Antibiotic that inhibits trans-translation blocks binding of EF-Tu to tmRNA but not to tRNA

ABSTRACT trans-Translation is conserved throughout bacteria and is essential in many species. High-throughput screening identified a tetrazole-based trans-translation inhibitor, KKL-55, that has broad-spectrum antibiotic activity. A biotinylated version of KKL-55 pulled down elongation factor thermo-unstable (EF-Tu) from bacterial lysates. Purified EF-Tu bound KKL-55 in vitro with a K d = 2 µM, confirming a high-affinity interaction. An X-ray crystal structure showed that KKL-55 binds in domain 3 of EF-Tu, and mutation of residues in the binding pocket abolished KKL-55 binding. RNA-binding assays in vitro showed that KKL-55 inhibits binding between EF-Tu and transfer-messenger RNA (tmRNA) but not between EF-Tu and tRNA. These data demonstrate a new mechanism for the inhibition of EF-Tu function and suggest that this specific inhibition of EF-Tu•tmRNA binding is a viable target for antibiotic development. IMPORTANCE Elongation factor thermo-unstable (EF-Tu) is a universally conserved translation factor that mediates productive interactions between tRNAs and the ribosome. In bacteria, EF-Tu also delivers transfer-messenger RNA (tmRNA)-SmpB to the ribosome during trans-translation. We report the first small molecule, KKL-55, that specifically inhibits EF-Tu activity in trans-translation without affecting its activity in normal translation. KKL-55 has broad-spectrum antibiotic activity, suggesting that compounds targeted to the tmRNA-binding interface of EF-Tu could be developed into new antibiotics to treat drug-resistant infections.

terminate on a reading frame within tmRNA, thereby rescuing the stalled ribosome (4).The tmRNA reading frame encodes a degradation signal that is appended to the nascent peptide chain, prompting rapid proteolysis of the released protein (10)(11)(12)(13).
Using a high-throughput screen of 663,000 candidate compounds, a small group of molecules were found to be effective trans-translation inhibitors (1).These mole cules inhibited trans-translation both in vitro and in vivo and exhibited broad-spectrum antibiotic activity against S. flexneri, Bacillus anthracis, and Mycobacterium smegmatis.KKL-55, a tetrazole-based compound, was one of the molecules that prevented Cterminal tagging and subsequent proteolysis of the nascent polypeptide.KKL-55 was later shown to be bactericidal to B. anthracis vegetative cells, germinants, and spores in vitro and after ex vivo infection of macrophages (14).Similarly, the minimum inhib itory concentration (MIC) of KKL-55 for Francisella tularensis was comparable to that of tetracycline (15).Notably, neither spontaneous nor UV-induced mutants of E. coli, S. flexneri, or N. gonorrhoeae that were resistant to KKL-55 could be recovered (1,2).Here, we show that KKL-55 inhibits trans-translation by binding elongation factor thermo-unstable (EF-Tu).
EF-Tu is an essential and universally conserved GTPase that is critical for protein synthesis.EF-Tu delivers aminoacyl-tRNAs (aa-tRNA) to the ribosomal aminoacyl (A) site during the elongation step of translation.Once the aa-tRNA docks to a cognate ribosomal A-site codon, EF-Tu exerts its GTPase activity, resulting in a conformational change that releases EF-Tu•GDP from the ribosome, and permits transfer of the nascent polypeptide from the P-site peptidyl-tRNA to the newly bound A-site aa-tRNA (16).Elongation factor thermo-stable (EF-Ts) binds EF-Tu and promotes the exchange of GDP with GTP, regenerating the EF-Tu•GTP complex to deliver another aa-tRNA (17).EF-Tu contains three structural domains.Domain 1 (amino acids 1-200) contains the GTPase center.Domains 2 (amino acids 209-299) and 3 (amino acids 301-393) together bind aa-tRNA; domain 3 also interacts with EF-Ts to promote nucleotide exchange (18).During trans-translation, EF-Tu delivers alanyl-tmRNA-SmpB to the empty ribosomal A site.SmpB binds to the mRNA channel and the ribosomal decoding center and facilitates the accommodation of tmRNA, so the nascent polypeptide chain can be transferred to its tRNA-like domain, and the tmRNA reading frame can be accepted as the new message (19).
Here, we perform structural and functional studies to understand how KKL-55 selectively inhibits trans-translation.We show that KKL-55 binds EF-Tu in vitro, and we solved a 2.2-Å X-ray crystal structure of the E. coli EF-Tu•KKL-55 complex, which shows that KKL-55 binds to a highly conserved pocket in domain 3.We demonstrate that the binding of KKL-55 has distinct effects on EF-Tu interaction with tRNA and tmRNA that explain the preferential inhibition of trans-translation.

EF-Tu is the molecular target of KKL-55
Previous modifications to the tetrazolyl benzamide KKL-55 suggested that large moieties could be added to the alkyl chain without impairing activity (14).To enable affinity purification of cellular molecules that bind to KKL-55, we designed KKL-201, an analog of KKL-55 that includes a biotin group (Fig. 1A).To ensure that KKL-201 retains the biochemical activity of KKL-55 and, therefore, is likely to bind the same target, we measured inhibition of trans-translation and translation by KKL-201 in vitro.In vitro transcription-translation assays using purified components from E. coli were program med with a gene-encoding DHFR with no stop codon at the 3′ end, and tmRNA-SmpB was added to the reactions.Transcription of the gene results in a nonstop mRNA, so translation and subsequent trans-translation result in a tagged DHFR protein (Fig. 1B).Like KKL-55, KKL-201 inhibited trans-translation in these reactions, resulting in a lower ratio of tagged to untagged protein.Neither KKL-55 nor KKL-201 inhibited protein synthesis when reactions were programmed with a gene encoding DHFR that included an in-frame stop codon (Fig. 1C).These results indicate that like KKL-55, KKL-201 specifically inhibits trans-translation and not normal translation.
To purify cellular molecules that bind to tetrazoyl benzamides, KKL-201 was incubated with a lysate from B. anthracis cells and the mixture was purified over Neutravidin resin.Affinity-purified proteins were visualized by SDS-PAGE, and inspection of the gel revealed two bands that appeared to be enriched in the eluate (Fig. 2A).Mass spectrometry identified these bands as pyruvate carboxylase and EF-Tu.Because pyruvate carboxylase contains a biotin co-factor which might allow it to bind to the Neutravidin resin inde pendently of KKL-201, we investigated EF-Tu as a potential target for KKL-55.

KKL-55 binds EF-Tu in vitro
We measured binding of KKL-55 with purified E. coli EF-Tu•GTP in vitro using microscale thermophoresis (MST) and observed concentration-dependent binding with an equili brium-binding constant (K d ) of 2.0 µM (Fig. 2B).No binding was observed between EF-Tu and the structurally distinct trans-translation inhibitor KKL-35 (Fig. S2), which has been shown to bind to the ribosome (2,8).EF-Tu•GDP bound KKL-55 with similar affinity as EF-Tu•GTP, indicating that the nucleotide state of EF-Tu is not important for KKL-55 binding (Fig. 2B).We measured the susceptibility of E. coli ∆tolC to KKL-55 using broth microdilu tion assays, and the MIC was 2.3 µM.The binding affinity for KKL-55 is, therefore, in the same range as the MIC, consistent with EF-Tu being the target responsible for growth inhibition by KKL-55.

KKL-55 binds domain 3 of EF-Tu
To determine the binding site of KKL-55, we solved a 2.2-Å X-ray crystal structure of EF-Tu co-crystallized with KKL-55 (Fig. 3; Fig. S4; Table 1).EF-Tu crystallized in the spacegroup P1 with two molecules per asymmetric unit.The structure was solved via molecular replacement using PDB code 6EZE as the starting model.The resulting difference F o −F c density allows for the placement of KKL-55 into domain 3 of EF-Tu.Domains 2 and 3 adopt β-barrel structures and form the binding surface for aa-tRNAs (Fig. S3), and domain 3 is also the binding site for EF-Ts, critical for nucleotide exchange.In our structure, KKL-55 binds EF-Tu domain 3 in a pocket formed by residues Gly317, Arg318, His319, and Glu378 (Fig. 3B).Arg318 is at the bottom of the binding pocket and packs against KKL-55 from the tetrazole to the benzylchloride, with the positively charged guanidino group oriented toward the electronegative arene ring.Gly317 and His319 form one side of the binding pocket, and Glu378 forms electrostatic interactions with the tetrazole ring of KKL-55 on the other side of the pocket (Fig. 3C).These EF-Tu residues are broadly conserved (Fig. 3D), suggesting that KKL-55 is likely to bind EF-Tu from many bacterial species.The propyl group on KKL-55 extends out of the binding pocket, consistent with the observation that large substitutions at this position do not compromise activity (14).The arene ring of KKL-55 places the meta-chlorine toward the solvent (positioned away from the EF-Tu pocket), suggesting that substitutions could be made on this ring to improve activity.
Based on the structure of EF-Tu-KKL-55, we made amino acid substitutions to Arg318, His319, and Glu378 and measured binding to KKL-55 using MST (Fig. 4).The R318A mutant had drastically lower affinity for KKL-55 (K d >30 µM), with a K d 15-fold higher than wild-type EF-Tu, consistent with the location of Arg318 in the KKL-55 binding pocket.Similarly, mutation of Arg318 to asparagine (R318N) dramatically reduced binding affinity (K d > 20 µM).However, the R318K mutant showed only a modest decrease in affinity (K d =2.6 µM), suggesting that the size and charge of the lysine residue allowed it to form similar interactions with KKL-55 as arginine at 318.The His319A had little impact on binding (Fig. 4), suggesting that the α-carbon of H319 is sufficient to form the pocket and the imidazole group is less important.Surprisingly, despite the electrostatic contacts between Glu378 and the tetrazole of KKL-55, the Glu378A mutation also had little effect on binding affinity (Fig. 4).The reason Glu378 appears to have no energetic contribution to binding is unclear, but in EF-Tu structures without KKL-55, Glu378 is oriented away from the KKL-55 binding pocket.Binding of KKL-55 results in the reorientation of Glu378 and compensating movement of the peptide backbone of residues 379 and 380 away from the pocket.This structural change might cost approximately the same energy gained by the electrostatic interaction with KKL-55, resulting in no net binding energy.Simultaneous mutation of Arg318, His319, and Glu378 to alanine abolished detectable binding with KKL-55 (Fig. S5).
Because the R318A mutant has substantially lower binding with KKL-55, cells with this version of EF-Tu might be resistant to KKL-55.However, Arg318 is strictly conserved in EF-Tu proteins across bacteria (Fig. 3D), so it might be important for other functions of EF-Tu.We were not able to construct an E. coli strain that had EF-Tu R318A or R318N as the only copy of EF-Tu in the cell, and linked marker co-transduction experiments confirmed that cells were not viable when the only copy of EF-Tu was the R318A or R318N mutant (Fig. S6).Conversely, cells could survive with EF-Tu R318K as the only copy of EF-Tu (Fig. S6).Consistent with the tight binding of the R318K mutant and KKL-55 observed in vitro, cells with R318K had the same MIC for KKL-55 as the wild type.These data indicate that Arg318 is important for viability in E. coli and explain why mutations at this position do not confer resistance to KKL-55.

KKL-55 binding alters EF-Tu binding to tmRNA
Because KKL-55 inhibits trans-translation but not normal translation, we hypothesized that KKL-55 alters or inhibits EF-Tu binding to tmRNA but has a smaller effect on the binding of EF-Tu to tRNA.To test this hypothesis, we used filter-binding assays to measure the impact of KKL-55 on binding between EF-Tu•GTP and Ala-tmRNA or Ala-tRNA Ala .In the absence of KKL-55, EF-Tu•GTP bound to Ala-tmRNA with a K d = 0.75 µM and EF-Tu•GTP bound to Ala-tRNA Ala with a K d = 0.28 µM (Fig. 5) similar to previously published data (20).In the presence of KKL-55, the binding affinity between EF-Tu•GTP and Ala-tmRNA was substantially reduced (Fig. 5A).Binding was weak enough that saturation could not be observed, but the K d >10 µM.Conversely, KKL-55 had little effect on the binding affinity between EF-Tu•GTP and Ala-tRNA Ala (Fig. 5B).This preferential inhibition explains how KKL-55 specifically inhibits trans-translation and not normal translation.

DISCUSSION
The data shown here demonstrate that KKL-55 binds EF-Tu at a site distinct from other antibiotics.Binding of KKL-55 to this site dramatically decreases the affinity of EF-Tu for tmRNA.Because EF-Tu is required for tmRNA-SmpB to efficiently interact with the ribosome, inhibition of EF-Tu binding to tmRNA results in inhibition of trans-translation.The conservation of EF-Tu residues in the KKL-55 binding pocket suggests that KKL-55 should inhibit trans-translation in most bacterial species.This inhibition, together with the requirement for trans-translation in many bacteria, is consistent with the broad-spec trum antibiotic activity of KKL-55 although we have not excluded the possibility that KKL-55 also has other cellular targets that contribute to antibiotic activity.Although the tRNA-like domain of tmRNA is very similar to a tRNA, EF-Tu binds to each RNA in subtly different ways at the elbow region, especially near the KKL-55 binding site in domain 3 (Fig. 6A).Overlaying EF-Tu•KKL-55 with tRNA and tmRNA reveals that KKL-55 binding to EF-Tu may alter interactions with each RNA substrate in slightly different ways.In the model of EF-Tu•tRNA•KKL-55, the arene end of KKL-55 would directly clash with the phosphate backbone of nucleotide 52 of the accepter arm of tRNA.Despite a predicted steric clash between KKL-55 and tRNA, KKL-55 has no effect on binding of EF-Tu to tRNA.The steric clash covers a single phosphate of the tRNA backbone, and presumably domain 3 of EF-Tu can move to prevent the clash without broadly disrupting other contacts with tRNA that make an important energetic contribution to binding.The same movement is clearly not possible when tmRNA is bound.Three factors may contribute to this difference.First, the total area of the steric clash is much larger for tmRNA (217.5 vs 140.4 Å 2 ) (Fig. 6B).Second, the clash covers two nucleobases of tmRNA (nucleotides 339 and 340) near the middle of the tmRNA acceptor arm axis, which might require a much larger adjustment to avoid.Third, the paths of the tRNA and tmRNA backbones are different in this region, with tmRNA passing much closer to KKL-55 (Fig. 6; Fig. S7).The closer proximity of tmRNA to EF-Tu and KKL-55 might sterically constrain the movement of domain 3 or force disruption of more contacts between EF-Tu and tmRNA to avoid the clash with KKL-55.The tight binding of EF-Tu to tRNA in the presence of KKL-55 and the lack of inhibition of translation in vitro after the addition of KKL-55 confirm that KKL-55 specifically disrupts trans-translation and not translation.
The subtle but mechanistically significant differences in how KKL-55 affects EF-Tu interaction with tmRNA and tRNA parallel subtle mechanistic differences in how EF-Tu delivers tRNA and tmRNA to the ribosome.During normal translation, EF-Tu•GTP•aa-tRNA enters the A site, and if there is no cognate base pairing between the tRNA anticodon and A-site codon, GTP hydrolysis is slow allowing EF-Tu•GDP•aa-tRNA to rapidly dissoci ate (21,22).If the tRNA anticodon has cognate base pairing with the A-site codon, GTP hydrolysis on EF-Tu is rapid, and EF-Tu•GDP dissociates, leaving the aa-tRNA accommoda ted in the A site (16).Delivery of Ala-tmRNA-SmpB does not appear to be strictly coupled to GTP hydrolysis on EF-Tu.Experiments with translating ribosomes that have mRNA extending 3′ past the leading edge of the ribosome show they are not substrates for trans-translation (23).However, when EF-Tu•GTP•Ala-tmRNA-SmpB enters these nonsubstrate ribosomes, GTP hydrolysis is stimulated on EF-Tu even though Ala-tmRNA-SmpB is not accommodated into the A site (24).In addition, mutations in SmpB that reduce the rate of GTP hydrolysis on EF-Tu do not affect the affinity of Ala-tmRNA-SmpB for the ribosomal A site, suggesting that Ala-tmRNA-SmpB is released from EF-Tu more easily than tRNAs.Likewise, kirromycin appears to have differential activity between trans-translation and normal translation (25).Specifically, kirromycin does not inhibit EF-Tu•Ala-tmRNA-SmpB dissociation from the ribosome in the same way it inhibits EF-Tu•aa-tRNA dissociation during canonical elongation, supporting a different EF-Tu activation pathway for EF-Tu•Ala-tmRNA-SmpB as compared to EF-Tu•aa-tRNA (25).These data suggest that EF-Tu binds Ala-tmRNA and aa-tRNA differently during interaction with the ribosome, and the specific inhibition of EF-Tu•Ala-tmRNA binding by KKL-55 indicates that these differences are present before interaction with the ribosome as well.
EF-Tu is targeted by several antibiotics other than KKL-55.EF-Tu-binding antibiotics, also called elfamycins, primarily inhibit canonical translation and can be broadly categorized into two classes based on how they inhibit EF-Tu activity (26).The first group, consisting of pulvomycin and GE2270A, binds in the cleft between domains 1 and 2 of EF-Tu to prevent the essential conformation switch necessary for binding aa-tRNAs (27,28).The second group, consisting of kirromycin and enacyloxin IIa, binds to the interface between domains 1 and 3 of EF-Tu, preventing EF-Tu from adopting its "open" conforma tion that is necessary to dissociate from the ribosome (29,30).The ability of KKL-55 to specifically target trans-translation by binding domain 3 of EF-Tu demonstrates a new mode of action by elfamycins and indicates that this binding site on EF-Tu could be targeted for the development of new antibiotics to kill bacteria that require transtranslation.

Reagents
Bacterial strains, plasmids, and primers used in this study are described in Tables S1 to S3.

Synthesis of KKL-201
The synthetic scheme is shown in Fig. S9.Compounds 1-4 were synthesized as previously described (14).All organic solvents and reagents were purchased from Sigma-Aldrich (St. Louis, MA, USA) unless otherwise stated.Chloroform-d was purchased from Cambridge Isotope Laboratories (Andover, MA, USA).Nuclear Magnetic Resonance analyses were conducted on a 400-MHz Bruker spectrophotometer.Triethylamine (551 mg, 5.5 mmol) was added to the flask containing a mixture of compound 4 (300 mg, 1.09 mmol), and biotinyl-N-hydroxysuccinimide (440 mg, 1.3 mmol) dissolved in 15 mL of dimethylformamide.This reaction mixture was stirred at room temperature for 18 h under an argon atmosphere.The resultant mixture was filtered and concentrated under reduced pressure.The crude product was purified by flash chromatography over silica gel using MeOH:CH 2 Cl 2 (1:20) as the mobile phase to obtain KKL-201 as a whitish powder in (recovery 78%).

In vitro translation and trans-translation assays
Translation assays were performed as previously described (1).Assays were set up using PURExpress in vitro protein synthesis kit (New England Biolabs, Ipswich, MA, USA) with a cloned full-length DHFR template, and protein synthesis was monitored by incorporation of 35 S-methionine.The inhibition of translation activity for each test compound was assessed with respect to the vehicle control from at least three independent assays.In vitro trans-translation was measured in a similar reaction that contained tmRNA-SmpB and employed a DHFR template missing two bases from the stop codon, as previously described (1).Due to the formation of a non-stop complex, tmRNA-SmpB can introduce an 11 amino acid tag on the DHFR protein which can be distinguished from the untagged DHFR protein on a SDS-PAGE gel.Test compounds were analyzed at 10 µM unless otherwise stated.Tagging efficiency was evaluated as the ratio of tagged DHFR to total DHFR from at least three repeats.

Minimum inhibitory concentration assay
MIC assays were performed by broth microdilution according to Clinical and Labora tory Standards Institute guidelines for determining the antimicrobial activity of the compounds as described previously (1).

Affinity chromatography
B. anthracis cells were grown in 5 mL lysogeny broth (LB) at 37°C overnight.This culture was grown in 1 L LB to a final OD 600 ~ 1.2.Cells were harvested by centrifugation at 14,000 × g for 10 min, re-suspended in 25 mL lysis buffer [20 mM Tris pH 7.5, 2 mM β-mercaptoethanol (β-Me), 1 mg/mL lysozyme], and lysed by sonication, and cell debris was removed by centrifugation at 28,000 × g.The lysate was concentrated using a 10K Amicon ultra centrifugal filter.Five hundred microliters of concentrated lysate was added to 500 µL KKL-201 (400 µM) in binding buffer (100 mM Tris-HCl pH 8.0, 150 mM NaCl, 1 mM EDTA).This mixture was gently shaken at 4°C for 2 h, added to NeutrAvidin agarose resin (ThermoFisher, Waltham, MA, USA) equilibrated in binding buffer, incubated for 2 h, and loaded onto a column.The column was washed with 10 volumes binding buffer, and bound proteins were eluted with 8 M guanidine hydrochloride.Fractions containing protein were combined, dialyzed against binding buffer, and concentrated using a 3K

Apo-EF-Tu and EF-Tu•GTP preparation
Purified GDP-bound 6His-tagged EF-Tu was incubated in buffer C (25 mM Tris-HCl pH 7.5, 50 mM NH 4 Cl, 10 mM EDTA) for 10 min at 37°C.Apo EF-Tu and GDP were separated by size exclusion chromatography on a Biorad ENRich SEC 70 column equilibrated in buffer F (25 mM Tris-HCl pH 7.5, 50 mM NH 4 Cl).GTP was added to 20 µM final concentration immediately after purification.

EF-Tu purification for structural studies
EF-Tu was overexpressed from E. coli BL21-Gold (DE3) cells containing the pQE60-tufA-6xHis plasmid.Overnight cultures were grown in LB supplemented with 100 µg/mL ampicillin and 10 µg/mL tetracycline and diluted 1:500 into 4 L media.The bacterial culture was grown at 37°C, 220 rpm to OD 600 ~ 0.6.Overexpression of EF-Tu was induced by the addition of 0.

Structural determination of the EF-Tu•KKL-55 complex
Crystallization trials of 11 mg/mL EF-Tu were performed using the Phoenix protein crystallization robot (Arts Robbins Instruments, Sunnyvale, CA, USA) using a 400 nL drop size with a 1:1 ratio of protein to reservoir condition in Intelli-Plate 96-3 LVR plates.
The mixture was then centrifuged to pellet any precipitation for 3 min at 20,627 × g, and the supernatant was used to set up sitting drops using 3.6 µL drop size in a 1:1 protein to reservoir ratio over 400 µL of reservoir volume at 20°C.Needle-like crystals grew within 1-2 days and were flash frozen after being cryoprotected stepwise with solutions of 5%/15%/25% PEG 400, 0.2 M (NH 4 ) 2 (SO 4 ), 20% PEG 5K MME, and 0.1 M MES pH 6.5.Data sets were integrated and scaled using XDS (31) and the North eastern Collaborative Access Team (NE-CAT) RAPD system.The structure was phased using molecular replacement with the EF-Tu model from PDB code 6EZE (32), and the model was iteratively refined in PHENIX (33) and built in Coot (34).The KKL-55 model was generated in ChemDraw, and refinement restraints were generated using eLBOW in PHENIX (35).Feature-enhanced maps were generated in PHENIX (36).Figures were generated in PyMOL.Surface area calculations were performed in ChimeraX (37).

Conservation analysis of EF-Tu and KKL-55 binding pocket
Conservation analyses of the KKL-55 binding pocket were performed on ~250 prokary otic EF-Tu protein sequences identified using BLASTp with the sequence from 6EZE against the UniProtKb protein sequence database (38).Amino acid sequences were downloaded from UniProt and duplicate sequences, and those of unknown function were removed.Protein sequences were aligned using MUSCLE (39).A frequency logo was created from aligned EF-Tu sequences using WebLogo (40).ConSurf analysis was performed with conservation score determined by Bayesian inference (41) (Fig. S8).

Site-directed mutagenesis of EF-Tu
All mutants were constructed using HiFi assembly (New England Biolabs).Two PCR products were generated for each mutant using primer pairs listed in Table S3, with pCA24N-His6-tufA as the template.The PCR products were assembled with the pCA24N-His6-tufA that had been digested with BamHI and NotI.

Microscale thermophoresis-binding assays
Three hundred nanomolars of 6His-tagged EF-Tu was incubated with 75 nM Red Tris NTA dye Generation 2 (NanoTemper Technologies, Watertown, MA, USA) in binding buffer (287 mM NaCl, 2.7 mM KCl, 10 mM Na 2 HPO 4 , 1.8 mM KH 2 PO 4 , 0.01% Tween 20) for 30 min followed by centrifugation at 21,000 × g for 10 min.The EF-Tu and dye complex was added to KKL-55 in a 1:1 ratio and incubated at room temperature for 2 h, and MST was measured in a Monolith NT.115 (NanoTemper Technologies).Plots of change in fluorescence vs the concentration of KKL-55 were fit to the hyperbolic function y = c/(1 + K d /x) to obtain the apparent binding constant.

Phage transduction
Tet R -tufA::Kan linked marker strain (KCK575) was constructed via recombination (42) of Tet R into the chromosome of tufA deletion strain, 15.6 kb away from tufA::Kan, and a P1 phage lysate (43) was prepared and used to transduce ΔtolC ΔtufB cells harbor ing pCA24N-His6-tufA, pCA24N-His6R318A, or pCA24NHis6R318N.Cells were plated on LB with oxytetracycline, chloramphenicol, and IPTG to select for transductants.The resulting colonies were tested for kanamycin resistance to determine the co-transduc tion frequency.

FIG 1 FIG 2
FIG 1 Tetrozoyl benzamides specifically inhibit trans-translation.(A) Chemical structure of KKL-55 and KKL-201.(B) In vitro trans-translation reactions in the presence of DMSO, 10 µM KKL-55, or 10 µM KKL-201.A gene-encoding DHFR with no stop codon was transcribed and translated in the presence of tmRNA-SmpB and a tetrazolyl benzamide or vehicle.Protein products were labeled by incorporation of 35 S-Met and analyzed by SDS-PAGE followed by autoradiography.A representative gel is shown with bands corresponding to DHFR and tagged DHFR indicated.The intensity of the DHFR and tagged DHFR bands was quantified, and the tagging efficiency was calculated as the percentage of total DHFR protein in the tagged DHFR band.Mean tagging efficiency with standard deviation for at least three biological repeats is shown.(C) In vitro translation reactions in the presence of DMSO, 100 µM chloramphenicol (chlor), 100 µM KKL-55, or 10 µM KKL-201.A gene-encoding DHFR with stop codon was transcribed and translated as in panel B but without the addition of tmRNA-SmpB.A representative gel is shown with the band corresponding to DHFR indicated.

FIG 4 FIG 5
FIG 4 EF-Tu Arg318 is important for binding KKL-55.MST-binding assays were performed with mutant versions of EF-Tu•GTP.One representative binding curve for each protein is shown.Data from three repeats were averaged and fit to the sigmoidal function to determine the dissociation constant.The mean dissociation constant with standard deviation for at least three repeats is shown for each protein.Data for wild-type (WT) EF-Tu from Fig. 2B are shown for comparison.

FIG 6
FIG 6 Comparison of EF-Tu binding interfaces with tRNA, tmRNA, and KKL-55.(A) Overlay of three structures of EF-Tu bound to tRNA (PDB code 1TTT), EF-Tu bound to tmRNA (PDB ID 7ABZ, EF-Tu in this structure not shown for clarity), and modeled KKL-55 as found in our structure in this study.Domains 2 and 3 of EF-Tu were used for alignments since domain 1 is in the "open, " GDP-bound conformation.Differences between tRNA and tmRNA upon EF-Tu are localized to the acceptor arm and specifically tRNA nucleotides 59-63 and tmRNA nucleotides 348-351.The interaction between tRNA and EF-Tu at this region is further apart as compared to the tmRNA-EF-Tu interaction (as denoted by the red arrow and text).(B) Surface representation of tRNA and tmRNA with KKL-55 (from PDBs above).The predicted clash area is larger between KKL-55 and tmRNA (217.5 Å 2 ) than KKL-55 and tRNA (140.4Å 2 ) consistent with stronger KKL-55 inhibition of tmRNA binding to EF-Tu.

TABLE 1
Data collection and refinement statistics for the crystal structure of EF-Tu bound to KKL-55 a a Values in parentheses are for highest-resolution shell.