Chimeras of Bacterial Translation Factors Tet(O) and EF-G

View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector

FEBS Letters 582 (2008) 1386–1390

Chimeras of bacterial translation factors Tet(O) and EF-G

Nehal S. Thakora,1, Roxana Nechiforb, Paul G. Scottb, Monika Keelanc, Diane E. Taylora,*, Kevin S. Wilsonb,* a Department of Medical Microbiology and Immunology, University of Alberta, Edmonton, Alberta, Canada T6G 2H7 b Department of Biochemistry, University of Alberta, Edmonton, Alberta, Canada T6G 2H7 c Department of Laboratory Medicine and Pathology, University of Alberta, Edmonton, Alberta, Canada T6G 2H7

Received 2 March 2008; revised 14 March 2008; accepted 14 March 2008

Available online 25 March 2008

Edited by Lev Kisselev

tion factor with GTP hydrolytic (GTPase) activity [7]. Tc does Abstract Ribosomal protection proteins (RPPs) confer bacterial resistance to tetracycline by releasing this antibiotic from not affect the subsequent steps of peptide bond formation and ribosomes stalled in protein synthesis. RPPs share structural ribosomal translocation during protein synthesis [2]. Ribo- similarity to elongation factor G (EF-G), which promotes ribosomal translocation, involving tRNA/mRNA movement in somal translocation during normal protein synthesis. We con- the ribosome, is regulated by elongation factor G (EF-G), structed and functionally characterized chimeric proteins of another universal GTPase translation factor [7]. Campylobacter jejuni Tet(O), the best characterized RPP, and Unfortunately, many different bacteria (both pathogenic Escherichia coli EF-G. A distinctly conserved loop sequence at and commensal) have acquired resistance to Tc, which is the tip of domain 4 is required for both factor-specific functions. becoming increasingly widespread [1]. One of the most com- Domains 3–5: (i) are necessary, but not sufficient, for functional mon means for acquiring Tc resistance is through the so- specificity; and (ii) modulate GTP hydrolysis by EF-G, while called ‘‘ribosomal protection proteins’’ (RPPs), harbored in minimally affecting Tet(O), under substrate turnover conditions. Ó 2008 Federation of European Biochemical Societies. Pub- the cytosol of resistant bacteria [8]. Eleven distinct RPP lished by Elsevier B.V. All rights reserved. genes have so far been identified from more than 90 bacterial genera [9]. The predicted amino acid sequences of all Keywords: Antibiotic resistance; Bacterial protein synthesis; RPP genes share significant similarity to all five domains Elongation factor; Ribosomal protection proteins; of EF-G [10,11], suggesting that RPPs may have arisen from Tetracycline an ancestral EF-G gene. The currently best characterized RPP is Tet(O) from Cam- pylobacter jejuni [8]. The GTPase activity of Tet(O), like EF-G, is strongly stimulated by its association with bacterial ribosomes [12]. In vitro GTP hydrolysis can be decoupled 1. Introduction from, but stimulates, the specific functions of both factors. Tet(O) catalyzes GTP-dependent dissociation of Tc from its Tetracycline (Tc) antibiotics have been used extensively primary binding site on stalled ribosome complexes [12]. against diverse bacterial pathogens of humans and animals Cryo-EM studies visualized Tet(O) on the Escherichia coli [1]. Tc inhibits bacterial cell growth by preventing binding of ribosome at 16 A˚ resolution, revealing that Tet(O) has a sim- aminoacyl-tRNA substrates to the ribosome during protein ilar shape as EF-G and binds to a similar site within the inter- synthesis [2]. Tc binds to a single ‘‘primary’’ site on the 30S subunit cavity of the ribosome [13]. subunit of the bacterial ribosome (KD 0.5 lM), as well as However, several observations suggest Tet(O) and EF-G an undefined number of ‘‘secondary’’ (lower affinity) binding interact specifically with the ribosome. Whereas domains 1–2 sites on both 30S and 50S ribosomal subunits [3–5]. The pri- of the two factors interact similarly with the ribosome, their mary site is located in a crevice between the head and shoulder domains 3–5 occupy distinct positions in the ribosomal cavity. of the 30S subunit, where Tc is believed to sterically block the Most intriguingly, the tip of domain 4 of EF-G, which is essen- productive accommodation of aminoacyl-tRNA substrates tial for ribosomal translocation [14,15], reaches deeper into the into the ribosomal A site [4,6]. This process is regulated by A site in the 30S subunit [16,17]. The tip of domain 4 of Tet(O) elongation factor Tu (EF-Tu), a universally conserved transla- is pointed toward the primary Tc binding site in the 30S subunit [13]. Tet(O) preferentially binds to Tc-stalled ribosome complexes in the posttranslocational state [18]. EF-G binds *Corresponding authors. to pretranslocational complexes, inducing a ratcheting motion E-mail addresses: [email protected] (D.E. Taylor), of the two ribosomal subunits, not observed with Tet(O) [13]. [email protected] (K.S. Wilson). Finally, a related factor, Tet(M), cannot functionally replace 1Present address: Apoptosis Research Centre, Department of Pediat- EF-G or EF-Tu in vivo [19]. rics, University of Ottawa, ChildrenÕs Hospital of Eastern Ontario In the present study, we wished to define the structural Research Institute, Ottawa, Ontario, Canada K1H 8L1. elements of Tet(O) that functionally distinguish it from Abbreviations: EF, elongation factor; GTPase, GTP hydrolytic activ- EF-G. Towards that end, we constructed a series of chimeras ity; MIC, minimal inhibitory concentration; RPPs, ribosomal protec- of the two factors and characterized their functions in assays tion proteins; Tc, tetracycline specific for Tet(O) and EF-G, in vivo and in vitro.

0014-5793/$34.00 Ó 2008 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.febslet.2008.03.023 N.S. Thakor et al. / FEBS Letters 582 (2008) 1386–1390 1387

2. Materials and methods

2.1. Chimeric gene constructions, protein expression, and puriﬁcation Tet(O)-EF-G chimeric gene constructions are described in the sup- plemental material. Chimeric proteins were expressed in E. coli and puriﬁed as described [20].

2.2. Activity assays The minimal inhibitory concentration (MIC) of Tc, preventing the growth of E. coli strains expressing each chimera, was determined as follows. Sterile LB agar plates were prepared containing increasing Tc concentration (1–256 lg/ml, at twofold increments). IPTG (1 mM) was spread onto these plates, to induce chimera expression. E. coli strains were grown in liquid LB + kanamycin (50 lg/ml) to mid-log phase (A600 nm 0.75). Aliquots (100 ll) of each culture were spread onto the plates, incubated (37 °C; 48 h), and checked for visible bacterial growth. [3H]Tc binding to E. coli ribosomes was measured by nitrocellulose ﬁlter binding [20]. Ribosomal translocation was monitored by toeprinting [21]. Hydrolysis of [c-32P]GTP was analyzed by TLC [20].

3. Results

3.1. Structural comparisons between RPP and EF-G factors In designing chimeric factors, we analyzed aligned amino acid sequences of RPP and EF-G proteins from representative Fig. 2. Tet(O)-EF-G chimeric proteins. Domains are defined with bacteria (Fig. S1). Atomic-resolution structures have been respect to T. thermophilus EF-G [22]. determined for Thermus thermophilus EF-G in several functional states [22,23], but no such structures are yet known of any RPP. Thus, we generated a homology-modeled structure chimeric proteins, Tet(O)E and EF-GT, were individually ex- of C. jejuni Tet(O) based on its 51% overall similarity to pressed in E. coli. Cell growth was assessed on solid growth T. thermophilus EF-G (Fig. 1). A similar model was also gen- media containing IPTG (to induce chimera expression) and erated for E. coli EF-G (not shown). These models allowed us increasing Tc concentration. Cells expressing either chimera to demarcate the putative domain boundaries within these pro- were sensitive to Tc (MIC of 2 lg/ml). This was in marked con- teins, from which we designed chimeric proteins. trast to cells expressing intact Tet(O) protein (MIC of 64 lg/ ml), but similar to cells expressing EF-G or no factor (MIC 3.2. Loop sequences at the tip of domain 4 are required for of 2 lg/ml). Tet(O) and EF-G specific functions As a control, we examined the chimeric protein expression At the distal tip of domain 4 of EF-G, a highly conserved levels in E. coli. Following cell lysis, soluble proteins in the ex- loop sequence H(E/D)VDSS (Fig. S1; S3A) is critical for ribo- tract were detected by immunoblotting, probing with antibody somal translocation [14,15]. The corresponding loop sequence against the His6-tag (Fig. 3A) [20]. Tet(O)E was present at sim- in RPPs is completely different and conserved as YSPVST, ilar levels as Tet(O). EF-GT was present at lower levels and while flanking sequences in domain 4 are similar. These obser- partially degraded, while EF-G was present at much higher vations suggested to us that the YSPVST sequence may also be levels. functionally important for RPPs, and functionally distinguish To characterize their activities in vitro, we purified these pro- them from EF-G. teins (Fig. 3B) [20]. We examined each protein for its ability to To address these questions, we swapped the loop sequences promote ribosomal translocation [21]. Whereas EF-G effi- of C. jejuni Tet(O) and E. coli EF-G (Fig. 2). The resulting ciently promoted ribosomal translocation, Tet(O) and both chimeras were completely inactive in this assay, even after 60 min at 37 °C(Fig. 4). We also assayed the ability of each factor to release Tc from E. coli ribosomes [12,20]. Whereas Tet(O) was efficient in this assay, EF-G and both chimeras caused little or no significant Tc release from ribosomes (Fig. 5). The above results provided evidence that both loop sequences at the tip of domain 4 of Tet(O) and EF-G are essential for their specific functions. However, simply transplanting these loops into the other factorÕs context was not sufficient to convert their functional specificities.

3.3. Domains 3–5 of Tet(O) and EF-G are necessary, but not Fig. 1. Homology-modeled structure of C. jejuni Tet(O). Tet(O) domains (color-coded) are superimposed onto the known structure of sufficient, for their specific functions T. thermophilus EF-G (gray) [22]. The Tet(O) structure was generated In addition to the above mentioned loops, several other ele- by using MODELLER (http://www.salilab.org/modeller/). ments in Tet(O) and EF-G are differentially conserved and 1388 N.S. Thakor et al. / FEBS Letters 582 (2008) 1386–1390

may contribute toward the functional specificities of these factors (Fig. S2). EF-G appears to be organized into two larger structural units, consisting of domains 1–2 and 3–5, which may move as rigid bodies depending on the functional state of EF-G [22,23]. Domain 1 is divided into G and G0 subdo- mains, which respectively hydrolyze GTP and regulate this reaction [24,25]. Domain 2 has not yet been functionally defined. Domain 3 affects both GTP hydrolysis and ribosomal translocation [26]. Domains 4 and 5 are individually required for ribosomal translocation, but not for GTP hydrolysis [15,27]. The functional roles of individual domains of Tet(O), or other RPPs, have not yet been defined. By analogy to EF-G, we imagined that elements within domains 3–5 of Tet(O) might function in catalyzing Tc release from the ribosome. We constructed four additional chimeras, swapping en masse domains 3–5 or 4–5 between Tet(O) and EF-G (Fig. 2). As the splice sites for these domain swappings, we chose interdomain sequences linking domains 2/3 and 3/4, which are partially conserved between the factors (Fig. S2; S3B). All four chimeras were expressed as soluble proteins in E. coli cells (Fig. 3A). Tet(O)12 and Tet(O)123 were expressed as stable proteins. EF-G12 and EF-G123 were partially degraded, but full-length proteins could still be detected in these strains. However, these cell strains were all sensitive to low levels of Tc (MIC of 2 lg/ml), suggesting that the expressed chimeras were not sufficiently active in releasing Tc from ribosomes in E. coli cells. Fig. 3. Expression and purification of Tet(O)-EF-G chimeras [20]. (A) The proteins were purified and further characterized in vitro Expressed proteins in E. coli, detected by immunoblotting (probe: anti- (Fig. 3B). Tet(O)12 and Tet(O)123 did not promote significant His6). (B) Purified proteins, analyzed by SDS–PAGE (stain: Coomas- sie). ribosomal translocation (Fig. 4), indicating that EF-G domains 3–5 were not sufficient to confer EF-G specificity. None of the four chimeras promoted significant Tc release from ribosomes (Fig. 5), indicating that Tet(O) domains 3–5 were necessary, but not sufficient, for Tet(O) specificity.

3.4. Tet(O)-EF-G chimeras are active in ribosome-dependent GTP hydrolysis The inactivity of the Tet(O)-EF-G chimeras might be trivi- Fig. 4. Ribosomal translocation, assessed by toeprinting [21]. Pre/ ally explained if they cannot bind properly to the ribosome. post = pre-/posttranslocational complexes. We tested their activities in ribosome-dependent GTP hydrolysis using excess GTP (50 lM) and equimolar protein factor and ribosome (0.5 lM). All factors displayed vigorous GTP hydrolysis under substrate-turnover conditions (Fig. 6). Tet(O) and EF-G were approximately equally active (V080 and 1 100 min ), near their Michaelis constants (KM 80 lM) for GTP binding [20]. The chimeras were less active than the

Fig. 5. Tetracycline release from vacant ribosomes, assessed by ﬁlter binding [20]. Bar graphs represent Tc equivalents bound per ribosome, averaged (±S.D.) from three independent reactions. Fig. 6. Kinetics of ribosome-dependent GTP hydrolysis [20]. N.S. Thakor et al. / FEBS Letters 582 (2008) 1386–1390 1389 wild-type proteins, but still very signiﬁcant compared to con- tional Science and Engineering Research Council of Canada (to D.E.T trol reactions containing factor or ribosome alone, showing and M.K.) and Canadian Institutes of Health Research (to K.S.W.). no detected GTP hydrolysis (data not shown). D.E.T. is a Research Scientist and K.S.W. is a Scholar of Alberta Her- itage Foundation for Medical Research. Interestingly, the GTPase activities of the chimeric factors fell into two distinct categories. The chimeras containing domain 1 of Tet(O) were two to sixfold less active than intact Appendix A. Supplementary data Tet(O). In contrast, the chimeras containing domain 1 of EF-G were 40- to 150-fold less active than intact EF-G. The Supplementary data associated with this article can be latter results suggest that domains 3–5 of Tet(O) have a nega- found, in the online version, at doi:10.1016/j.febslet.2008. tive impact on GTP hydrolysis when they are attached to do- 03.023. mains 1–2 of EF-G.

References

4. Discussion [1] Chopra, I. and Roberts, M. (2001) Tetracycline antibiotics: mode of action, applications, molecular biology, and epidemiology of In this study, we set out to construct chimeric Tet(O)-EF-G bacterial resistance. Microbiol. Mol. Biol. Rev. 65, 232–260. [2] Suarez, G. and Nathans, D. (1965) Inhibition of aminoacyl- proteins with altered functional specificity, guided by an exten- sRNA binding to ribosomes by tetracycline. Biochem. Biophys. sive literature on EF-G. Previous studies have shown that a Res. Comm. 18, 743–750. highly conserved loop sequence, H(E/D)VDSS, at the tip of [3] Epe, B. and Woolley, P. (1984) The binding of 6-demethylchlor- domain 4 of EF-G is critical for catalyzing ribosomal translo- tetracycline to 70S, 50S and 30S ribosomal particles: a quantita- tive study by fluorescence anisotropy. EMBO J. 3, 121–126. cation [14,15]. Here we have shown that a distinctly conserved [4] Brodersen, D.E., Clemons Jr,, W.M., Carter, A.P., Morgan- sequence, YSPVST, at the equivalent location in RPPs is re- Warren, R.J., Wimberly, B.T. and Ramakrishnan, V. (2000) The quired for releasing Tc from ribosomes. The former loop se- structural basis for the action of the antibiotics tetracycline, quence is located in the A site of the 30S ribosomal subunit pactamycin, and hygromycin B on the 30S ribosomal subunit. following ribosomal translocation [16,17]. The latter loop se- Cell 103, 1143–1154. [5] Pioletti, M., Schlunzen, F., Harms, J., Zarivach, R., Gluhmann, quence is expected to be proximal to the primary Tc binding M., Avila, H., Bashan, A., Bartels, H., Auerbach, T., Jacobi, C., site on the ribosome [13]. Hartsch, T., Yonath, A. and Franceschi, F. (2001) Crystal By swapping the loop sequences between Tet(O) and EF-G, structures of complexes of the small ribosomal subunit with we obtained chimeras inactive in both ribosomal translocation tetracycline, edeine and IF3. EMBO J. 20, 1829–1839. [6] Gordon, J. (1969) Hydrolysis of guanosine 50-triphosphate and Tc release, while still retaining ribosome-dependent associated with binding of aminoacyl transfer ribonucleic acid GTPase activity. These results may be interpreted in two ways, to ribosomes. J. Biol. Chem. 244, 5680–5686. which are not mutually exclusive: (i) multiple functional deter- [7] Wilson, K.S. and Noller, H.F. (1998) Molecular movement inside minants may reside within each factor; and/or (ii) the loop se- the translational engine. Cell 92, 337–349. [8] Connell, S.R., Tracz, D.M., Nierhaus, K.H. and Taylor, D.E. quences are incorrectly positioned in the ribosome when (2003) Ribosomal protection proteins and their mechanism of transplanted into the context of the other factor. tetracycline resistance. Antimicrob. Agents Chemother. 47, 3675– To distinguish between these possibilities, we were guided by 3681. EF-G studies indicating that its domains 3–5 are individually [9] Roberts, M.C. (2005) Update on acquired tetracycline resistance required for catalyzing ribosomal translocation [15,26,27], genes. FEMS Microbiol. Lett. 245, 195–203. [10] Sanchez-Pescador, R., Brown, J.T., Roberts, M. and Urdea, M.S. and its domains 1–2 and 3–5 appear to form structurally (1988) Homology of the TetM with translational elongation autonomous units [22,23]. Thus, we swapped domains 3–5 or factors: implications for potential modes of tetM-conferred 4–5 en masse between Tet(O) and EF-G. The resulting chime- tetracycline resistance. Nucleic Acids Res. 16, 1218. ras were again inactive in either factor-specific functions and [11] Taylor, D.E. and Chau, A. (1996) Tetracycline resistance mediated by ribosomal protection. Antimicrob. Agents Chemother. retained GTPase activity. 40, 1–5. We conclude that domains 3–5 are necessary, but insuffi- [12] Trieber, C.A., Burkhardt, N., Nierhaus, K.H. and Taylor, D.E. cient, for the functional specificities of both factors. GTP (1998) Ribosomal protection from tetracycline mediated by hydrolysis can be decoupled from both factor-specific func- Tet(O): Tet(O) interaction with ribosomes is GTP-dependent. tions which, at least for EF-G, is rapidly triggered upon inter- Biol. Chem. 379, 847–855. [13] Spahn, C.M., Blaha, G., Agrawal, R.K., Penczek, P., Grassucci, action of domain 1 with components of the 50S ribosomal R.A., Trieber, C.A., Connell, S.R., Taylor, D.E., Nierhaus, K.H. subunit [25]. These conclusions raise two further possibilities: and Frank, J. (2001) Localization of the ribosomal protection (i) additional functional determinants may reside in domains protein Tet(O) on the ribosome and the mechanism of tetracycline 1–2 of the factors; and/or (ii) the interface between domains resistance. Mol. Cell 7, 1037–1045. 1–2 and domains 3–5 may be critical for coupling GTP hydro- [14] Martemyanov, K.A., Yarunin, A.S., Liljas, A. and Gudkov, A.T. (1998) An intact conformation at the tip of elongation factor G lysis to the correct placement of domain 4 in the ribosome [23]. domain IV is functionally important. FEBS Lett. 434, 205–208. Our present findings lead us to believe RPPs and EF-G are [15] Savelsbergh, A., Matassova, N.B., Rodnina, M.V. and Winter- more divergent functionally than we had anticipated. Their meyer, W. (2000) Role of domains 4 and 5 in elongation factor G functional specificities appear to be determined by multiple functions on the ribosome. J. Mol. Biol. 300, 951–961. [16] Wilson, K.S. and Noller, H.F. (1998) Mapping the position of determinants on each factor working together, allowing each translational elongation factor EF-G in the ribosome by directed factor to carry out its specific function on bacterial ribosomes. hydroxyl radical probing. Cell 92, 131–139. [17] Agrawal, R.K., Penczek, P., Grassucci, R.A. and Frank, J. (1998) Acknowledgements: We thank S. Connell for helpful discussions, B. Visualization of elongation factor G on the Escherichia coli 70S Hazes for FPLC access, and J. Simala-Grant and L. Friis for com- ribosome: the mechanism of translocation. Proc. Natl. Acad. Sci. ments on the manuscript. This work was supported by grants from Na- USA 95, 6134–6138. 1390 N.S. Thakor et al. / FEBS Letters 582 (2008) 1386–1390

[18] Connell, S.R., Trieber, C.A., Dinos, G.P., Einfeldt, E., Taylor, Giesebrecht, S., Dabrowski, M., Mielke, T., Fucini, P., Yokoy- D.E. and Nierhaus, K.H. (2003) Mechanism of Tet(O)-mediated ama, S. and Spahn, C.M. (2007) Structural basis for interaction of tetracycline resistance. EMBO J. 22, 945–953. the ribosome with the switch regions of GTP-bound elongation [19] Burdett, V. (1991) Puriﬁcation and characterization of Tet(M), a factors. Mol. Cell 25, 751–764. protein that renders ribosomes resistant to tetracycline. J. Biol. [24] Borowski, C., Rodnina, M.V. and Wintermeyer, W. (1996) Chem. 266, 2872–2877. Truncated elongation factor G lacking the G domain promotes [20] Thakor, N.S., Wilson, K.S., Scott, P.G. and Taylor, D.E. (2007) translocation of the 3Õ end but not of the anticodon domain of An improved procedure for expression and puriﬁcation of peptidyl-tRNA. Proc. Natl. Acad. Sci. USA 93, 4202–4206. ribosomal protection protein Tet(O) for high-resolution structural [25] Nechifor, R., Murataliev, M. and Wilson, K.S. (2007) Functional studies. Protein Expr. Purif. 55, 388–394. interactions between the GÕ subdomain of bacterial translation [21] Wilson, K.S. and Nechifor, R. (2004) Interactions of translational factor EF-G and ribosomal protein L7/L12. J. Biol. Chem. 282, factor EF-G with the bacterial ribosome before and after mRNA 36998–37005. translocation. J. Mol. Biol. 337, 15–30. [26] Martemyanov, K.A. and Gudkov, A.T. (2000) Domain III of [22] Laurberg, M., Kristensen, O., Martemyanov, K., Gudkov, A.T., elongation factor G from Thermus thermophilus is essential for Nagaev, I., Hughes, D. and Liljas, A. (2000) Structure of a mutant induction of GTP hydrolysis on the ribosome. J. Biol. Chem. 275, EF-G reveals domain III and possibly the fusidic acid binding 35820–35824. site. J. Mol. Biol. 303, 593–603. [27] Martemyanov, K.A. and Gudkov, A.T. (1999) Domain IV of [23] Connell, S.R., Takemoto, C., Wilson, D.N., Wang, H., Muray- elongation factor G from Thermus thermophilus is strictly required ama, K., Terada, T., Shirouzu, M., Rost, M., Schuler, M., for translocation. FEBS Lett. 452, 155–159.