Arch Microbiol (2005) 184: 66–77 DOI 10.1007/s00203-005-0017-0
O R I GI N A L P A P E R
Jessica A. Silvers Æ W. Scott Champney
Accumulation and turnover of 23S ribosomal RNA in azithromycin-inhibited ribonuclease mutant strains of Escherichia coli Received: 25 March 2005 / Revised: 24 May 2005 / Accepted: 30 May 2005 / Published online: 12 August 2005 Springer-Verlag 2005
Abstract Ribosomal RNA is normally a stable molecule in bacterial cells with negligible turnover. Antibiotics which impair ribosomal subunit assembly promote the accumulation of subunit intermediates in cells which are then degraded by ribonucleases. It is predicted that cells expressing one or more mutated ribonucleases will degrade the antibiotic-bound particle less efficiently, resulting in increased sensitivity to the antibiotic. To test this, eight ribonuclease-deficient strains of Escherichia coli were grown in the presence or absence of azithromycin. Cell viability and protein synthesis rates were decreased in these strains compared with wild type cells. Degradation of 23S rRNA and recovery from azithromycin inhibition were examined by 3H-uridine labeling and by hybridization with a 23S rRNA specific probe. Mutants defective in ribonuclease II and polynucleotide phosphorylase demonstrated hypersensitivity to the antibiotic and showed a greater extent of 23S rRNA accumulation and a slower recovery rate. The results suggest that these two ribonucleases are important in 23S rRNA turnover in antibiotic-inhibited E. coli cells. Keywords Azithromycin Æ 50S ribosomal subunit Æ 23S rRNA Æ Ribonucleases Æ E. coli Æ Ribosomes Æ Antibiotic inhibition
Introduction The bacterial ribosome is an ancient structure and is an important target for numerous antimicrobial agents.
J. A. Silvers Æ W. S. Champney (&) Department of Biochemistry and Molecular Biology, J.H. Quillen College of Medicine, East Tennessee State University, Johnson City, TN 37614, USA E-mail:
[email protected] Tel.: +1-423-4392022 Fax: +1-423-4392030
Different antibiotics interfere with the process of translation by inhibiting different steps in the complicated process of protein biosynthesis (Vazquez 1979). Ribosomal RNA is the specific binding site for many of these compounds (Schroeder and Wallis 2001). Recent crystal structures of both 30S and 50S subunits with bound antibiotics have revealed precise interactions between different antimicrobial agents and parts of the RNA sequence in each particle (Broderson et al. 2000; Carter et al. 2000; Schlunzen et al. 2001; Hansen et al. 2002). It has been suggested from these studies that many of these compounds inhibit an essential conformational change in the particle, locking the structure into an irreversible conformation. Resistance to many commonly used translational inhibitors is increasing and new drug structures and targets need to be identified (Anderson 1999). In some clinical isolates, a change in a rRNA sequence is the result of a resistance mutation which reduces antibiotic binding to the ribosome (Canu et al. 2002). Knowledge of the mechanism of action of translational inhibitors is needed in order to better understand how these drugs function and how to design better antimicrobial agents (Chu et al. 1996). In addition to an inhibitory effect on translation, many ribosomal antibiotics also prevent the formation of the subunits in a specific fashion (reviewed in Champney 2003). Binding of antibiotics to rRNA during subunit assembly stalls the process and leads to the breakdown of rRNA in the incomplete precursor particle. Erythromycin has been shown to affect 50S synthesis specifically (Usary and Champney 2001) and both neomycin and paromomycin stop 30S subunit formation in cells (Mehta and Champney 2002, 2003). For several different antimicrobial agents we have shown a concentration dependent decline in 23S rRNA in mature particles and an increase in the accumulation of partially degraded rRNA oligonucleotides in inhibited cells (Champney and Burdine 1996, 1998; Champney and Miller 2002). A model outlining the steps in this process for the 50S subunit is illustrated in Fig. 1. In the absence
67 Fig. 1 Model for 50S subunit formation and inhibition by azithromycin. a Pathway for 50S subunit biosynthesis in E. coli cells. b Azithromycin inhibition of 50S formation, with accumulation of an assembly intermediate as a substrate for cellular ribonucleases. E antibiotic bound to existing 50S subunits; I antibiotic bound to a 50S subunit precursor
of the antibiotic, formation of the 50S subunit proceeds through two transient intermediate precursor particles before forming a functional 50S ribosomal subunit (Fig. 1a). The first intermediate sediments at 32S and is composed of certain ribosomal proteins, 23S rRNA, and 5S rRNA. The addition of other ribosomal proteins and a conformational change generates the second particle which sediments at 43S. Binding of all 34 proteins to the 23S and 5S rRNA and another conformational change results in a functional 50S ribosomal subunit (Nierhaus 1982). Antibiotic binding to an intermediate particle stalls the process and prevents normal 50S formation (Fig. 1b). The model predicts that cellular ribonucleases are needed for removal of the stalled precursors from inhibited cells. Identification of these enzymes would help to explain the mechanism of action of ribosomal antibiotics in preventing subunit assembly. We have shown previously that a strain of E. coli defective in the enzyme ribonuclease E was hypersensitive to erythromycin inhibition and accumulated more 50S precursor than a wild type strain (Usary and Champney 2001). This work was initiated to investigate the role of different cellular ribonucleases in the turnover of the antibiotic stalled intermediate of the 50S subunit generated by the azalide antibiotic azithromycin.
Azithromycin was chosen because it has a lower MIC for inhibition in E. coli than erythromycin (Zuckerman and Kaye 1995), it’s binding sites on the 50S subunit are known from crystallography (Hansen et al. 2002) and it is an inhibitor of 50S subunit assembly in this organism (Chittum and Champney 1995). Also, its binding to the 50S subunit of E. coli is stronger than the binding of erythromycin (Dinos et al. 2001). Eight mutant strains with defects in ribonuclease activity were examined for azithromycin sensitivity and aspects of 23S rRNA degradation. Hypersensitivity to azithromycin was determined by examining growth properties, protein synthesis rates, and 50S subunit amounts after azithromycin exposure. Recovery of the 50S subunit after azithromycin exposure and 23S rRNA accumulation were also examined. The enzymes RNase II and polynucleotide phosphorylase were found to be important in 23S rRNA turnover in azithromycin inhibited cells.
Methods and methods Materials The strains of Escherichia coli used in this study are listed in Table 1. Azithromycin was a gift from Pfizer
Table 1 Strains of Escherichia coli used Strain
Phenotype
Genotype
Reference
SK901 MRE 600 D10 SK7622 SK5665 CA244 SK4803 N7060 SK5704
wild type RNase I RNase I RNase III RNase E PNPase RNase II RNase I RNase II PNPase RNase II RNase E PNPase
F- malA thirna HfrH met- rna-1 relA F- thyA715 rncD38::kanR F- thyA715 rne1 F- Dpnp::kanR F- gal thi ton sup hasdR4 endAsbcB15 rnb296 F- metB1 tryA451 rpsl478 rna919rnb464 pnp13 F- thyA715 pnp7 rne1 rnb500 rph1
(Kushner et al. 1977) (Cammack and Wade 1965) (Gesteland 1966) (Babitzke et al. 1993) (Arraiano et al. 1988) (Reuven and Deutscher 1993) (Donovan and Kushner 1986) (Weatherford et al. 1972) (Arraiano et al. 1988)
68
Pharmaceuticals. 3H-uridine (1.0 mCi/mmol) and 35Smethionine (Tran35S-Label, 1,175 Ci/mmol) were purchased from ICN Pharmaceuticals.
MIC determination and analysis of cell growth and viability The minimal inhibitory concentration (MIC) for azithromycin in each mutant strain was determined by a dilution method as described (Champney et al. 1998). Cells were grown in a water bath at 32C in tryptone broth (TB) in the presence or absence of azithromycin at 2.5 lg/ml except for MRE600 and SK5704 (1 lg/ml). Media was supplemented with 50 lg/ml of thymine for strains SK5665, SK7622, and SK5704. Growth rates were measured by recording the increase in cell density over time using a Klett-Summerson colorimeter as previously described (Chittum and Champney 1995). Cell viability measurements were determined after serial dilution of the cells as described (Jett et al. 1997).
Analysis of the rate of protein synthesis Cell cultures were grown as described above. After two cell doublings in the presence or absence of azithromycin, 35S-methionine was added to the culture at a concentration of 1 lCi/ml. Three samples of 0.4 ml were removed at 5 min intervals after addition of the 35Smethionine. Cells were precipitated in 20% TCA, collected and washed on Whatman GF/A glass fiber filters. Radioactivity was measured by liquid scintillation counting.
Analysis of ribosomal subunit assembly Cells were grown to a Klett reading of 20 (1·108 cells/ ml) and azithromycin was added to the appropriate cultures. After 10 min of growth, 3H-uridine (1 lCi/ml) was added to the control and azithromycin treated samples. Following two cell doublings, an excess of uridine (50 lg/ml) was added to each culture. After a 15 min chase period, cells were collected by centrifugation, washed and the cell pellets were stored at 70C before cell lysis. Cell lysates were prepared by a lysozyme, freeze-thaw procedure as described (Chittum and Champney 1995). The samples were centrifuged through 5–20% sucrose gradients in S buffer (Champney et al. 1998) in a SW40 rotor at 39,000 rpm for 4.5 h or at 20,000 rpm for 18 h. After centrifugation, gradients were pumped through an ISCO Model UA-5 absorbance monitor set at 254 nm. Equivalent fractions of the gradient were collected into vials and mixed with 3 ml of Scintisafe gel. The incorporation of 3H-uridine into RNA was measured by liquid scintillation counting.
Analysis of 50S subunit recovery following azithromycin exposure Strains were grown with azithromycin and labeled with 2 lCi/ml 3H-uridine as described above. At a Klett reading of 80 the cells were centrifuged at 6,000 rpm for 12 min at room temperature. Cells were washed with 1 ml of TB at 32C and the cell pellets were resuspended in 20 ml of TB supplemented with 50 lg/ml uridine. Growth was resumed at 32C and 3 ml samples were removed from the culture at 5, 15, 30, 60, 90, and 120 min. Samples were centrifuged for 12 min at 6,000 rpm and pellets were stored at 70C. Cell lysis, centrifugation on 5–20% sucrose gradients, and gradient collection was performed as described above. RNA was isolated from samples at 60 and 120 min for slot blot analysis as described (Usary and Champney 2001). Construction of biotinylated 16S and 23S specific probes The polymerase chain reaction was used to amplify the 16S (241 bp) and 23S (146 bp) probes from plasmid pKK3535 DNA (Brosius et al. 1980). PCR reaction mixtures contained 45 ll PCR Supermix High Fidelity reagent (Gibco BRL), 1 ll of plasmid DNA (6.5 ng), 1 ll (10 pmol) of either the 16S or 23S forward primer, 1 ll (10 pmol) of either 16S or 23S reverse primer, and 2 ll sterile H2O. Primers used were 23S F: TAGGGGAGCGTTCTGTAAG and 23S R: CCCATTAA CGTTGGAC (nt. no. 1188-1334) and 16S F: GGAGG AAGGTGGGGATGACG and16S R: ATGGTGAC GGGCGGTGTG (nt. no. 1173-1414) from Life Technologies. Samples were amplified for 35 cycles under the following conditions: denaturation at 94C for 30 s, annealing at 57C for 30 s, and extension at 72C for 30 s. The PCR products were purified by extraction with an equal volume of phenol:CHCl3 and precipitated with 2 vols of ethanol. The pellets were dried at 44C for 15 min and then resuspended in 30 ll of sterile H2O. The purified DNA probes were labeled with biotin using the Label-IT biotin labeling kit (Mirus) following the manufacturer’s instructions. Northern hybridization analysis of 23S rRNA Five microgram of RNA from the top, 30S, and 50S gradient regions was denatured and separated on 1% agarose gels as described (Usary and Champney 2001; Champney et al. 2003). Following electrophoresis, RNA from the gels was blotted onto Nytran nylon membranes using a Turbo blot apparatus (S & S). The membranes were pre-hybridized at 42C for 30 min with 15 ml of 1· pre-hybridization solution (MRC, Inc.). The membranes were hybridized overnight at 42C with 7 ml hybridization buffer (50% formamide, 5· SSC, 0.1% Sarkosyl, 0.02% SDS and 200 lg/ml BSA) with 1· background quencher (MRC Inc.) and 4 pmol of
69
denatured 23S or 16S probe. Following hybridization, the membranes were washed and the probe was detected using the North2South chemiluminescent hybridization kit (Pierce Chemical Co.), according to the manufacturer’s instructions. For slot blot analysis, 3 lg of RNA from each gradient region was blotted to a Nytran membrane and processed as described. RNA samples were also separated on 2.5% agarose gels along with RNA size standards (Century Plus RNA Markers, Ambion), which were biotinylated with the Label-IT kit. After hybridization, RNA fragment sizes were estimated from a standard curve of the size markers.
Results Effects of azithromycin on cell viability and protein synthesis rates in E. coli strains This work was initiated to follow up on a previous observation that a strain defective in RNase E was hypersensitive to erythromycin and accumulated more 50S subunit precursor particles than did a control strain grown with the antibiotic (Usary and Champney 2001). Eight strains with single or multiple mutations in ribonuclease genes were investigated. Because some of these strains have different genetic backgrounds, they were compared for each characteristic during growth with and without azithromycin. The mutations in the RNase II and RNase E genes examined confer a temperature sensitive conditional lethal phenotype on the cells. For this reason, all of the strains were grown at 32C. The ribonuclease activity in each of these ts strains has been shown to be reduced at the permissive temperature, compared with a wild type organism (Weatherford et al. 1972; Donovan and Kushner 1983, 1986; Arraiano et al. 1988). The azithromycin MIC for the wild type strain SK901 and mutants D10, N7060, SK7622, SK5665 and CA244
Table 2 Inhibition of growth rate, cell viability and protein synthesis by azithromycin in wild type and ribonuclease mutant strains of E. coli
was 12 lg/ml. SK4803 had an MIC of 5 lg/ml while MRE600 and SK5704 were the most hypersensitive with an MIC of 2.5 lg/ml. Based on the MIC assays, azithromycin was used at 2.5 lg/ml for all strains except MRE600 and SK5704 which were grown at 1.0 lg/ml. Cells were grown at 32C in tryptone broth because cellular processes such as ribosome assembly are slowed under these conditions (Michaels 1972). This permits a greater inhibition of ribosome assembly and a larger accumulation of rRNA. Growth rates and cell viability were measured to examine the inhibitory effects of azithromycin on each strain. Table 2 summarizes the decrease in growth rate and cell viability for nine strains of E. coli growing with the antibiotic. The viability of each mutant strain was reduced more than the wild-type strain after growth with the drug. The effect of azithromycin on the rate of protein synthesis in the cells was examined by measuring the incorporation of 35S-methionine into cellular proteins. Table 2 shows the results from protein synthesis measurements with cells grown with and without azithromycin. When grown with the antibiotic, the rate of synthesis in the wild-type strain dropped to 83% of the control. All of the mutant strains showed greater decreases in translation rates by comparison. Strains MRE600 (RNase I), CA244 (PNPase) and SK5665 (RNase E) were the most inhibited, with a decline to less than 15% of the control rate. Specific inhibition of 50S ribosomal subunit assembly by azithromycin In addition to inhibiting translation, macrolide antibiotics specifically prevent the formation of 50S ribosomal subunits in growing E. coli cells (Champney et al. 1998). In a previous study we showed that an antibiotic-bound
Strain
Azithromycin Generation Cell viability Protein synthesis time (min) (cells/ml·107) (35S cpm/15 min)
SK 901(wt) + MRE 600 (RNase I) + D10 (RNase I) +
Values in parentheses indicate the fold increase in generation time or the percentage of the untreated cell number and protein synthesis rate for each strain grown with azithromycin. Growth rates, cell numbers and protein synthesis rates were determined as described in Methods. Results are the mean of three determinations with a SE of ±11%
SK7622 (RNase III) + SK5665 (RNase E) + CA244 (PNPase) + SK4803 (RNase II) + N7060 (RNase I RNase II PNPase) + SK5704 (RNase II RNase E PNPase) +
35 52 (1.5·) 35 150 (4.3·) 45 63 (1.4·) 65 93 (1.4·) 52 65 (1.3·) 48 137 (2.8·) 122 157 (1.3·) 43 73 (1.7·) 85 173 (2·)
151 90 (59%) 119 3 (3%) 168 55 (33%) 135 41 (30%) 96 21 (22%) 125 37 (30%) 40 15 (38%) 160 62 (39%) 30 20 (67%)
5446 4543 (83%) 3734 418 (11%) 5534 4254 (76%) 5201 2480 (48%) 10,069 1511 (15%) 5501 712 (13%) 23,354 15,959 (68%) 14,658 4286 (29%) 5622 1942 (35%)
70
50S precursor accumulated in cells growing with erythromycin. This was evidence that the antibiotic can bind to immature 50S subunits and prevent 50S subunit assembly (Usary and Champney 2001). In the present work, 50S ribosomal subunit amounts were quantitated in wild type and mutant cells growing in the presence and absence of azithromycin. Sucrose gradient profiles of ribosomal subunits labeled with 3H-uridine during growth with azithromycin indicated a reduction in 50S subunit amounts in most of the strains (Fig. 2). Except
Fig. 2 Sucrose gradient profiles of 3H-uridine labeled cell lysates. Cells were grown in the presence (filled square) or absence (open circle) of azithromycin and labeled with 1 lCi/ml 3H-uridine. Cell lysates were centrifuged on 5– 20% sucrose density gradients and fractions were collected. Gradient profiles from strains: a SK901(wt) b D10 (RNase I) c SK7622 (RNase III) d CA244 (PNPase) e SK5665 (RNase E) f SK4803 (RNase II)
for CA244 (PNPase), the expected 2:1 ratio of 50S to 30S subunits was found for each strain grown without the antibiotic (Table 3). The apparent excess of material in the 30S region for the drug-treated samples represents the accumulation of a 50S precursor with a similiar sedimentation coefficient (Usary and Champney 2001). A proportional increase in labeled RNA was found in the top gradient fractions of drug-treated cells. In each case a redistribution of the RNA radioactivity was seen in antibiotic-inhibited cells with the decline in 50S RNA
71 Table 3 Percentage of total gradient radioactivity in each region of sucrose gradients
Strain
Azithromycin
SK901 (wt) + MRE600 (RNase I) + D10 (RNase I) The percentage of the total 3H-uridine radioactivity in each gradient region was calculated for each strain from gradient profiles like those shown in Fig. 2. Top includes fractions 2–10, 30S includes fractions 14–19 and 50S includes fractions 24–30. Values in parentheses indicate the relative increase or decrease from the control values for cells grown with azithromycin
+ SK7622 (RNase III) + SK5665 (RNase E) + CA244 (PNPase) + SK4803 (RNase II) + N7060 (RNase I RNase II PNPase) + SK5704 (RNase II RNase E PNPase)
accompanied by an increase in RNA in the top gradient region. As Table 3 shows, 50S subunit assembly was most inhibited in strains SK4803 (RNase II), N7060 (RNase I RNase II PNPase), SK5704 (RNase II, RNase E PNPase) and MRE600 (RNase I) with reductions of 19% to 35% of the control amounts.
+
Total radioactivity (%) Top
30S
42.7 50.7 (+8) 30.6 56 (+25.4) 42.5 47 (+4.5) 42 47.7 (+5.7) 28.8 32.7 (+3.9) 63 64 (+1) 32 53.9 (+21.9) 45 63.5 (+18.5) 37.2 52.5 (+15.3)
17.7 24.7 21 21.9 17.3 25.5 18.7 18.3 21.9 23.9 13.7 13.1 21.4 19.1 18.8 22.4 20.6 21.2
50S (+7) (+0.9) (+8.2) (+0.4) (+2) ( 0.6) ( 2.3) (+3.6) (+0.6)
33.2 16.8 ( 16.4) 41.1 6 ( 35.1) 33.8 18.5 ( 15.3) 33.8 23.1 ( 10.7) 40.8 32.2 ( 8.6) 17.3 13.8 ( 3.5) 42.2 18.7 ( 23.5) 33.1 11 ( 22.1) 36.5 17.5 ( 19)
Fragments of about 1,150 and 585 nucleotides were found in wild type cells. The mutant strains also had two fragments varying in size from 950 to 1,250 and 400 to 700 bases. In strains D10 (RNase I), CA244 (PNPase), SK5665 (RNase E), SK4803 (RNase II) and SK5704 (RNase II, RNase E PNPase) a third RNA molecule of about 275 nucleotides was seen as well (data not shown).
Analysis of RNA by Northern hybridization analysis To confirm that the RNA accumulating in azithromycin treated cells was 23S rRNA from the 50S ribosomal subunit, hybridization experiments were performed. Figure 3 shows Northern blots of RNA hybridized with biotinylated 23S and 16S DNA probes. In the top gradient region (Fig. 3a), 23S rRNA sequences were found in all strains including wild-type, after growth with azithromycin. In the absence of the drug, wild type cells and six of the mutant strains had no 23S rRNA oligonucleotides present in this gradient region (Fig. 3a). Strains D10 (RNase I) and N7060 (RNase III) accumulated some 23S rRNA even in the absence of azithromycin. A similar observation was seen for 23S rRNA in the 30S gradient region (Fig. 3b). A precursor to the 50S subunit containing 23S and 5S rRNA accumulates here in erythromycin treated cells (Usary and Champney 2001). Azithromycin treatment promotes a similar accumulation. By contrast, comparable amounts of intact 23S rRNA were found in the 50S gradient region and were similar for antibiotic-treated and -free samples (Fig. 3d). As a control, rRNA from the 30S gradient region was hybridized with a 16S rRNA specific probe (Fig. 3c). Intact 16S rRNA was found in each case and the amount was unaffected by growth in the presence of azithromycin. The sizes of the 23S rRNA oligonucleotides in the top gradient regions were estimated by Northern blotting of 2.5% agarose gels calibrated with RNA size standards.
Recovery of 50S ribosomal subunits following azithromycin exposure Since azithromycin is bacteriostatic, treated cells are capable of recovery after removal of the antibiotic. To determine if the rate of recovery of the 50S subunit differed between the mutant strains, RNA in each gradient region was quantitated after the removal of azithromycin. Figure 4 shows the change in the percentage of radioactivity in the top, 30S, and 50S gradient regions as a function of time following azithromycin removal. For the wild-type strain, the 50S subunit was fully recovered to normal amounts after two hours, while RNA in the top gradient region decreased proportionally. Strains N7060, D10, and SK5665 showed comparable results. The rate of recovery for strains CA244, SK4803, and SK5704, all exoribonuclease mutants, was significantly slower. Strain SK7622, a RNase III mutant, also indicated a much slower rate of recovery. Very little change in 30S RNA amounts were seen. The t1/2 values for the changes in rRNA amounts are given in Table 4. The changes in the RNA levels during recovery were examined in more detail by hybridization of rRNA from strains SK901 (wt) and SK4803 (RNase II). Levels of 23S RNA in the top, 30S and 50S gradient regions were measured at 1 and 2 h after azithromycin removal. A significant decline in 23S rRNA oligonucleotides was seen with time in the top and 30S regions from the wild type strain along with a small increase in 50S region
72
Discussion
Fig. 3 Northern hybridization analysis of rRNA. RNA was isolated from the top, 30S and 50S regions of sucrose gradients from cells grown without and with azithromycin. The RNA was separated on a 1% agarose gel and a blot of the gel was hybridized with a biotin labeled 23S DNA probe (a, b, d) or a 16S DNA probe (c). a RNA from the top gradient region. Lane 1 SK901 - azi, lane 2 SK901 + azi; lane 3 D10 - azi, lane 4 D10 + azi; lane 5 CA244 azi, lane 6 CA244 + azi; lane 7 SK7060 - azi, lane 8 SK7060 + azi; lane 9 SK4803 - azi, lane 10 SK4803 + azi; lane 11 SK7622 - azi, lane 12 SK7622 + azi; lane 13 SK5704 - azi, lane 14 SK5704 + azi; lane 15 SK5665 - azi, lane 16 SK5665 + azi. b RNA from the 30S gradient region. Lane 1 SK901 - azi, lane 2 SK901 + azi; lane 3 D10 - azi, lane 4 D10 + azi; lane 5 SK5665 - azi, lane 6 SK5665 + azi; lane 7 CA244 - azi, lane 8 CA244 + azi; lane 9 SK7060 - azi, lane 10 SK7060 + azi; lane 11 SK4803 - azi, lane 12 SK4803 + azi. c RNA from the 30S gradient region, hybridized with a 16S DNA probe. Samples are the same as in b. d RNA from the 50S gradient region, hybridized with a 23S DNA probe. Samples are the same as in b
rRNA (Fig. 5a). By contrast very little change in the 23S rRNA distribution was seen in the RNase II mutant strain over this same time period (Fig. 5b).
Ribosomal and transfer RNAs are considered to be stable molecules in growing bacterial cells. They represent about 98% of the total cellular RNA and do not turnover during exponential growth (Deutscher 2003). Under starvation conditions, during the stationary phase or at very slow growth rates, some turnover of these RNAs has been observed (Kaplan and Apirion 1974). In addition, agents which damage the cell membrane permit the entry of the periplasmic enzyme RNase I which can partially degrade rRNA in subunits (Raziuddin et al. 1979). Mutational alteration of 16S and 23S rRNA leader sequences which impair subunit assembly is another condition which stimulates rRNA turnover (Liiv et al. 1996; Schaferkordt and Wagner 2001). The conditions for stable RNA turnover have been discussed in a recent review by Deutscher (Deutscher 2003). As this report demonstrates, antibiotic inhibition of subunit formation is another situation in which stable RNA breakdown is stimulated. The results of this work show that cells with defective RNases are hypersensitive to antibiotic inhibition as predicted and a show an enhance accumulation of rRNA oligo-nucleotides and a slower recovery after antibiotic removal. All of the mutant strains examined in this study showed reduced protein synthesis rates and enhanced lethality when growing with azithromycin, when compared with a wild type strain. Presumably this increased sensitivity comes from the excess accumulation of the 50S subunit precursor and the reduced degradation of this stalled product. The nuclease defective strains would have fewer functional 50S subunits under this condition and hence a reduced protein synthesis capacity. Strain D10 defective in RNase I was very similar in all regards to the wild type control strain. No specific effect of the RNase I mutation on 50S subunit formation or antibiotic hypersensitivity could be ascribed to the RNase I mutation. RNase I is active on tRNA and rRNA but is normally confined to the periplasmic space in E. coli (Raziuddin et al. 1979). Strain MRE600 is commonly used as a source of E. coli ribosomes because it lacks RNase I (Cammack and Wade 1965). However it is not a K-12 strain of E. coli and its ribosomes are different, containing for example a ribosomal protein S7 which is smaller than the protein in K12 strains (Held and Nomura 1973; Sun and Traut 1973). It is likely that the antibiotic hypersensitivity of this strain is due to more than the lack of RNase I. Strain SK7622 with a deletion in the RNase III gene was more affected by growth in the presence of the antibiotic than the wild type strain with slightly greater reductions in cell viability and protein synthesis rates (Table 2). It accumulated less 23S rRNA degradation products than the control and was generally less affected in 50S subunit assembly inhibition. By contrast, it required about three times as long to recover from azi-
73 80
80
A
B 60
% Total CPM
60
% Total CPM
Fig. 4 Recovery of 50S ribosomal subunits following azithromycin removal. The percentage of the total gradient radioactivity in the top (filled circle), the 30S subunit region (filled diamond) and the 50S subunit region (filled square) of sucrose gradients is shown. Recovery pattern in strains: a SK901 (wt) b D10 (RNase I) c SK7622 (RNase III) d CA244 (PNPase) e SK5665 (RNase E) f SK4803 (RNase II)
40
20
0
40
20
0 0
50 Time (min)
100
0
50 100 Time (min)
80
80
C
D 60
% Total CPM
% Total CPM
60
40
20
0
40
20
0 0
0
50 100 Time (min)
80
50 100 Time (min)
80
F
E 60
% Total CPM
% Total CPM
60
40
20
40
20
0
0 0
thromycin inhibition, possibly because of the important role for this enzyme in the processing of the precursor to both 16S and 23S rRNA (Srivastava and Schlessinger 1990). The mutation in SK5665 results in a temperaturesensitive RNase E enzyme (Arraiano et al. 1988). At 37C, this strain was hypersensitive to erythromycin and 50S subunit formation was reduced by 70% compared with a wild type RNase E strain (Usary and Champney 2001). In the present work, with cells grown at 32C, both cell viability and the rate of translation were reduced relative to the control strain (Table 2). Like the RNase III strain, SK5665 required significantly longer
50 Time (min)
100
0
50 100 Time (min)
to recover normal 50S subunit levels after azithromycin inhibition, also reflecting the major role for this enzyme in the processing of 5S and 23S rRNAs (Srivastava and Schlessinger 1990). Ribonuclease E is part of a multi enzyme complex in cells called the degradosome (Carpousis et al. 1994). It contains RNase E, PNPase, the RhlB helicase, the chaperone protein DnaK, GroEL, polynucleotide phosphate kinase and enolase (Py et al. 1994; Miczak et al. 1996). RNase E organizes the complex by interactions with its C-terminal protein sequences (Vanzo et al. 1998) and anchors the complex to the cell membrane by an N-terminal sequence interaction (Liou et al.
74 Table 4 Relative recovery rates after azithromycin removal Strain
SK901 (wt) D10 (RNase I) SK7622 (RNase III) SK5665 (RNase E) CA244 (PNPase) SK4803 (RNase II) N7060 (RNase I RNase II PNPase) SK5704 (RNase II RNase E PNPase)
t1/2 (min) Top
50S
180 145 360 150 250 740 130 620
75 95 240 280 100 ND 55 70
t1/2 values were calculated from best fit lines to recovery graphs like those in Fig. 4. Values indicate the t1/2 for decreases in the top gradient fractions and increases in the 50S region. Results are the mean of two experiments ND Not determined
2001). The N-terminal is also the site of the processing activity for conversion of 9S to 5S rRNA (Lopez et al. 1999). It has an important function in the turnover of mRNA during translation but has also been shown to function in rRNA turnover as well (Bessarab et al. 1998) RNase E has endonucleolytic activity on AU-rich loop regions of both 16S and 23S rRNAs in the degradosome particle. The largest effects of azithromycin were seen in the single mutants defective in RNase II and in polynucleotide phosphorylase which showed substantial reductions in cell numbers and protein synthesis rates (Table 2). In particular, strain CA244 had the largest reduction in growth rate, cell viability and translation rate of any K12 strain examined. 50S subunit precursors accumulated in this strain in the presence of azithromycin as detected by hybridization (Fig. 3b) and small 23S rRNA oligonucleotides were prominent in this strain after antibiotic inhibition. A general impairment in 50S subunit formation was also evident from the sucrose gradient analysis, which showed very reduced particle formation in the presence of azithromycin (Fig. 2d).
Fig. 5 Slot blot hybridization analysis of 23S rRNA during recovery. Samples were taken at 1 and 2 h after azithromycin removal from (a) strain SK901 (wt) and (b) strain SK4803 (RNase II). Ribosomal subunits were separated by sucrose gradient centrifugation and RNA from the top, 30S and 50S regions was transferred to a nylon membrane and hybridized with a 23S DNA probe
The strain with reduced RNase II activity (SK4803) was somewhat less susceptible to the growth effects of the antibiotic (Table 2). This organism demonstrated the largest reduction in 50S formation in response to the drug of any strain examined (except MRE600). Very substantial amounts of 23S rRNA oligonucleotides accumulated in this strain during antibiotic inhibition (Fig. 3a, b). Recovery from azithromycin inhibition was very prolonged in SK4803, with 50S synthesis prevented for more than 12 h (Table 4). Two triple mutant strains, N7060, affected in RNase II and PNPase (and RNase I) and SK5704 reduced in the activities of RNase II, RNase E and PNPase were also examined. Both strains were hypersensitive to azithromycin, with substantial reductions in both cell viability and protein synthesis rates during growth with the antibiotic (Table 2). They both showed large impairments in 50S subunit formation and each accumulated substantial amounts of 23S rRNA oligonucleotides during growth in the presence of the drug (Fig. 3a, b). The RNase II mutation seem to have the greater effect in combination, with changes more comparable to those seen with strain SK4803 (RNase II). The recovery from azithromycin inhibition was especially prolonged in the triple mutant strain SK5704 reflecting the need for both RNase II and PNPase to remove the accumulated 23S rRNA oligonucleotides (Table 4). We have shown previously that in E. coli cells growing with erythromycin, there was an antibiotic concentration dependent loss of RNA from the 50S region of sucrose gradients and a proportional increase in RNA in the top gradient fractions (Champney and Burdine 1996). This same proportional decrease and increase was also found for S. aureus cells growing with azithromycin, clarithromycin (Champney and Burdine 1998) and linezolid (Champney and Miller 2002). The results in Table 3 show a similar relationship for the redistribution of RNA in azithromycin-treated cells. In each case the decline in 50S associated RNA radioactivity was accompanied by an almost equal increase in
75
the 30S and top gradient regions. For example, SK901 cells growing with azithromycin lost 16% of the RNA in the 50S region compared with control cells, but gained 15% more radioactivity in the 30S and top gradient regions (Table 3). It is interesting that most of the mutant strains showed 23S rRNA degradation products of discrete sizes. This suggests that there are specific structural features of the molecule which are recognized by endonucleases leading to the initial formation of relatively large rRNA oligonucleotides. RNase E has been shown to cleave both 23S and 16S rRNA into large oligonucleotides (Bessarab et al. 1998). Exonucleolytic cleavage of these products would follow and may be a slower process. It is also important to point out the specificity of rRNA turnover under these conditions. We have shown repeatedly that 50S subunit inhibitors do not impede 30S particle formation (Champney 2003). The presence of intact 16S rRNA in azithromycin inhibited cells strengthens this previous observation. During the recovery from azithromycin inhibition, two processes are occurring. rRNA oligonucleotides in cells are being degraded at different rates depending on the RNase content of the cells. In wild type cells about 2 h were required to regain the control level of RNA fragments (Fig. 4). About this same time was found for RNase I and RNase E mutant strains. The other nuclease strains required substantially longer to reduce the levels of 23S rRNA oligonucleotides. The second process during recovery is the resynthesis of 50S particles from the stalled intermediates. Presumably, these lose the bound drug and mature into 50S subunits at different rates, depending on the RNase composition of the cells. 23S and 5S rRNA maturation requires the enzymes RNase III, E and T to generate mature rRNA sequences (Srivastava and Schlessinger 1990; Li et al. 1999). Maturation of the stalled intermediate will depend on the function of these enzymes or their substitutes. It is apparent from Table 4 that strains defective in RNases II, III and E required the longest period of time to recover to control levels of 50S subunits. Antibiotic-induced degradation of the 50S subunit precursor seems to be dependent on the exoribonuclease activity of RNase II and PNPase. RNase II defects impair 23S rRNA turnover most significantly. This enzyme has been implicated in the turnover of stable RNA during starvation along with RNase I and PNPase (Kaplan and Apirion 1974). It also has a role in the maturation of 17S to 16S rRNA and may function in 23S rRNA processing as well (Corte et al. 1971). Interestingly, the turnover of RNase II mRNA is controlled by the activity of RNases E and RNase III (Zilhao et al. 1995). By mutational analysis, Mohanty and Kushner (2000) have shown the specific involvement of RNase II in 23S rRNA turnover and the requirement of PNPase for 16S rRNA breakdown. The activity of these two exonucleases in rRNA turnover is facilitated by an indirect dependence on RNase E as well.
Our results help to explain the observation of McMurray and Levy (1987) that PNPase deletion strains are hypersensitive to several ribosomal antibiotics. PNPase is a major degradative enzyme in E. coli and unlike RNase II, it is dependent on phosphate as the nucleophile in RNA degradation (Deutscher and Li 2001). This enzyme has several roles in cells including mRNA degradation, tRNA repair and RNA synthesis under certain circumstances. The results of this work also suggest a major role in rRNA degradation. E. coli has two phosphate dependent exonucleases, PNPase and RNase PH. The activity of a least one of these proteins is needed for normal 50S subunit formation, presumably in the processing of 23S rRNA (Zhou and Deutscher 1997). Mutational loss of both activities reduces 50S particle formation substantially and promotes the accumulation of 23S rRNA fragments in cells. This situation mimics the findings for azithromycin inhibition of 50S formation demonstrated here. Other exoribonucleases not examined in this work include oligoribonuclease, RNase D, BN, PH and R. RNase R is an exonuclease which is 60% similar in sequence to RNase II (Nicholson 1999). Compared with RNase II, RNase R is more active on rRNA (Deutscher and Li 2001). In addition, both RNase R and PNPase are involved in the quality control of mature RNA structure and double mutants lacking both enzymes are inviable (Cheng and Deutscher 2003). An investigation of rRNA turnover in a RNase R mutant strain would be important. Our results indicate the importance of RNases II and PNPase in turnover of stalled 50S subunit intermediates in antibiotic-inhibited cells. Since inhibition of subunit assembly is equivalent to inhibition of translation for most ribosomal inhibitors (Champney 2003), this indicates the significance of these enzymes in reversing the significant inhibitory activity of these antimicrobial agents. Acknowledgements We are pleased to acknowledge the gift of strains from Murray Deutscher and Sidney Kushner. This research was supported in part by an AREA grant from the NIH.
References Anderson RM (1999) The pandemic of antibiotic resistance. Nat Med 5:147–149 Arraiano CM, Yancey SD, Kushner SR (1988) Stabilization of discrete mRNA breakdown products in ams pnp rnb multiple mutants of Escherichia coli K-12. J Bacteriol 170:4625–4633 Babitzke P, Granger L, Olsezewski J, Kushner SR (1993) Analysis of mRNA decay and rRNA processing in Escherichia coli multiple mutants carrying a deletion in RNase III. J Bacteriol 175:229–239 Bessarab DA, Kaberdin VR, Wei C-L, Liou G-G, Lin-Chao S (1998) RNA components of Escherichia coli degradosome: evidence for rRNA decay. Proc Natl Acad Sci US 95:3157–3161 Broderson DE, Clemons WM, Carter AP, Morgan-Warren RJ, Wimberly BT, Ramakrishnan V (2000) The structural basis for the action of the antibiotics tetracycline, pactamycin, and hygromycin B on the 30S ribosomal subunit. Cell 103:1143–1154
76 Brosius J, Dull TJ, Noller HF (1980) Complete nucleotide sequence of a 23S ribosomal RNA gene from Escherichia coli. Proc Natl Acad Sci US 77:201–204 Cammack KA, Wade HE (1965) The sedimentation behaviour of ribonuclease-active and -inactive ribosomes from bacteria. Biochem J 96:671–680 Canu A, Malbruny B, Coquemont M, Davies TA, Applebaum PC, Leclercq R (2002) Diversity of ribosomal mutations conferring resistance to macrolides, clindamycin, streptogramin, and telithromycin in Streptococcus pneumoniae. Antimicrobial Agents Chemother 46:125–131 Carpousis AJ, Houwe GV, Ehretsmann C, Krisch HM (1994) Copurification of E. coli RNAase E and PNPase: evidence for a specific association between two enzymes important in RNA procesing and degradation. Cell 76:889–900 Carter AP, Clemons WM, Broderson DE, Morgan-Warren RJ, Wimberly BT, Ramakrishnan V (2000) Functional insights from the structure of the 30S ribosomal subunit and its interactions with antibiotics. Nature 407:340–348 Champney WS (2003) Bacterial ribosomal subunit assembly is an antibiotic target. Curr Top Med Chem 3:929–947 Champney WS, Burdine R (1996) 50S ribosomal subunit synthesis and translation are equivalent targets for erythromycin inhibition in Staphylococcus aureus. Antimicrob Agents Chemother 40:1301–1303 Champney WS, Burdine R (1998) Azithromycin and clarithromycin inhibition of 50S ribosomal subunit formation in Staphylococcus aureus cells. Curr Microbiol 36:119–123 Champney WS, Miller M (2002) Linezolid is a specific inhibitor of 50S ribosomal subunit formation in Staphylococcus aureus cells. Curr Microbiol 44:350–356 Champney WS, Tober CL, Burdine R (1998) A comparison of the inhibition of translation and 50S ribosomal subunit formation in Staphylococcus aureus cells by nine different macrolide antibiotics. Curr Microbiol 37:412–417 Champney WS, Chittum H, Tober CL (2003) A 50S ribosomal subunit precursor particle is a substrate for the ermC methyltransferase in Staphylococcus aureus cells. Curr Microbiol 46:453–460 Cheng Z-F, Deutscher MP (2003) Quality control of ribosomal RNA mediated by polynucleotide phosphorylase and RNase R. Proc Natl Acad Sci US 100:6388–6393 Chittum HS, Champney WS (1995) Erythromycin inhibits the assembly of the large ribosomal subunit in growing Escherichia coli cells. Curr Microbiol 30:273–279 Chu DT, Plattner JJ, Katz L (1996) New directions in antibacterial research. J Med Chem 39:3853–3874 Corte G, Schlessinger D, Longo D, Venkov P (1971) Transformation of 17 s to 16 s ribosomal RNA using ribonuclease II of Escherichia coli. J Mol Biol 60:325–338 Deutscher MP (2003) Degradation of stable RNA in bacteria. J Biol Chem 278:45041–45044 Deutscher MP, Li Z (2001) Exoribonucleases and their multiple roles in RNA metabolism. Prog Nucleic Acid Res Mol Biol 66:67–105 Dinos GP, Michelinaki M, Kalpaxis DL (2001) Insights into the mechanism of azithromycin interaction with an Escherichia coli functional ribosomal complex. Molec Pharmacol 59:1441–1445 Donovan WP, Kushner SR (1983) Amplification of ribonuclease II (rnb) activity in Escherichia coli K-12. Nuc Acids Res 11:265– 275 Donovan WP, Kushner SR (1986) Polynucleotide phosphorylase and ribonuclease II are required for cell viability and mRNA turnover in Escherichia coli K-12. Proc Natl Acad Sci US 83:120–124 Gesteland RF (1966) Isolation and characterization of ribonuclease I mutants of Escherichia coli. J Mol Biol 16:67–84 Hansen JL, Ippolito JA, Ban N, Nissen P, Moore PB, Steitz TA (2002) The structures of four macrolide antibiotics bound to the large ribosomal subunit. Molecular Cell 10:117– 128
Held W, Nomura M (1973) Functional homology between the 30S ribosomal protein S7 from E. coli K12 and the S7 from E. coli MRE600. Mol Gen Genet 122:11–14 Jett BD, Hatter KL, Huycke M, Gilmore MS (1997) Simplified agar plate method for quantifying viable bacteria. BioTechniques 23:648–650 Kaplan R, Apirion D (1974) The involvement of ribonuclease I, ribonuclease II, and polynucleotide phosphorylase in the degradation of stable ribonucleic acid during carbon starvation in Escherichia coli. J Biol Chem 249:149–151 Kushner SR, Maples VF, Champney WS (1977) Conditionally lethal ribosomal protein mutants: characterization of a locus required for modification of 50S subunit proteins. Proc Natl Acad Sci USA 74:467–471 Li Z, Pandit S, Deutscher MP (1999) Maturation of 23S ribosomal RNA requires the exoribonuclease RNase T. RNA 5:139–146 Liiv A, Tenson T, Remme J (1996) Analysis of the ribosome large subunit assembly and 23S rRNA stability in vivo. J Mol Biol 263:396–410 Liou G-G, Jane W-N, Cohen SN, Lin-Chao S (2001) RNA degradosomes exist in Escherichia coli as multicomponent complexes associated with the cytoplasmic membrane via the Nterminal region of ribonuclease E. Proc Natl Acad Sci US 98:63–68 Lopez PJ, Marchand I, Joyce SA, Dreyfus M (1999) The C-terminal half of RNase E, which organizes the Escherichia coil degradosome, participates in mRNA degradation but not rRNA processing in vivo. Mol Microbiol 33:188–199 McMurry LM, Levy SB (1987) Tn5 insertion in the polynucleotide phosphorylase (pnp) gene in Escherichia coli increases susceptibility to antibiotics. J Bacteriol 169:1321–1324 Mehta R, Champney WS (2002) 30S ribosomal subunit assembly is a target for inhibition by aminoglycosides in Escherichia coli. Antimicrob Agents Chemother 46:1546–1549 Mehta R, Champney WS (2003) Neomycin and paromomycin inhibit 30S ribosomal subunit assembly in Staphylococcus aureus. Curr Microbiol 43:237–243 Michaels GA (1972) Ribosome maturation of Escherichia coli growing at different growth rates. J Bacteriol 107:385–387 Miczak A, Kaberdin VR, Wei C-L, Lin-Chao S (1996) Proteins associated with RNase E in a multicomponent ribonucleolytic complex. Proc Natl Acad Sci US 93:3865–3869 Mohanty BK, Kushner SR (2000) Polynucleotide phosphorylase, RNaseII and RNase E play different roles in the in vivo modulation of polyadenylation in Escherichia coli. Mol Microbiol 36:982–994 Nicholson A (1999) Function, mechanism and regulation of bacterial ribonucleases. FEMS Microbiol Rev 23:371–390 Nierhaus K (1982) Structure, assembly and function of ribosomes. Curr Topics Microbiol Immunol 97:81–155 Py B, Causton H, Mudd EA, Higgins CF (1994) A protein complex mediating mRNA degradation in Escherichia coli. Mol Microbiol 14:717–729 Raziuddin, Chatterji D, Ghosh S, Burma DP (1979) Site of action of RNase I on the 50 S ribosome of Escherichia coli and the association of the enzyme with the partially degraded subunit. J Biol Chem 254:10575–10578 Reuven NB, Deutscher MP (1993) Multiple exoribonucleases are required for the 3¢ processing of Escherichia coli tRNA precursors in vivo. FASEB J 7:143–148 Schaferkordt J, Wagner R (2001) Effects of base change mutations within an Escherichia coli ribosomal RNA leader region on rRNA maturation and ribosome formation. Nuc Acids Res 29:3394–3403 Schlunzen F, Zarivach R, Harms J, Bashan A, Tocilj A, Albrecht R, Yonath A, Franceschi F (2001) Structural basis for the interaction of antibiotics with the peptidyltransferase centre in eubacteria. Nature 413:814–821 Schroeder R, Wallis MG (2001) RNA-binding antibiotics. Landes Bioscience, Georgetown
77 Srivastava AK, Schlessinger D (1990) Mechanism and regulation of bacterial ribosomal RNA processing. Ann Rev Microbiol 44:105–129 Sun TT, Traut RR (1973) The functional and structural homology of ribosomal protein S7 of E. coli strains K and MRE600. Mol Gen Genet 122:1–9 Usary J, Champney WS (2001) Erythromycin inhibition of 50S ribosomal subunit formation in Escherichia coli cells. Mol Microbiol 40:951–962 Vanzo NF et al (1998) Ribonuclease E organizes the protein interactions in the Escherichia coli degradosome. Genes Dev 12:2770–2781 Vazquez D (1979) Inhibitors of protein biosynthesis. Springer, Berlin Heidelberg New York
Weatherford SC, Rosen L, Gorelic L, Apirion D (1972) Escherichia coli strains with thermolabile ribonuclease II activity. J Biol Chem 247:5404–5408 Zhou Z, Deutscher MP (1997) An esential function for the phosphate-dependent exoribonucleases RNase PH and polynucleotide phosphorylase. J Bacteriol 179:4391–4395 Zilhao R, Regnier P, Arraiano CM (1995) The role of endonucleases in the expression of ribonuclease II in Escherichia coli. FEMS Microbiol Lett 130:237–244 Zuckerman JM, Kaye KM (1995) The newer macrolides. Azithromycin and clarithromycin. Infect Dis Clin North Am 9:731–745
