Pentameric Organization of the Ribosomal Stalk Accelerates Recruitment of Ricin A Chain to the Ribosome for Depurination

Abstract

Ribosome inactivating proteins (RIPs) depurinate a universally conserved adenine in the α-sarcin/ricin loop (SRL) and inhibit protein synthesis at the translation elongation step. We previously showed that ribosomal stalk is required for depurination of the SRL by ricin toxin A chain (RTA). The interaction between RTA and ribosomes was characterized by a two-step binding model, where the stalk structure could be considered as an important interacting element. Here, using purified yeast ribosomal stalk complexes assembled in vivo, we show a direct interaction between RTA and the isolated stalk complex. Detailed kinetic analysis of these interactions in real time using surface plasmon resonance (SPR) indicated that there is only one type of interaction between RTA and the ribosomal stalk, which represents one of the two binding steps of the interaction with ribosomes. Interactions of RTA with the isolated stalk were relatively insensitive to salt, indicating that nonelectrostatic interactions were dominant. We compared the interaction of RTA with the full pentameric stalk complex containing two pairs of P1/P2 proteins with its interaction with the trimeric stalk complexes containing only one pair of P1/P2 and found that the rate of association of RTA with the pentamer was higher than with either trimer. These results demonstrate that the stalk is the main landing platform for RTA on the ribosome and that pentameric organization of the stalk accelerates recruitment of RTA to the ribosome for depurination. Our results suggest that multiple copies of the stalk proteins might also increase the scavenging ability of the ribosome for the translational GTPases.

Introduction

Ricin, a type II ribosome-inactivating protein (RIP)² from the castor bean plant Ricinus communis depurinates a critical adenine (A3027 in Saccharomyces cerevisiae) in the conserved α-sarcin/ricin loop (SRL) of large rRNA, resulting in translation inhibition (1–3). Ricin holotoxin contains one enzymatically active A chain (RTA) and one cellular binding B chain, which facilitates cell entry but must be separated from the A chain for RTA activity (4). RTA is thought to interact with 60S ribosomes at a site overlapping that of the elongation factors 1 and 2 (eEF1 and eEF2), and eEF2 can partially protect the ribosome from RTA damage (5–8). The eEF2 and other translational GTPases (tGTPases) interact with the ribosomes at the GTPase associated center (GAC), which includes the ribosomal stalk and the SRL (9). The stalk is a key player in the recruitment of tGTPases and in the stimulation of factor-dependent GTP hydrolysis during translation (10), and together with the SRL, it plays an active role in GTPase activation of translation factors and phosphate release (11, 12). Recent results indicate that A2660 in the bacterial SRL, which is targeted by RIPs, is critical for triggering the GTPase activity of elongation factor G (13).

The stalk is a multimeric complex, but its composition is variable among species. In eukaryotes, the stalk occurs in a pentameric configuration P0-(P1/P2)₂ (14, 15), where P0 constitutes the base of the stalk and anchors two P1/P2 dimers (16). The N-terminal domain of the P0 protein is responsible for attachment of the stalk to the large rRNA, and the P-domain anchors the P1/P2 proteins at two separate binding sites (17–19). The P1/P2 proteins form a heterodimer (20–22), with the N-terminal domain responsible for dimerization and linking the dimer to the P0 protein (23, 24), whereas the C-terminal part is regarded as a functional element involved in factor recruitment (19). In S. cerevisiae, the P1 and P2 proteins are further subdivided into P1α and P1β and P2α and P2β, and they form two heterodimers, P1α/P2β and P1β/P2α (24, 25). In eukaryotes, the pentameric organization of the stalk is found exclusively. In contrast, a heptamer is described in Archaea (26), whereas bacteria have pentameric as well as heptameric organization (27). It is not clear why the ribosomal stalk contains multiple copies of the dimers. It was suggested that the protein synthesis machinery may function more efficiently with multiple copies of the dimers (18, 28–30), and in eukaryotes, it is postulated that the stalk may have a regulatory role in gene expression, adjusting cell metabolism to changing environmental conditions (31, 32).

The individual P proteins of the ribosomal stalk or their C termini have been reported to interact with the RIPs, such as trichosanthin (33–36), Stx1 (Shiga-like toxin 1) (37), and maize ribosome-inactivating protein (38). However, these studies provided only a fragmentary view of the ribosome interactions because individual P proteins or short peptides were used. Our previous work, using intact ribosomes, showed that RTA interacts with the ribosomal stalk to access its substrate, the SRL in yeast, and that this interaction is important for the cytotoxicity of RTA (39). The interaction of RTA with ribosomes did not follow a simple 1:1 interaction model and was characterized by a two-step binding model (40). According to this model, the slower, nonspecific interactions concentrate RTA molecules on the surface of ribosomes and facilitate faster, more specific interactions with the stalk (40). Based on this model, we predicted that there should be only one type of interaction between RTA and the isolated stalk complex. Here, we present the binding kinetics and affinity of the interactions between RTA and the native stalk isolated from the yeast, S. cerevisiae. Our results show that RTA interacts with the isolated stalk pentamer in a simple 1:1 interaction model, consistent with our previously proposed model (40). We further show that the association rate of RTA with the stalk pentamer is higher than with either trimer, demonstrating that multiple copies of stalk proteins accelerate the recruitment of RTA to the ribosome for depurination.

EXPERIMENTAL PROCEDURES

Yeast Strains and Ribosomal Stalk Pentamer and Trimer Purification

Two yeast strains, P0-TH199 (MATa; RPP0-TH199; his3Δ1; leu2Δ0; lys2Δ0; ura3Δ0), and P0-TH230 (MATa; RPP0-TH230; his3Δ1; leu2Δ0; lys2Δ0; ura3Δ0) were used in the purification of the stalk pentamer and trimer as described previously (17). A new strain, P0ΔH₂-TH199 (MATa; RPP0_Δ231–258-TH199; his3Δ1; leu2Δ0; lys2Δ0; ura3Δ0) was prepared according to the previously described protocol (17). In all strains, the gene encoding the ribosomal P0 protein (RPP0) had been modified to introduce a thrombin cleavage site (Fig. 1). Yeast strain P0-TH199, in which a thrombin cleavage site and a His₆ peptide were inserted after Pro-199 of the P0 protein, was used to isolate the stalk pentamer TH199 (P0_200–312-P1α/P2β-P1β/P2α). The yeast strain P0-TH230, in which the thrombin cleavage site plus a His₆ were inserted at the position after Pro-230, was used to isolate the stalk trimer, TH230 (P0_231–312-P1β/P2α). The yeast strain P0ΔH₂-TH199, in which the thrombin cleavage site plus the His₆ were the same as strain TH199, but the binding site for P1β/P2α on the P0 (Δ231–258) was deleted, was used to isolate the stalk trimer, TH199ΔH₂ (ΔH₂P0_200–312-P1α/P2β). There were no significant differences in growth rates between wild type strain and the strains carrying the modified RPP0 gene under normal growth conditions. The purity of stalk complexes was verified by four different biochemical methods: size exclusion chromatography, isoelectrofocusing, native PAGE, and nondenaturing mass spectrometry (data not shown). All complexes were shown to be stable entities; in particular, TH199 represented an exceptionally stable structure (15).

FIGURE 1.

Schematic representation of the stalk complexes prepared from the genetically modified yeast strains. The yeast strain P0-TH199, in which a thrombin cleavage site and a His₆ peptide were inserted after Pro-199 of the P0 protein, was used to isolate the stalk pentamer TH199. The yeast strain P0-TH230, in which the thrombin cleavage site plus a His₆ were inserted at the position after Pro-230, was used to isolate the stalk trimer TH230. The yeast strain P0ΔH₂-TH199, in which the thrombin cleavage site plus the His₆ were the same as strain TH199, but the binding site for P1β/P2α on the P0 (Δ231–258) was deleted, was used to isolate the stalk trimer TH199ΔH₂. The compositions of the isolated stalk complexes are indicated at the top with the calculated molecular mass.

Recombinant RTA

N-terminal His-tagged recombinant RTA was obtained from BEI Resources (Manassas, VA). The C-terminal His-tagged RTA was purified from Escherichia coli as described previously (40). Both proteins showed a single band on SDS-PAGE by Coomassie Blue staining and by immunoblot analysis and were active in in vitro translation inhibition and ribosome depurination assays (40).

Analysis of Interaction between RTA and Isolated Ribosomal Stalk Complexes by Coimmunoprecipitation

Protein A-agarose (20 μl) (Santa Cruz Biotechnology, Santa Cruz, CA) was incubated with polyclonal rabbit anti-RTA serum (10 μl) overnight at 4 °C in 70 μl of Biacore running buffer (10 mm HEPES, 150 mm NaCl, 10 mm MgOAc, pH 7.4, with 0.005% surfactant P20) and washed three times with running buffer. Stalk complexes (50 pmol) were mixed with 500 pmol N-His-tagged RTA in 50 μl of Biacore running buffer and incubated overnight at 4 °C. The washed protein A-agarose was added to the RTA-stalk complexes in 400 μl of final volume, and the mixture was rotated at 4 °C for 3 h. The agarose was washed three times with PBS. The proteins were eluted with 20 μl 2× SDS-PAGE sample buffer, and 10 μl of each protein was separated by 15% SDS-PAGE. The P proteins were detected by monoclonal antibody against the C termini of all P proteins (3BH5) (41) at a 1:4 dilution. The blot was stripped with 8 m guanidine-HCl and reprobed with monoclonal antibody against RTA (MAB 3) from BEI Resources at a 1:1000 dilution.

Binding Kinetics of Interaction between RTA and Isolated Ribosomal Stalk Complexes

The kinetic studies were conducted using surface plasmon resonance (SPR) with either a Biacore 3000 or a Biacore T100 (GE Healthcare). We first tried to capture the stalk complexes on activated nitrilotriacetic acid (NTA) chips as the ligand because the stalk complexes contained a His-tag at the N-terminal end of truncated P0. The stalk complexes were captured on NTA chips to <5 resonance units (RU) when their concentrations were varied up to 30 nm, indicating that the activated NTA chip was not able to capture the stalk complexes efficiently. As our previous results indicated that N-His-tagged RTA or C-His tagged RTA was readily captured by the NTA chip (40), we immobilized N-His or C-His tagged RTA on the NTA chip as the ligand in subsequent experiments. RTA was captured on the NTA chip in the detection channel at 1000 RU, and the same amount of His-tagged enhanced green fluorescent protein (EGFP), which did not interact with the isolated stalk complexes, was captured on the reference channel as the control. The purified yeast stalk pentamer or the trimer complexes were passed over the reference channel and the detection channel as the analyte, and the binding was measured in real time. When Biacore 3000 was used, multicycle kinetic analysis was performed where each sample concentration was run in a separate cycle, and the surface was regenerated between each cycle with NTA regeneration buffer and 0.3% SDS. The association was monitored for 3 min and dissociation for 5 min. In contrast, in Biacore T100, kinetic analysis can be performed using single-cycle kinetics, in which a concentration series of the analyte is injected in a single-cycle without regenerating the surface between injections. Single-cycle kinetics approach was used with a Biacore T100 by evaluating five different stalk complex concentrations in each cycle and then regenerating the surface with NTA regeneration buffer and 0.3% SDS. The association was monitored for 5 min, and the final dissociation was monitored for 10 min. Standard NTA running buffer (0.01 m HEPES, 0.15 m NaCl, 50 μm EDTA, and 0.005% surfactant P20, pH 7.4) containing different concentrations of MgOAc and KCl were used. Different flow rates were analyzed to examine mass transport limitation (supplemental Fig. S1) and a flow rate of 60 μl/min was selected for all kinetic measurements.

Statistical Evaluation

The data presented in Table 2 was analyzed by analysis of variance using SAS software (version 9.1.3, SAS Institute, Inc.). Mean comparisons were performed using Fisher's least significant difference.

TABLE 2.

Kinetic parameters of the interaction of RTA with the stalk pentamer and the stalk trimers using single-cycle kinetic analysis

Single-cycle kinetics was used for the interaction analysis using a Biacore T100. Kinetic profiles of the interaction of RTA with the stalk pentamer and the stalk trimers are shown in Fig. 5. N-His-RTA was immobilized on the NTA chip in running buffer, containing 10 mm HEPES-HCl, pH 7.4, 150 mm NaCl, 10 mm MgOAc, 50 μm EDTA, 0.005% surfactant P20, and the same amount of EGFP was immobilized on the reference channel. The association time was 5 min, and the final dissociation time was 10 min. The stalk complexes were used at 0.2–25 nm. Data are shown as mean ± S.E. (n = 4).

Complex stoichiometry	Stalk pentamer, TH199 (P0200–312-P1α/P2β-P1β/P2α)	Stalk trimer, TH230 (P0231–312-P1β/P2α)	Stalk trimer, TH199ΔH2 (ΔH2P0200–312-P1α/P2β)
kon (m−1 s−1)	(2.81 ± 0.17) × 106a	(1.42 ± 0.04) × 106b	(1.21 ± 0.15) × 106b
koff (s−1)	(1.20 ± 0.05) × 10−3b	(0.99 ± 0.04) × 10−3b	(1.69 ± 0.38) × 10−3a
KD (m−1)	(4.28 ± 0.16) × 10−10b	(6.98 ± 0.35) × 10−10b	(13.6 ± 1.34) × 10−10a

^a,b Means corresponding to k_on and K_D followed by the same letter in each row are not significantly different (Fisher's least significant difference; p ≤ 0.001). Means corresponding to k_off followed by the same letter in each row are not statistically different (Fisher's least significant difference; p ≤ 0.05).

RESULTS

RTA Interacts with Stalk Complexes from Yeast

The configurations of the three different stalk complexes prepared from yeast are shown in Fig. 1. The pentameric complex, P0_200–312-P1α/P2β-P1β/P2α (TH199) and the trimeric complex, P0_231–312-P1β/P2α (TH230) were described previously (15). To construct the third complex, ΔΗ₂P0_200–312-P1α/P2β (TH199ΔH₂), the amino acid region corresponding to hypothetical α-helix 2 (residues 231–258), which is responsible for P1β/P2α binding, was removed (17). The complexes were purified to homogeneity (supplemental Figs. S2, A and B). Nondenaturing mass spectrometry showed that all complexes represented stable protein entities, and their measured masses closely corresponded to the theoretical one, indicating proper stoichiometric organization (data not shown) (15).

To determine whether RTA interacts with the purified yeast stalk complexes, polyclonal RTA antibodies were used to coimmunoprecipitate RTA with the different stalk complexes. The RTA-stalk complexes pulled down with anti-RTA were separated by SDS-PAGE, transferred to nitrocellulose, and probed with monoclonal antibodies against the conserved C termini of P proteins (41). As shown in Fig. 2A, the three different stalk complexes were each immunoprecipitated only when mixed with RTA and not when they were incubated alone. When RTA was coimmunoprecipitated with the pentamer (TH199), the intensity of the lower band, which corresponds to four different P1/P2 proteins, was less than the intensity of the upper band, which corresponds to P0, possibly because the monoclonal antibody (41) does not recognize the C termini of all four P proteins equally well, as shown in the supplemental Fig. S2B. When the blot was stripped and probed with monoclonal anti-RTA, RTA was detected in the lanes that contained the stalk proteins (Fig. 2B), indicating that RTA interacts with the three different stalk complexes in a stable manner.

FIGURE 2.

Coimmunoprecipitation of RTA with stalk complexes. A, stalk complexes (50 pmol) were incubated with N-His-RTA (500 pmol) and immunoprecipitated using polyclonal anti-RTA. Immunoprecipitated proteins (10 μl) were separated on a 15% SDS-PAGE, transferred to nitrocellulose, and probed with monoclonal antibody against the C termini of P proteins (3BH5, 1:4 dilution) (41). B, the blot was stripped using 8 m guanidine-HCl and reprobed with monoclonal anti-RTA antibody (MAB3, 1:1000 dilution). MM, molecular mass.

RTA and Stalk Pentamer TH199 (P0_200–312-P1α/P2β-P1β/P2α) Interaction Fits a Simple 1:1 Interaction Model

To characterize the RTA and stalk complex interaction, we examined the binding kinetics of the interaction between RTA and the TH199 complex using a Biacore 3000 under the same conditions that were previously used to study the interaction of RTA with ribosomes (40). As shown in Fig. 3A, the interaction of RTA with the stalk pentamer could fit well with a simple 1:1 interaction model, indicating that only one type of interaction (AB) was observed between RTA and the stalk pentamer. Both association (k_on) and dissociation (k_off) rate constants were slower than the previously reported AB1 interactions thought to be specific for the stalk but faster than the stalk-nonspecific AB2 interactions between RTA and ribosomes (Table 1, 200 mm salt). To further analyze the interaction between RTA and the stalk pentamer, we increased the total salt concentration in the running buffer as we did in the RTA-ribosome interaction studies (40). Surprisingly, we found that the interaction between RTA and the stalk pentamer was not very sensitive to the salt concentration. Even in the presence of 300 or 400 mm salt (Fig. 3, B and C), the interaction fit well with the single 1:1 interaction model, indicating that monovalent ions did not exert a strong effect on the RTA-pentamer interaction. Although the equilibrium dissociation constant (K_D) of the RTA-stalk pentamer interaction did not change much when the salt concentration was increased, the association (k_on) and the dissociation (k_off) rate constants decreased slightly with increasing salt concentrations (Table 1). The decrease in k_on and k_off with increasing salt concentrations is shown in Fig. 3D, where interaction of the stalk complex with RTA was compared at three different salt concentrations. These results indicated that the interaction of RTA with the isolated stalk pentamer is strong and is not dominated by the electrostatic interactions. In contrast, previous analyses showed that RTA-ribosome interaction is salt-dependent and sensitive to high salt concentrations (40). We further checked the effect of the divalent ion Mg²⁺, which is critical for stabilizing the ribosome structure. Kinetic studies in the presence of various concentrations of Mg²⁺ (5, 10, and 20 mm) showed that K_D did not change much with increasing Mg²⁺ concentrations (supplemental Fig. S3 and Table S1). However, both rate constants (k_on and k_off) decreased moderately for the stalk pentamer with increasing Mg²⁺ concentrations (supplemental Fig. S3 and Table S1) as observed for the monovalent ions (Table 1).

FIGURE 3.

Kinetic profiles of the interaction of RTA with the stalk pentamer complex (TH199) at different salt concentrations. Biacore 3000 was used for the kinetic measurements using previously described conditions (40). The kinetic parameters are shown in Table 1. N-His-RTA was immobilized on the NTA chip at a concentration of 1000 RU in running buffer containing 10 mm HEPES-HCl, pH 7.4, 150 mm NaCl, 50 μm EDTA, 0.0005% surfactant P20, with 5 mm MgOAc, and the same amount of EGFP was immobilized on the reference channel. The stalk complexes (0.1–15 nm) were passed over the immobilized RTA and EGFP in running buffer containing 50, 150, or 200 mm KCl. The association time was 3 min, and the dissociation time was 5 min. The surface was regenerated after each analyte concentration. The graph shows experimental data (colored lines) and the kinetic fit for 1:1 interaction model (black lines). A, 200 mm total salt; B, 300 mm total salt; C, 400 mm total salt; D, comparison of the interaction data at 200, 300, and 400 mm total salt with the stalk pentamer at 8 nm. Resp. Diff., response difference.

TABLE 1.

Comparison of the kinetic parameters of the interaction of RTA with yeast ribosomes and with the isolated stalk pentamer

	Total salt
	200 mm	300 mm	400 mm
Stalk pentamer (P0200–312-P1α/P2β-P1β/P2α)a
kon (m−1 s−1)	2.06 ± 0.30 × 106	1.42 × 106	1.33 × 106
koff (s−1)	2.80 ± 0.33 × 10−3	2.62 × 10−3	1.81 × 10−3
KD (m−1)	1.38 ± 0.12 × 10−9	1.84 × 10−9	1.36 × 10−9
Ribosomesb
kon (m−1 s−1)
AB1c	1.75 × 107c	5.45 × 106	No interaction
AB2	7.77 × 105
koff (s−1)
AB1c	1.02 × 10−1c	5.82 × 10−3	No interaction
AB2	1.13 × 10−3
KD (m−1)
AB1c	5.83 × 10−9c	1.07 × 10−9	No interaction
AB2	1.45 × 10−9

^a Kinetic profiles of the interaction of RTA with the stalk pentamer using multicycle kinetics with Biacore 3000 are shown in Fig. 3A. The data shown for the stalk pentamer at 200 mm total salt condition is average ± S.E. from three independent measurements. The Chi² of three different individual fitting for the stalk pentamer were below the 10% of R_max. The fitting parameters are R_max = 173–190 and Chi² = 8.7 for the 300 mm total salt condition and R_max = 157–207 and Chi² = 7.8 for the 400 mm total salt condition. N-His-RTA was immobilized on the NTA chip in running buffer, containing 10 mm HEPES-HCl, pH 7.4, 150 mm NaCl, 50 μm EDTA, 0.0005% surfactant P20, with 5 mm MgOAc. The stalk complexes (0.1–15 nm) were passed over the immobilized RTA and EGFP in running buffer, containing different concentrations (50, 150, or 250 mm) of KCl. The association time was 3 min, and dissociation time was 10 min.

^b Data for ribosomes are from a previous publication (40).

^c The kinetic parameters of the AB1 interaction in the presence of 200 mm total salt are shown.

Stalk Pentamer Interacts with C Terminus of RTA

We previously showed that the C terminus of RTA had to be exposed on the surface of the NTA chip for RTA to interact with ribosomes (40). We examined the effect of the orientation of RTA on its interaction with the stalk pentamer. The N- or C-terminally His-tagged RTA was immobilized on the NTA chip, allowing the C- or N-terminal domain of RTA, respectively, to be exposed to the solvent. When the stalk pentamer was passed over the chip, interaction was observed with the C-terminal domain-exposed RTA (N-His-RTA) (Fig. 4). Very little interaction was observed at the high stalk concentration with the N-terminal domain-exposed RTA (C-His-RTA) (Fig. 4). These results indicate that as observed with the RTA-ribosome interaction (40), the RTA-stalk pentamer interaction involves the C-terminal domain of RTA.

FIGURE 4.

Kinetic profiles of the interaction of stalk pentamer with the N- or C-terminal His-tagged RTA. The N- or C-His-RTA was immobilized on the NTA chip 1000 RU using Biacore 3000. The same amount of EGFP was immobilized on the reference channel. The stalk pentamer (1, 4, or 10 nm) was passed over the surface containing the immobilized N- or C-His-RTA as the analyte. The experimental conditions were the same as described in the legend to Fig. 3. Resp. Diff., response difference.

RTA Interacts with Stalk Trimers at a Slower Rate than with Stalk Pentamer

To compare the binding kinetics of RTA with the whole native stalk pentamer and the stalk trimers, we used genetically modified yeast strains to prepare the stalk complexes with individual dimers together with the C terminus of the P0 protein. The kinetic parameters of the interaction between RTA and each stalk trimer and the pentamer were studied with a Biacore T100 using single-cycle kinetic analysis. Because the surface is not regenerated between injections in the single-cycle approach, multiple samples can be analyzed in a shorter period of time, allowing comparison among different samples under identical conditions. The kinetic parameters of the interaction of RTA with the stalk pentamer and each trimer were measured in a single run. The experiments were repeated four times by changing the sample order each time. As shown in Fig. 5, B and C, the interaction between RTA and the stalk trimers fit very well with a simple 1:1 interaction model. The association rate constant of the RTA-stalk pentamer interaction was about twice that of either the RTA-stalk trimer interaction and the difference between k_on for the pentamer, and the trimers was highly significant (Table 2). The dissociation rates only had small differences between TH199 and TH230 and were not significantly different but were significantly faster for TH199ΔH₂. Comparison of the equilibrium dissociation constants (K_D) indicated that RTA had similar affinity for the stalk pentamer and the stalk trimer, TH230, but had lower affinity for the stalk trimer, TH199ΔH₂ (Table 2).

FIGURE 5.

Kinetic profiles of RTA binding to stalk pentamer or trimers using single-cycle kinetics. Sensorgrams show blank and reference subtracted data with kinetic fit for 1:1 interaction model using Biacore T100. Dashed lines represent the experimental data, and solid lines represent the kinetic fit. Five different analyte concentrations (0.2–25 nm) were used in each cycle, and the surface was regenerated after each cycle. The association was measured for 5 min, and the final dissociation time was 10 min. N-His-RTA was immobilized on an NTA chip at 1000 RU, and the same amount of EGFP was immobilized on the reference channel in the running buffer, containing 10 mm HEPES, pH 7.4, 150 mm NaCl, 10 mm MgOAc, 50 μm EDTA, and 0.005% surfactant P20. The kinetic parameters are shown in Table 2. A, stalk pentamer, TH199; B, stalk trimer, TH230; C, stalk trimer, TH199ΔH₂. Resp. Diff., response difference.

To further confirm these results, we measured the interaction kinetics using multicycle kinetic analysis with a Biacore 3000. In multicycle analysis, each sample concentration was run in a separate cycle, and analyte was removed by regeneration of the surface after each cycle. The overlaid sensorgrams collected from individual analysis cycles are shown in the supplemental Fig. S4, and the kinetic parameters are shown in the supplemental Table S2. Although the rate constants differed slightly from the values obtained using the single-cycle kinetic method (Table 2), the trends were very similar. The association rate of RTA with the pentamer was about twice its association rate with each trimer and TH199 and TH230 had higher affinity for RTA than TH199ΔH₂ (supplemental Table S2). The faster association with the stalk pentamer than with the stalk trimers suggests that ribosomes that contain two dimers recruit RTA twice as fast as ribosomes that contain a single dimer. As indicated by the smooth interaction curves observed for the RTA-stalk pentamer and RTA-stalk trimer interactions (Figs. 3 and 5 and supplemental Fig. S4), the stalk pentamer and trimers were stable under the conditions used for the kinetic measurements. However, the influence of salt on the interaction of RTA with the stalk trimers could not be measured because the trimers showed lower intrinsic stability than the stalk pentamer in the presence of higher salt (15).

DISCUSSION

Kinetic Parameters of Interaction of RTA with Isolated Stalk Pentamer Differ from Its Interaction with Intact Ribosomes

Previous studies examined the interaction of RIPs with individual P proteins or short peptides (33–38), which are not found in vivo or with intact ribosomes (39, 40). Here, using natively assembled stalk complexes from yeast, we show that RTA interacts directly with the yeast stalk complexes and provide the first detailed kinetic view of the interactions between RTA and the yeast ribosomal stalk. The kinetic data show that the stalk constitutes the main binding site for RTA on the ribosome. However, we observed differences between interactions of RTA with ribosomes and with the isolated stalk complexes. The major differences were that the association and dissociation rates were faster with ribosomes than with the isolated pentameric stalk and the salt sensitivity of the RTA-ribosome and RTA-stalk complexes was different. In our previous work, only AB1 interactions, thought to be specific for RTA-stalk interplay, were observed at 300 mm salt (40). The association and dissociation rates of the interaction between RTA and ribosomes were about two to four times faster than the association and dissociation rates of the interaction of RTA with the isolated stalk pentamer at 300 mm salt concentration and were even higher at 200 mm salt concentration (Table 1). However, comparison of the equilibrium dissociation constants at 300 mm salt indicated that the affinity of RTA for the isolated stalk pentamer and for the ribosome (AB1 interactions) was similar. Based on these results, the stalk complex could be considered as a primary target for RTA. We previously showed that RTA interacts with ribosomes via two different types of interactions, the slower, nonspecific AB2 interactions, which might concentrate RTA molecules on the surface of ribosomes and the faster, AB1 interactions, which are attributed to RTA-stalk interplay. Unlike the interaction with ribosomes, the kinetic profiles of the interaction between RTA and the isolated stalk fit a simple 1:1 interaction model, indicating that there is only one type of interaction with the stalk. The kinetic parameters indicated that this interaction was slower than the AB1 interaction but faster than the AB2 interaction with ribosomes (40). Therefore, the AB2 interactions, which create a pool of RTA on the ribosome, may contribute to the faster association rate of RTA with ribosomes than with the isolated stalk (40). The faster dissociation rate of RTA with ribosomes than with the isolated stalk might be determined by the ribosomal particle, where dynamics of the stalk may deliver RTA to the SRL by undergoing conformational changes, as proposed for the interaction between the C termini of the stalk proteins and the translation factors (10, 42).

The electrostatic character of the ribosomal surface was shown to be critical for its interaction with a ribotoxin (43) and for the enzymatic activity of RIPs (44). Similarly, we showed that electrostatic interactions dominated the interaction between RTA and ribosomes, enabling rapid target localization, and accounted for much higher depurination rate of intact ribosomes than of isolated rRNA (40). At high salt concentration (400 mm), there was no interaction between RTA and ribosomes (40), but interaction between RTA and isolated stalk was observed (Table 1), indicating that binding of RTA to the ribosome occurs in a complex manner, where perturbations in RTA-ribosome interaction caused by high concentrations of salt abolish the direct interaction between RTA and the stalk. In the isolated system, the interaction between RTA and the stalk complex, being insensitive to high salt, is dominated by nonelectrostatic interactions and/or shielded from the negative influence of mono- and divalent ions. Interestingly, although the RTA-stalk complex was not salt-sensitive, both the k_on and k_off decreased gradually with increasing salt concentration, whereas the equilibrium dissociation constants did not show much change. These results underline the fact that the stalk is a major docking site for RTA on the ribosome and that the higher salt slows down both the association and dissociation rates without affecting the overall affinity. Salt may have a negative effect on the association process, but once the RTA and stalk pentamer are together, the binding becomes stable even in the presence of the higher salt concentration. Therefore, it is possible that when RTA and the stalk complex are far from each other, the electrostatic interactions bring them together on the ribosome, as shown previously (40). However, once RTA and the stalk complex are in closer proximity, the interaction is dominated by nonelectrostatic interactions. This observation is supported by structural analysis indicating that electrostatic and hydrophobic interactions were critical for a complex between trichosanthin and a peptide corresponding to the C termini of the P protein (36), suggesting that cooperation of several types of interactions may contribute to the overall interaction between RTA and the ribosome. In our previous report (40), we have shown that RTA interacts with the ribosome in a stalk-specific and nonspecific manner, suggesting that the stalk is the interacting partner for RTA. Using a similar experimental setup, but with much lower complexity, we show here that the ribosomal stalk represents the primary landing platform for RTA, and these two partners have exceptionally high affinity for each other. We further show that RTA uses its C-terminal domain to interact with the isolated stalk pentamer. Therefore, the interaction of RTA with the isolated stalk complexes described here is the interaction observed previously with the stalk complex on intact ribosomes (40).

Association of RTA with Stalk Pentamer Is Faster than Its Association with Stalk Trimers

In all living organisms, the ribosomal stalk contains more than two copies of P1/P2 protein dimers (9). The biological significance of the multiple copies of P1/P2 protein dimers and distinct roles for each dimer are not known. To determine how multiplication of the stalk proteins affects the interaction of the stalk with RTA, we have developed a system to measure the interaction of RTA with each heterodimer separately and showed that RTA can bind to either heterodimer on the C termini of P0 in the absence of the other heterodimer. Using single-cycle kinetics, we were able to compare the interaction of RTA with the stalk pentamer to its interaction with the stalk trimers under identical conditions. The assay was repeated four times with reproducible results. The differences in the association rates between RTA-stalk pentamer versus RTA-stalk trimer interactions were highly significant. We confirmed these results using a different SPR instrument and a different interaction method. As observed by the two methods, the rate of association of RTA with either trimer was about half of its rate of association with the stalk pentamer. The dissociation rates were not significantly different between TH199 and TH230 but were significantly faster for TH199ΔH₂. Because the dissociation constant (K_D) is determined by the dissociation rate divided by the association rate, the K_D for TH230 versus TH199 was not significantly different. However, the affinity of RTA for TH199ΔH₂ was lower with 3-fold higher K_D than the pentameric complex. Our data are also consistent with a model where the two trimers bind RTA essentially independently. In such a case, the on-rate of the pentamer would be expected to be twice that of the individual on-rates for the trimers and the off-rate for the pentamer would be half of the individual off-rates. As a result, the ratio of K_D trimer/K_D pentamer would be ∼4:1, as we indeed observed for TH199ΔH₂ (Table 2). There is some evidence that TH230 binds slightly more tightly than TH199ΔH₂. These results indicate that multiple copies of the P proteins increase the association rate of the interaction between the stalk and RTA and demonstrate that the pentameric configuration of the stalk is optimal for recruitment of RTA to the ribosome for depurination. Our data are in accordance with earlier reports, which showed that the presence of the C-terminal region of the stalk proteins, rather than the numbers of proteins, was responsible for the GTPase turnover of elongation factors (26, 30). Recently, it was shown that only one bacterial stalk protein interacts with elongation factor G and is implicated in stimulating catalysis (45), suggesting that the redundancy in stalk proteins is primarily related to tGTPase scavenging. Our results are also consistent with the recent finding that the bacterial stalk is involved in recognition and recruitment of the IF2 (initiation factor 2) and IF2-mediated association of the ribosomal subunits (46).

The P1α/P2β heterodimer is closer to the stalk base than the P1β/P2α heterodimer in yeast ribosomes, and in vitro reconstitution experiments showed that the P1α/P2β heterodimer is more important for stalk formation than the P1β/P2α heterodimer (29). Therefore, the difference in the role of these heterodimers in the formation of the stalk may have an influence on RTA binding. We cannot exclude that the lower affinity of RTA for the TH199ΔH₂ complex may be due to a perturbation in its conformation. However, the high association and dissociation rates and the K_D value indicate that this interaction has significant biological meaning. RTA may interact better with the P1β/P2α heterodimer because this dimer which is further away from the stalk base on P0 (47) may be more accessible to RTA. In any case, observed differences in the way the two pairs of heterodimers recruit RTA indicate that they have an overlapping function with respect to RTA, but the stalk pentamer is the optimum protein configuration.

In summary, in a previous study, using yeast mutants defective in the stalk proteins, we proposed a two-step interaction model to describe the interaction of RTA with wild type ribosomes from yeast (40). Based on this model, we predicted that RTA should interact with the ribosomal stalk pentamer in a simple 1:1 model. The results presented here provide direct evidence that the ribosomal stalk is the main docking site for RTA on the ribosome and that there is only one type of interaction between RTA and the stalk. Furthermore, we demonstrate that the interaction between RTA and the isolated stalk pentamer is not salt-sensitive, and therefore cooperative interplay of electrostatic and nonelectrostatic interactions likely play the most important role in complex stabilization. Analysis of the trimeric complexes provided evidence that the number of P protein dimers increases the association rate of the stalk for RTA and the two trimers possess independent ability to bind RTA. Finally, we conclude that the stalk constitutes a dominant interaction interface for the depurination activity of RTA. Our results suggest that the eukaryotic stalk complex, which is responsible for ribosome specificity of eukaryotic translation factors (16), may also be the basis for the ability of RTA to depurinate the eukaryotic but not the prokaryotic ribosomes (48). Moreover, the binding mechanism presented here may be broadly applicable to the interactions of the stalk with other interacting elements, such as the tGTPases. Multiplication of the stalk proteins on the ribosome might increase scavenging ability of the ribosome for the tGTPases.