trans complementation by rna of defective foot-and-mouth

JOURNAL OF VIROLOGY, Feb. 1994, p. 697-703 Vol. 68, No. 20022-538X/94/$04.00+0Copyright © 1994, American Society for Microbiology

trans Complementation by RNA of Defective Foot-and-MouthDisease Virus Internal Ribosome Entry Site Elements

JEFF DREW AND GRAHAM J. BELSHAM*AFRC Institute for Animal Health, Pirbright, Woking, Surrey GU24 ONF, United Kingdom

Received 6 July 1993/Accepted 29 October 1993

A region of about 435 bases from the 5' noncoding region of foot-and-mouth disease virus RNA directsinternal initiation of protein synthesis. This region, termed the internal ribosome entry site (IRES), ispredicted to contain extensive secondary structure. Precise deletion of five predicted secondary structurefeatures has been performed. The mutant IRES elements have been constructed into vectors which expressbicistronic mRNAs and assayed within cells. Each of the modified IRES elements was defective in directinginternal initiation when assayed alone. However, coexpression of an intact foot-and-mouth disease virus IREScomplemented four of these defective elements to an efficiency of up to 80o of wild-type activity. Nocomplementation was observed with the structurally analogous element from encephalomyocarditis virus. Therole of RNA-RNA interactions in the function of the picornavirus IRES is discussed.

The initiation of protein synthesis on picornavirus RNAsoccurs at AUG codons which are 600 to 1300 nucleotides (nt)downstream from the 5' termini of the molecules. No cap

structure is present at the 5' terminus of the viral RNA, incontrast to cellular mRNAs. A number of features suggest thatthe initiation of protein synthesis on these RNAs must occur bya mechanism which is distinct from that described for cellularmRNAs (see reviews in references 11 and 16). The long 5'noncoding regions (NCRs) of picornavirus RNAs containseveral unused AUG codons, and these regions are predictedto form extensive secondary structure. Translation of many

picornavirus RNAs also has to occur when cap-dependenttranslation is abolished following the cleavage of the cap-

binding complex (eIF-4F) component p220. In representativesof each of the genera of these viruses, evidence has beenobtained for the presence of an element which directs internalinitiation of protein synthesis (2, 5, 12, 17, 26). This element isusually termed the internal ribosome entry site (IRES).There is considerable conservation of these structures within

certain groups of picornaviruses. The IRES elements fromaphthoviruses (foot-and-mouth disease viruses [FMDV]) andcardioviruses (e.g., encephalomyocarditis virus [EMCV]) are

about 50% identical in sequence, and their predicted second-ary structures are very similar (29) (Fig. 1). These structureshave been defined on the basis of phylogenetic comparisonsand biochemical probing of these RNA molecules. The entero-virus (e.g., poliovirus [PV]) and rhinovirus IRES elements alsoshow extensive similarity in their predicted secondary struc-tures (30, 33). However these elements are very different insequence and predicted structure from the aphthovirus andcardiovirus elements. A conserved feature across all of theseelements is the presence of a pyrimidine tract about 25 ntupstream from an AUG codon. In the case of the cardiovirusesand aphthoviruses, this AUG codon is an initiation site ofprotein synthesis. In contrast, in the enteroviruses and rhino-viruses, a region of up to 120 nt is present between the AUGcodon associated with the pyrimidine tract and the initiationcodon. Mutagenesis of this cryptic AUG codon in PV has

* Corresponding author. Mailing address: Department of Biochem-istry, McIntyre Medical Sciences Building, McGill University, 3655Drummond St., Montreal, Quebec, Canada H3G 1Y6. Phone: (514)398-7275. Fax: (514) 398-1287.

shown that it is not essential for virus viability. However,uniquely among the upstream AUG codons, its modification isdeleterious to the virus. The mutant has a small-plaquephenotype (25). In FMDV, two initiation codons, some 84 ntapart, are used to initiate protein synthesis. It has been shownthat some ribosomes scan through the region following the firstinitiation codon, presumably following recognition of the RNAat a point upstream of the first initiation codon (1).

Considerable attention has been focused on determining themechanism by which the picornavirus IRES elements function.Several studies have identified cellular proteins which interactwith them. The IRES elements of FMDV and EMCV cross-link principally to a protein of 57 or 58 kDa. Two binding siteshave been mapped for this protein on the FMDV IRES (18).Only a single site has been identified in EMCV (13). The latteris analogous to the 5' proximal site in FMDV. A distinctprotein of 52 kDa has been shown to cross-link to the PV IRES(21), although some evidence for interaction with the 57-kDaprotein has also been presented (28). Recently evidence hasbeen obtained (3) that the 58-kDa protein is the polypyrimi-dine tract-binding protein (PTB), which has been previouslyassigned a function within the nucleus in the process ofpre-mRNA splicing (24). The 52-kDa protein has been re-cently identified as La, a protein involved in maturation ofRNA polymerase III transcripts, also located within the nu-cleus (20). The role of both of these proteins in IRES functionis not yet clear. Mutagenesis of the 58-kDa protein binding sitein the EMCV IRES indicated that loss of protein binding wasassociated with loss of ability to direct internal initiation ofprotein synthesis (13). However, other studies (7) suggestedthat this region of the IRES is not essential for the IRES tofunction, and hence the importance of this interaction is notcertain.

Extensive analysis of the PV IRES has been performed. Thesequence from nt 140 to 630 was initially defined as beingnecessary to direct internal initiation of protein synthesis.However, more recent studies (22, 27) have shown that twopredicted stem-loop structures within this sequence are notessential for IRES activity. Indeed, the 3'-terminal structure,containing the cryptic AUG, can be deleted from full-lengthvirion RNA without abolishing infectivity. However, the vi-ruses recovered have regenerated an AUG codon in anappropriate position (31).

697

on January 30, 2018 by guesthttp://jvi.asm

.org/D

ownloaded from

http://jvi.asm.org/

698 DREW AND BELSHAM

FIG. 1. Secondary structure prediction for the FMDV IRES. Thenomenclature used for the different predicted elements within thestructure is indicated. The structure was generated and presented byusing the FOLD and SQUIGGLES programs (6). The precise se-quences removed by each of the indicated four domain deletions andthe hammerhead structure (deletion 5) are indicated in Table 1. The 5'terminus of the IRES is indicated by the filled circle, and the 3'terminus is represented by the filled oval.

Surprisingly, studies of the PV IRES showed that coexpres-sion of full-length PV RNA within cells containing defectivePV IRES elements partially restored activity to the defectiveelements (27). Furthermore, these studies recently have beenextended to show that merely coexpression of the 5' NCR ofPV with the defective IRES elements efficiently restoredfunction (35). These studies suggested that complementationin trans between IRES elements occurred.

In this study, we have investigated the requirement for eachof the major predicted stem-loop structures in the FMDVIRES for efficient internal initiation of protein synthesis.Furthermore, we have sought further evidence for the ability ofan intact IRES element to complement defective IRES ele-ments. Since the FMDV IRES is very different in sequence andpredicted structure from the PV IRES, equivalent resultswould strengthen the view that such complementation mayhave a role in the biology of picornaviruses. We now show thatfour of five defective FMDV IRES elements tested can beefficiently complemented in trans by an intact FMDV IRES.The loss of the largest motif renders the IRES defective whenassayed alone or in the presence of an intact IRES.

MATERIALS AND METHODS

Plasmid constructions. Standard methods (32) were usedfor the manipulation and growth of plasmid DNAs, which werepurified by CsCl density gradient centrifugation prior to use inthe transient expression assay. Plasmid pKSGC (Fig. 2) ex-presses a bicistronic mRNA transcript, encoding 3-glucuroni-dase (GUS) (14) and chloramphenicol acetyltransferase(CAT) (10) reporter proteins, under the control of the T7promoter. The FMDV IRES, as an EcoRI-Clal fragmentderived from pKSMRHMRI.Cla (2), was inserted into pKS+(Stratagene) to produce pKSRCla and used as the template fordeletion analysis. The same fragment was also inserted intoEcoRI- and Clal-digested pSK+ to produce pSKRCla, so thatan antisense transcript is expressed from the T7 promoter (Fig.2). The SmaI-Clal fragment from pKSRCla containing thewild-type (wt) FMDV IRES was blunt ended and inserted intothe vector pKSGC, previously digested with HindlIl and bluntended, to produce pKSGSCC (Fig. 2).The PCR was performed essentially as described previously

GUS CAT

pKSGC=

(Smal) IRES

pKSGSCC OEZIMEEcoRI (Clal)

Hindlll

pKSGA1-5C5C(EcoRI) (Clal)

pKSRCIa E _lEcoRI Cla

Hindlil

pSKRCIa

Clal EcoRI

pSKEMCRBEcoRl BamHI

FIG. 2. Structure of the plasmids expressing bicistronic mRNAsand the FMDV and EMCV IRES elements. The T7 promoter isindicated by filled circles. Note the opposite orientation of the FMDVIRES (black rectangles) in pKSRCla (sense orientation) and pSKRCIa(antisense orientation).

(32). The EMCV IRES element was produced by usingprimers GTGGCCATATAAGGATCCTG7'FF'ITTC and GGGAGACCGGAATTCCGCC, which create BamHI andEcoRI restriction sites near the termini of the PCR fragment(about 550 bases in length), using plasmid pE5LVPO (23) asthe template. The BamHI site is at the position of AUG10,some 8 nt upstream of the initiation site. The EcoRI-BamHI-digested PCR product was ligated into similarly digestedpSK+ to produce pSKEMCRB (Fig. 2). This EMCV elementfunctions efficiently to direct internal initiation of proteinsynthesis when assayed within cells in the context of a bicis-tronic mRNA transcript (data not shown).The precise deletion of regions within the FMDV IRES was

performed by the PCR in two stages. One set of reactions wasperformed with the forward sequencing primer (GTAAAACGACGGCCAGT) together with specific primers 1 (Table 1),which introduced the required deletions. Separate reactionsusing the complementary mutagenic primers (primers 2; Table1) together with the reverse sequencing primer (AACAGCTATGACCATG) were also performed. The half-reaction prod-ucts were purified by agarose gel electrophoresis and mixedappropriately, and further PCRs were performed with only theforward and reverse primers. The fragments obtained were gelpurified, digested with EcoRI and Clal, and ligated intoEcoRI- and Clal-cut pSK+ or pKS+. Modified FMDV ele-ments were constructed into the pKSGC vector as describedabove for the wt element except that blunt-ended EcoRI-ClaIfragments were used. All mutants were sequenced by directdouble-stranded DNA sequencing using a T7 DNA poly-merase-based sequencing system (Amersham).

Transient expression assays. Plasmids were assayed bytransfection, using Lipofectin (Life Technologies), into HTK-143 cells infected with the recombinant vaccinia virus vTF7-3(8), which expresses the T7 RNA polymerase as describedpreviously (2). After 20 h, cell extracts were prepared andassayed for GUS activity in a fluorescence assay (14) and forCAT protein in a quantitative enzyme-linked immunosorbentassay (ELISA) (Boehringer Mannheim).

J . VIROL.


.org/D

ownloaded from

http://jvi.asm.org/

IRES COMPLEMENTATION 699

TABLE 1. Deletions introduced into the FMDV IRES

Plasmid Mutagenic oligonucleotide Sequence deleted"

pKSGSCC NonepKSGllC 1, ACGAATTCCGCTGTGCTTGA -432 to -379

2, CGAGCGTGGAGTCAAGCACAGCGGAATTCGTpKSGA2C 1, CACTTTGTACTGCTGGTGACAGGC -375 to -164

2, CATCCTTAGCCTGTCACCAGCAGTACAAAGTGpKSGA3C 1, AGCACTGGTGACGAATAGGTGACC - 155 to -48

2, GCCTCCGGTCACCTATTCGTCACCAGTGCTpKSGA4C 1, TACGCCTGAATATTTCTTTTATA -42 to -23

2, GTGGTTATAAAAGGAAATATTCAGGCGTApKSGA5C 1, GGAGCATGGTGGCGGCATGCCCATCATGTGTG - 302 to - 269,

2, CACACATGATGGGCATGCCGCCACCATGCTCC -266 to -258, and -253 to -244

"The numbering system used positions the A of the first initiation codon as nucleotide 1.

RESULTS

Determination of secondary structure motifs required forFMDV IRES function. The secondary structure prediction forthe FMDV IRES, essentially as derived previously (29), hasbeen used as the basis for precisely deleting predicted domainswithin the IRES (Fig. 1). Very similar models were proposedfor the FMDV IRES and for the element from EMCV.Different laboratories have used alternative nomenclatures (7,13) for the various predicted secondary structure elementswithin the cardiovirus and aphthovirus IRES. A simpler systemis shown in Fig. 1 and was used throughout this study. Analternative model for the EMCV IRES (7) is similar in manyrespects. However, at the 3' terminus, several short stem-loopstructures (termed L, M, and N) which are not compatible withthe FMDV sequence are predicted by the latter model. Withinthis region, the latter model (7) also includes the polypyrimi-dine tract in secondary structure, in contrast to the modelspredicted for FMDV and EMCV previously (29). Alternativefoldings of short sequences within the largest domain (domain2) remain unresolved but do not affect the experiments under-taken in this study.

a)pKSGGIC 1

pKSG.62C 100

pKSGG3C 39pKSGtbC

pKSGSCC

pKSGC

C)pKSG Al C

pKSGQ2CpKSGA3C

pKSGLACpKSG.5CpKSGSCC

0 20 40 60 81

_ 102

_100- 125

pKSGC 5

0 20 40 60 80 100 120 140

Each of the deletions indicated Fig. 1 and Table 1 wasconstructed by PCR as described in Materials and Methods.Each deletion was designed to precisely delete a predictedsecondary structure feature. Note that deletion 5 removes onlythe hammerhead structure at the end of domain 2. The PCRfragments generated were cloned, sequenced, and constructedinto vectors expressing bicistronic mRNAs transcribed fromthe T7 RNA polymerase promoter. The vectors contain theGUS gene as an indicator of cap-dependent translation, andthis activity also verifies the expression of the plasmid withinthe cells. The efficient expression of CAT requires internalinitiation of protein synthesis and is dependent on an activeFMDV IRES.Each of the plasmids expressing bicistronic mRNAs contain-

ing the wt or mutant IRES elements efficiently expressed GUSactivity (Fig. 3a). However, marked differences in the level ofCAT expression were observed (Fig. 4a). The expression ofCAT directed by the wt IRES (plasmid pKSGSCC) served asa positive control, whereas the lack of expression from aplasmid (pKSGC) containing the GUS and CAT genes butlacking any IRES element (Fig. 2) served as a control for any

b)pKSG61C

pKSGL22C

pKSGA3C 7

pKSGA4C _5

pKSGA5C 6

pKSGSCC

pKSGC _

0 20 40 60d)pKSG 1CpKSG62C

pKSG,5C

pKSGA5C

pKSGSCC

pKSGC

I*+pKSRCla

0 20 40 60 80 100 120 140

FIG. 3. Effects of coexpression of IRES elements on the expression of GUS activity. The indicated plasmids were transfected into

vTF7-3-infected cells alone (a) or with pKSRCla (wt FMDV IRES; b), pSKRCla (antisense FMDV IRES; c) or pSKEMCRB (EMCV IRES; d).After 20 h, cell extracts were prepared and assayed for GUS activity. GUS activity from pKSGSCC assayed alone was set at 100% in each

experiment, and other values were related to it. Mean values from four independent determinations are shown except for panel d, in which valuesare from a single experiment repeated twice with similar results.

* 100 120 140

f1 7 |*+pSKRCIa

4

*19

in2

*1

15i

80 100 120 140

-38 *+pSKEMCRB_48

2

N 2

-34_8

ND

r-6

pKSGu,n3G

VOL. 68, 1994


.org/D

ownloaded from

http://jvi.asm.org/


a)pKSG A1 C

pKSGt62CpKSGA3CpKSGA4C

pKSGt65CpKSGSCC 100

pKSGC

C)pKSGA1C

pKSG 02C

pKSGA3C

pKSG 64C

pKSGt65CpKSGSCC

pKSGC

b)pKSGA1 C m

pKSG62C

pKSGA3CpKSGt64CpKSG&5C

pKSGSCC 1

0

pKSGC

d)pKSGA1CpKSG&lC

pKSGb3C

pKSGMnCpKSG65C

pKSGSCC

pKSGC

0 20 40 60 80 100 120 0 20 40

0O

FIG. 4. Complementation of defective FMDV IRES elements by coexpression of the wt FMDV IRES. Cell extracts prepared as described forFig. 3 were also assayed for CAT expression by using a quantitative ELISA. CAT expression from plasmid pKSGSCC (containing the wt FMDVIRES) assayed alone was set at 100%, and CAT expression from other constructs was compared with this value in each experiment. Appropriatedilutions of cell extracts were used so that the assays were performed within the range of the standard curve. The results presented are the meanvalues obtained from four independent determinations except for panel d, in which values are from a single experiment repeated twice with similarresults.

expression resulting from reinitiation. Essentially no CATexpression was detected from this construct (Fig. 4a). Individ-ual deletion of domains 2, 3, and 5 from the FMDV IRESreduced CAT expression to almost the background level.However, significant but low-level expression (ca. 3% of the wtlevel) of CAT was observed when domains 1 and 4 were

individually deleted (Fig. 4a).Complementation of mutant FMDV IRES elements. To

determine whether the defects in the ability of the defectiveIRES elements were complementable in trans, each of theplasmids expressing the bicistronic mRNAs was cotransfectedwith a plasmid (pKSRCla) expressing the intact FMDV IRES.The results obtained for the IRES-directed expression of CATare presented in Fig. 4. The results, mean values from fourdeterminations, are presented as a percentage of the CATexpression obtained from plasmid pKSGSCC, which containsthe wt FMDV IRES. The coexpression of the FMDV IRES,unlinked to any open reading frame, from plasmid pKSRClaenhanced expression of CAT from mutants with domains 1, 3,4, and 5 individually deleted (Fig. 4b). The level of CATexpression observed was between 12 and 81% of that observedfrom the wt IRES. The deletion of domain 1 was particularlyefficiently complemented. No enhanced CAT activity was ever

observed from the construct lacking the large domain 2 or

lacking all IRES sequences. It is noteworthy that the plasmidcontaining deletion 5 (deletion of only part of domain 2 [thehammerhead]) was very efficiently complemented. It may bethat deletion of all of domain 2 interferes with the ability of therest of the IRES to adopt its native conformation.

Coexpression of the intact FMDV IRES had some inhibitoryeffect on the expression of GUS from the plasmids (compareFig. 3a and b), but this effect was very similar in each case andprobably reflects competition for cellular components.The high efficiency of the complementation and also the

inability to complement the deletion of domain 2 both arguestrongly against the effect resulting from some form of recom-bination event between the plasmid DNAs. As a furthercontrol against this possibility, cotransfection with a plasmid

(pSKRCla) containing the FMDV IRES in the oppositeorientation to the T7 promoter was also performed. Underthese conditions, no complementation, as indicated by en-

hanced CAT expression, of the defective mutants lackingdomain 1, 3, 4, or 5 was observed (Fig. 4c). Very surprisingly,however, a consistent enhancement of CAT expression fromthe domain 2 deletion construct was observed. The antisenseIRES also inhibited GUS expression (Fig. 3c). The latterfinding was observed previously (35) in analogous studies withthe PV IRES. The inhibition of GUS activity by the antisenseIRES was most marked on the RNAs which contained theFMDV IRES. A substantially less severe inhibition of GUSexpression from pKSGC was observed. This finding indicatesthat the interaction with the antisense RNA reduces theactivity of the bicistronic mRNA in translation, perhaps byreducing its stability.To determine whether the structurally related EMCV IRES

was able to complement the defective FMDV IRES elements,we constructed a plasmid (pSKEMCRB; Fig. 2) containing thiselement unlinked to any open reading frame. The elementused extends from the poly(C) tract to AUG10 just 8 basesupstream from the initiation codon. This EMCV IRES was

unable to complement any of the defective FMDV IRESelements (Fig. 4d) but could complement deletions within theEMCV IRES when assayed with analogous plasmids (2a). Ithad an inhibitory effect similar to that of the FMDV IRES onthe expression of GUS activity (Fig. 3d).

DISCUSSION

The FMDV and EMCV IRES are strikingly similar inpredicted secondary structure. There is considerable agree-ment among several laboratories that a segment of about 450bases is sufficient to direct efficient internal initiation of proteinsynthesis. However, some contradictory data on the precise 5'terminus of the element have been presented. Studies on theFMDV element (2) and the EMCV element (13) suggestedthat loss of sequence within domain 1 grossly impaired the

0.1

0.2

*12

*13

_J1

*+pKSRCla0.2

20 40 60 80 100 1

1.3

0

.8

1.2

_27

1.3 |*+pSKRCIa

0 20 40 60 80 100 121

, .4

.3

.3

.5

:0.3

I; 9

:ND | *+pSKEMCRB

d)~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

11 -A A . .11nrs 14on60 80 100 120

J. VIROL.

I'll


.org/D

ownloaded from

http://jvi.asm.org/


activity of the element. However, other studies (17) reportedthat partial loss of domain 1 of FMDV only had a modest effect(50% loss) on translational efficiency. Similarly, it has beenreported that partial loss of sequence from this terminalstem-loop structure in EMCV impaired activity by about 70%,while complete loss of the element (and loss of a few basesfrom the next stem) resulted in a deficiency of only about 35%compared with the wt element (7). These results are ofimportance since this terminal structure is reported to be themajor site of binding of the 58-kDa protein (PTB) whichinteracts with each of these elements. Resolution of thesevarious results is not entirely clear. The data suggesting aminimal role for domain 1 have been obtained by using in vitrotranslation reactions on monocistronic RNAs, whereas theevidence for a requirement for this structure has been obtainedby using bicistronic mRNAs in cell culture. It has beensuggested that partial motifs may interfere with other struc-tures and hence give false answers (7). The presence of theupstream open reading frame in the bicistronic mRNAs mayalso perturb the structure of partial motifs. The precise dele-tion of this element in this study, without any predicted effecton the neighboring structures, severely inhibited the ability ofthe FMDV IRES to direct internal initiation, although somelow-level (ca. 3% of the wt level) activity persisted. This resultwas consistent with our previous observations that the partialloss of this element made it severely defective (2).Only limited mutagenesis of domains 2 and 3 has been

performed previously. Linker insertions into the hammerheadstructure at the end of domain 2 (as deleted in the domain 5deletion) and into domain 3 completely abolished the functionof the EMCV IRES (7). The results presented here show thatprecise deletion of the hammerhead structure and domain 3within the FMDV IRES also totally abolished function.We have used the secondary structure model for the 3' end

of the IRES proposed by Pilipenko et al. (29), since the L, M,and N motifs suggested by Duke et al. (7) for the EMCVsequence are not compatible with the FMDV sequences. Thestem-loop structure proposed by Pilipenko et al. (29) is com-patible with both FMDV and EMCV sequences and leaves thepolypyrimidine tract unstructured. Precise deletion of thepredicted 3'-terminal motif (domain 4), leaving the polypyri-midine tract intact, also severely inhibited the activity of theIRES, although the activity was above the background level.Complementation of defective IRES elements. The FMDV

IRES is substantially different from that of PV. The sequencehas little similarity, and the secondary structure prediction iscompletely different. However, they do have the same biolog-ical activity. In each case, not surprisingly, it has been shownthat precise deletion of predicted secondary structure motifscan abolish the activity of the element when assayed alone.Deletion of domain IV from the PV IRES produces anelement which retains significant activity, but none of majorfeatures of the FMDV IRES appear dispensable. More signif-icantly, upon coexpression with an intact homologous element,unlinked to any other gene, enhanced activity is observed fromthe defective elements within bicistronic mRNAs. Recombina-tion between the plasmid DNAs is ruled out by three lines ofevidence. First, the efficiency of this unselected event is veryhigh; up to 80% of the wt activity is observed from thedefective elements when coexpressed with the intact IRES.Second, no enhanced CAT expression is observed when aplasmid containing the identical FMDV cDNA sequence, butin the opposite sense to the T7 promoter, is used (the oneexception is discussed below). This plasmid should be able torecombine with the defective IRES elements in an equivalentmanner to the construct containing the IRES in the opposite

orientation. Third, the construct lacking domain 2 did notexhibit enhanced CAT expression when the wt IRES wascoexpressed. If recombination was responsible for the restora-tion of CAT expression from the defective elements, onewould expect that this defect would also be corrected.

It may be significant that the deletion of domain 1 is themost efficiently complemented (up to 81% of wt IRES activi-ty). This is the region of the IRES to which different labora-tories attribute very distinct properties (as discussed above)and to which p57/p58 (PTB) binds. It may be that this terminalregion has an enhancing function as proposed for the stem-loop structure only on the 5' side of this domain in EMCV(13), which may also constitute a binding site for p57. Thus, itsactivity may be fully revealed only under stringent assayconditions, such as when the RNA is assayed in competitionwith other RNAs (cf. in vitro translation assays). However,since the EMCV IRES failed to complement even the deletionof this motif from the FMDV IRES, this element must domore than simply bind PTB.The structurally related EMCV IRES was unable to com-

plement any of the defective FMDV elements. This closelymirrors the inability of the rhinovirus 5' NCR to complementthe defective PV IRES elements and the very limited ability ofthe more closely related coxsackievirus B4 5' NCR to performthe same function (35). These elements are about 60 and 70%,respectively, identical to the PV IRES over this region of thegenome. This is a higher degree of identity than between theEMCV and FMDV IRES (about 50%).

Deletion of any sequence up to nt 528 within the PV 5' NCR(742 nt long) has proved to be complementable, but deletion ofsequences at the 3' end of the IRES, including the region fromnt 528 to 620, were not complemented by the cotransfection ofthe intact PV genome (27). Four independent deletions of theFMDV IRES have been complemented. However, it was notpossible to complement, with the FMDV IRES, the deletion ofthe largest element, domain 2. It may be that the tertiarystructure of the IRES is severely compromised by this largedeletion. It is of interest that a single point mutation at thebase of this large domain enhanced the activity of the type CFMDV IRES (19), suggesting that this domain is key to thestructure and/or activity of the IRES.The strangest result that we observed was the enhancement

of CAT expression directed by the defective element lackingdomain 2 when coexpressed with the antisense transcript of theFMDV IRES. We have no very satisfactory explanation forthis phenomenon at present. Antisense oligonucleotides havebeen shown previously to enhance the translation of maskedmRNAs (34). It was proposed that this resulted from compe-tition with cellular proteins which were binding to the RNAand inhibiting its translation. It is difficult to adapt this conceptto the enhanced activity of the defective IRES, however. Analternative possibility is that the antisense transcript, whenfolded by itself, forms a structure which allows the defectiveIRES to partially anneal to it and form a scaffold for theresidual features of the defective IRES to function.

All of the data so far obtained indicating complementationof defective IRES elements by the coexpression of an intactIRES have been obtained by using the vaccinia virus-T7 RNApolymerase expression system (this report and reference 35).The cellular environment is clearly modified by the vacciniavirus infection. However, there is considerable accord betweenthe results obtained in this assay system for other aspects ofIRES function and those obtained by other workers usingalternative strategies (2, 17, 22, 27, 35). An advantage of thevaccinia virus system is that it produces high levels of cytoplas-

VOL. 68, 1994


.org/D

ownloaded from

http://jvi.asm.org/


mic transcripts as obtained with a natural picornavirus infec-tion.Two models for complementation of the PV IRES were

suggested (35). One model suggests an RNA-RNA interactionin which the defective (and missing) element in one IRES issubstituted by that domain from the intact element so that afunctional unit is re-formed. This would be analogous to theregeneration in vitro of a functional EMCV IRES from twomolecules by the melting and annealing of overlapping tran-scripts (4). The cosynthesis of the RNA transcripts within cellsin our studies could obviate the need for such a meltingprocedure. The second model involves the transfer of a

functional initiation complex from the functional IRES ele-ment to some point on the defective IRES structure in trans. Inthe case of FMDV and EMCV, this seems particularly attrac-tive since previous studies (15) suggest that the viral RNA isrecognized with respect to its ability to initiate protein synthe-sis from a point within 8 nt of the initiation codon itself. Thus,the region of the RNA in the immediate vicinity of theinitiation codon may serve as an acceptor site for a preformedinitiation complex. Very recently, it was shown (9) that a

nonlinear migration of ribosomes on cauliflower mosaic virusRNA, which was demonstrated on individual RNA molecules,could also be achieved between two RNA molecules. Theseauthors suggested analogous donor/acceptor regions for theribosome shunt mechanism proposed.

Since the complementation of defective IRES elements hasbeen observed for both major types of element found withinthe picornavirus family, it seems likely that this activity isimportant in the life cycle of these viruses.

ACKNOWLEDGMENTS

We thank Julia Brangwyn for excellent technical assistance.J.D. acknowledges receipt of an IAH studentship.

REFERENCES

1. Belsham, G. J. 1992. Dual initiation sites of protein synthesis onfoot-and-mouth disease virus RNA are selected following internalentry and scanning of ribosomes in vivo. EMBO J. 11:1105-1110.

2. Belsham, G. J., and J. K. Brangwyn. 1990. A region of the 5'noncoding region of foot-and-mouth disease virus RNA directsefficient internal initiation of protein synthesis within cells: in-volvement with the role of L protease in translational control. J.Virol. 64:5389-5395.

2a.Belsham, G. J., J. K. Brangwyn, and A. van der Velden. Unpub-lished data.

3. Borman, A., M. T. Howell, J. G. Patton, and R. Jackson. 1993. Theinvolvement of a spliceosome component in internal initiation ofhuman rhinovirus RNA translation. J. Gen. Virol. 74:1775-1788.

4. Borovjagin, A. V., M. V. Ezrokhi, V. M. Rostapshov, T. Y. Ugarova,T. F. Bystrova, and I. N. Shatsky. 1991. RNA-protein interactionswithin the internal translation initiation region of encephalomyo-carditis virus RNA. Nucleic Acids Res. 19:4999-5005.

5. Brown, E. A., S. P. Day, R. W. Jansen, and S. M. Lemon. 1991. The5' nontranslated region of hepatitis A virus RNA: secondarystructure and elements required for translation in vitro. J. Virol.65:5828-5838.

6. Devereux, J., P. Haeberli, and 0. Smithies. 1984. A comprehensiveset of sequence analysis programs for the VAX. Nucleic AcidsRes. 12:387-395.

7. Duke, G. M., M. Hoffman, and A. C. Palmenberg. 1992. Sequenceand structural elements that contribute to efficient encephalomyo-carditis viral RNA translation. J. Virol. 66:1602-1609.

8. Fuerst, T. R., E. G. Niles, F. W. Studier, and B. Moss. 1986.Eukaryotic transient expression system based on recombinantvaccinia virus that synthesizes bacteriophage T7 RNA polymerase.Proc. Natl. Acad. Sci. USA 83:8122-8126.

9. Futterer, J., Z. Kiss-Laszlo, and T. Hohn. 1993. Nonlinear ribo-

some migration on cauliflower mosaic virus 35S RNA. Cell73:789-802.

10. Gorman, C. M., B. M. Moffat, and B. H. Howard. 1982. Recom-binant genomes which express chloramphenicol acetyltransferasein mammalian cells. Mol. Cell. Biol. 2:1044-1051.

11. Jackson, R. J., M. T. Howell, and A. Kaminski. 1990. The novelmechanism of picornavirus translation. Trends Biochem. Sci.15:477-483.

12. Jang, S. K., H.-G. Krausslich, M. J. H. Nicklin, G. M. Duke, A. C.Palmenberg, and E. Wimmer. 1988. A segment of the 5' nontrans-lated region of encephaloymyocarditis virus RNA directs internalentry of ribosomes during in vitro translation. J. Virol. 62:2636-2643.

13. Jang, S. K., and E. Wimmer. 1990. Cap-independent translation ofencephalomyocarditis virus RNA: structural elements of the inter-nal ribosome entry site and involvement of a 57kD RNA bindingprotein. Genes Dev. 4:1560-1572.

14. Jefferson, R. A. 1987. Assaying chimeric genes in plants: the GUSgene fusion system. Plant Mol. Biol. Rep. 5:387-405.

15. Kaminski, A., M. T. Howell, and R. J. Jackson. 1990. Initiation ofencephalomyocarditis virus RNA translation: the authentic initia-tion site is not selected by a scanning mechanism. EMBO J.9:3753-3759.

16. Kozak, M. 1989. The scanning model for translation: an update. J.Cell Biol. 108:229-241.

17. Kuhn, R., N. Luz, and E. Beck. 1990. Functional analysis of theinternal initiation site of foot-and-mouth disease virus. J. Virol.64:4625-4631.

18. Luz, N., and E. Beck. 1991. Interaction of a cellular 57-kilodaltonprotein with the internal initiation site of foot-and-mouth diseasevirus. J. Virol. 65:6486-6494.

19. Martinez-Salas, E., J.-C. Saiz, M. Davila, G. J. Belsham, and E.Domingo. 1993. A single nucleotide substitution in the internalribosome entry site of foot-and-mouth disease virus leads toenhanced cap-independent translation in vivo. J. Virol. 67:3748-3755.

20. Meerovitch, K., H. S. Lee, Y. Svitkin, D. J. Kenan, E. K. L. Chan,V. I. Agol, J. D. Keene, and N. Sonenberg. 1993. La autoantigenenhances and corrects translation of poliovirus RNA in reticulo-cyte lysate. J. Virol. 67:3798-3807.

21. Meerovitch, K., J. Pelletier, and N. Sonenberg. 1989. A cellularprotein that binds to the 5' non-coding region of poliovirus RNA:implications for internal translation initiation. Genes Dev. 3:1026-1034.

22. Nicholson, R., J. Pelletier, S.-Y. Le, and N. Sonenberg. 1991.Structural and functional analysis of the ribosome landing pad ofpoliovirus type 2: in vivo translation studies. J. Virol. 65:5886-5894.

23. Parks, G. D., G. M. Duke, and A. C. Palmenberg. 1986. Encepha-lomyocarditis 3C protease: efficient cell-free expression fromclones which link viral 5' noncoding sequences to the P3 region. J.Virol. 60:376-384.

24. Patton, J. G., S. A. Mayer, P. Tempst, and B. Nadal-Ginard. 1991.Characterization and molecular cloning of a polypyrimidine tract-binding protein: a component of a complex necessary for pre-mRNA splicing. Genes Dev. 5:1237-1251.

25. Pelletier, J., M. E. Flynn, G. Kaplan, V. R. Racaniello, and N.Sonenberg. 1988. Mutational analysis of upstream AUG codons ofpoliovirus RNA. J. Virol. 62:4486-4492.

26. Pelletier, J., and N. Sonenberg. 1988. Internal initiation of trans-lation of eukaryotic mRNA directed by a sequence derived frompoliovirus RNA. Nature (London) 334:320-325.

27. Percy, N., G. J. Belsham, J. K. Brangwyn, M. Sullivan, D. M.Stone, and J. W. Almond. 1992. Intracellular modifications in-duced by poliovirus reduce the requirement for structural motifs inthe 5' noncoding region of the genome involved in internalinitiation of protein synthesis. J. Virol. 66:1695-1701.

28. Pestova, T. V., C. U. T. Hellen, and E. Wimmer. 1991. Translationof poliovirus RNA: role of an essential cis-acting oligopyrimidineelement within the 5' nontranslated region and involvement of acellular 57-kilodalton protein. J. Virol. 65:6194-6204.

29. Pilipenko, E. V., V. M. Blinov, B. K. Chernov, T. M. Dmitrieva,

J. VIROL.


.org/D

ownloaded from

http://jvi.asm.org/


and V. I. Agol. 1989. Conservation of the secondary structureelements of the 5'-untranslated region of cardio- and aphthovirusRNAs. Nucleic Acids Res. 17:5701-5711.

30. Pilipenko, E. V., V. M. Blinov, L. I. Romanova, A. N. Sinyakov,S. V. Maslova, and V. I. Agol. 1989. Conserved structural domainsin the 5'-untranslated region of picornaviral genomes: an analysisof the segment controlling translation and neurovirulence. Virol-ogy 168:201-209.

31. Pilipenko, E. V., A. P. Gmyl, S. V. Maslova, Y. V. Svitkin, A. N.Sinyakov, and V. I. Agol. 1992. Procaryotic-like cis elements in thecap-independent internal initiation of translation on picomavirusRNA. Cell 68:119-131.

32. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecularcloning: a laboratory manual, 2nd ed. Cold Spring Harbor Labo-

ratory, Cold Spring Harbor, N.Y.33. Skinner, M. A., V. R. Racaniello, G. Dunn, J. Cooper, P. D. Minor,

and J. W. Almond. 1989. New model for the secondary structure ofthe 5' non-coding RNA of poliovirus is supported by biochemicaland genetic data that also show that RNA secondary structure isimportant in neurovirulence. J. Mol. Biol. 207:379-392.

34. Standart, N., M. Dale, E. Stewart, and T. Hunt. 1990. Maternalmessenger RNA from clam oocytes can be specifically unmaskedin vitro by antisense RNA complementary to the 3' untranslatedregion. Genes Dev. 4:2157-2168.

35. Stone, D. M., J. W. Almond, J. K. Brangwyn, and G. J. Belsham.1993. trans complementation of cap-independent translation di-rected by poliovirus 5' noncoding region deletion mutants: evi-dence for RNA-RNA interactions. J. Virol. 67:6215-6223.

VOL. 68, 1994


.org/D

ownloaded from

http://jvi.asm.org/

trans complementation by rna of defective foot-and-mouth

Documents