Negamycin Binds to the Wall of the Nascent Chain Exit Tunnel of the 50S Ribosomal Subunit

Abstract

Negamycin, a small-molecule inhibitor of protein synthesis, binds the Haloarcula marismortui 50S ribosomal subunit at a single site formed by highly conserved RNA nucleotides near the cytosolic end of the nascent chain exit tunnel. The mechanism of antibiotic action and the function of this unexplored tunnel region remain intriguingly elusive.

Negamycin is a translation inhibitor that prevents the release of complete proteins from the ribosome and thus causes an accumulation of polysomes. Negamycin also induces miscoding (1, 15, 23, 26) but does not interfere with peptide initiation or elongation (7, 25, 28). Because the inhibition of peptide release and miscoding vary independently in negamycin derivatives (27), negamycin may interfere with ribosome function in two distinct ways. Here we report the results of crystallographic experiments which show that negamycin binds to the wall of the peptide exit tunnel of the large ribosomal subunit. This interaction may account for negamycin's inhibition of peptide release.

The 50S ribosomal subunits were prepared from wild-type Haloarcula marismortui cells and crystallized as described previously (3). Negamycin was soaked into the crystals following previously described protocols (11, 16, 20, 21). Five millimolar negamycin is approximately five times the MIC of negamycin for this organism. Crystals were equilibrated at least twice with solutions containing 5 mM negamycin dissolved in buffer B (12% [wt/vol] polyethylene glycol 6000, 20% [vol/vol] ethylene glycol, 1.7 M NaCl, 0.5 M NH₄Cl, 1 mM CdCl₂, 100 mM potassium acetate, 6.5 mM acetic acid [pH 6.0], 30 mM MgCl₂ or 100 mM SrCl₂, 5 mM β-mercaptoethanol) at 4°C over the course of 1 to 4 h, during which SrCl₂ was included in soak solutions, and over 16 to 24 h, during which MgCl₂ was included in soak solutions.

Data collected at the Advanced Light Source (ALS), using beamline 8.2.2, were processed as previously described (3). Unbiased-difference electron density maps were computed for each drug-ribosome complex, using as amplitudes [|F_o(hkl)|_{drug/ribosome} − |F_o(hkl)|_{best native}], where the |F_o(hkl)|s are measured amplitudes of the (hkl) reflection in the data collected from the two crystals whose structures were being compared, and phases obtained by rigid-body refinement of the native structure (Protein Data Bank code 1S72; 12) into the drug-ribosome data set. Structure refinements were done with CNS software (4) and included rigid-body, energy minimization, and B-factor refinement. In the final rounds of refinement, only those atoms within a 25- to 35-Å-diameter sphere centered on the antibiotic binding site were allowed to vary in position. The coordinates and structure factors for the negamycin complex described here may be found in the Protein Data Bank (code 2QEX; http://www.rcsb.org/pdb/home/home.do).

The functional significance of the negamycin binding site identified in the crystal structure was investigated by site-directed mutagenesis and random targeted mutagenesis and by the selection of naturally occurring mutations in Escherichia coli, and Thermus thermophilus. A pUC derivative was created containing a fragment of the E. coli rrnB cistron obtained from pLK35 between restriction sites for SphI and XbaI. It includes all of domain 1 and part of domain 2 of the 23S rRNA. Site-directed mutations were introduced into the pUC construct using a QuikChange kit (Stratagene) and following the manufacturer's protocol. Random mutations were introduced using error-prone PCR (5). These mutated DNA fragments were used to replace the corresponding sequences in pLK35, and the resulting plasmids were transformed into E. coli POP2136, which carries a temperature-sensitive λ repressor. Mutated plasmids were propagated at 30°C and assayed for viability and negamycin sensitivity at 42°C. Mutant pLK35 plasmids were also transformed into the Squires strain SQZ10 or SZ7, and sucrose selection was used to exchange the mutated plasmids for the plasmid carrying the single wild-type rRNA cistron on which these strains depend for survival (2; S. Qwan, personal communications).

Selections for naturally occurring mutations were performed with the E. coli Squires strains SQZ10 and SZ7 at 25 μg/ml negamycin, which is twice the MIC, and with T. thermophilus strains HB8 and IB21 at 100 μg/ml or 200 μg/ml negamycin or using 250 μg disks under growth conditions that were otherwise standard for both organisms (2, 10; S. Qwan, personal communication). The MIC of negamycin for H. marismortui was measured in liquid culture and proved to be so high that resistance mutations could not be selected in H. marismortui because the amount of drug available was too small.

Difference electron density maps computed using data obtained from two different crystals at resolutions of around 3 Å show a single positive feature large enough to accommodate the drug. These maps indicate that the drug binds to a cleft in the wall of the polypeptide exit tunnel, close to its cytosolic end, in an extended conformation (Fig. 1A; Table 1). However, the maps' resolutions are not high enough to determine the orientation of the drug in its binding site unambiguously. The orientation shown here was chosen because it maximizes the number of hydrogen bonds the drug forms with the ribosome and avoids electrostatic repulsions (Fig. 1D). It is also consistent with structure-activity relationship studies, which show the importance of the hydroxyl group, the stereochemistry at position 3, and both termini of negamycin for its activity (12, 13, 17-19, 27). The drug interacts primarily with RNA backbone atoms. The N1 and N4 amino groups of the drug hydrogen bond to the nonbridging phosphate oxygens of HmG518(Ec512) and HmU517(Ec511) (where Hm and Ec refer to H. marismortui and E. coli, respectively, the letter after “Hm” identifies the nucleotide type, and the number after “Hm” is the position of that nucleotide in the Hm 23S RNA sequence and the number in parenthesis is the corresponding sequence number in Ec), and its O4 and N3 atoms interact with a cation, possibly a magnesium ion, that has two ribosomal ligands, the O2P oxygen of HmU517(Ec511) and the N7 of HmG518(Ec512).

FIG. 1.

(A) The position of the negamycin bound in the peptide exit tunnel. Negamycin, in yellow, is displayed on one half of the large subunit, which is shown sliced along the lumen of the peptide exit tunnel. The ribosomal proteins are displayed in blue, rRNAs in gray, and a modeled P site tRNA in orange. (B) Difference electron density map for negamycin contoured at 4 sigma. (C) Chemical structure of negamycin. (D) Interactions of negamycin with rRNA and a metal ion (green) at the binding site in the peptide tunnel.

TABLE 1.

Crystallographic statistics

Phase	Parameter	Data or range (upper limit)a
Data collection	Negamycin resolution (Å)	50.0-3.22 (3.32-3.22)
	I/σb	10.2 (2.7)
	Rmergec	0.132 (0.81)
	Completenessd	99.7 (100)
	Rede	6
Model refinement	Negamycin resolution (Å)	50.0-3.20
	Rcrystf	0.204
	Rfreeg	0.262
	Bond (Å)h	0.006
	Angle (°)i	1.06
	PDB no.	2QEX

^aResolution range of X-ray diffraction data. The resolution range of the highest-resolution data bin included in the data is indicated in parentheses. The upper-limit statistic is the single value (in parentheses) for that highest-resolution bin.

^bI/σ, average intensity/average standard error of the intensity.

^cR_merge = {∑_hkl∑_iI_i(hkl) − i(hkl)}/{∑_hkl∑_iI_i(hkl)}, where i(hkl) is the average value of the intensity of reflection (hkl) in the data set, and I_i(hkl) is the intensity of the ith observation of that reflection.

^dCompleteness, percentage of the possible number of reflections that was actually observed.

^eRed, average redundancy of the data set.

^fR_cryst = {∑_hklF_c(hkl) − F_o(hkl)/{∑_hklF_o(hkl)}, where F_c(hkl) is the structure factor for the reflection (hkl) computed from the structural model, and F_o(hkl) is the structure factor measured for that reflection.

^gR_free is computed the same way as R_cryst, using a small but randomly chosen subset of measured reflections that is not included in the data set used for refinement.

^hBond, the root-mean-squared differences between the actual and expected bond lengths.

ⁱAngle, the root-mean-squared differences between the actual and expected bond angles in degrees.

The inhibitor binds to the irregularly structured region in domain 1 of the 23S rRNA that connects helices 24 and 2 [HmA509(Ec503)-HmG512(Ec506); and HmC515(Ec509)-HmG518(Ec512)]. The sequence in this region is >99% conserved across all species (6, 22), consistent with negamycin's lack of species specificity and also consistent with the hypothesis that this region is functionally significant. However, its function is unknown. These nucleotides are unreactive in chemical modification experiments (8), and no mutations have been reported in this region (24). However, low-resolution cryoelectron microscopy images suggest that this part of the 23S rRNA could be dynamic (7, 14).

To test the significance of the site, E. coli mutant strains were prepared in which EcG512(Hm518) was changed to U or deleted. Both strains retained sensitivity to negamycin, even though EcG512 is an important component of its binding site. Deletion of either Ec510-512 or Ec508 (Hm516-518 or Hm512) provided no insight because both were dominant lethal mutations. Finally, no rRNA mutations that conferred negamycin resistance emerged from selection experiments carried out using a single-operon strain of E. coli (2) or the HB9 and I21 strains of T. thermophilus (10). Finally, we note that the lumen of the tunnel is so large in the region where the drug binds that its binding should not interfere sterically with the passage of nascent polypeptides down the tunnel. Thus, further work will have to be done to determine the functional significance of the negamycin binding site reported here and to elucidate the mechanism of action of negamycin.