Cryoelectron microscopy structures of theribosome complex in intermediate statesduring tRNA translocationJie Fua, James B. Munrob, Scott C. Blanchardb, and Joachim Franka,c,d,1

aDepartment of Biochemistry and Molecular Biophysics, Columbia University, New York, NY 10032; bDepartment of Physiology and Biophysics,Weill Cornell Medical College, New York, NY 10065; cHoward Hughes Medical Institute, Department of Biochemistry and Molecular Biophysics,Columbia University, New York, NY 10032; and dDepartment of Biological Sciences, Columbia University, New York, NY 10027

Contributed by Joachim Frank, January 28, 2011 (sent for review December 28, 2010)

mRNA–tRNA translocation is a central and highly regulated processduring translational elongation. Along with the mRNA, tRNAmoves through the ribosome in a stepwise fashion. Using cryoelec-tron microscopy on ribosomes with a P-loop mutation, we haveidentified novel structural intermediates likely to exist transientlyduring translocation. Our observations suggest a mechanism bywhich the rate of translocation can be regulated.

The elongation cycle of protein synthesis is an iterative processthat can be divided into three steps, tRNA incorporation, pep-tidyl transfer, and mRNA–tRNA translocation (1). During tRNAincorporation, an aminoacyl-tRNA (aa-tRNA) is delivered tothe ribosomal A site as part of a ternary complex with elongationfactor Tu (EF-Tu) and GTP. In the peptidyl-transfer reaction, thenascent peptide chain bound to the P-site tRNA is transferred tothe N-terminus of the aminoacyl group of the A-site aa-tRNA,and thus the polypeptide chain is extended by one amino acid.The resulting ribosome complex bearing a peptidyl-tRNA in theA site and a deacylated tRNA in the P site is termed the pretran-slocational (PRE) complex. In the translocation process, themRNA, along with the tRNAs, are moved by one codon with re-spect to the ribosome. The resulting posttranslocation (POST)complex contains a peptidyl-tRNA at the P site and a deacylatedtRNA at the E site. The new downstream mRNA codon is pre-sented at the A site, which enables the incorporation of the nextcognate aa-tRNA. Translocation is catalyzed by elongation factorG (EF-G). However, factor-free, spontaneous translocationoccurs, as well, albeit at much lower rates (2).

To a first approximation translocation of tRNA and mRNAsubstrates takes place in two steps: First, the acceptor arms ofthe A-site and P-site tRNAs move to the P site and E site on the50S subunit, respectively, forming the so-called hybrid A/P andP/E tRNA configurations; second, the anticodon arms of theA/P and P/E tRNA, along with the mRNA, move to the 30S sub-unit’s P and E sites, respectively (3). The hybrid configurations ofthe tRNAs have been directly visualized in three-dimensionalstructures of the ribosome obtained by cryoelectron microscopy(cryo-EM) and single-particle reconstruction (4–6). Coupled withthe formation of the hybrid tRNA configurations, as seen inthese cryo-EM reconstructions, are a number of conformationalchanges of the ribosome, including the so-called ratchet-like mo-tion (a counterclockwise rotation of the 30S subunit with respectto the 50S subunit), and the movement of the L1 stalk toward themain body of the ribosome where it makes contact with the elbowof the hybrid P/E tRNA. These conformational changes of theribosome are thought to play a direct role in the translocationmechanism, by helping break and reestablish substrate-ribosomeinteractions during the movement (5, 7). Recently, the structureof the PRE ribosome, with tRNAs replaced by anticodon stemloops, was solved by crystallography in two intermediate confor-mations (8).

Increasingly, the technique of single-molecule fluorescenceresonance energy transfer (sm-FRET) is being used to obtain

real-time dynamic information about the translocation process(9). These studies have established that ribosome and tRNAsubstrates fluctuate between two or more structurally and kine-tically distinct states (10–14). In particular, one sm-FRET studyprovided evidence that the tRNA duplex in the PRE complexoscillates between three states: (i) the classical state (A/A andP/P), (ii) the hybrid state (A/P and P/E), and (iii) a previouslyunidentified partially hybrid state (A/A and P/E), in which onlyone tRNA has moved (11). Occupancy of this additional statewas enriched by introducing a point mutation, G2252C, on theP-loop of the peptidyl-transferase center on the 23S rRNA. Thismutation has been shown to promote the formation of the P/Ehybrid tRNA, and at the same time prevent the formation ofthe A/P hybrid tRNA configuration (15).

Here, we used cryo-EM and single-particle reconstruction tostudy this G2252C PRE complex. We set out to visualize inter-mediate states of the tRNAs in an attempt to understand howthe global conformation of the ribosome changes as the tRNAsswitch between those states. As a result, we were able to identifytwo intermediate states in which the tRNAs are in unique posi-tions, with the P/E tRNA apparently primed to latch on to theL1 stalk. We also show that the observed positions of the L1 stalkcan be characterized as the result of different directions of move-ment, each of which is associated with a different tRNA config-uration. These observations lead to a hypothesis of how the L1stalk may be employed to regulate the translocation of tRNA.

ResultsObservation of Multiple Conformations of the PRE Complex. We col-lected approximately 100,000 projection images of the G2252CPRE complex. A large degree of conformational heterogeneitywas anticipated because of the dynamic nature of this complex,as characterized by the sm-FRET study (11). Thus, we appliedmaximum likelihood (ML) classification (16) to generate sevenclasses (see Methods). We then obtained a reconstruction fromprojections falling in each of the classes using standard angularrefinement. Among the resulting density maps, one showed in-complete density for the tRNA, suggesting the existence of resi-dual heterogeneity. This map was discarded in further analysis. Inthe remaining six density maps, two pairs were found to be struc-turally quite similar, and thus the underlying data subsets werepairwise combined for reconstruction. In this way we obtained

Author contributions: J. Fu, J.B.M., S.C.B., and J. Frank designed research; J. Fu and J.B.M.performed research; J. Fu and J. Frank analyzed data; and J. Fu, J.B.M., S.C.B., and J. Frankwrote the paper.

The authors declare no conflict of interest.

Freely available online through the PNAS open access option.

Data deposition: The cryo-EM maps reported in this paper have been deposited with theElectron Microscopy Data Bank, http://emdatabank.org/ (accession no. EMD-5262).1To whom correspondence should be addressed. E-mail: jf2192@columbia.edu.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1101503108/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1101503108 PNAS ∣ March 22, 2011 ∣ vol. 108 ∣ no. 12 ∣ 4817–4821

BIOPH

YSICSAND

COMPU

TATIONALBIOLO

GY

Dow

nloa

ded

by g

uest

on

July

9, 2

021

http://emdatabank.org/http://emdatabank.org/http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1101503108/-/DCSupplementalhttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1101503108/-/DCSupplementalhttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1101503108/-/DCSupplementalhttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1101503108/-/DCSupplementalhttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1101503108/-/DCSupplementalhttp://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1101503108/-/DCSupplemental

a total of four density maps of the G2252C PRE complex (Fig. 1and Table S1). Finally, we applied real-space refinement (17) toobtain atomic models that optimally fit into the maps.

Using the crystal structure of the Escherichia coli ribosome[Protein Data Bank (PDB) IDs 212U and 212V (18)] as refer-ence, we measured the extent of the counterclockwise rotationof the 30S subunit with respect to the 50S subunit (Fig. 2 A–Cand Table 1). We term the structure with the least amount ofrotation (approximately 1.0°) the nonrotated structure (NRS),the one with 3.2° of rotation the intermediate rotated structure1 (IRS1), with 4.9°, the intermediate rotated structure 2 (IRS2),and, with 7.5°, the rotated structure (RS). We compared thesestructures with those found by X-ray crystallography of the ribo-some bound with tRNA anticodon stem loop fragments (8).The two intermediate structures (3I1Q/3I1R and 3I1Z/3I20)obtained by these authors have rotation angles that are similarto those of IRS2 (Table 1).

The movement of the L1 stalk can be characterized by a com-bination of two directions of rotation around a hinge point

located approximately at residue 2,197 of helix 76, one in theplane separating the subunits, and the other perpendicular to thatplane, toward the 30S subunit (Fig. 3). Wemeasured the angles ofthese two rotations (Table 1). It can be readily seen that the L1stalk displays a gradual movement inward, into the intersubunitspace, and the degree of this movement increases with theextent of the ratchet-like intersubunit rotation. The perpendicu-lar movement, toward the 30S subunit, is observed in the IRS2and RS structures. The L1 stalks in the crystal structures of theintermediate states (3I1M/3I1N, 3I1Q/3I1R, and 3I1Z/3I20, (8)do not display the inward rotation, but, compared to the nonro-tated structures (3I1O/3I1P, 3I1S/3I1T, 3I21/3I22), they also dis-play a rotation toward the 30S subunit (Table 1).

Novel Configurations of the P-site tRNA. The A-site tRNA remainsat the A site in all four structures, whereas the P-site tRNAmovesfrom the P-site to the P/E-hybrid position through intermediatesteps. In NRS, the P-site tRNA remains at the classical P/P posi-tion (Fig. 1A). In IRS1, even though part of the tRNA density is

A B C D50S

30S

30S

MS2 hp

h

bk

sp

P/P

A/A A/A A/A A/A

PF1 PF2 P/EE F G H

Fig. 1. The four cryo-EM density maps of the G2252C PRE complex produced by maximum-likelihood classification in XMIPPS followed by SPIDER refinement(Table S1). (Upper) Density maps of the 70S structures. (Lower) Density maps of the 30S subunit and intersubunit ligands isolated computationally. (A) and (E)Density map of the nonrotated structure (NRS). (B) and (F) Intermediate rotated structure 1 (IRS1). (C) and (G) Intermediate rotated structure 2 (IRS2). (D) and(H) Rotated structure (RS). Landmarks: bk, beak; sp, spur; MS2 hp: MS2-binding hairpin inserted into helix 98 of the large subunit for purification purpose.

A B C

D E F

bk

sp

PF1PF1

PF2PF2

P/EP/P

30S

h

Fig. 2. Comparison of the different states of theG2252C PRE complex. (Upper) Superimposition of the density maps of the 30S subunit, depicting therotation of the 30S. (Lower) Superimposition of the density maps of the P-site tRNA, depicting the conformational change of the P-site tRNA. (A) and(D) Comparison between the nonrotated structure (NRS) and intermediate rotated structure 1 (IRS1). (B) and (E) Comparison between the intermediaterotated structure 1 (IRS1) and the intermediate rotated structure 2 (IRS2). (C) and (F) Comparison between the intermediate rotated structure 2 (IRS2)and the rotated structure (RS).

4818 ∣ www.pnas.org/cgi/doi/10.1073/pnas.1101503108 Fu et al.

Dow

nloa

ded

by g

uest

on

July

9, 2

021

http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1101503108/-/DCSupplemental/pnas.1101503108_SI.pdf?targetid=ST1http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1101503108/-/DCSupplemental/pnas.1101503108_SI.pdf?targetid=ST1

weak, the orientation of the tRNA is unambiguous, providingbasis for fitting the crystal structures of tRNA into the cryo-EMdensity. Compared to the classical P/P position, the tRNA hasrotated clockwise (as viewed from the 50S subunit’s intersubunitside) around the anticodon arm by approximately 95°, havingflipped the acceptor arm toward the E site on the 50S subunit(Figs 2A and 3A). We term this position P-flip 1 (PF1). Thetip of the CCA end, as shown in the atomic model generated byreal-space refinement, lies underneath helices 78 and 77, whichare situated above helix 76 of 23S rRNA (Fig. 3).

The displacement of the tRNA in IRS2, compared to that inIRS1, can be characterized by an approximately 15° tilt towardthe E site, with the anticodon stem loop fixed on the 30S subunit’sE site, and an approximately 13° clockwise rotation (viewed fromthe 50S intersubunit side) around the anticodon arm (Figs. 1Cand 2B). We term this structure P-flipped 2 (PF2).

In RS, the P-site tRNA is in the hybrid P/E position (Fig. 1D)that has been observed in previous cryo-EM studies (5, 6).Because the A-site tRNA is in its classical A/A position, the con-figuration of the tRNA duplex in RS is thus a combination of A/A

and P/E hybrid positions, just as suggested by the previous FRETstudy of the G2252C mutant (11).

Correlation with the sm-FRET Results on the G2252C PRE Complex.Wemeasured the distance between the residues where the fluores-cent labels were attached to the tRNA (a distance we termdtRNAs) in the related sm-FRET study (11). The distances are41 Å in NRS, 59 Å in IRS1, 66 Å in IRS2, and around 69 Åin RS. These measurements agree very well with the three-statessituation inferred in the sm-FRET study: the dtRNAs measuredfor IRS2 and RS are very similar; thus, the low-FRET state ob-served in the sm-FRETstudy may be an aggregate of signals fromIRS2 and RS. Moreover, the small differences in dtRNAs betweenthe IRS1, IRS2, and RS configurations are consistent with theappearance in the histogram of the low-FRET population, whichhas a wide distribution. We also measured dtRNAs for the A/P andP/E tRNA duplex, based on the atomic model generated in a pre-vious cryo-EM study (6). Here it is around 64 Å, which is similarto those in the IRS1, IRS2, and RS states of the ribosome.

Table 1. Conformational changes observed in the four structures and related crystal structures (see ref. 8)

Rotation angle (°)L1 stalk movement Distance between

the tRNAs (Å)Distance between L1and the tRNA (Å)

Inward rotation (°) Forward rotation (°)

2I2U-2I2V 0 0 0NR 1.0 7 −10 41 93IRS1 3.2 12 −10 59 69IRS2 4.9 21 −5 66 60RS 7.5 35 12 69 373I1O -3I1P 0.8 1 −9

N/A

3I1M -3I1N 6.3 −4 23I1S -3I1T −0.9 3 −33I1Q -3I1R 5.0 0 03I21 -3I22 −0.6 3 −53I1Z -3I20 5.1 0 0

“Rotation angle” refers to the counterclockwise rotation of the 30S subunit with respect to the 50S subunit, viewed from the 30Ssubunit’s solvent side. “Larger angle”means larger degree of rotation. The “inward rotation” refers to the inward movement of the L1stalk toward the main body of the ribosome. The “forward rotation” refers to the movement of the L1 stalk toward the 30S subunit ofthe ribosome. “Distance between tRNAs” refers to the distance between residue 47 of the A-site tRNA and residue 8 of the P-site tRNA,measured based on atomic models obtained trough real-space refinement. “Distance between L1 and the P-site tRNA” refers to thedistance between residue 55 of the L1 stalk protein and residue 8 of the P-site tRNA. The model of the L1 stalk protein was obtained byaligning the L1 stalk of 3KNI, which contains the L1 stalk protein, with the atomic models of L1 stalk base obtained by real-spacerefinement.

L1 h

bk

sp

L1

sp

A C

B D

Fig. 3. Interaction of the L1 stalk with the P-site tRNA in the intermediate rotated structure 1 (IRS1) and the intermediate rotated structure 2 (IRS2). (A) and(C) Interaction of the L1 stalk with the P-flipped 1 (PF1) tRNA observed in the intermediate rotated structure 1 (IRS1). (B) and (D) Interaction of the L1 stalkwith the P-flipped 2 (PF2) tRNA observed in the intermediate rotated structure 2 (IRS2). The L1 stalk is depicted in blue, and tRNA in green. Two columnsshow the interactions from two different views. The thumbnail on the upper left corner of each column indicates the view and region from which thestructures are shown.

Fu et al. PNAS ∣ March 22, 2011 ∣ vol. 108 ∣ no. 12 ∣ 4819

BIOPH

YSICSAND

COMPU

TATIONALBIOLO

GY

Dow

nloa

ded

by g

uest

on

July

9, 2

021

Another sm-FRET study, showing that the motions of the L1stalk and the tRNAs are not tightly coupled (19), is also consis-tent with our experimental results and interpretation. We in-ferred the position of L1 stalk protein by aligning the crystalstructure of the helices 76–78 of 23S rRNA, along with that ofthe L1 protein in a ternary complex [PDB ID 3FIK (20)], tothe atomic models of NRS, IRS1, IRS2, and RS obtained byreal-space refinement. We then measured the distance betweenthe L1 stalk protein (residue 55) and the P-site tRNA elbowwhere the dye is attached. This distance, which we denote asdL1-tRNA, is found to be 93 Å in NRS, 69 Å in IRS1, 60 Å inIRS2, and 37 Å in RS. This set of distances agrees well withthe interpretation from the sm-FRETstudy, which identified fourFRET states (11). By correlating the measurements of dL1-tRNAwith dtRNAs (Table 1), we can see that their variations are indeedonly loosely coupled. It is worth noting, though, that the L1 stalkprotein may change its conformation and relative position withrespect to the L1 stalk, potentially introducing further complexityto the behavior of dL1-tRNA.

DiscussionThrough the investigation of a purified mutant ribosomal com-plex bearing the G2252Cmutation in the P loop of the 23S rRNA,we have obtained structures of the ribosome in three previouslyuncharacterized intermediate states, in which the 30S ribosomalsubunit is in intermediate rotational positions and the tRNAsare in previously uncharacterized positions. By establishing a cor-relation between tRNA positions and defined ribosome config-urations, our observations lend further credence to the notionthat tRNA repositioning goes hand in hand with distinct confor-mational changes of the ribosome.

One of the unique structures contains the (A/A, P/E) tRNAduplex, a configuration previously suggested by sm-FRETexperi-ments on the G2252C PRE sample (11). Whereas it cannot beruled out that this intermediate configuration is attained solelyas the consequence of the G2252C mutation, sm-FRET studiesof both wild-type and mutant PRE complexes suggest such inter-mediate configuration exists in the wild-type, as well, and thatthe reason we observe it by cryo-EM is that it is more populatedin the G2252C mutant. There are two possible reasons why this(A/A, P/E) configuration has not been observed before by cryo-EM. First, the structural difference between tRNAs in A/A versusA/P configurations is rather small (6). Second, the difference inthe angle of ratchet-like rotation between RS (approximately7.5°) and the PRE complex bearing the A/P and P/E duplex (ap-proximately 8.5°) is quite small, as well. Thus, it would be difficultto extract the RS state we observed here by less sensitive classi-fication methods. Our interpretation of the unique configurationsis that the point mutation may just decrease the frequency or like-lihood of transitions of RS into the fully rotated hybrid state,characterized by tRNAs situated in A/P and P/E.

In the other two unique structures revealed in our study, thedeacylated tRNA adopts distinct, flipped configurations, inwhich the acceptor arm is rotated toward the E site aroundthe anticodon arm. This poses a problem because the tRNAin the flipped position evidently must maintain its codon-antic-odon interaction. Crystallography studies showed that the antic-odon loop of the tRNA could adopt conformations that wouldaccommodate some degree of rotation of the P-site tRNA(21–23), though the size of this variation falls short of whatwe observe. Another possible source of flexibility comes fromthe mRNA, which has also been shown by crystallography studiesto have some degree of flexibility (24).

Because the movement of the P-site tRNA goes hand in handwith conformational changes of the ribosome, it can be concludedthat these unique configurations are the result of discrete inter-mediate steps that occur between the classical P-site position ofthe deacylated tRNA and formation of the P/E hybrid tRNA con-

figuration. As the G2252C point mutation is unlikely to changethe trajectory of the tRNA movement, we interpret the two ob-served intermediate positions as being physiologically relevant.These states, which have escaped detection in previous structuralstudies, may be more populated as a consequence of the mutationthrough destabilization of the classical, P/P configuration. As aresult, the number of molecules occupying normally short-livedintermediate states would be increased. Importantly, the flippedP-site tRNA configurations are present in the partially rotatedstructures, and part of the tRNA density is not well resolved,suggesting that the motions of the tRNA and the ribosomemay not be perfectly coupled. Thus, the flipped P-site tRNA con-figurations would likely be averaged out when traditional super-vised classification methods are used, which compare particlesto either preselected, classical/nonrotated, or hybrid/rotatedstates reference structures.

Probably of critical importance for the present findings, theMg2þ concentration in our specimen is 15 mM. High Mg2þ con-centrations reduce the rates of tRNA dynamics in the wild-typePRE complex (25) and favor classical A- and P-site tRNA con-figurations in the wild-type PRE complex (11, 25, 26) and hybridstate tRNA configurations in the G2252C PRE complex (11).

The L1 stalk plays an important role in translocation (5, 10,11). As far as the structure is concerned, cryo-EM studies haveshown that the L1 stalk interacts with the elbow of the P/E hybridtRNA and thereby serves to stabilize the hybrid position (5).Here, we show that, through its CCA ends, the tRNA interactswith the L1 stalk even before it forms the P/E position (Fig. 3).We believe that these interactions might be an important prere-quisite for the formation of the P/E hybrid state tRNA. Asdescribed above, in IRS1, helices 77 and 78 of the 23S rRNAlatch on to the PF1 tRNA above the tip of the CCA end (Fig. 3A).This interaction may stabilize the tRNA in the intermediate,flipped position, which would readily explain why this state issufficiently populated to be captured by classification. In theISR2 structure, the position of the L1 stalk (Fig. 3B), still allowsthe CCA end of the PF2 tRNA to interact closely with the L1stalk. Further L1 stalk movement in the direction toward the 30Ssubunit would allow helices 77 and 78 to slide over the CCA endof the tRNA, disengaging these interactions, presumably to allowthe P-site tRNA to reach the P/E hybrid state, as observed in theRS structure. The existence of discernible tRNA-L1 stalk inter-actions, combined with destabilization of the classical P/P state asa consequence of the G2252C mutation, may explain why flippedstates of the P-site tRNA are sufficiently populated (i.e., in the10% range) to be captured by classification.

Combining our results with those from previous sm-FRETexperiments (10, 11, 19, 27), we propose a mechanism for the L1stalk inducing the formation of the hybrid state (see Fig. S1). Inthis mechanism, the formation of the hybrid state is the conse-quence of a productive interaction between the L1 stalk and theCCA end of the tRNA, as seen in configurations IRS1 and IRS2.Before this productive interaction, the random motion of the L1stalk in and out of the intersubunit space provides opportunitiesfor chance encounters with the tRNA that fluctuates indepen-dently, but these encounters do not result in binding eventswith high probability. The tRNA is committed to form the hybridP/E position only when its CCA end is latched onto the L1 stalk.One piece of evidence supporting this interpretation comesfrom the related sm-FRET study (11). In the L1-stalk depletedmolecules, the rate of transitions of the tRNA duplex fromthe classical to the hybrid state is strongly decreased, suggestingthat the L1 stalk may not only stabilize the P/E state of the tRNA,but also induce its formation. The increased population in thehybrid state in the G2252C mutant can be explained by theincrease of the chances of encounters between the tRNA andthe L1 stalk: The mutation has the effect that the PF1 configura-tion is sampled more frequently by the deacylated tRNA, increas-

4820 ∣ www.pnas.org/cgi/doi/10.1073/pnas.1101503108 Fu et al.

Dow

nloa

ded

by g

uest

on

July

9, 2

021

http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1101503108/-/DCSupplemental/pnas.1101503108_SI.pdf?targetid=SF1http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1101503108/-/DCSupplemental/pnas.1101503108_SI.pdf?targetid=SF1

ing the chances for the L1 stalk to catch the CCA end of thetRNA, and thus increasing the probability of the ribosome tran-siting to the second setting.

As this study was completed, we became aware of a study ofback-translocation by Fischer et al. (28), in which multiple inter-mediate conformational states of the ribosome–tRNA complexwere identified. None of the states identified in the present studyof forward-translocation has a counterpart in the results pre-sented by Fischer and et al. although, conceivably, there mightbe a partial overlap in the states visited during the two processes.

MethodsPreparation of Ribosome Complexes. Ribosome complex was prepared aspreviously described (11).

Electron Microscopy and Image Processing. The ribosome specimen wasdiluted into a final concentration of 42 nM. A cryogrid (carbon-coatedQuantifoil 2∕4 grid, Quantifoil Micro Tools GmbH) was prepared as previousdescribed (29). Microscopy was performed under low-dose conditions onan FEI Tecnai Polara instrument operating at 300 kV and a nominal magni-fication of 59,000×. Micrographs were recorded on a 4k × 4k 16-bit TVIPS

TemCam-F415 CCD camera with a physical pixel size of 15 μm by using theautomated data collection system AutoEMation (30). The effective magnifi-cation on the CCD was 100,000× due to a postmagnification ratio of 1.7, thusmaking the pixel size 1.5 Å on the object scale. Micrographs were subse-quently decimated to a 3 Å pixel size to boost the low-spatial frequency sig-nal and reduce the processing time. Reconstruction was done followingstandard SPIDER protocols for reference-based reconstruction (31). Maxi-mum-likelihood classification (16) was performed using the XMIPP package.

Real-space refinement fitting was done as described previously (17, 32).The initial atomic models were from the crystal structure of the E. coli ribo-some [PDB IDs 2AVYand 2AW4 (33)]. The head of H38 was from PDB structure1PNY (34). The initial atomic model of A-site tRNA used was from PDB struc-ture 2WDG (35).

ACKNOWLEDGMENTS. We thank Jianlin Lei and Bob Grassucci for assistancein the microscopy, Lila Iino-Rubinstein and Melissa Thomas for assistancein the preparation of the illustrations, Eduard Schreiner for calculatingthe 30S rotation angles in Table 1, and Jesper Pallesen and Yaser Hashemfor a critical reading of the manuscript. This work was supported by theHoward Hughes Medical Institute and National Institutes of Health R01GM29169 (J. Frank), P01 GM064692 (to Robert M. Glaeser), and R01GM079238 (to S.C.B.)

1. Dunkle JA, Cate JH (2010) Ribosome structure and dynamics during translocation andtermination. Annu Rev Biophys 39:227–244.

2. Pestka S (1968) Studies on the formation of trensfer ribonucleic acid-ribosome com-plexes. V. On the function of a soluble transfer factor in protein synthesis. Proc NatlAcad Sci USA 61:726–733.

3. Moazed D, Noller HF (1989) Intermediate states in the movement of transfer RNA inthe ribosome. Nature 342:142–148.

4. Frank J, Agrawal RK (2000) A ratchet-like inter-subunit reorganization of the ribosomeduring translocation. Nature 406:318–322.

5. Valle M, et al. (2003) Locking and unlocking of ribosomal motions. Cell 114:123–134.6. Agirrezabala X, et al. (2008) Visualization of the hybrid state of tRNA binding

promoted by spontaneous ratcheting of the ribosome. Mol Cell 32:190–197.7. Frank J, Gao H, Sengupta J, Gao N, Taylor DJ (2007) The process of mRNA-tRNA trans-

location. Proc Natl Acad Sci USA 104:19671–19678.8. Zhang W, Dunkle JA, Cate JH (2009) Structures of the ribosome in intermediate states

of ratcheting. Science 325:1014–1017.9. Blanchard SC (2009) Single-molecule observations of ribosome function. Curr Opin

Struc Biol 19:103–109.10. Fei J, Kosuri P, MacDougall DD, Gonzalez RL, Jr (2008) Coupling of ribosomal L1 stalk

and tRNA dynamics during translation elongation. Mol Cell 30:348–359.11. Munro JB, Altman RB, O’Connor N, Blanchard SC (2007) Identification of two distinct

hybrid state intermediates on the ribosome. Mol Cell 25:505–517.12. Munro JB, et al. (2009) Spontaneous formation of the unlocked state of the ribosome

is a multistep process. Proc Natl Acad Sci USA 107:709–714.13. Cornish PV, Ermolenko DN, Noller HF, Ha T (2008) Spontaneous intersubunit rotation in

single ribosomes. Mol Cell 30:578–588.14. Ermolenko DN, et al. (2007) Observation of intersubunit movement of the ribosome in

solution using FRET. J Mol Biol 370:530–540.15. Dorner S, Brunelle JL, Sharma D, Green R (2006) The hybrid state of tRNA binding is an

authentic translation elongation intermediate. Nat Struct Mol Biol 13:234–241.16. Scheres SH, et al. (2006) Disentangling conformational states of macromolecules in

3D-EM through likelihood optimization. Nat Methods 4:27–29.17. Gao H, et al. (2003) Study of the structural dynamics of the E. coli 70S ribosome using

real-space refinement. Cell 113:789–801.18. Berk V, Zhang W, Pai RD, Cate JH (2006) Structural basis for mRNA and tRNA position-

ing on the ribosome. Proc Natl Acad Sci USA 103:15830–15834.19. Munro JB, et al. (2009) Spontaneous formation of the unlocked state of the ribosome

is a multistep process. Proc Natl Acad Sci USA 107:709–714.20. Villa E, et al. (2009) Ribosome-induced changes in elongation factor Tu conformation

control GTP hydrolysis. Proc Natl Acad Sci USA 106:1063–1068.

21. Barraud P, Schmitt E, Mechulam Y, Dardel F, Tisne C (2008) A unique conformationof the anticodon stem-loop is associated with the capacity of tRNAfMet to initiateprotein synthesis. Nucleic Acids Res 36:4894–4901.

22. Stuart JW, Koshlap KM, Guenther R, Agris PF (2003) Naturally-occurring modificationrestricts the anticodon domain conformational space of tRNA(Phe). J Mol Biol334:901–918.

23. Rees B, Cavarelli J, Moras D (1996) Conformational flexibility of tRNA: structuralchanges in yeast tRNA(Asp) upon binding to aspartyl-tRNA synthetase. Biochimie78:624–631.

24. Yusupova G, Jenner L, Rees B, Moras D, Yusupov M (2006) Structural basis for messen-ger RNA movement on the ribosome. Nature 444:391–394.

25. Blanchard SC, Kim HD, Gonzalez RL, Jr, Puglisi JD, Chu S (2004) tRNA dynamics on theribosome during translation. Proc Natl Acad Sci USA 101:12893–12898.

26. Kim HD, Puglisi JD, Chu S (2007) Fluctuations of transfer RNAs between classical andhybrid states. Biophys J 93:3575–3582.

27. Fei J, et al. (2009) Allosteric collaboration between elongation factor G and theribosomal L1 stalk directs tRNA movements during translation. Proc Natl Acad SciUSA 106:15702–15707.

28. Fischer N, Konevega AL, Wintermeyer W, Rodnina MV, Stark H (2010) Ribosomedynamics and tRNA movement by time-resolved electron cryomicroscopy. Nature466:329–333.

29. Grassucci RA, Taylor DJ, Frank J (2007) Preparation of macromolecular complexes forcryo-electron microscopy. Nat Protoc 2:3239–3246.

30. Lei J, Frank J (2005) Automated acquisition of cryo-electron micrographs forsingle particle reconstruction on an FEI Tecnai electron microscope. J Struct Biol150:69–80.

31. Shaikh TR, Trujillo R, LeBarron JS, Baxter WT, Frank J (2008) Particle-verification forsingle-particle, reference-based reconstruction using multivariate data analysis andclassification. J Struct Biol 164:41–48.

32. Chapman M (1995) Restrained real-space macromolecular atomic refinement using anew resolution-dependent electron-density function. Acta Crystallogr A 51:69–80.

33. Schuwirth BS, et al. (2005) Structures of the bacterial ribosome at 3.5 Å resolution.Science 310:827–834.

34. Vila-Sanjurjo A, et al. (2003) X-ray crystal structures of the WT and a hyper-accurateribosome from Escherichia coli. Proc Natl Acad Sci USA 100:8682–8687.

35. Voorhees RM, Weixlbaumer A, Loakes D, Kelley AC, Ramakrishnan V (2009) Insightsinto substrate stabilization from snapshots of the peptidyl transferase center of theintact 70S ribosome. Nat Struct Mol Biol 16:528–533.

Fu et al. PNAS ∣ March 22, 2011 ∣ vol. 108 ∣ no. 12 ∣ 4821

BIOPH

YSICSAND

COMPU

TATIONALBIOLO

GY

Dow

nloa

ded

by g

uest

on

July

9, 2

021