Conformational variability in Escherichia coli 70S ribosome as revealed by 3D cryo-electron microscopy

The International Journal of

PERGAMON

The International Journal of Biochemistry & Cell Biology 31 (1999) 243±254

Biochemistry & Cell Biology

Conformational variability in Escherichia coli 70S ribosome as revealed by 3D cryo-electron microscopy Rajendra K. Agrawal a, *, Ramani K. Lata a, 1, Joachim Frank b, c a

Wadsworth Center, New York State Department of Health, Empire State Plaza, P.O. Box 509, Albany, NY 12201-0509, USA b Howard Hughes Medical Institute, Wadsworth Center, State University of New York at Albany, Empire State Plaza, P.O. Box 509, Albany, NY 12201-0509, USA c Department of Biomedical Sciences, State University of New York at Albany, Empire State Plaza, P.O. Box 509, Albany, NY 12201-0509, USA

Abstract During protein biosynthesis, ribosomes are believed to go through a cycle of conformational transitions. We have identi®ed some of the most variable regions of the E. coli 70S ribosome and its subunits, by means of cryo-electron microscopy and three-dimensional (3D) reconstruction. Conformational changes in the smaller 30S subunit are mainly associated with the functionally important domains of the subunit, such as the neck and the platform, as seen by comparison of heat-activated, non-activated and 50S-bound states. In the larger 50S subunit the most variable regions are the L7/L12 stalk, central protuberance and the L1-protein, as observed in various tRNA-70S ribosome complexes. Di€erence maps calculated between 3D maps of ribosomes help pinpoint the location of ribosomal regions that are most strongly a€ected by conformational transitions. These results throw direct light on the dynamic behavior of the ribosome and help in understanding the role of these ¯exible domains in the translation process. # 1999 Elsevier Science Ltd. All rights reserved.

1. Introduction Abbreviations: 3D, three-dimensional, Cryo-EM, cryoelectron microscopy, PRE, pre-translocational, POST, posttranslocational. * Corresponding author. 1 Present address: Laboratory of Structural Biology, National Institute of Arthritis, Musculoskeletal and Skin Diseases, National Institutes of Health, Bethesda, MD 20892, USA.

Ribosomes are complex and dynamic ribonucleoprotein assemblies which provide the framework for protein biosynthesis in all organisms. The E. coli 70S ribosome comprises of two unequal subunits: the smaller 30S subunit is composed of 16S ribosomal RNA (rRNA) and 21 di€erent proteins whereas the larger 50S subunit

1357-2725/99/$ - see front matter # 1999 Elsevier Science Ltd. All rights reserved. PII: S 1 3 5 7 - 2 7 2 5 ( 9 8 ) 0 0 1 4 9 - 6

244

R.K. Agrawal et al. / The International Journal of Biochemistry & Cell Biology 31 (1999) 243±254

is composed of two rRNA molecules, 5S and 23S, and 34 di€erent proteins (for reviews, see Refs. [1, 2]). An understanding of the detailed structure of ribosomes and its individual subunits is the key to understanding the mechanism of translation. Approximate positions of most of the proteins have been mapped on the ribosome using neutron scattering [3, 4] and immunoelectron microscopy (reviewed in Ref. [5]) techniques. Extensive work has gone into the elucidation of 3D models of rRNAs, the main sca€olding materials of both the subunits and enormous progress has been made, especially in the case of smaller 30S subunit [6±9]. Because of the large size (more than 2 MD) and the structural complexity of the ribosome, progress in obtaining its crystal structure has been rather slow, although crystals of 70S ribosome and both of its subunits have been reported which di€ract to a resolution range of 3 AÊ [10]. Meanwhile, cryo-electron microscopy (cryo-EM) and three-dimensional (3D) reconstruction from single particles has been the most successful method to date to study the structure of the ribosome at a lower resolution range of 20±25 AÊ [11± 13]. More recently, a 3D cryo-map of the 70S ribosome has been obtained at a resolution better than 15 AÊ [14]. In the last couple of years signi®cant progress has been made in locating the 3D interaction sites of some of the ligands on the ribosome using the cryo-EM and 3D reconstruction technique, e.g. tRNA [14±16] and elongation factors (EF-Tu ternary complex: Ref. [17] EF-G: Ref. [18]). During protein biosynthesis, the ribosome interacts with various ligands, such as mRNA, tRNAs, initiation factors, elongation factors and release factors, and is thought to undergo a series of conformational changes to facilitate the various steps of the process. Elongation factors (EFG and EF-Tu) interact with the ribosome to promote two di€erent crucial steps of the elongation cycle: EF-Tu delivers the aminoacyl tRNA to the A site in the form of a ternary complex of aminoacyl-tRNA, EF-Tu and GTP, whereas EF-G promotes the movement of A-site tRNA along with the mRNA codon to the P site. Both EFs require a GTP molecule as the source of energy

to complete their tasks. However, the ribosome has the intrinsic property to perform both steps even without the help of EFs [19, 20], a feat that is not conceivable unless one assumes that some conformational changes take place. In the past, various biochemical techniques have been employed to study the conformational changes of the ribosomes under various conditions. However, none of these has given a direct clue about the exact topographical location and magnitude of the postulated changes. Earlier works have indicated that ribosomes are present in two di€erent conformational states [21±26], which were thought to be related to two main physiological states of the ribosome, the pre(PRE) and post-translocational (POST) states. These studies suggested that the 50S subunit has a more variable structure than the 30S subunit and that it is its conformational variability that is mainly responsible for the existence of the two conformational states in the 70S ribosome. Among the morphological features of the 50S subunit, the L7/L12 stalk is thought to be the most ¯exible part, which is composed of extended and folded forms of the four copies of L7/L12 proteins (see Ref. [27]). Cryo-electron microscopy is the obvious method of choice to study these postulated or inferred changes by direct visualization. Without being the explicit focus of the various cryo-EM studies, accounts of conformational changes are scattered throughout the relatively young literature. The present article is a ®rst attempt to bring these observations together and to discuss the possible implications in the various steps of protein synthesis. Information of the type provided here is essential in establishing a link between structure and function of the translational machinery and also in providing necessary constraints for consideration in future model building attempts. The resolution limitations and the apparent sensitivity of the ribosome's conformation to variations in bu€er conditions make it necessary to restrict the interpretation to mass movements in the 3D cryo-map that are substantial. For instance, di€erence maps obtained in all e€orts to localize tRNA bound to the ribosome typically show several minor density peaks within

R.K. Agrawal et al. / The International Journal of Biochemistry & Cell Biology 31 (1999) 243±254

245

Fig. 1. 3D cryo-EM maps of non-activated (a) and heat-activated (b) 30S subunits shown from intersubunit space side. (c) Two subunits superimposed to reveal the regions conformational changes associated with heat activation. The heat-activated subunit is shown as wiremesh structure.

the 30S subunit portion of the 70S map [14±16] that are too small to be signi®cant. 2. Results 2.1. Conformational ¯exibility of the 30S subunit The 30S subunit provides the site for the interaction between the mRNA codon and the anticodon of tRNA. It also interacts with three initiation factors, IF1, IF2 and IF3, during the initiation step of the protein synthesis. IF3, which is also known as antiassociation factor, binds to the 30S subunit and prevents its association with the 50S subunit. In in vitro conditions, the association of the two subunits can be prevented, without the involvement of IF3, by manipulation of Mg2+ concentration. At lower Mg2+ concentration (E1 mM), 30S and 50S subunits do not associate to form the 70S ribosome. On the other hand, the association of the two subunits are induced by changing temperature, e.g. by incubation at 378C, and higher Mg2+ concentration (4±10 mM). Thus, in vitro some of the physiologically important steps can be controlled and manipulated by change of ionic conditions and temperature [28]. 3D-Cryo maps of 30S subunit in heat-activated and non-activated conditions [29] for the ®rst time provided gross topographical information on the location of

conformationally variable parts of the subunit under physiologically relevant conditions. Here we present a more detailed comparison of these maps together with the map of the 30S subunit in its 50S subunit-bound state [11]. This comparison unveils both the locations and the magnitudes of the conformational changes. The di€erence maps show that the conformational changes are mainly localized in the neck, the platform and the shoulder regions, indicating that these are the most variable regions of the 30S subunit. 2.2. Comparison between cryo-maps of heatactivated and non heat-activated 30S subunits Our comparison of the 30S structure in these two states, non-activated (Fig. 1a) and heat-activated (Fig. 1b), shows a band of mass where changes are located passing through the neck from shoulder to the platform (Fig. 1c). This is the main region of the subunit that is exposed in the intersubunit space, which is formed upon the association of 50S subunit (see Ref. [30]). The band of mass is located in the region of the cleft that is mainly involved in the interaction of the 30S subunit with the tRNA molecule at various stages of the elongation cycle [14±16]. Positive (Fig. 2a) and negative (Fig. 2b) di€erence maps highlight the regions that are highly a€ected by the heat activation. Looking at Fig. 1(c) and

246

R.K. Agrawal et al. / The International Journal of Biochemistry & Cell Biology 31 (1999) 243±254

be shifted and merged into the platform in the activated state. 2.3. Comparison between heat-activated and 50Sbound 30S subunits

Fig. 2. Positive and negative di€erence maps between maps of non-activated (yellow) and heat-activated (blue) 30S subunits. (a) Di€erence map obtained by subtracting the 3D map of non-activated from that of heat-activated 30S subunit, superimposed on the 3D map of the non-activated one. (b) Negative di€erence map superimposed on the 3D map of heat-activated 30S subunit.

Fig. 2(a) and (b), one can see that there is a bending of the head, narrowing the gap between head and shoulder of the subunit by approximately 10 AÊ from the non-activated to the activated state. At the same time, the shoulder rises slightly (by approximately 5 AÊ) towards the head and the platform moves towards the head, narrowing the angle between head and platform by approximately 108. All these changes have a pronounced e€ect at the junction point of the head, platform and the main body. In the non-activated state we observe a hump between the base of the platform and the shoulder on the interface side. This hump (marked * in Fig. 2b) appears to

Signi®cant conformational changes were observed when comparing 3D maps of heat-activated (Fig. 1b and Fig. 3a) and 50S-bound 30S subunit (Fig. 3b). The latter was isolated by applying a cutting plane at the 30S±50S interface of the 70S map [11, 12, 29]. As compared to the non-activated 30S subunit, the 3D map of the heat-activated 30S subunit shows greater resemblance with that of the 50S-bound state (also see Ref. [29]). Our comparison shows that shoulder and head of the 50S-bound 30S subunit are much closer to each other than seen in either of the unbound states. The head has drastically bent toward the shoulder and the interface side (by approximately 138), to the extent that it merges with the shoulder. This movement produces a channel between the neck and the shoulder, which is the proposed conduit for mRNA [11, 15, 29]. The 30S body, below the shoulder, appears to undergo a conformational change (see Fig. 3c); however, this change is not as pronounced in the upper shoulder region as seen in the comparison between activated and non-activated subunit (compare Fig. 2a and

Fig. 3. 3D cryo-EM maps of heat-activated (a) and 50S-bound (b) 30S subunits. (c) Two subunits superimposed to reveal the regions conformational changes associated with subunit-subunit association. In this case, the 50S-bound subunit is shown as wiremesh structure.

R.K. Agrawal et al. / The International Journal of Biochemistry & Cell Biology 31 (1999) 243±254

Fig. 4. Positive and negative di€erence maps between maps of heat-activated (blue) and 50S-bound (pink) 30S subunits. (a) Di€erence map obtained by subtracting the 3D map of heatactivated from that of 50S-bound 30S subunit superimposed on the 3D map of heat-activated one. (b) Negative di€erence map superimposed on the 3D map of 50S-bound 30S subunit.

Fig. 4a). The platform, mainly in the tip portion, curls inward, toward the head, further narrowing the angle between the head and the platform by approximately 58. Positive (Fig. 4a) and negative (Fig. 4b) di€erence maps obtained by subtracting the 3D maps

247

of the heat-activated and the 50S-bound 30S subunit show a signi®cant shift of mass of density from the upper part of the head to the lower part. The di€erence map (Fig. 4a) also shows the mass of density at the tip of the platform, which is due to the curled shape of the platform in the 50S-bound state. A mass of density is observed near the neck, on the solvent side (marked by * in Fig. 4a). This is due to a more prominent back lobe structure [29] in the 50S-bound subunit. Few small masses are also seen on the lower parts of the 30S subunit body, which probably represent the conformational changes related to subunit-subunit association. 2.4. Conformational ¯exibility of the 50S subunit In addition to the L7/L12-stalk region, which was thought to be mainly responsible for the conformational changes of the 70S ribosome, we have identi®ed some other regions of the 50S

Fig. 5. (a) 3D cryo-EM map of fMet-tRNAMet ±70S ribosome complex at 15 AÊ resolution [14] shown in 30S±50S side view. The f 30S subunit is on the left-hand side, the 50S subunit on the right-hand side. The dashed line indicates the plane used for the separation of the 30S subunit portion from the 50S subunit (shown in b). (b) The 50S subunit (isolated by applying a cutting plane to the 70S map as shown in a) is shown in the intersubunit space face view. Di€erence masses (shown in pink, green and yellowish brown) represent highly variable regions of the 50S subunit. The pink mass, in the immediate vicinity of the L7/L12 stalk, was obtained by subtracting the 3D map of fMet-tRNAMet ±70S ribosome complex form that of a POST complex (see Section 4). The f green and the yellowish brown masses, in the vicinity of central protuberance and the L1-protein, respectively, were obtained by subtracting the 3D map of the naked 70S control from that of fMet-tRNAMet ±70S ribosome complex. The yellowish brown mass f has been presented at slightly lower threshold value than the green mass. All three di€erence masses represent the most prominent di€erence peaks, other than the tRNA peaks (not shown), associated with conformational changes.

248

R.K. Agrawal et al. / The International Journal of Biochemistry & Cell Biology 31 (1999) 243±254

subunit that appear to be highly variable. These are the central protuberance and the L1-protein regions. All three regions interact with the tRNA molecule during the various phases of elongation cycle. According to general wisdom, the L7/L12 stalk region is close to the A site, while the L1protein region is close to the E site of the ribosome and the P site tRNA is located in between, i.e. close to the central protuberance region (e.g. Ref. [31]). Here we will present some of the results obtained from various tRNA±ribosome complexes that have been analyzed by the 3D cryo-EM technique. In our analysis of the ®rst tRNA±ribosome complex [15], we observed a conformational change in the L7/L12-stalk region, where the stalk was found to be extended as compared to its position in the 3D map of the naked control 70S ribosome [11]. Subsequent analyses of PRE and POST complexes showed that the stalk assumes di€erent degrees of extension. In addition to densities corresponding to the P- and the E-site tRNAs, the POST complex shows a mass of density in the immediate vicinity of the L7/L12 stalk, on the interface canyon side (Fig. 5b; shown in pink). This mass of density, which is a part of the di€erence map obtained by ±70S subtracting the 3D map of fMet-tRNAMet f complex from the 3D map of POST complex and is found in the overlapping position of the A-site tRNA in the PRE complex, re¯ects a movement of the stalk region associated with the translocation process. The elbow region of the P-site tRNA has been found in direct contact with the intersubunitspace side of the central protuberance (CP) [14]. In the same study, it was observed that binding of tRNA at the P site induces a prominent conformational change (Fig. 5b; shown in green) at the contact point between the head of 30S subunit and the CP of 50S subunit [11, 29], on the intersubunit-space side. This conformational change possibly results from a conformational change in the 5S rRNA, which is present in the CP, where both 3 0 and 5 0 ends of 5S rRNA have been localized [32, 33]. Variability of the same CP region is also observed in many other tRNA±ribosome complexes.

The L1-protein feature of the 50S subunit, visible in the cryo-EM map as a globular mass supported on a stem [14], appears to be a strongly variable region of the ribosome. The conformation of this region is altered upon binding of a ligand in its immediate vicinity (e.g. E-site tRNA binding) as well as upon binding of a ligand at a distant part of ribosome (e.g. binding of tRNA at the P site or binding of a ligand in the L7/L12 stalk region). In all the di€erence maps computed from various ligand-binding experiments that we have conducted so far, an extra mass of density is found in the immediate vicinity of L1 (Fig. 5b; shown in yellowish brown color). 3. Discussion 3.1. Conformational changes at/near the decoding region of the 30S subunit Earlier studies have clearly shown that the reactivity of 16S RNA is dramatically altered between heat-activated and non-activated states. Nuclease accessibility, chemical modi®cation, cross-linking and complementary oligodeoxynucleotide binding data have been used to determine the state of the regions of 16S RNA, in terms of whether they are exposed, single- or double-stranded [34±40] in the 30S subunit in various functional states. Similarly, the reactivity of several 30S ribosomal proteins [41, 42] is also altered. It has been suggested [28, 36] that the observed changes due to thermal activation of the 30S subunit might mimic a natural process in vivo and thus cannot be described as a simple loosening or unfolding of the native structure. However, this hypothesis needed direct experimental veri®cation. Ericson and Wollenzien [39], in their long-distance crosslinking experiments, have observed changes during inactive to active transition in intra-molecular crosslinks in the 16S rRNA. They studied two regions that are placed at the junction of the head, platform and the main body of the 30S subunit. It is important to note that most of the highly conserved and functionally important 16S RNA domains have been mapped within these regions [43±47]. Our results

R.K. Agrawal et al. / The International Journal of Biochemistry & Cell Biology 31 (1999) 243±254

clearly support the conformational ¯exibility of these regions. Accessibility of complementary oligonucleotides to the 16S RNA regions, which have been located in platform and neck regions of the 30S subunit, were di€erent in heat-activated as compared to non-activated 30S subunits [40]. Furthermore, Agrawal and Burma [26] found di€erential accessibility of complementary oligonucleotides in tight and loose couple ribosomes, which presumably represent the pre- and post-translocational states of the ribosome, respectively. In this context it is interesting to note that Stark et al. [16] found a substantial conformational change in the platform region of the 30S subunit in the transition from the pre- to the posttranslocational state. Both head and platform of the 30S subunit are attached to the main body of the subunit by only minimal connections. The head of the subunit is connected to the body of the ribosome by a single strand of RNA [6±9], which explains its capacity for independent movement. Moreover, the site of convergence of the body, head and platform is at the cleft and neck of the subunit, precisely where the anticodon arms of the A- and Psite tRNAs have been localized [14±16]. The decoding region [43] of 16S RNA (i.e. C-1400 region) has been localized in the cleft near the neck [46, 48]; the anti-Shine-Dalgarno region, responsible for positioning mRNA during the initiation of prokaryotic protein synthesis [49], in a region on the platform [45, 47] and the 530 loop region, a 16S RNA region known to play an important role in the tRNA interaction at the ribosomal A-site [50±53], near the shoulder [44]. Conformational changes observed upon comparison between heat-activated and 50S-bound 30S subunits, which can be seen in Fig. 3(c), a€ect the topography of the neck and shoulder regions of the bound subunit in a signi®cant manner, probably creating a completely new environment of ribosomal proteins and rRNA precisely in the region where the anticodon of A-site tRNA has been found to interact with the 30S subunit [15, 16]. The conformational rearrangement of this region has the consequence that the anticodon of the P-site tRNA encounters a di€erent environment when comparing the initiation

249

and the elongation phases. Since the tRNA does not bind to isolated 30S subunit at the A site, it is likely that the A site on the 30S subunit is actually created in the course of the observed structural rearrangement, i.e. only after the subunits have associated, at the very beginning of the ®rst elongation cycle. Variability in the regions of the 30S subunit that are mostly in the immediate vicinity of the tRNA and mRNA binding region strongly support the idea that conformational rearrangements associated with these speci®c regions might have an active role in vivo, facilitating various phases of initiation and elongation. If the movement of the tRNA were coordinated with the movement of one or more of these structural elements of the small subunit, translocation could be accomplished while intermolecular contacts between the mRNA±tRNA complex and the 30S subunit are transiently maintained. Furthermore, during the elongation cycle, when the mRNA is no longer in contact with the anti-Shine-Dalgarno region of the 16S RNA, the proposed mRNA channel, which is apparently formed on the 30S subunit upon 30S and 50S subunits association [29], must be playing an important role in maintaining proper alignment for the mRNA. Chemical crosslinking data, incorporated into the 3D model of 16S RNA [54] using the cryo-EM map of the 50S-bound 30S subunit as envelope constraint, is highly supportive of the existence of such a channel for mRNA at this particular location. 3.2. Is the 30S subunit involved in the translocation step? EF-G, responsible for promoting translocation step, has traditionally been assigned to the large subunit, as far as its action in EF-G-dependent GTP hydrolysis is concerned. However, threedimensional mapping of EF-G [18, 55] clearly shows that domain IV of EF-G acts on the small subunit to e€ect the movement of tRNA from the A site. The domain that binds to the anticodon region of A-site tRNA or, alternatively, to the tip of domain IV of EF-G [18], appears to be one of the most ¯exible domains of the 30S sub-

250

R.K. Agrawal et al. / The International Journal of Biochemistry & Cell Biology 31 (1999) 243±254

unit. Since the ribosome can perform the translocation step even without help from EF-G [19, 20], the observation of conformational changes in this 30S region is intriguing. Furthermore, a recent mutation study of 16S rRNA [56] supports the active involvement of the 30S subunit in the translocation process. The study shows that there are two alternative base pairings for the threebase pair sequence of nucleotides 910±912 (CUC) of 16S rRNA, both of which are required for ribosome function. The CUC sequence pairs either with nucleotides 888±890 (GAG) or with the immediately adjacent nucleotide 885±887 (GGG) sequence. A tRNA-dependent conformational change in this region of 16S rRNA has been suggested earlier [57]. Although these ®ndings imply that the triplet switch is primarily involved in tRNA selection, the nature of the switch itself and the movement of the CUC sequence by precisely three nucleotides between adjacent complementary triplets has an interesting resemblance to the process of translocation. This 16S rRNA region has been placed [9] at the junction point of body and the platform of the 30S subunit, close to the region where the ¯exible hump region described earlier (* in Fig. 2b) has been found. 3.3. Conformational changes associated with 50S subunit and translocation In the 50S subunit, three main structural regions, L7/L12 stalk, CP, and L1-protein, have been identi®ed as the most ¯exible domains. Interestingly, these structural elements are directly associated with A-, P- and E-site tRNAs, respectively. Involvement of L7/L12 stalk in the translocation has been implicated in the past by several laboratories [23, 27, 58±62] both directly and indirectly, through involvement of elongation factor G (EF-G). In a recent study, we have observed multiple points of contact between the stalk base and various domains of EF-G [18]. In the same study, the stalk was found in an unusually stable, well extended conformation. The mass of density observed at the base of the stalk, on the interface canyon side, in the POST com-

plex (Fig. 5b; pink) clearly shows the involvement of that region in the translocation step. The L1-protein region has been found to be one of the most ¯exible region of the 50S subunit. In another cryo-EM study it has been shown that L1 protein conformation is highly sensitive to bu€er condition as well as to tRNA binding. In domain V of the 23S RNA, the nucleotide region 2090±2200 has been implicated in the binding of the L1 protein [63, 64]. Of these, the 2109±2119 nucleotide region [65] and nucleotides 2112, 2116 and 2169 [66] have been shown to be involved in the binding of E-site tRNA. There is evidence supporting the existence of a conformational change near the exit site of the E. coli ribosome upon tRNA binding [65, 67]. Such conformational changes occur even upon binding of peptidyl-tRNA [14] (as shown here) as well as deacylated-tRNA [54] at the P site. In fact, tRNAs at the ribosomal P and E sites are crosslinked with this protein or with other ribosomal components present in its immediate neighborhood, e.g. the 2111±2112 nucleotide region of 23S RNA [68] and protein L33 [69]. It should be mentioned that Stark et al. [16] found no evidence of gross morphological changes accompanying the transition from the pre- to the posttranslocational state. Since both subunits are required for the translocation reaction, a rhythmic coordinated movement of the ¯exible regions from both the subunits appears to play a major role in the process. Several models for the mechanism of translocation have been proposed over the years (for reviews see Refs. [70, 71]). One of the most favored models, the hybrid site model [72], suggests a movement of the two ends of tRNA in two independent steps and incorporates a relative movement of the two subunits with respect to each other, while another model, the movable domain model, also known as the a±e model [73], suggests movement of small ribosomal domains associated with the tRNA binding during translocation. Our observation that ¯exible domains are associated primarily with three primary tRNA binding sites is consistent with the movable domain model, however, without ruling out the hybrid site model.

R.K. Agrawal et al. / The International Journal of Biochemistry & Cell Biology 31 (1999) 243±254

251

4. Materials and methods

Acknowledgements

Ribosomes and ribosomal subunits were isolated from MRE-600 strain of E. coli. Methods for the preparation and 3D reconstruction of 30S subunit have been described earlier [29]. The resolution of 3D cryo-EM map of 30S subunit in both non-activated and heat-activated 30S subunits were 37 AÊ. The 50S-bound 30S subunit map was obtained by applying a cutting plane on a 25 AÊ resolution map of 70S ribosome [11, 12] as already described by Lata et al. [29]. Positive and negative di€erence maps were calculated by subtracting 3D cry-EM map of one volume from the other. A negative di€erence mass appears wherever the mass of the volume being subtracted exceeds the mass of the other volume. Before calculating the di€erence maps, shown in Fig. 4, the 3D cryo-EM map of the 50S-bound 30S subunit was ®ltered to the resolution of 3D map of heatactivated 30S subunit, i.e. 37 AÊ. Preparation and 3D reconstruction of the ±70S ribosome complex is fMet-tRNAMet f described elsewhere [14]. The resolution of this complex was 15 AÊ. The preparation and 3D reconstructions of the naked control 70S and the at the E POST 70S complex (having tRNAMet f Phe at the P site) have site and the AcPhe-tRNA been described in detail in a separate article. These two volumes have a resolution of 25 AÊ and were used to produce di€erence maps presented in Fig. 5(b). Di€erence maps were calculated by subtracting the 3D map of fMet±70S ribosome complex form that of tRNAMet f the POST complex and by subtracting the 3D map of naked 70S control from that of fMet±70S ribosome complex. Before calcutRNAMet f lating di€erence maps, the 3D cryo-EM map of ±70S ribosome complex was the fMet-tRNAMet f ®ltered to the resolution of 25 AÊ. Three prominent di€erence peaks shown in Fig. 5(b) were isolated by applying spherical (48 AÊ radius) masks on the prominent density regions, other than those related to tRNAs, of the respective di€erence maps (Fig. 5, ®gure legend).

We thank Amy B. Heagle for help with illustrations. This work was supported by grants NIH R37 GM29169, R01 GM55440 and NSF BIR 9219043.

References [1] W.E. Hill, P. Moore, A. Dahlberg, R. Schlessinger, R. Garrett, J. Warner, The Ribosome: Structure, Function and Evolution, American Society for Microbiology, Washington, DC, 1990. [2] A. Matheson, P. Dennis, J. Davies, W. Hill, Frontiers in Translation, National Research Council Canada, Ottawa, Canada, 1995. [3] M.S. Capel, D.M. Engelman, B.R. Freeborn, M. Kjeldgaard, J.A. Langer, V. Ramakrishnan, D.G. Schindler, D.K. Schneider, B.P. Schoenborn, I.Y. Sillers et al, A complete mapping of the proteins in the small ribosomal subunit of Escherichia coli, Science 238 (1987) 1403±1406. [4] R.P. May, V. Nowotny, P. Nowotny, H. Voss, K.H. Nierhaus, Inter-protein distances within the large subunit from Escherichia coli ribosomes, EMBO J. 11 (1992) 373±378. [5] M. StoȂer-Meilicke, G. StoȂer, Topography of the ribosomal proteins from Escherichia coli within the intact subunits as determined by immunoelectron microscopy and protein±protein cross-linking, in: W.E. Hill, A.E. Dahlberg, R.A. Garrett, P.B. Moore, D. Schlessinger, J.R. Warner (Eds.), The Ribosome: Structure, Function, Evolution, American Society for Microbiology, Washington DC, 1990, pp. 123±133. [6] R. Brimacombe, J. Atmadja, W. Stiege, D. SchuÈler, A detailed model of the three-dimensional structure of Escherichia coli 16S ribosomal RNA in situ in the 30S subunit, J. Mol. Biol. 199 (1988) 115±136. [7] S. Stern, B. Weiser, H.F. Noller, Model for the threedimensional folding of 16S ribosomal RNA, J. Mol. Biol. 204 (1988) 447±481. [8] A. Malhotra, S.C. Harvey, A quantitative model of the Escherichia coli 16 S RNA in the 30S ribosomal subunit, J. Mol. Biol. 240 (1994) 308±340. [9] F. MuÈller, R. Brimacombe, A new model for the threedimensional folding of Escherichia coli 16S ribosomal RNA. I. Fitting the RNA to a 3D electron microscopic map at 20 AÊ, J. Mol. Biol. 271 (1997) 524±544. [10] A. Yonath, Approaching atomic resolution in crystallography of ribosomes, Annu. Rev. Biophys. Biomol. Struct. 21 (1992) 77±93. [11] J. Frank, J. Zhu, P. Penczek, Y. Li, S. Srivastava, A. Verschoor, M. Radermacher, R. Grassucci, R.K. Lata, R.K. Agrawal, A model of protein synthesis based on

252

[12]

[13]

[14]

[15]

[16]

[17]

[18]

[19] [20]

[21]

[22]

[23]

R.K. Agrawal et al. / The International Journal of Biochemistry & Cell Biology 31 (1999) 243±254 cryo-electron microscopy of the E. coli ribosome, Nature 376 (1995) 441±444. J. Frank, A. Verschoor, Y. Li, J. Zhu, R.K. Lata, M. Radermacher, P. Penczek, R. Grassucci, R.K. Agrawal, S. Srivastava, A model of the translational apparatus based on a three-dimensional reconstruction of the Escherichia coli ribosome, Biochem. Cell Biol. 73 (1995) 757±765. H. Stark, F. MuÈller, E.V. Orlova, M. Schatz, P. Dube, T. Erdemir, F. Zemlin, R. Brimacombe, M. van Heel, The 70S Escherichia coli ribosome at 23 AÊ resolution: ®tting the ribosomal RNA, Structure 3 (1995) 815±821. A. Malhotra, P. Penczek, R.K. Agrawal, I. Gabashvili, R.A. Grassucci, N. Burkhardt, R. JuÈnemann, K.H. Nierhaus, J. Frank, E. coli 70S ribosome at 15 AÊ resolution by cryo-electron microscopy: localization of fMettRNAMet and ®tting of L1 protein, J. Mol. Biol. 280 f (1998) 103±115. R.K. Agrawal, P. Penczek, R.A. Grassucci, Y. Li, A. Leith, K.H. Nierhaus, J. Frank, Direct visualization of A-, P-, and E-site transfer RNAs in the Escherichia coli ribosome, Science 271 (1996) 1000±1002. H. Stark, E.V. Orlova, A.J. Rinke, N. Junke, F. MuÈller, M. Rodnina, W. Wintermeyer, R. Brimacombe, M. van Heel, Arrangement of tRNAs in pre- and post-translocational ribosomes revealed by electron cryomicroscopy, Cell 88 (1997) 19±28. H. Stark, M.V. Rodnina, J. Rinke-Appel, R. Brimacombe, W. Wintermeyer, M. van Heel, Visualization of elongation factor Tu on the Escherichia coli ribosome, Nature 389 (1997) 403±406. R.K. Agrawal, P. Penczek, R.A. Grassucci, J. Frank, Visualization of elongation factor G on the Escherichia coli 70S ribosome: the mechanism of translocation, Proc. Natl. Acad. Sci. USA 95 (1998) 6134±6138. S. Pestka, Studies on the formation of transfer ribonucleic acid±ribosome complexes, J. Biol. Chem. 244 (1969) 1533±1539. L.P. Gavrilova, O.E. Kostiashkina, V.E. Koteliansky, N.M. Rutkevitch, A.S. Spirin, Factor-free (`non-enzymic') and factor-dependent systems of translation of polyuridylic acid by Escherichia coli ribosomes, J. Mol. Biol. 101 (1976) 537±552. O.P. van Diggelen, L. Bosch, The association of ribosomal subunits of Escherichia coli. 1. Two types of association products di€ering in their apparent sedimentation coecient, Eur. J. Biochem. 39 (1973) 499±510. O.P. van Diggelen, H. Oostrom, L. Bosch, The association of ribosomal subunits of Escherichia coli. 2. Two types of association products di€ering in sensitivity to hydrostatic pressure generated during centrifugation, Eur. J. Biochem. 39 (1973) 511±523. D.P. Burma, S. Srivastava, A.K. Srivastava, S. Mahanti, D. Dash, Conformational change of 50S ribosomes during protein synthesis, in: B. Hardesty, G. Kramer (Eds.), Structure Function and Genetics of Ribosomes, Springer-Verlag, New York, 1986, pp. 438±453.

[24] H.J. Rheinberger, K.H. Nierhaus, Allosteric interactions between the ribosomal transfer RNA-binding sites A and E, J. Biol. Chem. 261 (1986) 9133±9139. [25] J.R. Mesters, A.P. Potapov, J.M. de Graaf, B. Kraal, Synergism between the GTPase activities of EF-Tu
GTP and EF-G
GTP on empty ribosomes: elongation factors as stimulators of the ribosomal oscillation between two conformations, J. Mol. Biol. 242 (1994) 644±654. [26] R.K. Agrawal, D.P. Burma, Sites of ribosomal RNAs involved in the subunit association of tight and loose couple ribosomes, J. Biol. Chem. 271 (1996) 21285± 21291. [27] R.R. Traut, D. Dey, D.E. Bochkarlov, A.V. Oleinikov, G.G. Jokhadze, B. Hamman, D. Jameson, Location and domain structure of Escherichia coli ribosomal protein L7/L12: site-speci®c cysteine crosslinking and attachment of ¯uorescent probes, Biochem. Cell Biol. 73 (1995) 949± 958. [28] A. Zamir, R. Miskin, D. Elson, Inactivation and reactivation of ribosomal subunits: aminoacyl-transfer RNA binding activity of the 30S subunit of Escherichia coli, J. Mol. Biol. 60 (1971) 347±364. [29] K.R. Lata, R.K. Agrawal, P. Penczek, R. Grassucci, J. Zhu, J. Frank, Three-dimensional reconstruction of the Escherichia coli 30S ribosomal subunit in ice, J. Mol. Biol. 262 (1996) 43±52. [30] J. Frank, R.K. Agrawal, The movement of tRNA through the ribosome, Biophys. J. 74 (1998) 589±594. [31] R.A. Zimmerman, C.L. Thomas, J. Wower, Structure and function of rRNA in the decoding domain and the peptidyltransferase center, in: W.E. Hill, A. Dahlberg, R.A. Garrett, P.B. Moore, D. Schlessinger, R. Warner (Eds.), American Society for Microbiology, Washington DC, 1990, pp. 331±347. [32] I.N. Shatsky, A.G. Evsta®eva, T.F. Bystrova, A.A. Bogdanov, V.D. Vasiliev, Topography of RNA in the ribosome: location of the 3 0 -end of 5S RNA on the central protuberance of the 50S subunit, FEBS Lett. 121 (1980) 97±100. [33] A.G. Evsta®eva, I.N. Shatsky, A.A. Bogdanov, V.D. Vasiliev, Topography of RNA in the ribosome: location of the 5S RNA residues A39 and U40 on the central protuberance of the 50S subunit, FEBS Lett. 185 (1985) 57± 62. [34] D.A. Brow, H.F. Noller, Protection of ribosomal RNA from kethoxal in polyribosomes. Implication of speci®c sites in ribosome function, J. Mol. Biol. 163 (1983) 27± 46. [35] B.J. Van Stolk, H.F. Noller, Chemical probing of conformation in large RNA molecules. Analysis of 16S ribosomal RNA using diethylpyrocarbonate, J. Mol. Biol. 180 (1984) 151±177. [36] D. Moazed, B.J. Van Stolk, S. Douthwaite, H.F. Noller, Interconversion of active and inactive 30S ribosomal subunits is accompanied by a conformational change in the decoding region of 16S rRNA, J. Mol. Biol. 191 (1986) 483±493.

R.K. Agrawal et al. / The International Journal of Biochemistry & Cell Biology 31 (1999) 243±254 [37] N. Meier, R. Wagner, E€ects of the ribosomal subunit association on the chemical modi®cation of the 16S and 23S RNAs from Escherichia coli, Eur. J. Biochem. 146 (1985) 83±87. [38] F. Baudin, M. Mougel, P. Romby, F. Eyermann, J.P. Ebel, B. Ehresmann, C. Ehresmann, Probing the phosphates of the Escherichia coli ribosomal 16S RNA in its naked form, in the 30S subunit, and in the 70S ribosome, Biochemistry 28 (1989) 5847±5855. [39] G. Ericson, P. Wollenzien, An RNA secondary structure switch between the inactive and active conformations of the Escherichia coli 30S ribosomal subunit, J. Biol. Chem. 264 (1989) 540±545. [40] J.W. Weller, W.E. Hill, Probing dynamic changes in rRNA conformation in the 30S subunit of the Escherichia coli ribosome, Biochemistry 31 (1992) 2748± 2757. [41] I. Ginzburg, R. Miskin, A. Zamir, N-ethyl maleimide as a probe for the study of functional sites and conformations of 30S ribosomal subunits, J. Mol. Biol. 79 (1973) 481±494. [42] I. Ginzburg, A. Zamir, Characterization of di€erent conformational forms of 30S ribosomal subunits in isolated and associated states: possible correlations between structure and function, J. Mol. Biol. 93 (1975) 465±476. [43] J.B. Prince, B.H. Taylor, D.L. Thurlow, J. Ofengand, R.A. Zimmermann, Covalent crosslinking of tRNA1Val to 16S RNA at the ribosomal P site: identi®cation of crosslinked residues, Proc. Natl. Acad. Sci. USA 79 (1982) 5450±5454. [44] L.S. Lasater, L. Montesano-Roditis, P.A. Cann, D.G. Glitz, Localization of an oligodeoxynucleotide complementing 16S ribosomal RNA residues 520±531 on the small subunit of Escherichia coli ribosomes: electron microscopy of ribosome±cDNA±antibody complexes, Nucl. Acids Res. 18 (1990) 477±485. [45] M.I. Oakes, J.A. Lake, DNA-hybridization electron microscopy. Localization of ®ve regions of 16S rRNA on the surface of 30S ribosomal subunits, J. Mol. Biol. 211 (1990) 897±906. [46] M.I. Oakes, M.W. Clark, E. Henderson, J.A. Lake, DNA hybridization electron microscopy: ribosomal RNA nucleotides 1392±1407 are exposed in the cleft of the small subunit, Proc. Natl. Acad. Sci. USA 83 (1986) 275±279. [47] M.I. Oakes, L. Kahan, J.A. Lake, DNA-hybridization electron microscopy tertiary structure of 16S rRNA, J. Mol. Biol. 211 (1990) 907±918. [48] P. Gornicki, K. Nurse, W. Hellmann, M. Boublik, J. Ofengand, High resolution localization of the tRNA anticodon interaction site on the Escherichia coli 30S ribosomal subunit, J. Biol. Chem. 259 (1984) 10493± 10498. [49] J. Shine, L. Dalgarno, The 3 0 -terminal sequence of Escherichia coli 16S ribosomal RNA: complementarity to nonsense triplets and ribosome binding sites, Proc. Natl. Acad. Sci. USA 71 (1974) 1342±1346.

253

[50] D. Moazed, H.F. Noller, Binding of tRNA to the ribosomal A and P sites protects two distinct sets of nucleotides in 16S rRNA, J. Mol. Biol. 211 (1990) 135±145. [51] T. Powers, H.F. Noller, A functional pseudoknot in 16S ribosomal RNA, EMBO J. 10 (1991) 2203±2214. [52] T. Powers, H.F. Noller, The 530 loop of 16S rRNA: a signal to EF-Tu?. Trends Genet. 10 (1994) 27±31. [53] D.I. Van Ryk, A.E. Dahlberg, Structural changes in the 530 loop of Escherichia coli 16S rRNA in mutants with impaired translational ®delity, Nucl. Acids Res. 23 (1995) 3563±3570. [54] F. MuÈller, H. Stark, M. van Heel, J. Rinke-Appel, R. Brimacombe, A new model for the three-dimensional folding of Escherichia coli 16 S ribosomal RNA. III. The topography of the functional center, J. Mol. Biol. 271 (1997) 566±587. [55] K.W. Wilson, H.F. Noller, Mapping the position of EFG in the ribosome by directed hydroxyl radical probing, Cell 92 (1998) 131±139. [56] J.S. Lodmell, A.E. Dahlberg, A conformational switch in Escherichia coli 16S ribosomal RNA during decoding of messenger RNA, Science 277 (1997) 1262±1267. [57] D. Moazed, H.F. Noller, Transfer RNA shields speci®c nucleotides in 16S ribosomal RNA from attack by chemical probes, Cell 47 (1986) 985±994. [58] D. Dey, A.V. Oleinikov, R.R. Traut, The hinge region of Escherichia coli ribosomal protein L7/L12 is required for factor binding and GTP hydrolysis, Biochimie 77 (1995) 925±930. [59] A. Liljas, A.T. Gudkov, The structure and dynamics of ribosomal protein L12, Biochimie 69 (1987) 1043±1047. [60] W. MoÈller, Hypothesis: ribosomal protein L12 drives rotational movement of tRNA, in: W.E. Hill et al. (Ed.), The Ribosome: Structure, Function, and Evolution, American Society for Microbiology, Washington DC, 1990, pp. 380±389. [61] W. MoÈller, P.I. Schrier, J.A. Maassen, A. Zantema, E. Schop, H. Reinalda, A.F.M. Cremers, J.E. Mellema, Ribosomal protein L7/L12 of Escherichia coli: localization and possible molecular mechanism in translation, J. Mol. Biol. 163 (1983) 553±573. [62] R.R. Traut, D.S. Tewari, A. Sommer, G.R. Govino, H.M. Olson, D.G. Glitz, Protein topography of ribosomal functional domains: e€ects of monoclonal antibodies to di€erent epitopes in Escherichia coli protein L7/L12 on ribosome function and structure, in: B. Hardesty, G. Kramer (Eds.), Structure Function and Genetics of Ribosomes, Springer-Verlag, New York, 1986, pp. 286± 308. [63] C. Branlant, V. Korobko, J.P. Ebel, RNA sequences associated with proteins L1, L9 and L5, L18, L25 in ribonucleoprotein fragments isolated form 50S subunit of Escherichia coli ribosomes, Eur. J. Biochem. 70 (1976) 471±482. [64] R. Brimacombe, J. Atmadja, A. Kyriatsousis, W. Stiedge, RNA structure and RNA-protein neighborhoods in the ribosome, in: B. Hardesty, G. Kramer (Eds.),

254

[65]

[66] [67]

[68]

R.K. Agrawal et al. / The International Journal of Biochemistry & Cell Biology 31 (1999) 243±254 Structure Function and Genetics of Ribosomes, Springer-Verlag, New York, 1986, pp. 184±202. J.S. Lodmell, W. Tapprich, W.E. Hill, Evidence for a conformational change in the exit site of the Escherichia coli ribosome upon tRNA binding, Biochemistry 32 (1993) 4067±4072. D. Moazed, H.F. Noller, Interaction of tRNA with 23S rRNA in the ribosomal A, P, and E sites, Cell 57 (1989) 585±597. J.M. Bullard, M.A. van Waes, D.J. Bucklin, W.E. Hill, Regions of 23S ribosomal RNA proximal to transfer RNA bound at the P and E sites, J. Mol. Biol. 163 (1995) 27±46. J. Rinke-Appel, N. JuÈnke, M. Osswald, R. Brimacombe, The ribosomal environment of tRNA: crosslinks to rRNA from positions 8 and 20:1 in the central fold of tRNA located at the A, P, or E site, RNA 1 (1995) 1018±1028.

[69] J. Wower, P. Sche€er, L.A. Sylvers, W. Wintermeyer, R.A. Zimmermann, Topography of the E site on the E. coli ribosome, EMBO J. 12 (1993) 617±623. [70] A.S. Spirin, Ribosomal translocation: facts and models, Prog. Nucl. Acid Res. Mol. Biol. 32 (1985) 75±114. [71] K.H. Nierhaus, Solution of the ribosome riddle: how the ribosome selects the correct aminoacyl-tRNA out of 41 similar contestants, Mol. Microbiol. 9 (1993) 661±669. [72] D. Moazed, H.F. Noller, Intermediate states in the movement of transfer RNA in the ribosome, Nature 342 (1989) 142±148. [73] K.H. Nierhaus, D. Beyer, M. Dabrowski, M.A. Schafer, C.M. Spahn, J. Wadzack, J.U. Bittner, N. Burkhardt, G. Diedrich, R. JuÈnemann et al, The elongating ribosome: structural and functional aspects, Biochem. Cell Biol. 73 (1995) 1011±1021.