guanosine and precludesbinding to cytosine and uridine. In support of this,tRNALys has been shown to decode preferentiallythe AAA codon and has little af®nity for AAG(Lustig et al., 1981). Tsuchihashi (1991) has rational-ised the high-ef®ciency frameshift signal of thednaX gene on the basis that the slippery sequence,AAAAAAG, is poorly recognised by tRNALys,which slips ef®ciently at the AAG codon due to aweak wobble base-pair. This concept is supportedby the experiments reported by Tsuchihashi &Brown (1992), who were able to inhibit frameshift-ing at the dnaX site by expressing an additionaltRNALys with anticodon 50 CUU 30. In these cells,frameshifting at the AAAAAAG slippery sequenceis thought to have been prevented by the morestable recognition of the AAG codon by tRNALys

CUU. We believe that the ef®cient frameshift ob-served at the IBV signal in E. coli (at UUUAAAG)is also a result of a restricted capacity of tRNALys

to pair with AAG. The fact that the UUUAAAAslippery sequence is also highly permissivesuggests that recognition of AAA by tRNALys isalso unusual when compared, for example, withthe decoding of AAC or AAU by tRNAAsn.

Yokoyama et al. (1985) have suggested that boththe 2 and 5-substituents contribute to the confor-


mational rigidity of tRNALys, with the major inputfrom the 2-thiocarbonyl group. The NMR studiesby Agris and colleagues (Sierzputowska-Graczet al., 1987; Agris et al., 1992) indicate a role foronly the 2-position thiolation. A second interaction,a hydrogen bond, has been proposed, which formsbetween the amino group of 50-substitutents con-taining an aminomethyl moiety (mnm5 andcmnm5) and the 20-OH residue of the adjacent, un-modi®ed U33 base in the U-turn structure of theanticodon loop (Hillen et al., 1978; Yokoyama &Nishimura, 1995; and see Figure 4). Circular di-chroism analysis of the tRNA suggests that the an-ticodon has an unusual conformation in which thewobble base is buried in the anticodon loop, poss-ibly interacting via its 50-substituent with the N6threonylcarbamoyl (t6) modi®cation of base A37across the loop (Watanabe et al., 1993). Support forthis idea comes from studies of a chemically syn-thesised short oligoribonucleotide comprising U33-mnm5s2U-U-U-t6A37 (cited by Agris, 1996). In thisdoubly modi®ed pentamer, a unique interactionoccurs between the two modi®cations that isthought to be between the amine group of mnm5

and the amino acid residue of t6A, by hydrogen orionic bonding. Furthermore, model-building stu-dies predict that the anticodon domain U-turn intRNALys is at the mnm5s2U34 base rather than atthe usual invariant residue U33.

The structural studies described above suggestthat the intrinsic shiftyness of tRNALys arises fromits inability to recognise ef®ciently lysine codons asa consequence of the anticodon modi®cations,which are required to prevent erroneous decodingof asparagine codons. In this light, we expectedthat gross alterations in the modi®cation status (ab-sence of thiolation in asuE, absence of the 50-substi-tuent in trmE) would result in a tRNA that wouldbe less restricted and perhaps pair more readilywith lysine codons. Following this logic, frame-shifting would be reduced in these mutants, sincethe codon-anticodon interaction would be morestable. Clearly this was not the case, and at presentwe are unable to explain these observations satis-factorily. The most likely explanation is that pos-session of either of the two modi®cations issuf®cient to reduce recognition of lysine codonsand allow ef®cient frameshifting. The structuralstudies imply a role for the amino group of the 50-substituent of tRNALys, most likely in inter-loopcontact with t6A37. The increase in frameshiftingseen with trmE, which does not possess the aminogroup, may well re¯ect the loss of such an inter-action, resulting in altered recognition of the AAAcodon (and to a lesser extent the AAG codon) bythe mutant tRNA. It seems unlikely that the effectseen with trmE is related to how well the mutanttRNA is aminoacetylated; ef®cient aminoacetyla-tion appears to be correlated with the presence ofthe s2 moiety in tRNAs of this class (Rogers et al.,1995). The lack of effect of the trmC1 and trmC2mutations, which retain the amino group, maysuggest that the conformation of tRNALys in these

strains is not suf®ciently different to in¯uencecodon recognition. In the case of trmC1(cmnm5s2U; TH69) and trmC2 (nm5s2U; TH49),model-building studies (Lim, 1994) support theidea that the modi®ed tRNAs would be expectedto have similar decoding properties as the wild-type tRNA. Mutants defective in asuE, trmE andtrmC all decrease the ef®ciency of the ochre sup-pressor tRNA supG, which is a derivative oftRNALys with anticodon 50 mnm5s2UUA 30 (re-viewed by BjoÈrk, 1992). Whether this re¯ects a re-duced af®nity for the nonsense codon or anotherstep in the suppression pathway is not clear,although the trmC1 mutation is known not to affectthe binding properties of tRNALys to AAA or AAGprogrammed ribosomes (Elseviers et al., 1984).

In conclusion, our data are not wholly consistentwith either of the hypotheses advanced to explainthe role of tRNA anticodon modi®cation in frame-shifting. The observed effects of hypomodi®edtRNAAsn on frameshifting at UUUAAAU/C aremost consistent with the idea that the strength ofthe codon-anticodon interaction determines frame-shift ef®ciency rather than being mediated primar-ily by hypomodi®ed tRNAs. However, the dataobtained with variantly modi®ed tRNAsLys are notreadily explainable in terms of either model.Although the increased frameshifting seen in trmEstrains is consistent with a role for hypomodi®edtRNAs, the inherent slipperiness of the wild-typehypermodi®ed tRNALys argues strongly againstthis hypothesis. Nevertheless, the data obtained forthe other tRNALys modi®cation mutants are dif®-cult to interpret solely in terms of the predictedstrength of codon-anticodon recognition. All ef®-cient ÿ1 frameshift sites in E. coli employ tRNALysas the A-site decoding tRNA and in all likelihood,this tRNA has a very unusual anticodon structure.In attempting to compare the role of tRNA antico-don modi®cation in frameshifting between prokar-yotic and eukaryotic systems, we remain mindfulthat tRNALys may be a special case and that therole of modi®ed bases in eukaryotic frameshiftingneeds to be tested directly. Such studies are under-way.

Materials and Methods

Site-specific mutagenesis

Site-directed mutagenesis was carried out by a pro-cedure based on that of Kunkel (1985) as described(Brierley et al., 1989). Mutants were identi®ed by dideoxysequencing of single-stranded templates (Sanger et al.,1977).

Construction of plasmids

The starting plasmids pFS7, pFS8 or mutant deriva-tives (Brierley et al., 1989, 1991), which contain the IBVORF 1a/1b frameshift signal ¯anked by portions of thein¯uenza virus A/PR8/34 PB1 gene (Young et al., 1983),were subjected to site-directed mutagenesis. In mostcases, this was to change the IBV slippery sequence from


UUUAAAC to UUUAAAG. Following mutagenesis,585 bp NheI-EcoRI fragments encompassing the frame-shift region were subcloned from the mutated plasmidsinto BamHI-cleaved pET3xc, an E. coli expression vector(Studier et al., 1990). Both fragments and vector wereend-®lled using the Klenow fragment of DNA polymer-ase I prior to ligation with phage T4 DNA ligase. The re-sulting plasmids, which comprise the pMM series, areshown in Figure 1. The PB1:1a/1b:PB1 fragment is lo-cated downstream of, and in frame with, the ®rst 783 bpof coding sequence of the bacteriophage T7 gene 10 andthe ensemble is under the control of the bacteriophageT7 promoter.

Bacterial strains and culture conditions

Bacteria were grown at 37�C in rich medium (LB;Maniatis et al., 1982) or in minimal salt medium (M9;Maniatis et al., 1982) containing thiamine (5 mg/l), glu-cose (0.2%, w/v) and the required amino acids(50 mg/l). The M9 medium was prepared using re-agents of high purity (Aldrich Chemical Company). Thegenotypes and origins of the E. coli strains used aregiven in Table 1. TH78 was constructed by transferringthe trmE allele from DEV 16 into XA10B by phage P1transduction, essentially according to Miller (1972).Transductants were selected for valine resistance andco-transduction of trmE monitored by screening for anArgÿ phenotype. In this screen, the presence of trmEwas observed as an antisuppresor activity of supB,which suppresses poorly the amber mutation in argE. Alack of mnm5s2U and the presence of s2U in tRNA fromTH78 was veri®ed by HPLC analysis according toGehrke & Kuo (1990). TH79 is an isogenic trmE strain.The isolation of the asuE107 mutation used in this study(TH159) will be described elsewhere. TH160 is an iso-genic asuE strain. Transfer RNA puri®ed from TH159was shown to contain mnm5U instead of mnm5s2U bycombined liquid chromatography-mass spectrometry(LC/MS; unpublished results).

Frameshifting in E. coli BL21

The sequence requirements for IBV frameshifting inE. coli were investigated by expressing the pMM plasmidseries in E. coli BL21/DE3/pLysS cells (Figure 1, Table 1),largely as described by Studier et al. (1990). These cellscontain, under the control of the lacUV5 promoter, thegene for T7 RNA polymerase inserted within the intgene of the prophage DE3, a l derivative. Expression ofT7 RNA polymerase in BL21 cells can be induced by ad-dition of isopropyl-b,D-thiogalactopyranoside (IPTG).Freshly transformed BL21 cells prepared by the methodof Hanahan (1983) and containing plasmids of the pMMseries were inoculated into 1.5 ml LB cultures containingampicillin (50 mg/ml) and chloramphenicol (30 mg/ml).After three hours at 37�C, IPTG was added to 0.4 mMand incubation continued for a further three hours. Cellswere pelleted, resuspended in 150 ml of lysis buffer(25 mM Tris (pH 8), 10 mM EDTA, 50 mM sucrose,2 mg/ml lysozyme), placed on ice for 30 minutes andtreated with DNase I (30 mg/ml) for a further 30 minutesat 37�C in the presence of 8 mM MgCl2 and 1 mMMnCl2. Detergent solution (300 ml of 20 mM Tris (pH7.5), 2 mM EDTA, 0.2 M NaCl, 1% (w/v) deoxycholicacid, 1% (v/v) Nonidet P-40) was added and the insolu-ble material harvested by centrifugation at 8000 g fortwo minutes. The pellet was washed three times with0.5% (v/v) Triton X-100, 1 mM EDTA and the Triton-in-

soluble material, which contained predominantly thepMM expression products, dissolved in 100 ml of samplebuffer (Laemmli, 1970). Aliquots were analysed on SDS/15% (w/v) polyacrylamide gels according to standardprocedures (Hames, 1991). Proteins were stained withCoomassie brilliant blue R (0.05%, w/v) in 10% (v/v)acetic acid, 50% (v/v) methanol, and destained in 10%acetic acid, 20% methanol. The relative abundance ofnon-frameshifted or frameshifted products was esti-mated (Adobe Photoshop and NIH Image software) byscanning densitometry and adjusted to take into accountthe differential size of the products. Scans were per-formed on gels whose proteins were stained to an inten-sity at the centre of the dynamic range of the scanner(Microtek IIXE Scanmaker). The frameshift ef®cienciesquoted in the text and summarised in Table 2 are theaverage of three to ®ve independent measurements thatvaried by less than 5%, i.e. a measurement of 40% frame-shift ef®ciency was between 38% and 42%.

Frameshifting in modification-deficient E. coli strains

The in¯uence of tRNA anticodon modi®cation on IBVframeshifting was probed by expressing pMM plasmidsin modi®cation-de®cient E. coli strains (Table 1). Sincethe strains were non-lysogenic for DE3, expression of T7RNA polymerase was achieved by infecting plasmid-bearing cells with l CE6 (AMS Biotechnology UK Ltd.)>This phage contains the gene for T7 RNA polymeraseunder the control of the PL and PI promoters. Stocks ofCE6 were prepared in E. coli LE392 by plate lysis (Mania-tis et al., 1982), collected by centrifugation (100,000 g fortwo hours) and resuspended at 1012 pfu/ml. Modi®-cation-de®cient strains harbouring pMM plasmids weregrown until the absorbance at 600 nm was between 0.6and 0.8, and infected with CE6 at 10 to 20 pfu/cell. Afterthree hours, the cells were harvested and pMM ex-pression products puri®ed and analysed as above.

Acknowledgements

We thank Professor Glenn BjoÈrk for his support andadvice, Professor Helga Kersten and members of her lab-oratory for the kind gift of E. coli strains and Dr PaulDigard for helpful comments on the manuscript. Thiswork was supported by the Medical Research Council,UK, the Swedish Cancer Society (project no. 680 toGlenn R. BjoÈrk) and the Swedish Natural Science Re-search Council (project no. BBU2930-150 to Glenn R.BjoÈrk).

References

Agris, P. F. (1996). The importance of being modi-®ed: roles of modi®ed nucleosides and Mg2 inRNA structure and function. Prog. Nucl. Acid Res.Mol. Biol. 53, 79±129.

Agris, P. F., Sierzputowska-Gracz, H., Smith, W.,Malkiewicz, A., Sochacka, E. & Nawrot, B. (1992).Thiolation of uridine carbon-2 restricts the motionaldynamics of the transfer RNA wobble positionnucleoside. J. Am. Chem. Soc. 114, 2652±2656.

BjoÈrk, G. G. (1992). The role of modi®ed nucleosides intRNA interactions. In Transfer RNA in Protein Syn-thesis (Hat®eld, D. L., Lee, B. J. & Pirtle, R. M., eds),pp. 23±85, CRC Press, Boca Raton, FL.


Blinkowa, A. L. & Walker, J. R. (1990). Programmedribosomal frameshifting generates the Escherichiacoli DNA polymerase III g subunit from within thet subunit reading frame. Nucl. Acids Res. 18, 1725±1729.

Boursnell, M. E. G., Brown, T. D. K., Foulds, I. J., Green,P. F., Tomley, F. M. & Binns, M. M. (1987). Com-pletion of the sequence of the genome of the coro-navirus avian infectious bronchitis virus. J. Gen.Virol. 68, 57±77.

Brierley, I. (1995). Ribosomal frameshifting on viralRNAs. J. Gen. Virol. 76, 1885±1892.

Brierley, I., Boursnell, M. E. G., Binns, M. M., Bilimoria,B., Blok, V. C., Brown, T. D. K. & Inglis, S. C.(1987). An ef®cient ribosomal frame-shifting signalin the polymerase-encoding region of the corona-virus IBV. EMBO J. 6, 3779±3785.

Brierley, I., Digard, P. & Inglis, S. C. (1989). Characteris-ation of an ef®cient coronavirus ribosomal frame-shifting signal: requirement for an RNApseudoknot. Cell, 57, 537±547.

Brierley, I., Rolley, N. J., Jenner, A. J. & Inglis, S. C.(1991). Mutational analysis of the RNA pseudoknotcomponent of a coronavirus ribosomal frameshift-ing signal. J. Mol. Biol. 220, 889±902.

Brierley, I., Jenner, A. J. & Inglis, S. C. (1992). Mutationalanalysis of the ``slippery sequence'' component of acoronavirus ribosomal frameshifting signal. J. Mol.Biol. 227, 463±479.

Chandler, M. & Fayet, O. (1993). Translational frame-shifting in the control of transposition in bacteria.Mol. Microbiol. 7, 497±503.

Chen, E. Y. & Roe, B. A. (1978). The nucleotide sequenceof rat liver tRNAAsn. Biochem. Biophys. Res. Commun.82, 235±246.

Debarros, J. P. P., Keith, G., Eladlouni, G., Glasser, C.,Mack, A. L., Dirheimer, G. & Desgres, J. (1996). 20-O-Methyl-5-formylcytidine (F(5)CM), a new modi-®ed nucleotide at the wobble position of 2 cyto-plasmic tRNAs (Leu) (NAA) from bovine liver.Nucl. Acids Res. 24, 1489±1496.

Dinman, J. D. & Wickner, R. B. (1992). Ribosomal frame-shifting ef®ciency and gag/gag-pol ratio are criticalfor yeast M1 double-stranded RNA viruspropagation. J. Virol. 66, 3669±3676.

Dinman, J. D., Icho, T. & Wickner, R. B. (1991). A ÿ1ribosomal frameshift in a double-stranded RNAvirus of yeast forms a gag-pol fusion protein. Proc.Natl Acad. Sci. USA, 88, 174±178.

Escoubas, J., Prere, M., Fayet, O., Salvignol, I., Galas, D.,Zerbib, D. & Chandler, M. (1991). Translational con-trol of transposition activity of the bacterial inser-tion sequence IS1. EMBO J. 10, 705±712.

Elseviers, D., Petrullo, L. A. & Gallagher, P. J. (1984).Novel E. coli mutants de®cient in the biosynthesisof 5-methylaminomethyl-2-thiouridine. Nucl. AcidsRes. 12, 3521±3534.

Farabaugh, P. J. (1996). Programmed translationalframeshifting. Microbiol. Rev. 60, 103±134.

Flower, A. M. & McHenry, C. S. (1990). The gammasubunit of DNA polymerase III holoenzyme ofEscherichia coli is produced by ribosomalframeshifting. Proc. Natl Acad. Sci. USA, 87, 3713±3717.

Gehrke, C. W. & Kuo, K. C. (1990). Ribonucleoside anal-ysis by reversed-phase high performance liquidchromatography. In Chromatography and Modi®cationof Nucleosides. Part A: Analytical Methods for Major

Modi®ed Nucleosides (Gherke, C. W. & Kuo, K. C. T.,eds), pp. 3±27, Elsevier, Amsterdam.

Gibson, T., Stockwell, P., Ginsburg, M. & Barrell, B.(1984). Homology between 2 EBV early genes andHSV ribonucleotide reductase and K-38 genes. Nucl.Acids Res. 12, 5087±5099.

Hagervall, T. G. & BjoÈ rk, G. R. (1984). Genetic mappingand cloning of the gene (trmC) responsible for thesynthesis of tRNA (mnm5s2U)methyltransferase inEscherichia coli K12. Mol. Gen. Genet. 196, 201±207.

Hames, B. D. (1991). An introduction to polyacrylamidegel electrophoresis. In Gel Electrophoresis of Proteins:A Practical Approach (Hames, B. D. & Rickwood, D.,eds), pp. 1±19, IRL Press, Oxford.

Hanahan, D. (1983). Studies on transformation of Escher-ichia coli with plasmids. J. Mol. Biol. 166, 557±580.

Hat®eld, D., Feng, Y.-X., Lee, B. J., Rein, A., Levin, J. G. &Oroszlan, S. (1989). Chromatographic analysis ofthe aminoacyl-tRNAs which are required for trans-lation of codons at and around the ribosomal fra-meshift sites of HIV, HTLV-1 and BLV. Virology,173, 736±742.

Hat®eld, D., Levin, J. G., Rein, A. & Oroszlan, S. (1992).Translational suppression in retroviral geneexpression. Advan. Virus Res. 41, 193±239.

Hillen, W., Egert, E., Lindner, H. J. & Gassen, H. G.(1978). Restriction or ampli®cation of wobblerecognition. FEBS Letters, 94, 361±364.

Hizi, A., Henderson, L. E., Copeland, T. D., Sowder,R. C., Hixson, C. V. & Oroszlan, S. (1987). Charac-terisation of mouse mammary tumor virus gag-progene products and the ribosomal frameshift site byprotein sequencing. Proc. Natl Acad. Sci. USA, 84,7041±7045.

Hors®eld, J. A., Wilson, D. N., Mannering, S. A.,Adamski, F. M. & Tate, W. P. (1995). Prokaryoticribosomes recode the HIV-1 gag-pol ÿ1 frameshiftsequence by an E/P site post-translocation simul-taneous slippage mechanism. Nucl. Acids Res. 23,1487±1494.

Inokuchi, H. & Yamao, F. (1995). Structure and ex-pression of prokaryotic tRNA genes. In tRNA:Structure, Biosynthesis and Function (SoÈ ll, D. &RajBhandary, U. L., eds), pp. 17±30, ASM Press,Washington, DC.

Jacks, T. & Varmus, H. E. (1985). Expression of the Roussarcoma virus pol gene by ribosomal frameshifting.Science, 230, 1237±1242.

Jacks, T., Power, M. D., Masiarz, F. R., Luciw, P. A.,Barr, P. J. & Varmus, H. E. (1988a). Characterisationof ribosomal frameshifting in HIV-1 gag-polexpression. Nature, 331, 280±283.

Jacks, T., Madhani, H. D., Masiarz, F. R. & Varmus,H. E. (1988b). Signals for ribosomal frameshifting inthe Rous sarcoma virus gag-pol region. Cell, 55,447±458.

Kramer, G. F. & Ames, B. N. (1988). Isolation andcharacterisation of a selenium metabolism mutantof Salmonella typhimurium. J. Bacteriol. 170, 736±743.

Kunkel, T. A. (1985). Rapid and ef®cient site-speci®cmutagenesis without phenotypic selection. Proc.Natl Acad. Sci. USA, 82, 488±492.

Laemmli, U. K. (1970). Cleavage of structural proteinsduring the assembly of the bacteriophage T4.Nature, 227, 680±685.

Larsen, B., Wills, N. M., Gesteland, R. F. & Atkins, J. F.(1994). rRNA-mRNA base-pairing stimulates a pro-grammed ÿ1 ribosomal frameshift. J. Bacteriol. 176,6842±6851.


Levin, M. E., Hendrix, R. W. & Casjens, S. R. (1993). Aprogrammed translational frameshift is required forthe synthesis of a bacteriophage l tail assemblyprotein. J. Mol. Biol. 234, 124±139.

Lim, V. I. (1994). Analysis of action of wobble nucleo-side modi®cations on codon-anticodon pairingwithin the ribosome. J. Mol. Biol. 240, 8±19.

Lustig, F., Elias, P., Axberg, T., Samuelsson, T.,Tittawella, I. & Lagerkvist, U. (1981). Codon read-ing and translational error. Reading of glutamineand lysine codons during protein synthesis in vitro.J. Biol. Chem. 256, 2635±2643.

Maniatis, T., Fritsch, E. F. & Sambrook, J. (1982). Molecu-lar Cloning: A Laboratory Manual, Cold Spring Har-bor Laboratory Press, Cold Spring Harbor, NY.

Miller, J. (1972). Experiments in Molecular Genetics, ColdSpring Harbor Laboratory Press, Cold Spring Har-bor, NY.

Miller, J. H., Ganem, D., Lu, P. & Schmitz, A. (1977).Genetic studies of the lac repressor I. Correlation ofmutational sites with speci®c amino acid residues:construction of a colinear gene-protein map. J. Mol.Biol. 109, 275±301.

Nam, S. H., Copeland, T. D., Hatanaka, M. & Oroszlan,S. (1993). Characterisation of ribosomal frameshift-ing for expression of pol gene products of humanT-cell leukemia virus type I. J. Virol. 67, 196±203.

Reuter, K., Slany, R., Ullrich, F. & Kersten, H. (191).Structure and function of Escherichia coli genesinvolved in biosynthesis of the deazaguanine de-rivative queuine, a nutrient factor for eukaryotes.J. Bacteriol. 173, 2256±2264.

Rogers, K. C., Crescenzo, A. T. & SoÈ ll, D. (1995). Ami-noacetylation of transfer RNAs with 2-thiouridinederivatives in the wobble position of the anticodon.Biochimie, 77, 66±74.

Sanger, F., Nicklen, S. & Coulson, A. R. (1977). DNAsequencing with chain-terminating inhibitors. Proc.Natl Acad. Sci. USA, 74, 5463±5467.

Sekine, Y. & Ohtsubo, E. (1989). Frameshifting isrequired for production of the transposase encodedby insertion sequence 1. Proc. Natl Acad. Sci. USA,86, 4609±4613.

Sekine, Y. & Ohtsubo, E. (1992). DNA sequencesrequired for translational frameshifting in pro-duction of the transposase encoded by IS1. Mol.Gen. Genet. 235, 317±324.

Sekine, Y., Eisaka, N. & Ohtsubo, E. (1994). Transla-tional control in production of transposase and intransposition of insertion sequence IS3. J. Mol. Biol.235, 1406±1420.

Sierzputowska-Gracz, H., Sochacka, E., Malkiewicz, A.,Kuo, K., Gehrke, C. W. & Agris, P. F. (1987). Chem-istry and structure of modi®ed uridines in the antic-odon, wobble position of transfer RNA aredetermined by thiolation. J. Am. Chem. Soc. 109,7171±7177.

Studier, F. W. & Moffatt, B. A. (1986). Use of bacterio-phage T7 RNA polymerase to direct selective highlevel expression of cloned genes. J. Mol. Biol. 189,113±130.

Studier, F. W., Rosenberg, A. H., Dunn, J. J. &Dubendorff, J. W. (1990). Use of T7 RNA polymer-ase to direct expression of cloned genes. MethodsEnzymol. 185, 60±89.

ten Dam, E. B., Pleij, C. W. A. & Bosch, L. (1990). RNApseudoknots: translational frameshifting and read-through on viral RNAs. Virus Genes, 4, 121±136.

ten Dam, E., Verlaan, P. & Pleij, C. (1995). Analysis ofthe role of the pseudoknot component in the SRV-1gag-pro ribosomal frameshift signal: loop lengthsand stability of the stem regions. RNA, 1, 146±154.

Tsuchihashi, Z. (1991). Translational frameshifting in theEscherichia coli dnaX gene in vitro. Nucl. Acids Res.19, 2457±2462.

Tsuchihashi, Z. & Brown, P. O. (1992). Sequence require-ments for ef®cient translational frameshifting in theEscherichia coli dnaX gene and the role of an un-stable interaction between tRNALys and an AAGlysine codon. Genes Dev. 6, 511±519.

Tsuchihashi, Z. & Kornberg, A. (1990). Translational fra-meshifting generates the gamma subunit of DNApolymerase III holenzyme. Proc. Natl Acad. Sci.USA, 87, 2516±2520.

VoÈgele, K., Schwartz, E., Welz, C., Schiltz, E. & Rak, B.(1991). High level ribosomal frameshifting directsthe synthesis of IS150 gene products. Nucl. AcidsRes. 19, 4377±4385.

Watanabe, K., Hayashi, N., Oyama, A., Nishikawa, K.,Ueda, T. & Miura, K. (1993). Unusual anticodonloop structure found in E. coli lysine tRNA. Nucl.Acids Res. 22, 79±87.

Weiss, R. B., Dunn, D. M., Shuh, M., Atkins, J. F. &Gesteland, R. F. (1989). E. coli ribosomes re-phaseon retroviral frameshift signals at rates rangingfrom 2 to 50 per cent. Nature New. Biol. 1, 159±169.

Wittwer, A. J. & Ching, W.-M. (1989). Selenium contain-ing tRNAGlu and tRNALys from Escherichia coli: puri-®cation, codon speci®city and translationalspeci®city. Biofactors, 2, 27±34.

Wittwer, A. J. & Stadtman, T. C. (1986). Biosynthesis of5-methylaminomethyl-2-selenouridine, a naturaloccurring nucleoside in E. coli tRNA. Arch. Biochem.Biophys. 248, 540±550.

Yokoyama, S. & Nishimura, S. (1995). Modi®ed nucleo-sides and codon recognition. In tRNA: Structure,Biosynthesis and Function (SoÈ ll, D. & RajBhandary,U. L., eds), pp. 207±223, ASM Press, Washington,DC.

Yokoyama, S., Watanabe, T., Murao, K., Ishikura, H.,Yamaizumi, Z., Nishimura, S. & Miyazawa, T.(1985). Molecular mechanism of codon recognitionby tRNA species with modi®ed uridine in the ®rstposition of the anticodon. Proc. Natl Acad. Sci. USA,82, 4905±4909.

Young, J. F., Desselberger, U., Graves, P., Palese, P.,Shatzman, A. & Rosenberg, M. (1983). Cloning andexpression of in¯uenza virus genes. In The Origin ofPandemic In¯uenza Viruses (Laver, W. G., ed.),pp. 129±138, Elsevier Science, Amsterdam.

Edited by J. Karn

(Received 28 February 1997; received in revised form 8 May 1997; accepted 9 May 1997)

Expression of a Coronavirus Ribosomal Frameshift Signal in Escherichia coli: Influence of tRNA Anticodon Modification on ...IntroductionResultsFigure 1Figure 2Figure 3Figure 4Figure 5Table 1Table 2Sequence requirements for IBV frameshifting in E. coliSlippery sequence requirementsRNA secondary structure requirements

The role of tRNA anticodon modification in frameshiftingThe role of QThe role of mnm 5 s 2 U

DiscussionSignals for frameshifting in E. coliInfluence of tRNA anticodon modification on frameshifting

Materials and MethodsSite-specific mutagenesisConstruction of plasmidsBacterial strains and culture conditionsFrameshifting in E. coli BL21Frameshifting in modification-deficient E. coli strains

AcknowledgementsReferences

expression of a coronavirus ribosomal frameshift signal in … expression... · 2020. 2. 1. ·...

Documents