Translocation of a tRNA with an Extended Anticodon Through the Ribosome

doi:10.1016/j.jmb.2006.05.016 J. Mol. Biol. (2006) 360, 610–622

Translocation of a tRNA with an Extended AnticodonThrough the Ribosome

Steven S. Phelps1†, Cyril Gaudin2†, Satoko Yoshizawa2

Cesar Benitez1, Dominique Fourmy2 and Simpson Joseph1⁎
1Department of Chemistry andBiochemistry, University ofCalifornia, San Diego9500 Gilman Dr., La JollaCA 92093-0314, USA2Laboratoire de RMN a HautChamp, ICSN CNRS1 ave de la terrasse91190 Gif-sur-Yvette, France
† S.S.P. and C.G. contributed equaPresent address: S. Joseph, 4104 U

of Chemistry and Biochemistry, UniSan Diego, 9500 Gilman Dr., La Jolla,Abbreviations used: ASL, anticod

elongation factor G; NOE, nuclear OCOSY, correlated spectroscopy; TOCdouble quantum filtered.E-mail address of the correspondi

[email protected]

0022-2836/$ - see front matter © 2006 E

Coordinated translocation of the tRNA–mRNA complex by the ribosomeoccurs in a precise, stepwise movement corresponding to a distance of threenucleotides along the mRNA. Frameshift suppressor tRNAs generallycontain an extra nucleotide in the anticodon loop and they subvert thenormal mechanisms used by the ribosome for frame maintenance. Themechanism by which suppressor tRNAs traverse the ribosome duringtranslocation is poorly understood. Here, we demonstrate translocation of atRNA by four nucleotides from the A site to the P site, and from the P site tothe E site. We show that translocation of a punctuated mRNA is possiblewith an extra, unpaired nucleotide between codons. Interestingly, the NMRstructure of the four nucleotide anticodon stem–loop reveals a conformationdifferent from the canonical tRNA structure. Flexibility within the loop mayallow conformational adjustment upon A site binding and for interactingwith the four nucleotide codon in order to shift the mRNA reading frame.

© 2006 Elsevier Ltd. All rights reserved.

Keywords: ribosome; tRNA; translocation; extended anticodon stem–loop;suppressor frameshift
*Corresponding author
Introduction

Ribosomes are an extremely complex arrange-ment of RNA and proteins, which interact to create amacromolecular machine that is capable of translat-ing the genetic code into proteins with high fidelity.In Escherichia coli, the larger 50 S ribosomal subunitand the smaller 30 S subunit associate to create thepeptidyl (P) site, aminoacyl (A) site and exit (E) site.A purine-rich sequence within the mRNA, known asthe Shine–Dalgarno sequence, guides the initiationcodon to the P site of the ribosome throughinteraction with its complementary sequence in the16 S rRNA of the 30 S subunit. Initiator fMet-tRNA-fMet binds to the P site, followed by binding of anaminoacyl-tRNA to its matching codon at the A site.The 50 S subunit catalyzes peptide bond formation,

lly to this work.rey Hall, Departmentversity of California,CA 92093-0314, USA.on stem–loop; EF-G,verhauser effect;SY, total COSY; DQF,

ng author:

lsevier Ltd. All rights reserve

resulting in a deacylated tRNA in the P site and apeptidyl-tRNA in the A site. After formation of thepeptide bond, movement of the acceptor ends ofboth deacylated and peptidyl tRNAs occurs withrespect to the 50 S subunit. This movement placesthe deacylated tRNA in the P/E state and thepeptidyl tRNA in the A/P state in preparation fortranslocation.1 Elongation factor G (EF-G) thenbinds to the ribosome and hydrolyzes GTP totrigger translocation of the tRNA–mRNA complex.2Translocation by three nucleotides, a feature that isuniversal to all organisms, results in placement ofthe anticodon ends of both the deacylated tRNA intothe E site and the peptidyl tRNA into the P site forthe next round of amino acid incorporation.The ribosome is capable of maintaining high-

efficiency triplet decoding. Under normal circum-stances, frameshift errors are rare, occurring at a rateof less than 10−5 per codon.3 On occasion, aninsertion mutation will lead to a +1 shift in thereading frame. In this case, the ribosome willcontinue past the insertion in the standard threenucleotide reading frame, resulting in non-sensetranslation of a protein that is usually truncated anddegraded prematurely.4Mutant tRNAs capable of suppressing this type of

frameshift were originally characterized in bacteria.5These suppressor tRNAs have an extra nucleotide

d.

mailto:[email protected]

http://dx.doi.org/10.1016/j.jmb.2006.05.016

611Translocation of a tRNA with an Extended Anticodon

inserted into the anticodon, which allows forinteraction with the extra nucleotide present in thecodon. These findings supported the yardstickmodel of translocation,6 in which the tRNA isresponsible for determining the distance traversedby the ribosome during translocation. Surprisingly,some mutant tRNAs having an extra nucleotideinserted between positions 33 and 34 (5′ side of theanticodon; position 33.5) are capable of suppressionwithout formation of a base-pair at the fourthposition, although base-pairing does enhancesuppression.7 These results suggested that suppres-sor tRNAs undergo quadruplet translocation to theP site, where they are capable of sterically interferingwith binding of the in-frame tRNA to the A site.8Mutant tRNAs having an extra nucleotide insertedbetween positions 36 and 37 (3′ side of theanticodon) have been found, but were shown tonot be involved in quadruplet translocation.9 Analternative model was proposed, in which translo-cation to the P site occurs by the standard threenucleotides, followed by a shift into the +1 frame.Protein engineering experiments have revealed

that chemically misacylated suppressor tRNAs arean extremely effective tool for the incorporation ofnon-natural amino acids into proteins.10 An under-standing of the structure of the extended anticodonstem–loop (ASL) and the functional role of sup-pressor tRNAs in frameshifting is essential forour understanding of in-vivo suppression, andfor the optimization of in vitro translation for pro-tein engineering experiments. However, little hasbeen done to determine the structure of a fournucleotide ASL, or to evaluate its interactions withthe mRNA or other tRNAs on the ribosome duringtranslocation.11Here, we utilize an in vitro assay to demonstrate

translocation of a tRNA with a four nucleotideanticodon from the A site to the P site, and from theP site to the E site on the ribosome. Our results showthat the suppressor tRNA is capable of forming apotential fourth base-pair with its codon in the A siteand translocation to the P site by four nucleotides.Accommodation of the four nucleotide tRNA in theP site and subsequent translocation to the E siteeffectively prevents translocation of the following A-site-bound tRNA in the zero reading frame. Addi-tionally, we demonstrate translocation of tRNAswhere an extra unpaired nucleotide exists betweenthe P site and the A site codons. Finally, usingheteronuclear nuclear magnetic resonance (NMR)spectroscopy we investigate the effect of an addi-tional nucleotide in the anticodon loop structure.

Results

Translocation of tRNACCG and tRNACCCG fromthe A site to the P site

Translocation of tRNAs by the ribosome has beenmonitored under a variety of conditions using the

toeprinting method.12,13 The toeprint assay is avaluable tool that can be used to study themechanism of frameshift suppression. Frameshiftsuppression has been shown to occur in well-characterized mutants of Salmonella typhimurium,5Saccharomyces cerevisiae,7 and E. coli.14 In yeast, oneof the tRNAGly isoacceptors was identified ashaving an extra cytosine in the anticodon. Thisinsertion creates the four nucleotide extendedanticodon GCCC (note that both codon and antico-don sequences are displayed in the 5′ to 3′direction), and this tRNA was shown to be anefficient frameshift suppressor. More recently,tRNAs with quadruplet anticodons have beendeveloped to incorporate non-natural amino acidanalogs into proteins.10 For efficient incorporation ofthe amino acid analog, it is imperative that there areno endogenous tRNAs available to interfere with Asite binding to the ribosome. It was determined thatthe CGGG codon will eliminate competition withendogenous tRNAs and generate the most efficientincorporation.10 The corresponding tRNACCCG hasbeen used extensively for incorporation of non-natural amino acids,10 yet our understanding of thestructure of the extended ASL and the mechanismby which it is translocated through the ribosome ispoor. Therefore, tRNAs were transcribed with atRNAPhe body and the anticodon sequence CCCGor CCG (as a control) to analyze translocation of atRNA having a four nucleotide anticodon (Figure1(a) and (b)). The toeprint assay was used to monitormRNA translocation directly, and the methodindirectly reveals the extent of tRNA binding andtranslocation. More importantly, the toeprint assayalso allows the investigation of the extent of theinteraction between the first position of the extendedanticodon (position 33.5) and the fourth position ofthe extended codon at the P site.Initially, tRNAfMet, tRNACCG, or tRNACCCG were

bound directly to the P site to establish a toeprintposition for each of the tRNAs (Figure 2(a), lanes 1–3). When tRNACCCG is bound to the P site, thetoeprint is shifted by one nucleotide towards the 3′end of the mRNA compared to the toeprint of thecontrol tRNACCG. This indicates that tRNACCCG isable to position the fourth nucleotide of the codon inthe P site, resulting in a +1 shift in themRNA readingframe. For translocation experiments, tRNAfMet wasbound to the P site, followed by the addition of eithertRNACCG, or tRNACCCG to the A site (Figure 2(a),lanes 4 and 6, respectively). With tRNACCG bound tothe A site, EF-G•GTP was added and a bandcorresponding to translocation by three nucleotidesappeared (Figure 2(a), lane 5) (for quantification, seeSupplementary Data Figure 1). In the reaction wheretRNACCCGwas bound to theA site, when EF-G•GTPwas added, a band corresponding to a four nuc-leotide shift appeared (Figure 2(a), lane 7).As a control measure to verify the necessity of the

fourth base-pair in the codon–anticodon interactionfor quadruplet translocation, the CGGG codon wasreplaced with a CGGC codon, creating a C-Cmismatch at the fourth position (Figure 1(c)). With

Figure 1. Arepresentation of thedifferent ribosomal complexes. (a)Ribosomes programmed withmRNACGGG, tRNAfMet in the Psite, and tRNACCG in the A site. (b)Ribosomes programmed withmRNACGGG, tRNAfMet in theP site, and tRNACCCG in the A site.(c) Ribosomes programmed withmRNACGGC, tRNAfMet in the Psite, and tRNACCCG in the A site.(d) Ribosomes programmed withmRNACGGG, tRNACCG in the P site,and tRNAPhe in the A site. (e)Ribosomes programmed withmRNACGGG, tRNACCCG in the Psite, and tRNAPhe in the A site. (f)Ribosomes programmed withmRNACGGG, tRNACCG in the Psite, and tRNAVal in the A site. (g)Ribosomes programmed withmRNACGGG, tRNACCCG in the Psite, and tRNAVal in the A site.Position 4 of the extended codon ishighlighted in grey and the possibleinteraction with position 33.5 in theanticodon is indicated by the ovalshape.

612 Translocation of a tRNA with an Extended Anticodon

tRNAfMet pre-bound to the P site, tRNACCCG wasbound to the CGGC codon in the A site, (Figure 2(b),lane 5). Addition of EF-G•GTP results in inefficienttranslocation by primarily three nucleotides (Figure2(b), lane 6) indicating that the fourth base-pair isimportant for efficient quadruplet translocation.Interestingly, addition of tRNACCCG directly to theCGGC codon in the P site (Figure 2(b), lane 4) createsa toeprint that corresponds to the tRNACCCGbinding by four nucleotides. It is possible that,although a C-C mismatch has been introduced, theextended anticodon structure is capable of position-ing the P site of the ribosome over the fourthposition by steric interference creating the +1 shift inthe toeprint. However, additional toeprint bandswere observed, indicating that the mRNA is notproperly positioned within some of the ribosomalcomplexes, possibly due to the base-pair mismatch.

Translocation of tRNACCG and tRNACCCG fromthe P site to the E site

To evaluate translocation of tRNACCG andtRNACCCG from the P site, ribosomal complexeswere created with either tRNA pre-bound to the Psite, followed by the addition of E. coli N-acetylatedPhe-tRNAPhe to the A site (Figure 1(d) and (e)). WithtRNACCG bound to the P site, N-acetylated Phe-tRNAPhe was subsequently bound to the A site(Figure 2(c), lane 1). After the addition of EF-G•GTP,an enhancement in the post-translocation band isapparent (Figure 2(c), lane 2). This result is intrigu-

ing, considering that an extra, unpaired nucleotidemay exist between the two codons. With tRNACCCGbound to the P site, N-acetylated Phe-tRNAPhe wassubsequently added to the A site (Figure 2(c), lane3). Translocation by three nucleotides was inducedby the addition of EF-G•GTP (Figure 2(c), lane 4),which is consistent with the expected mechanismwhereby suppressor tRNAs restore the readingframe.A nitrocellulose filter-binding assay was used to

evaluate binding of tRNAPhe to the A site in the +1frame with tRNACCG or tRNACCCG bound to the Psite (Figure 3(a), rows A and B). The results showthat there is little difference in binding of tRNAPhe tothe A site in the +1 frame when either tRNACCG ortRNACCCG is bound to the P site. Consequently, it islikely that the extra nucleotide is accommodatedbetween the P site and the A site so as to present theUUU codon in the A site for binding of tRNAPhe.

P site binding of tRNACCCG preventstranslocation of tRNAVal in the zero readingframe

Binding of tRNACCCG to the P site causes tRNAPhe

to bind to the A site in the +1 frame, completing theframeshift. However, in vivo suppression efficiencyis not always 100%, indicating that the tRNA in thenormal reading frame is able to compete with thetRNA in the +1 frame for binding to the A site. Toexamine this situation, E. coli N-acetyl Val-tRNAVal

was added to the A site after tRNACCG or tRNACCCG

Figure 2. Toeprint assay of translocation by four nucleotides. (a) A site binding and translocation of tRNACCG andtRNACCCG to the P site. Lanes 1–3 demonstrate P site binding of the indicated tRNA. Lanes 4–7 show tRNAfMet bound tothe P site and tRNACCG or tRNACCCG added to the A site (as in Figure 1(a) and (b)). (+) and (−) indicate the presence andthe absence of EF-G, respectively. Arrows indicate toeprint bands that correspond to positioning of the specified tRNA inthe P site. (b) Binding and translocation of tRNACCCG with an mRNA having a C-Cmismatch at the fourth position. Lanes1–3 are repeated from (a) as a control. Lane 4 demonstrates P site binding of tRNACCCG. Lanes 5 and 6 show tRNAfMet

bound to the P site and tRNACCCG bound to the A site (as in Figure 1(c)). (+) and (−) indicate the presence and the absenceof EF-G, respectively. Arrows indicate toeprint bands that correspond to positioning of the tRNACCCG on the CGGCmRNA. (c) P site binding and translocation of tRNACCG and tRNACCCG to the E site. Lanes 1 and 2 show tRNACCG boundto the P site and N-acetyl Phe-tRNAPhe added to the A site (as in Figure 1(d)). (+) and (−) indicate the presence and theabsence of EF-G, respectively. Lanes 3 and 4 show tRNACCCG bound to the P site and N-acetyl Phe-tRNAPhe added to theA site (as in Figure 1(e)). (+) and (−) indicate the presence and the absence of EF-G, respectively. Lanes 5 and 6 showtRNACCG bound to the P site and N-acetyl Val-tRNAVal added to the A site (as in Figure 1(f)). (+) and (−) indicates thepresence and the absence of EF-G, respectively. Lanes 7 and 8 show tRNACCCG bound to the P site and N-acetyl Val-tRNAVal added to the A site (as in Figure 1(g)). (+) and (−) indicate the presence and the absence of EF-G, respectively.Arrows indicate toeprint bands that correspond to positioning of the specified tRNA in the P site. (d) A site binding andtranslocation of ASL5CCG and ASL5CCCG to the P site. Lanes 1–3 demonstrate P site binding of the indicated ASL. Lanes 4–7 show tRNAfMet bound to the P site and ASL5CCG or ASL5CCCG added to the A site. (+) and (−) indicate the presence andthe absence of EF-G, respectively. Arrows indicate toeprint bands that correspond to positioning of the specified ASL inthe P site.


were pre-bound to the P site (Figure 1(f) and (g)).A strong toeprint is created with tRNACCG ortRNACCCG bound to the P site, followed by additionof N-acetyl Val-tRNAVal to the A site (Figure 2(c),lanes 5 and 7). With the addition of EF-G•GTP,translocation occurs in the reaction where tRNACCGis bound to the P site (Figure 2(c), lane 6). However,no translocation occurswhere tRNACCCG is bound tothe P site (Figure 2(c), lane 8). This result indicatesthat tRNACCCG prevents translocation of tRNAVal inthe zero reading frame.Binding of tRNACCG to the P site andN-acetyl Val-

tRNAVal to the A site appears to trigger a shift in thetoeprint of the pre-translocation complex that is notobserved when N-acetyl Phe-tRNAPhe is bound tothe A site (Figure 2(c), compare lanes 1 and 5). This

phenomenon is likely due to conformational changesthat occur in the ribosome during the formation of apre-translocation state, and the post-translocationtoeprint represents a cross section of a population ofribosomes, only a small majority of which translo-cate to a position that correlates with the valinecodon.A simple explanation for the lack of translocation

observed when tRNACCCG occupies the P site isthat the extended ASL is impinging upon the A siteand preventing N-acetyl Val-tRNAVal from bindingin the zero reading frame. To verify this, A sitebinding of tRNAVal was evaluated with tRNACCGor tRNACCCG bound to the P site (Figure 3(a), rowsC and D). Overall, the binding of tRNAVal to the Asite is reduced compared to the binding of tRNAPhe

Figure 3. A site binding of tRNAPhe and tRNAVal withtRNACCG or tRNACCCG bound to the P site. (a) Thenegatively charged filter used for retention of ribosomecomplexes is shown at the top and the positively chargedfilter used for retention of RNAs is shown at the bottom.RowA, tRNACCG bound to the P site with tRNAPhe boundto the A site. Row B, tRNACCCG bound to the P site withtRNAPhe bound to the A site. Row C, tRNACCG bound tothe P site with tRNAVal bound to the A site. Row D,tRNACCCG bound to the P site with tRNAVal bound to theA site. Column 1, control reactions containing allcomponents, except ribosomes. Columns 2–4, replicatesfor each reaction. (b) Extent of A site binding of tRNAPhe

and tRNAVal with tRNACCG or tRNACCCG bound to the Psite. Fraction bound refers to the fraction of ribosomeshaving the specified tRNA bound to the A site.


to the A site, and the fraction bound correlates withthe extent of translocation. An unexpected result isthat binding of tRNAVal to the A site with tRNACCCGbound to the P site remains unchanged, whencompared to binding with tRNACCG bound to the Psite (Figure 3(b)). Apparently, binding of tRNAVal to

the A site in the zero reading frame is permitted,even though tRNACCCG is bound to the P site. SincetRNAVal will bind to the A site, but is preventedfrom translocating to the P site, it is possible thatthe ribosome plays a role in frameshifting bymonitoring contacts in the P and the A sites andinhibits translocation of specific tRNA–mRNAcomplexes.

Translocation of ASLCCG and ASLCCCG from theA site to the P site

ASL analogs of the corresponding tRNACCG andtRNACCCG were synthesized and assayed for bothP site binding and translocation from the A site.For translocation experiments, tRNAfMet wasbound to the P site. Addition of ASL5CCG to theA site, followed by EF-G•GTP, resulted in translo-cation by three nucleotides (Figure 2(d), lanes 4and 5). When ASL5CCCG is bound to the A site,addition of EF-G•GTP results in a shift in thetoeprint by four nucleotides (Figure 2(d), lanes 6and 7). The appearance of two bands for both ASLsin both pre and post-translocation lanes likelyreflects the disordered nature of both ASLCCG andASLCCCG.It is interesting to note that tRNACCG and

tRNACCCG create strong P site toeprints; conversely,the ASLs create a very weak toeprint, indicatingpoor P site occupancy (Figure 2(d), lanes 2 and 3).Translocation of ASL5CCG and ASL5CCCG from theA site to the P site remains relatively efficient. Thisresult suggests that direct P site binding of theseASLs is prevented either by a more stringent P siteor a reduced level of mRNA binding to theribosome. In fact, it is likely that both of thesesituations affect binding efficiency. The less restric-tive A site would allow for easier entry by thedisordered ASL, and formation of three G-C codon-anticodon base-pairs in the A site might restrict theASL conformation to one that is acceptable for entryinto the P site only during translocation. ASL5CCGand ASL5CCCG show differing extents of quadruplettranslocation, possibly reflecting a transition be-tween different conformations during the shift fromthe A site to the P site.

Sensitivity of chemical shifts to pH and saltconcentration suggests conformationalflexibility in the loop of ASLCCCG

We next characterized the solution structure ofASLCCCG, using a 22 nt RNA (Figure 4) prepared byin vitro transcription from an oligonucleotide tem-plate. NMR samples of unlabeled, (13C, 15N)ade-nine-labeled, and (13C, 15N)-labeled RNAs wereprepared. Complete 1H assignments were obtainedusing standard homonuclear and heteronuclear(1H, 15N, 13C, 31P) multidimensional NMR experi-ments. Assignments were obtained at pH 6.4 in10 mM sodium phosphate and in experimental con-ditions similar to the toeprinting assay at pH 7.5 inthe absence or in the presence of either 50 mM NaCl

Figure 4. Sequence and secondary structure of theASLCCCG RNA. tRNA nucleotides are shown in bold,while those added for transcription efficiency are outlined.The extra cytosine in the anticodon is labeled 33.5.


or 50 mM NaCl and 3 mM MgCl2 (see Supplemen-tary Data Tables 1, 2, 3, 4). In all different conditions,the aromatic-H1′ connectivity walk demonstrateshelical stem formation. The resonances from loopnucleotides were found to be very sensitive to pHand addition of NaCl or MgCl2 (see below).Difficulties in assigning proton resonances from

adenines were resolved by preparing a sample withadenine specifically (13C, 15N)-labeled. Three of thefour adenine residues (A29, A31, and A37) werereadily assigned (see Supplementary Data Figure 2).However, the A38 C2 resonance was missing fromthe canonical region of adenine C2 resonances (150–155 ppm). A very broad resonance was detectedsignificantly upfield (145.8 ppm), which is charac-teristic of an adenine residue with a protonatedN1.15 This result would suggest the existence of aA38H+-C32 base-pair, in the loop region of theASLCCCG at low pH. (For a comparison of two-dimensional 1H,13C heteronuclear single quantumcoherence spectra of the ASLCCCG RNA collectedat different pH values, see Supplementary DataFigure 2). Upon reducing the pH to 5.5, the A38 C2resonance sharpens as it becomes completely pro-tonated. At pH 6.4, A38 was found to be inintermediate exchange between a protonated andunprotonated state. At pH 7.5, the A37 C2 and A38C2 and C8 resonances are no longer detected,suggesting conformational exchange on the NMRexperimental time-scale (∼103 s−1). We then inves-tigated the effect of the presence of 50 mMNaCl and3 mM MgCl2 on the proton resonances of the RNAloop in order to probe the RNA structure inconditions closer to those of the toeprinting assay.At pH 7.5, the chemical shifts and nuclear Over-hauser effect (NOE) patterns observed for the RNAwere affected significantly by addition of 50 mMNaCl, and the changes are even larger with furtheraddition of 3 mM MgCl2 (Figure 5). These resultssuggest that the extended ASL has structuralflexibility.

Structure of the ASL

The structure was determined in two differentexperimental conditions, in buffer conditions (pH7.5, with 50 mM NaCl and 3 mM Mg2+) similar tothose of the toeprinting assay and in 10 mMphosphate buffer at pH 6.4. Structures were calcu-lated using restrained molecular dynamics in asimulated annealing protocol. At pH 7.5, the overallstructure is well defined by the NMR data (Table 1),with a heavy-atom root-mean-square deviation(rmsd) of 1.79 Å for the 20 structures (Figure 6(a)and (b)). Like in tRNA, the anticodon loop is closedby a U39-A31 base-pair. However, the structure ofthe ASLCCCG, as it exists free in solution (Figure7(a)), lacks the U-turn motif and is likely to bedifferent from the canonical anticodon loop struc-ture, which can be observed interacting with thecodon in the crystal structure (Figure 7(b)). The ASLlacks the canonical NOEs found in the U-turnconformation. Instead of having a sharp change inthe direction of the RNA backbone between nucleo-tides 33 and 34 of tRNA, the stacking of nucleotideson the 5′ strand is extended in the ASLCCCG andinvolves the anticodon base C33.5. The turn occursat nucleotides C34 and C35 which have a C2′-endopucker in the ribose moiety. The anticodon basesC34, C35, and G36 are disordered, with G36 existingin a highly disordered state (i.e. a very low numberof internucleotide NOEs from G36). Similar totRNA, nucleotides A37 to U39 are continuouslystacked.At pH 6.4, ASLCCCG folds as a more compact

structure (data not shown), stabilized by theformation of A38H+-C32 and U33-A37 base-pairs.Internucleotide NOEs between non-exchangeableprotons indicate the existence of a Watson–CrickU33-A37 base-pair. However, the imino proton ofU33 is in fast exchange with the solvent and couldnot be observed even at low temperature, indicat-ing a very short life-time for this base-pair. Nohydrogen bonding restraint for U33-A37 andA38H+-C32 base-pairs was included in the struc-ture calculations. With the A38H+-C32 and U33-A37 base-pairs closing the 8 nt ASLCCCG loop, thefour remaining nucleotides C33.5 to G36 form aknown CCCG tetraloop. The structure is similar tothat of a CCCG tetraloop,16 and belongs to theUNCG tetraloop family.17 At pH 6.4, the conservedguanine is in the syn conformation. However, atpH 7.5 this guanine is in the anti conformation(Figure 7(a)).

Discussion

Translocation by four nucleotides

Translocation by three nucleotides is a hallmark oftranslation of the genetic code. Recent modelsdevised to explain a known +1 frameshifting eventdemonstrate that the standard three nucleotide

Figure 5. NOESY spectrum of the ASLCCCG RNA. The experiments were recorded at 30 °C with a mixing time of400 ms. (a) pH 6.4, (b) pH 7.5, (c) pH 7.5 with 50 mM NaCl, (d) pH 7.5 with 50 mM NaCl and 3 mM MgCl2. Base-H1′sequential connectivities from A29 to C35 and G36 to A38 are shown. A break in the sequential connectivities is observedbetween C35 and G36.


translocation can be subverted through a variety ofmechanisms. These mechanisms are characterizedby a 2 nt or a 3 nt codon–anticodon interaction at theA site followed by translocation to the P site, where a+1 frameshift occurs.18 This +1 frameshift is depen-dent upon a ‘‘slippery’’ mRNA sequence, whichtends to promote the frameshift by allowing the Psite tRNA to re-pair with a favorable codon in the +1frame.In addition to sequence-specific effects at the

codon in the P site, factors such as tRNAmodification,19 levels of tRNA charging,20 andinteractions between the mRNA and 16 S rRNA21

have a direct role in frameshifting efficiency.Alternatively, it has been demonstrated previousythat re-pairing in the +1 frame at the P site is notpossible, since the codon existing in the +1 readingframe presents an unfavorable sequence for Wat-son–Crick base-pairing. In this case, it was sug-gested that a tRNAwill bind to the A site in the +1

frame, without slippage of the P site-bound pepti-dyl-tRNA.22,23The in vitro study employed here is designed so

that re-pairing events are unlikely due to the factthat the tRNAs present do not have an alternativecognate or near-cognate site for re-pairing. Abenefit of the in vitro study is that only specifictRNAs are added to the system, which preventserrors due to background binding of unanticipatedendogenous tRNAs. Therefore, using a systemwhere re-pairing of tRNAs to the mRNA andcompetition between tRNAs for binding to theribosome does not occur, we are able to show thattRNACCCG translocates by four nucleotides fromthe A site to the P site. These results are consistentwith an independent study carried out by Walkerand Fredrick (see the accompanying paper).24Furthermore, the structure of the 30 S subunitcomplexed with ASLCCCG in the A site was solvedrecently by X-ray crystallography (C. Dunham, M.

Table 1. NMR and refinement statistics for ASL RNA

ASL RNA Loop (C32–A38)

A. NMR distance and dihedral constraintsDistance restraints

Total NOE 339 170Intra-residue 147 62Inter-residue 154 108Sequential (∣i–j∣=1) 140 76Non-sequential (∣i–j∣>1?) 14 32

Hydrogen bonds 38Total dihedral angle restraints 73 24

Base-pair 7Sugar pucker 22 8Backbone 51 16

B. Structure statisticsViolations (mean and s.d.) ±

Distance constraints (Å) 0.017±0.003Dihedral angle

constraints (deg.)0.58±0.13

Max. dihedral angleviolation (deg.)

6.3

Max. distanceconstraint violation (Å)

0.26

Deviations fromidealized geometryBond lengths (Å) 0.0013±0.00006Bond angles (deg.) 0.56±0.006Impropers (deg.) 0.30±0.003

Average r.m.s.d. to themean structure a

All RNA heavy (Å) 1.79 1.92Disordered loop

nucleotides b (Å)2.01

Stem–loop orderednucleotides c (Å)

0.91

a Among (at least 20) refined structures.b C33.5–A37.c G25–U33, A38–C45.


Selmer and V. Ramakrishnan, personal communi-cation). The structure shows that the expandedanticodon is able to interact with position 4 ofthe codon in an unconventional manner, poten-tially initiating the shift in the mRNA readingframe in the ribosomal A site. In agreement withthis result, our toeprint analysis shows thatchanging G to C at position 4 of the codon,which disrupts interactions with the expandedanticodon, decreases the level of four nucleotidetranslocation by tRNACCCG (Figure 2(b)).

Translocation with an unpaired nucleotidebetween codons

The analysis of translocation of a four nucleotidecodon from the P site to the E site has yielded anunexpected result. Interestingly, the ribosome willbind the control tRNACCG to the P site and allowbinding of tRNAPhe directly to the +1 reading frame.The result is translocation with an extra, presumablyunpaired, nucleotide between the A site and the Psite codons.An alternative interpretation of the toeprint data

presented here would suggest a +1 frameshift of theP site-bound tRNACCG before A site binding of

tRNAPhe and translocation. This frameshift mightoccur through re-pairing of anticodon nucleotidesC34 and C35 to codon nucleotides G6 and G7, assuggested by previously presented frameshiftingmechanisms.18 This mechanism requires a reducedlevel of codon–anticodon stability to induce slip-page of the P site-bound tRNA into the +1 frame.The P site-bound tRNA used here, tRNACCG, has astrong interaction with its codon and is unlikely toslip into the +1 frame. Additionally, the formation ofa doublet in the pre-translocation band of thetoeprint (Figure 2(c), lane 1) may suggest +1slippage before translocation. However, this doubletis observed also in the P site toeprint of tRNAfMet.Slippage by +1 of tRNAfMet on the initiation codon ishighly unlikely. Therefore, the doublet observed ispresumably the result of a conformational changethat occurs upon initiation.25Considering these results, it may be possible to

create an in vivo +1 frameshift through the compe-tition of endogenous tRNAs (having a normal 7 ntloop) for binding to the zero or +1 reading frames atthe A site. In fact, translational accuracy has beenlinked directly to reading frame maintenance,26 andnon-slippery frameshifting events occur with nor-mal tRNAs,27 or might require a special peptidyl-tRNA.28Binding and translocation of tRNAPhe in the +1

frame is significant, because it shows that theribosome is capable of allowing an unpairednucleotide between codons, and because it showsthat the ribosome has the ability to translate anmRNA that contains a punctuated message. It haslong been assumed that translation of the geneticcode does not require punctuation.29 However, ourresults show that the ribosome is capable of trans-lating a punctuated message when it is providedunder regulated conditions.The structural implications are intriguing, con-

sidering that universally conserved rRNA nucleo-tides A1492, A1493, and G530 are involveddirectly in decoding at the first and second posi-tions of the A site codon–anticodon interaction.30The extra, presumably unpaired, nucleotide be-tween A site and P site codons needs to be accom-modated. It is possible that the structure of theribosome is capable of accommodating an extranucleotide between codons in the A and the P sitesby flipping the nucleotide out of the codon–anticodon helix.

tRNAs bound to overlapping codons do nottranslocate

Current mechanisms hold that the four nucleotideextended anticodon interacts at the A site andtranslocates to the P site, while maintaining a threenucleotide codon–anticodon interaction. Only aftertranslocation to the P site does a spontaneous isom-erization with the mRNA occur in which the fourthbase-pair is made.9 It is apparent from the resultsthat when tRNACCCG is bound directly to the P site,it is entirely capable of preventing translocation in

Figure6. Structureof theASLCCCG.(a) Best-fit superposition of 20 finalsimulated annealing structures ofthe ASLCCCG RNA. The heavyatoms of the RNA have been super-imposed. Bases are shown in blueand the backbone in light grey. (b)View of a single representativestructure of the ASLCCCG RNA. Allheavy atoms are displayed. Colorsare as in (a), and phosphate atomsare in orange. (c) Stereo view of theASLCCCG anticodon loop. Colorsand numbering are as in (a).


the zero reading frame. Therefore, translocationfrom the A site to the P site is not necessary forformation of the fourth base-pair, leading to the +1frameshift. Other types of suppressor tRNAs, whichpossess varying degrees of efficiency, may requirethe formation of a base-pair in the P site only aftertranslocation.

Figure 7. Comparison of the ASLCCCG and the tRNA anticoof the crystal structure of the tRNAPhe ASL.50 All heavy atom

The result that binding of tRNACCCG to theP site prevents translocation of tRNAVal in thezero reading frame, but allows translocation oftRNAPhe in the +1 reading frame, indicates thatthe additional codon–anticodon base-pair createdby the four nucleotide anticodon of tRNACCCGforms a boundary at the 5′ end of the A site.

don loops. (a) Aview of the ASLCCCG structure. (b) Aviews are displayed. Colors are as in Figure 6(a).


Here, we see that binding of tRNAVal in the zeroreading frame is permitted, possibly demonstrat-ing a transient interaction between the firstnucleotide of the extended anticodon and thefourth nucleotide of its corresponding codon inthe P site.Overlapping binding of tRNACCCG and tRNAVal

to their codons is the product of several factors thatenhance codon sampling by the incoming tRNA atthe A site. These factors include 20 mM Mg2+ and alack of competition between tRNAs in the zero and+1 reading frames. It is likely that steric crowdingoccurs when two tRNAs occupy overlappingcodons in the A and P sites. The ribosome maysense the situation with overlapping tRNAs andprevent translocation. Perhaps the same mechanisminvolved in decoding by rRNA nucleotides G530,A1492 and A1493 is responsible for the preventionof translocation.

The four nucleotide ASL in solution isstructurally different from a tRNA anticodonloop

It has been shown that the 7 nt tRNAASL adopts astructure allowing high decoding efficiency andtranslocation on the ribosome. In particular, theU-turn motif of the ASL is important for P sitebinding,31 as well as A site binding and trans-location17. Here, we found that the ASLCCCG, free insolution, lacks the conserved tRNA U-turn motif.This result is not surprising and a conformationdistinct from the U-turn motif has been observedfor an unmodified tRNAPhe ASL.32 Additionally,chemical shifts in the 8 nt loop were found to beparticularly sensitive to experimental conditionssuch as pH, salt concentration, and temperature.Changes in these conditions result in differentstructures for the ASL at pH 6.4, or pH 7.5, in thepresence of NaCl and MgCl2. In agreement withthis observation, the pyrimidine-rich anticodon is,in the structure, quite disordered and nucleotideG36 is highly disordered. In fact, RNAmodificationshave been shown to stabilize the pyrimidine-richASLLys

UUU by pre-ordering the loop into a structurethat is more suitable for presentation to the codon inthe decoding center.33Phosphorothioate substitutions in the canonical

ASL34 have shown that disruption of the contactbetween the N3 of uridine 33 and the pro-RP non-bridging phosphate oxygen atom at position 36 haslittle effect on A site binding but severely inhibitstranslocation to the P site. Destabilization of the U-turn motif does not affect codon-dependent A sitebinding and this correlates with the finding that alack of a stabilized U-turn motif in the ASLCCCG alsohas little effect on A site binding. However,translocation by four nucleotides with tRNACCCGsuggests that upon A site binding and codon–anticodon recognition, the extended anticodon loopchanges conformation to accommodate the U-turnmotif necessary for efficient translocation. Thus, theribosome can reconfigure the extended tRNA anti-

codon to enhance binding to the A site and translo-cation to the P site.

Materials and Methods

Preparation of ribosomes, mRNAs, tRNAs, and ASLs

Tight couple 70 S ribosomes were isolated from E. coliMRE 600 cells.35 sup2 mRNA was transcribed from alinearized plasmid containing an insert that places thecggg codon directly between the aug initiation codon andthe uuu phenylalanine codon of gene 32mRNA. tRNACCGand tRNACCCG were transcribed from a plasmid encodingtRNAPhe (3′-CA) which was altered to create the correct 3′end, and at the sequence corresponding to the anticodon,where gaa was replaced with ccg or cccg. ASL5CCG andASL5CCCG, both having the same sequence as tRNAPhe

except in the anticodon, were purchased from Dharma-con. E. coli tRNAfMet, tRNAPhe, and tRNAVal werepurchased from Sigma. PheRS and ValRS were preparedaccording to an established procedure.36 tRNAPhe andtRNAVal were aminoacylated, N-acetylated, and purifiedby HPLC.37

Translocation assay

Toeprinting assays were used to monitor translo-cation.12,34 Ribosomes (0.40 μM), tRNACCG andtRNACCCG (or ASL5) (6.0 μM), and charged (2.0 μM) oruncharged (6.0 μM) tRNAPhe and tRNAVal were preparedindividually in 80 mM potassium cacodylate (pH 7.2),20 mM MgCl2, 150 mM NH4Cl, 3 mM 2-mercaptoethanol.A 32P-labeled AL2 primer was annealed to the mRNA(0.8 μM) and added to pre-programmed and activatedribosomes.13 P site binding of the appropriate tRNA orASL occurred for 30 min at 37 °C. A tRNA, ASL, or 1Xbinding buffer (where P site binding is demonstrated),was subsequently added to the mixture and A site bindingoccurred for 30 min at 37 °C. E. coli EF-G (2.0 μM) and GTP(300 μM) were prepared in binding buffer and added tobring the final reaction volume to 25 μl. Followingincubation at 37 °C for 30 min, an aliquot was removedfrom each sample and mixed with AMV-reverse tran-scriptase (Seikagaku). After primer extension, extensionproducts corresponding to the pre and post-translocationcomplexes were separated on a denaturing 10% poly-acrylamide gel and visualized by autoradiography. Alltranslocation data are the result of at least three inde-pendent experiments.

A site binding assay

Saturation binding assays were used to determine thefraction of ribosomes having an occupied A site.34Conditions used for the binding assay are the same asthose for the translocation assay above, except E. colitRNAPhe and E. coli tRNAVal were mixed with a tracequantity of the [32P]pCp end-labeled tRNA. Non-specificbinding controls (i) without ribosome and (ii) withoutmRNA were determined to be less than 5% and 20%,respectively. Nitrocellulose filters were placed within the96-well filtration apparatus, and a positively chargednylon sheet filter was placed under the nitrocellulose toretain the total number of counts for each sample well.Total counts were monitored also by removing 1 μl before


filtration. The radioactive intensity for each filter wasquantified and the fraction of ribosomes with an A site-bound ASL was calculated as follows: (Counts bound tonitrocellulose–background)/ (Total counts expected at100% occupancy). All binding data are the result of atleast three independent experiments.

NMR sample preparation

RNAs were prepared unlabeled and 13C,15N-labeledby in vitro transcription from an oligonucleotide tem-plate as described.38 Labeled NTPs were purchased fromSpectra Gases Inc. RNAs were dialyzed first against10 mM sodium phosphate (pH 6.4) and finally 10 mMsodium phosphate (pH 5.7, 6.4, 7.5, or 8.4). NMRsamples (of 0.5–1 mM) were prepared in ShigemiNMR tubes.

NMR spectroscopy

NMR experiments were recorded at 800 MHz or 600MHz on Bruker DRX spectrometers equipped with tripleresonance, z-gradient cryoprobes. NMR experimentsinvolving 31P were recorded on a Varian Inova spec-trometer at 500 MHz. NMR data were processed andanalyzed using XWINNMR (Bruker), Gifa,39 andSPARKY‡ software packages. 15N and 13C chemicalshifts were referenced with the known chemical shifts ofthe ribose carbon atom or the base nitrogen atom fromthe 5′ and 3′ end regions.40 1H, 13C, 15N, and 31Passignments were obtained using standard homonuclearand heteronuclear methods.41 NMR data were acquiredat 20 °C, 30 °C, and 35 °C, and data for exchangeableprotons were collected at 5 °C, 10 °C, and 15 °C. Two-dimensional NOESY spectra in 95% H2O/5% 2H2O wereacquired with mixing times of 50 ms, 150 ms, and300 ms. Hydrogen bonding patterns of base-pairswere observed directly from heteronuclear J(N,N)-HNN correlated spectroscopy (COSY) experiments in90% H2O/10% 2H2O.42 Heteronuclear NMR spectrawere measured at 30 °C in 99.9% 2H2O. The assignmentof the non-exchangeable protons of the labeled RNAwas completed using constant-time heteronuclearsingle quantum coherence,43 3D-HCCH- total COSY(TOCSY),44 and HCN experiments.45 Sequential connec-tivities between nucleotides were obtained by HP-COSYexperiments,46 and HCP experiments.47 NOEs wereclassified into four distance bound ranges: strong 1.8–3.0 Å, medium 2.0–4.0 Å, weak 3.0–5.0 Å, and very weak3.5–6.0 Å. Very weak NOE restraints observed in theloop or bulge regions were loosened to 3.5–7.0 Å. RNAdihedral restraints were obtained from double quantumfiltered (DQF)-COSY, TOCSY, and HP-COSY experimentsand were assigned following a previously describedstrategy.17The β dihedral angles were restrained from estimates

of the 3JP-H5′, 3JP-H5″ and 3JP-C4′ coupling constants fromthe HP-COSY experiment on the unlabeled RNA. ε wasalso restrained from estimates of 3JH3′-P, 3JC2′-P and 3JC4′-Pfrom the HP-COSY. The ε dihedral angles were con-strained to 210(±30)° (trans) or 260(±20)° (gauche–) or235(±55)° (when the trans or gauche– conformations couldnot be distinguished).The γ dihedral angles were constrained where possible

using estimates of 3JH4′-H5′ and 3JH4′-H5″ coupling constants

‡ http://www.cgl.ucsf.edu/home/sparky/

from the 31P decoupled DQF-COSY. For a gauche+ con-formation, γ was constrained to 55(±20)°. The ribosesugar pucker was estimated from analysis of the H1′-H2′coupling constants in the 31P decoupled DQF-COSYspectrum. Nucleotides with a H1′-H2′ coupling constantof >8 Hz in the COSY spectrum were classified as C2′-endo (δ=150(±30)°). Nucleotides with no COSY andTOCSY crosspeaks between the H1′-H2′ protons (J <3 Hz) were classified as C3′-endo (δ=85(±20)°). Somenucleotides had weak H1′-H2′ crosspeaks in the TOCSYspectrum, but not in the COSY. When a mixedpopulation of C2′/C3′-endo conformations was observed,the ribose puckers for these nucleotides were restrainedin the range of the C2′ + C3′-endo conformations duringmolecular dynamics.

Structure determination

The CNS package was used to calculate structures usingNOE distance and dihedral restraints.48 The structurecalculations closely followed the default values for NMRstructure determination of nucleic acids with CNS 1.1.First, an extended unfolded structure was generated, fromwhich 100 structures were calculated from random initialvelocities. Structures were subjected to 60 ps (15 fstimesteps) of restrained molecular dynamics in torsionangle space, followed by 90 ps of slow cooling. Finally, 30ps (5 fs timesteps) of restrained molecular dynamics inCartesian coordinate space were performed. For Watson–Crick base-pairs that were detected by trans-hydrogenbond experiments, only very weak planarity restraints (10kcal mol−1 Å−2) were enforced during calculations andhydrogen bonds were maintained by distance restraints.No intrasugar distance restraint was used for structurecalculation, except 1′–3′ and 1′–4′ distances. Structuresshown in Figure 6 had the lowest total energy andrestraint violation energies. Structures were analyzedusing MolMol,49 and color Figures were generated withInsightII (Biosym Technologies, San Diego, CA).

Protein Data Bank accession code

Atomic coordinates have been deposited in the ProteinData bank under accession code 2GV3.

Acknowledgements

We thank M. Sisido for providing a plasmidencoding tRNACCCG (-CA), T. Ueda for providingcells that over-express a His-tagged form of PheRSand ValRS, J.L. Leroy for helpful discussions andassistance with some NMR experiments, V. Ramak-rishnan, K. Fredrick, C. Dunham and M. Selmer forcomments on the manuscript. This work wassupported by a grant to D.F. and S.J. from theHuman Frontier Science Program.

Supplementary Data

Supplementary data associated with this articlecan be found, in the online version, at doi:10.1016/j.jmb.2006.05.016

http://www.cgl.ucsf.edu/home/sparky


References

1. Moazed, D. & Noller, H. F. (1989). Intermediate statesin the movement of transfer RNA in the ribosome.Nature, 342, 142–148.

2. Savelsbergh, A., Katunin, V. I., Mohr, D., Peske, F.,Rodnina, M. V. & Wintermeyer, W. (2003). Anelongation factor G-induced ribosome rearrangementprecedes tRNA-mRNA translocation. Mol. Cell, 11,1517–1523.

3. Atkins, J. F., Elseviers, D. & Gorini, L. (1972). Lowactivity of -galactosidase in frameshift mutants ofEscherichia coli. Proc. Natl Acad. Sci. USA, 69,1192–1195.

4. Kurland, C. G. (1992). Translational accuracy and thefitness of bacteria. Annu. Rev. Genet. 26, 29–50.

5. Riddle, D. L. & Carbon, J. (1973). Frameshiftsuppression: a nucleotide addition in the anticodonof a glycine transfer RNA. Nature New Biol. 242,230–234.

6. Holley, R. W. (1965). Structure of an alanine transferribonucleic acid. JAMA, 194, 868–871.

7. Gaber, R. F. & Culbertson, M. R. (1984). Codonrecognition during frameshift suppression in Saccha-romyces cerevisiae. Mol. Cell Biol. 4, 2052–2061.

8. Bossi, L. & Smith, D. M. (1984). Suppressor sufJ: anovel type of tRNA mutant that induces transla-tional frameshifting. Proc. Natl Acad. Sci. USA, 81,6105–6109.

9. Qian, Q., Li, J. N., Zhao, H., Hagervall, T. G.,Farabaugh, P. J. & Bjork, G. R. (1998). A new modelfor phenotypic suppression of frameshift mutationsby mutant tRNAs. Mol. Cell, 1, 471–482.

10. Ninomiya, K., Minohata, T., Nishimura, M. & Sisido,M. (2004). In situ chemical aminoacylation with aminoacid thioesters linked to a peptide nucleic acid. J. Am.Chem. Soc. 126, 15984–15989.

11. Smith, D. & Yarus, M. (1989). tRNA-tRNA interactionswithin cellular ribosomes. Proc. Natl Acad. Sci. USA,86, 4397–4401.

12. Joseph, S. & Noller, H. F. (1998). EF-G-catalyzedtranslocation of anticodon stem-loop analogs oftransfer RNA in the ribosome. EMBO J. 17, 3478–3483.

13. Phelps, S. S., Jerinic, O. & Joseph, S. (2002). Univer-sally conserved interactions between the ribosomeand the anticodon stem-loop of A site tRNA impor-tant for translocation. Mol. Cell, 10, 799–807.

14. Atkins, J. F., Weiss, R. B., Thompson, S. & Gesteland,R. F. (1991). Towards a genetic dissection of the basisof triplet decoding, and its natural subversion:programmed reading frame shifts and hops. Annu.Rev. Genet. 25, 201–228.

15. Legault, P. & Pardi, A. (1994). In situ probing ofadenine protonation in RNA by 13C-NMR. J. Am.Chem. Soc. 116, 8390–8391.

16. Finger, L. D., Trantirek, L., Johansson, C. & Feigon, J.(2003). Solution structures of stem-loop RNAs thatbind to the two N-terminal RNA-binding domains ofnucleolin. Nucl. Acids Res. 31, 6461–6472.

17. Allain, F. H. & Varani, G. (1995). Structure of the P1helix from group I self-splicing introns. J. Mol. Biol.250, 333–353.

18. Hansen, T. M., Baranov, P. V., Ivanov, I. P., Gesteland,R. F. & Atkins, J. F. (2003). Maintenance of the correctopen reading frame by the ribosome. EMBO Rep. 4,499–504.

19. Urbonavicius, J., Qian, Q., Durand, J. M., Hagervall,T. G. & Bjork, G. R. (2001). Improvement of reading

frame maintenance is a common function for severaltRNA modifications. EMBO J. 20, 4863–4873.

20. Gao, W., Jakubowski, H. & Goldman, E. (1995).Evidence that uncharged tRNA can inhibit aprogrammed translational frameshift in Escherichiacoli. J. Mol. Biol. 251, 210–216.

21. Weiss, R. B., Dunn, D. M., Dahlberg, A. E., Atkins, J. F.& Gesteland, R. F. (1988). Reading frame switchcaused by base-pair formation between the 3′ end of16S rRNA and the mRNA during elongation ofprotein synthesis in Escherichia coli. EMBO J. 7,1503–1507.

22. Farabaugh, P. J., Zhao, H. & Vimaladithan, A. (1993).A novel programed frameshift expresses the POL3gene of retrotransposon Ty3 of yeast: frameshiftingwithout tRNA slippage. Cell, 74, 93–103.

23. Pande, S., Vimaladithan, A., Zhao, H. & Farabaugh,P. J. (1995). Pulling the ribosome out of frame by +1 ata programmed frameshift site by cognate binding ofaminoacyl-tRNA. Mol. Cell Biol. 15, 298–304.

24. Walker, S. E. & Fredrick, K. (2006). Recognition andpositioning of mRNA in the ribosome by tRNAs withexpanded anticodons. J. Mol. Biol. accompanying article.

25. Jerinic, O. & Joseph, S. (2000). Conformationalchanges in the ribosome induced by translationalmiscoding agents. J. Mol. Biol. 304, 707–713.

26. Farabaugh, P. J. & Bjork, G. R. (1999). How transla-tional accuracy influences reading frame mainte-nance. EMBO J. 18, 1427–1434.

27. Atkins, J. F., Gesteland, R. F., Reid, B. R. & Anderson,C. W. (1979). Normal tRNAs promote ribosomalframeshifting. Cell, 18, 1119–1131.

28. Vimaladithan, A. & Farabaugh, P. J. (1994). Specialpeptidyl-tRNA molecules can promote translationalframeshifting without slippage. Mol. Cell Biol. 14,8107–8116.

29. Crick, F. H., Barnett, L., Brenner, S. &Watts-Tobin, R. J.(1961). General nature of the genetic code for proteins.Nature, 192, 1227–1232.

30. Ogle, J. M., Brodersen, D. E., Clemons, W. M., Jr, Tarry,M. J., Carter, A. P. & Ramakrishnan, V. (2001).Recognition of cognate transfer RNA by the 30 Sribosomal subunit. Science, 292, 897–902.

31. Uhlenbeck, O. C., Lowary, P. T. & Wittenberg, W. L.(1982). Role of the constant uridine in binding of yeasttRNAPhe anticodon arm to 30 S ribosomes. Nucl.Acids Res. 10, 3341–3352.

32. Cabello-Villegas, J., Winkler, M. E. &Nikonowicz, E. P.(2002). Solution conformations of unmodified and A(37)N(6)-dimethylallyl modified anticodon stem-loopsof Escherichia coli tRNA(Phe). J. Mol. Biol. 319,1015–3104.

33. Murphy, F. V. t., Ramakrishnan, V., Malkiewicz, A. &Agris, P. F. (2004). The role of modifications in codondiscrimination by tRNA(Lys)UUU. Nature Struct. Mol.Biol. 11, 1186–1191.

34. Phelps, S. S. & Joseph, S. (2005). Non-bridgingphosphate oxygen atoms within the tRNA anticodonstem-loop are essential for ribosomal A site bindingand translocation. J. Mol. Biol. 349, 288–301.

35. Powers, T. & Noller, H. F. (1990). Dominant lethalmutations in a conserved loop in 16S rRNA. Proc. NatlAcad. Sci. USA, 87, 1042–1046.

36. Shimizu, Y., Inoue, A., Tomari, Y., Suzuki, T.,Yokogawa, T., Nishikawa, K. & Ueda, T. (2001). Cell-free translation reconstituted with purified compo-nents. Nature Biotechnol. 19, 751–755.

37. Odom, O. W., Deng, H. Y. & Hardesty, B. (1988).Fluorescence labeling and isolation of labeled RNA


and ribosomal proteins. Methods Enzymol. 164,174–187.

38. Gaudin, C., Mazauric, M. H., Traikia, M., Guittet, E.,Yoshizawa, S. & Fourmy, D. (2005). Structure of theRNA signal essential for translational frameshifting inHIV-1. J. Mol. Biol. 349, 1024–1035.

39. Pons, J. L., Malliavin, T. & Delsuc, M. A. (1996). Acomplete package for NMR data set processing.J. Biomol. NMR, 8, 445–452.

40. Yoshizawa, S., Fourmy, D. & Puglisi, J. D. (1998).Structural origins of gentamicin antibiotic action.EMBO J. 17, 6437–6448.

41. Furtig, B., Richter, C., Wohnert, J. & Schwalbe, H.(2003). NMR spectroscopy of RNA. ChemBiochem. 4,936–962.

42. Dingley, A. J. & Grzesiek, S. (1998). Direct observationof hydrogen bonds in nucleic acid base-pairs byinternucleotide 2JNN couplings. J. Am. Chem. Soc. 120,8293–8297.

43. Santoro, J. & King, G. (1992). A constant-time 2Doverbodenhausen experiment for inverse correlationof isotopically enriched species. J. Magn. Reson. 97,202–207.

44. Clore, G. M., Bax, A., Driscoll, P. C., Wingfield, P. T. &Gronenborn, A. M. (1990). Assignment of the side-chain 1H and 13C resonances of interleukin-1β usingdouble- and triple-resonance heteronuclear three-dimensional NMR spectroscopy. Biochemistry, 29,8172–8184.

45. Riek, R., Pervushin, K., Fernandez, C., Kainosho, M. &Wüthrich, K. (2001). [13C,13C]- and [13C, 1H]-TROSY ina triple resonance experiment for ribose-base andintrabase correlations in nucleic acids. J. Am. Chem.Soc. 123, 658–664.

46. Sklenar, V., Miyashiro, H., Zon, G., Miles, H. T. & Bax,A. (1986). Assignment of the 31P and 1H resonances inoligonucleotides by two-dimensional NMR spectros-copy. FEBS Letters, 208, 94–98.

47. Marino, J. P., Schwalbe, H., Anklin, C., Bermel, W.,Crothers, D. M. & Griesinger, C. (1995). Sequentialcorrelation of anomeric ribose protons and interven-ing phosphorus in RNA oligonucleotides by a 1H, 13C,31P triple resonance experiment: HCP-CCH-TOCSY.J. Biomol. NMR, 5, 87–92.

48. Brunger, A. T., Adams, P. D., Clore, G. M., DeLano,W. L., Gros, P., Grosse-Kunstleve, R. W. et al. (1998).Crystallography and NMR system: A new softwaresuite for macromolecular structure determination.Acta Crystallog. sect. D, 54, 905–921.

49. Koradi, R., Billeter, M. & Wuthrich, K. (1996).MOLMOL: a program for display and analysis ofmacromolecular structures. J. Mol. Graph. 14,51–55.

50. Sussman, J. L., Holbrook, S. R., Warrant, R. W.,Church, G. M. & Kim, S. H. (1978). Crystalstructure of yeast phenylalanine transfer RNA. I.Crystallographic refinement. J. Mol. Biol. 123,607–630.

Edited by D. E. Draper

(Received 31 March 2006; received in revised form 4 May 2006; accepted 4 May 2006)Available online 22 May 2006

Translocation of a tRNA with an Extended Anticodon Through the Ribosome

Documents

Transcript of Translocation of a tRNA with an Extended Anticodon Through the Ribosome