DOI: 10.1126/science.1249410, 1485 (2014);343 Science

et al.Alexey AmuntsStructure of the Yeast Mitochondrial Large Ribosomal Subunit

This copy is for your personal, non-commercial use only.

clicking here.colleagues, clients, or customers by , you can order high-quality copies for yourIf you wish to distribute this article to others

  here.following the guidelines

can be obtained byPermission to republish or repurpose articles or portions of articles

  ): March 28, 2014 www.sciencemag.org (this information is current as of

The following resources related to this article are available online at

http://www.sciencemag.org/content/343/6178/1485.full.htmlversion of this article at:

including high-resolution figures, can be found in the onlineUpdated information and services,

http://www.sciencemag.org/content/suppl/2014/03/27/343.6178.1485.DC1.html can be found at: Supporting Online Material

http://www.sciencemag.org/content/343/6178/1485.full.html#relatedfound at:

can berelated to this article A list of selected additional articles on the Science Web sites

http://www.sciencemag.org/content/343/6178/1485.full.html#ref-list-1, 28 of which can be accessed free:cites 70 articlesThis article

http://www.sciencemag.org/content/343/6178/1485.full.html#related-urls1 articles hosted by HighWire Press; see:cited by This article has been

http://www.sciencemag.org/cgi/collection/biochemBiochemistry

subject collections:This article appears in the following

registered trademark of AAAS. is aScience2014 by the American Association for the Advancement of Science; all rights reserved. The title

CopyrightAmerican Association for the Advancement of Science, 1200 New York Avenue NW, Washington, DC 20005. (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week in December, by theScience

on

Mar

ch 2

9, 2

014

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

on

Mar

ch 2

9, 2

014

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

on

Mar

ch 2

9, 2

014

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

on

Mar

ch 2

9, 2

014

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

on

Mar

ch 2

9, 2

014

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

on

Mar

ch 2

9, 2

014

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

13. K. D’Onise, J. W. Lynch, M. G. Sawyer, R. A. McDermott,Soc. Sci. Med. 70, 1423–1440 (2010).

14. J. Sparling, I. Lewis, LearningGames for the First ThreeYears: A Guide to Parent/Child Play (Berkley Books,New York, 1979).

15. J. Sparling, I. Lewis, LearningGames for Threes and Fours(Walker and Company, New York, 1984).

16. F. A. Campbell, C. T. Ramey, Child Dev. 65, 684–698 (1994).17. F. A. Campbell, C. T. Ramey, Am. Educ. Res. J. 32,

743–772 (1995).18. F. A. Campbell et al., Early Child. Res. Q. 23, 452–466 (2008).19. A. J. Reynolds, J. A. Temple, S. R. Ou, I. A. Arteaga,

B. A. White, Science 333, 360–364 (2011).20. K. Davis, K. K. Christoffel, Arch. Pediatr. Adolesc. Med.

148, 1257–1261 (1994).21. M. A. T. Flynn et al., Obes. Rev. 7 (suppl. 1), 7–66 (2006).22. E. Waters et al., Cochrane Database Syst. Rev. 12,

CD001871 (2011) Review.23. J. P. Romano, M. Wolf, J. Am. Stat. Assoc. 100, 94–108

(2005).24. J. Johnston, J. E. DiNardo, Econometric Methods.

(McGraw-Hill, New York, NY, ed. 4, 1997).25. D. G. Horvitz, D. J. Thompson, J. Am. Stat. Assoc. 47,

663–685 (1952).26. A. V. Chobanian et al., Hypertension 42, 1206–1252 (2003).27. K. Strong, R. Bonita, Surveillance of Risk Factors Related

To Noncommunicable Diseases: Current Status of GlobalData (World Health Organization, Geneva, 2003).

28. Treated males also report a lower bp in the age 30interview (0.190 versus 0.408) and in the history sectionof the mid-30s medical visit (0.158 versus 0.444); inneither of these cases are the differences statisticallysignificant, although the rates are remarkably similar tothose obtained for hypertension in the physical exam.

29. National Cholesterol Education Program (NCEP) ExpertPanel on Detection, Evaluation, and Treatment of HighBlood Cholesterol in Adults (Adult Treatment Panel III),Circulation 106, 3143–3421 (2002).

30. American Diabetes Association, Diabetes Care 36(suppl. 1), S11–S66 (2013).

31. M. F. Holick, T. C. Chen, Am. J. Clin. Nutr. 87,1080S–1086S (2008).

32. WHO Consultation, Definition, Diagnosis andClassification of Diabetes Mellitus and its Complications(World Health Organization, Geneva, 1999).

33. S. M. Grundy et al., Circulation 112, 2735–2752 (2005).34. P. W. F. Wilson et al., Circulation 97, 1837–1847 (1998).35. J. Cawley, The Economics of Obesity (Oxford University

Press, Oxford, 2011).36. J. Tsevat, M. C. Weinstein, L. W. Williams, A. N. Tosteson,

L. Goldman, Circulation 83, 1194–1201 (1991).

37. A. Peeters et al., Ann. Intern. Med. 138, 24–32 (2003).38. J. M. Fletcher, M. R. Richards, Health Aff. 31, 27–34 (2012).39. T. Minor, Econ. Hum. Biol. 11, 534–544 (2013).40. K. J. Hunt, R. G. Resendez, K. Williams, S. M. Haffner,

M. P. Stern, San Antonio Heart Study, Circulation 110,1251–1257 (2004).

41. H. Levy, D. Meltzer, in Health Policy and the Uninsured,C. G. MacLaughlin, Ed. (Urban Institute Press,Washington, DC, 2004), pp. 179-204.

42. T. J. Cole, Eur. J. Clin. Nutr. 44, 45–60 (1990).43. T. J. Cole, P. J. Green, Stat. Med. 11, 1305–1319 (1992).44. The LMS method is an age- and gender-dependent

transformation that translates BMI measures into zscores. The transformation is applied to two widely usedpopulation-based reference data: (i) the 2000 CDCgrowth standards; and (ii) the 2006 WHO child growthstandards scale. The LMS method transforms theinformation on the median (M), coefficient of variation(S), and skewness (L) of the BMI distribution of thesepopulation-based reference data into a Box-Cox powerfunction. This information is then used to transform BMImeasures into z scores.

45. S. E. Barlow; Expert Committee, Pediatrics 120(suppl. 4), S164–S192 (2007).

46. E. M. Taveras et al., Pediatrics 123, 1177–1183 (2009).47. D. E. Frisvold, J. C. Lumeng, J. Hum. Resour. 46,

373–402 (2011).48. P. Carneiro, R. Ginja, Long Term Impacts of

Compensatory Preschool on Health and Behavior:Evidence from Head Start (Institute for the Study ofLabor, Bonn, 2012).

49. A reduction in BMI can be caused by either betternutrition and/or more physical exercise. Physicalenvironment is certainly important: the Frank PorterGraham (FPG) Child Development Institute included alarge open space, and outside play was a part of thedaily routine (11:30 am and 3:00 pm); however, this islikely to have played an important role only after thetoddlers had started walking. Of course, better nutritiondoes not necessarily come only in the daycare center:The treated children could have enjoyed more nutritiousfood at home as well, both because their preferences forfood might have been affected and because their parentswere counseled by the pediatricians on site during thephysical exams. Finally, (47) provides evidence—basedon the What We Eat in America 2003–2004, combinedwith the National Health and Nutrition Examination Survey(NHANES) 2003–2004—that Head Start participantsconsume similar level of calories as non–Head Startparticipants during evenings and weekends but fewer caloriesin the morning and in the afternoon during the week.

50. J. J. Heckman, R. Pinto, P. A. Savelyev, Am. Econ. Rev.103, 2052–2086 (2013).

51. S. W. Barnett, L. N. Masse, Econ. Educ. Rev. 26, 113–125(2007).

Acknowledgments: The order of authorship names does notreflect the degree of contribution. G.C., J.J.H., S.H.M., andR.P. contributed equally to the design, method, andeconomic analysis presented in the paper. F.C., E.P., and Y.P.generated the medical data for this work and engaged inrepeated and productive discussions with G.C., J.J.H., S.H.M.,and R.P. over the content of the intervention and theinterpretation of the findings. Background data were generatedover many years by researchers at the University of NorthCarolina at Chapel Hill. We thank others at the Frank PorterGraham Child Development Institute at the University ofNorth Carolina (UNC) at Chapel Hill who contributed to thiswork: C. Bynum and E. Gunn, both of whom assisted indata collection and entry. In addition, thanks are due toT. Keyserling of the UNC School of Medicine, an investigatoron the original grant; G. Steen, a consultant to the medicalgrant who contributed in numerous ways to the effort; andL. Powell-Tillman of Chapel Hill Internal Medicine, whoconducted the physical examinations. We also thank M. Griffinfor excellent research assistance. Funding given to FPG/UNCfrom grant 5RC1MD004344 (National Center on MinorityHealth and Health Disparities, NIH) was used to collectthese data. The research was supported in part by theAmerican Bar Foundation, the Pritzker Children's Initiative,the Buffett Early Childhood Fund, NIH grants NICHD5R37HD065072 and 1R01HD54702, an anonymous funder, aEuropean Research Council grant hosted by University CollegeDublin (DEVHEALTH 269874), and a grant from the Institute forNew Economic Thinking (INET) to the Human Capital and EconomicOpportunity Global Working Group (HCEO)—an initiative ofthe Becker Friedman Institute for Research in Economics (BFI).The views expressed in this paper are those of the authorsand not necessarily those of the funders or persons named here.The data have been deposited by the principal investigator ofthe Abecedarian follow-up studies at the Inter-UniversityConsortium for Political and Social Research (ICPSR) at theUniversity of Michigan under identification number 34918.

Supplementary Materialswww.sciencemag.org/content/343/6178/1478/suppl/DC1Materials and MethodsSupplementary TextTables S1 to S26References (52–112)

12 November 2013; accepted 4 March 201410.1126/science.1248429

Structure of the Yeast MitochondrialLarge Ribosomal SubunitAlexey Amunts,* Alan Brown,* Xiao-chen Bai,* Jose L. Llácer,* Tanweer Hussain, Paul Emsley,Fei Long, Garib Murshudov, Sjors H. W. Scheres,† V. Ramakrishnan†

Mitochondria have specialized ribosomes that have diverged from their bacterial and cytoplasmiccounterparts. We have solved the structure of the yeast mitoribosomal large subunit using single-particlecryo–electron microscopy. The resolution of 3.2 angstroms enabled a nearly complete atomic modelto be built de novo and refined, including 39 proteins, 13 of which are unique to mitochondria,as well as expansion segments of mitoribosomal RNA. The structure reveals a new exit tunnel pathand architecture, unique elements of the E site, and a putative membrane docking site.

Mitochondria are organelles in eukaryoticcells that play a major role in metabo-lism, especially the synthesis of aden-

osine triphosphate (ATP). During evolution,mitochondria have lost or transferred most of theirgenes to the nuclei, substantially reducing the sizeof their genome (1). In yeast, all but one of thefew remaining protein-coding genes encode sub-

units of respiratory chain complexes, whose syn-thesis involves insertion into the innermitochondrialmembrane along with incorporation of prostheticgroups (2). For the translation of these genes,mitochondriamaintain their own ribosomes (mito-ribosomes) and translation system. The mitochon-drial ribosomal RNA (rRNA) and several transferRNAs (tRNAs) are encoded by the mitochondrial

genome, whereas all but one of its ribosomalproteins are nuclear-encoded and imported fromthe cytoplasm.Mitoribosomes have diverged great-ly from their counterparts in the cytosol of bacterialand eukaryotic cells and also exhibit high varia-bility depending on species (table S1) (3). Severalgenetic diseases map to mitoribosomes (4). In ad-dition, the toxicity of many ribosomal antibiotics,in particular aminoglycosides, is thought to be dueto their interaction with the mitoribosome (5).

Mitochondrial translation in the yeast Saccharo-myces cerevisiae (6) has been used as a model tostudy humanmitochondrial diseases (7). The 74Syeast mitoribosome has an overall molecularweight of 3 MD, some 30% greater than that ofits bacterial counterpart. It consists of a 54S largesubunit (1.9MD) and a 37S small subunit (1.1MD).

MRC Laboratory of Molecular Biology, Cambridge CB2 0QH, UK.

*These authors contributed equally to this work.†Corresponding author. E-mail: scheres@mrc-lmb.cam.ac.uk (S.H.W.S.); ramak@mrc-lmb.cam.ac.uk (V.R.)

www.sciencemag.org SCIENCE VOL 343 28 MARCH 2014 1485

RESEARCH ARTICLES

High-resolution crystal structures have been solvedfor bacterial (8, 9) and eukaryotic ribosomes (10–12),as well as for an archaeal large subunit (13). Incomparison, the only structures to date for mito-ribosomes come from cryo–electron microscopy(cryo-EM) reconstructions at 12 to 14 Å (14–16).

The very low yields of mitoribosomes frommost tissues have impeded their crystallization.However, the advent of high-speed direct electrondetectors (which allow accounting for particlemovement during imaging) (17, 18) and improvedalgorithms for the classification and alignment ofparticles (19) allowed us to determine the struc-ture of the large subunit of the yeast mitoribosome(hereafter referred to as 54S) to an overall reso-lution of 3.2 Å, resulting in a nearly complete,stereochemically refined atomic model.

Structure DeterminationRibosomes purified from yeast mitochondria werecontaminated with cytoplasmic ribosomes andfree 54S (fig. S1A). The use of maximum like-lihood image classification techniques (20) en-abled sorting of these classes of particles intoseparate groups (fig. S1B). A preliminary re-construction of the 74S mitochondrial ribosomeyielded a density map with overall resolution of3.6 Å. However, the small subunit was found toadopt a wide range of orientations with respectto the large subunit, resulting in markedly worseresolution for this region (~5 Å). Attempts tofurther classify the mitoribosomes into separategroups failed to yield maps that were suitablefor de novo model building of the small subunit.We focused our subsequent efforts on the largesubunit only. Density for this region was im-proved further by reconstructions that combinedthe large subunit regions of particles consisting ofthe 74S mitoribosome with those of the free 54Ssubunit, in a refinement in which the small sub-unit was masked away. The resulting total of47,124 particles yielded an overall resolutionof 3.2 Å by the “gold standard” Fourier shellcorrelation (FSC) = 0.143 criterion (fig. S2A) (21).

The local resolution (22) is better than 3 Å formost of the interior of the 54S subunit (Fig. 1 andfig. S2B) but decreases toward the periphery, es-pecially for elements suchas theL1andL7/L12stalks,which are often disordered even in crystal structuresof the ribosome. Even at the surface, the resolu-tionmostly appears to be better than 4Å (fig. S2C).

The high resolution for much of the moleculeallowed us to construct a nearly complete modelfor the large subunit, including one-third of themolecule for which no previous structural homol-ogy was known. In some regions, the quality ofthe map allowed identification of amino acids di-rectly from side-chain densities without any ad-ditional information, leading to an unambiguousassignment of specific mitoribosomal proteinsto their corresponding densities. For other regions,the identification of specific proteins required acombination of secondary structure prediction,homology modeling, and molecular replacement(23). X-ray crystallographic tools such as Coot

(24) and REFMAC (25) were modified for mod-el building, refinement, and cross-validation (23).The final model (Fig. 2A) has good stereochem-istry and agreement with the density (table S2).

The model does not include surface exposedregions for which only partial density is visible,such as the L1 and L7/L12 stalks. In addition, someof the mitochondria-specific rRNA expansions arehighly mobile and thus could not be completelyresolved. Therefore, we included only the parts forwhich clear density is apparent in appropriatelysharpened maps. The relative orientation of the twosubunits in the major class is shown in fig. S3.

Overall StructureBoth protein and RNA moieties are substantiallylarger than their bacterial counterparts. From thetotalmassof 1.9MD,0.4MDare frommitochondria-specific proteins and extensions of known bacte-rial homologs, 0.2MD are from rRNA expansionsegments (ESs), and 1.3 MD have structural equiv-alents in bacteria.Most of these additional elementsoccur on the surface of the conserved core (Fig. 2B),giving the 54S a markedly different overall shapefrom the large subunits of cytoplasmic ribosomes.

In particular, there is an expansion of the cen-tral protuberance (CP) and a new protuberanceformed at the membrane-facing surface. In bothcases, the rRNA expansion serves as a scaffoldfor the addition of new proteins. Such progressiveexpansion shares principles with the developmentof modern ribosomes themselves from primordialancestors (26). In addition to these changes at thesurface, there is a substantial remodeling of thetunnel through which the nascent protein chain pro-ceeds as it emerges from the ribosome.

rRNA Expansion SegmentsThe 21S RNA of the 54S subunit contains 3296nucleotides, of which 2714 have beenmodeled in

our structure. A secondary structure diagram wasconstructed using base-pairing information di-rectly from the structure, supplemented by pre-dictions for the unmodeled regions (Fig. 3 andfigs. S4 and S5). Conserved rRNA helices arenumbered corresponding to those in Escherichiacoli (27), with ESs denoted by the preceding con-served helix. Despite the lower sedimentationcoefficient relative toE. coli 23S, there has been aconsiderable expansion of rRNA, resulting in 392additional nucleotides spread over 11 ESs. Only afew insertions correlate with those of the eukaryoticribosome. However, as in the eukaryotic case, manyof the insertions are flexible expansion segmentsthat project out from the core and are not well re-solved in the map. It is not clear whether this sim-ilarity is a coincidental growth of complexity byneutralmeans (28) orwhether both evolved indepen-dently under selective pressure for functional reasons.

Mitochondria-Specific Protein ElementsOf the 39 proteins built into the 54S, 26 sharehomology with bacterial ribosomes, 8 are presentin both yeast and mammalian mitochondria, and5 are unique to yeast mitochondria (table S3 andfig. S6). Following a recent proposal by the ribo-some community, proteins universal to all do-mains of life have a “u” prefix and bacterialnumbering; those with only a bacterial homologhave a “b” prefix and bacterial numbering; andproteins unique to mitochondria have an “m” pre-fix. Bacterial ribosomal proteins L18, L20, andL25 are not present in the yeast mitoribosome.However, a homolog of L35 that was previouslythought not to be part of the mitoribosome (29)was clearly identified in the maps. Mitochondrialhomologous recombination protein 1 (MHR1)wasalso found to be a component of the mitoribosome(29). Density is absent formL53,mL54, andmL61,which were present in our sample as judged by

Fig. 1. Examples of the density in cryo-EM maps of the 54S subunit. (A) RNA showing a doublehelix (top) and a base pair (bottom). (B) An a helix and a b turn in proteins showing side chains. (C) Mgion that coordinates a tight turn in RNA.

28 MARCH 2014 VOL 343 SCIENCE www.sciencemag.org1486

RESEARCH ARTICLES

mass spectrometry, indicating that they are dis-ordered or dissociated during grid preparation.

Most of the homologous proteins are consid-erably larger, with an average mass of ~25 kDcompared to ~13.5 kD for their bacterial counter-parts (table S3 and fig. S6). Similar to eukaryoticribosomes, many of the extensions to bacterialproteins consist of long tails and loops extendingfrom globular domains. For example, uL24 hasapproximately three times the number of residuesand extends 115 Å across the surface of the mito-ribosome (fig. S7A). As in the eukaryotic ribo-some (11, 12), many of these extensions contactother extensions and unique proteins and rRNA(Fig. 2), suggesting a role in the assembly ofmitoribosome-specific components. Some of theflexible extensions are highly basic and probablyassist rRNA folding by neutralizing negative charge,as suggested for bacterial and archaeal ribosomes(13, 30). Only the extensions to uL17 and bL27form compact domains. uL17 is characterized bya four–helix bundle fold, whereas bL27 forms a

domain with no homology to known structures(fig. S7B); both domains interact with rRNA.

As a result of protein extensions, the numberof protein-protein contacts in the 54S is not onlylarger than in bacteria, but also exceeds that ob-served in the eukaryotic large subunit (11). Eachprotein in the mitoribosome has on average 4.5neighbors (fig. S8), in comparison with only 1.5in bacteria. Despite many more proteins overall,the number that contact only rRNA is halvedrelative to bacteria. Some of the new interactionsinclude bridging different sites of the subunit (nor-mally performed by 5S rRNA in cytoplasmicribosomes), whereas others are involved in thearchitecture of the exit tunnel.

All of the 13 mitochondria-specific proteinsinteract with rRNA, including those that have do-mains not previously shown to have a role inbinding nucleic acids. Of these, all except mL43and mL49 are bound to rRNA ESs. Analyses ofthe unique proteins show that three of them havesimilaritywithRNAbindingproteins (mL44,mL49,

andmL57), twowithDNAbinding proteins (mL58and mL60), two with nucleotide-binding proteins(mL38 and mL46), and two with lipid-bindingproteins (mL38 and mL50; table S3). These pro-teins have lost their catalytic or ligand-binding resi-dues and perform other roles in the mitoribosome.

The Central Protuberance (CP)One site of major expansion is the CP, whichbecause of its location is thought to facilitatecommunication between various functional sitesof the ribosome in a manner that is not fully un-derstood (31, 32). In bacterial, archaeal, and eu-karyotic ribosomes, the CP consists of a separate5S rRNA core interacting with several ribosomalproteins and 23S rRNA.

Yeast 54S lacks 5S rRNA and the 5S RNA-binding proteins L18 and L25. It is extensivelyremodeled (Fig. 4A) and consists ofmitochondria-specific proteins (mL38,mL40, andmL46), RNAexpansion clusters (82-ES and 84-ES1), and theextensions of the five homologous proteins (uL5,uL16, bL27, bL31, and bL33). These elementsdo not structurally mimic 5S rRNA but occupyits location in bacterial ribosomes (Fig. 4B). Thebottom “toe” of 5S rRNA is replaced by 44-ES1,which interacts with uL16 and mL60. The largeextension of bL27 binds the 82-ES1-3 to form partof the CP, illustrating how expansions in bothproteins and rRNA have been used to reform theCP (fig. S7B). Despite the absence of 5S rRNA,the CP is triple the volume of the bacterial version.There is a larger network of interactions with therest of the large subunit, as well as potential newcontactswith the head of the small subunit (fig. S3).

A Membrane-Facing ProtuberanceA second site enriched with mitochondria-specificproteins and rRNA is located adjacent to themouth of the exit tunnel and thus must be closeto the point of attachment of the ribosome to thetranslocon in the membrane (fig. S9). This protu-berance consists of mitochondria-specific mL43,44, 50, 57, and 58 (Fig. 2A) and rRNA expansionsegments 0-ES1, 0-ES2, and 44-ES3. Two inac-tive members of the ribonuclease III family, mL44and mL57, form a heterodimer and bend the0-ES by an angle of 90°, forming an “L-shaped”helix (Fig. 5, B and C). The rest of the protu-berance involves a second rRNA ES (44-ES3),the C-terminal part of mL50, N- and C-terminalextensions of uL22 and mL43, which binds pre-dominantly to helix 46.

Yeast mitoribosomes have been shown to bepermanently associated with the inner mitochon-drial membrane in both translationally active andinactive states (33, 34). Moreover, mitoribosomesremain docked to the membrane even in the ab-sence of the protein insertion machinery associ-ated with the tunnel exit, indicating that additionaland as yet uncharacterized anchors exist (35).The location of the membrane-facing protuberancemakes it a candidate to be involved in the tether-ing of mitoribosomes to the membrane, indepen-dent of nascent peptide insertion.

Fig. 2. Overview of the yeast 54S ribosomal subunit. (A) Cytoplasmic and side views of final modelwith 21S rRNA colored in gray. (B) Cytoplasmic and interface views colored showing regions conservedwith bacteria (blue), found in both yeast and mammalian mitochondria (red) and present only in yeastmitochondria (yellow).

www.sciencemag.org SCIENCE VOL 343 28 MARCH 2014 1487

RESEARCH ARTICLES

E-Site tRNAThe yeast mitoribosomes copurified with a tRNAbound in the exit (E) site. The density for thecomplete tRNA is relatively weak, presumably be-cause of a combination of partial occupancy andmixed species of tRNA (fig. S10A). Nonetheless,we were able to model the acceptor stem in-cluding the conserved 3′CCA.The terminal adenineintercalates between bases 2688 (E. coli: 2421)and 2689 (2422) in a manner identical to bacteriaand eukaryotes (fig. S10B). The tRNA appears tobe stabilized by an additional contact not seen inbacteria. A C-terminal extension in bL33 createsa highly electropositive patch that stabilizes thephosphate backbone of the tRNA acceptor stem(fig. S10C). The extension of bL33 is stabilizedby a change in the rRNApath prior to helix 82-ES1,which is itself partially stabilized by the uniqueprotein mL59. Thus, the E site of ribosomes fromarchaea (36), bacteria, (9) andmitochondria (thiswork)all bind the terminal A76 in an essentially identicalmanner by intercalating between two purines ofrRNA,while at the same time the rest of the E sitehas diverged among these classes of ribosomes.

The Exit TunnelDuring translation, the growing nascent chain pro-ceeds from the peptidyl transferase center (PTC)through a tunnel to the opposite end of the largesubunit where it emerges ~100 Å away. Givenits universality, the tunnel is highly conservedamong previously studied ribosomes. Despite thisconservation, a superposition with bacterial ribo-somes showed that the path of the bacterial tunnelis blocked in the 54S by a section of a longmitochondria-specific extension to uL23 that iswell anchored in the 54S on either side and thusunlikely to be displaced (Fig. 6A). Instead, a newtunnel path is seen that deviates from the bacterialone ~60 Å from the PTC (Fig. 6B). This is theonly channel in the yeast mitoribosome that ex-tends from the PTC to the exterior of the ribo-some. The boundaries of the new tunnel aredefined by extensions of uL22, uL23, and uL24,and by rRNA 3-ES. The tunnel leads to a newexit site located roughly 35 Å away from wherethe bacterial tunnel would have emerged (Fig. 6,B and C, and fig. S9) and is surrounded by twoadditional unique proteins,mL44 andmL50,whichare also involved in forming the membrane-facingprotuberance (Fig. 6B). This suggests that theattachment to a translocon-like entity for inser-tion of the nascent peptide into the mitochondrialmembrane must be very different in the 54S thanin bacterial ribosomes. In support of this idea,uL24 was chemically found to form cross-linkswith uL23, mL44, andmL50 (37). The exit of thetunnel is also wider than in bacteria, possibly toallow cotranslational assembly of oxidative phos-phorylation complexes (35, 38).

The entrance to the tunnel just beyond thePTC is noticeably narrower than in all previouslydetermined structures of the large subunit (Fig.6D). In part this is due to a base pair betweenU2877 and A1958 that appears unique to yeast

mitochondria and could only be identified on thebasis of the structure. This pair brings opposingstrands of 21S rRNA closer together by about8 Å. This location is the binding site of macrolideantibiotics, which, unlike many aminoglycosides,do not bind to mitochondrial ribosomes (39) andthus do not have the same toxicity. A superposi-tion of the macrolide erythromycin in its bindingsite shows that it would clash with the narrowertunnel entrance in mitochondria, so in yeast mito-ribosomes there is an additional mechanism forresistance to these antibiotics beyond sequencevariation of the binding pocket. Beyond the en-trance, the tunnel is lined by conserved elements

(Fig. 6C), including the constriction site formed byuL22 and uL4. In addition, near the point wherethe path diverges from the path of the bacterialtunnel, two aromatic residues of uL22 and uL24narrow the tunnel to a diameter of ~9Å (Fig. 6C).

This work shows that recent advances incryo-EM can be used to determine structures ofcomparable quality to x-ray crystallography butfrom much smaller amounts of more heteroge-neous material. It also reveals that this highlydivergent ribosomal subunit differs substantiallyfrom those of bacteria, eukaryotes, and archaea,with several unique features including an alteredpath of the exit tunnel.

Fig. 4. Central protuberance (CP) of the 54S subunit. (A) Proteins involved in the CP. (B) Elementsthat replace 5S rRNA (gray) in the CP.

Fig. 3. RNA of the 54S subunit. (A) Schematic diagram of the secondary structure of 21S rRNA. (B) Twoviews of tertiary structure showing mitochondria-specific ESs in red labeled after the conserved helixaccording to standard numbering (see text).

28 MARCH 2014 VOL 343 SCIENCE www.sciencemag.org1488

RESEARCH ARTICLES

References and Notes1. C. G. Kurland, S. G. Andersson, Microbiol. Mol. Biol. Rev.

64, 786–820 (2000).2. M. Ott, J. M. Herrmann, Biochim. Biophys. Acta 1803,

767–775 (2010).3. R. K. Agrawal, M. R. Sharma, Curr. Opin. Struct. Biol. 22,

797–803 (2012).

4. P. Smits, J. Smeitink, L. van den Heuvel, J. Biomed.Biotechnol. 2010, 737385 (2010).

5. C. N. Jones, C. Miller, A. Tenenbaum, L. L. Spremulli,A. Saada, Mitochondrion 9, 429–437 (2009).

6. J. M. Herrmann, M. W. Woellhaf, N. Bonnefoy, Biochim.Biophys. Acta 1833, 286–294 (2013).

7. A. Barrientos, IUBMB Life 55, 83–95 (2003).

8. B. S. Schuwirth et al., Science 310, 827–834 (2005).9. M. Selmer et al., Science 313, 1935–1942 (2006).

10. J. Rabl, M. Leibundgut, S. F. Ataide, A. Haag, N. Ban,Science 331, 730–736 (2011).

11. S. Klinge, F. Voigts-Hoffmann, M. Leibundgut,S. Arpagaus, N. Ban, Science 334, 941–948 (2011).

12. A. Ben-Shem et al., Science 334, 1524–1529 (2011).13. N. Ban, P. Nissen, J. Hansen, P. B. Moore, T. A. Steitz,

Science 289, 905–920 (2000).14. M. R. Sharma et al., Cell 115, 97–108 (2003).15. J. A. Mears et al., J. Mol. Biol. 358, 193–212 (2006).16. M. R. Sharma, T. M. Booth, L. Simpson, D. A. Maslov,

R. K. Agrawal, Proc. Natl. Acad. Sci. U.S.A. 106,9637–9642 (2009).

17. X.-C. Bai, I. S. Fernandez, G. McMullan, S. H. Scheres,eLife 2, e00461 (2013).

18. X. Li et al., Nat. Methods 10, 584–590 (2013).19. S. H. Scheres, J. Struct. Biol. 180, 519–530 (2012).20. S. H. Scheres, J. Mol. Biol. 415, 406–418 (2012).21. S. H. Scheres, S. Chen, Nat. Methods 9, 853–854 (2012).22. A. Kucukelbir, F. J. Sigworth, H. D. Tagare, Nat. Methods

11, 63–65 (2013).23. See supplementary materials on Science Online.24. P. Emsley, B. Lohkamp, W. G. Scott, K. Cowtan,

Acta Crystallogr. D 66, 486–501 (2010).25. G. N. Murshudov et al., Acta Crystallogr. D 67, 355–367

(2011).26. G. E. Fox, Cold Spring Harb. Perspect. Biol. 2, a003483

(2010).27. P. Maly, R. Brimacombe, Nucleic Acids Res. 11, 7263–7286

(1983).28. M. W. Gray, J. Lukes, J. M. Archibald, P. J. Keeling,

W. F. Doolittle, Science 330, 920–921 (2010).29. P. Smits, J. A. M. Smeitink, L. P. van den Heuvel,

M. A. Huynen, T. J. G. Ettema, Nucleic Acids Res. 35,4686–4703 (2007).

30. B. T. Wimberly et al., Nature 407, 327–339 (2000).31. O. A. Dontsova, J. D. Dinman, Int. J. Biomed. Sci. 1, 1–7

(2005).32. A. Ben-Shem, L. Jenner, G. Yusupova, M. Yusupov,

Science 330, 1203–1209 (2010).33. M. Prestele, F. Vogel, A. S. Reichert, J. M. Herrmann,

M. Ott, Mol. Biol. Cell 20, 2615–2625 (2009).34. M. Ott et al., EMBO J. 25, 1603–1610 (2006).35. M. Keil et al., J. Biol. Chem. 287, 34484–34493

(2012).36. T. M. Schmeing, P. B. Moore, T. A. Steitz, RNA 9, 1345–1352

(2003).37. S. Gruschke et al., J. Biol. Chem. 285, 19022–19028

(2010).38. S. Gruschke et al., J. Cell Biol. 193, 1101–1114 (2011).39. N. D. Denslow, T. W. O’Brien, Eur. J. Biochem. 91,

441–448 (1978).

Acknowledgments: We thank J. Walker for co-initiatingthis project and his support in the early phase of the work;G. McMullan, X. Chen, and C. Savva for help with datacollection; I. S. Fernández and D. Tourigny for help withpreliminary modeling; S. Peak-Chew and M. Skehel formass spectrometry; J. Grimmett and T. Darling for help withcomputing; and R. Nicholls for help with ProSMART. Supportedby UK Medical Research Council grants MC_U105184332(V.R.), MC_UP_A025_1013 (S.H.W.S.), and MC_UP_A025_1012(G.M.); a Wellcome Trust Senior Investigator award (WT096570),the Agouron Institute, and the Jeantet Foundation (V.R.);and fellowships from Human Frontiers Science Program(A.A.), EU FP7 Marie Curie (X.-C.B.), FEBS ( J.L.L.), and EMBO(T.H.). Cryo-EM density maps have been deposited withthe EMDB (accession no. EMD-2566) and coordinates havebeen deposited with the Protein Data Bank (entry code 3J6B).

Supplementary Materialswww.sciencemag.org/content/343/6178/1485/suppl/DC1Materials and MethodsFigs. S1 to S10Tables S1 to S3References (40–73)

6 December 2013; accepted 14 February 201410.1126/science.1249410

Fig. 5. A membrane-facing protuberance. (A) Site of expansion of the 54S relative to bacteria. (B)ES0 of 21S rRNA serves as an anchor for the protuberance. (C) Two views of the protuberance showing howmitochondria-specific proteins bind to the expansion segment.

Fig. 6. The putative exit tunnel in the 54S subunit. (A) The bacterial tunnel (top, path in orange) isblocked by a mitochondria-specific extension to uL23 (detail in bottom panel). (B) The 54S tunnel (bluemesh) has been extensively remodeled by mitochondria-specific protein elements (colored) (top), re-sulting in a tunnel exit (bottom) that is substantially different from that in bacteria. In (A) and (B) the lowerpanels are 90° rotations of the upper panels. (C) Conservation of elements that make up the entrance andtop part of the tunnel. The constriction at the top is narrower, and two aromatic amino acids that line thetunnel make a close approach. The path of the bacterial tunnel (orange) would exit 35 Å away. (D) Narrowingof the 54S tunnel entrance (yellow) due to an AU base pair not found in bacteria (white) that would causea steric clash with macrolides such as erythromycin (pink).

www.sciencemag.org SCIENCE VOL 343 28 MARCH 2014 1489

RESEARCH ARTICLES