doi:10.1016/j.jmb.2010.04.009 J. Mol. Biol. (2010) 399, 145–153

Available online at www.sciencedirect.com

The Ribosomal Stalk Plays a Key Role in IF2-Mediated Association of the Ribosomal Subunits

Chenhui Huang, Chandra Sekhar Mandava and Suparna Sanyal⁎

Department of Cell and Ribosomal “stalk” protein L12 is known to activate translational Molecular Biology, Uppsala EF-G and EF-Tu, but not much is known about its role in relation to other University, BMC, Box -596, two translational G factors, IF2 and RF3. Here, we have clarified the role S-75 124 Uppsala, Sweden of L12 in IF2-mediated initiation of bacterial protein synthesis. With fast kinetics measurements, we have compared L12-depleted 50S subunits Received 12 March 2010; with the native ones in subunit association, GTP hydrolysis, P (inorganic received in revised form i phosphate) release and IF2 release assays. L12 depletion from 50S subunit 5 April 2010; slows the subunit association step significantly (∼40 fold) only when accepted 6 April 2010 IF2·GTP is present on the 30S preinitiation complex. This demonstrates Available online that rapid subunit association depends on a specific interaction between 10 April 2010 the L12 stalk on the 50S subunit and IF2·GTP on the 30S subunit. L12 depletion, however, did not affect the individual rates of the subsequent steps including GTP hydrolysis on IF2 and Pi release. Thus, L12 is not a GTPase activating protein (GAP) for IF2 unlike as suggested for EF-G and EF-Tu. © 2010 Elsevier Ltd. All rights reserved. Edited by D. E. Draper Keywords: ribosome; translation initiation; L12; IF2; subunit association

Introduction guanosine triphosphatases (GTPases).5 The third domain is a flexible hinge that connects the NTD The ribosomal stalk is a structural landmark on with the CTD and can attain different conforma- tions, from a compact helix to an extended random the bacterial ribosome, as it forms a finger-like 6 extension from the 50S subunit.1,2 It is highly flexible coil. The flexibility of the hinge region allows for different locations of the CTDs in relation to the and composed of L12 protein dimers bound to the 7 protein L10, varying from two to three in different NTDs of L12 on the ribosome. It was proposed bacterial species. In Escherichia coli, about half of the that the L12 monomers are arranged in an L12 proteins are present in N-terminally acetylated antiparallel fashion in the dimer, with one hinge form, historically known as L7. Therefore, the in compact and the other in extended conformation, and that they often switch between these two ribosomal stalk is also known as the L7/L12 stalk. 1 For simplicity, we will refer to L7/L12 hereinafter as conformations in alternative phase. Rather recent- ly, direct evidence supporting this hypothesis L12 only. 8 The L12 protein has three main domains. The N- became available. Although proposed as involved “ ”2 terminal domain (NTD), with two short α-helices, is in factor recruitment, the functional relevance of responsible for the formation of the L12 dimer and such dynamic behavior of the L12 proteins is not for its binding to L10.3,4 The C-terminal domain yet understood. (CTD), with a well-conserved, globular structure, is In bacteria, four GTPase factors modulate the involved in the interaction with the translational main steps in translation. IF2 catalyzes the initia- tion step leading to a fast assembly of 70S initiation complex (70S IC); EF-Tu loads the aminoacyl *Corresponding author. E-mail address: tRNAs in the A site of the ribosome; EF-G [email protected]. translocates peptidyl tRNAs from the A site to Abbreviations used: NTD, N-terminal domain; CTD, the P site; and RF3 recycles class I release factors C-terminal domain; 70S IC, 70S initiation complex; 30S (RF1 and 2) during the termination process. preIC, 30S preinitiation complex; GTPase, guanosine Among these, EF-Tu and -G have been shown to triphosphatase; GAP, GTPase-activating protein; GTP, be activated by L12 proteins for guanosine 9,10 guanosine triphosphate; GDP, guanosine diphosphate; Pi, triphosphate (GTP) hydrolysis, and such activa- inorganic phosphate; PBP, phosphate-binding protein. tion is hindered by mutations in the factor

0022-2836/$ - see front matter © 2010 Elsevier Ltd. All rights reserved. 146 L12–IF2 Interaction Triggers Subunit Association interaction site of L12 CTD.11,12 However, nothing Results is known so far about the role of L12 in IF2 and RF3 functions. We have recently mapped the inter- action sites of these four GTPases on L12 using L12 depletion from 50S subunit and its effect on heteronuclear magnetic resonance and found that subunit association all of these factors interact essentially with the same site on the conserved CTD of the L12 70S ribosomes and 50S subunits lacking (85–90%) 5 – protein. This common interaction site may signify L12 proteins were prepared by NH4Cl ethanol 10,22 a general role of L12 in the recruitment of the treatment. In addition, reconstitution of L12 translation factors or in their GTPase activation. proteins on L12-depleted 70S ribosomes and 50S Experimental verification of these hypotheses subunits was achieved by incubating those with remains to be done. excess L12 at 37 °C for 15 min. Association of the 30S In the present work, we have clarified the role and 50S subunits was monitored by Rayleigh light of the L12 proteins in IF2-mediated initiation of scattering after rapid mixing in a stopped-flow 23 bacterial protein synthesis. IF2 plays a pivotal role instrument. in the association of the ribosomal subunits.13,14 The depletion of L12 from the 50S subunit did not First, IF2 with GTP binds to the 30S subunit affect the association of the naked subunits, (i.e., together with mRNA, initiator tRNA, and initia- without mRNA, tRNA and initiation factors) tion factors IF1 and IF3 forming the 30S preinitia- (Fig. 1a). The addition of MLIstop mRNA, fMet- tion complex (30S preIC). Next, the 50S subunit tRNAfMet and initiation factors IF1 and IF3 to the associates with it, forming 70S IC, and the 30S subunit, one at a time or together, influenced the initiation factors leave the complex following rate of subunit association in general, but showed – GTP hydrolysis on IF2.15 17 The kinetic steps of virtually no difference between the L12-depleted the process have been explored in detail in several and native 50S subunits (Fig. 1b). However, when – – recent articles.13 16,18 20 There is some controversy IF2·GTP was present on 30S preIC, the fast rate of μ − 1 − 1 in the literature regarding the GTP hydrolysis step. subunit association, ka =130±10 M s for native One model suggests that GTP hydrolysis triggers 50S subunits, decreased about 40-fold to 3.5± − 1 − 1 the release of IF2 from the 70S initiation complex13 0.5 μM s by L12 depletion, which could be − 1 − 1 or leads to the elongation step.19,21 The other view restored to 110±10 μM s only by reconstitution is that GTP hydrolysis is primarily required for of L12 to the 50S subunit core (Fig. 1c). Similar stable association of the ribosomal subunits.14 This reduction in the rate of subunit association was also rather recent model needs to be tested with other seen when IF2 was excluded from the 30S preIC μ − 1 − 1 approaches. (ka =2.8±0.5 M s )(Fig. 1c) or when IF2 was present but without any guanine nucleotide (k =4± To study the role of the L12 proteins in IF2- − − a 0.8 μM 1 s 1) or with guanosine diphosphate (GDP) mediated initiation, we have chemically depleted − − 10 μ 1 1 L12 proteins from the 50S subunits and com- (ka =3±0.2 M s )(Table 1). Interestingly, when pared the L12-depleted 50S subunits with the L12 was missing on the 50S subunit the rate of subunit ∼ μ −1 −1 normal ones in different steps of initiation by association was invariably slow (ka 3 M s ), conducting fast-kinetics measurements using irrespective of the nucleotide bound on IF2 (Table 1). quench-flow and stopped-flow techniques. We These data suggest that a specific interaction have, in particular, measured the rate constants of between L12 on the 50S subunit and IF2·GTP on subunit association, GTP hydrolysis on IF2, inor- the 30S preIC is essential for rapid subunit associ- ganic phosphate (Pi) release, and the release of IF2 ation in the initiation of bacterial protein synthesis. from the 70S IC with these two types of 50S Addition of IF3 to 30S preIC made the IF2-mediated μ − 1 − 1 subunits. As a control, we have used 50S subunits association much slower (ka =10±1 M s ), but reconstituted by addition of the L12 proteins on the the relative reduction of the association rate (40–50 originally L12-depleted ones. Our work has times) upon L12 depletion remained virtually revealed that rapid subunit association depends unchanged (Fig. 1d). on a specific interaction between the L12 stalk on the 50S subunit and IF2·GTP on 30S preIC. When The effect of L12 depletion on IF2-GTPase L12 was depleted from the 50S subunit, the activation by naked 70S association of the subunits became significantly slower and, consequently, subsequent steps such as We have followed GTP hydrolysis and Pi release GTP hydrolysis, Pi release and IF2 release also by IF2 in the presence of free L12 proteins as well as appeared to be slower. However, when individual L12-containing and -lacking naked 70S ribosome. rate constants for these subsequent steps were GTP hydrolysis was measured from the relative determined, no significant difference was seen amounts of [3H]GDP and [3H]GTP separated in a between those with the native 50S subunit and mono-Q column. Pi release was monitored by the the L12-depleted ones. Thus, in summary, we increase in fluorescence of an MDCC-labeled phos- conclude that L12 is not a GAP (GTPase-activating phate-binding protein (PBP).24 protein) for IF2 but plays a key role in IF2- The intrinsic GTPase activity of IF2 was very low b − 1 mediated subunit association during initiation of (kobs 0.01 s ) and remained unaltered upon protein synthesis. addition of L12 protein alone, even in huge excess L12–IF2 Interaction Triggers Subunit Association 147

Fig. 1. Subunit association with different 50S subunits. The time course of association of normal (black trace), L12- depleted (red trace) and L12 reconstituted (green trace) 50S subunit (0.5 μM), with (a) naked 30S subunit (0.5 μM), (b) 30S preIC containing IF1 and IF3, but not IF2, (c) 30S preIC with IF2 but without IF3, and (d) 30S preIC with IF1, IF2 and IF3 was followed by the increase in light scattering at 430 nm. The blue traces in (c) and (d) reflect association of the subunits in the absence of IF2 in otherwise identical conditions. The rates were calculated by fitting the curves (continuous lines) as second-order 30S+50S subunits binding reaction following the procedure described by Antoun et al.23

(Fig. 2a and b). Naked 70S particles stimulated the L12 depletion and its effect on binding of IF2 to GTPase activity of IF2 moderately, resulting in k naked 70S ribosome − obs of 0.43±0.02 s 1 for GTP hydrolysis and 0.41± − 1 0.02 s for Pi release in multiple turnover condition. The defect seen in the GTPase activation of IF2 for L12-depleted 70S was strongly impaired in promot- L12 depletion could arise from its inefficient binding ing IF2-mediated GTP hydrolysis and Pi release, but to the L12-depleted variants of 70S. To check this its stimulatory activity was regained by the recon- possibility, we followed IF2 binding to naked 70S stitution with L12 proteins (Fig. 2a and b). No burst ribosome using tetramethyl rhodamine-5-maleimide of GTP hydrolysis or Pi release was seen (Fig. 2a, (rhodamine)-labeled IF2 (IF2-rho) (Fig. 2c). IF2-rho inset) under our experimental condition. was equally active as the unlabeled IF2 in subunit association (Fig. 2c, inset). The binding of IF2-rho to 70S ribosome showed a biphasic increase in its 14 Table 1. The rate constant of association (ka) of native or fluorescence at 590 nm as seen in an earlier report. L12-depleted 50S subunit with the 30S preIC without or The first phase reflecting IF2-rho binding was with IF2, in combination with different guanine compared with L12-proficient and -deficient 70S nucleotides ribosome. As shown in Fig. 2c, the rate of binding ∼ μ −1 −1 of IF2 to L12-depleted 70S was much slower ( 10 50S subunit 30S preinitiation complex ka ( M s ) −1 times) than to the native 70S (kobs around 43±5 s ) Native (+L12) +IF2·GTP 130±10 and was further regained by reconstitution of L12 −IF2 2.5±1 +IF2 (no GXP) 4±0.8 proteins on the 70S core (Fig. 2c). Since IF2 binding to +IF2·GDP 3±0.2 70S ribosome must take place prior to its GTPase L12 depleted +IF2·GTP 3.5±0.5 activation by the ribosome, our result suggests that −IF2 3±0.5 L12 depletion primarily slows IF2 binding to 70S +IF2 (no GXP) 2.5±0.5 +IF2·GDP 2.5±0.5 ribosome, as a consequence of which 70S-assisted GTPase activation on IF2 appears to be inhibited. 148 L12–IF2 Interaction Triggers Subunit Association

Fig. 2. The effect of L12 depletion on GTPase activity, binding and release of IF2. The GTPase activity of IF2 was studied by GTP hydrolysis (a and d) and Pi release (b and e) assays, respectively. (a and b) The translation uncoupled GTPase activity of IF2 (0.2 μM) is measured without any modulator (light blue trace) or with L12 protein (10 μM, dark blue trace), naked 70S (1 μM, black trace), L12-depleted 70S (1 μM, red trace) and L12 reconstituted 70S (1 μM, green trace). In the inset (a), early time points of 70S modulated GTP hydrolysis by IF2 (1 μM) are shown. (d and e) The GTPase activity of IF2 (2 μM) in the context of translation initiation was measured, where 30S preIC (0.5 μM) (see Materials and Methods, for composition of mix A) was rapidly mixed with 0.5 μM native 50S (black trace), L12-depleted 50S (red trace, also insets in d and e), and L12-reconstituted 50S (green trace) subunits, respectively. (c) The binding of IF2 to naked 70S λ ribosome was followed by monitoring the fluorescence of IF2-rho at 590 nm ( Ex =560 nm). For this purpose, IF2-rho (0.4 μM) was rapidly mixed with 1 μM naked 70S (black trace), L12-depleted 70S (red trace) and L12-reconstituted 70S (green trace) ribosome in a stopped-flow instrument. The inset shows the time course of subunit association with IF2-rho (black trace) or unlabeled IF2 (green trace). (f) The binding and release of IF2 during initiation was followed with IF2-rho in an experimental setup similar to that in Fig. 1b. (d, e and f) Data were fitted with a single-exponential equation (continuous lines) using Origin 8 (Origin Lab Co). L12–IF2 Interaction Triggers Subunit Association 149

Effect of L12 depletion on GTP hydrolysis and Pi aswell,therealrateforGTPhydrolysiswas release in translation initiation estimated after subtraction of the average time spent in the association step from the average time The GTPase activity of IF2 in initiation of protein spent until GTP hydrolysis (Fig. 3). For Pi release, the synthesis was studied in the experimental setup average time spent for GTP hydrolysis was also essentially identical to the subunit association assay, taken into account. Surprisingly, these estimated where 50S subunits were rapidly mixed with 30S rates (kest) were found only marginally reduced (two preIC containing MLIstop mRNA, fMet-tRNAfMet to three times) for L12-lacking 50S subunits com- and IF1 and IF2 in quench-flow (for GTP hydrolysis) pared to the L12-containing ones (Fig. 3), thus or stopped-flow (for Pi release) instruments. In fact, suggesting that L12 is not directly involved in the Pi release was monitored simultaneously with the GTPase activation on IF2. subunit association using two cutoff filters (475 nm long pass and 395±25 nm band pass) on two parallel The effect of L12 depletion on IF2 release from detectors in the stopped-flow instrument. When 70S IC native 50S subunit was used, a strong stimulation of IF2 GTPase, significantly higher than that with After the formation of 70S IC, IF2 is released and naked 70S ribosome, was obtained with rate the ribosome is prepared for peptide synthesis.13 To − 1 constant kobs =25±4 and 16±2 s for GTP hydro- check whether the depletion of L12 from the 50S lysis and Pi release, respectively (Fig. 2d and e). L12- subunit has any effect on this step, the release of IF2 depleted 50S subunits, on the other hand, showed was followed with fluorescent IF2-rho, which ∼ − 1 kobs 0.9 s in both assays. Re-addition of L12 showed a decrease in fluorescence at 590 nm, proteins on the 50S core restored the rate of GTP indicating IF2 release from the ribosomal complex. hydrolysis and P release close to the native values This experiment was done in the same setup as that i − (23±2 and 15±1 s 1, respectively) (Fig 2d and e). of the subunit association, with 30S preIC in one From these observed rates it seemed that L12 syringe and 50S subunit in the other. IF2-rho release depletion caused a decrease in the stimulation of and subunit association were monitored simulta- GTPase activity of IF2 in initiation. However, since neously with parallel detectors equipped with these reactions involved a subunit association step suitable cutoff filters (590 nm long pass and 395±

Fig. 3. Kinetic scheme of the steps in translation initiation. Kinetic scheme of the steps in initiation for native and L12- − depleted 50S subunits. kobs is the observed rate. taverage (average time spent in a step)=1/kobs (this step) 1/kobs (previous step) ′ ′ ′ and estimated rate kest =1/taverage. Symbols used for L12-depleted 50S subunit are k obs, k est and t average, respectively; and shown in red in the cartoon. 150 L12–IF2 Interaction Triggers Subunit Association

25 nm band pass) in the stopped-flow instrument. initiator tRNA, IF1 and IF3, and the performance of The apparent rate of IF2 release, k , decreased from these components are independent of the presence − obs − 17±3 s 1 for native 50S subunits to 0.86±0.14 s 1 by of the L12 proteins on the ribosome. L12 depletion, and L12 reconstitution restored the − 1 kobs value to 16±2 s (Fig. 2f). However, when the L12 proteins interact with IF2·GTP for rapid actual rate for IF2 release was estimated after subunit association subtraction of average time spent in the steps before it, namely, subunit association and GTP hydrolysis, In the light-scattering experiments, the most the rates of IF2 release for native and L12-depleted striking result was that the depletion of L12 from 50S subunits were found not very different from the 50S subunit significantly hampered subunit each other. In fact, the reduction in the rate of IF2 association only when IF2 was present on 30S release (three to fourfold) due to L12 depletion was preIC (Fig. 1c). The reverse situation, that is, IF2 quite similar to that seen in GTP hydrolysis and Pi lacking from 30S preIC when L12 was present on release, and insignificant compared to the reduction 50S subunit, also had a similar effect (Fig. 1c and seen in the rate of subunit association (∼40 fold) Table 1). These observations led us to the proposi- (Fig. 3). tion that L12 and IF2 act as recognition signals on the 50S subunit and 30S preIC, respectively, and a direct interaction between these two are needed for Discussion fast subunit association. It is interesting to note that when L12 was With the dramatic development in the ribosome removed from the 50S subunit, the rate of subunit field in past 10 years, the mechanism of the association became indifferent to the nucleotide individual steps of bacterial protein synthesis has bound to IF2 in 30S preIC (Table 1). However, in become increasingly lucid with the roles of the the presence of L12, only IF2·GTP (or GTP analo- translation factors understood with molecular gues) promoted fast subunit association (Table 1). It details. One of the remaining questions in this field has been reported that IF2·GTP and IF2·GDP is the function of the flexible ribosomal stalk, conformations are fairly different from each particularly in relation to the GTPase factors. Here, other.27 Thus, possibly, L12 scrutinizes the nucleo- we have deciphered the role of the L12 proteins in tide-bound state of IF2 in 30S preIC and selects only IF2 function. The conclusions drawn from our the ones with IF2 in the GTP state for stable experiments are discussed below. association with the 50S subunit.

L12 proteins are not essential for association of L12 is not a GAP for IF2 GTPase the naked subunits The stalk protein L12 is known to be involved in – Stable assembly of the 70S ribosome needs the the GTPase activation of EF-G and EF-Tu.9 12 These formation of the intersubunit bridges in a precise observations were often extrapolated to other manner.25,26 It is not surprising that the stalk protein translational GTPases such as IF2 and RF3, although L12, due to its flexible nature, is not a part of these direct evidence is still missing. Our data show that bridges. This observation, drawn from structural L12 protein by itself does not have any stimulatory studies, becomes evident from our fast kinetics action on the GTPase activity of IF2, but its presence measurements. The rate of association of the naked on naked 70S is essential for ribosome-mediated 30S and 50S subunits remained unaffected by the GTPase activation of this factor (Fig. 2a and b). A presence or absence of the L12 proteins on the 50S similar observation was reported earlier with EF- subunit (Fig. 1a). This result leads to the conclusion Tu,10 although the mechanism of GTPase stimula- that L12 proteins do not constitute a structural tion on these two factors might be different. Our component of the 50S subunit that is essential for its data show that 70S ribosomes lacking L12 proteins association with the 30S subunit. are primarily defective in IF2 binding, which, being the rate-limiting step, affects the following steps L12 proteins do not have direct interaction with (GTP hydrolysis and Pi release) accordingly. mRNA, initiator tRNA, IF1 and IF3 Surprisingly, a similar mechanism was also seen for the enhancement of the GTPase activity of IF2 in During initiation of protein synthesis, 30S subunit the context of translation initiation, although in that first binds to the mRNA on its Shine–Dalgarno case IF2 first mounts on the 30S preIC and then 50S sequence, and then fMet-tRNAfMet binds to its P site subunit binds to it. The L12-depleted 50S subunits with the help of the initiation factors, thereby are highly defective in binding to IF2 on the 30S forming 30S preIC. When subunit association was preIC; consequently, subunit association slows followed by the addition of the 30S preIC containing significantly. However, the effective rates estimated all factors and components other than IF2, one at a for different steps in initiation (summarized in Fig. time or in different combinations, no defect was seen 3) show that L12 depletion from the 50S subunit has for the absence of L12 from the 50S subunit (Fig. 1b). only a small effect on the individual rates of GTP Thus we conclude that the L12 proteins do not have hydrolysis by IF2 and subsequent Pi release. Thus any functional or structural interaction with mRNA, L12 protein, in solution or on the ribosome, is not a L12–IF2 Interaction Triggers Subunit Association 151

GAP for IF2 GTPase. The molecular mechanism for translational GTPase IF2. The importance of activation of translational GTPases is somewhat IF2·GTP in subunit association has been reported – unclear. One possibility is that the sarcin–ricin loop in many earlier studies.14 18 However, nothing was (SRL) of 23S rRNA is involved in this process. known about the complementary element on the 50S We would like to add a note explaining why we subunit that would recognize the correct state of IF2 used an indirect method for determination of the rates on 30S preIC for association. Our data identify the of the individual steps in initiation from single-point stalk protein (L12) to be the key element on the 50S measurements with respect to the 50S subunit subunit for this job. Since IF2 adheres tightly to the concentration (Fig. 3), instead of measuring those 30S preIC in the presence of GTP,13 the recognition directly with increasing concentration of the 50S and interaction between L12 and IF2 brings the 50S subunit. Although 50S concentration dependency of subunit and 30S preIC close to each other, which in the individual rates would be most natural to measure, turn triggers rapid subunit association (Fig. 4). An this is technically impossible to execute with L12- indirect support for this proposition can be obtained depleted 50S subunits. This is because the chemical from the cryoelectron microscopy reconstruction of depletion method usually leaves 10–15% L12 mole- the 70S IC containing IF2·GDPNP,28 where density cules on the 50S subunit. Thus, if L12-depleted 50S for the CTD of L12 was seen close to the G domain of subunits are added in higher excess, the population of IF2 in the subunit interface. Detailed mutational the L12-containing 50S subunits would also increase, analysis of L12 CTD and IF2 will be needed to which may in turn mask the effect of L12 depletion. To understand the molecular mechanism of the inter- bypass that problem, we measured and compared the action between L12 and IF2, which is the key for fast rates of the initiation steps with a fixed concentration formation of 70S IC during initiation. of L12-containing and -lacking 50S subunits.

L12 does not influence the release of IF2·GDP Materials and Methods from the 70S IC

Since the presence of the L12 protein was found to Chemicals and buffers be crucial on the 50S subunit for its binding to IF2 in All experiments were carried out in Hepes–polymix solution or on the 30S preIC, it was important to 29 check whether the release of IF2 from the ribosome buffer (pH 7.5). was also dependent on the presence of L12. As shown in Fig. 3, the rate of release of IF2·GDP from Components of the translation system 70S IC was not changed much by the removal of L12 from the 50S subunit. Most probably, L12 interacts All components are from E. coli. Ribosomal particles fMet with IF2 only in the GTP form and thus allows the (70S, 50S and 30S), fMet-tRNA , mRNA fMet-Leu- release of IF2·GDP from 70S IC without affecting it. Ile-stop (gggaauucgggcccuuguu aacaauuaaggagguauacu- AUGCUGAUCUAAuugcagaaaaaaaaaaaaaaaaaaaaa; the – L12–IF2 recognition plays key role for rapid Shine Dalgarno sequence is underlined and coding sequence is in capital letters) and his-tagged initiation subunit association factors were purified as described by Antoun et al.23 Mutations V400G and H448E were introduced in IF2 Our data, for the first time, present clear biochem- (cloned in pET21b) with Quik-Change mutagenesis (Qia- ical evidence for the role of the ribosomal stalk gen). Mutant IF2s were purified similarly as for other protein L12 in the recognition and recruitment of a initiation factors. To study the binding and release of IF2,

Fig. 4. L12–IF2 interaction triggers subunit association during initiation. A hypothetical cartoon illustrating how L12– IF2 interaction (left) leads to the association of the 50S subunit with the 30S preIC (right). The positioning of the NTDs of the L12 protein on the 50S subunit as well as the interaction point between one L12 CTD and IF2 are drawn following the cryoelectron microscopy reconstruction of 70S IC by Allen and Frank.30 L12 dimers are drawn on the 50S subunit as proposed by Diaconu et al.2 152 L12–IF2 Interaction Triggers Subunit Association it was labeled with IF2-rho according to the Invitrogen Acknowledgements manual and was purified in a G25 gel-filtration column. The labeling incurred no loss in IF2 activity. The research grants from Göran Gustafsson Stiftelse, Carl Tryggers Stiftelse, and Swedish L12 depletion and reconstitution Research Council (VR: NT and Linnestöd) to S.S. are thankfully acknowledged. The authors have no – L12 was depleted from 70S by NH4Cl ethanol treatment conflict of interest. on ice and reincorporated by incubating it 10 times in excess at 37 °C for 15 min according to the method of Mohr et al.10 L12-depleted and -reconstituted 70S were further purified by ultracentrifugation on 30% sucrose cushion. References 1. Chandra Sanyal, S. & Liljas, A. (2000). The end of the Components for fast kinetic measurements beginning: structural studies of ribosomal proteins. Curr. Opin. Struct. Biol. 10, 633–636. All kinetic measurements were performed at 37 °C by 2. Diaconu, M., Kothe, U., Schlunzen, F., Fischer, N., rapidly mixing two mixes with the following composi- Harms, J. M., Tonevitsky, A. G. et al. (2005). Structural tions unless otherwise mentioned: mix A, either 0.5 μM basis for the function of the ribosomal L7/12 stalk in μ factor binding and GTPase activation. Cell, 121, 30S subunit alone or in different combinations with 2 M – IF1, 2 μM IF2, 1 μM IF3, 2 μM mRNA(MLI), 2 μM fMet- 991 1004. tRNAfMet, 100 μM GTP (or other GXP); mix B, 0.5 μM 50S 3. Mulder, F. A., Bouakaz, L., Lundell, A., Venkataramana, subunit. M., Liljas, A., Akke, M. & Sanyal, S. (2004). Confor- mation and dynamics of ribosomal stalk protein L12 in solution and on the ribosome. Biochemistry, 43, Subunit association 5930–5936. 4. Bocharov, E. V., Gudkov, A. T., Budovskaya, E. V. & The association of ribosomal subunits was followed by Arseniev, A. S. (1998). Conformational independence of an increase in Rayleigh light scattering (430 nm) in a N- and C-domains in ribosomal protein L7/L12 and in stopped-flow instrument (Applied Photophysics). The the complex with protein L10. FEBS Lett. 423,347–350. data were fitted as a second-order 30S+50S subunit 5. Helgstrand, M., Mandava, C. S., Mulder, F. A., Liljas, binding reaction (continuous lines in Fig. 1) as elaborated A., Sanyal, S. & Akke, M. (2007). The ribosomal stalk by Antoun et al. using Origin8 (Origin Lab Co).23 binds to translation factors IF2, EF-Tu, EF-G and RF3 via a conserved region of the L12 C-terminal domain. J. Mol. Biol. 365, 468–479. GTP hydrolysis 6. Wahl, M. C., Bourenkov, G. P., Bartunik, H. D. & Huber, R. (2000). Flexibility, conformational diversity μ For translation uncoupled GTP hydrolysis, 0.4 M IF2 and two dimerization modes in complexes of ribo- μ 3 was mixed with 100 M[H]GTP in the absence or somal protein L12. EMBO J. 19, 174–186. μ presence of 2 M 70S. For GTPase activity of IF2 in 7. Traut, R. R., Dey, D., Bochkariov, D. E., Oleinikov, translation initiation, an otherwise similar mix A contain- A. V., Jokhadze, G. G., Hamman, B. & Jameson, D. μ 3 ing 20 M[H]GTP was mixed with B in a quench-flow (1995). Location and domain structure of Escherichia instrument. The reactions were quenched with 50% ice- coli ribosomal protein L7/L12: site specific cysteine cold formic acid and centrifuged for 10 min at 14,000 rpm. crosslinking and attachment of fluorescent probes. The supernatant was loaded in a 1 ml mono-Q column Biochem. Cell Biol. 73, 949–958. 3 3 attached to FPLC; [ H]GDP and [ H]GTP were eluted 8. Moens, P. D., Wahl, M. C. & Jameson, D. M. (2005). – with 0 200 mM NaCl gradient. GTP hydrolysis curves Oligomeric state and mode of self-association of were fitted with a single-exponential equation using Thermotoga maritima ribosomal stalk protein L12 in Origin8 (Origin Lab Co). solution. Biochemistry, 44, 3298–3305. 9. Savelsbergh, A., Mohr, D., Wilden, B., Wintermeyer,

Pi release W. & Rodnina, M. V. (2000). Stimulation of GTPase activity of translation G by ribosom- al protein L7/L12. J. Biol. Chem. 275, 890–894. Pi release was monitored by gradual increase in fluorescence of MDCC-labeled PBP (a kind gift from 10. Mohr, D., Wintermeyer, W. & Rodnina, M. V. (2002). λ GTPase activation of elongation factors Tu and G on Prof. Martin Webb, London) at 464 nm ( Ex =425 nm) after 41 – fast mixing of A and B in a stopped-flow instrument at the ribosome. Biochemistry, , 12520 12528. 37 °C.24 To remove preaccumulated phosphates, a 11. Kothe, U., Wieden, H. J., Mohr, D. & Rodnina, M. V. phosphate mop containing purine nucleoside phosphor- (2004). Interaction of helix D of elongation factor Tu ylase (0.1 U/mL) and 500 μM u7-methylguanosine was with helices 4 and 5 of protein L7/12 on the ribosome. J. Mol. Biol. 336, 1011–1021. added. Pi-release data were fitted with a single-exponen- tial equation using Origin8 (Origin Lab Co). 12. Savelsbergh, A., Mohr, D., Kothe, U., Wintermeyer, W. & Rodnina, M. V. (2005). Control of phosphate release from elongation factor G by ribosomal protein IF2 binding and release L7/12. EMBO J. 24, 4316–4323. 13. Antoun, A., Pavlov, M. Y., Andersson, K., Tenson, T. IF2-rho was used in a subunit association assay and its & Ehrenberg, M. (2003). The roles of initiation factor 2 binding and release was followed by the increase or and guanosine triphosphate in initiation of protein decrease in rhodamine fluorescence at 590 nm synthesis. EMBO J. 22, 5593–5601. λ μ ( Ex =560 nm). For IF2 binding to naked 70S, 0.4 M IF2- 14. Grigoriadou, C., Marzi, S., Kirillov, S., Gualerzi, C. O. rho was rapidly mixed with 2 μM 70S. & Cooperman, B. S. (2007). A quantitative kinetic L12–IF2 Interaction Triggers Subunit Association 153

scheme for 70 S translation initiation complex forma- with T and G factors and with guanosine triphos- tion. J. Mol. Biol. 373, 562–572. phate. J. Biol. Chem. 247, 805–814. 15. Antoun, A., Pavlov, M. Y., Lovmar, M. & Ehrenberg, 23. Antoun, A., Pavlov, M. Y., Tenson, T. & Ehrenberg, M. (2006). How initiation factors tune the rate of M. M. (2004). Ribosome formation from subunits initiation of protein synthesis in bacteria. EMBO J. 25, studied by stopped-flow and Rayleigh light scatter- 2539–2550. ing. Biol. Proced. Online, 6,35–54. 16. Grigoriadou, C., Marzi, S., Pan, D., Gualerzi, C. O. & 24. Brune, M., Hunter, J. L., Corrie, J. E. & Webb, M. R. Cooperman, B. S. (2007). The translational fidelity (1994). Direct, real-time measurement of rapid inor- function of IF3 during transition from the 30 S ganic phosphate release using a novel fluorescent initiation complex to the 70 S initiation complex. probe and its application to actomyosin subfragment J. Mol. Biol. 373, 551–561. 1 ATPase. Biochemistry, 33, 8262–8271. 17. Milon, P., Konevega, A. L., Gualerzi, C. O. & Rodnina, 25. Selmer, M., Dunham, C. M., Murphy, F. V. t., M. V. (2008). Kinetic checkpoint at a late step in Weixlbaumer, A., Petry, S., Kelley, A. C. et al. (2006). translation initiation. Mol. Cell, 30, 712–720. Structure of the 70S ribosome complexed with mRNA 18. Antoun, A., Pavlov, M. Y., Lovmar, M. & Ehrenberg, and tRNA. Science, 313, 1935–1942. M. (2006). How initiation factors maximize the 26. Gao, N. & Frank, J. (2006). A library of RNA bridges. accuracy of tRNA selection in initiation of bacterial Nat. Chem. Biol. 2, 231–232. protein synthesis. Mol. Cell, 23, 183–193. 27. Myasnikov, A. G., Marzi, S., Simonetti, A., Giuliodori, 19. Tomsic, J., Vitali, L. A., Daviter, T., Savelsbergh, A., A. M., Gualerzi, C. O., Yusupova, G. et al. (2005). Spurio, R., Striebeck, P. et al. (2000). Late events of Conformational transition of initiation factor 2 from translation initiation in bacteria: a kinetic analysis. the GTP- to GDP-bound state visualized on the EMBO J. 19, 2127–2136. ribosome. Nat. Struct. Mol. Biol. 12, 1145–1149. 20. Milon, P., Konevega, A. L., Peske, F., Fabbretti, A., 28. Allen, G. S., Zavialov, A., Gursky, R., Ehrenberg, M. & Gualerzi, C. O. & Rodnina, M. V. (2007). Transient Frank, J. (2005). The cryo-EM structure of a translation kinetics, fluorescence, and FRET in studies of initia- initiation complex from Escherichia coli. Cell, 121, tion of translation in bacteria. Methods Enzymol. 430, 703–712. 1–30. 29. Gao, H., Zhou, Z., Rawat, U., Huang, C., Bouakaz, L., 21. Marshall, R. A., Aitken, C. E. & Puglisi, J. D. (2009). Wang, C. et al. (2007). RF3 induces ribosomal GTP hydrolysis by IF2 guides progression of the conformational changes responsible for dissociation ribosome into elongation. Mol. Cell, 35,37–47. of class I release factors. Cell, 129, 929–941. 22. Hamel, E., Koka, M. & Nakamoto, T. (1972). Require- 30. Allen, G. S. & Frank, J. (2007). Structural insights on ment of an Escherichia coli 50 S ribosomal protein the translation initiation complex: ghosts of a univer- component for effective interaction of the ribosome sal initiation complex. Mol. Microbiol. 63, 941–950.