|Home | About | Journals | Submit | Contact Us | Français|
The EngA protein is a conserved and essential bacterial GTPase of largely enigmatic function. While most investigations of EngA have suggested a role in ribosome assembly, the protein has also been implicated in diverse elements of physiology including chromosome segregation, cell division and cell cycle control. Here, we have probed additional phenotypes related to ribosome biogenesis on depletion of EngA in Escherichia coli to better understand its role in the cell. Depletion of EngA resulted in cold sensitive growth and stimulation of a ribosomal rRNA promoter, both phenotypes associated with the disruption of ribosome biogenesis in bacteria. Amongst antibiotics that inhibit translation, depletion of EngA resulted in sensitization to the aminoglycoside class of antibiotics. EngA bound the alarmone ppGpp with equally high affinity as it bound GDP. These data offer additional support for a role in ribosome biogenesis for EngA, possibly in maturation of the A-site of the 50S subunit.
In prokaryotes, ribosome biogenesis involves the production, folding and assembly of 54 proteins and 3 rRNA species to form the 30S and 50S subunits. This is a cooperative process, which requires several classes of trans-acting factors, including RNases, helicases, modification enzymes, folding chaperones and GTPases (Shajani et al. 2011). There are six conserved subfamilies of GTPases in bacteria (the EngA, Era, YihA, Obg, YchF and HflX subfamilies) that are associated with ribosome biogenesis (Verstraeten et al. 2011). All of the putative ribosome biogenesis GTPases have also been associated with diverse non-ribosomal processes such as cell cycle progression, DNA replication and metabolism (Verstraeten et al. 2011; Britton 2009). This has led to uncertainty about the cellular function of this group of GTPases in bacteria.
EngA is an essential bacterial GTPase with tandem GTP-binding domains and a C-terminal KH (K homology) domain. We and others have previously shown that cells that are depleted of EngA show a decrease in the level of 70S ribosomes and an accumulation of 30S and 50S ribosomal subunits (Bharat et al. 2006; Hwang & Inouye 2006; Tomar et al. 2009). Depletion of EngA also led to accumulation of precursors to the 16S and 23S rRNA (Hwang & Inouye 2006). When the 30S, 50S and 70S ribosomal species in cell lysates were resolved on sucrose gradients, EngA cofractionated with the 50S subunit (Bharat et al. 2006; Hwang & Inouye 2008). Overexpression of engA rescued the defective ribosome profile of a deletion of rrmJ, which methylates residue U2552 of 23S rRNA on a loop that contacts the aminoacyl end of the A-site tRNA (Tan et al. 2002; Widerak et al. 2005). EngA was able to restore wildtype levels of 70S without restoring this methylation near the A-site (Tan et al. 2002).
Non-ribosomal phenotypes of depletion of EngA have also been reported. In E. coli, a temperature sensitive mutant displayed cell filamentation and abnormal chromosome segregation (Hwang & Inouye 2001). In B. subtilis, reduction of expression of the EngA orthologue led to cell curvature and apparent nucleoid condensation (Morimoto et al. 2002). Lastly, overexpression of EngA in E. coli resulted in clear spaces at the cell poles and a portion of overexpressed EngA protein was partially associated with membranes (Lee et al. 2011). If EngA is indeed a ribosome biogenesis factor, it is possible that loss of EngA results in pleiotropic effects due to a general decrease in protein synthesis.
To further test the link between EngA and ribosome biogenesis, we examined two cell-based phenotypes that are associated with disruption of ribosome biogenesis: (i) cold sensitivity and (ii) increased transcription of rRNA. Lower temperature reduces the thermal energy that is necessary for rRNA remodeling during ribosome biogenesis (Shajani et al. 2011). Cold sensitivity has been reported for mutants of ribosomal structural genes and ribosome biogenesis factors, including helicases, modification enzymes and RNA binding proteins (Shajani et al. 2011). Transcription of rRNA is the rate-limiting step of ribosome biogenesis and is subject to multiple mechanisms of regulation to ensure that the level of ribosomes matches the protein synthesis requirements of the cell (Paul et al. 2004). Disruption of ribosome biogenesis or translation initiation is associated with increased transcription of rRNA (Jinks-Robertson et al. 1983; Takebe et al. 1985).
Chemical genetics is an approach to studying gene function where small molecule inhibitors are used to mimic the effects of genetic lesions. The use of antibiotics as chemical probes of protein function has provided important insight into many aspects of biological complexity (Falconer et al. 2011). Indeed, loss of the putative ribosome biogenesis GTPase, YloQ, in B. subtilis led to sensitization to inhibitors of the ribosomal decoding centre at the P-site, which was later supported by cryo-EM structures showing distortion near the decoding centre of the 30S subunit (Campbell et al. 2005; Jomaa et al. 2011). Similarly, early evidence of ribosomal functions for the ribosome rescue factor, SsrA and the GTPase of unknown function, HflX, were provided by observations of sensitization to ribosomal P-site antibiotics (la Cruz & Vioque 2001; Shields et al. 2009). In the work presented here, we took a chemical-genetic approach to probing the action of EngA on ribosomes using ribosomal antibiotics, including aminoglycosides, macrolides and tetracycline derivatives.
Here, we report that depletion of EngA led to cold sensitive growth and stimulation of a fluorescent reporter of the rRNA promoter, PrrnH. Cells that were depleted of EngA were sensitized to aminoglycoside antibiotics, which inhibit decoding at the A-site of the ribosome. EngA bound ppGpp, a nucleotide that helps to regulate production of ribosomes, with similar affinity as for GTP and GDP. These results support a role for EngA in ribosome biogenesis and suggest that EngA may act on the 50S subunit near the A-site.
EB68 is wildtype E. coli MG1655 (F− λ− ilvG− rfb-50 rph-1). EB1209 (araBAD::engA-kanR, engA::CmR) and EB2354 (araBAD::engA-tetR, engA::CmR) are two versions of a strain containing a deletion of engA at its native locus and a rescue copy of engA at the arabinose-inducible araBAD locus. Expression of engA in EB1209 and EB2354 was induced by 1% L-arabinose.
EB68 was grown in LB and EB2354 was grown in LB with or without 1% L-arabinose at 37°C for 16 h. Each sample was subcultured to an initial OD600 of 0.002 and grown with rotation at either 37°C or 15°C in Sunrise plate readers (Tecan, Morrisville, NC). OD600 was monitored for 20 h at 37°C and for 40 h at 15°C to allow all cultures to reach stationary phase.
Low copy reporter plasmids were obtained from a library of known and predicted E. coli promoters fused to the fast-folding GFP variant, GFPmut2 in plasmid pUA66 (Zaslaver et al. 2006). Reporter plasmids for the rRNA promoter, PrrnH, or the control promoter, PlexA, were transformed into EB2354 and EB68. Overnight cultures of cells harboring reporter plasmids were subcultured in LB with 50 ug mL−1 kanamycin to an initial OD600 of 0.001 and grown at 37°C with rotation. OD600 and fluorescence intensity (λEx 485 nm and λEm 520 nm) were continuously monitored in a Synergy™ HT Multi-Mode microplate reader (Biotek, Winooski, VT).
Minimum inhibitory concentrations (MIC) of ribosomal antibiotics were obtained in EngA-depleted and EngA-replete conditions, similarly to the method described by Campbell et al (Campbell et al. 2005). An overnight culture of EB1209 was diluted to an OD600 of 0.001 in LB with or without 1% L-arabinose. Antibiotics were added to the first columns of 96-well plates at concentrations of 8-fold above each predicted MIC and 2-fold serial dilutions were carried out. Cultures were grown at 37°C without shaking for 16 h and OD600 was measured on a SpectraMax spectrophotometer (Molecular Devices, Sunnyvale, CA).
Recombinant untagged EngA was purified as previously described (Bharat et al. 2006). All nucleotide-binding assays were carried out using GDP labeled at the 2′ or 3′ position of the ribose ring with 4,4-difluoro-4-bora-3a,4a-diaza-s-indacene (BODIPY FL) (Life Technologies, Burlington, ON) in black low-volume 96-well plates. Fluorescence polarization (λEx 485 nm and λEm 520 nm) was measured on an EnVision® Multilabel microplate reader (Perkin Elmer, Waltham, MA).
To find the dissociation constant (Kd) of EngA for GDP, the binding of BODIPY FL-GDP (0.1 μM) to EngA (0–25 μM) was measured by the increase in fluorescence polarization. The Kd of EngA for BODIPY FL-GDP was determined by fitting the data to a hyperbolic function using SigmaPlot 12 software (Systat Software Inc, San Jose, CA).
We measured the displacement of 0.1 μM BODIPY-GDP from 1 μM EngA by various concentrations (0.01 μM–1000 μM) of unlabeled GTP, GDP, ppGpp or the nonhydrolyzable GTP analogs, guanosine-5′-[(β,γ)-methyleno] triphosphate (GMPPCP) and guanosine-5′-[(β,γ)-imido] triphosphate (GMPPNP). The competitive displacement curves were fitted to a four parameter logistic model with SigmaPlot 12 software using Eq. 1, where FP is fluorescence polarization, min and max are the fitted minimum and maximum polarization, Hillslope is the slope of the curve at its midpoint and EC50 is the concentration of competitive ligand that produced 50% displacement.
The inhibition constant (Ki) of each nucleotide was calculated with Eq. 2 using the EC50 of each competitive nucleotide (determined with Eq. 1). L is the concentration of the labeled ligand (BODIPY-GDP) and Kd is the dissociation constant of BODIPY-GDP (determined above).
Sensitivity to growth at low temperature is a common phenotype of mutation of ribosome biogenesis factors and ribosomal structural proteins (Shajani et al. 2011). EB2354 was grown in the presence or absence of arabinose, which induces expression of engA from a rescue copy of the gene at the araBAD locus. At 37°C, EB2354 initially grew slower in the absence of arabinose compared to wildtype EB68 or EB2354 grown with arabinose; however, all three cultures achieved similar final densities (Fig. 1). At 15°C, EB2354 grown without arabinose achieved a final density that was 10-fold lower than either EB68 or EB2354 grown with arabinose (Fig. 1). Thus, at 15°C, EngA-depleted cells displayed cold sensitivity of growth by a factor of 10-fold.
Cold sensitivity is not a definitive phenotype of ribosome function since other cellular features, such as membrane fluidity, are also affected (Phadtare 2004). This phenotype is, however, very common among mutants of ribosomal genes. Cold-sensitive mutations have been isolated in rRNA, ribosomal structural proteins and ribosome assembly factors (Dammel & Noller 1995; Guthrie et al. 1969; Jones et al. 1996; Shajani et al. 2011). It is thought that low temperature leads to a greater propensity for ribosomal RNA to form secondary structures leading to kinetic or thermodynamic traps that may hinder ribosome assembly (Herschlag 1995; Talkington et al. 2005).
The rate of transcription of rRNA is influenced by the level of functioning ribosomes in the cell (Jinks-Robertson et al. 1983; Takebe et al. 1985). We measured the transcriptional response of an rRNA promoter, PrrnH, and a negative control promoter, PlexA, to depletion of EngA. The promoter for the lexA gene, which is involved in DNA repair, was used as a negative control because this process appears to be unrelated to ribosome biogenesis. Reporter plasmids were obtained from a library of E. coli promoters that were fused to the GFP variant GFPmut2, which folds in minutes (Zaslaver et al. 2006).
Upon depletion of EngA, EB2354 showed profound slow growth compared to wildtype (Fig. 2A, 2B). The control promoter, PlexA, produced levels of fluorescence in EB2354 and EB68 that were proportional to the cell densities of these strains (Fig. 2D). The rRNA promoter, PrrnH, however, produced a level of fluorescence in EngA-depleted cells that approached the level of fluorescence in wildtype cells, despite the difference in cell densities (Fig. 2C). After 10 h, the cell-density normalized fluorescence of the control promoter, PlexA, was nearly identical in EngA-depleted and wildtype cells, while the cell density-normalized fluorescence of PrrnH was 5-fold higher in EngA-depleted cells compared to wildtype cells (Fig. 2E, 2F). Thus, the rRNA promoter, PrrnH, was activated 5-fold upon depletion of EngA.
This stimulation of rRNA transcription is consistent with the model of feedback regulation whereby disruption of ribosome biogenesis or translation initiation leads to an increase in production of rRNA (Jinks-Robertson et al. 1983; Paul et al. 2004). For example, perturbation of ribosome biogenesis by overexpression of the repressor ribosomal protein, S4, caused a 50% increase in transcription of rRNA (Takebe et al. 1985). This feedback is mediated by alteration of the level of the alarmone ppGpp, which inhibits transcription of rRNA (Paul et al. 2004). Among ribosome-related promoters, the rRNA promoters are more suitable than ribosomal protein promoters for use in transcriptional reporter assays because rRNA is regulated at the level of transcription, whereas ribosomal proteins are mainly regulated at the level of translation (Dennis et al. 2004).
Where loss of EngA was previously associated with an accumulation of 30S and pre-50S particles, we were interested to use ribosome-related antibiotics to phenotypically probe the function of ribosomes produced in the absence of EngA. The MIC of each antibiotic was determined in EngA-depleted and EngA-replete conditions. The fold-sensitization was determined by the decrease in MIC upon depletion of EngA. Depletion of EngA caused 4-fold to 8-fold sensitization to 7 of the 9 aminoglycosides that were tested (Fig. 3). The aminoglycosides were the only class of ribosomal antibiotics that showed clear chemical genetic interactions with EngA.
The two aminoglycoside-like antibiotics that did not show sensitization, viomycin and hygromycin B, are quite structurally distinct from the majority of aminoglycosides. Aminoglycoside antibiotics inhibit decoding at the A-site of ribosomes (Carter et al. 2000). Previously, it was reported that overexpression of EngA suppressed the loss of RrmJ, which methylates a residue of 23S rRNA that contacts the A-site tRNA (Tan et al. 2002; Widerak et al. 2005). Thus, EngA may be important for maturation of this region of the ribosome. The observation that aminoglycosides further decrease growth upon depletion of EngA suggests that some of the subunits produced in the absence of EngA can progress in the biogenesis cycle but that these ribosomes may be functionally different from wildtype.
The function of a GTPase is normally regulated by its GTPase cycle, whereby nucleotide binding and release dictates its interaction with an effector molecule. The affinity of EngA for guanine nucleotides was tested to better understand how ribosome binding might be regulated. Fluorescently labeled BODIPY-GDP bound EngA with a dissociation constant (Kd) of 2 μM (Fig. 4A). The displacement of BODIPY-GDP from EngA by unlabeled GTP, GDP, ppGpp or the nonhydrolyzable GTP analogs, GMPPCP and GMPPNP, was measured by the decrease in fluorescence polarization (Fig. 4B). E. coli EngA bound nucleotides with an inhibition constant (Ki) of 6 μM for GTP and 2 μM for GDP. These values are consistent with the previously reported affinity of Salmonella thyphimurium EngA for GTP (12.7 μM) and GDP (3 μM and 0.6 μM when fitted to a two-site model) (Lamb et al. 2007). The value we obtained for GTP likely reflects the affinity of EngA for both GTP and the hydrolysis product, GDP. No binding was observed for the two non-hydrolyzable analogs, GMPPNP and GMPPCP. Previously, it was reported that GMPPNP increased the binding of EngA to ribosomes in cell lysates; however, similar in vitro experiments required the use of either 15–45 mM urea or a Y134A variant, presumably to unfasten the connection between the first GTP-binding domain and the C-terminal KH domain (Tomar et al. 2009). There may be an unknown factor in cell lysates that facilitates the binding of EngA to GMPPNP.
We were interested in the affinity of EngA for ppGpp due to a previous report that overexpression of the ppGpp synthase relA suppressed both the slow growth and the ribosome profile defect of an engA mutant (Hwang & Inouye 2008). EngA bound ppGpp as tightly as GDP, with both displaying inhibition constants of 2 μM (Fig. 4B). When nutrient and amino acid availability is poor, the level of ppGpp in the cell is higher and the requirement for ribosomes is lower. Our observation of binding of EngA to ppGpp, may represent one mechanism of fine-tuning ribosome biogenesis to match nutrient availability.
Depletion of EngA resulted in cold sensitivity, stimulation of an rRNA promoter and sensitization to aminoglycosides. These phenotypes of depletion of EngA suggest a ribosome-related function for this essential GTPase. EngA displayed strong affinity for ppGpp, a molecule that has an important role in regulating the production of ribosomes. Our observation of sensitization to A-site binding aminoglycosides, together with a previous report that EngA suppressed the loss of RrmJ, which methylates a residue near the A-site, may suggest that EngA is important for maturation of this region. Structural studies of the decoding region of ribosomes from EngA-depleted cells will be useful for testing this hypothesis.
This work was supported by an operating grant to E.D.B. from the Canadian Institutes of Health Research [grant number MOP-64292] and by scholarship funds from the Canadian Institutes of Health Research to A.B. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.