RESEARCH LETTER–Physiology&Biochemistry Antimicrobial ...mcbl.iisc.ac.in/Varshney/docs/2018/fusidic_acid.pdf · pTrcEcoEFG E. coli fusA clonedintopTrc99c. RaoandVarshney(2001) pTrc-ssrA

FEMS Microbiology Letters, 365, 2018, fny133

doi: 10.1093/femsle/fny133Advance Access Publication Date: 25 May 2018Research Letter

RESEARCH LETTER –Physiology & Biochemistry

Antimicrobial activity of fusidic acid in Escherichia coliis dependent on the relative levels of ribosomerecycling factor and elongation factor GShreya Ahana Ayyub1, Kuldeep Lahry1, Divya Dobriyal1, Sanjay Mondal1

and Umesh Varshney1,2,∗

1Department of Microbiology and Cell Biology, Indian Institute of Science, Bangalore 560012, India and2Jawaharlal Nehru Centre for Advanced Scientific Research, Jakkur, Bangalore, 560064∗Corresponding author: Tel: +918022932686; Fax: +918023602697; E-mail: [email protected] sentence summary: The antimicrobial activity of the antibiotic fusidic acid in Escherichia coli increases with an increase in the cellular levels ofelongation factor G with respect to ribosome recycling factor.Editor: Geertje van Keulen

ABSTRACT

During protein synthesis, elongation factor G (EFG) participates at the steps of translocation and ribosome recycling. Fusidicacid (FA) is a bacteriostatic antibiotic, which traps EFG on ribosomes, stalling them on mRNAs. How the bacterialsusceptibility to FA is determined, and which of the two functions of EFG (translocation or ribosome recycling) is morevulnerable, has remained debatable. The in vivo studies addressing these aspects of FA mediated inhibition of proteinsynthesis are lacking. Here, we used a system of Escherichia coli strains and their complementation/supplementation withthe plasmid borne copies of the inducible versions of EFG and ribosome recycling factor (RRF) genes. Additionally, weinvestigated FA sensitivity in a strain with increased proportion of stalled ribosomes. We show that the cells with highEFG/RRF (or low RRF/EFG) ratios are more susceptible to FA than those with low EFG/RRF (or high RRF/EFG) ratios. Our in vivoobservations are consistent with the recent in vitro reports of dependence of FA susceptibility on EFG/RRF ratios, and thenotion that an overriding target of FA is the translocation function of EFG. An applied outcome of our in vivo study is that FAmediated growth inhibition could be facilitated by depletion or inactivation of cellular RRF.

Keywords: fusidic acid; antibiotic; translocation; ribosome recycling; EFG; RRF

INTRODUCTION

The mechanism of action of antibiotics can broadly be dividedinto three classes: disruption of cell wall andmembrane biosyn-thesis, inhibition of protein synthesis and disruption of nu-cleic acid repair or metabolism (Gale 1963). Protein synthesis in-hibitors work by either directly binding to rRNA or by binding totranslation factors. Fusidic acid (FA) is a bacteriostatic steroid-like translation factor inhibitor that is used to treat infections byGram positive bacteria (Godtfredsen et al. 1962). It is widely ac-

cepted that FA functions by inhibiting the activity of elongationfactor G (EFG), which is crucial for translocation and ribosomerecycling. FA stalls translating ribosomes by decreasing the rateof dissociation of EFG-GDP from the ribosome (Willie et al. 1975).Recent studies have shown that FA binds after Pi release, andinhibits later steps of translocation and release of E-site boundtRNA (Belardinelli and Rodnina 2017).

In our earlier studies, we showed that excess productionof EFG in Escherichia coli, causes hypersensitivity to FA (Raoand Varshney 2001). More recently, in vitro studies from the

Received: 29 December 2017; Accepted: 23 May 2018C© FEMS 2018. All rights reserved. For permissions, please e-mail: [email protected]

1

Downloaded from https://academic.oup.com/femsle/article-abstract/365/13/fny133/5004850by Indian Institute of Science useron 23 August 2018

http://www.oxfordjournals.org

mailto:[email protected]

mailto:[email protected]

2 FEMS Microbiology Letters, 2018, Vol. 365, No. 13

Table 1. Description of plasmids and strains used in this study.

Plasmid/Strain Relevant features Source

pTrcEcoRRF E. coli frr cloned into pTrc99c. Rao and Varshney (2001)pTrcEcoEFG E. coli fusA cloned into pTrc99c. Rao and Varshney (2001)pTrc-ssrA E. coli ssrA cloned into pTrc99c Singh and Varshney (2004)TG1 supE hsd�5 thi �(lac-proAB) F‘ [traD36 proAB+ lacIq lacZ�M15] Sambrook (1989)LJ14 MC1061 containing frrts allele. Has very low levels of RRF. Janosi et al. (1998)TG1dnRRF A derivative of E. coli TG1 harbouring a ChloramphenicolR cassette in the

promoter region of frr gene, leading to decreased expression of RRFSingh and Varshney (2004)

TG1�ssrA::kan The ssrA deletion was moved to TG1 from E. coli AA7852 �ssrA::kan. Singh et al. (2008)

Ehrenberg group have demonstrated that a relative increase inEFG (with respect to ribosome recycling factor (RRF)), or decreasein RRF (with respect to EFG) result inmore unrecycled ribosomesand consequently, greater FA sensitivity (Borg et al. 2015). Forproductive ribosome recycling, RRF must bind to the ribosomebefore EFG. Only in the presence of RRF does the GTP hydrolysisactivity of EFG lead to the splitting of ribosomes. In the event thatEFG binds to the ribosome before RRF, EFG hydrolyzes GTP in afutilemannerwithout splitting ribosomes (Prabhakar et al. 2017).

Since EFG performs the dual functions of translocation andribosome recycling in protein synthesis, in vitro studies havebeen unable to agree on which of these functions is the majortarget of FA. According to studies from the Rodnina group, theantimicrobial activity of FA is almost entirely due to inhibitionof ribosome recycling (Savelsbergh, Rodnina and Wintermeyer2009). They observed that the inhibitory concentration of FA forribosome recycling was very close to MIC50 values observed invivo. Themost recent studies on themechanism of FA inhibitionwere performed by the Ehrenberg group using stopped-flowand quench-flow experiments (Borg, Pavlov and Ehrenberg2016) where it was noted that FA binds less efficiently to theFA-susceptible states of the ribosome in translocation thanin recycling. However, considering the fact that for every 400translocation events there is only one recycling event led theEhrenberg group to conclude that less than 10% of the growthinhibition by FA is due to inhibition of ribosome recycling andmore than 90% is due to inhibition of translocation (Borg, Pavlovand Ehrenberg 2016). Importantly, the in vivo contribution ofthe inhibition of translocation or ribosome recycling to theantimicrobial activity of FA has never been studied.

Here, we have carried out an investigation of the in vivo ef-fect of perturbations of the relative amounts of EFG and RRF onFA sensitivity. For this, we have used a system of genomic RRFmutants in E. coli and overexpressed EFG and RRF in these back-grounds. By studying the FA sensitivity in a strain with an in-creased proportion of stalled ribosomes, we have also attemptedto find out the contribution of the inhibition of ribosome recy-cling and translocation towards the sensitivity of E. coli to FA.

MATERIALS AND METHODS

Strains and plasmids

Strains and plasmids used in this study have been listed inTable 1. E. coli was grown in LB and LB agar plates (Difco). E.coli cells were made transformation competent by the Hana-han method and transformation was carried out by the heatshock method (Sambrook 1989). Unless mentioned otherwise,media were supplemented with ampicillin (Amp, 100 μg/mL),

kanamycin (Kan, 25 μg/mL) or tetracycline (Tet, 7.5 μg/mL) asrequired.

Preparation of cell-free extracts (for SDS PAGE andimmunoblotting)

Cells were harvested by centrifugation at 13 000 rpm (Kub-ota RA2724) for 1 min. The cell pellets were resuspended in400 μL TME buffer containing 25 mM Tris-HCl pH 8, 2 mM β-mercaptoethanol, 1 mM Na2EDTA and subjected to sonication.The conditions of sonication were as follows: amplitude 40%,pulse 2 s, on/off for 25 s 3 times with 1 min gap at each inter-val. Cells were harvested at 14 000 rpm for 20 min. Supernatantand pellet were obtained. The pellet was resuspended in 200 μLof TME buffer. The sonicator was manufactured by Sonics andMaterials Inc, Danbury, Connecticut, USA.

Immunoblot analysis

Requisite amounts of protein were loaded on a 12% or 15%SDS PAGE and transferred onto polyvinylidene difluoride mem-brane at 15 V for 1 h using the Bio-Rad semi-dry transfer ap-paratus. The membrane was soaked in blocking solution [5%skimmedmilk prepared in TBST (20mMTris-HCl, 0.9% NaCl and0.1% Tween-20)], kept at room temperature for 1 h and washedthrice with TBST buffer for 10 min each. Primary antibody wasadded and the blot was incubated overnight under rocking con-ditions at 4◦C followed by three washes of 10 min each withTBST. The blot was treated with anti-rabbit goat IgG-ALP con-jugate (GENEI) secondary antibody (1:3000) for 1 h under rockingconditions, washed thrice with TBST, and equilibrated with 50mL 0.1 M Tris-HCl buffer pH 9 for 10 min. The blot was devel-oped in darkness using 20 mL 0.1 M Tris-HCl buffer pH 9, 200μL 5 mg/mL 5-bromo-4-chloro-3-indolyl phosphate [prepared in100% dimethyl formamide (DMF)], 200 μL 30 mg/mL nitro bluetetrazolium (prepared in 70% DMF) and 80 μL 2 M MgCl2. Fordetection of RRF, rabbit polyclonal anti-RRF antiserum (1:5000)was used and for detection of EFG, rabbit polyclonal anti-EFGantiserum (1:5000).

RESULTS

Cellular levels of RRF and EFG in various E. coli strainsand the regulation of their relative ratios

The E. coli strain TG1dnRRF bears the frr1 mutation containingan insertion in the promoter region of the gene encoding RRF(Singh and Varshney 2004), leading to about three-fold decreasein the cellular levels of RRF [Fig. 1a]. Also, we exploited the E. coli


Ayyub et al. 3

Figure 1. Immunoblot analyses with anti-EcoRRF and anti-EcoEFG antibodies of cell extracts (with total protein content of 40 μg) from indicated strains to measure

expression of RRF and EFG (panel a, analysis of RRF in the strains TG1 and TG1dnRRF harbouring pTrc99c vector; panel b, analysis of RRF and EFG in strains TG1 and LJ14harbouring pTrc99c vector alone, or pTrc99c based expression construct for EcoRRF or EcoEFG). SDS-PAGE analyses of the cell extracts are also shown. Lanes marked as−/+ denote either no induction or induction of the plasmid borne copies of the RRF and EFG by addition of 1 mM IPTG in the strains harbouring the relevant expression

constructs. The white arrows in SDS-PAGE indicate the bands corresponding to EFG (top) or RRF (lower). The pixel values of the relevant bands were quantitated in theimmunoblots using Multi Gauge V2.3, and the ratios of RRF/EFG intensities are shown below the immunoblots.

LJ14 strain bearing an frrts allele (Janosi et al. 1998). Even at thepermissive temperature for its growth (30 ◦C), LJ14 has very low(barely detectable) levels of RRF (Fig. 1b, lanes 6, 9 and 10) pro-viding an extremely low RRF/EFG ratio as compared to that ofE. coli TG1, a strain wild type for RRF (Fig. 1b, lane 1) or the RRFcomplemented LJ14 strain (Fig. 1b, lanes 7 and 8). Introductionof pTrcEcoRRF or pTrcEcoEFG with leaky basal level expression ofRRF (lanes 2 and 7), EFG (lanes 4 and 9) or IPTG induced levels ofRRF (lanes 3 and 8) and EFG (lanes 5 and 10) allow us to regulatethe RRF/EFG ratios even further.

Increased relative levels of EFG over RRF render E. colihypersensitive to FA

To investigate the effect of FA and its dependence on the rel-ative levels of RRF/EFG, we carried out growth curve analy-ses. We observed that the strain TG1dnRRF that has only aslightly lowered ratio of RRF/EFG (as compared to the TG1control strain), is unaffected in its sensitivity to FA (as com-pared to the TG1 control strain) [Fig. 2c–f, compare curves 1(TG1/pTrc99c) and 4 (TG1dnRRF/pTrc99c)]. However, in both theTG1 and the TG1dnRRF strains, overexpression of EFG leads toinhibition of growth in the presence of FA [Fig. 2d–f, comparecurves 1 (TG1/pTrc99c) and 4 (TG1dnRRF/pTrc99c) with curves 3(TG1/pTrcEcoEFG) and 5 (TG1dnRRF/pTrcEcoEFG)]. This inhibitionof growth upon increase in the relative amount of EFG (with re-spect to RRF) is in agreement with the in vivo data from our lab(Rao and Varshney 2001) and the in vitro work from the Ehren-berg group (Borg et al. 2015). As per the expectations from thein vitro experiments, LJ14, which has a very low RRF/EFG ratioeven at its permissive temperature of 30◦C, is very sensitive toFA as compared to the RRF complemented strain [Fig. 3c–f, com-pare curves 1 (LJ14/pTrcEcoRRF) and 2 (LJ14/pTrc99c)]. EFG over-expression in the background of LJ14 (Fig. 1b, lanes 9 and 10) ren-

ders the strains hypersensitive to FA [Fig. 3d, compare curves 2(LJ14/pTrc99c) and 3 (LJ14/pTrcEcoEFG)].

Increased levels of stalled ribosomes lead tohypersensitivity to FA in E. coli

Occasionally, ribosomes stall on mRNAs prior to completionof the polypeptide chain. The tmRNA (encoded by ssrA) pos-sesses a tRNA-like domain and a short mRNA region encodinga short peptide (ANDENYALAA, in E. coli) followed by a termina-tion codon and is recruited to the A site of the stalled ribosome(Keiler, Waller and Sauer 1996). The peptidyltransferase activ-ity of the ribosome transfers the peptide from the P-site boundpeptidyl-tRNA to the alanine on the -CCA end of the tmRNA.The tRNA that was sequestered as peptidyl-tRNA in the stalledcomplex is freed, and the peptide extended at the C-terminalwith the tmRNA encoded sequence is released by the activityof release factors, and subjected to degradation by cellular pro-teases. The absence of ssrA leads to increased accumulation ofstalled ribosomes (Singh et al. 2008). Interestingly, FA resistancein Staphylococcus aureus is mediated by the FusB-type proteinsthat rescue stalled ribosomes formed by FA by binding to EFG onthe ribosome and releasing the stalled EFG-GDP complexes (Coxet al. 2012). Since rescuing stalled ribosomes is a method of alle-viating the toxicity of FA, we were interested in studying the FAresistance of strains lacking ssrA.

From our growth curve analyses, we see that the strain lack-ing tmRNA is hypersensitive to FA and this sensitivity is res-cued by plasmid-borne complementation with ssrA [Fig. 4 c andd, compare curves 3 (TG1�ssrA/pTrc99c) and 4 (TG1�ssrA/pTrc-ssrA)]. Therefore, increased dependence on ribosome recyclingin E. coli in the presence of FA affects bacterial growth. In fact,the absence of tmRNA would lead to an increase in stalled ribo-somes, which in turn would also decrease the pool of actively



Figure 2. Growth of the indicated strains of E. coli TG1 (TG1 and TG1dnRRF) in the presence of the indicated plasmids. Overnight cultures grown in LB containingampicillin were diluted 100-fold in the same medium without or with 1 mM IPTG, and their growth was monitored at 37◦C in the indicated concentrations of FA usinga Bioscreen C growth reader. Four replicates of all strains were studied. The lines join the mean, and the vertical bars denote the standard deviation.

translating ribosomes requiring EFG for translocation purpose,thereby indirectly raising the relative levels of EFG (with respectto the translating ribosomes).

DISCUSSION

The biochemical data on FA inhibition of translocation and re-cycling were used to design a study of the growth inhibitory ef-fect of FA on bacterial populations. We see that overexpressionof EFG in wild type and low RRF strains (TG1dnRRF, LJ14) leadsto increased sensitivity to FA (Figs. 2 and 3). Furthermore, LJ14 ismore sensitive to FA as compared to E. coliwild type for RRF. And,the LJ14 strain with EFG overexpression is hypersensitive to FA(Fig. 3). In a strain like LJ14, where the intrinsic levels of RRF inthe cells are low, EFG is more likely to bind to the ribosome be-fore RRF. This phenomenon is further exacerbated when EFG isoverexpressed in LJ14. Thismimics a translocation-like situationwhere EFG binds to the ribosome in the absence of RRF. There-fore,we conclude that FA inhibition of translocation provides thecentral basis of bacterial growth reduction. Our in vivo observa-tions are also in complete agreement with in vitro data from theEhrenberg group, where it was predicted that a decrease in thephysiological RRF/EFG ratio would lead to increased sensitivityto FA (Borg et al. 2015). As RRF levels in the cell are already high,

increasing the RRF levels further does not alter the physiologi-cal response to FA to any detectable levels [Fig. 2e and f comparecurves 1 (TG1/pTrc99c) and 2 (TG1/pTrcEcoRRF)]. Furthermore,the Ehrenberg group suggests that the KI value for translocationwas overestimated in an earlier study (Savelsbergh, Rodnina andWintermeyer 2009; Borg, Pavlov and Ehrenberg 2016), leading toa conclusion that the ribosome recycling function is the primarytarget of FA.

Our in vivo data also indicates that an E. coli strain lacking tm-RNA is hypersensitive to FA (Fig. 4). How do the cells deleted forssrA become hypersensitive to FA?We believe that a large popu-lation of stalled ribosomes (due to the lack of ssrA, which is themajor mechanism for recycling such ribosomes) leaves behindonly a smaller population of ribosomes active in protein synthe-sis, requiring EFG for the purposes of translocation. Thus, evenwith an unchanged level of overall EFG in the cell, such a phe-nomenon of ribosomal stalling could lead to an increase in therelative EFG levels in the cell (i.e. with respect to the translatingribosomes). Since an increase in EFG leads to increased sensitiv-ity to FA (as seen in Figs. 2 and 3), deletion of ssrA would ren-der the cell hypersensitive to FA. The Ehrenberg group has alsosuggested that while FA generally affects the translocation ac-tivity of EFG, inhibition of its recycling activity also contributesto the overall inhibition of protein synthesis by FA. However, theinhibition of the recycling activity of EFG may assume greater


Ayyub et al. 5

Figure 3. Growth of the indicated strains of E. coli LJ14 in the presence of the indicated plasmids. Overnight cultures grown in LB containing ampicillin were diluted100-fold in the same medium without or with 1 mM IPTG, and their growth was monitored at 30◦C in the indicated concentrations of FA using a Bioscreen C growth

reader. Four replicates of all strains were studied. The lines join the mean, and the vertical bars denote the standard deviation.

significance for smaller ORFs, where the ratio of translocationevents to recycling would decrease.

The tmRNA rescues ribosomes by the process of trans-translation when it is recruited to the empty A-site of stalledribosomes (Keiler, Waller and Sauer 1996). Therefore, in the ab-sence of tmRNA, there would be an increase in the population ofstalled ribosomeswith peptidyl-tRNA in the P-site and an emptyA-site (in non-rotated state). EF-G has been shown to bind to ri-bosomes in the rotated (pre-translocation) state (Prabhakar et al.2017). Even though the stalled post-translocation ribosomes inthe non-rotated state may not, at least in vivo, bind EFG withhigh enough affinity, given that FA can target multiple stagesof EFG binding to ribosome (Borg et al. 2015), they may still bea substrate for the drug mediated inhibition of the bacterialgrowth. Furthermore, the fact that some of the stalled ribosomescould also be a substrate for recycling by RRF and EFG (Heurgue-Hamard et al. 1998; Rao and Varshney 2001), the stalled ribosomewould also increase the events for recycling by RRF. Both of thesescenarios (i.e. increase in EFG/RRF ratio and increase in riboso-mal population requiring recycling) would lead to increased FAsensitivity.

Of late, there has been an emergence of spontaneous FA re-sistance by means of chromosomal fusA mutations, plasmid-mediated decreased cell membrane permeability and FusB-type

proteins (Dobie and Gray 2004; Cox et al. 2012). Therefore, in-creasing the potency of FA is the need of the hour. Since we haveshown that FA susceptibility increases with a decrease in theRRF/EFG ratio, the effectiveness of FA may be increased by us-ing it in conjunction with compounds that may lower the activelevels of RRF. Development of such supplementary compoundsmay prolong the usability of FA as an effective antibiotic.

ACKNOWLEDGEMENT

We thank our laboratory colleagues for their suggestions on themanuscript.

FUNDING

This work was supported by the Department of Biotechnology(DBT), Ministry of Science and Technology; Department of Sci-ence and Technology (DST), Ministry of Science and Technology;DST J.C. Bose Fellowship (to U.V.). The authors also acknowledgethe DBT-IISc partnership programme, University Grants Com-mission, New Delhi for the Centre of Advanced Studies, and theDST-FIST level II infrastructure supports to carry out this work.

Conflict of interest. None declared.



Figure 4. Growth of the indicated strains of E. coli TG1 (TG1 and TG1�ssrA) in the presence of the indicated plasmids. Overnight cultures grown in LB containing

ampicillin were diluted 100-fold in the same medium without or with 1 mM IPTG, and their growth was monitored at 37◦C in the indicated concentrations of FA usinga Bioscreen C growth reader. Four replicates of all strains were studied. The lines join the mean, and the vertical bars denote the standard deviation.

REFERENCES

Belardinelli R, Rodnina MV. Effect of fusidic acid on the kineticsof molecular motions during EF-G-induced translocation onthe ribosome. Sci Rep 2017;7:10536.

Borg A, Pavlov M, Ehrenberg M. Mechanism of fusidic acid inhi-bition of RRF- and EF-G-dependent splitting of the bacterialpost-termination ribosome. Nucleic Acids Res 2016;44:3264–75.

Borg A, HolmM, Shiroyama I et al. Fusidic acid targets elongationfactor G in several stages of translocation on the bacterialribosome. J Biol Chem 2015;290:3440–54.

Cox G, Thompson GS, Jenkins HT et al. Ribosome clearance byFusB-type proteins mediates resistance to the antibiotic fu-sidic acid. Proc Natl Acad Sci 2012;109:2102–7.

Dobie D, Gray J. Fusidic acid resistance in Staphylococcus aureus.Arch Dis Child 2004;89:74–7.

Gale EF. Mechanisms of antibiotic action. Pharmacol Rev1963;15:481–530.

Godtfredsen WO, Jahnsen S, Lorck H et al. Fusidic acid: a newantibiotic. Nature 1962;193:987.

Heurgue-Hamard V, Karimi R, Mora L et al. Ribosome re-lease factor RF4 and termination factor RF3 are involved indissociation of peptidyl-tRNA from the ribosome. EMBO J1998;17:808–16.

Janosi L, Mottagui-Tabar S, Isaksson LA et al. Evidence for invivoribosome recycling, the fourth step in protein biosynthesis.EMBO J 1998;17:1141–51.

Keiler KC, Waller PR, Sauer RT. Role of a peptide tagging systemin degradation of proteins synthesized from damaged mes-senger RNA. Science 1996;271:990–3.

Prabhakar A, Capece MC, Petrov A et al. Post-termination ribo-some intermediate acts as the gateway to ribosome recy-cling. Cell Rep 2017;20:161–72.

Rao AR, Varshney U. Specific interaction between the ribosomerecycling factor and the elongation factor G from Mycobac-terium tuberculosis mediates peptidyl-tRNA release and ri-bosome recycling in Escherichia coli. EMBO J 2001;20:2977–86.

Sambrook J. Molecular Cloning: A Laboratory Manual. NY: ColdSpring Harbor,1989.

Savelsbergh A, Rodnina MV, Wintermeyer W. Distinct functionsof elongation factor G in ribosome recycling and transloca-tion. RNA 2009;15:772–80.

Singh NS, Varshney U. A physiological connectionbetween tmRNA and peptidyl-tRNA hydrolase func-tions in Escherichia coli. Nucleic Acids Res 2004;32:6028–37.

Singh NS, Ahmad R, Sangeetha R et al. Recycling of ribosomalcomplexes stalled at the step of elongation in Escherichia coli.J Mol Biol 2008;380:451–64.

Willie GR, Richman N, Godtfredsen WP et al. Translocation. XV.Characteristics of and structural requirements for the inter-action of 24,25-dihydrofusidic acidwith ribosome.elongationfactor G complexes. Biochemistry 1975;14:1713–8.


Documents

RESEARCH LETTER–Physiology&Biochemistry Antimicrobial ...mcbl.iisc.ac.in/Varshney/docs/2018/fusidic_acid.pdf · pTrcEcoEFG E. coli fusA clonedintopTrc99c. RaoandVarshney(2001) pTrc-ssrA